Citation
Comparative processing of cyanogenic glycosides and a novel cyanide inhibitory enzyme in Heliconius butterflies (Lepidoptera: Nymphalidae: Heliconiinae)

Material Information

Title:
Comparative processing of cyanogenic glycosides and a novel cyanide inhibitory enzyme in Heliconius butterflies (Lepidoptera: Nymphalidae: Heliconiinae)
Creator:
Hay-Roe, Mirian Medina
Publication Date:
Language:
English
Physical Description:
xvi, 122 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Butterflies -- Host plants ( lcsh )
Butterflies -- Metabolism ( lcsh )
Dissertations, Academic -- Entomology and Nematology -- UF
Entomology and Nematology thesis, Ph. D
Glycosides ( jstor )
Butterflies ( jstor )
Cyanides ( jstor )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 2004.
Bibliography:
Includes bibliographical references.
General Note:
Printout.
General Note:
Vita.
Statement of Responsibility:
by Mirian Medina Hay-Roe.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
022437028 ( ALEPH )
880637413 ( OCLC )
880438633 ( OCLC )
Classification:
LD1780 2004 .H4137 ( lcc )

Downloads

This item has the following downloads:

AA00029850_00001 ( .pdf )

comparativeproce00hayr_0121.txt

comparativeproce00hayr_0081.txt

comparativeproce00hayr_0108.txt

comparativeproce00hayr_0004.txt

AA00029850_00001_0025.txt

comparativeproce00hayr_0019.txt

comparativeproce00hayr_0064.txt

comparativeproce00hayr_0006.txt

comparativeproce00hayr_0098.txt

AA00029850_00001_0068.txt

AA00029850_00001_0139.txt

comparativeproce00hayr_0054.txt

AA00029850_00001_0093.txt

AA00029850_00001_0051.txt

AA00029850_00001_0110.txt

AA00029850_00001_0123.txt

AA00029850_00001_0001.txt

comparativeproce00hayr_0106.txt

comparativeproce00hayr_0109.txt

comparativeproce00hayr_0003.txt

AA00029850_00001_0039.txt

AA00029850_00001_0022.txt

comparativeproce00hayr_0067.txt

comparativeproce00hayr_0077.txt

comparativeproce00hayr_0115.txt

AA00029850_00001_0057.txt

comparativeproce00hayr_0035.txt

AA00029850_00001_0134.txt

AA00029850_00001_0109.txt

AA00029850_00001_0016.txt

comparativeproce00hayr_0025.txt

AA00029850_00001_0120.txt

AA00029850_00001_0037.txt

comparativeproce00hayr_0058.txt

AA00029850_00001_0131.txt

AA00029850_00001_0045.txt

AA00029850_00001_0066.txt

AA00029850_00001_0033.txt

AA00029850_00001_0002.txt

comparativeproce00hayr_0133.txt

AA00029850_00001_0003.txt

comparativeproce00hayr_0032.txt

AA00029850_00001_0105.txt

comparativeproce00hayr_0042.txt

comparativeproce00hayr_0061.txt

AA00029850_00001_0050.txt

AA00029850_00001_0020.txt

comparativeproce00hayr_0122.txt

comparativeproce00hayr_0022.txt

comparativeproce00hayr_0128.txt

comparativeproce00hayr_0034.txt

comparativeproce00hayr_0038.txt

comparativeproce00hayr_0089.txt

comparativeproce00hayr_0113.txt

AA00029850_00001_0129.txt

comparativeproce00hayr_0023.txt

AA00029850_00001_0047.txt

AA00029850_00001_0090.txt

AA00029850_00001_0064.txt

AA00029850_00001_0082.txt

AA00029850_00001_0007.txt

AA00029850_00001_0015.txt

comparativeproce00hayr_0070.txt

comparativeproce00hayr_0005.txt

AA00029850_00001_0036.txt

comparativeproce00hayr_0060.txt

comparativeproce00hayr_0031.txt

AA00029850_00001_0034.txt

comparativeproce00hayr_0017.txt

comparativeproce00hayr_0021.txt

AA00029850_00001_0071.txt

comparativeproce00hayr_0049.txt

AA00029850_00001_0083.txt

AA00029850_00001_0073.txt

AA00029850_00001_0133.txt

comparativeproce00hayr_0062.txt

AA00029850_00001_0094.txt

comparativeproce00hayr_0002.txt

AA00029850_00001_0119.txt

AA00029850_00001_0108.txt

AA00029850_00001_0132.txt

comparativeproce00hayr_0065.txt

AA00029850_00001_0092.txt

AA00029850_00001_0046.txt

comparativeproce00hayr_0063.txt

AA00029850_00001_0096.txt

comparativeproce00hayr_0071.txt

comparativeproce00hayr_0057.txt

AA00029850_00001_0124.txt

AA00029850_00001_0044.txt

AA00029850_00001_0026.txt

AA00029850_00001_0077.txt

AA00029850_00001_0038.txt

AA00029850_00001_0135.txt

comparativeproce00hayr_0126.txt

comparativeproce00hayr_0110.txt

comparativeproce00hayr_0030.txt

comparativeproce00hayr_0029.txt

comparativeproce00hayr_0130.txt

comparativeproce00hayr_0097.txt

AA00029850_00001_0080.txt

AA00029850_00001_0028.txt

AA00029850_00001_0100.txt

AA00029850_00001_0055.txt

comparativeproce00hayr_0024.txt

AA00029850_00001_0104.txt

comparativeproce00hayr_0102.txt

comparativeproce00hayr_0056.txt

comparativeproce00hayr_0059.txt

comparativeproce00hayr_0132.txt

comparativeproce00hayr_0073.txt

AA00029850_00001_0041.txt

AA00029850_00001_0091.txt

AA00029850_00001_0014.txt

comparativeproce00hayr_0127.txt

AA00029850_00001_0084.txt

AA00029850_00001_0121.txt

comparativeproce00hayr_0100.txt

comparativeproce00hayr_0048.txt

comparativeproce00hayr_0015.txt

comparativeproce00hayr_0033.txt

AA00029850_00001_0070.txt

AA00029850_00001_0107.txt

comparativeproce00hayr_0085.txt

AA00029850_00001_0060.txt

comparativeproce00hayr_0093.txt

AA00029850_00001_0122.txt

AA00029850_00001_0058.txt

AA00029850_00001_0128.txt

comparativeproce00hayr_0137.txt

comparativeproce00hayr_0014.txt

comparativeproce00hayr_0119.txt

comparativeproce00hayr_0001.txt

AA00029850_00001_0130.txt

AA00029850_00001_0081.txt

comparativeproce00hayr_0007.txt

comparativeproce00hayr_0016.txt

AA00029850_00001_0052.txt

AA00029850_00001_0103.txt

AA00029850_00001_0011.txt

comparativeproce00hayr_0139.txt

comparativeproce00hayr_0076.txt

comparativeproce00hayr_0037.txt

comparativeproce00hayr_0105.txt

AA00029850_00001_0013.txt

comparativeproce00hayr_0129.txt

AA00029850_00001_0029.txt

comparativeproce00hayr_0011.txt

comparativeproce00hayr_0052.txt

AA00029850_00001_0017.txt

AA00029850_00001_0111.txt

AA00029850_00001_0076.txt

AA00029850_00001_0089.txt

comparativeproce00hayr_0046.txt

comparativeproce00hayr_0036.txt

AA00029850_00001_0049.txt

AA00029850_00001_0074.txt

AA00029850_00001_0112.txt

AA00029850_00001_0063.txt

comparativeproce00hayr_0072.txt

comparativeproce00hayr_0111.txt

comparativeproce00hayr_0103.txt

AA00029850_00001_0085.txt

comparativeproce00hayr_0078.txt

AA00029850_00001_0098.txt

AA00029850_00001_0030.txt

AA00029850_00001_0042.txt

comparativeproce00hayr_0124.txt

comparativeproce00hayr_0092.txt

AA00029850_00001_pdf.txt

comparativeproce00hayr_0000.txt

AA00029850_00001_0067.txt

AA00029850_00001_0023.txt

comparativeproce00hayr_0094.txt

AA00029850_00001_0117.txt

AA00029850_00001_0137.txt

AA00029850_00001_0086.txt

AA00029850_00001_0075.txt

comparativeproce00hayr_0020.txt

AA00029850_00001_0053.txt

comparativeproce00hayr_0091.txt

comparativeproce00hayr_0086.txt

comparativeproce00hayr_0051.txt

comparativeproce00hayr_0045.txt

AA00029850_00001_0115.txt

comparativeproce00hayr_0075.txt

comparativeproce00hayr_0117.txt

AA00029850_00001_0069.txt

comparativeproce00hayr_0041.txt

comparativeproce00hayr_0040.txt

AA00029850_00001_0138.txt

AA00029850_00001_0012.txt

AA00029850_00001_0106.txt

comparativeproce00hayr_0131.txt

comparativeproce00hayr_0055.txt

comparativeproce00hayr_0090.txt

comparativeproce00hayr_0008.txt

AA00029850_00001_0088.txt

comparativeproce00hayr_0112.txt

comparativeproce00hayr_0028.txt

comparativeproce00hayr_0114.txt

comparativeproce00hayr_0134.txt

comparativeproce00hayr_0095.txt

comparativeproce00hayr_0125.txt

comparativeproce00hayr_0053.txt

comparativeproce00hayr_0082.txt

AA00029850_00001_0126.txt

comparativeproce00hayr_0047.txt

AA00029850_00001_0114.txt

AA00029850_00001_0035.txt

AA00029850_00001_0056.txt

AA00029850_00001_0078.txt

AA00029850_00001_0010.txt

comparativeproce00hayr_0018.txt

AA00029850_00001_0140.txt

comparativeproce00hayr_0107.txt

comparativeproce00hayr_0044.txt

AA00029850_00001_0125.txt

comparativeproce00hayr_0039.txt

AA00029850_00001_0099.txt

AA00029850_00001_0095.txt

comparativeproce00hayr_0104.txt

AA00029850_00001_0087.txt

comparativeproce00hayr_0074.txt

AA00029850_00001_0127.txt

AA00029850_00001_0065.txt

comparativeproce00hayr_0120.txt

AA00029850_00001_0101.txt

AA00029850_00001_0059.txt

comparativeproce00hayr_0087.txt

AA00029850_00001_0048.txt

comparativeproce00hayr_0027.txt

comparativeproce00hayr_0068.txt

comparativeproce00hayr_0123.txt

AA00029850_00001_0018.txt

comparativeproce00hayr_0099.txt

AA00029850_00001_0118.txt

AA00029850_00001_0031.txt

AA00029850_00001_0032.txt

comparativeproce00hayr_0010.txt

comparativeproce00hayr_0080.txt

comparativeproce00hayr_0050.txt

AA00029850_00001_0061.txt

AA00029850_00001_0024.txt

comparativeproce00hayr_0079.txt

AA00029850_00001_0079.txt

AA00029850_00001_0072.txt

EY32K9BI9_H9HZJR_xml.txt

AA00029850_00001_0040.txt

comparativeproce00hayr_0069.txt

comparativeproce00hayr_0012.txt

comparativeproce00hayr_0118.txt

comparativeproce00hayr_0009.txt

comparativeproce00hayr_0096.txt

comparativeproce00hayr_0135.txt

AA00029850_00001_0054.txt

AA00029850_00001_0062.txt

comparativeproce00hayr_0066.txt

AA00029850_00001_0043.txt

AA00029850_00001_0097.txt

comparativeproce00hayr_0101.txt

comparativeproce00hayr_0116.txt

comparativeproce00hayr_0043.txt

AA00029850_00001_0005.txt

AA00029850_00001_0113.txt

comparativeproce00hayr_0084.txt

comparativeproce00hayr_0088.txt

AA00029850_00001_0027.txt

AA00029850_00001_0004.txt

AA00029850_00001_0008.txt

AA00029850_00001_0116.txt

comparativeproce00hayr_0026.txt

AA00029850_00001_0021.txt

AA00029850_00001_0019.txt

AA00029850_00001_0006.txt

AA00029850_00001_0136.txt

AA00029850_00001_0009.txt

AA00029850_00001_0102.txt

comparativeproce00hayr_0136.txt

comparativeproce00hayr_0083.txt

comparativeproce00hayr_0138.txt

comparativeproce00hayr_0013.txt


Full Text











COMPARATIVE PROCESSING OF CYANOGENIC GLYCOSIDES AND A NOVEL
CYANIDE INHIBITORY ENZYME IN HELICONIUS BUTTERFLIES
(LEPIDOPTERA: NYMPHALIDAE: HELICONIINAE)

















By

MIRIAN MEDINA HAY-ROE












A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2004
































Copyright 2004

by

Mirian Medina Hay-Roe


































To all the people that love science and helped me through the development of this research.














ACKNOWLEDGMENTS

I would like to thank all of the members of my committee, especially my chairman, Dr. T. C. Emmel, for financial support for a field trip to Per' and Panama, for providing laboratory space and two greenhouses where I kept plants and butterfly colonies,and for his friendship and trust. Dr. J. L. Nation was kind enough to guide me throughout many of the experiments performed for this study. I would also like to thank him for his laboratory space, which became my "home" for many years. Thanks go to Dr. H. J. MacAuslane, for allowing me to use an entire compartment of her ultrafreezer. I am very grateful for her friendship and her support. Thanks go to Dr. D. A. Jones and Dr. F. Slansky for providing me with enzymes, and some cyanogens used as standards. Dr. Jones kindly allowed me to use his extensive library containing the most important literature on passionvines.

I extend my special thanks to several people who were not members of my

committee, but who provided equipment for the chemical analyses and whose comments were crucial in completing my dissertation. My deepest gratitude is expressed to Dr. J. Tlumlinson and Dr. P. Teal from the Center for Medical, Agricultural and Veterinary Entomology (CMAVE), U. S. Department of Agriculture (USDA), without whose help my dissertation would not have been possible. I thank them for their trust and support. Special thanks as well are due to Dr. C. Lait, with whom I spent a summer through a USDA Student Intership Program during which I had the opportunity to practice various




iv








techniques for protein purification. I thank Dr. S. Yu and Dr. D. Boucias for discussing enzymology work with me and for the use of their laboratory equipment.

Special thanks to the National Science Foundation (NSF) and the National High 'Magnetic Field Laboratory External User Program via Dr. P. Teal, CMAVE, USDA for .the use of the Nuclear Magnetic Resonance (NMR) at the Advanced Magnetic Resonance Imaging and Spectroscopy (AMRIS) facility in the McKnight Brain Institute of the University of Florida. I am very grateful to Mr. Jim Rocca from for his help in the chemical elucidation of my compounds through (NMR) analysis. I thank Dr. J. Johnson, who runs the HPLC/Mass Spectroscopy unit in the Chemistry Department. His reports on the analysis performed are the best! I also appreciate his friendship and long discussions on the HPLC/MS analysis. Dr. I. Ghiviriva kindly ran the NMR analysis to corroborate the chemical composition of two plants: P. auriculata and P. coriacea. Dr. H. Alborn from the USDA generously ran certain samples in the mass spectrometry and also gave me advice on some of the techniques I used. Special thanks are due to Dr. B. Torto, for his friendship.

I owe special thanks to Dr. J. Jarozewsky from the Royal Danish School of Pharmacy in Copenhagen, Denmark, for sending me several milligrams of purified gynocardin and tetraphyllin B.

Thanks are due to Dr. R. Ferl, Dr. P. Sehnke, and B. Laughter from the

Horticultural Science Department, who allowed me to use the lab ultracentrifuge and electrophoresis equipment. Beth taught me all of the tricks for getting the best electrophoretic gel. Working with her was very enjoyable.





V








My former M. A. supervisor and professor, Dr. L. Gilbert from the University of Texas at Austin, allowed me to take all the cuttings I needed from his great collection of identified Passiflora. My deepest gratitude is also given to Mr. R. Boender from Butterfly World for providing me with some Passiflora cuttings, adult plants, and butterflies, and for financial support at key times. I also thank Dr. J. McDougall for the identification of the Passiflora plants.

I extend thanks to Jacob Olander in Quito, Ecuador, for providing me with H. e. cyrbia and H. himera.

The Institute of Natural Resources (INRENA) in Lima, Peru, provided me with

collection and export permits for two seasons of work in Tingo Maria, Huanuco, where I collected Heliconius eratofavorinus. Thanks are also due to the Smithsonian Tropical Research Institute in Panama for providing me with permits to collect H. e. demophon.

I am very grateful for the financial support provided by the Delores A. Auzenne Scholarship, the minority programs at the University of Florida, the Mulrennan Scholarship, and for the various semesters on teaching assistantship, which I received in the Entomology and Nematology Department.

I also extend heartfelt thanks to Christine Eliazar, who was always there to listen and give me comfort during stressful times, and to Arif, Steve, David, Alfredo, Matthew and Chris for helping me with the greenhouse maintenance chores. Special thanks are due to Alfredo Rios for scientific discussions on butterfly issues, and for raising caterpillars to maintain the greenhouse colonies. Steve and Jim Schlachta and Jerry Wenzel were always there when I had electrical problems in the greenhouses and environmental chambers and fixed them quickly to help prevent any harm to my cultures.




vi









And finally, my thanks go to my husband, Keith, for his love and willingness in

helping me with some of this research and especially for being the outstanding editor that he is, and for being the father of my daughter, Kylie, who shares my love for butterflies and enjoys being my best helper.











































vii
















TABLE OF CONTENTS
page

ACKNOWLEDGMENTS ........................................ ................................................. iv

L IST O F T A B LE S................................................. ....................................................... xi

LIST OF FIGURES ........................................ xiii

A B ST R A C T ..................................................... ............................................................ xv

CHAPTER

1 IN TR O D U C T IO N ................................................................................................... 1

2 LIFE HISTORIES AND NUTRITIONAL EFFECTS ON SIZE OF HELICONIUS
ERA TO FEEDING ON DIFFERENT PASSIFLORA HOST PLANTS ...................... 6

M aterials and M ethods ....................................................................................... 8
Museum Specimens .......................................................... .................... 8
Living Insects and Host Plants ........................................................................... 9
Feeding Experiments................................................. ..................................9
Life H istory Studies ................................................................................. 10
Egg m easurem ents .............................................................. ........... 11
Larval developm ent .................................................................................. 11
Pupae and adult offspring measurement ........................................ ..... 12
Statistical Analysis ..............................................................12
R esults ... ................................................................................................ ...... 13
Analysis of Specimens from the Museum Collection.............................. 13
Life History Studies in H. e. favorinus and H. e. cyrbia................. 13
Eggs............. ............. 13
L arv ae .......................................................................................................... 14
P upae ................................ ........ ......................................................... 15
A dult .......................... ..... ...................................... 15
M ortality .............................................................................................................. 16
Discussion and Conclusions ............................................................................. 16

3 CYANOGENESIS IN HELICONIUS ERATO AND PASSIFLORA HOST
P L A N T S ................................................................................. ..............................29

Occurrence of Cyanogenesis in Plants ......................................... .......29
Occurrence of Cyanogenesis in Lepidoptera...........................................................30


viii









Cyanogenic Glycosides as a Defensive Mechanism ....................................... 31
M aterials and M ethods ..................... .............. ..... ................................... 33
Plant Growth and Insect Rearing .................................................33
Experimental Design and Treatments .............................................34
B utterflies ........................................... .................................................... 34
Temporal cyanide quantification of Passiflora plants .............................. 34
Quantitative Determination of Cyanide ........................................................... 35
Colorimetric analysis ...................................................... 35
Standard curve .......................................................... 36
Predator-Prey Interaction .......................................................37
Statistical A nalysis .......................................... .... .......................................37
Results ..................................................................................... 37
Temporal Variation of Cyanide Release in Passiflora Plants.......................... 37
Cyanide Quantification in Heliconius erato Fed Their Natural Host Plants ......38
H. e. cyrbia fed on P. rubra and P. punctaa ...................................... 38
H. e. ftavorinus fed P. trifasciata......................................39
H. e. demophon fed on natural and alternative hosts plants......................... 39
Cyanide Correlation Between Heliconius erato Fed Natural Host Plants..........40 Predator-Prey Interaction ....................................................... 40
Discussion and Conclusions ...................................................... ................... 40
Cyanide Concentrations in Heliconius erato fed Passiflora host plants ..........40 Temporal Variation in Cyanide in Passiflora Plants ..................................... 43
Implications of Cyanide Concentration for Palatability........................... 44
Implications of Cyanide Concentration for Mimicry ..................................... 44

4 CYANOGENIC GLYCOSIDES IN PASSIFLORA PLANTS AND
B U T TER FL IE S .................................................................................................... 52

Mechanisms Used by Insects to Process Toxic Compounds from Plants ........... 53
Cyanogenic Glycosides in the Family Passifloraceae ...................................... 54
Cyanogenic Glycosides in Butterflies ....................... .............. 55
Synthesis D e N ovo .......................... ..............................................................55
Sequestration of Cyanogenic Glycosides ..................................... 56
E n zym es ....................... ... .................. .... ... ................. ...... ........................5 7
Materials and Methods ..............................................................58
Butterflies and Plants ....................................................... .......... ..........58
Plant and Butterfly Cyanogen Extraction ...................................... ....... 59
P-Glucosidase Preparations ................................ ................ 59
Purification of Cyanogenic Glycoside from Plants and Butterflies ................. 60
Purification by liquid chromatography ...................................... .....60
Purification by thin layer chromatography ..................................... ...61
Purification by high performance liquid chromatography (HPLC) .............62
Identification of Cyanogenic Glycosides from Passiflora Plants and
B utterflies .................................................................................................. 62
Nuclear magnetic resonance (NMR)............................ 62
HPLC/ (+) ESI-MS experimental ...................... ......64
Amino Acid Analysis....................................... ...............65


ix









R esu lts ........................................................................................................................6 6
Cyanogenic Glycosides in Passiflora Plants......................................66
Identification of cyclopentanoid cyanogenic in Passiflora rubra ............66
Identification of cyclopententanoid cyanogenic in Passiflora trifasciata ... 67 Cyanogen profile of Passiflora auriculata ............................................67
Cyanogen profile of Passiflora biflora ......................................................... 67
Cyanogen profile of Passiflora punctata .................................................. 68
Cyanogenic Glycosides of Butterflies Fed Natural Host Plants ......................... 68
D e novo synthesis .................................................................................. ...... 68
Sequestration of cyanogenic glycosides ..............................................69
Amino Acids Analysis in Butterfly Extracts....................................... 69
T hiol A nalysis ................................................... ............................................ 70
D iscussion and Conclusions .................................................................................... 70
Sequestration, De Novo Synthesis and Metabolism of Cyanogenic
G lycosides ....................................................................................................... 70
Amino Acids and their Importance For Heliconius Biology ...........................72

5 STUDIES OF CYANIDE INHIBITION BY P-GLUCOSIDASE FROM
IELICONIUS LARVAL GUTS ..................................................................... 92

M aterials and M ethods ......................................................................................... 94
Study O rganism .................................................. ...........................................94
Insect R earing Procedure ................................................................... ................. 95
Tissue Preparation .............................................................95
Protein Purification ........................................................................................... 95
Enzymatic activity of P-glucosidase ...................................... .....96
Protein concentration measurements .........................................................97
In vitro Cyanide Quantification of P-Glycosidase from H. himera Larval
Midgut and Various Cyanogenic Glycosides as Substrates .........................97
Sodium Dodecyl Sulfate (Sds)-Polyacrylamide Gel Electrophoresis (PAGE)...97 Protein Identification................................................. ............. ..............98
Results ................................................................................. 98
In vitro Cyanide Quantification of -Glycosidase from H. himera Larval
Midgut and Various Cyanogenic Glycosides as Substrates .........................98
3-glucosidase Activity in the Midgut of H. himera ..................................... 98
Protein Purification and Protein Identification ................................. ...99
Discussion and Conclusions .............................................................................. 99

6 C O N C LU SIO N S ..................................................................................................... 106

LIST O F REFEREN CES.............................................................................................. 108

BIOGRAPH ICAL SKETCH .............................................. .................................... 122






x















LIST OF TABLES

Table page

2-1. Average wing length and standard error of geographical races of Heliconius erato
from museum specimens in different regions .........................20

2-2. Repeated measurements ANOVA of head capsule size with race, host and family
as factors. ................................................................. ...........................................20

2-3. Repeated measurements ANOVA of larval developmental time with race and host
as factors. ........................................ .... ................. ..................... ......................2 1

2-4. Sexual differences in larval developmental time (LDT), growth rate (GR) and
wing length (WL) in H. e. favorinus and H. e. cyrbia feeding on five different
host plants. .......................................................... ................................................2 1

2-5. Mortality in different stages of the life history of H. e. favorinus and H. e. cyrbia
feeding on natural and suitable host plants .............................................22

3-1. Heliconius erato fed natural and alternative Passiflora host plant and sample size
used in the study................................................................................................... 46

3-2. Passiflora plants tested for cyanogenesis in this study. All plants are used by
Heliconius erato in different geographical areas ...................................... ...46

4-1. Qualitative analysis of the extracts of Passiflora plants. Number 0=no cyanide
release, l=light blue, slightly cyanogenic, 2=blue, cyanogen and 3=dark blue,
very cyanogenic. .................................................... .............................................75

4-2. 1'H NMR(") spectrum of a cyanogen, identified as passicapsin, from P. rubra in
CD30D. Chemical shifts (ppm) and coupling constants (Hz). Degree of splitting
is identified by: (s) single, (d) doublet, (q) quartet, (m) multiplet ........................ 76

4-3. "1C NMR'a) spectrum of a cyanogen, identified as passicapsin, from P. rubra in
CD30D.......................... ......................... ..................................... 77

4-4. H1- NMR a) spectrum of a cyanogen, identified as passitrifasciatin, from
tritasciata in CD30D. Chemical shifts (ppm) and coupling constants (Hz).
Degree of splitting is identified by: (d) doublet, (t) triplet, (m) multiplet. ...........78




xi









4-5. '3C NMR(a) spectrum of a cyanogen, identified as passitrifasciatin from P.
trifiasciata in CD30D................................................ 79

4-6. Qualitative analysis of cyanide release of whole body extracts tested with different
(3-glucosidases by Feigl-Anger strips monitored over time in H. e. favorinus and
H. e. cyrbia. Numbers indicate, 0=no cyanide release, 1=light blue, slightly
cyanogenic, 2=blue, cyanogen and 3=dark blue, very cyanogenic......................... 80

4-7. Peak areas counts of cyanogenic glycosides detected by HPLC/MS in H. erato and
H him era butterflies. ...................................................................... ................... 81

4-8. Percent of cyanogenic glycosides detected by HPLC/MS in H. erato and H. himera
butterflies. ......................................................... ................................................. 81

4-9. 'H NMR(a) spectra of cyanogen, epivolkenin, in an extract from female H. e. cyrbia
fed on P. auriculata in CD30D at 27 C. Chemical shifts (ppm), coupling
constants (Hz) and multiplicities. Multiplicity is identified by: (d) doublet, (m)
m ultiplet. ................................................................................................................. 82

4-10. Peak areas of measured free amino acids in butterfly samples .............................83

4-11. Percent total peak areas of measured free amino acids in butterfly samples........... 84

5-1. Experimental treatments performed with 13-glycosidase .................................... 102

5-2. Cyanide concentration when larval f-glycosidase from Heliconius himera were
tested against different cyanogenic substrates. .................................... 102
























xii










LIST OF FIGURES

Figure page

2-1. Life history of Heliconius butterflies .............................................................. 23

2-2. Museum specimens of H. erato from the eastern (left column) and western sides
(right column) of the Andes that were included in the analysis. (A) H. e. phyllis,
(B) H. e. favorinus, (C) H. e. magnificus, (D) H. e.colombina, (E) hydara,
(F) H. e. cyrbia, (G) H. e. petiveranus ....................... ........... 24

2-3. Study organisms. (A) H. e. favorinus from Tingo Maria, Huanuco, Peri and
(B) H. e. cyrbia from Pichincha, Ecuador. ....................................... ......25

2-4. Growth trajectories of (A) I. e. favorinus and (B) H. e. cyrbia feeding on five
different host plants. Each point represents the mean value of larvae from the
first to fifth instar. Bars on each point indicate standard error. ...........................26

2-5. Growth trajectories of H. e. favorinus and H. e. cyrbia fed their natural host plants
(P. trifusciata and P. rubra, respectively). Each point represents the mean value
of larvae from the first instar to the fifth. Bars on each point indicate standard
error... ............ .................................................................................. . .... 27

2-6. Pupal weight of H. e. favorinus (A) and H. e. cyrbia (B) fed on natural and
alternative host plants. Different letters indicate significant differences between
means (Tukey HSD-test). Bars in each point indicate standard error .................. 28

2-7. Average wing length of H. e. favorinus and H. e. cyrbia fed natural and alternative
host plants. Different letters indicate significant differences between means
(Tukey HSD-test). Bar in each point indicate standard error. .............................28

3-1. General process of enzymatic hydrolysis of cyanogenic glycosides. .....................47

3-2. Cyanide quantification of Passiflora plants in the (A) summer and (B) fall seasons.
Bars indicate standard error ....................................................... 48

3-3. Cyanide quantification of male and female H. e. cyrbia fed P. rubra. Different
letters indicate significant differences between means (Tukey HSD-test). Bars
indicate standard error .............................................................................................. 49

3-4. Cyanide quantification of male and female H. e. cyrbia fed P. punctata. Different
letters indicate significant differences between means (Tukey HSD-test). Bars
indicate standard error ............................................................................................ .. 49

3-5. Cyanide quantification of male and female H. e. favorinus fed P. trifasciata.
Different letters indicate significant differences between means (Tukey HSD-test).
Bars indicate standard error .......................................................50


xiii








3-6. Cyanide quantification of male and female H. e. demophon fed natural (P. biflora)
and alternative host plants. Bars indicate standard error. ....................................50

3-7. Regression of the mean cyanide concentration of H. erato over the mean cyanide
concentration of the natural Passiflora host plants upon which they fed. (The
means were used in the figure to enhance the clarity of presentation.) ................ 51

4-1. Cyanogenic glycosides in Passifloraceae plants (A) Type I, (B) Type II.
(C) Type III and (D) Type IV (Jaroszewski et al. 2002)............................... 85

4-2. Heliconius used in this study (A) H. e. cyrbia, (B) H. e. favorinus, and (C) H.
him era ....................................................................................... ......... 86

4-3. Passiflora plants used in this study (A) P. auriculata, (B) P. punctata, (C) P. rubra.
(D ) P. biflora. (E) P. trifasciata ........................................................................ ... 87

4-4. Proton/Carbon heteronuclear multiple quantum coherence (HMQC) spectrum of
cyanogen passicapsin in CD3OD from P. rubra ......................................... 88

4-5. Proton/Carbon heteronuclear multiple quantum coherence (HMQC) spectrum of
cyanogen passitrifasciatin in CD30D from Passiflora sp........................................ 89

4-6. Potential enzymatic hydrolysis products of passibiflorin. ....................................90

4-7. Potential enzymatic hydrolysis products of passicapsin. ...................................... 91

5-1. Heliconius himera from Vilcabamba, Loja Province, Ecuador ............. 103

5-2. Electrophoresis in SDS-10% gel slab. The gel shows different levels of purification
of the protein studied. CE: is the cytosolic crude extract, Q sepharose elutes:
Q1-Q3 fractions 31-33, Mono-Q elutes: M1-M2, fractions 23 and 24 from Qsepharose fraction 31, and M3-M4, fractions 24 and 25 from Q-sepharose
fraction 32. RM is the rainbow marker ........................................ 104

5-3. Comparison of amino acid sequence of Spodopterafrugiperda (Lepidoptera:
Noctuidae) P-glucosidase with mass spectroscopy identified peptides from H.
himera. The letters in bold indicate the homologous amino acids sequencies
from H. himera that match with S. frugiperda. A. Mr 55.000 and B. Mr 65,000. 105













xiv















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

COMPARATIVE PROCESSING OF CYANOGENIC GLYCOSIDES AND A NOVEL
CYANIDE INHIBITORY ENZYME IN HELICONIUS BUTTERFLIES
(LEPIDOPTERA: NYMPHALIDAE: HELICONIINAE) By

Mirian Medina Hay-Roe

May 2004

Chair: Thomas C. Emmel
Major Department: Entomology and Nematology

Heliconius erato, a neotropical butterfly known to specialize in larval feeding on cyanide-producing Passiflora plants, has evolved different strategies for dealing with cyanogenic compounds. By combining ecological and biochemical data, I investigated some of these mechanisms. Subspecies of Heliconius erato from the eastern and the western side of the Andes differ in body size. The subspecies from the eastern side of the Andes proved to be consistently larger than their relatives from the western side of the Andes. These differences in body size are genetically based. When two subspecies (H. e. cyrbia and H. e. favorinus) were fed different host plants (natural and alternative host), plastic growth trajectories resulted which were related to the host plant fed on by the larvae. Feeding on one of the host plants, P. rubra, resulted in lower growth rates in both subspecies and higher mortality in H. e. cyrbia, but not in H. e. favorinus. As a result, the few individuals from the former subspecies that survived while feeding on this highly toxic cyanogenic plant were better protected against frog predators because they


xv








accumulated higher concentrations of cyanogenic glycosides in their bodies, a condition which in turn generates higher concentrations of hydrogen cyanide. H. e. favorinus, on the other hand, turned out to be less cyanogenic and therefore less well-protected against predation by frogs, demonstrating that chemical defenses in H. erato vary between geographical races and that lower cyanide concentrations may lead to higher predation rates.

H. erato not only synthesizes de novo aliphatic cyanogenic glycosides, but also

sequesters both simple and complex cyclopentenoid glycosides from their host plants. In addition, during the larval stage H. erato metabolized the complex cyclopentenoid glycosides into simple cyclopentenoid glycosides. Analysis of amino acids in the adults revealed that these butterflies are storing not only cyanogenic compounds for defense, but also essential amino acids from metabolized cyclopentenoids gathered during the larval stage.

Finally, a novel cyanide detoxification enzyme was tested in vitro. Two 1glucosidases were isolated from larval midgut; they had some homologies with Spodopterafrugiperda -glucosidase. The enzymes are specific to simple cyclopentenoid glycosides and proved to inhibit the production of cyanide when cyanogenic glycosides, plant P-glucosidase and larval midgut P-glucosidase were tested. This discovery leads to the conclusion that this detoxification mechanism is related to the ability of Heliconius to metabolize different Passiflora cyanogen types. However, more experiments are warranted to describe the characteristics of the enzymes. The importance of this new mechanism of cyanide detoxification is discussed.





xvi














CHAPTER 1
INTRODUCTION

Heliconius erato is a brightly colored, highly variably patterned butterfly that is distributed from Central to South America. H erato is adapted to altitudes that range from sea level to 1,600 m. It is frequently found flying in open pastures in disturbed forest and secondary growth (DeVries 1987). This species is unpalatable to predators and shows strong Miillerian mimicry with other distasteful species. Geographical races have different patterns of coloration that range from orange and black, to red, yellow and black, to iridescent blue and pink (this pattern of coloration is shared with the co-mimic H. melpomene). The races have diversified broadly and there are now approximately 30 parapatric races that are able to mate randomly in narrow hybridization zones (Emsley 1964. Mallet 1986).

In this study, three races of erato were observed: H. e. cyrbia from Pichincha, Ecuador; H e. favorinus from Tingo Maria, Huanuco, Peru; and H e. demophon from Panama. Heliconius himera (used in experiments described in Chapter 5) constitutes an intermediate step in the transition from a race to a species (McMillan et al. 1997). It had previously been considered a race of erato (Lamas 1976, Brown 1979, Sheppard et al. 1985), but recent mtDNA studies suggest a divergence approximately 1 million years ago (Brower 1994; see also Emsley 1965 for arguments that it is actually a separate species).

Phylogenetic studies of H erato group the races into two main clades. One clade groups races that are located on the eastern side of the Andes, while the second clade can be found on the western side of the Andes (Brower 1994). The ecology (Brown 1979,


1






2

1981, Gilbert 1991, Benson et al. 1976), behavior (Chai 1986, 1990, 1996), and physiology (Dunlap-Pianka et al. 1977, Boggs 1987) of this species have been studied extensively. However, very little is known about the chemical ecology of this species.

Heliconius erato is known to specialize on Passiflora plants (Passifloraceae). The females lay eggs individually on new growth or tendrils. In many areas of South America, this species has been considered an oligophagous species, as larvae feed on many different Passiflora species, but only from the subgenus Decaloba (Benson et al. 1976, Menna-Barreto and Araujo 1985, Hay-Roe 1996). In Central America, however, this species tends to be monophagous, selecting only one or two host plant species at any one location (Gilbert 1991, Benson et al. 1976). Diet breadth has been associated with competition for resources within members of different species of Heliconius that are distributed in each geographical region. This drives the partitioning of the resources among competing species.

With respect to its adult diet, Heliconius has developed a special adaptation

allowing it to feed on pollen from Psiguria and Gurania flowers (Cucurbitaceae) by externally regurgitated digestive enzymes (Gilbert 1972). Pollen plays an important role in reproduction because butterflies assimilate the amino acids released from the pollen for egg development and long-term survival (Dunlap-Pianka 1979). Adult H. erato also feed on liquid nectar from Lantana flowers which provide carbohydrates in their nectar as well as in pollen (Boggs 1981).

Ehrlich and Raven (1964) have determined that secondary plant compounds play a critical role in determining patterns of plant utilization. Insects appear to follow their host plants' adaptive radiation, dispersal, and elaboration of secondary chemical






3


constituents. Host ranges diverge and insects specialize with respect to their host plants as they radiate. In the Heliconius-Passiflora system, coevolution is thought to be the process that has generated this tight insect-plant interaction (Benson et al. 1976, Futuyma and Keese 1992).

Passiflora plants form a large genus which has been reviewed by many botanists (e.g., Harms 1924, Killip 1938, De Wilde 1971, 1974). The most recent review, by McDougal and Feuillet (unpublished), divided the genus into four subgenera and various super sections comprising a total of 521 species. Characters such as plant chemistry (Spencer 1984) and ecological data nicely fit the taxonomic data, specifically within each of the sections of Passiflora.

Passiflora plants have developed a series of morphological characteristics to avoid predation by herbivores such as Heliconius larvae (Benson et al. 1976). The development of a great diversity of leaf shapes by the same plant (a phenomenon called heteroblastia) is notable in the genus. Gilbert (1983) speculated that the difference in leaf appearance assists the plant in escaping detection by female Heliconiinae butterflies as they search for appropriate oviposition substrates. Mechanical defense in Passiflora species may occur in the form of protective hairs on the leaves, especially in the section Pseudodysosmia. The hooks of Passiflora adenopoda penetrate the cuticle of Heliconius larvae, causing immobilization and death from desiccation (Gilbert 1971). Passiflora also develop deciduous filiform stipules which resemble tendrils, and which may function to stimulate egg placement by Heliconius, later causing the eggs to fall off with the deciduous stipules. Various Passiflora produce novel structures on the stipules, leaves or new shoots which resemble the eggs of Heliconius butterflies in shape and color, and





4


cause female butterflies to avoid "further" oviposition on such plants (Williams and Gilbert 1981). For chemical protection against herbivores, several strategies are used by the Passiflora. Most of the Passiflora produce sugary secretions from extrafloral nectaries on petioles, leaf margins, surface, tips, bracts and stipules, which attract ants that actively defend the plants. Ants then eat heliconian eggs, and attack and carry off small larvae (Lanza 1988). But most importantly, Passiflora plants are protected from most herbivores by their production of toxic cyanide-releasing secondary compounds called cyanogenic glycosides (Spencer 1988). The Passifloraceae are among the few plants producing cyanogenic glycosides with a cyclopentene moiety (Spencer 1988). This characteristic is shared only by members of five other closely related families: Flacortiaceae, Malesherbiaceae, Turneraceae, Achariaceae and Caricaceae.

Few Heliconius species have been investigated with regard to their chemical

interaction with Passiflora cyanogens. It is known that the larvae, pupae, and adults of Heliconius and other related genera biosynthesize de novo two simple aliphatic cyanogens, linamarin and lotaustralin (Nahrstedt and Davis 1983). Also, it has recently been demonstrated that Heliconius sara, a specialist on P. auriculata, sequesters cyanogens (epivolkenin) from its host plant and metabolizes most of this compound into the corresponding thiol derivative (sarauriculatin) by replacing the nitrile group which prevents cyanide release (Engler et al. 2000). Adult butterflies of the species Acraea horta (Nymphalidae: Acraeinae) store the cyclopentyl cyanogen, gynocardin, when they develop as larvae on Passiflora plants (Raubenheimer 1989).

In order to understand the phenomena of synthesis and sequestration, of chemical compounds in Lepidoptera, it is important to understand the life histories of these species






5


(Nishida 2002). Vital clues to the biochemistry of defensive and other chemicals may be revealed by ecological information gathered in the field or in live captive cultures.

In this dissertation I will combine ecological and biochemical data in order to understand the patterns of host utilization and detoxification in geographical races of Heiiconius erato. In Chapter 2, I compare the life histories of two races of Heliconius erato (one from the eastern and one from the western side of the Andes) that differ in body size and describe the effects of natural host plants vs. alternate Passiflora plants on their growth, nutrition and mortality. In Chapter 3, I analyze cyanide concentrations in butterflies and host plants, in order to understand whether there are any differences in defense mechanisms between races. In Chapter 4, I describe the isolation and identification of cyanogenic glycosides in both butterflies and plants in order to learn how the butterflies process these toxic compounds. In Chapter 5, I present preliminary results on the analysis of 13-glucosidase from larval guts in order to test the hypothesis that this enzyme will inhibit cyanide release by Passiflora plants.













CHAPTER 2
LIFE HISTORIES AND NUTRITIONAL EFFECTS ON SIZE OF HELICONIUS
ERA TO FEEDING ON DIFFERENT PASSIFLORA HOST PLANTS

In the last two decades, many studies of life history evolution and evolution of

body size have been published. These studies present a retrospective analysis of patterns of variation in size and age at maturity in an attempt to understand the mechanisms responsible for generating phenotypic variation in maturation. However, there are important aspects that these studies did not take into account, such as invariant age and/or size thresholds that occur during instar transitions in larval development (Nijhout 1974, 1975). To study the life history of a population and/or species, one must, first of all, study the egg because it constitutes potential maternal (Dunlap-Pianka et al. 1977, Sinervo et al. 1992) and paternal (Boggs and Gilbert 1979, Wiklund and Kaitala 1993) contributions to the offspring's fitness. The second stage, juvenile development, is probably the most important stage in the life history of an arthropod. At this stage, three different characters may be responsible for the age and size at maturity: the number of instars, the larval developmental time or molting interval, and the change in size at ecdysis. These characters can be plastic or canalized (Higgins and Rankin 1996) and might differ, depending on the different post-embryonic pathways of the arthropod. In the end, the variation among these stages may affect the final adult size in the species or population (Figure 2-1).

Recent models are focusing not only on the analysis of growth but also on age, and other complex and combined characters, which reflect more accurately how adaptive



6






7


maturation phenotypes are achieved by different organisms (Bernardo 1993, Higgins and Rankin, 1996, Kirkpatrick and Lofsrold 1992, Nylin 1992, Nylin et al. 1989, Steam and Koella 1986). Thus, age and size at maturity are the results of a growth and developmental history which reflect both genetic and environmental factors. These factors act throughout the juvenile developmental time. The relative contribution of genetic and environmental effects to age and size are likely to be different for individuals within and between populations and species.

Plastic juvenile development may play an important role in the interaction between the life cycle and the habitat. The phenotypic value in this context is best explained using the quantitative genetics formula from Falconer (1989): P=G+E+M

where P is the phenotypic value; G is the genotypic value; E is the environmental deviation; and M is the contribution from maternal effects (non-genetic effects).

The latest phylogeny of H. erato (Brower 1994) grouped different subspecies of H. erato (geographical races) into two main clades. One of the clades grouped races that are geographically distributed on the eastern side of the Andes (the eastern clade), while the second grouped races distributed on the western side of the Andes (the western clade).

In a previous study (Hay-Roe 1996), differences in body size between two

subspecies of H. erato were investigated. One subspecies from the western side of the Andes (H. e. petiveranus) is very specific in its host plant feeding, while a second subspecies (H. e. phyllis) from the eastern side of the Andes is a broad specialist on various subgenera of the Passiflora. Differences in body size were genetically based, H e. phyllis is bigger than H. e. petiveranus. Differences in growth trajectories between the








races were plastic in H e. petiveranus (due to variation in larval developmental time), but canalized in H. e. phyllis. The results of this study led me to hypothesize that there may be a difference in patterns of growth within subspecies from the western vs. the eastern side of the Andes. I also wondered whether the members of the subspecies located on the eastern side of the Andes were always bigger in size than the members of the western side of the Andes.

In this chapter of the dissertation, I have first examined museum specimens of the most representative members of each region to see whether there is a common variation in size between different races of erato from the eastern and the western sides of the Andes. Second, I have studied the life histories of two geographical races, H. e. favorinus from the eastern side of the Andes and H. e. cyrbia from the western side of the Andes. For each race, phenotypic effects were analyzed based on environmental, genotypic, and maternal effects. Growth trajectories, final size and mortality were compared between groups fed natural host plant vs. alternative host plants.

Materials and Methods

Museum Specimens

The Lepidoptera collection of the Allyn Museum of Lepidoptera, Sarasota, Florida, was used to record the wing lengths of different subspecies of H. erato (Figure 2-2). Wing length was measured from the base of the wing to the apex. A Vernier caliper (Spi 2000) graduated to 0.1 mm was used to make the measurements.

Four subspecies, H. e. colombina, H. e. hydara, H. e. cyrbia, and H e. petiveranus, are representatives from the western side of the Andes. These subspecies were compared with H. e. phyllis, H. e. favorinus, H. e. magnificus from the eastern side of the Andes.





9


Living Insects and Host Plants

Two subspecies were maintained in temperature controlled Lord & Burnham glass houses (5 x 8 m) at the University of Florida, Department of Entomology and Nematology. H. e. favorinus were the descendents of 60 individuals collected at Tingo Maria, Huanuco, Peru, which is located in the valley of the upper Huallaga river, at the base of the western slopes of the eastern chain of the Andes (locally known as "Cordillera Azul"). H. e cyrbia were the descendents of 50 individuals collected at Pichincha, Ecuador, located on the western chain of the Andes. The two subspecies were kept in separate cages to avoid hybridization. Butterflies in each colony were provided with their natural host plants. Passiflora rubra, P. punctata, and P. auriculata from Ecuador are the natural host plants of H. e. cyrbia, while P. trifasciata is one of the natural host plants of H. e. favorinus. Although this Passiflora species is reported here as P. trifasciata, it is actually an undescribed sister species of P. tricuspis and P. trifasciata (McDougal personal communication).

P. biflora from Costa Rica is an alternative host plant for both subspecies. Lantana and Pentas plants were used as adult nectar sources, while Psiguria spp. flowers were used for pollen feeding.

Cups containing 20% sugar-water and red and orange "Tuffy" sponges sprayed with sugar water and an adult artificial diet of an amino acid solution that simulates the amino acid content of Psiguria flowers (Gilbert unpublished) were also provided to ensure an adequate food supply to the adults. Feeding Experiments

Seventeen H. e. favorinus females were used for the host plant feeding experiment. Twelve H. e. cyrbia females were used to represent the Ecuadorian race. The colony was






10


obtained from a company in Ecuador called Heliconius Works Quito, Ecuador. Some H. e. cyrbia were also obtained from other colonies in the USA. Since I was not sure whether this race was fed on its natural host plant as larvae. I raised H. e. cyrbia on their natural host plants P. punctata and P. rubra for at least three generations prior to starting my experiments.

The offspring from each female are referred to as family and were so tracked through the experiments. This identification allowed me to recognize any life history differences between offspring from different female butterflies. Such differences would indicate genetic variations within populations for the trait in question. Analyzing the egg mass reveals potential maternal effects.

Isolated females were allowed to lay eggs on their natural host plants. The eggs

were then collected and taken to environmental chambers for the hatching larvae to grow under constant laboratory conditions (270C, 75% humidity and a 14L: 10D photoperiod). Life History Studies

Different aspects of the life histories were observed and contrasted between races from the time the egg was laid until adult emergence from the pupae.

As the larvae were raised, they were checked at least three times a day (morning, afternoon, night) for the presence of head capsule apolysis, or for evidence of ecdysis. Larval developmental time in each instar was recorded. When larvae in the last instar were observed in the position typical of pupation, that day was recorded as the first day of pupation. Pupal developmental time was recorded to have ended when the adult butterfly emerged from the pupal case.








Egg measurements

Eggs were collected every day from the experimental cages and placed in a small SOLO cup. Each egg was weighed on an electro balance graduated to 0.01 mg (Mettler AC 100, Mettler Instruments Corp. Hightstown, New Jersey). Each egg was placed in an individual cup with the family identification number (ID), offspring ID, host plant and oviposition date. Fresh shoots were placed into the cups, usually on the second and third day when I expected the larvae to hatch from the egg. A small wet ball of Kimwipe tissue was also placed in the cup to provide eggs with adequate moisture. Larval development

Upon hatching, larvae were removed from the cups with a fine artist's brush and placed on shoots of the experimental host plants (P. auriculata, P. punctata, P. rubra. P. trifasciata and P. biflora).

The base of each plant cutting (approximately 15 cm in length) was placed in a glass vial with distilled water. The bottle with the plant was then placed in a plastic aquarium (one gallon) that was labeled with the following data: family ID, host plant used, geographical race of erato, and the offspring ID.

As a larva develops, it periodically molts, which is first signaled by apolysis of the head capsule. The head capsule is a strongly sclerotized part of the larval body and is therefore a good indicator of growth. Consequently, I measured head capsule dimensions. Head capsule width and height were measured. Measurements were made with an Olympus microscope and an ocular micrometer. All measurements were done at 20X magnification. Relying on head capsule width has one drawback. In the last instar, the epicranial suture splits the head capsule during the molt, so these measurements are not accurate indicators of growth at this instar. Head capsule height, however, is not





12

affected by the split, so one can use the height of the last instar head capsule to indicate growth during the last instar. The following formula was used to calculate growth rate: Growth rate = (In (Hf) In (Hi))/t

where Hf is the measurement of the fifth instar head capsule height; Hi is the measurement of the first instar head capsule height; and t is the larval developmental time. A natural logarithm transformation was performed as recommended by LaBarbera (1989).

Pupae and adult offspring measurement

Each pupa was weighed two days after pupation. Measurements of pupa length and width also were taken.

For the adult offspring, I recorded the following data: sex, wing length, total body length (measured from the head to the tip of the abdomen), and fresh weight. Measurements were made in the afternoon to allow time for the new adults to dry their wings after emergence. A Vernier caliper (Spi 2000) graduated to 0.1 mm was used for those measurements. Mass measurements were made with an electro balance. Teneral butterflies were then stored at -850C for further analysis. Statistical Analysis

Systat program V 10 was used for the statistical analysis. Since the variation in the head capsule size between instars was large, the measurements were converted to a logarithmic scale to do the statistical analysis. Repeated measures ANOVA was used for analyzing growth trajectories between races. The statistical test was performed on both head capsule width (using the measurement from the first to the fourth instar) and head capsule height. Since both yielded the same results, I present here the results of the analysis of head capsule height. Finally ANOVA was used for the rest of the analysis. In





13

all cases, family was nested within race to observe either maternal or genetic variation. Also, sex was used as factor and if the interaction was not statistically significant it was excluded from the analysis. A Tukey test was used for multiple comparisons.

Results

Analysis of Specimens from the Museum Collection

The ANOVA of the measurements taken from the museum specimens supports the hypothesis that subspecies from the eastern side of the Andes tend to be larger than the subspecies from the western side of the Andes (Table 2-1). Regions alone had a large effect on size (F=254.2; df=- 1, 179; p<0.001), and different races of erato within the regions are also different (F=14.4; df= 5, 179; p<0.001).

The subspecies from the eastern region are relatively constant with respect to their body size. In contrast, the subspecies from the western side of the Andes show more variability in their sizes.

Life History Studies in H. e. favorinus and H. e. cyrbia

Data collected in the field (by myself and museum specimens) indicated that H. e. favorinus is larger than H. e. cyrbia (ANOVA for wing length F= 283.8, p<0.001 and body length F=45.3, p<0.001) (Figure 2-3). Eggs

Heliconius erato egg mass was not significantly different between races. However, egg mass was strongly influenced by family within each race (F= 5.69, p<0.001). This result indicates maternal effects for egg mass. Both races laid eggs singly, mostly on young growth (tendrils, and the tips of developing natural Passiflora host plants). The eggs are cylindrical, yellow, with a pointed apex resembling a crown. I did not find any difference in egg morphology between races. First instars hatch from the





14


egg on the third day following oviposition. All first instars in both races ate their egg shells.

Larvae

Both subspecies underwent five instars of development. No supernumerary instars were detected during the experiment.

Three different analyses were carried out with head capsule measurements. The

first was the comparison of head capsule size and larval developmental time between and within geographical races feeding on natural vs. alternative host plants.

Repeated measurements ANOVA for head capsule size showed a significant variation in size between the two races (Table 2-2). Analysis within H. e. favorinus showed that host plants directly affect head capsule size (Table 2-2). Tukey comparison test indicates that individuals fed on P. auriculata tend to be bigger than individuals fed P. trifasciata and P. rubra during the second, third and fourth instars of larval development. No significant differences in head capsule size were found among members of H. e. cyrbia. There was no size variation among full-sib within races.

Differences in larval developmental time between and within races were

statistically significant between family and between individuals fed different host plants (Table 2-3). On average, H. e. favorinus larvae fed longer on P. auriculata and P. rubra (Table 2-4). Both H. e. favorinus, and H. e. cyrbia also showed significant differences between larval developmental time between full-sibs, indicating genetic variation within them and between individuals fed different hosts (Table 2-3). Longer larval developmental time was experienced by both male and female cyrbia fed on rubra, as well as females fed auriculata and males fed trifasciata (Table 2-4).






15


Both H. e. fhivorinus and H. e. cyrbia had plastic growth trajectories (Figure 2-4 AB) (Higgins and Rankin 1996). However, as indicated above in H e. favorinus the plasticity is due to variation of head capsule size and developmental time, while in H. e. cyrbia it is only due to variation in larval developmental time. Differences in growth trajectories for the two races on their natural host plants can be observed in Figure 2-5.

The last part of the analysis was to calculate growth rates and compare them

between races. Growth rates were significantly different between races (F=6.18; df=l, 227; p<0.01). Host plants also affected growth rate (F=12.11; df-4, 227; p<0.001) between races. Lower growth rates in female H. e. favorinus fed P. auriculata, and in males and females fed P. rubra, were due to prolonged development. Within H. e. cyrbia, low growth rates occurred when males and females were fed P. rubra (Table 2-4). Pupae

The average pupal developmental time for both races was eight days. When H. e. favorinus fed on different host plants, it resulted in a significant variation in pupal mass (F=4.57; df=4, 145; p<0.001). Multiple comparison test indicated that individuals fed P. trifasciata were significantly heavier than individuals fed P. punctata, P. auriculata and P. rubra (Figure 2-6).

Adult

Final size under laboratory conditions corresponds to the data gathered in the field and from museum specimens. Regardless of the host plant, H. e. favorinus was greater in wing length than H. e cyrbia (F= 271.63; df=l, 230; p<0.001) (Table 2-4).

Analysis within H. e. favorinus indicates that host plant selection caused a slight but significant variation in wing length (F= 2.34; df-4, 145; p=0.05). Multiple comparison showed that individuals that fed on P. punctata were smaller in size than the






16


ones fed P. trifasciata. No other host caused a significant variation in wing length. On the other hand, H. e. cyrbia wing length did not differ significantly between individuals fed on different hosts. There was no size difference between sexes (Figure 2-7). Mortality

In general, higher mortality occurred throughout the life cycle in H e. cyrbia. I started with 148 individuals from H. e. cyrbia and 143 individuals from H e. favorinus. At the end of the experiment I had a total life cycle mortality of 38.74% in H. e. cyrbia and 13.26% in H. e. favorinus. During the experiments, 8 eggs from H. e. favorinus and 10 from H. e. cyrbia did not hatch; some were dried, others were bitten by ants. High mortality during the larval stage was observed in H. e. cyrbia fed its host plant P. rubra. In all cases, larvae died during the early instars (Table 2-5).

Discussion and Conclusions

The results presented here support the preliminary hypothesis that there are

differences in body sizes between races of Heliconius erato from the western and eastern sides of the Andes. While the group of races from the eastern side of the Andes reveal low variations in body size, the group of races from the western side of the Andes present much greater variation. In the eastern region, stabilizing selection probably acts against extreme phenotypes and favors the more common variant. This mode of selection may reduce variation and maintains the status of this particular large phenotypic character. The races of the western side of the Andes are perhaps more exposed to ecological pressures, such as competitive interactions, for example, as discussed by Benson et al. (1976). As Brower (1994) has explained, the vicariant separation of the races could have occurred during the Cenozoic era, specifically the Pleistocene, when the formation of large mountains around the world (including the Andes) occurred.






17

In this chapter, I demonstrated through life history studies of the two races of Heliconius erato that differences in body size seem to be genetically based and size ,differences will remain, regardless of the host plant provided. Within each population, there were maternal effects confirmed by egg mass variation between full-sibs. Also, genotypic variation occurred in each stage of the life history within each race.

H. e. favorinus proved to have more plastic growth trajectories, due to variation in their average larval developmental time and variation in size when the larvae were fed different host plants. This result demonstrates that both developmental pattern and growth are influenced by host plants (chemical/nutritional). It also indicates that maturation is not simply a consequence of growth, and that environments can have correlated effects on both growth and developmental rates. Within this race, I also proved that whenfavorinus fed on their natural host plant, P. trifasciata, as well as the alternative host P. punctata, they generate earlier-born progeny. Earlier-born progeny might have a significant impact on population growth (Fisher 1930 and Charlesworth 1980) because they might increase reproductive success and fitness in stable populations. This particularly occurred in H. e. favorinus fed its natural host P. trifasciata, which produced larger individuals of both sexes. Larger individuals will again have a direct effects on fecundity and competitive ability (mating success), especially in competitions for pupal mating (Deinert et al. 1994). However, the alternative host plants (P. punctata) produced smaller females, on average, which directly affects fitness, because smaller females will tend to produce fewer eggs on average (Dunlap-Pianka 1979 and Boggs 1979). Interestingly, fast growing individuals become heavier than individuals that grow more slowly. This is probably due to faster food consumption (not measured here). Final






18


adult size indicates that there is only a slight difference (p=0.05) in size offavorinus when fed its natural host plant compared to other host plants. This result indicates that there is a strong selection for being the "right size" as well as having a short developmental time.

On the other hand, H. e. cyrbia also presented plastic growth across host plants. However, the most striking result was the high mortality and low growth rate when individuals fed on their natural host plant, P. rubra. Although not significant, there was also a lower growth rate when female larvae fed on P. auriculata and they generated smaller females. This particular race has been investigated (Jiggins et al. 1997) in relationship to their hybridization zones and the recent speciation of H. himera. In contrast to their results, my greenhouse colony of cyrbia displayed a marked preference for oviposition on P. punctata. H. e. cyrbia was exposed to both natural host plants for more than 10 generations. Two important observations regarding host plant preferences were made during these studies (Hay-Roe and Rios, personal observations), a. On average, we collected 3 to 5 eggs per day on P. rubra and 10 to 12 eggs per day
on P. punctata.

b. On various occasions, P. punctata were defoliated when we allowed the larvae to
grow on the plants inside the cage. P. rubra were never defoliated.

Regardless of the observations of oviposition behavior, the results of this study

suggest that there is a difference in survival of larvae fed P. rubra vs. P. punctata, a fact that Jiggins et al. (1997) did not examine explicitly (also, Mallet personal communications). Perhaps, one ought to revive the hypothesis that differences in host plant usage between the sister species H. e. cyrbia and H. himera could be one cause that induced the recent speciation of H. himera.





19


Several factors might be operating during the life history of these subspecies,

creating selective pressures affecting larval growth rates and trajectories (e.g. temporal differences in host availability, chemical differences among hosts that affect survival ability, etc.) The results of this study suggest that H. e. favorinus employed different strategies than H e. cvyrbia did to overcome the intake of different chemicals and/or nutrients from the host plants. This is suggested by differences in larval performance and survival. In the following chapters I shall explore the importance of plant chemistry as related to the coevolutionary interaction between H. erato and Passiflora host plants.






20


Table 2-1. Average wing length and standard error of geographical races of Heliconius
erato from museum specimens in different regions.
Geographical races Regions N WL (cm) X+ se
H e. magnificus Eastern 15 3.85 + 0.06 H. e. favorinus Eastern 57 3.94 0.03 H. e. phyllis Eastern 32 3.89 0.04 H. e. hydara Western 15 3.30 0.06 H. e. petiveranus Western 25 3.01 0.08 H. e. cyrbia Western 31 3.55 0.03 H. e. colombina Western 15 3.24 0.03




Table 2-2. Repeated measurements ANOVA of head capsule size with race, host and family as factors.
Source Df F P Between races
Race 1 751.96 0.001 Host 4 3.86 0.005 Race x Host 4 1.18 0.323 Error 223
Within races
H. e. favorinus
Family 16 1.47 0.120 Host 4 3.58 0.008 Error 129
H. e cvrbia
Samil 9 1.52 0.159 Host 4 1.32 0.272 Error 69






21


'Table 2-3. Repeated measurements ANOVA of larval developmental time with race and
host as factors.
Source Df F P Between races
Race 1 9.68 0.002 Host 4 15.75 0.001 Race x Host 4 0.62 0.649 Error 223 Within races
SH. e. favorinus
a Family 16 2.68 0.001 Host 4 33.89 0.001 Error 129 H. e. cyrbia
i Family 9 3.27 0.002 Host 4 3.41 0.013 Error 69


'Table 2-4. Sexual differences in larval developmental time (LDT), growth rate (GR) and
wing length (WL) in H. e. favorinus and H e. cyrbia feeding on five different
host plants.
Race H ost Plants N Sex LDT (days) GR WL (cm) Both X se (mm/days) X se sexes } X se I H. e. auriculata 24 9.18 0.26 0.197 0.006 3.95 0.044 tavorinus 9.54 0.31 0.186 0.007 3.91 0.040
biflora 25 (S 8.37 0.18 0.217 0.006 3.98 0.027
8.64 0.10 0.205 0.003 3.92 0.034 punctata 26 6 8.07 0.12 0.219 0.004 3.94 0.033
8.36 0.15 0.208 0.004 3.81 0.090 rubra 23 8 10.46 0.28 0.174 0.007 3.90 0.060 Y+ 9.75 0.25 0.188 0.006 3.92 0.046 trifasciata 26 r3 8.84 0.21 0.205 0.006 4.00 0.034 SY 8.79 0.18 0.208 0.005 4.00 0.030 I He cyrbia auriculata 19 9.13 + 0.30 0.211 0.007 3.58 :0.060 9 9.70 0.73 0.203 0.012 3.48 0.085 biflora 19 8.38 + 0.26 0.226 0.006 3.68 0.047
_ 9.00 0.88 0.220 0.014 3.55 0.036 punctata 17 9.38 0.99 0.212 0.016 3.53 0.067
8.22 0.22 0.230 0.008 3.58 0.036 rubra 17 ( 9.80 0.33 0.190 + 0.007 3.54 0.051 9 10.00 0.26 0.188 0.006 3.54 0.032 trifasciata 18 6 9.63 0.60 0.198 0.014 3.51 0.063 I 9.00 0.3i 0.215 0.007 3.61 10.029






22


Table 2-5. Mortality in different stages of the life history of H. e. favorinus and H e
cyrbia feeding on natural and suitable host plants.
Race Egg Host plant Larvae Pupae Adult
N i%-7%%% N %
dead dead dead dead H ie. 8 5.6 auriculata 1 0.7 0 0 0 0

biflora 2 1.4 2 1.4 0 0 punctata 1 0.7 0 0 0 0 rubra 5 3.5 0 0 0 0 trifasciata 0 0 0 0 0 0 e. 10 6.75 auriculate 2 1.4 2 1.35 1 1.1 cyrbia
biflora 5 3.4 2 1.4 0 0 punctata 1 0.7 3 2.0 0 0 rubra 20 13.5 2 1.4 2 1.4
___t_ __ri fasciata 7 4.7 0 0 1 0.7











EGG LARVAL DEVELOPMENT PUPA ADULT




/ ,,

%~
"-- -- ---- .7\;'
Si ..;i"


I --_ I


ca. 3 ca. 15 ca. 8 ca. 3
Days-*O- Days Days"*k Months-*" Figure 2-1. Life history of Heliconius butterflies.





24



D

A





















G








Figure 2-2. Museum specimens of H. erato from the eastern (left column) and western
sides (right column) of the Andes that were included in the analysis. (A) H. e.
phyllis, (B) H. e. favorinus, (C) H. e. magnificus, (D) H. e.colombina, (E)
hydara, (F) H. e. cyrbia, (G) H. e. petiveranus.






25





A

































Figure 2-3. Study organisms. (A) H. e. favorinus from Tingo Maria, Huanuco, Per6 and
(B) H. e. cyrbia from Pichincha, Ecuador.






26


4.5

4.0 A 1 a
E
E 3.5 3.0

2.5
3 2.00 + auicdata
w 1.5 a biflora
o
'a 1.0 punctata w Irubra
I 0.5
trifasciata
0.0 1
0 1 2 3 4 5 6 7 8 9 10 11 Larval developmental time (days)



4.5

4.0 B

E 3.5 -L
E 3.0

2.5

2.0 J
S+. auricata
,, 1.5 b a

1.0- punctata
rubra
X 0.5 "
trifasciata
0.0 ..
0 1 2 3 4 5 6 7 8 9 10 11 Larval developmental time (days)



Figure 2-4. Growth trajectories of (A) H. e. favorinus and (B) H. e. cyrbia feeding on five
different host plants. Each point represents the mean value of larvae from the
first to fifth instar. Bars on each point indicate standard error.






27




4.0

3.5

E3.0
.
.L 2.5
c
S2.0

1.5
0 1.0 =.t

favorinus
: 0.5
Scyrbia
0.0
0 1 2 3 4 5 6 7 8 9 10 11 Larval developmental time (days)


Figure 2-5. Growth trajectories of H. e. favorinus and H. e. cyrbia fed their natural host
plants (P. trifasciata and P. rubra, respectively). Each point represents the
mean value of larvae from the first instar to the fifth. Bars on each point
indicate standard error.






28




0.6
B AB B B A
0.5

0.4



0.2
I 0 A female

0.1
0 B female D B male aunculate billore punctata rubra trifesciate
Host plant


Figure 2-6. Pupal weight of H. e.favorinus (A) and H. e. cyrbia (B) fed on natural and
alternative host plants. Different letters indicate significant differences
between means (Tukey HSD-test). Bars in each point indicate standard error.




4.1
AB AB A
4.0 B AB
3.9
3.8 O favorinus 3.8
3.7 cyrbia

3.6 C 3.5
S3.4
3.3 3.2
auriculata bilora punctata rubra trifasciata
Host Plants

Figure 2-7. Average wing length of H. e. favorinus and H. e. cyrbia fed natural and
alternative host plants. Different letters indicate significant differences
between means (Tukey HSD-test). Bar in each point indicate standard error.













CHAPTER 3
CYANOGENESIS IN HELICONIUS ERA TO AND PASSIFLORA HOST PLANTS

Cyanogenic glycosides are secondary metabolites derived from amino acids. Insects or plants that contain cyanogenic glycosides undergo a process called cyanogenesis, which occurs during tissue disruption by feeding herbivores, predators or pathogens. The process occurs in two steps: first, the enzyme 13-glucosidase hydrolyzes the cyanogenic glycosides, generating a cyanohydrin (a-hydroxynitrile) and a sugar moiety (usually D-glucose, but the products may include other sugars, e. g., gentibiose). The second step involves the degradation of the cyanohydrin and occurs rapidly at normal cell pH. This is thermodynamically favored, and in many instances, but not always, a second type of enzyme (a-hydroxynitrile lyase) catalyzes the dissociation of the cyanohydrin to a carbonyl compound and hydrogen cyanide (H6sel 1981) (Figure 3.1).

Occurrence of Cyanogenesis in Plants

Normally, the cyanogenic glycosides and the enzymes that hydrolyze them to

produce cyanide are spatially separated within the plant tissues to prevent autotoxicity (H6sel 1981, Poulton 1988). In plants, the cyanogenic glycosides are found in the epidermal layer of the plant tissues, whereas both hydrolyzing enzymes are found in the mesophyll cells (Kojima et al. 1979, Saunders et al. 1977, Poulton 1988).

Cyanogenic plants show high variability in hydrogen cyanide (HCN) production, which has been related to low moisture stress (Foulds and Grime 1972, Abbot 1977), frost or low temperatures (Ellis et al. 1977), or presence of herbivores (Cooper-Driver and Swain 1976). This variation reflects phenotypic and genotypic diversity in both the


29






30


production of cyanogenic glycosides and the existence or absence of their degrading

-nzymes (Vetter 2000, Jones 1977, Hughes 1981. Schappert and Shore 1995, 2000). Plants with such genetic variation constitute an important source of information on the direct effects of this type of compound in defense against herbivores that prey upon the plants (Jones 1971).

Occurrence of Cyanogenesis in Lepidoptera

Cyanogenic glycosides have been found in all genera of Heliconiinae, including Cethosia biblis and C. hypsea (Heliconiinae: Nymphalidae) (Nahrstedt and Davis 1983), Acraea horta (Acraeinae: Nymphalidae) (Raubenheimer 1989), the lycaenids Brephidium exillis and Polyommatus icarus (Nahrstedt 1987), and in moths of the family Zygaenidae (Franzl and Naumann 1985, Muhtasib and Evans 1987). The location within the body of cyanogenic glycosides has been studied in Zygaena moths and Heliconius butterflies. In both, the cyanogenic compounds were found in the cuticular part of the integument, although in zygaenids, 30% of the cyanogenic compounds were also found in the hemolymph (Davis and Nahrstedt 1985, Nahrstedt and Davis 1983, 1985). The enzyme 3-glucosidase, on the other hand, is mainly stored in the hemolymph (Nahrstedt 1988).

All stages in the life history of Heliconiinae have been tested and the highest

concentrations of cyanide have been found at the egg stage. The concentrations in the larval and pupal stages do not vary greatly, but cyanide concentration drops in adults (Nahrstedt and Davis 1985).

The variation of the cyanogenic glycoside concentrations within body parts of adult Hteliconius, as related to their host plants, has not been explored. Within insects that are chemically protected, it is expected that the greater concentrations of the toxic compounds will be found in areas of the body with which the predators come in contact






31

first. For example, toxic compounds are more concentrated in the wings of monarch butterflies (Brower and Glazier 1975). Chrysomelid beetles (of the genus Oreina) :rapidly distribute toxic compounds stored in pronotal and elytral glands onto the surface of the cuticle (Pasteels et al. 1995, 1979) when disturbed. Variations in the concentration of defensive compounds among body parts may be due to selection by predators for more ,effective protection and/or may depend on the mode of action or properties of the toxic ,compounds (Bowers 1988).

Cyanogenic Glycosides as a Defensive Mechanism

It is generally accepted that cyanogenic glycosides act as a defense mechanism against many potential herbivores and predators (Jones 1988, Seigler 1991, Nahrstedt 1992, 1988, Gleadow and Woodrow 2002). In the last decade, there has been some controversy over the effectiveness of cyanogenesis as a herbivore defense (Kassarov 1999, 2001, Hruska 1988), but the utilization of cyanogenic glycoside in plant defense has recently been proven by transferring the pathway for cyanogenic glycoside biosynthesis into the acyanogenic Arabidopsis thaliana (Tattersall et al. 2001). There is also some doubt about whether HCN or the aglycone (aldehydes or ketones), which are simultaneously released upon cyanogenesis, are the source of toxicity (Conn 1979, Jones 1988, Spencer 1987). For example, regurgitate of enteric fluids by the eastern tent caterpillar, Malacosoma americana fed black cherry contained both HCN and the carbonyl compound, benzaldehyde. When these compounds were tested on ants, they tolerated the HCN but were repelled by the benzaldehyde (Peterson et al. 1987). Also, Jones (1988) revealed that the carbonyl compounds, acetone and 2-butanone, acted as deterrents to feeding on Lotus corniculatus and Trifolium repens by mollusks. Despite





32


this controversy, what is clear is that feeding is required in order to provoke tissue disruption and enzymatic hydrolysis.

In Lepidoptera, defensive compounds have been classified in relation to how

predators react to them (Brower 1984). Class I are compounds that are noxious due to their ability to irritate or poison; they may or may not stimulate the olfactory or gustatory receptors of predators. Class II are harmless chemicals that stimulate the predator's olfactory and/or gustatory receptors (Brower 1984). Thus cyanogenic glycosides are considered a Class I compound; however, Heliconius also possess Class II pungent odors emitted through their abdominal glands, (Gilbert 1976), and pyrazines produced in some unknown region of the body (Hay-Roe and McAuslane unpublished).

Laboratory experiments have demonstrated that insect predators such as birds, frogs, and lizards can learn to avoid a particular class of prey, depending on the unpalatability of that prey (Brower 1958, Brower et al. 1963, 1971; Chai 1990, 1996). Chai (personal communication) described how inexperienced birds will attack almost anything. However, after tasting the unpalatable butterflies, they quickly become conditioned to the pattern of coloration (Chai 1988) and flying styles of distasteful prey (Srygley and Chai 1990). In the Costa Rican rainforest, Heliconius erato was rejected 41% of the time when tested with jacamar birds (Galbula ruficauda) (Chai 1986).

Various experiments have shown that some butterflies and moths that are generally considered to be highly toxic can be consumed by birds without any rejection behavior (Collins and Watson 1983, Chai, 1986). Within a single species, unpalatability cannot be absolute; the amount and types of chemical defense can vary among individuals, producing a palatability spectrum (Brower et al. 1967).






33

For this chapter, the concentrations of cyanogenic glycosides in various body parts of races of Heliconius erato fed natural host plants were examined. The same analysis was performed on separate parts of Passiflora plants (young leaves, mature leaves and stems). Roots, fruits and flowers were not included in the analysis because Heliconius larvae rarely feed upon these tissues. Cyanide quantification analysis of the plants was performed over two seasons (summer and fall). Finally, cyanide concentrations in the plant were correlated with the concentrations in H erato which fed upon them. The discussion addresses the costs and benefits of allocating cyanogens as a defensive mechanism in the different body parts of this species and the implications of this variation on the palatability of this butterfly and on its mimetic relationships.

Materials and Methods

Plant Growth and Insect Rearing

This study was conducted at the University of Florida in Gainesville with butterfly colonies that were raised in Lord & Burnham glass houses (5 x 8 m). During the first two years of experimentation the colonies were unable to survive the winter, but improvement in the glass houses' environmental conditions made breeding several generations possible in the last years of research. H. e. favorinus was collected in Tingo Maria, Huanuco, Peru, while H. e. cyrbia was collected in Pichincha, Ecuador, and H. e. demophon was collected in Soberania National Park, Panama.

Most of the Passiflora plants began as cuttings brought from the collection of Dr. L. E. Gilbert at the University of Texas at Austin; a few cuttings were also donated by R. Boender, Butterfly World, Fort Lauderdale. P. trifasciata, the natural host plant ofH. e. favorinus, was collected in Tingo Maria by the author of this research. All vouchers of Passiflora plants used for this study are deposited at the Missouri Botanical Garden, St.





34

Louis, Missouri and the identification of the plants has been confirmed by Dr. John MacDougal, an expert in the systematics of the family Passifloraceae. Experimental Design and Treatments Butterflies

H. e. favorinus, H. e. cyrbia and H. e. demophon larvae were raised in

environmental chambers under controlled temperatures (270C), humidity (78%) and photoperiod (1 4L: 1 OD).

The first two races were fed only natural host plants, Passiflora trifasciata for H. e. favorinus, and P. punctata and P. rubra for H. e. cyrbia. H. e. demophon, on the other hand, was fed both natural host plants (P. biflora) and alternative host plants (Table 3.1). The experimental procedure used was as follows: a. Specimens were selected from both H e. cyrbia and H. e. favorinus. b. Each individual male and female was tested separately. c. For each specimen, the head, thorax, abdomen and wings were tested separately.

The procedure was slightly altered for the testing of H e. demophon (the full procedure had not been designed at the time that this testing took place). For this subspecies the head, thorax and abdomen were not tested separately, only the wings were. Temporal cyanide quantification of Passiflora plants

Various Passiflora plants were tested during the summer (July-August) and fall

(September-October) of 1999 (Table 3.2). All the plants tested were potential host plants of Heliconius erato in different geographical areas, such as Brazil, Costa Rica, and Ecuador.

The analysis was performed on three different plant parts: young leaves, mature leaves, and stems. Plant parts not normally eaten by Heliconius larvae were not tested.





35


All Passiflora plants were subjected to the same environmental conditions

(humidity, temperature and natural photoperiod) in the greenhouses. None of the plants tested was exposed to herbivores.

Quantitative Determination of Cyanide Colorimetric analysis

Cyanogenesis was quantified by the Lambert procedure (Lambert et al. 1975) which was modified by Brinkler and Seigler (1989, 1992).

The plant or butterfly samples were ground to a fine powder in liquid nitrogen. The powdered plant or butterfly tissue was poured into 10 ml vials containing 0.1 M phosphate buffer (pH 6.8). The volume of liquid was kept to a minimum, as cyanide is very soluble in water. A 5 ml vial containing fresh IM NaOH solution was then placed inside the 10 ml vial and the 10 ml vial was stoppered tightly. The samples were prepared quickly to avoid loss of HCN. The vials were then incubated at 37C overnight. During the incubation, HCN released from tissue is trapped in the NaOH solution, and forms NaCN.

After the incubation, aliquots of the solution from the small vial were transferred to test tubes and adjusted to 0.1 M NaOH with distilled water. Typically, 0.1 ml aliquots of

1 M NaOH solution are removed and diluted to 1 ml; if the total sample volume is less,

0.1 M NaOH is added to bring the volume up to 1 ml.

The assay involved sequential addition of the following reagents: 1.0 M acetic acid to neutralize the NaOH; a succinimide/N-chlorosuccinimide oxidizing reagent, and a barbituric acid/pyridine coupling reagent (Lambert et al. 1975; Brinker and Seigler 1989).






36


The tubes were vortexed vigorously and the solution was allowed to sit for 10 min. The absorbance was then measured at 580 nm in a spectrophotometer (Spectro 22; Labomed, Culver City, CA).

Standard curve

For the standard curve, a 0.02 M solution of NaCN in 0.1 N NaOH (0.980g NaCN/I) was prepared. The exact amount of CN- in solution was determined by a modified Liebig titration method as follows:

A standard 0.1 M AgNO3 solution was prepared by diluting dry AgNO3 (at 1000C for 3 h) in 1 L deionized water in a volumetric flask. The solution was stored in the dark. Twenty five ml of the NaCN solution was placed in a 50 ml flask equipped with a stirring bar. Into this solution, 1.5 ml 6 N NH40H and 0.5 ml 10% KI were added. This mixture was then titrated with the standard AgNO3 solution, until the mixture became turbid, which occurred at approximately 2.4 ml of AgNO3. The endpoint was easier to see against a black background.

To determine the amount of CN in the 25 ml sample, the amount of AgNO3 used to reach the endpoint was multiplied by 0.005204 g, then divided by 25. The result reflected the g CN /ml.

Aliquots (1 ml) of the NaCN solution prepared above (close to 500 pg CN-/ml) were then diluted in a volumetric flask with 0.1 N NaOH to 100 ml to reach a final concentration of approximately 5 [ig CN /ml.

The standard curve was then prepared using aliquots of the standardized NaCN solution. Between 0 and 0.2 ml of NaCN are increased to 1 ml with 0.1 M NaOH. This gives CN concentrations ranging from 0 to approximately I lg CN- /ml.






37

After the analysis, the ground plant or insect tissue was dried in crucibles for

several days in a drying oven at 650C, until they reached a constant dry weight (approx. 6 days). I recorded the tg CN /g dry weight of plant tissue or butterfly body tissue. Predator-Prey Interaction

During the present study I consistently observed the presence of tree-frogs (Hyla sp.) in the cage containing the colony of H e. favorinus. In order to learn whether this vertebrate was preying on the larvae, I collected frogs and confined them in a separate 1 gallon empty aquaria until the frog had excreted. No permit by the Institutional Animal Care and Use Committee (UF) was required to conduct this experiment. Statistical Analysis

Two-way analysis of variance (ANOVA) was performed using the statistical

program SYSTAT 10. Some of the data was not normally distributed and the data was subjected to square root transformation to compensate for increasing variation associated with larger mean values (Zar 1996). However, the results using transformed data did not vary significantly from the raw data, so the results of the raw data are reported here. A Tukey test was used for multiple comparisons.

In order to learn whether cyanide concentration in butterflies is correlated with

cyanide concentration in Passiflora host plants, a linear regression was done on the mean cyanide concentration of Passiflora host plant with mean cyanide concentration of H. erato fed natural host plants.

Results
Temporal Variation of Cyanide Release in Passiflora Plants

Cyanide concentrations varied significantly between the two analyzed seasons (F=129.95; df=l, 388; p<0.001). In the fall, cyanide concentrations increased to






38


approximately two times the amount released in the summer. Also, there was a significant difference in cyanide concentrations between host plants tested (F=45.06; df=-6, 388; p<0.001) and plant parts (F=131.83; df=2, 388; p<0.001). The concentration of cyanide was always higher in the new leaves, and this value differed significantly from the cyanide concentration in both mature leaves and stems. Cyanide concentration in mature leaves and stems did not follow a constant pattern, and varied according to the species (Figure 3.2).

Cyanide Quantification in Heliconius erato Fed Their Natural and Alternative Host
Plants

Cyanide concentrations varied significantly between the races (F=29.90; df=l, 139; p<0.001). H e. cyrbia released greater amounts of cyanide than H e. favorinus. There were also differences in cyanide concentration between the sexes (F=8.26; df-l, 139; p<0.005) and also between body parts (F=8.14; df=3, 139; p<0.001). H. e. cyrbia fed on P. rubra and P. punctata

Differences in host plant utilization resulted in different concentrations of cyanide being released from individuals of this race (F= 10.02; df=l, 97 p<0.002). When H. e. cyrbia fed on P. rubra (Table 3-3), the butterflies released higher concentrations of cyanide, as compared to the butterflies that fed on P. punctata (Figure 3.4). Also, males and females of this subspecies differed significantly in the quantity of cyanide released (F=5.68; df=l, 97; p<0.025). The females were able to concentrate higher amounts of cyanide in their body when fed P. rubra. Furthermore, the concentrations of cyanide were stored differently within the body parts (F=7.32; df-3, 97; p<0.001). The thorax had higher concentrations of cyanide, which differed significantly from the amounts stored in the head, the abdomen, and the wings. When H e. cyrbia was fed P. punctata,






39


the cyanide concentrations did not differ between the sexes. However, there was a significant difference in concentrations between body parts (F=12.31; df=3, 52; p<0.001). The thorax and the head had higher cyanide concentrations, as compared to the abdomen and wings (Figure 3.4).

H. e. favorinus fed P. trifasciata

Low concentrations of cyanide were found within the body of this race. Although the female thorax and head seemed to have higher cyanide concentration than the males, the statistical analysis did not show any significant differences between these values. Wings and abdomen had the lowest concentrations of cyanide. More males than females were tested because more than 50% of the males tested negative. Addition of exogenous 1-glucosidase did not favor the release of cyanide, indicating that cyanide release is not due to the lack of the enzyme, but to the lower concentrations of cyanogenic glycosides within the butterfly's body (Figure 3.5).

H. e. demophon fed on natural and alternative hosts plants

Host plant selection had a significant effect on the cyanide concentration found in Ii. e. demophon (F=10.15; df=3, 50; p<0.001). Butterflies fed P. rubra and P. biflora exhibited significantly lower concentrations of cyanide than butterflies fed P. auriculata and P. trifasciata. Also, the interaction between host plant and sex was statistically significant (F=7.34; df=3, 52; p<0.001). The natural host plant, P. biflora, is associated with lower cyanide concentrations in the male butterflies. Furthermore, cyanide concentrations varied between body parts (F=4.96; df=-l, 52; p=0.03). In general, the body (which in these samples includes the head, thorax and abdomen) possessed higher cyanide concentrations than the wings (Figure 3.6).





40


Cyanide Correlation Between Heliconius erato Fed Natural Host Plants

The regression line in Figure 3.7 indicates that 68% of the variation in cyanide

concentration in Heliconius butterflies is due to differences in the cyanide concentration in the host plants.

Predator-Prey Interaction

At least 10 frogs were collected within the captive H. e. favorinus butterflies' colonies and the feces was analyzed during this study. I found, on average, 4 head capsules and many spines of H. erato in each frog's excreta. The Heliconius head capsules corresponded to the third and fourth instar of development. Usually, I found the frogs camouflaged on the main stem of the plant or on leaves. Occasionally, I found frogs on the cage walls. Frog feces also contained body parts of other insects such as cockroach legs and ant heads.

Discussion and Conclusions

Cyanide Concentrations in Heliconius erato fed Passiflora host plants

Host plant selection was the main reason for the variation in cyanide concentrations between the different races of Heliconius erato. In natural habitats, resource availability and competitive interactions are among the factors that play important roles in this variation. Within each race, genetic variation is probably responsible for differences in the ability to sequester cyanogenic glycosides.

The results of the analysis of the different body parts of H. erato are related to the allocation of the defensive compounds in the butterflies and the mode of action of the chemicals involved in defense. The thorax and the head were the body parts with higher cyanide concentrations. These results did not follow the patterns of cardenolide toxicity reported for monarch butterflies (Brower and Glazier 1975). Since the cyanogenic






41


defensive mechanism requires hydrolysis to function, we expected to find higher concentrations of cyanide release at the interface of the hemolymph containing 3glucosidase and the cuticle. The thorax is a heavily sclerotized body part, filled with hemolymph (Nation 2002). These two properties allow a heavy deposition of the cyanogenic glycosides on the cuticle. By mixing these glycosides with hemolymph containing P-glucosidase, the butterfly is able to generate high concentrations of cyanide. A similar rationale also enables us to understand finding high concentrations of cyanide in the head. The head is small in mass and releases smaller amounts of cyanide, as compared to the thorax, but since the head has a large surface area to volume ratio the volatile cyanide will be released faster. Low cyanide concentrations within the wings and abdomen might be the result of low levels of hemolymph in the former, and low levels of sclerotized area in the latter, which implies a low level of hemolymph containing 3glucosidase in the wings and low levels of cyanogenic glycosides in the cuticle of the abdomen.

Brower et al. (1988) hypothesized that higher cuticular concentrations of

cardenolides may have evolved due to pre-adaptation from biochemical deposition in the developing cuticle, rather than having evolved due to selection from predation or parasitism.

Cyanogenesis in Heliconius butterflies probably functions as an alert signal to reinforce memory of toxicity, also called recall-to-mind mimicry (Rothschild 1984). Under this assumption it is not necessary for selection to favor highly cyanogenic wings, because low levels of HCN will be enough to trigger recall of experiences of the bad taste and smell of HCN and carbonyl compounds, which are formed during cyanogenesis. A






42

naive predator will rapidly learn to avoid feeding on that particular pattern of wing coloration (Chai 1988), and flight (Srygley and Chai 1990) and the members that constitute a Milllerian ring will benefit from it. The selection for belonging to the ring's common pattern will be reinforced. A similar mechanism of avoidance is likely used for the larval stage, however a definitive test was not run. Hyla frogs were more consistently found in the cage of H e. favorinus, than in the cages of H e. cyrbia and H. himera, suggesting that larvae of H. e. favorinus were more palatable. On the other hand, being highly unpalatable has a negative aspect as well, as we observed with the members of H e. cyrbia fed with P. rubra. Members that fed on this particular host plant gained the benefit of being distasteful, but at the same time, there was a high cost paid for this benefit, because individuals feeding on P. rubra spent a longer period feeding on the host, which increases the time they are exposed to parasitoids. On average, individuals feeding on such plants are slightly smaller in size than comparable insects feeding on other host plants.

H. e. favorinus, seems to be better adapted to detoxifying and tolerating different chemicals in the plants, probably due to more efficient degradative enzymatic activity, or higher tolerance of chemicals, for example by having a target-site insensitivity (Brattsten 1979, Lindroth and Weisbrod 1991) because excreta was not found to be cyanogenic. The differences between strategies for coping with cyanogenic compounds probably reflect genetic variation between the races of H. erato.

Lastly, it has been demonstrated that, at least for H. e. cyrbia, there is a limit range of toxicity because of high mortality at the larval stage, implying that there is a level of effectiveness of cyanogenic defenses in the plants in relation to a herbivore. However,






43


there are other factors such as specialization, mode of feeding, etc, that herbivores could also use to cope with plant cyanogenesis.

Temporal Variation in Cyanide in Passiflora Plants

Considerable variation in cyanide concentration was found among Passiflora tested over two seasons (summer and fall). The concentration of cyanide in the fall was almost double the amount found in the summer season. I could not find any literature related to seasonal variation in cyanogenic glycosides. Most of the studies are linked to herbivore presence (Cooper-Driver et al. 1977), or to altered climatic conditions such as temperature (Ellis et al. 1977, Rammani and Jones 1985, and Jones and Rammani 1985), which, it is suggested, contribute to seasonal changes in cyanogenic glycosides levels. However, in this study, none of the above factors is a contributing factor in the differences in cyanide concentration, since the biotic and abiotic conditions were similar. The only environmental factor that might be responsible for the variation is the difference in photoperiod due to seasonal change. A study of seasonal variation in flavonoid concentration performed on temperate-zone Alliaria plants (Haribal and Renwick 2001) produced results similar to those obtained in this study. (Flavonoid concentrations were also lower in the summer and higher in the fall.) Because all of these Passiflora are tropical species where photoperiod varies only 30 minutes during a year from a 12 light: 12 dark cycle, this is a puzzling finding that deserves further investigation. Also, it would be interesting to see how this variation in cyanide concentration correlates to the survival rates of larvae and/or adults of Heliconius species.

Finally, different Passiflora parts were found to possess different cyanide

concentrations. Higher concentrations of cyanide are found in the new growth. The concentration within the plant parts followed the same pattern of alteration between the





44


summer and fall seasons. The result of the distribution of cyanogens within the plant is consistent with the optimal allocation theory (McKey 1974) that states that vulnerable plant tissues and organs are defended more than older senescing ones. It has already been demonstrated elsewhere that for several cyanogenic species, young developing tissues contain the highest amounts of cyanogenic glycosides (Conn 1981 a). Implications of Cyanide Concentration for Palatability

It seems that H e. favorinus is better adapted than H. e. cyrbia to handle different compounds because its larval survival rate is higher, whether feeding on natural or alternative host plants (as has been demonstrated in Chapter 2). H. e. favorinus' superior ability to handle different chemical compounds has both benefits and costs in terms of survival. The survival benefit is reduced larval mortality due to plant toxicity as the larvae is less likely to be killed by the plant's chemical defenses. The survival cost is increased palatability (the larvae may be less toxic), which increases the potential for larval mortality from predation (e.g., by Hyla frogs). H. e. cyrbia, on the other hand, has the advantage, perhaps, of being more toxic, implying better protection against predation, but at the cost of high mortality during larval development. Implications of Cyanide Concentration for Mimicry

Different concentrations of cyanide within a butterfly's body produce a palatability spectrum (Brower et al. 1967). The palatability spectrum theory states that if members of a Millerian mimicry complex are not equally distasteful, then the presence of individuals that are less distasteful would lead to an increase in predation on the models, and this will result in an overall weakening of the Millerian mimicry.

H. erato and its co-mimic H. melpomene are members of a Mtillerian mimicry ring of races in the South and Central America rainforests. There is some disagreement as to






45


whether erato or melpomene is the model for this ring, but Mallet (1999) listed three different reasons that indicate that erato is the model. The first is that erato is more abundant in the areas where it coexists with H. melpomene; second, erato has a broader geographical distribution, and third, erato has a broader habitat than melpomene. If H e. Javorinus in Tingo Maria, Huanuco is less palatable, then this species is in fact a weak model. We should then expect that this will directly affect the presence of the less common H. melpomene. This would be true if P. trifasciata were the only host plant used by H. e. favorinus in the area. However, P. auriculata (which is also the host plant of Heliconius sara), P. laurifolia, P. tricuspis, P. rubra, and P. vespertilio have been recorded in Huanuco (Brako and Zarucchi 1993) and can be potential host plants for H. e. favorinus.

In the last four years, doubts have been raised with regards to the existence of the classical Mtillerian mimic. Instead, some have proposed a new term 'quasi-Batesian mimicry' (Speed and Turner 1999) as an alternative explanation for some types of polymorphism associated with mimicry. Quasi-Batesian mimicry describes Millerian mimics as weakly unpalatable, which benefit from the presence of more unpalatable conspecifics.

In this particular case, members of H. e. favorinus that fed on P. trifasciata would be considered quasi-Batesian mimics, within their race. The males get protection from the presence of more unpalatable conspecifics (females or other conspecifics that did not feed on P. trifasciata). However, the females will be affected, since this sex is probably more exposed to the attack of predators, particularly during oviposition and while searching for host plants.






46






Table 3-1. Heliconius erato fed natural and alternative Passiflora host plant and sample
size used in the study.
Geographical race Host plant N Male Female
H e. cyrbia P. rubra 6 6 P. punctata 7 7 H. e. favorinus P. trifasciata 5 4 H e. demophon P. biflora 5 5 P. trifasciata' 3 3 P. auriculata' 5 5 P. rubra' 2 2 Alternative host plants from other regions




Table 3-2. Passiflora plants tested for cyanogenesis in this study. All plants are used by
Heliconius erato in different geographical areas.
Subgenus Section Passiflora Country of N species Origin Summer Fall Decaloba Decaloba P. biflora Costa Rica 22 12 (Plectostemma)
Decaloba P. misera Brazil 12 6 Decaloba P. punctata Ecuador 6 0 Decaloba P. trifasciata Peru 17 6 Auriculata P. auriculata Ecuador 22 10 Xerogona P. capsularis Brazil 6 14 Xerogona P. rubra Ecuador 12 0






47














p-glucosidase OH NC OH NC + sugar(s)

H O-sugar H OH Cyanogenic glycoside a-hydroxynitrile
Toxic Unstable


a-hydroxynitrile lyase




OH + HCN
H
aldehyde/ketone
Toxic Toxic

Figure 3-1. General process of enzymatic hydrolysis of cyanogenic glycosides.






48



1200 A

1 1000

800 New leaws 600 Old leaws 400 Stem
400

200

0 -
auriculaa biflora capsularis misera pandata rbra trifasciata Passiflora plants









1200

S1000

800
U New leas
c- 600
600 U Old leaws



0 Stem



auriculata biflora capsularis misera trifasciata Passiflora plants Figure 3-2. Cyanide quantification of Passiflora plants in the (A) summer and (B) fall
seasons. Bars indicate standard error.






49






A
600

E500- B

400 B
S40 Female S300 B Male 200

S100

0
Head Thorax Abdomen Wings Body parts

Figure 3-3. Cyanide quantification of male and female H. e. cyrbia fed P. rubra.
Different letters indicate significant differences between means (Tukey HSDtest). Bars indicate standard error.








600 500

t 400 A
U Female ,. 300 B B B
200




Head Thorax Abdomen Wings Body parts

Figure 3-4. Cyanide quantification of male and female H. e. cyrbia fed P. punctata.
Different letters indicate significant differences between means (Tukey HSDtest). Bars indicate standard error.






50





600

S500

400
U Female S300 A

200 B B B

S100
0
Head Thorax Abdomen Wings Body parts

Figure 3-5. Cyanide quantification of male and female H. e. favorinus fed P. trifasciata.
Different letters indicate significant differences between means (Tukey HSDtest). Bars indicate standard error.




600

e 500 S400

'- 300 i

200 100




trifasciata rubra biflora auriculata

Figure 3-6. Cyanide quantification of male and female H. e. demophon fed natural (P.
biflora) and alternative host plants. Bars indicate standard error.






51











1000
o y = 0.8726x- 356.67 H. e. cyrbia a 900 R = 0.6808 & P. rubra
C 800
S 700 H a e. cyrbia
68 oo& P. puntata
z H. e. demophon
) 500 & P. biflora
>Z
0 400
=. 300*300 H.e. fatinus
.0 200 & P. trifesciata
g 100
0
600 800 1000 1200 1400
Mean plant CN concentration (pg CNIg tissue)

Figure 3-7. Regression of the mean cyanide concentration of H. erato over the mean
cyanide concentration of the natural Passiflora host plants upon which they
fed. (The means were used in the figure to enhance the clarity of
presentation.)













CHAPTER 4
IDENTIFICATION OF CYANOGENIC GLYCOSIDES IN PASSIFLORA AND THEIR FATE WHEN INGESTED BY HELICONIUS LARVAE

In the tropical rainforest, the evolutionary relationships between insects and their host plants have driven a variety of interactions and adaptations. Insect pressure has led to the evolution of mechanical and chemical defenses in plants. Herbivores, in turn, have evolved to overcome all of these defenses and are a strong selective force in the evolution and diversification of plant secondary compounds. This evolutionary process, called coevolution, was first introduced by Mode (1958) and was developed by Ehrlich and Raven (1964), who proposed the theory of radiation and escape between plants and butterflies.

A few life history studies have examined the correlations between offspring size, maintenance, and growth, but usually did not relate the resulting integrated life history strategies to toxic secondary compounds in plants. In chapter 2, the life histories of H. erato fed natural host plants vs. alternative host plants were investigated and in chapter 3 it has been shown that butterflies feeding on different host plants generate different concentrations of toxic compounds which are used for their own defense. The compounds found in insect bodies are synthesized de novo, and/or sequestered from plants.

In plants, from which butterflies and moths often sequester toxic secondary

compounds, there are three principal building blocks for these compounds: (1) acetate,

(2) amino acids, and (3) shikimic acids (Trigo 2000). Passiflora plants containing



52






53

cyanogenic glycosides use amino acids as building blocks. These toxic substances take part in a plant's chemical defense against insects that feed upon it.

Cyanogenic glycosides are secondary metabolites that possess intermediate polarity and are water soluble. Chemically, they are defined as O-P-glycosides of ahydroxynitriles (cyanohydrins), biosynthetically derived from amino acids. Cyanogenic glycosides generally co-occur with P-glycosidases, which are specific to the type of cyanogenic glycosides and are spatially separated in the plant or animal tissue to avoid autotoxicity. The enzymatic cleavage that occurs upon disruption of the tissue releases HCN plus sugar, and ketones or aldehydes. It is thought that the high degree of specificity of the glycosidase is due to the structure of the aglycone (Hdsel 1981).

Subspecies of Heliconius erato (Nymphalidae: Heliconiinae) are aposematic

butterflies that have a tight relationship with Passiflora plants. They specialize in the group of Passiflora plants of the subgenus Decaloba. Heliconius erato evidently managed to counteract the effects of the cyanogens in their specific Passiflora host plants, but their mechanism for overcoming these cyanogens are unknown. (See, however the new mechanism of detoxification in Heliconius sara described below.)

Toxic plant compounds would set off many reactions in the insect's digestive system. The mechanisms employed by insects to counteract these chemicals vary between species and can be used solely or in combinations as described below.

Mechanisms Used by Insects to Process Toxic Compounds from Plants

Blum (1983) suggests three ways in which an insect can cope with host allelochemicals: a. Excretion and egestion: Compounds may be eliminated from the insect's body, i.e.,
nicotine is eliminated without any transformation from the body of Trichoplusia ni
and Heliothis virescens (Self et al. 1964).






54


b. Metabolism: Toxic compounds may be converted into less harmful derivatives
through solubilization or conjugation. The metabolites are then reused in the
intermediary metabolic pathway of the organism, or they can be converted to a
metabolic product that can be stored or excreted. The first alternative is common
for toxic compounds that are highly water soluble such as toxic amino acids,
whereas lipid soluble toxins are converted to water soluble metabolites and then
eliminated as such, rather than being recycled (Brattsten 1992).

c. Sequestration: Secondary compounds are deposited intact in particular sites of the
body. The substance to be sequestered is absorbed through the gut membrane,
transported by the hemolymph and stored in body tissues or special glands (Nishida
2002). Sequestration of plant secondary compounds, in general, is used as a
defense against predators and it has been described in many herbivorous insects
such as the Lepidoptera, Coleoptera, Hemiptera and Hymenoptera (Brower 1984,
Bowers 1992, Rothschild 1972). Most of the insects found to sequester toxic
compounds specialize on one or a very few plants, from which they extract their
defensive compounds (Bowers 1992).

Cyanogenic Glycosides in the Family Passifloraceae

The chemistry of Passiflora plants has been studied since the 1960s. Some

Passiflora have been found to contain simple harmane alkaloids (Copeland and Slaytor

1974), flavonoids (McCormick and Mabrick 1981, Ulubelen et al. 1982), a distinctive

triterpene glycoside (Bombardelli et al. 1975), and tannins (Smiley and Wisdom 1985).

However, the Passiflora are unusual in producing cyanogenic glycosides with a

cyclopentene moiety (Spencer and Seigler 1984, 1985a, 1985b, Olafsdottir et al. 1989a,

Jaroszewski et al. 2002).

Chemically, the cyanogenic glycosides in the Passifloraceae have been divided into

four distinct structural classes (Jarozewsky et al. 2002)

a. Type I (called by Spencer 1988, simple cyclopentenoid). These glycosidic
molecules are peculiar because they usually (if not always) occur as pairs of 13 -D
glucopyranosides having enantiomeric aglycones, i.e., tetraphyllin B co-occurs
with volkenin; epivolkenin co-occurs with taraktophyllin; and tetraphyllin A cooccurs with deidaclin (Figure 4-1A).

b. Type II (includes the sulfated cyanogens and complex cyclopentenoid (Spencer
1988)). These glycosides are similar in structure to Type I, but they either have an
unusual sugar molecule, sulfate group, or additional oxygenation of the






55


cyclopentene ring. None of the glycosides in Type II co-occur with another
enantiomer, as it occurs in glycosides of Type I. Also, all the glycosides in Type II have the same stereochemistry in position C-1. The cyanogenic glycoside grouped within Type I and Type II are believed to be biosynthesized from the amino acid 2cyclopentenyl glycine via oxidation, oxidative decarboxylation, dehydration, and/or glycosylation (Conn 198 1b). However, Olafsdottir et al. (1992) do not exclude the
possibility that they originated from oxygenated amino acids (Figure 4-1B). c. Type III (called aliphatic cyanogenic glycoside by Spencer 1988). These
glycosides do not contain a cyclopentene ring. They are derived from the amino
acids valine and isoleucine. This group includes the glycosides linamarin and
lotaustralin, the cyanogens found most often in nature (Figure 4-1C).

d. Type IV. This group of glycosides include the aromatic cyanogens, for example
amygdalin, which is present in seeds of bitter almonds and prunasin found in
Prunus species. The glycosides within this type originated from the amino acid
phenylalanine (See Figure 4-1D).

No other plant group is as diversified in its types of cyanogenic glycosides as the genus Passiflora (Spencer 1988, Jaroszewski et al. 2002).

Cyanogenic Glycosides in Lepidoptera

Unpalatable insects can acquire their defensive toxins either by sequestering plant derived substances or by synthesis de novo (Blum 1981, Rothschild et al. 1970). In butterflies and moths feeding on cyanogenic plants, both acquisition methods have been described.

Synthesis De Novo

De novo synthesis has been described in many species of the family Zygaenidae (Jones et al. 1962, Davis and Nahrstedt 1979), in three subfamilies of the Nymphalidae (Heliconiinae, Nymphalinae and Acraeinae), and in one subfamily of the Lycaenidae (Nahrstedt and Davis 1985).

Based on radiolabeled amino acids, Nahrstedt and Davis (1983, 1985)

demonstrated that larvae and adult Heliconius butterflies synthesize the cyanogenic glycoside themselves, linamarin and lotaustralin from valine and isoleucine, respectively.






56

However, there was an unresolved issue in their studies. The cyanogenic glycoside of larvae and adults increased proportionally to the administration of labeled amino acids; however, within the control, the adult butterflies were also able to synthesize these same cyanogenic glycosides within 14 days after emergence. Nahrstedt and Davis hypothesized that substrates for cyanogenic glycosides biosynthesis can be derived from catabolism of tissue proteins in the absence of dietary supplies of amino acids. During the adult stage, the acquisition of additional amino acids through pollen feeding would provide relevant precursors for the production of additional linamarin and lotaustralin (Gilbert 1991).

The biosynthesis of linamarin and lotaustralin from the amino acids valine and

isoleucine in both plants and insects seems to follow a common process. This process is achieved using a glucosyltransferase enzyme system (Jaroszewski et al. 1988; Olafsdottir et al. 1992).

Sequestration of Cyanogenic Glycosides

Intake of cyanogenic glycosides from a plant has been described for the larvae of the African Acraea horta (Nymphalidae: Acraeinae) which sequesters the cyclopentyl cyanogenic glycoside (gynocardin) from its host plant Kiggelaria africana (Flacortiaceae) (Raubenheimer 1987). Also, Parnassius spp. (Papilionidae), the cotton moth Abraxas spp. (Geometridae), Pryeria sinica (Zygaenidae), and Yponomeuta hexabolous (Yponomeutidae) sequester the cyanoglucoside sarmentosin when fed on their host plants in the families Crassulaceae and Celastraceae (Nishida 1994, Nishida and Rothschild 1995, Nishida et al. 1994). However, the species in the genus Abraxas synthesize sarmentosin even if this compound is not present in any of the plants they fed






57


upon. This fact suggests that these species synthesize their own cyanogenic glycoside, and also sequester sarmentosin (Nishida 1994).

Moreover, it has recently been demonstrated that a cyclopentyl monoglycoside (epivolkenin) was sequestered by Heliconius sara when it fed upon its host plant P. auriculata (Engler et al. 2000). This species converts most of this compound into the corresponding thiol derivative (sarauriculatin), suggesting that the replacement of the nitrile group by a thiol would prevent cyanide release from the host plant and release valuable nitrogen into the insect's primary metabolism. This finding suggested that a unique enzymatic mechanism for dealing with plant cyanogenic glycoside exists in this species (Engler et al. 2000).

Enzymes

Plant cyanogens co-occur with hydrolyzing P-glucosidase enzymes, which are very specific in the type of compound with which they can co-occur (Robinson 1930, Spencer 1987). The aglycone, together with the sugar part of the substrate, play an important role in this specificity, although the aglycone seems to be the main determinant of the specificity (H1-sel 1981).

In this chapter, the fate of cyanogenic glycosides from Passiflora plants after ingestion by Heliconius erato larvae was studied. The chemical compounds found in Passiflora plant will be discussed first, and then the cyanogens found in larval and adult samples. The mechanisms used by H. erato and H. himera species to counteract host allelochemicals will be described. Finally, since amino acids are the precursors of cyanogenic glycosides, I determined the essential amino acids composition derived from larval feeding.






58


Materials and Methods

Butterflies and Plants

This study was conducted at the University of Florida at Gainesville with butterfly colonies that were maintained in Lord & Burnham glass houses (5 x 8 m). Each geographical race was confined in separate cages to avoid hybridization. Wild butterflies were collected and brought from the field, and stock was replenished with new live specimens every six to twelve months. During the first two years of experimentation, the colonies were unable to survive the winters, but improvement in greenhouse environmental settings made winter breeding possible in the last several years of research.

Most of the butterfly stocks were raised and propagated in environmental chambers under controlled temperatures (27oC), humidity (78%) and photoperiods (14L: 10D).

Adult butterfly colonies were provided with natural host species Passiflora plants for oviposition, potted Pentas lanceolata and Lantana camara flowers, sugar water and Psiguria flowers. When Psiguria flowers were not available, a supplemental amino acid solution was sprayed on Tuffy plastic scrubbing pads, whose orange and yellow color attracted the butterflies.

Two races of Heliconius erato were used in this experiment: H. e. favorinus from Peru and H. e. cyrbia from Ecuador. H. himera, a recently named species which has been considered a race of erato, was also included in this study (Figure 4-2). Five Passiflora plants were used in the analysis: P. punctata, P. rubra, P. auriculata, P. trifasciata and P. biflora.






59


Plant and Butterfly Cyanogen Extraction

Fresh plant (approximately 50 mg) and insect samples (in all cases six individuals of each sex were used, except for H. e. favorinus fed P. trifasciata in which twelve individuals of each sex were used) were ground separately in liquid nitrogen. The powdery ground material was extracted with 80% cold aqueous methanol (HPLC grade, Fisher Scientific) and allowed to sit in the solvent overnight in the refrigerator. The extract was then filtered under vacuum through a Btichner funnel, and dried on a rotary evaporator at 400C. The syrup was partitioned between chloroform and water several times to remove lipophilic substances.

A 50-gl aliquot of the aqueous cyanogenic fraction was combined with an equal amount of various P-glucosidase enzymes (see below) in 4.5 x 1.3 cm glass vials. Each vial was corked with a (5 x 1 cm) freshly prepared Feigl-Anger cyanide test strip (Feigl and Anger 1966) and checked for cyanogenesis reaction for 24 h. Color changes in the test strips were ranked as 0, 1, 2 or 3, according to the intensity of blue. Exogenous 1glucosidase was used when the sample was acyanogenic (blue intensity of 0). The aqueous layer was then concentrated to a thick syrup on a rotary evaporator and stored in an ultra freezer at -850C, until further purification could be done. j-Glucosidase Preparations

Fresh plant material was ground to a fine powder in liquid nitrogen in a pestle and mortar and extracted with acetone (HPLC grade, Fisher Scientific). The suspension was filtered with Whatman #1 filter paper under vacuum through a Biichner funnel and rinsed with acetone until the solid residue had lost its color and had dried in the filter. The dry residue was resuspended in pH 6.8 phosphate buffer (0.02M, 500 ml) and stirred at 40C for 1 h and filtered with Whatman #1. The filtrate was dialyzed using a membrane tubing






60


(Spectra/Por Regenerated Cellulose) against pH 6.8 phosphate buffer every 12 h. The buffer was changed 5 to 6 times until the HCN had been removed, as determined by the Feigl-Anger test strip. In order to dehydrate and reduce salt content, the final product was concentrated to 50 ml with Aquacide I (Calbiochem) and the hydrolytic activity was tested with the appropriate standard cyanogens and the Feigl-Anger test.

Because P-glucosidases are specific to cyanogenic glycoside types, the following 3glucosidases were prepared as explained above and used for the detection and general identification of the cyanogens in extracts, fractions of the first purification method and larval pellets: P-glucosidase from P. coriacea was used for identified Type I glycosides, P-glucosidase from Passiflora biflora to test Type II glycosides, and linamarase from linseed for identified Type III glycosides. Two commercially available P-glucosidases were also used: sulfatase and emulsin from almonds (Sigma Chemical Co., St. Louis, Missouri) to test sulfated cyanogens and Type IV glycosides, respectively. However, emulsin is not a very specific enzyme as it can react with other cyanogenic glycoside types (Brimer et al. 1983). For the samples purified by TLC and HPLC, the enzyme 3glucuronidase Type H-1 from Helix pomatia (Sigma Chemical Co., St. Louis, Missouri) was used because this enzyme is not specific and can detect all cyanogenic glycoside types.

Purification of Cyanogenic Glycoside from Plants and Butterflies Purification by liquid chromatography

The thick syrup obtained from the extraction procedure was applied on top of a 3:1 Whatman CF 11 and microcrystalline cellulose column (25 cm x 3 cm) packed with isopropanol-butanol-water (6:3:1). The column was eluted with the same solvent mixture. The column was connected to a peristaltic pump and a fraction collector.






61

Seventy 8-ml fractions were collected. Fractions were monitored by testing 50-tl of each fraction with P-glucosidase enzymatic test with the Feigl-Anger strip to test for cyanogenesis. Appropriate fractions were combined and evaporated in a speed vacuum (Savant radiant cover PC210B, Savant Instruments Inc. Holbrook, NY)). Purification by thin layer chromatography

A second purification method, thin layer chromatography (TLC), was then used. In this case, the right edge of the TLC plate was used as a control and the material to be purified was loaded on the rest of the plate. Merck silica gel 60 TLC plate (60-200 pm) was the stationary phase and ethyl acetate-acetone-chloroform-methanol-water (40:30:12:10:8) constituted the mobile phase. Cyanogenic compounds were monitored by the sandwich method (Brimer et al. 1983) as follows. After development, the portion of the plate containing the cyanogenic glycosides was covered with glass, and the control side of the TLC plate was sprayed with an enzyme solution (aqueous solution of 3glucuronidase) obtained from Sigma. A pre-coated Polygram ion exchange sheet (Polygram ionex 25-SB-Ac) impregnated with three different solutions (a saturated solution of picric acid in water, a 1 M aqueous sodium carbonate solution and 2 % (w/v) ethanolic 1-hexadecanol solution) was used to capture the cyanogenic band overnight. Location of the cyanogenic spot in the control side of the plate allowed me to scrape the purified material on the covered side of the plate from the silica gel. The scraped area ivas placed in a test tube and the cyanogen deabsorbed with methanol-water mixtures (50%), vortexed and centrifuged for 20-30 min. The supernatant was collected, filtered and evaporated in a speed vacuum. Rf values from the control cyanogenic band were calculated.






62


Purification by high performance liquid chromatography (HPLC)

The dry sample from the TLC purification was diluted in 1 ml 20% methanol in water. From this mix, 60 pl was further mixed with 40 ptl methanol and injected into a C18 reverse phase column (Econosil, 10 p.m, 250 x 10 mm, Alltech). The column was maintained at room temperature with a flow rate of 3.5 ml/min. The HPLC column was eluted with 20% methanol in water and the separations were monitored with a differential refractometer connected to an integrator. The individual components were collected and tested for cyanogenesis (Brimer et al. 1983). Identification of Cyanogenic Glycosides from Passiflora Plants and Butterflies

Two methods were used for the identification of the cyanogenic glycosides in plants and butterflies.

Nuclear magnetic resonance (NMR) experimental

All NMR spectra were acquired at the Advanced Magnetic Resonance Imaging and Spectroscopy (AMRIS) facility in the McKnight Brain Institute of the University of Florida. Proton ('H) NMR spectra, as well as two-dimensional 'H /'H -correlation (COSY) spectra and 'H/13C-heteronuclear multiple quantum coherence (HMQC) spectra, were acquired on Bruker Avance spectrometers equipped with 2.5 mm or 5 mm inverse detection (TXI) probes operating at 500 or 600 MHz using standard pulse programs and techniques. Carbon (13C) spectra were acquired with continuous, composite (Waltz-16), proton decoupling on the Bruker Avance-500 spectrometer equipped with a 5 mm broadband observe (BBO) probe operating at 125.8 MHz.

NMR analysis of Passiflora rubra. The cyanogen (4.3 mg) was dissolved in approximately 0.22 ml of methanol-D4, CD30D (Acros Organics, 99.96% atom %-D) and placed in a Wilmad 520-lA, 2.5 mm NMR tube for analysis. The 'H-NMR spectrum





63


(500.4 MHz) of this sample was recorded at 170C, and the chemical shift axis was referenced to internal, residual 'HCD2OD which was assigned to 3.3 ppm. Subsequent addition of tetramethylsilane (TMS) to this solution verified that the assignment of the residual methanol signal at 3.31ppm was valid. The 2D-COSY and 2D-HMQC spectra were acquired with the same sample under the same conditions. The 13C-NMR spectrum (125.8 MHz) of the glycoside was recorded at 270C on a larger sample (7.5 mg), obtained by the addition of more chromatographic material to the original sample, and this required a larger solution volume of 0.60 ml of CD30D, as well as a larger 5mm NMR tube. The chemical shift axis was referenced to the nitrile carbon (CN), which was assigned 120.4 ppm for comparison to the data of Olafsdottir et al. 1989b.

NMR analysis of Passiflora sp. The cyanogenic glycoside (1.2 mg) was dissolved in approximately 0.14 ml of methanol-D4, CD3OD (Aldrich, 99.95% atom %-D) and placed in a Wilmad 520-lA 2.5 mm NMR tube for analysis. The 'H-NMR spectrum (500.4 MHz) of the glycoside was recorded at 220C, and the chemical shift axis was referenced to internal, residual 'HCD2OD which was assigned to 3.3 ppm, as above. The 2D-COSY and 2D-HMQC spectra were acquired with the same sample under the same conditions. The '3C-NMR spectrum (125.8 MHz) of the glycoside was also recorded on the same sample, but at a temperature of 270C. The chemical shift axis was referenced to the nitrile carbon (CN), which was assigned 120.5 ppm for comparison to the data of Olafsdottir et al. 1991.

NMR analysis of Heliconius erato cyrbia fed on Passiflora auriculata. The partially purified cyanogen (about 2.4 mg) was dissolved in approximately 0.4 ml of methanol-D4, CD3OD (Acros Organics, 99.96% atom %-D) and placed in a 5 mm NMR






64


tube for analysis. The 'H-NMR spectrum (600.13 MHz) of the glycoside was originally recorded at 270C and then subsequently at 170C, in order to move the OH resonance of the solvent away from the H-4 resonance of the glycoside. The chemical shift axis was referenced to internal, residual 'HCD20D which was assigned to 3.3 ppm. 2D-COSY spectra were acquired with the same sample at both 270C and 170C. HPLC/ (+) ESI-MS experimental

MS: All mass spectrometric data were acquired with a Thermo-Finnigan (San Jose, CA) LCQ "Classic" quadrupole ion trap mass spectrometer operated in the electrospray ionization (ESI) mode, typically under the following conditions: sheath gas (N2) = 60; aux gas (N2) = 5; spray voltage = 3.3 kV; capillary temperature = 2500C; capillary voltage = 15 V; tube lens offset = 0 V. In order to enhance the sensitivity and obtain dependent MSn (tandem mass spectrometry) data on different components, the (+) ESInormal mass spectra were normally acquired over two ranges: for example, m/z 75-210 for expected amino acids and m/z 205-1200 for higher molecular weight components, including cyanogenic glycosides. Tandem mass spectrometry (MSn, n=2, 3,..) was normally performed in a dependent fashion (i.e., under software control) on the most intense ion from the preceding normal mass spectrum. The MS" spectra were normally acquired with a parent ion isolation of 4-6 u (u=atomic mass unit) and a normalized % CID (collision induced dissociation) energy of 35-40%.

HPLC: High performance liquid chromatography (HPLC) was conducted with an Agilent HPl 100 series binary pump system. Mobile phase A was 0.5% formic acid (Fisher Scientific, 88%; Certified, A.C.S.) and 5 mM ammonium formate (Fisher Scientific, Certified) in water (Fisher Scientific, HPLC grade). Mobile phase B was






65


either (a) 0.5% formic acid in methanol (Burdick & Jackson (Muskegon, MI), B&J Brand High Purity Solvent) or (b) methanol without any modifiers. The column used was a Phenomenex (Torrance, CA) Synergi Hydro-RP column (2 x 150 mm; 4t; 80 A) with the equivalent guard column (2 x 4 mm). A number of different gradients were used to achieve the best separation of all components of the butterfly extracts. The following gradient was used at a mobile phase flow of 0.15 ml/min:

A:B(min) = 100:0(0-5) => 60:40(45) => 0:100(75-120) => 100:0(130-165)

UV: During some of the initial analyses, an Applied Biosystems (Foster City, CA) Model 785A, programmable absorbance detector was interfaced between the column effluent and the mass spectrometer. The absorbance at 254 nm was monitored. After the initial analyses, the UV was no longer used and the column effluent was connected directly to the mass spectrometer.

Post-column mobile phase modification: The efficiency of electrospray

ionization was low when spraying 100% aqueous mobile phase. Often, methanol or 0.5% formic acid in methanol was added to the column effluent via a PEEK tee-union. The post-column modifier was either provided by Applied Biosystems model 400 solvent delivery system at 0.1 ml/min or via the LCQ's syringe pump at 20 Pl/min. The latter was found to be more reproducible, as the Applied Biosystems pump did not provide steady flow at 0.1 ml/min and under.

Amino Acid Analysis

The following standard amino acids were used for the amino acid analysis:

isoleucine, leucine, phenylalanine, tryptophan, tyrosine and valine. The amino acids were combined with linamarin, which was used as a marker, because the retention time






66


was known. The analysis was performed by HPLC/MS under the same conditions described above.

Results

Cyanogenic Glycosides in Passiflora Plants

The P-glucosidases classification test of plant extracts revealed that all Passiflora species tested possess cyanogenic glycoside Type II, with the exception of P. auriculata, which possess cyanogenic glycoside Type I. Interestingly, P. rubra reacted with all Pglucosidases. This means that the aglycone in P. rubra was not specific to any Pglucosidase. Emulsin was also able to hydrolyze the cyanogenic glycoside from P. trifasciata and P. auriculata. Cyanogenesis occurred very rapidly in the extracts of P. rubra with high release of cyanide within 30 minutes of being tested, followed by P. punctata and P. auriculata (Table 4-1). Cyanogenesis occurred within one hour in P. biflora and within two hours in P. trifasciata. Cyanide release then increased over time. Identification of cyclopentanoid cyanogenic glycoside in Passiflora rubra

The chemistry of this plant has not been analyzed before. One peak was isolated from Passiflora rubra by HPLC and tested for cyanogenesis by TLC (Rf = 0.2-0.37). The molecular weight was verified by HPLC-MS (MW 417).

'H NMR spectral data of the cyanogen in P. rubra showed chemical shifts and

signal splitting similar to those reported for passicapsin in CD30D (Table 4-2). '3C NMR (Table 4-3) also showed characteristic peaks for passicapsin (Olafsdottir et al. 1989b). It was not clear whether the resonances for H2, H3, C2, and C3 had been assigned in Olafsdottir's data. Those 'H assignments are given in Table 4-2 by analogy to the data in Table-4-4, and the coupling constants for these resonances appear consistent, too. The






67


corresponding 13C resonances (Table 4-3) were assigned by way of a 2D-HMQC spectrum (Figure 4-4).

Identification of cyclopentanoid cyanogenic glycoside in Passiflora sp.

Although this Passiflora species is reported throughout this dissertation as P.

trifasciata, it is actually an undescribed sister species of P. tricuspis and P. trifasciata (McDougal personal communication).

A peak was isolated from Passiflora sp. by HPLC and tested for cyanogenesis by TLC (Rf= 0.16-0.24). The molecular weight was verified by HPLC-MS (MW 433). 'H NMR spectral of this species showed the coupling constants (Hz) and chemical shifts (ppm) listed in Table 4-4. '3C NMR also showed characteristic peaks for passitrifasciatin (Table 4-5) (Olafsdottir et al. 1991). The resonances for C2, C3, Cl', and Cl" had been assigned in Table 4-5 by correlations with their corresponding 'H resonances (Table 4-4), which were assigned by Olafsdottir et al. 1991; this was accomplished using the 2DHMQC spectrum (Figure 4-5).

Cyanogen profile of Passiflora auriculata

Two peaks were isolated from P. auriculata by HPLC and tested for

cyanogenesis by TLC, which resulted in a Rf, (0.11, 0.2) and Rf2 (0.28, 0.37). The molecular weight for both was verified by HPLC-MS (MW 287). Fraction #1 was identified by HPLC/MS as epivolkenin MW=287, eluting at 17.76 min. Fraction #2 was identified as taraktophyllin MW=287 and eluted at 2.73 min. Cyanogen profile of Passiflora biflora

P. biflora contained passibiflorin as described by Olafsdottir et al. (1989b). The Rf values obtained by TLC were 0.16-0.23 and the MW=433.





68


Cyanogen profile of Passiflora punctata

This plant is currently being analyzed by Dr J. Jaroszewsky (personal

communication). However, from the HPLC/MS analysis, we know that this plant contains passibiflorin with a MW of 433 and a retention time of 34.00 min. Cyanogenic Glycosides of Butterflies Fed Natural Host Plants

A P-glucosidase test indicated that there were no differences in cyanogenic

glycoside types between the sexes in either of the two races. All butterflies possessed aliphatic cyanogenic glycoside, except for H. e. favorinus that tested acyanogenic when fed on P. rubra and P trifasciata. Based on the results presented in Chapter 3, it is known that H. e. favorinus contains low amounts of cyanogenic glycosides. Aliphatic and Type I glycosides were present when males and females from the two races were fed P. auriculata. Cyanogenesis occurred within 30 minutes in extracts of H. e. favorinus fed P. punctata and P. auriculata, while cyanogenesis reaction occurred within one hour in almost all tested butterflies samples (Table 4-6).

No glycosides were found in larval fecal pellets, suggesting that the ingested

glycosides were absorbed by the larvae, and, if excreted, metabolized prior to excretion.

P-glucosidase classification tests were not able to detect small quantities of

aliphatic and Type II glycosides in butterfly samples, but these were later detected by the HPLC/MS method. This suggests that the P-glucosidase test is not conclusive and it should be used only to detect high concentrations of cyanogenic glycosides. De novo synthesis

HPLC/MS showed that all butterflies contained aliphatic cyanogenic glycosides

(linamarin and lotaustralin), even the ones that tested acyanogenic with the -glucosidase enzymatic test (Table 4-6, 4-7 and 4.8). There were variations among the aliphatic






69


compounds found in butterfly samples that were related to host plant feeding. When butterflies were fed plants containing monoglycoside cyclopentenoids, the percentage of lotaustralin was very low. Also, H. e. cyrbia and H. e. favorinus have a similar percent ratio of linamarin:lotaustralin, while H. himera generally have a higher ratio of linamarin to lotaustralin.

Sequestration of cyanogenic glycosides

Butterflies developing from larvae fed P. auriculata sequestered monoglycoside cyclopentenoids. 'H NMR analysis showed that the female H. e. cyrbia contained epivolkenin (Table 4-9), linamarin, and another monoglycoside (perhaps taraktophyllin, based on H1' and H4). The male sample also contained epivolkenin, linamarin and another glycoside (perhaps taraktophyllin, based on HI' and H4). Sarauriculatin was not found in the samples. The results of the HPLC/MS (Table 4-7 and 4-8) showed the percent values of each compound. However, taraktophyllin was not distinguishable from epivolkenin, under the conditions in which this analysis was conducted.

Small percentages of the cyanogen epivolkenin, passicapsin and MW 433

(passibiflorin or passitrifasciatin) were found when butterflies were fed diglycoside cyclopentenoids, demonstrating both sequestration and metabolism of this type of cyanogenic compound (Table 4-7 and 4-8).

The entire butterfly extracts for the cyanogenic and amino acids analysis had a blank in between each sample in order to avoid carry-over. Amino Acids Analysis in Butterfly Extracts

The amino acids valine, isoleucine/leucine, tyrosine, phenylalanine, and tryptophan, were measured in butterfly samples (Table 4-10 and 4-11).






70

All butterflies fed plants containing diglycoside cyclopentenoids (Type II) had levels of valine greater than 13 %, except for H e. cyrbia fed P. rubra. Unfortunately, given the conditions used in the HPLC/MS analysis, leucine and isoleucine could not be separated. However, the combined percentages of these amino acids were always very high, especially when H e. cyrbia was fed P. rubra. A further amino acid analysis will be run under different conditions in order to separate leucine and isoleucine.

Another interesting result from the amino acid analysis is the low percentage of valine in H. e. cyrbia fed on auriculata (in both male and female samples), and the presence of high percentages of tyrosine and tryptophan in the female, but not in the male. Phenylalanine, on the other hand, was present in the males but not in the female samples. Also, high percentages of phenylalanine were found in H. e. cyrbia fed biflora, H. e. favorinus fed trifasciata and H himera fed punctata and rubra. Thiol Analysis

As mentioned above, sarauriculatin was not found in samples of H. erato fed P. auriculata. Also. thiol and sulfates were not found in these samples, demonstrating that these may not be major paths for metabolism of cyanogenic glycoside by H. erato fed P. auriculata.

Discussion and Conclusions

Sequestration, De Novo Synthesis and Metabolism of Cyanogenic Glycosides

Although sequestration of simple cyclopentenoid cyanogens by H. erato fed P.

auriculata was reported based on a P-glycosidase enzymatic test (Engler et al. 2000), the compounds were not identified. H. e. cyrbia fed P. auriculata possessed linamarin, lotaustralin, epivolkenin and, possibly, taraktophylin. This butterfly did not metabolize sarauriculatin as reported for H sara by Engler et al. (2000).






71

Also, H. erato and H. himera butterflies sequestered small amounts of complex

cyclopentenoids (Type II) from host plants and contained various products of metabolism of these cyanogenic glycosides. The presence of epivolkenin can be explained as a hydrolysis product of both passibiflorin and passicapsin, as shown in Figures 4-6 and 4-7. Since passibiflorin and passicapsin contain the epivolkenin structure, the two former compounds can be hydrolyzed to form the latter (Fisher et al. 1982, Olafsdottir et al. 1989b).

Passicapsin was also found in butterflies fed on plants containing passibiflorin and passitrifasciatin. Perhaps the MW 433 cyanogenic glycoside underwent a loss of an O to yield this compound as follows: R-OH ===--> RH, difference = 16 units. Notice that this is the only difference between passibiflorin and passicapsin.

Sequestration of secondary compounds has been described in many insects, including the Lepidoptera. For example, Junonia coenia (Nymphalidae) sequesters iridoid glycosides from Plantago lanceolata (Bowers and Stamp 1997), Ideopsis similes (Nymphalidae: Danainae) sequesters phenanthroindolizidine alkaloids from Tilophora tanakae, an asclepiad plant (Abe et al. 2001), and the lycaenids Eumaeus and Taenaris spp. take in cycasin from cycads (Bowers and Larin 1989, also see Nishida 2002). On the other hand, the cases of mixed strategies combining de novo synthesis and sequestration are scarce, as mentioned in the introduction. In this study, it was found that when simple and complex cyclopentenoid glycosides are sequestered by the butterflies, linamarin and lotaustralin were always present. However, lotaustralin was present in small amounts when the larvae were fed host plant containing simple cyclopentenoids. In general, a mixed strategy (de novo synthesis and sequestration) may increase





72

distastefulness of Heliconius to predators. This discovery could have important consequences for the theory of defense mechanisms in these insects.

As was discussed in the introduction, sequestration has been associated with defense against natural predators (Brower 1984). If an organism both sequesters and synthesizes its own toxic compound, it is expected that the individual will be better protected against predation, but only if the quantities of the stored and metabolized compound are high.

Although metabolic processes similar to those in H. sara may occur in H. erato, I discard the possiblility of the formation of sarauriculatin because no evidence of thiols and/or sulfates was found in H erato fed P. auriculata. Amino Acids and their Importance For Heliconius Biology

Since none of the butterflies in the experiment was ever fed on Psiguria or other pollen containing flowers as an adult, I conclude that all of the reported amino acids are essential amino acids that are generated from larval feeding. The percentage of amino acids found were related to the larval host plant. For example, butterflies that fed as larvae on diglycoside cyclopentenoid-containing plants gathered higher percentages of valine and leucine/isoleucine, while butterflies fed as larvae on plants containing monoglycoside cyclopentenoids produced higher percentages of leucine/isoleucine and tyrosine in the males. Nahrstedt and Davis (1985) illustrated the proportional increases of linamarin and lotaustralin related to their amino acid precursors valine and isoleucine; however, they were unable to explain how their control butterflies were able to synthesize linamarin within 14 days after emergence, despite the absence of dietary valine. Here, we demonstrate that butterflies are potentially able to synthesize linamarin from amino acids stored in their bodies during their larval development. Furthermore, butterfly






73

ecologists recognize the importance that pollen feeding has for Heliconius butterflies (Gilbert 1991). Pollen feeding is correlated not only with unpalatability (Cardoso, umpublished), but also with nutrition for adult somatic maintenance and extended reproductive longevity (Dunlap-Pianka et al. 1977, Dunlap-Pianka 1979), and nuptial gifts (Boggs and Gilbert 1979). Thus, in the Heliconius system, fourteen days are spent searching for Psiguria flowers after emergence from the puparium. Adult butterflies keep synthesizing the aliphatic glycosides from the amino acids gained during the larval stage, until they are able to collect new amino acids via pollen feeding. Therefore, the ability of teneral adults to invest in reproduction and defense is constrained by the resources that the individual butterfly accumulates as a larva.

Phenylalanine was found in high percentages in adult H. erato fed P. biflora or P. trifasciata as larvae and in H. himera fed P. punctata and P. rubra. Tyrosine, on the other hand, was found in males of H. e. cyrbia fed P. auriculata as larvae. Tyrosine and phenylalanine are aromatic amino acids which are important for insect development and reproduction. Phenylalanine is an essential amino acid for all insects, whereas tyrosine is synthesized from phenylalanine by hydroxylation (Gilmour 1965). Tyrosine plays an important role in morphogenetic processes and protein synthesis and, with phenylalanine, it is involved in the hardening and tanning process of newly formed cuticle (Zografou et al. 2001). Furthermore, tyrosine and tryptophan are precursors of the pigments ommochrome and melanin, respectively (Nijhout 1991).

In plants, there is evidence suggesting that cyanogenic compounds may also

function in the metabolism and transport of nitrogen to form proteins (Selmar et al. 1988, Nahrstedt 1992). For example, cyanide is not released from the seeds of the rubber tree,






74


Hevea brasiliensis, during germination due to the cyanide detoxification enzyme 3cyanoalanine synthase and hydrase, which cause the transport of nitrogen to be used in protein synthesis (Selmar et al. 1998). If Heliconius is able to sequester cyanogenic glycoside, this means that there must be a way to cope with cyanide release. This could be achieved with an enzymatic mechanism. In H. sara, for example, the cyanogenic glycoside is metabolized within the insect, by means of an unknown mechanism, and the nitrogen is recovered and used in protein synthesis (Engler et al. 2000). If metabolism of sequestered compounds is useful for nitrogen metabolism and reallocation for nutritional needs, Heliconius is gaining in the coevolutionary race, because butterflies are able to gain nitrogen for their primary metabolic needs via nutrition from the plant, and can also use it for defense. A similar mechanism, with other compounds, has been described in the leaf beetle, Chrysomela confluens (Kearsley and Witman 1992).

This study investigated the basis of the importance of cyanogenic glycosides in relation to amino acids. Future experiments could proceed in two ways, the first using radiolabeled amino acids to reveal the pathway of cyanide derived from sequestered cyanogens in Heliconius butterflies, and the second using HPLC/MS to learn the contribution of different cyanogenic glycosides towards amino acid composition in the adult butterfly.






75




Table 4-1. Qualitative analysis of the extracts of Passiflora plants. Number 0=no cyanide
release, 1=light blue, slightly cyanogenic, 2=blue, cyanogen and 3=dark blue,
very cyanogenic.
Passiflora Less 30min 1 h 2 h 3 h 4 h 5 h 12 h 24 h
than 30
min
auriculata 0 1 1 1 1 2 3 2 2 biflora 0 0 1 1 1 1 2 2 2 punctata 0 1 1 1 2 2 3 2 2 rubra 2 2 2 3 3 3 3 3 3 trifasciata 0 0 0 1 1 1 2 2 2








Table 4-2. 'H NMR(a) spectrum of a cyanogen, identified as passicapsin, from P. rubra in CD3OD. Chemical shifts (ppm) and
coupling constants (Hz). Degree of splitting is identified by: (d) doublet, (q) quartet, (m) multiple, r) broad.
Cyanogen H2 H3 H4 H5A & HI' H2'-H5' H6'A & HI" H2"A & H2"B H3" H4" H5" CH3 H5B H6'B
Multiplicity dd dd m dd d m,m,m,m dd, dd dd dt, ddd br, q dq d Passicapsinb)
8 ppm 6.12 6.36 4.87 2.34, 3.04 4.59 3.33-3.51 3.67 4.89 1.70 3.96 3.20 4.00 1.25
3.86 1.83
J values 1.3,5.5 2.1,5.5 4.7,14.8 7.7 c 5.2,12.0 2.5,9.5 2.5,13.5 3.5 1.5,6.6 6.6
7.1,14.8 2.0,12.0 3.2,9.5,13.5 P. rubra
8 ppm 6.14 6.37 4.88 2.34,3.05 4.58 3.22-3.39 3.67 4.89 1.71 3.96 3.20 3.99 1.25
3.86 1.82
J values 1.0,5.5 2.0,5.5 4.7,14.8 7.8 c 5.3,12.0 2 ,9.5 2.1,13.6 3.2 1.1,6.6 6.6
7.1,14.8 2.0,12.0 3.1,9.9,13.3
,a) 500.4 MHz; see Figure 4-1 for a numbered structure.
(b) Olafsdottir et al. 1989b. "N
(c) Too complex for J-analysis.









Table 4-3. 3C NMR(a) spectrum of a cyanogen, identified as passicapsin, from P. rubra in CD30D. Cyanogen Cl C2 C3 C4 C5 CN Cl' C1" Remaining resonances Passicapsin(b) 82.1 132.7 142.4 81.7 46.9 120.4 101.0 99.7 17.2, 34.9, 62.9, 70.6 (two), 71.3, 71.7. 74.9, 78.3, 78.4
P. rubra() 82.0 132.6 142.3 81.6 46.5 120.4 101.0 99.6 17.2, 34.9, 62.8, 70.6, 70.5, 71.3, 71.6. 74.9, 78.2, 78.4
(a) 125.8 MHz; see Figure 4-1 for a numbered structure.
(h) Olafsdottir et al. 1989b.
(c) Nitrile carbon (CN) assigned 120.4 ppm for comparison to spectrum (b).








Table 4-4. 'H NMR(a) spectrum of a cyanogen, identified as passitrifasciatin, from Passiflora sp.(referred throughout this dissertation
as P. trifasciata) in CD30D. Chemical shifts (ppm) and coupling constants (Hz). Degree of splitting is identified by: (d)
doublet, (t) triplet, (m) multiplet.
Cyanogen H2 H3 H4 H5A&H5B HI' H2' H3'- H5' H6'A& HI" H2" H3" H4" H5" CH3 H6'B
Multiplicity dd dd dd, dd d dd m,m,m dd, dd d m t dd dq d Passitrifasciatin(b)
6 ppm 6.16 6.36 4.86 2.49,3.0 4.61 3.21 3.25-3.40 3.66,3.85 4.69 3.25- 4.0 3.15 3.75 1.25
3.40
J values 5.5,1.0 5.5,2.1 15.0, 4.0 7.7 7.7,9.0 c 12.1,5.1 8.0 c 9.5,2.8 9.5,6.3 6.3
15.0, 7.0 12.1,1.9 Passiflora sp.
6 ppm 6.17 6.36 e 2.50,3.01 4.62 3.22 3.25-3.40 3.67,3.85 4.7 3.25- 4.01 3.16 3.76 1.26
3.40
J values 5.5,1.0 5.5,2.1 15.0, 4.1 7.8 7.9,9.0 c 12.0,5.1 8.0 c 2.9 9.5,2.8 9.5,6.2 6.2
15.0, 7.1 12.1,1.7
(a) 500.4 MHz; see Figure 4-1 for a numbered structure.
(b) Olafsdottir et al. 1991.
(c) Too complex for J-analysis.
(e) Obscured by OH resonance.










Table 4-5. "C NMRa) spectrum of a cyanogen, identified as passitrifasciatin, from Passiflora sp.(referred throughout this dissertation
as P. trifasciata) in CD30D.
Cyanogen Cl C2 C3 C4 C5 CN C1' Cl" C6' CH3 Remaining resonances Passitrifasciatin(b) 82.1 133 142 82.4 46.2 120.5 100.8 101.5 62.8 18.5 71.0, 71.6, 72.7, 73.0, 74.5, 74.9, 78.2, 78.3
Passiflora sp. C) 82.1 133 142 82.4 46.2 120.5 100.8 101.5 62.8 18.4 71.0, 71.7, 72.7, 73.1, 74.6, 74.9, 78.2, 78.4
(a) 125.8 MHz; see Figure 4-1 for a numbered structure.
(b) Olafsdottir et al. 1991.
(c) Nitrile carbon (CN) assigned 120.5 ppm for comparison to spectrum (b).





80


Table 4-6. Qualitative analysis of cyanide release of whole body extracts tested with
different 3-glucosidases by Feigl-Anger strips monitored over time in H. e.
favorinus and H. e. cyrbia. Numbers indicate, 0=no cyanide release, l=light
blue, slightly cyanogenic, 2=blue, cyanogen and 3=dark blue, very
cyanogenic.
N 30min 1 h 2h 3 h 4h 5h 12 h 24 h H. e.
favorinus
Spunctata 6 0 0 1 1 1 1 0 0 0punctata 6 1 2 2 2 2 1 0 0 c rubra 6 0 0 0 0 0 0 0 0 I rubra 6 0 0 0 0 0 0 0 0 c auriculata 6 0 1 1 2 2 2 1 1 9 auriculata 6 1 1 1 1 1 1 1 1 Strifasciata 6 0 0 0 0 0 0 0 0 9 trifasciata 6 0 0 1 0 0 0 0 0 biflora 6 0 0 2 2 2 2 1 1 9 biflora 6 0 0 1 1 1 1 0 0 H. e. cyrbia
d punctata 6 0 3 3 3 3 2 2 1 9 punctata 6 0 3 3 3 3 2 2 1 d rubra 6 0 3 3 3 3 3 2 1 o rubra 6 0 1 2 2 2 2 2 1 d auriculata 6 0 1 2 3 3 2 1 1 ?auriculata 6 0 1 2 2 2 3 2 1 c trifasciata 6 0 1 2 3 3 3 2 1 Y trifasciata 6 0 1 2 3 3 3 2 1 Sbiflora 6 0 1 2 2 2 2 2 1 Y biflora 6 0 1 2 2 2 2 2 1










Table 4-7. Peak areas counts of cyanogenic glycosides detected by HPLC/MS in tH. erato and H. himera butterflies.
Peak area of cyanogenic glycosides x 106
Butterfly sample Total
Linamarin Lotaustralin Epivolkenin Passicapsin MW 433 Cyanogenic glycoside
SH. e. cyrbia fed auriculata 307.70 4.59 681.24 0 0 993.54 ( H. e. cyrbia fed auriculata 396.37 3.25 650.00 0.10 0 182.91 SH. e. cyrbia fed punctata 560.98 573.63 4.41 0.67 1.83 1141.51 SH. e. cyrbia fed rubra 69.62 66.64 12.21 0.49 0.54 149.51 SH. e. cyrbia fed biflora 766.52 693.98 2.60 0.26 0.23 1463.59 YH. e. favorinus fed trifasciata 117.53 110.10 12.12 0.74 5.90 246.39 H. himera fed punctata 1106.17 536.34 18.85 0.60 0.40 1662.35 H. himera fed rubra 271.24 198.94 22.47 0.45 0 493.11


Table 4-8. Percent of cyanogenic glycosides detected by HPLC/MS in H. erato and H. himera butterflies.
Percent total peak area of cyanogenic glycosides
Butterfly sample Total
Linamarin Lotaustralin Epivolkenin Passicapsin MW 433 Cyanogenic glycoside
Y H. e. cyrbia fed auriculata 30.97 0.46 68.57 0.00 0.00 100.00 d H e. cyrbia fed auriculata 37.76 0.31 61.92 0.01 0.00 100.00 SH. e. cyrbia fed punctata 49.14 50.25 0.39 0.06 0.16 100.00 9 H. e. cyrbia fed rubra 46.57 44.57 8.17 0.31 0.36 100.00 SH. e. cyrbia fed biflora 52.37 47.42 0.18 0.02 0.02 100.00 HI. e. favorinus fed trifasciata 47.70 44.68 4.92 0.30 2.40 100.00 YH. himera fed punctata 66.54 32.26 1.13 0.04 0.02 100.00 YH. himera fed rubra 55.01 40.34 4.56 0.09 0.00 100.00








Table 4-9. 'H NMRa) spectrum in CD30D at 27'C of a cyanogen, identified as epivolkenin, from female H. e. cyrbia fed on P.
auriculata. Chemical shifts (ppm), coupling constants (Hz) and multiplicities. Multiplicity is identified by: (d) doublet,
(m) mul tiplet.
Cyanogen H2 H3 H4 H5A H5B HI' H2' H3'-H5' H6A' H6B' Multiplicity dd dd m dd dd d dd m,m,m dd dd Epivolkenin-b)
8 ppm 6.14 6.27 4.80 2.25 3.02 4.62 3.22 3.3-3.4 3.68 3.86
J values 5.5,1.3 5.5,2.0 14.6,4.8 14.6,7.1 7.7 9.1,7.8 c 12,5.5 12,2
H. e. cyrbia
6 ppm 6.13 6.26 ( 2.25 3.01 (3.02) 4.62 3.22 3.3-3.4
(6.14) (6.26) (4.81) (2.25) (4.62) (3.85) J values 5.5,1.3 5.5,2.0 14.6,4.8 14.6,7.1 7.7 9.1,7.8 c 12,2
(a) 600.13 MHz; see Figure 4-1 for a numbered structure.
(b) From Jaroszewski et al. 1987a, b.
(c) Too complex for J-analysis.








'Table 4-10. Peak areas of measured free amino acids in butterfly samples.
Areas of detected amino acids
Butterfly sample Valine Leu/IsoLeu() Tyrosine Phenylalanine Tryptophan Total amino acids
H. e. cyrbiafed auriculata 2079517 39047401 15421636 298840 9648436 66495830 d H. e. cyrbia fed auriculata 208897 24267233 0 2623200 0 27099330 SH. e. cyrbia fedpunctata 117372033 166137488 7019235 50572236 29184233 370285225 H. e. cyrbia fed rubrar 8207693 190182200 218414 3928763 0 202537070 H. e. cyrbia fed flora 62722048 266334406 18532600 142270940 8758244 498618238 H. e.favorinus fedtrasciata 27204475 91750748 2622139 35121022 13495640 170194024 YH. himera fed punctata 256841130 517108138 7361421 535650216 10175322 1327136227 LH. himera fed rubra 215165754 456333019 27780824 206786183 14364120 920429900
Cannot be distinguished under the given conditions.








-Table 4-11. Percent total peak areas of measured free amino acids in butterfly samples.
% Total peak area of measured amino acids
Butterfly sample Valine Leu/IsoLeu) Tyrosine Phenylalanine Tryptophan Total amino acids
H. e. cyrbia fed auriculata 3.13 58.72 23.19 0.45 14.51 100.00 g H. e. cyrbia fed auriculata 0.77 89.55 0.00 9.68 0.00 100.00
H. e. cyrbia fed punctata 31.70 44.87 1.90 13.66 7.88 100.00 H. e. cyrbia fed rubra- 4.05 93.90 0.11 1.94 0.00 100.00 H. e. cyrbia fed biflora 12.58 53.41 3.72 28.53 1.76 100.00 2H. e.favorinus fed trifasciata 15.98 53.91 1.54 20.64 7.93 100.00 H. himera fed punctata 19.35 38.96 0.55 40.36 0.77 100.00 ,H. himera fed rubra 23.38 49.58 3.02 22.47 1.56 100.00
Cannot be distinguished under the given conditions.


00
oP.




Full Text
9
Living Insects and Host Plants
Two subspecies were maintained in temperature controlled Lord & Burnham glass
houses (5 x 8 m) at the University of Florida, Department of Entomology and
Nematology. H. e. favorinus were the descendents of 60 individuals collected at Tingo
Maria, Huanuco, Peru, which is located in the valley of the upper Huallaga river, at the
base of the western slopes of the eastern chain of the Andes (locally known as "Cordillera
Azul"). H. e cyrbia were the descendents of 50 individuals collected at Pichincha,
Ecuador, located on the western chain of the Andes. The two subspecies were kept in
separate cages to avoid hybridization. Butterflies in each colony were provided with their
natural host plants. Pass (flora rubra, P. punctata, and P. aur ¡culata from Ecuador are
the natural host plants of H. e. cyrbia, while P. trifasciata is one of the natural host plants
of H. e. favorinus. Although this Passiflora species is reported here as P. trifasciata, it is
actually an undescribed sister species of P. tricuspis and P. trifasciata (McDougal
personal communication).
P. biflora from Costa Rica is an alternative host plant for both subspecies. Lantana
and Pentas plants were used as adult nectar sources, while Psiguria spp. flowers were
used for pollen feeding.
Cups containing 20% sugar-water and red and orange Tuffy" sponges sprayed
with sugar water and an adult artificial diet of an amino acid solution that simulates the
amino acid content of Psiguria flowers (Gilbert unpublished) were also provided to
ensure an adequate food supply to the adults.
Feeding Experiments
Seventeen H. e. favorinus females were used for the host plant feeding experiment.
Iwelve H. e. cyrbia females were used to represent the Ecuadorian race. The colony was


CHAPTER 4
IDENTIFICATION OF CYANOGENIC GLYCOSIDES IN PASSIFLORA AND THEIR
FATE WHEN INGESTED BY HELICONIUS LARVAE
In the tropical rainforest, the evolutionary relationships between insects and their
host plants have driven a variety of interactions and adaptations. Insect pressure has led
to the evolution of mechanical and chemical defenses in plants. Herbivores, in turn, have
evolved to overcome all of these defenses and are a strong selective force in the evolution
and diversification of plant secondary compounds. This evolutionary process, called
coevolution, was first introduced by Mode (1958) and was developed by Ehrlich and
Raven (1964), who proposed the theory of radiation and escape between plants and
butterflies.
A few life history studies have examined the correlations between offspring size,
maintenance, and growth, but usually did not relate the resulting integrated life history
strategies to toxic secondary compounds in plants. In chapter 2, the life histories of H.
erato fed natural host plants vs. alternative host plants were investigated and in chapter 3
it has been shown that butterflies feeding on different host plants generate different
concentrations of toxic compounds which are used for their own defense. The
compounds found in insect bodies are synthesized de novo, and/or sequestered from
plants.
In plants, from which butterflies and moths often sequester toxic secondary
compounds, there are three principal building blocks for these compounds: (1) acetate,
(2) amino acids, and (3) shikimic acids (Trigo 2000). Passiflora plants containing
52


I certify that I have read this study and that in my opinion it conforms to acceptable
standards of scholarly presentation and is fully adequate, in scope and quality, as a
dissertation for the degree of Doctor of Philosophy.
Thomas C. Emmel, Chair
Professor of Entomology and Nematology
I certify that 1 have read this study and that in my opinion it conforms to acceptable
standards of scholarly presentation and is fully adequate, in scope and quality, as a
dissertation for the degree of Doctor of Phil
tames L. Nation
Professor of Entomology and Nematology
1 certify that I have read this study and that in my opinion it conforms to acceptable
standards of scholarly presentation and is fully adequate, in scope and quality, as a
dissertation for the degree of Doctor of Philosophy.
Heather J. McAuslane
Associate Professor of Entomology and
Nematology
I certify that I have read this study and that in my opinion it conforms to acceptable
standards of scholarly presentation and is fully adequate, in scope and quality, as a
dissertation for the degree of Doctor of Philosophy.
Frank Slansky, Jr.
Professor of Entomology and Nematology
I certify that I have read this study and that in my opinion it conforms to acceptable
standards of scholarly presentation and is fully adequate, in scope and quality, as a
dissertation for the degree of Doctor of Philosophy.
David A. Jones
Professor of Botany


Cyanogen
Cl
C2
j
C3
C4
C5
CN
Cl'
Cl"
Remaining resonances
Passicapsim
82.1
132.7
142.4
81.7
46.9
120.4
101.0
99.7
17.2, 34.9, 62.9, 70.6 (two), 71.3, 71.7. 74.9, 78.3, 78.4
P. rubra(c)
82.0
132.6
142.3
81.6
46.5
120.4
101.0
99.6
17.2, 34.9, 62.8, 70.6, 70.5, 71.3, 71.6. 74.9, 78.2, 78.4
Ta' 125.8 MHz; see Figure 4-1 for a numbered structure.
ic) Nitrile carbon (CN) assigned 120.4 ppm for comparison to spectrum (b).


35
All Passiflora plants were subjected to the same environmental conditions
(humidity, temperature and natural photoperiod) in the greenhouses. None of the plants
tested was exposed to herbivores.
Quantitative Determination of Cyanide
Colorimetric analysis
Cyanogenesis was quantified by the Lambert procedure (Lambert et al. 1975)
which was modified by Brinkler and Seigler (1989, 1992).
The plant or butterfly samples were ground to a fine powder in liquid nitrogen.
The powdered plant or butterfly tissue was poured into 10 ml vials containing 0.1 M
phosphate buffer (pH 6.8). The volume of liquid was kept to a minimum, as cyanide is
very soluble in water. A 5 ml vial containing fresh 1M NaOH solution was then placed
inside the 10 ml vial and the 10 ml vial was stoppered tightly. The samples were
prepared quickly to avoid loss of HCN. The vials were then incubated at 37C overnight.
During the incubation, HCN released from tissue is trapped in the NaOH solution, and
forms NaCN.
After the incubation, aliquots of the solution from the small vial were transferred to
test tubes and adjusted to 0.1 M NaOH with distilled water. Typically, 0.1 ml aliquots of
1 M NaOH solution are removed and diluted to 1 ml; if the total sample volume is less,
0.1 M NaOH is added to bring the volume up to 1 ml.
The assay involved sequential addition of the following reagents: 1.0 M acetic acid
to neutralize the NaOH; a succinimide/N-chlorosuccinimide oxidizing reagent, and a
barbituric acid/pyridine coupling reagent (Lambert et al. 1975; Brinker and Seigler 1989).


94
it is unknown whether the specificity is due to the sugar or the aglycone components of
the glycosides.
Insect P-glucosidases may be divided into three classes (Ferreira et al. 1998):
a. Class 1 enzymes with glycosyl P-glucosidase and aryl P-glucosidases activity, e.g.,
the sphingid moth, Erinnyis ello (Santos and Terra 1985).
b. Class 2 enzymes with only glycosyl P-glucosidase activity, particularly active
against disaccharides or oligosaccharides, e.g., the migratory locust, Locusta
migratoria (Morgan 1975).
c. Class 3 enzymes with only aryl (or alkyl) P-glucosidases activity, mainly active
against monosaccharides e.g., p-nitrophenyl p-glucosidase (NPpGluc), p-
nitrophenyl P-galactosidase (NPpGal) and salicin, e.g., Thaumetopoea pytiocampa
(Praviel-Sosa et al. 1987).
This chapter constitutes a preliminary study of the P-glycosidase found in
Heliconius himera. The objective of this chapter was to test whether or not larval P-
glycosidase per se inhibits cyanogenesis. Also, I wanted to document whether inhibition
was specific to one type of cyanogenic glycoside, or common to all types. Finally, I
wished to see whether there was only one or many glycosidases in the larval gut.
Materials and Methods
Study Organism
H. himera (Figure 5-1) is an aposematic butterfly that is distributed over
southwestern Ecuador, northwestern Peru and the Maraon Valley on the eastern side of
the Andes. The status of this species has been controversial for many years. Some
authors considered this species to be a race of Heliconius erato (Eltringham 1916, Lamas
1976, Brown 1979, Sheppard et al. 1985), while others consider this a separate species
(Kaye 1916, Emsley 1965). In the last few years, there has been a tendency to believe
that H. himera is a separate species, based on mitochondria DNA phylogenetic studies
(Brower 1994), which suggest a divergence between H. erato and H. himera


107
also sequestered both simple cyclopentenoid (Type I) efficiently, and small amounts of
complex cyclopentenoid (Type II) from their host plants. In addition, during the larval
stage the butterfly metabolized the complex cyanogenic glycosides into simple
cyclopentenoids which then were stored in the butterfly body. Because H. erato were
able to sequester simple cyclopentenoid glycosides efficiently, these butterflies might
have a way to deal with these toxic compounds, perhaps by an enzymatic mechanism.
Amino acid analysis of the adults showed that these butterflies are not only storing
cyanogenic compounds for defense but also essential amino acids gathered during the
larval stage that are probably used for its primary metabolism, nutrition and possibly for
the further synthesis of aliphatic glycosides during its adult stage.
Finally, a novel enzyme inhibiting cyanide release was discovered which is specific
for simple cyclopentenoid glycosides. The enzyme was purified, tested in vitro, and
isolated for amino acid sequencing. Two P-glucosidases were isolated from larval
midgut. The sequencing of these proteins produced some homologies with the complete
sequence for P-glucosidase of the moth Spodoptera frugiperda (Noctuidae). In vitro
testing of P-glucosidase from the plant, and cyanogenic glycoside and P-glucosidase from
the larval gut showed that the protein was specific to simple cyclopentenoid glycosides
and that it inhibited the production of cyanide. This discovery leads to the conclusion
that this detoxification mechanism is related to the ability of Heliconius to inhibit the
metabolism of different Passiflora cyanogen types. However, more experiments are
warranted to describe the characteristics of the enzymes. The possible genetic
manipulation of this new mechanism of cyanide detoxification through bioengineering
may have important ramifications for agricultural crop improvement.


COMPARATIVE PROCESSING OF CYANOGENIC GLYCOSIDES AND A NOVEL
CYANIDE INHIBITORY ENZYME IN HELICONIUS BUTTERFLIES
(LEPIDOPTERA: NYMPHALIDAE: HELICONIINAE)
By
MIRIAN MEDINA HAY-ROE
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
2004


LARVAL DEVELOPMENT
Figure 2-1. Life history of Heliconius butterflies.
PUPA
\? \
ri:.
ca. 8
Days"
ADULT
ca. 3
Months


CHAPTER 2
LIFE HISTORIES AND NUTRITIONAL EFFECTS ON SIZE OF HELICONIUS
ERATO FEEDING ON DIFFERENT PASSIFLORA HOST PLANTS
In the last two decades, many studies of life history evolution and evolution of
body size have been published. These studies present a retrospective analysis of patterns
of variation in size and age at maturity in an attempt to understand the mechanisms
responsible for generating phenotypic variation in maturation. However, there are
important aspects that these studies did not take into account, such as invariant age and/or
size thresholds that occur during instar transitions in larval development (Nijhout 1974,
1975). To study the life history of a population and/or species, one must, first of all,
study the egg because it constitutes potential maternal (Dunlap-Pianka et al. 1977,
Sinervo et al. 1992) and paternal (Boggs and Gilbert 1979, Wiklund and Kaitala 1993)
contributions to the offsprings fitness. The second stage, juvenile development, is
probably the most important stage in the life history of an arthropod. At this stage, three
different characters may be responsible for the age and size at maturity: the number of
instars, the larval developmental time or molting interval, and the change in size at
ecdysis. These characters can be plastic or canalized (Higgins and Rankin 1996) and
might differ, depending on the different post-embryonic pathways of the arthropod. In
the end, the variation among these stages may affect the final adult size in the species or
population (Figure 2-1).
Recent models are focusing not only on the analysis of growth but also on age, and
other complex and combined characters, which reflect more accurately how adaptive
6


41
defensive mechanism requires hydrolysis to function, we expected to find higher
concentrations of cyanide release at the interface of the hemolymph containing (3-
glucosidase and the cuticle. The thorax is a heavily sclerotized body part, filled with
hemolymph (Nation 2002). These two properties allow a heavy deposition of the
cyanogenic glycosides on the cuticle. By mixing these glycosides with hemolymph
containing (3-glucosidase, the butterfly is able to generate high concentrations of cyanide.
A similar rationale also enables us to understand finding high concentrations of cyanide
in the head. The head is small in mass and releases smaller amounts of cyanide, as
compared to the thorax, but since the head has a large surface area to volume ratio the
volatile cyanide will be released faster. Low cyanide concentrations within the wings and
abdomen might be the result of low levels of hemolymph in the former, and low levels of
sclerotized area in the latter, which implies a low level of hemolymph containing P-
glucosidase in the w'ings and low levels of cyanogenic glycosides in the cuticle of the
abdomen.
Brower et al. (1988) hypothesized that higher cuticular concentrations of
cardenolides may have evolved due to pre-adaptation from biochemical deposition in the
developing cuticle, rather than having evolved due to selection from predation or
parasitism.
Cyanogenesis in Heliconius butterflies probably functions as an alert signal to
reinforce memory of toxicity, also called recall-to-mind mimicry (Rothschild 1984).
Under this assumption it is not necessary for selection to favor highly cyanogenic wings,
because low levels of HCN will be enough to trigger recall of experiences of the bad taste
and smell of HCN and carbonyl compounds, which are formed during cyanogenesis. A


118
Olafsdottir, E. S., C. Cornett, and J. W. Jaroszewski. 1989b. Cyclopentenoid cyanohydrin
glycosides with unusual sugar residues. Acta Chem. Scand. 43: 51-55.
Pasteis, J. M., S. Dobler, M. Rowell-Rahier, A. Ehmke, and T. Hartmann. 1995.
Distribution of autogenous and host derived chemical defense in Oreina leaf
beetles (Coleptera: Chrysomelidae) J. Chem. Ecol. 21(8): 1163-1179.
Pasteis, J. M., D. Daloze, W. van Dorsser, and J. Roba. 1979. Cardiac glycosides in the
defensive secretion of Chrysolina herbcea. Identification, biological role and
pharmacological activity. Comp. Biochem. Physiol. 63C: 117-121.
Peterson, S. C., N. D. Johnson, and J. L. LeGuyader. 1987. Defensive regurgitation of
allelochemicals derived from host cyanogenesis by eastern tent caterpillars.
Ecology 68(5): 1268-1272.
Poulton, J. E. 1988. Localizations and catabolism of cyanogenic glycosides, pp. 67-91, in
D. Evered and S. Harnett (Eds.), Cyanide in Biology, Academic Press, London.
Praviel-Sosa, F., S. Clermont, F. Percheron, and C. Chararas. 1987. Studies of
glycosidases and glucanases in Thaumetopoea pytiocampa larvae. II. Purification
and some properties of a broad specific P-D glucosidase. Comp. Biochem. Physiol.
86B: 173-178.
Rammani, A. D. and D. A. Jones. 1985. Flexibility in cyanogenic phenotype of Lotus
corniculatus L. in response to low fluctuating temperatures. Pak. J. Bot. 17(1): 9-
23.
Raubenheimer, D. 1989. Cyanoglycoside gynocardin from Acraea horta (L)
(Lepidoptera: Acraeinae): possible implications for evolution of Acraeine host
choice. J. Chem Ecol. 15(8): 2177-2189.
Rhoades, D. F. and R. G. Cates. 1976. Toward a general theory of plant anti-herbivore
chemistry. Recent Adv. Phytochem. 10: 168-213.
Robinson, M. E. 1930. Cyanogenesis in plants. Biol. Rev. 5: 126-142.
Rothschild, M. 1972. Secondary plant substances and warning colouration in insects.
Symp. Roy. Entomol. Soc. London 6: 59-83.
Rothschild, M. 1984. Aide mmoire mimicry. Ecol. Entomol. 9: 311-319.
Rothschild, M., R. T. Reichstein, J. V. Euw, R. T. Aplin, and R. R. M. Harman. 1970.
Toxic Lepidoptera. Toxicon 8: 293-299.
Santos, C. D. and W. R. Terra. 1985. Physical properties, substrate specificities and a
probable mechanism for a P-D glucosidase (cellobiase) from midgut cells of the
cassava homworm (Erinnys ello). Biochim. Biophys. Acta 831: 179-185.


93
digestive enzymes, which have limited glycosidase activity, prevent expected
cyanogenesis in reaction mixtures of Passiflora cyanogenic glycoside/p-glycosidase.
Unfortunately, no data on these observations have ever been published. However,
Spencer (1984) published interesting data regarding P-glycosidase inhibition, when
combined two P-glucosidases (from plant or commercial origin) and a range of different
cyanogenic substrates. In all cases, cyanide release (measured by the Fiegl-Anger
method) was inhibited. Based on all of these data, Spencer (1987) proposed that P-
glycosidases of insect origin bind with a plant substrate-enzyme complex, either during
or after complex formation, in a competitive manner. The one-substrate, two-enzyme
complex then precipitates from solution (Spencer 1987).
P-glycosidases (E. C. 3.2.1) are hydrolytic enzymes specialized to break glycosidic
bonds to O-, N-, or S- groups on the aglycone. These exoenzymes hydrolyze the terminal
p-linked monosaccharide from the corresponding glycoside (di and/or oligosaccharides).
Depending on the monosaccharide that is removed, the P-glycosidase is named P*
glucosidase (glucose), P-galactosidase (galactose), and so on. Some insects have three or
four digestive P-glycosidases with different substrate specificities. In this case, P-
glucosidase is used as the generic name for all enzymes which remove glucose efficiently
(Terra and Ferreira 1994). Among the P-glycosidase substrates found in insects are toxic
p-glycosides, which are produced by plants and function to avoid or reduce herbivore
attack. Some insects are able to feed on these toxic plants without any apparent harm,
whereas others will show slower development and growth. These differences in
performance may be due to a detoxification mechanism acting after glycoside hydrolysis
or may be based on differential P- glycosidase specificity (Ferreira et al. 1997). However


accumulated higher concentrations of cyanogenic glycosides in their bodies, a condition
which in turn generates higher concentrations of hydrogen cyanide. H. e. favorinus, on
the other hand, turned out to be less cyanogenic and therefore less well-protected against
predation by frogs, demonstrating that chemical defenses in H. erato vary between
geographical races and that lower cyanide concentrations may lead to higher predation
rates.
H erato not only synthesizes de novo aliphatic cyanogenic glycosides, but also
sequesters both simple and complex cyclopentenoid glycosides from their host plants. In
addition, during the larval stage H. erato metabolized the complex cyclopentenoid
glycosides into simple cyclopentenoid glycosides. Analysis of amino acids in the adults
revealed that these butterflies are storing not only cyanogenic compounds for defense, but
also essential amino acids from metabolized cyclopentenoids gathered during the larval
stage.
Finally, a novel cyanide detoxification enzyme was tested in vitro. Two (3-
glucosidases were isolated from larval midgut; they had some homologies with
Spodoptera frugiperda P-glucosidase. The enzymes are specific to simple
cyclopentenoid glycosides and proved to inhibit the production of cyanide when
cyanogenic glycosides, plant P-glucosidase and larval midgut P-glucosidase were tested.
This discovery leads to the conclusion that this detoxification mechanism is related to the
ability of Heliconius to metabolize different Passiflora cyanogen types. However, more
experiments are warranted to describe the characteristics of the enzymes. The
importance of this new mechanism of cyanide detoxification is discussed.
xvi


104
Figure 5-2. Electrophoresis in SDS-10% gel slab. The gel shows different levels of
purification of the protein studied. CE: is the cytosolic crude extract, Q
sepharose elutes: Q1-Q3 fractions 31-33, Mono-Q elutes: Ml-M2, fractions
23 and 24 from Q-sepharose fraction 31, and M3-M4, fractions 24 and 25
from Q-sepharose fraction 32. RM is the rainbow marker.


21
Table 2-3. Repeated measurements ANOVA of larval developmental time with race and
host as factors.
Source
Df
F
P
Between races
j Race
1
9.68
0.002
j Host
4
15.75
0.001
| Race x Host
4
0.62
0.649
Error
223
Within races
H. e. favorinus
Family
16
2.68
0.001
Host
4
33.89
0.001
Error
129
H. e. cyrbia
Family
9
3.27
0.002
Host
4
3.41
0.013
Error
69
fable 2-4. Sexual differences in larval developmental time (LDT). growth rate (GR) and
wing length (WL) in H. e. favorinus and H. e. cyrbia feeding on five different
host plants.
Race
Host Plants
N
Both
sexes
Sex
LDT (days)
^ise
GR
(mm/days)
Jise
WL (cm)
X1 se
H. e.
auriculata
24
8
9.1810.26
0.19710.006
3.95 1 0.044
favorinus
9
9.5410.31
0.18610.007
3.91 10.040
biflora
25
0
8.3710.18
0.21710.006
3.9810.027
$
8.6410.10
0.205 1 0.003
3.92 1 0.034
punctata
26
e
8.0710.12
0.21910.004
3.9410.033
9
8.3610.15
0.208 1 0.004
3.81 10.090
rubra
23
8
10.4610.28
0.17410.007
3.9010.060
9
9.7510.25
0.18810.006
3.92 1 0.046
trifasciata
26
cJ
8.84 10.21
0.205 1 0.006
4.0010.034
9
8.7910.18
0.208 1 0.005
4.0010.030
H. e cyrbia
auriculata
19
8
9.1310.30
0.211 10.007
3.58 10.060
9
9.7010.73
0.203 10.012
3.4810.085
biflora
19
8
8.3810.26
0.2261 0.006
3.68 1 0.047
9-
9.0010.88
0.22010.014
3.55 10.036
punctata
17
6
9.3810.99
0.21210.016
3.53 1 0.067
9
8.22 1 0.22
0.23010 008
3.58 1 0.036
rubra
17
8
9.8010.33
0.19010.007
3.5410.051
9
10.0010.26
0.1881 0.006
3.54 1 0.032
trifasciata
18
8
9.63 1 0.60
0.19810.014
3.51 10.063
.
1
?
9.0010.31
0.21510.007
3.61 10.029


115
Jones, D. A. 1977. On the polymorphism of cyanogenesis in Lotus corniculatus L. VII.
The distribution of the cyanogenic form in Western Europe. Heredity 39: 27-44.
Jones. D. A. 1971. Chemical defense mechanisms and genetic polymorphism. Science
173: 945.
Jones, D. A., J. Parsons, and M. Rothschild. 1962. Release of hydrocyanic acid from
crushed tissues of all stages in the life-cycle of species of the Zygaeninae
(Lepidoptera). Nature 193(4819): 52-53.
Kaye, W. J. 1916. A reply to Dr. Eltringham's paper on the genus Heliconius. Trans. Ent.
Soc. London. 1916:149-155.
Kassarov, L. 2001. Do cyanogenic glycosides and pyrrolizidine alkaloids provide some
butterflies with a chemical defense against their bird predators? A different point of
view. Behavior 138:45-67.
Kassarov, L. 1999. Are birds able to taste and reject butterflies based on beak mark
tasting? A different point of view. Behavior 136: 965-981.
Kearsley, M. J. C. and T. G. Whitham. 1992. Guns and butter: a no cost defense against
predation for Chrysomela confluens. Oecologia 92: 556-562.
Killip, E. P. 1938. The American species of Passifloraceae. Field Mus. Nat. Hist. Bot.
Ser. 19: 1-613.
Kirkpatrick, K. and D. Lofsvold. 1992. Measuring selection and constraint in the
evolution of growth. Evolution 46(4): 954-971.
Kojima, M., J. E. Poulton, S. S. Thayer, and E. E. Conn. 1979. Tissue distributions of
dhurrin and of enzymes involved in its metabolism in leaves of Sorghum bicolor.
Planta Physiol. 63: 1022-1028.
Lamas, G. 1976. Notes on Peruvian butterflies (Lepidoptera). II New Heliconius from
Cusco and Madre de Dios. Revista Peruana de Entomologa 19: 1-7.
Lambert, J. L.. J. Ramasamy, and J. V. Paukstells. 1975. Stable reagents for the
colorimetric determination of cyanide by modified Konig reactions. Anal. Chem.
47(6): 916-918.
Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of
the bacteriophage T4. Nature 227: 680-685.
l anza, J. 1988. Ant preference for Passiflora nectar mimics that contain amino acids.
Biotropica 20(4): 341-344.
Lindroth. R. L. 1988. Hydrolysis of phenolic glycosides by midgut P-glucosidase in
Papilio glaucus subspecies. Insect Biochem. 18: 789-792.


CHAPTER 3
CYANOGENESIS IN HELICONIUS ERATO AND PASS1FLORA HOST PLANTS
Cyanogenic glycosides are secondary metabolites derived from amino acids.
Insects or plants that contain cyanogenic glycosides undergo a process called
cyanogenesis, which occurs during tissue disruption by feeding herbivores, predators or
pathogens. The process occurs in two steps: first, the enzyme P-glucosidase hydrolyzes
the cyanogenic glycosides, generating a cyanohydrin (a-hydroxynitrile) and a sugar
moiety (usually D-glucose, but the products may include other sugars, e. g., gentibiose).
The second step involves the degradation of the cyanohydrin and occurs rapidly at
normal cell pH. This is thermodynamically favored, and in many instances, but not
always, a second type of enzyme (a-hydroxynitrile lyase) catalyzes the dissociation of the
cyanohydrin to a carbonyl compound and hydrogen cyanide (Hosel 1981) (Figure 3.1).
Occurrence of Cyanogenesis in Plants
Normally, the cyanogenic glycosides and the enzymes that hydrolyze them to
produce cyanide are spatially separated within the plant tissues to prevent autotoxicity
(Hosel 1981, Poulton 1988). In plants, the cyanogenic glycosides are found in the
epidermal layer of the plant tissues, whereas both hydrolyzing enzymes are found in the
mesophyll cells (Kojima et al. 1979, Saunders et al. 1977, Poulton 1988).
Cyanogenic plants show high variability in hydrogen cyanide (HCN) production,
which has been related to low moisture stress (Foulds and Grime 1972, Abbot 1977),
frost or low temperatures (Ellis et al. 1977), or presence of herbivores (Cooper-Driver
and Swain 1976). This variation reflects phenotypic and genotypic diversity in both the
29


50
O)
Vi
600
500
400
300
200
>>
la
o
M
O
D£
z
u
M 100
B
L
B
B
Head
Thorax Abdomen Wings
Body parts
Female
Male
Figure 3-5. Cyanide quantification of male and female H. e.favorirtus fed P. trifasciata.
Different letters indicate significant differences between means (Tukey HSD-
test). Bars indicate standard error.
3
o
,W)
z
u
or
Body
Wings
Figure 3-6. Cyanide quantification of male and female H. e. demophon fed natural (P.
biflora) and alternative host plants. Bars indicate standard error.


17
In this chapter, I demonstrated through life history studies of the two races of
Heliconius erato that differences in body size seem to be genetically based and size
differences will remain, regardless of the host plant provided. Within each population,
there were maternal effects confirmed by egg mass variation between full-sibs. Also,
genotypic variation occurred in each stage of the life history within each race.
//. e.favorinus proved to have more plastic growth trajectories, due to variation in
their average larval developmental time and variation in size when the larvae were fed
different host plants. This result demonstrates that both developmental pattern and
growth are influenced by host plants (chemical/nutritional). It also indicates that
maturation is not simply a consequence of growth, and that environments can have
correlated effects on both growth and developmental rates. Within this race, I also
proved that when favorinus fed on their natural host plant, P. trifasciata, as well as the
alternative host P. punctata, they generate earlier-born progeny. Earlier-bom progeny
might have a significant impact on population growth (Fisher 1930 and Charlesworth
1980) because they might increase reproductive success and fitness in stable populations.
This particularly occurred in H. e. favorinus fed its natural host P. trifasciata, which
produced larger individuals of both sexes. Larger individuals will again have a direct
effects on fecundity and competitive ability (mating success), especially in competitions
for pupal mating (Deinert et al. 1994). However, the alternative host plants (P. punctata)
produced smaller females, on average, which directly affects fitness, because smaller
females will tend to produce fewer eggs on average (Dunlap-Pianka 1979 and Boggs
1979). Interestingly, fast growing individuals become heavier than individuals that grow
more slowly. This is probably due to faster food consumption (not measured here). Final


Copyright 2004
by
Mirian Medina Hay-Roe


To all the people that love science and helped me through the development of this
research.


89
ppm
20
40
60
80
100
120
140
Figure 4-5. Proton/Carbon heteronuclear multiple quantum coherence (HMQC) spectrum
of cyanogen passitrifasciatin in CD3OD from Passiflora sp.


34
Louis. Missouri and the identification of the plants has been confirmed by Dr. John
MacDougal, an expert in the systematics of the family Passifloraceae.
Experimental Design and Treatments
Butterflies
H. e favorinus, H. e. cyrbia and H. e. demophon larvae were raised in
environmental chambers under controlled temperatures (27C), humidity (78%) and
.photoperiod (14L:10D).
The first two races were fed only natural host plants, Passiflora trifasciata for H. e.
favorinus, and P. punctata and P. rubra for H. e. cyrbia. H. e. demophon, on the other
hand, was fed both natural host plants (P. biflora) and alternative host plants (Table 3.1).
The experimental procedure used was as follows:
a. Specimens were selected from both H. e. cyrbia and H. e. favorinus.
b. Each individual male and female was tested separately.
. For each specimen, the head, thorax, abdomen and wings were tested separately.
The procedure was slightly altered for the testing of H. e. demophon (the full
procedure had not been designed at the time that this testing took place). For this
subspecies the head, thorax and abdomen were not tested separately, only the wings were.
Temporal cyanide quantification of Passiflora plants
Various Passiflora plants were tested during the summer (July-August) and fall
(September-October) of 1999 (Table 3.2). All the plants tested were potential host plants
of Heliconius erato in different geographical areas, such as Brazil, Costa Rica, and
Ecuador.
The analysis was performed on three different plant parts: young leaves, mature
leaves, and stems. Plant parts not normally eaten by Heliconius larvae were not tested.


4
cause female butterflies to avoid further oviposition on such plants (Williams and
Gilbert 1981). For chemical protection against herbivores, several strategies are used by
the Passiflora. Most of the Passiflora produce sugary secretions from extrafloral
nectaries on petioles, leaf margins, surface, tips, bracts and stipules, which attract ants
that actively defend the plants. Ants then eat heliconian eggs, and attack and carry off
small larvae (Lanza 1988). But most importantly, Passiflora plants are protected from
most herbivores by their production of toxic cyanide-releasing secondary compounds
called cyanogenic glycosides (Spencer 1988). The Passifloraceae are among the few
plants producing cyanogenic glycosides with a cyclopentene moiety (Spencer 1988).
1his characteristic is shared only by members of five other closely related families:
Flacortiaceae, Malesherbiaceae, Tumeraceae, Achariaceae and Caricaceae.
Few Heliconius species have been investigated with regard to their chemical
interaction with Passiflora cyanogens. It is known that the larvae, pupae, and adults of
Heliconius and other related genera biosynthesize de novo two simple aliphatic
cyanogens, linamarin and lotaustralin (Nahrstedt and Davis 1983). Also, it has recently
been demonstrated that Heliconius sara, a specialist on P. auriculata, sequesters
cyanogens (epivolkenin) from its host plant and metabolizes most of this compound into
the corresponding thiol derivative (sarauriculatin) by replacing the nitrile group which
prevents cyanide release (Engler et al. 2000). Adult butterflies of the species Acraea
horta (Nymphalidae: Acraeinae) store the cyclopentyl cyanogen, gynocardin, when they
develop as larvae on Passiflora plants (Raubenheimer 1989).
In order to understand the phenomena of synthesis and sequestration, of chemical
compounds in Lepidoptera, it is important to understand the life histories of these species


113
Fraenkel, G. 1959. The raison detre of secondary plant substances. Science 129:1466-
1470.
Franzl, S. and C. M. Naumann. 1985. Cuticular cavities: storage chambers for
cyanoglucoside-containing defensive secretions in larvae of zygaenid moth. Tissue
and Cell 17(2): 267-278.
Futuyma, D. J. 1983. Evolutionary interactions among herbivorous insects and plants, pp.
207-231, in D. J. Futuyma and M. Slatkin (Eds.). Coevolution. Sinauer Associates.
Massachussets.
Futuyma, D. J. and M. C. Keese. 1992. Evolution and coevolution of plants and
phytophagous arthropods, pp. 439-475, in G. A. Rosenthal and M. R. Berenbaum
(Eds.), Herbivores, Their Interactions With Secondary Plant Metabolites Vol II,
Academic Press, San Diego.
Gilbert, L. E. 1991. Biodiversity of a Central American Heliconius community: Patterns,
process, and problems, pp. 403-427, in P. W. Price, T. M. Lewinsohn. G. W.
Fernandes, and W. W. Benson (Eds.), Plant-Animal Interactions: Evolutionary
Ecology in Tropical and Temperate Regions, John Wiley and Sons, New York.
Gilbert, L. E. 1983. Coevolution and mimicry, pp. 263-285, in D. Futuyma and D. M.
Slatkin (Eds), Coevolution, Sinauer, Sunderland, MA.
Gilbert, L. E. 1976. Post mating female odor in Heliconius butterflies: a male-contributed
antiphrodisiac. Science 193: 419-420.
Gilbert, L. E. 1972. Pollen feeding and the reproductive biology of Heliconius butterflies.
Proc. Nat. Acad Sci. USA. 69: 1403-1407.
Gilbert, L. E. 1971. Butterfly-plant coevolution: Has Passiflora adenopoda won
selectional race with Heliconiinae butterflies? Science 172: 585-586.
Gilmour, D. 1965. The Metabolism of Insects. W. H. Freeman and Company. San
Francisco.
Gleadow. R. M. and I. E. Woodrow. 2002. Constraints on effectiveness of cyanogenic
glycosides in herbivore defense. J. Chem. Ecol. 28(7): 1301-1313.
Haribal, M. and J. A. A. Renwick. 2001. Seasonal and population variation in flavonoid
and alliarinoside content of Alliariapetiolata. J. Chem. Ecol. 27(8): 1585-1594.
Harms, H. 1924. Passifloraceae, pp. 69-94, in A. Engler & K. Prantyl (Eds.), Die
naturlichen Pflanzenfamilien III (6A).
Hay-Roe, M. M. 1996. Growth rate plasticity in two races of Heliconius erato that differ
in body size. M. A. Thesis, University of Texas at Austin.


31
first. For example, toxic compounds are more concentrated in the wings of monarch
butterflies (Brower and Glazier 1975). Chrysomeiid beetles (of the genus Oreina)
rapidly distribute toxic compounds stored in pronotal and elytral glands onto the surface
of the cuticle (Pasteis et al. 1995, 1979) when disturbed. Variations in the concentration
of defensive compounds among body parts may be due to selection by predators for more
effective protection and/or may depend on the mode of action or properties of the toxic
compounds (Bowers 1988).
Cyanogenic Glycosides as a Defensive Mechanism
It is generally accepted that cyanogenic glycosides act as a defense mechanism
against many potential herbivores and predators (Jones 1988, Seigler 1991, Nahrstedt
1992, 1988, Gleadow and Woodrow 2002). In the last decade, there has been some
controversy over the effectiveness of cyanogenesis as a herbivore defense (Kassarov
1999, 2001, Hruska 1988), but the utilization of cyanogenic glycoside in plant defense
has recently been proven by transferring the pathway for cyanogenic glycoside
biosynthesis into the acyanogenic Arabidopsis thaliana (Tattersall et al. 2001). There is
also some doubt about whether HCN or the aglycone (aldehydes or ketones), which are
simultaneously released upon cyanogenesis, are the source of toxicity (Conn 1979, Jones
1988, Spencer 1987). For example, regurgitate of enteric fluids by the eastern tent
caterpillar, Malacosoma americana fed black cherry contained both HCN and the
carbonyl compound, benzaldehyde. When these compounds were tested on ants, they
tolerated the HCN but were repelled by the benzaldehyde (Peterson et al. 1987). Also,
Jones (1988) revealed that the carbonyl compounds, acetone and 2-butanone, acted as
deterrents to feeding on Lotus corniculatus and Trifolium repens by mollusks. Despite


74
Hevea brasiliensis, during germination due to the cyanide detoxification enzyme P-
cyanoalanine synthase and hydrase, which cause the transport of nitrogen to be used in
protein synthesis (Selmar et al. 1998). If Heliconius is able to sequester cyanogenic
glycoside, this means that there must be a way to cope with cyanide release. This could
be achieved with an enzymatic mechanism. In H. sara, for example, the cyanogenic
glycoside is metabolized within the insect, by means of an unknown mechanism, and the
nitrogen is recovered and used in protein synthesis (Engler et al. 2000). If metabolism of
sequestered compounds is useful for nitrogen metabolism and reallocation for nutritional
needs, Heliconius is gaining in the coevolutionary race, because butterflies are able to
gain nitrogen for their primary metabolic needs via nutrition from the plant, and can also
use it for defense. A similar mechanism, with other compounds, has been described in
the leaf beetle, Chrysomela confluens (Kearsley and Witman 1992).
This study investigated the basis of the importance of cyanogenic glycosides in
relation to amino acids. Future experiments could proceed in two ways, the first using
radiolabeled amino acids to reveal the pathway of cyanide derived from sequestered
cyanogens in Heliconius butterflies, and the second using HPLC/MS to learn the
contribution of different cyanogenic glycosides towards amino acid composition in the
adult butterfly.


pg CN/g of dry tissue pg CN/g of dry tissue
48
Passiflora plants
auri culata biflora capsularis misera trifasciata
Passiflora plants
Figure 3-2. Cyanide quantification o Passiflora plants in the (A) summer and (B) fall
seasons. Bars indicate standard error.


66
was known. The analysis was performed by HPLC/MS under the same conditions
described above.
Results
Cyanogenic Glycosides in Passiflora Plants
The (3-glucosidases classification test of plant extracts revealed that all Passiflora
species tested possess cyanogenic glycoside Type II, with the exception of P. auriculata,
which possess cyanogenic glycoside Type I. Interestingly, P. rubra reacted with all p-
glucosidases. This means that the aglycone in P. rubra was not specific to any P-
glucosidase. Emulsin was also able to hydrolyze the cyanogenic glycoside from P.
trifasciata and P. auriculata. Cyanogenesis occurred very rapidly in the extracts of P.
rubra with high release of cyanide within 30 minutes of being tested, followed by P.
punctata and P. auriculata (Table 4-1). Cyanogenesis occurred within one hour in P.
biflora and within two hours in P. trifasciata. Cyanide release then increased over time.
Identification of cyclopentanoid cyanogenic glycoside in Passiflora rubra
The chemistry of this plant has not been analyzed before. One peak was isolated
from Passiflora rubra by HPLC and tested for cyanogenesis by TLC (Rf = 0.2-0.37).
The molecular weight was verified by HPLC-MS (MW 417).
'id NMR spectral data of the cyanogen in P. rubra showed chemical shifts and
signal splitting similar to those reported for passicapsin in CD3OD (Table 4-2). I3C NMR
(Table 4-3) also showed characteristic peaks for passicapsin (Olafsdottir et al. 1989b). It
was not clear whether the resonances for H2, H3, C2, and C3 had been assigned in
Olafsdottirs data. Those 'H assignments are given in Table 4-2 by analogy to the data in
Table-4-4, and the coupling constants for these resonances appear consistent, too. The


And finally, my thanks go to my husband, Keith, for his love and willingness in
helping me with some of this research and especially for being the outstanding editor that
he is, and for being the father of my daughter, Kylie, who shares my love for butterflies
and enjoys being my best helper.
vii


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
COMPARATIVE PROCESSING OF CYANOGENIC GLYCOSIDES AND A NOVEL
CYANIDE INHIBITORY ENZYME IN HELICONIUS BUTTERFLIES
(LEPIDOPTERA: NYMPHALIDAE: HELICONIINAE)
By
Mirian Medina Hay-Roe
May 2004
Chair: Thomas C. Emmel
Major Department: Entomology and Nematology
Heliconius erato, a neotropical butterfly known to specialize in larval feeding on
cyanide-producing Passiflora plants, has evolved different strategies for dealing with
cyanogenic compounds. By combining ecological and biochemical data, I investigated
some of these mechanisms. Subspecies of Heliconius erato from the eastern and the
western side of the Andes differ in body size. The subspecies from the eastern side of the
Andes proved to be consistently larger than their relatives from the western side of the
Andes. These differences in body size are genetically based. When two subspecies (H.
e. cyrbia and H. e. favorinus) were fed different host plants (natural and alternative host),
plastic growth trajectories resulted which were related to the host plant fed on by the
larvae. Feeding on one of the host plants, P. rubra, resulted in lower growth rates in both
subspecies and higher mortality in H. e. cyrbia, but not in H. e. favorinus. As a result, the
few individuals from the former subspecies that survived while feeding on this highly
toxic cyanogenic plant were better protected against frog predators because they
xv


20
Table 2-1. Average wing length and standard error of geographical races of Heliconius
erato from museum specimens in different regions.
Geographical races
Regions
N
WL (cm)
X se
H e. magnificus
Eastern
15
3.85 0.06
i H. e.favorinus
Eastern
57
3.94 0.03
H. e. phyllis
Eastern
32
3.89 0.04
H. e. hydara
Western
15
3.30 0.06
H. e. petiveranus
Western
25
3.01 0.08
H. e. cyrbia
Western
31
3.55 0.03
j H. e. colombina
Western
15
3.24 0.03
Table 2-2. Repeated measurements ANOVA of head capsule size with race, host and
familv as factors.
Source
Df
F
P
Between races
Race
1
751.96
0.001
Host
4
3.86
0.005
Rac x Host
4
1.18
0.323
Error
223
Within races
H. e. favorinus
Family
16
1.47
0.120
Host
4
3.58
0.008
Error
129
H. e cyrbia
family
9
1.52
0.159
Host
4
1.32
0.272
Error
69


18
adult size indicates that there is only a slight difference (p=0.05) in size offavorinus
when fed its natural host plant compared to other host plants. This result indicates that
there is a strong selection for being the right size" as well as having a short
developmental time.
On the other hand, H. e. cyrbia also presented plastic growth across host plants.
However, the most striking result was the high mortality and low growth rate when
individuals fed on their natural host plant, P. rubra. Although not significant, there was
also a lower growth rate when female larvae fed on P auriculata and they generated
smaller females. This particular race has been investigated (Jiggins et al. 1997) in
relationship to their hybridization zones and the recent speciation of H. himera. In
contrast to their results, my greenhouse colony of cyrbia displayed a marked preference
for oviposition on P. punctata. H. e. cyrbia was exposed to both natural host plants for
more than 10 generations. Two important observations regarding host plant preferences
were made during these studies (Hay-Roe and Rios, personal observations),
a. On average, we collected 3 to 5 eggs per day on P. rubra and 10 to 12 eggs per day
on P. punctata.
b. On various occasions, P. punctata were defoliated when we allowed the larvae to
grow on the plants inside the cage. P. rubra were never defoliated.
Regardless of the observations of oviposition behavior, the results of this study
suggest that there is a difference in survival of larvae fed P. rubra vs. P. punctata, a fact
that Jiggins et al. (1997) did not examine explicitly (also, Mallet personal
communications). Perhaps, one ought to revive the hypothesis that differences in host
plant usage between the sister species H. e. cyrbia and H. himera could be one cause that
induced the recent speciation of H. himera.


55
cyclopentene ring. None of the glycosides in Type II co-occur with another
enantiomer, as it occurs in glycosides of Type I. Also, all the glycosides in Type II
have the same stereochemistry in position C-l. The cyanogenic glycoside grouped
within Type I and Type II are believed to be biosynthesized from the amino acid 2-
cyclopentenyl glycine via oxidation, oxidative decarboxylation, dehydration, and/or
glycosylation (Conn 1981b). However, Olafsdottir et al. (1992) do not exclude the
possibility that they originated from oxygenated amino acids (Figure 4-1B).
c. Type III (called aliphatic cyanogenic glycoside by Spencer 1988). These
glycosides do not contain a cyclopentene ring. They are derived from the amino
acids valine and isoleucine. This group includes the glycosides linamarin and
lotaustralin, the cyanogens found most often in nature (Figure 4-1C).
d. Type IV. This group of glycosides include the aromatic cyanogens, for example
amygdalin, which is present in seeds of bitter almonds and prunasin found in
Prunus species. The glycosides within this type originated from the amino acid
phenylalanine (See Figure 4-ID).
No other plant group is as diversified in its types of cyanogenic glycosides as the
genus Passiflora (Spencer 1988, Jaroszewski et al. 2002).
Cyanogenic Glycosides in Lepidoptera
Unpalatable insects can acquire their defensive toxins either by sequestering plant
derived substances or by synthesis de novo (Blum 1981, Rothschild et al. 1970). In
butterflies and moths feeding on cyanogenic plants, both acquisition methods have been
described.
Synthesis De Novo
De novo synthesis has been described in many species of the family Zygaenidae
(Jones et al. 1962, Davis and Nahrstedt 1979), in three subfamilies of the Nymphalidae
(Heliconiinae, Nymphalinae and Acraeinae), and in one subfamily of the Lycaenidae
(Nahrstedt and Davis 1985).
Based on radiolabeled amino acids, Nahrstedt and Davis (1983, 1985)
demonstrated that larvae and adult Heliconius butterflies synthesize the cyanogenic
glycoside themselves, linamarin and lotaustralin from valine and isoleucine, respectively.


67
corresponding l3C resonances (Table 4-3) were assigned by way of a 2D-HMQC
spectrum (Figure 4-4).
Identification of cyclopentanoid cyanogenic glycoside in Passiflora sp.
Although this Passiflora species is reported throughout this dissertation as P.
trifasciata, it is actually an undescribed sister species of P. tricuspis and P. trifasciata
(McDougal personal communication).
A peak was isolated from Passiflora sp. by HPLC and tested for cyanogenesis by
TLC (Rf = 0.16-0.24). The molecular weight was verified by HPLC-MS (MW 433). 'H
NMR spectral of this species showed the coupling constants (Hz) and chemical shifts
(ppm) listed in Table 4-4. I3C NMR also showed characteristic peaks for passitrifasciatin
(Table 4-5) (Olafsdottir et al. 1991). The resonances for C2, C3, CL, and Cl" had been
assigned in Table 4-5 by correlations with their corresponding ]H resonances (Table 4-4),
which were assigned by Olafsdottir et al. 1991; this was accomplished using the 2D-
HMQC spectrum (Figure 4-5).
Cyanogen profile of Passiflora auriculata
Two peaks were isolated from P. auriculata by HPLC and tested for
cyanogenesis by TLC, which resulted in a Rf (0.11, 0.2) and Rf2 (0.28, 0.37). The
molecular weight for both was verified by HPLC-MS (MW 287). Fraction #1 was
identified by HPLC/MS as epivolkenin MW=287, eluting at 17.76 min. Fraction #2 was
identified as taraktophyllin MW=287 and eluted at 2.73 min.
Cyanogen profile of Passiflora biflora
P. biflora contained passibiflorin as described by Olafsdottir et al. (1989b). The Rf
values obtained by TLC were 0.16-0.23 and the MW=433.


57
upon. This fact suggests that these species synthesize their own cyanogenic glycoside,
and also sequester sarmentosin (Nishida 1994).
Moreover, it has recently been demonstrated that a cyclopentyl monoglycoside
(epivolkenin) was sequestered by Heliconius sara when it fed upon its host plant P.
auriculata (Engler et al. 2000). This species converts most of this compound into the
corresponding thiol derivative (sarauriculatin), suggesting that the replacement of the
nitrile group by a thiol would prevent cyanide release from the host plant and release
valuable nitrogen into the insects primary metabolism. This finding suggested that a
unique enzymatic mechanism for dealing with plant cyanogenic glycoside exists in this
species (Engler et al. 2000).
Enzymes
Plant cyanogens co-occur with hydrolyzing P-glucosidase enzymes, which are very
specific in the type of compound with which they can co-occur (Robinson 1930, Spencer
1987). The aglycone, together with the sugar part of the substrate, play an important role
in this specificity, although the aglycone seems to be the main determinant of the
specificity (Hosel 1981).
In this chapter, the fate of cyanogenic glycosides from Passiflora plants after
ingestion by Heliconius erato larvae was studied. The chemical compounds found in
Passiflora plant will be discussed first, and then the cyanogens found in larval and adult
samples. The mechanisms used by H. erato and H. himera species to counteract host
ailelochemicals will be described. Finally, since amino acids are the precursors of
cyanogenic glycosides, I determined the essential amino acids composition derived from
larval feeding.


117
Nahrstedt, A. and R. H. Davis. 1983. Occurrence, variation, and biosynthesis of the
cyanogenic glucosides linamarin and lotaustralin in the species of Heliconiini
(Insects: Lepidoptera). Comp. Biochem. Physiol. 75B: 65-73.
Nation, J. L. 2002. Insect Physiology and Biochemistry. CRC Press. Washington D.C.
Nijhout, H. F. 1991. The Development and Evolution of Butterfly Wing Patterns.
Smithsonian Institution Press. Washington, DC.
Nijhout, H. F. 1975. A threshold size for metamorphosis in the tobacco hornworm,
Manduca sexta (L). Biol. Bull. 149: 214-225.
Nijhout, H. F. 1974. Control of moulting and metamorphosis in the tobacco hornworm,
Manduca sexta (L): Growth of the last instar larva and the decision to pupate. J.
Exp. Biol. 61:481-491.
Nishida, R. 2002. Sequestration of defensive substances from plants by Lepidoptera.
Annu. Rev. Entomol. 47: 57-92.
Nishida, R. 1994. Sequestration of plant secondary compounds by butterflies and moths.
Chemoecology 5/6: 127-138.
Nishida, R. and M. Rothschild. 1995. A cyanoglucoside stored by a Sedum-feeding
Apollo butterfly, Parnassius phoebus. Experientia 51: 267-269.
Nishida. R., M. Rothschild, and R. Mummery. 1994. A cyanoglucoside, sarmentosin,
from the magpie moth. Abraxas grossularia, Geometridae: Lepidoptera.
Phytochemistry 36: 37-38.
Nylin, S. 1992. Seasonal plasticity in life history traits growth and development in
Polygonia c-album (Lepidoptera: Nymphalidae). Biol. J. Linn. Soc. 47: 301-323.
Nylin, S., P.O. Wickman, and C. Wiklund. 1989. Seasonal plasticity in growth and
development of the speckled wood butterfly, Pararge aegeria (Satyrinae). Biol. J.
Linn. Soc. 38: 155-171.
Olafsdottir, E. S., L. B. Jorgensen, and J. W. Jaroszewsky. 1992. Substrate specificity in
the biosynthesis of cyclopentanoid cyanohydrin glucosides. Phytochemistry 31:
4129-4134.
Olafsdottir, E. S., J. Jaroszewski, and D. Seigler 1991. Cyanohydrin glycosides with
unusual sugar residues: revised structure of passitrifasciatin. Phytochemistry 30(3):
867-869.
Olafsdottir, E. S., J. V. Andersen, and J. W. Jaroszewski. 1989a. Cyanohydrin glycosides
of Passifloraceae. Phytochemistry 28 (1): 127-132.


Table 4-4. *H NMR(a) spectrum of a cyanogen, identified as passitrifasciatin, from Passijlora sp. (referred throughout this dissertation
as P. trifasciata) in CD3OD. Chemical shifts (ppm) and coupling constants (Hz). Degree of splitting is identified by: (d)
Cyanogen
H2
H3
H4
H5A&H5B
HI'
H2'
H3'- H5'
H6'A&
H6'B
HI"
H2"
H3"
H4"
H5"
CHj
Multiplicity
dd
dd
dd, dd
d
dd
m,m,m
dd, dd
d
m
t
dd
dq
d
Passitrifasciatin(b)
8 ppm
6.16
6.36
4.86
2.49,3.0
4.61
3.21
3.25-3.40
3.66,3.85
4.69
3.25-
3.40
4.0
3.15
3.75
1.25
J values
5.5,1.0
5.5,2.1
15.0,4.0
15.0, 7.0
7.7
7.7,9.0
c
12.1,5.1
12.1,1.9
8.0
c
-
9.5,2.8
9.5,6.3
6.3
Passijlora sp.
8 ppm
6.17
6.36
e
2.50,3.01
4.62
3.22
3.25-3.40
3.67,3.85
4.7
3.25-
3.40
4.01
3.16
3.76
1.26
J values
5.5,1.0
5.5,2.1
"
15.0,4.1
15.0, 7.1
7.8
7.9,9.0
c
12.0,5.1
12.1,1.7
8.0
c
2.9
9.5,2.8
9.5,6.2
6.2
M Too complex for J-analysis.


103
Figure 5-1. Heliconius himera from Vilcabamba, Loja Province, Ecuador.


CHAPTER 5
STUDIES OF CYANIDE INHIBITION BY P-GLUCOSIDASE FROM HELICONIUS
LARVAL GUTS
In the course of evolutionary time, herbivores have adapted to different degrees to
plant hosts in their struggle to obtain the nutrients required for growth, development,
reproduction and other biological processes (Fraenkel 1959, Feeny 1976, Rhoades and
Cates 1983, Futuyma 1983). Plants, in turn, protect themselves by either physical or
chemical means. Plant defensive compounds are diverse but are not always toxic to all
insects (Schoonhaven 1972, Slansky 1992). Obviously, many phytophagous insects are
able to overcome these otherwise toxic allelochemicals by using different enzymes to
transform them (Brattsten 1992).
For herbivores feeding on plants containing cyanogenic glycosides, two different
detoxification pathways for hydrogen cyanide have been demonstrated (Witthoun and
Naumann 1987). One way involves the thiosulfate sulfur transferase, rhodanase, which
converts hydrogen cyanide to thiocyanate. The other pathway is based on two enzymes,
P-cyanoalanine synthase and p-cyanoalanine hydrase, which convert hydrogen cyanide to
P-cyanoalanine (Conn 1981b, Narshtedt 1988, Jones 1988). Some data suggest that
Heliconius butterflies may have evolved a tolerance to cyanide through the possession of
P-cyanoalanine synthase and p-cyanoalanine hydrase (Conn 1981b, Narshtedt 1988,
Witthoum and Naumann 1987). Some Heliconius may prevent cyanogenesis altogether
in their host plant via larval gut P-glucosidase, which may inhibit Passiflora P-
glucosidase (Spencer 1988). Spencer (1988) observed that crude extracts of larval
92


116
Lindroth, R. L. and A. V. Weisbrod. 1991. Genetic variation in response of the gypsy
moth to aspen phenolic glycosides. Biochem. Syst. Ecol. 19(2): 97-103.
Mallet, J. 1999. Causes and consequences of a lack of coevolution in Mullerian mimicry.
Evol. Ecol. 13: 777-806.
Mallet, J. 1986. Hybrid zones in Heliconius butterflies in Panama, and the stability and
movement of warming colour dines. Heredity 56: 191-202.
McCormick, S. and T. J. Mabry. 1981. Flavonoids of Passiflorapavonis. J. Nat. Prod.
44:623-624.
McKey, D. 1974. Adaptive patterns in alkaloid physiology. Am. Nat. 108:305-320.
McMillan, W. O., C. D. Jiggins, and J. Mallet 1997. What initiates speciation in passion-
vine butterflies? Proc. Nat. Acad. Sci. USA. 94: 8628-8633.
Menna-Barreto, Y. and A. M. Araujo 1985. Evidence for host plant preference in
Heliconius erato phyllis from southern Brazil (Nymphalidae). J. Res. Lepid. 24(1):
41-46.
Mode, C. J. 1958. A mathematical model for the co-evolution of obligate parasites and
their hosts. Evolution 12(2): 158-165.
Morgan, M. R. J. 1975. Relationships between gut cellobiase, lactase, aryl [3-glucosidase
and aryl P-galactosidase activities of Locusta migratoria. Insect Biochem. 5: 609-
617.
Muhtasib, H. and D. L. Evans. 1987. Linamarin and histamine in the defense of adult
Zygaena filipndula. J. Chem. Ecol. 13(1): 133-142.
Nahrstedt, A. 1992. The biology of the cyanogenic glycosides: new developments, pp.
249-269, in K. Mengel and D. J. Pilbeam (Eds.) Nitrogen Metabolism of Plants.
Clarendon Press, Oxford.
Nahrstedt, A. 1988. Cyanogenesis and the role of cyanogenic compounds in insects, pp.
131-150, in D. Evered and S. Harnett (Eds), Cyanide Compounds in Biology, Ciba
Foundation Symposium, Wiley, Chichester.
Nahrstedt, A. 1987. Recent developments in chemistry, distribution and biology of the
cyanogenic glycosides, pp. 213-234, in K. Hostettmann and P. J. Lea (Eds.),
Biochemically Active Natural Products. Clarenton Press. Oxford.
Nahrstedt, A. and R. H. Davis. 1985. Biosynthesis and quantitative relationships of the
cyanogenic glycosides, linamarin and lotaustralin, in genera of the Heliconiini
(Insecta: Lepidoptera). Comp. Biochem. Physiol. 82B: 745-749.


Table 4-2. !H NMR'2 spectrum of a cyanogen, identified as passicapsin, from P. rubra in CD3OD. Chemical shifts (ppm) and
coupling constants (Hz).
Degree of splitting is identified
?y: (d) dou
?let, (q) quartet, (m) multiplet, (br) broad.
Cyanogen
H2
H3
H4
H5A &
H5B
Hr
H2'-H5'
H6'A &
H6'B
HI"
H2"A & H2"B
H3"
H4"
H5"
ch3
Multiplicity
dd
dd
m
dd
d
m,m,m,m
dd, dd
dd
dt, ddd
br, q
dlL
d
Passicapsin(b)
8 ppm
6.12
6.36
4.87
2.34,3.04
4.59
3.33-3.51
3.67
3.86
4.89
1.70
1.83
3.96
3.20
4.00
1.25
J values
1.3,5.5
2.1,5.5
4.7.14.8
7.1.14.8
7.7
c
5.2,12.0
2.0,12.0
2.5,9.5
2.5.13.5
3.2.9.5.13.5
3.5
1.5,6.6
6.6
P. rubra
8 ppm
6.14
6.37
4.88
2.34,3.05
4.58
3.22-3.39
3.67
3.86
4.89
1.71
1.82
3.96
3.20
3.99
1.25
J values
1.0,5.5
2.0,5.5
4.7.14.8
7.1.14.8
7.8
c
5.3,12.0
2.0,12.0
2 ,9.5
2.1,13.6
3.1,9.9,13.3
3.2
1.1,6.6
6.6
(c) Too complex for J-analysis.
On


30
production of cyanogenic glycosides and the existence or absence of their degrading
enzymes (Vetter 2000. Jones 1977, Hughes 1981. Schappert and Shore 1995, 2000).
Plants with such genetic variation constitute an important source of information on the
direct effects of this type of compound in defense against herbivores that prey upon the
plants (Jones 1971).
Occurrence of Cyanogenesis in Lepidoptera
Cyanogenic glycosides have been found in all genera of Heliconiinae, including
Cethosia biblis and C. hypsea (Heliconiinae: Nymphalidae) (Nahrstedt and Davis 1983).
Acraea horta (Acraeinae: Nymphalidae) (Raubenheimer 1989), the lycaenids Brephidium
exillis and Polyommatus icarus (Nahrstedt 1987), and in moths of the family Zygaenidae
(Franzl and Naumann 1985, Muhtasib and Evans 1987). The location within the body of
cyanogenic glycosides has been studied in Zygaena moths and Heliconius butterflies. In
both, the cyanogenic compounds were found in the cuticular part of the integument,
although in zygaenids, 30% of the cyanogenic compounds were also found in the
hemolymph (Davis and Nahrstedt 1985, Nahrstedt and Davis 1983, 1985). The enzyme
jj-glucosidase, on the other hand, is mainly stored in the hemolymph (Nahrstedt 1988).
All stages in the life history of Heliconiinae have been tested and the highest
concentrations of cyanide have been found at the egg stage. The concentrations in the
iarval and pupal stages do not vary greatly, but cyanide concentration drops in adults
(Nahrstedt and Davis 1985).
The variation of the cyanogenic glycoside concentrations within body parts of adult
Heliconius, as related to their host plants, has not been explored. Within insects that are
chemically protected, it is expected that the greater concentrations of the toxic
compounds will be found in areas of the body with which the predators come in contact


80
Table 4-6. Qualitative analysis of cyanide release of whole body extracts tested with
different P-glucosidases by Feigl-Anger strips monitored over time in H. e.
favorinus and H. e. cyrbia. Numbers indicate, 0=no cyanide release, l=light
blue, slightly cyanogenic, 2=blue, cyanogen and 3=dark blue, very
N
30min
1 h
2 h
3 h
4 h
5 h
12 h
24 h
H. e.
favorinus
S punctata
6
0
0
1
1
1
1
0
0
9 punctata
6
1
2
2
2
2
1
0
0
0 rubra
6
0
0
0
0
0
0
0
0
$ rubra
6
0
0
0
0
0
0
0
0
S auriculata
6
0
1
1
2
2
2
1
1
9 auriculata
6
1
1
1
1
1
1
1
1
S trifasciata
6
0
0
0
0
0
0
0
0
2 trifasciata
6
0
0
1
0
0
0
0
0
0 biflora
6
0
0
2
2
2
2
1
1
9 biflora
6
0
0
1
1
1
1
0
0
H. e. cyrbia
S punctata
6
0
3
3
3
3
2
2
1
2 punctata
6
0
3
3
3
3
2
2
1
S rubra
6
0
3
3
3
3
3
2
1
9 rubra
6
0
1
2
2
2
2
2
1
S auriculata
6
0
1
2
3
3
2
1
1
P auriculata
6
0
1
2
2
2
3
2
1
S trifasciata
6
0
1
2
3
3
3
2
1
9 trifasciata
6
0
1
2
3
3
3
2
1
S biflora
6
0
1
2
2
2
2
2
1
9 biflora
6
0
1
2
2
2
2
2
1


LIST OF REFERENCES
Abbot, R. J. 1977. A quantitative association between soil moisture content and the
frequency of the cyanogenic forms of Lotus corniculatus L. at Birsay, Orkney.
Heredity 38: 397-400.
Abe, F., T. Yamauchi, K. Honda, H. mura, and N. Hayashi. 2001. Sequestration of
phenanthroindolizidine alkaloids by an Asclepiadaceae-feeding danaid butterfly,
Ideopsis similes. Phytochemistry 56: 697-701.
Benson. W. W., K. S. Brown, and L. E. Gilbert. 1976. Coevolution of plant and
herbivores: passion flower butterflies. Evolution 29(4):659-680.
Bernardo, J. 1993. Determinants of maturation in animals. Trends Ecol. Evolut. 8(5):
166-173.
Blum, M. S. 1981. Chemical Defense of Arthropods. Academic Press, New York. 562 pp.
Blum. M. S. 1983. Detoxification, deactivation, and utilization of plant compounds by
insects, pp. 265-275, in P. Hedin (Ed.) Plant Resistance to Insects. American
Chemical Society. Washington D.C.
Boggs, C. L. 1987. Ecology of nectar and pollen feeding in Lepidoptera, pp. 369-391, in
F. Slansky Jr. and J. G. Rodriguez (Eds.) Nutritional Ecology of Insects, Mites and
Spider, John Wiley and Sons. New York.
Boggs. C. L. 1981. Nutritional and life-history determinants of resource allocation in
holometabolous insects. Am. Nat. 117: 692-709.
Boggs. C. L. 1979. Resource Allocation and Reproductive Strategies in Several
Heliconiine Butterfly Species. Ph. D. Dissertation, University of Texas at Austin.
Boggs, C. L. and L. E. Gilbert. 1979. Male contribution to egg production in butterflies:
evidence for transfer of nutrients at mating. Science 206: 83-84.
Boinbardelli, E., A. Bonati. B. Gabetta, E. Martinelli, G. Mustich, and B. Danieli. 1975.
Passiflorine, a new glycoside from Passiflora edulis. Phytochemistry 14: 2661-
2665.
Bowers, M. D. 1988. Chemistry and coevolution: Iridoid glycosides, plants and
herbivorous insects, pp. 133-165, in K. Spencer (Ed.) Chemical Mediation of
Coevolution. Academic Press, London.
108


Pupal weight (g)
28
auriculata biflora punctata rubra
Host plant
A female
A male
B female
B male
trifasciata
Figure 2-6. Pupal weight of H. e.favorinus (A) and H. e. cyrbia (B) fed on natural and
alternative host plants. Different letters indicate significant differences
between means (Tukey HSD-test). Bars in each point indicate standard error.
auriculata biflora punctata rubra trifasciata
Host Plants
Figure 2-7. Average wing length of H. e.favorinus and H. e. cyrbia fed natural and
alternative host plants. Different letters indicate significant differences
between means (Tukey HSD-test). Bar in each point indicate standard error.


10
obtained from a company in Ecuador called Heliconius Works Quito, Ecuador. Some H.
e. cyrbia were also obtained from other colonies in the USA. Since I was not sure
whether this race was fed on its natural host plant as larvae. I raised H. e. cyrbia on their
natural host plants P. punctata and P. rubra for at least three generations prior to starting
my experiments.
The offspring from each female are referred to as family and were so tracked
through the experiments. This identification allowed me to recognize any life history
differences between offspring from different female butterflies. Such differences would
indicate genetic variations within populations for the trait in question. Analyzing the egg
mass reveals potential maternal effects.
Isolated females were allowed to lay eggs on their natural host plants. The eggs
were then collected and taken to environmental chambers for the hatching larvae to grow
under constant laboratory conditions (27C, 75% humidity and a 14L:10D photoperiod).
Life History Studies
Different aspects of the life histories were observed and contrasted between races
from the time the egg was laid until adult emergence from the pupae.
As the larvae were raised, they were checked at least three times a day (morning,
afternoon, night) for the presence of head capsule apolysis, or for evidence of ecdysis.
Larval developmental time in each instar was recorded. When larvae in the last instar
were observed in the position typical of pupation, that day was recorded as the first day
of pupation. Pupal developmental time was recorded to have ended when the adult
butterfly emerged from the pupal case.


61
Seventy 8-ml fractions were collected. Fractions were monitored by testing 50-pl of each
fraction with P-glucosidase enzymatic test with the Feigl-Anger strip to test for
cyanogenesis. Appropriate fractions were combined and evaporated in a speed vacuum
(Savant radiant cover PC210B, Savant Instruments Inc. Holbrook, NY)).
Purification by thin layer chromatography
A second purification method, thin layer chromatography (TLC), was then used. In
this case, the right edge of the TLC plate was used as a control and the material to be
purified was loaded on the rest of the plate. Merck silica gel 60 TLC plate (60-200 pm)
was the stationary phase and ethyl acetate-acetone-chloroform-methanol-water
(40:30:12:10:8) constituted the mobile phase. Cyanogenic compounds were monitored
by the sandwich method (Brimer et al. 1983) as follows. After development, the portion
of the plate containing the cyanogenic glycosides was covered with glass, and the control
side of the TLC plate was sprayed with an enzyme solution (aqueous solution of 13-
glucuronidase) obtained from Sigma. A pre-coated Polygram ion exchange sheet
(Polygram ionex 25-SB-Ac) impregnated with three different solutions (a saturated
solution of picric acid in water, a 1 M aqueous sodium carbonate solution and 2 % (w/v)
ethanolic 1 -hexadecanol solution) was used to capture the cyanogenic band overnight.
Location of the cyanogenic spot in the control side of the plate allowed me to scrape the
purified material on the covered side of the plate from the silica gel. The scraped area
was placed in a test tube and the cyanogen deabsorbed with methanol-water mixtures
(50%), vortexed and centrifuged for 20-30 min. The supernatant was collected, filtered
and evaporated in a speed vacuum. Rf values from the control cyanogenic band were
calculated.


22
Table 2-5. Mortality in different stages of the life history of H. e favorinus and H. e
Race
Egj
y
Host plant
Larvae
Pupae
Adult
N
%
dead
N
%
dead
N
%
dead
N
%
dead
H. e.
favorinus
8
5.6
auriculata
1
0.7
0
0
0
0
biflora
2
1.4
2
1.4
0
0
punctata
1
0.7
0
0
0
0
rubra
5
3.5
0
0
0
0
trifasciata
0
0
0
0
0
0
H. e.
cyrbia
10
6.75
auriculata
2
1.4
2
1.35
1
1.1
biflora
5
3.4
2
1.4
0
0
punctata
1
0.7
3
2.0
0
0
rubra
20
13.5
2
1.4
2
1.4
trifasciata
7
4.7
0
0
1
0.7


119
Saunders, J. A., E. E. Conn, H. L. Chin, and C. R. Stocking. 1977. Subcellular location of
the cyanogenic glycoside of Sorghum by autoradiography. Plant Physiol. 59: 647-
652.
Schappert, P. J and J. S. Shore. 2000. Cyanogenesis in Turnera ulmifolia L.
(Tumeraceae): II. Developmental expression, heritability and cost of cyanogenesis.
Evolutionary Ecology Research 2: 337-352.
Schappert, P. J and J. S. Shore. 1995. Cyanogenesis in Turnera ulmifolia L.
(Tumeraceae). I. Phenotypic distribution and genetic variation for cyanogenesis on
Jamaica. Heredity 74: 392-404.
Schoonhaven, L. M. 1972. Secondary plant substances and insects. Rec. Adv. Phytochem.
5: 197-224.
Seigler, D. S. 1991. Cyanide and cyanogenic glycosides, pp. 35-77, in G. A. Rosenthal,
M. R. Berenbaum (Eds.), Herbivores: Their Interactions with Secondary' Plant
Metabolites, Vol. I: The Chemical Participants, Academic Press, San Diego.
Self, L. S., F. E. Guthrie and E. Hodgson. 1964. Metabolism of nicotine by tobacco
feeding insects. Nature 204: 300-302.
Selmar, D., R. Lieberei, and B. Biehl. 1988. Mobilization and utilization of cyanogenic
glycosides. Plant Physiol. 86: 711-716.
Sheppard, P. M., J. R. G. Turner, K. Brown, W. W. Benson, and M. C. Singer. 1985.
Genetic and the evolution of muellerian mimicry in Heliconius butterflies.
Phil.Trans. Roy. Soc. Lond. 308B: 433-613.
Sinervo, B., P. Doughty, R. B. Huey, and K. Zamudio. 1992. Allometric engineering: a
causal analysis of natural selection on offspring size. Science 258: 1927-1930.
Slansky, F. 1992. Allelochemical-nutrient interactions in herbivore nutritional ecology,
pp. 135-174, in G.A. Rosenthal and M. R. Berenbaum (Eds.) Herbivores, Their
interactions with Secondary Plant Metabolites Vol II, Academic Press, San Diego.
Smiley, J. T. and C. S. Wisdom. 1985. Determinants of growth rate on chemical
heterogeneous host plants by specialist insects. Biochem. Syst. Ecol. 13:305-312.
Speed, M. P. and J. R. G. Turner. 1999. Learning and memory in mimicry: II. Do we
understand the mimicry spectrum? Biol. J. Linn. Soc. 67: 281-312.
Spencer, K. C. 1988. Chemical mediation of coevolution in the Passiflora-Heliconius
interaction, pp. 167-240, in K. C. Spencer (Ed.), Chemical Mediation of
Coevolution, Academic Press, New York.
Spencer, K. C. 1987. Specificity of action of allelochemicals: diversification of
glycosides. ACS Symp. Series 330: 275-288.


64
tube for analysis. The H-NMR spectrum (600.13 MHz) of the glycoside was originally
recorded at 27C and then subsequently at 17C, in order to move the OH resonance of
the solvent away from the H-4 resonance of the glycoside. The chemical shift axis was
referenced to internal, residual 'HCD20D which was assigned to 3.3 ppm. 2D-COSY
spectra were acquired with the same sample at both 27C and 17C.
HPLC/ (+) ESI-MS experimental
MS: All mass spectrometric data were acquired with a Thermo-Finnigan (San Jose,
CA) LCQ Classic quadrupole ion trap mass spectrometer operated in the electrospray
ionization (ESI) mode, typically under the following conditions: sheath gas (N2) = 60;
aux gas (N2) = 5; spray voltage = 3.3 kV; capillary temperature = 250C; capillary
voltage = 15 V; tube lens offset = 0 V. In order to enhance the sensitivity and obtain
dependent MS" (tandem mass spectrometry) data on different components, the (+) ESI-
normal mass spectra were normally acquired over two ranges: for example, m/z 75-210
for expected amino acids and m/z 205-1200 for higher molecular weight components,
including cyanogenic glycosides. Tandem mass spectrometry (MSn, n=2, 3,..) was
normally performed in a dependent fashion (i.e., under software control) on the most
intense ion from the preceding normal mass spectrum. The MSn spectra were normally
acquired with a parent ion isolation of 4-6 u (u=atomic mass unit) and a normalized %
CID (collision induced dissociation) energy of 35-40%.
HPLC: High performance liquid chromatography (HPLC) was conducted with an
Agilent HP 1100 series binary pump system. Mobile phase A was 0.5% formic acid
(Fisher Scientific, 88%; Certified, A.C.S.) and 5 mM ammonium formate (Fisher
Scientific, Certified) in water (Fisher Scientific, HPLC grade). Mobile phase B was


12
affected by the split, so one can use the height of the last instar head capsule to indicate
growth during the last instar. The following formula was used to calculate growth rate:
Growth rate = (In (Hf) In (H¡))/t
where Hf is the measurement of the fifth instar head capsule height; H¡ is the
measurement of the first instar head capsule height; and t is the larval developmental
time. A natural logarithm transformation was performed as recommended by LaBarbera
(1989).
Pupae and adult offspring measurement
Each pupa was weighed two days after pupation. Measurements of pupa length
and width also were taken.
For the adult offspring, I recorded the following data: sex, wing length, total body
length (measured from the head to the tip of the abdomen), and fresh weight.
Measurements were made in the afternoon to allow time for the new adults to dry their
wings after emergence. A Vernier caliper (Spi 2000) graduated to 0.1 mm was used for
those measurements. Mass measurements were made with an electro balance. Teneral
butterflies were then stored at -85C for further analysis.
Statistical Analysis
Systat program V 10 was used for the statistical analysis. Since the variation in the
head capsule size between instars w'as large, the measurements were converted to a
logarithmic scale to do the statistical analysis. Repeated measures ANOVA was used for
analyzing growth trajectories between races. The statistical test was performed on both
head capsule width (using the measurement from the first to the fourth instar) and head
capsule height. Since both yielded the same results, I present here the results of the
analysis of head capsule height. Finally ANOVA was used for the rest of the analysis. In


Table 4-i 1. Percent total peak areas of measured free amino acids in butterfly samples.
Butterfly sample
% Total peak area of measured amino acids
Valine
Leu/IsoLeu*1
Tyrosine
Phenylalanine
Tryptophan
Total amino
acids
$ H. e. cyrbia fed auriculata
3.13
58.72
23.19
0.45
14.51
100.00
S H. e. cyrbia fed auriculata
0.77
89.55
0.00
9.68
0.00
100.00
$ H. e. cyrbia fed punctata
31.70
44.87
1.90
13.66
7.88
100.00
9 H. e. cyrbia fed rubra
4.05
93.90
0.11
1.94
0.00
100.00
9 H. e. cyrbia fed biflora
12.58
53.41
3.72
28.53
1.76
100.00
9H. e.favorinus fed trifasciata
15.98
53.91
1.54
20.64
7.93
100.00
9H. himera fed punctata
19.35
38.96
0.55
40.36
0.77
100.00
9H. himera fed rubra
23.38
49.58
3.02
22.47
1.56
100.00
Cannot be distinguished under the given conditions.
OO


39
the cyanide concentrations did not differ between the sexes. However, there was a
significant difference in concentrations between body parts (F=12.31; df=3, 52; p<0.001).
The thorax and the head had higher cyanide concentrations, as compared to the abdomen
and wings (Figure 3.4).
H. e. favorinus fed P. trifasciata
Low concentrations of cyanide were found within the body of this race. Although
the female thorax and head seemed to have higher cyanide concentration than the males,
the statistical analysis did not show any significant differences between these values.
Wings and abdomen had the lowest concentrations of cyanide. More males than females
were tested because more than 50% of the males tested negative. Addition of exogenous
p-glucosidase did not favor the release of cyanide, indicating that cyanide release is not
due to the lack of the enzyme, but to the lower concentrations of cyanogenic glycosides
within the butterflys body (Figure 3.5).
H. e. demophon fed on natural and alternative hosts plants
Host plant selection had a significant effect on the cyanide concentration found in
//. e. demophon (F=10.15; df=3, 50; pO.OOl). Butterflies fed P. rubra and P. biflora
exhibited significantly lower concentrations of cyanide than butterflies fed P. auriculata
and P. trifasciata. Also, the interaction between host plant and sex was statistically
significant (F=7.34; df=3, 52; pO.OOl). The natural host plant, P. biflora, is associated
with lower cyanide concentrations in the male butterflies. Furthermore, cyanide
concentrations varied between body parts (F=4.96; df=l, 52; p=0.03). In general, the
body (which in these samples includes the head, thorax and abdomen) possessed higher
cyanide concentrations than the wings (Figure 3.6).


88
Figure 4-4. Proton/Carbon heteronuclear multiple quantum coherence (HMQC) spectrum
of cyanogen passicapsin in CD3OD from P. rubra.


25
Figure 2-3. Study organisms. (A) H. e.favoritms from Tingo Maria, Huanuco, Per and
(B) H. e. cyrbia from Pichincha, Ecuador.


75
Table 4-1. Qualitative analysis of the extracts of Passiflora plants. Number 0=no cyanide
release, l=light blue, slightly cyanogenic, 2=blue, cyanogen and 3=dark blue,
very cyanogenic. .
Passiflora
Less
than 30
min
30min
1 h
2 h
3 h
4 h
5 h
12 h
24 h
auriculata
0
1
1
1
1
2
3
2
2
biflora
0
0
1
1
1
1
2
2
2
punctata
0
1
1
1
2
2
3
2
2
rubra
2
2
2
3
3
3
3
3
3
trifasciata
lo-
0
0
1
1
1
2
2
2


3-6. Cyanide quantification of male and female H. e. clemophon ted natural (P. biflora)
and alternative host plants. Bars indicate standard error 50
3-7. Regression of the mean cyanide concentration of H. erato over the mean cyanide
concentration of the natural Passiflora host plants upon which they fed. (The
means were used in the figure to enhance the clarity of presentation.) 51
4-1. Cyanogenic glycosides in Passifloraceae plants (A) Type I. (B) Type II.
(C) Type III and (D) Type IV (Jaroszewski et al. 2002) 85
4-2. Heliconius used in this study (A) H. e. cyrbia, (B) H. e.favorinus, and (C) H.
himera 86
4-3. Passiflora plants used in this study (A) P. auriculata, (B) P. punctata, (C) P. rubra.
(D) P. biflora, (E) P. trifasciata 87
4-4. Proton/Carbon heteronuclear multiple quantum coherence (HMQC) spectrum of
cyanogen passicapsin in CD3OD from P. rubra 88
4-5. Proton/Carbon heteronuclear multiple quantum coherence (HMQC) spectrum of
cyanogen passitrifasciatin in CD3OD from Passiflora sp 89
4-6. Potential enzymatic hydrolysis products of passibiflorin 90
4-7. Potential enzymatic hydrolysis products of passicapsin 91
5-1. Heliconius himera from Vilcabamba, Loja Province, Ecuador 103
5-2. Electrophoresis in SDS-10% gel slab. The gel shows different levels of purification
of the protein studied. CE: is the cytosolic crude extract, Q sepharose elutes:
Q1-Q3 fractions 31-33, Mono-Q elutes: Ml-M2, fractions 23 and 24 from Q-
sepharose fraction 31, and M3-M4, fractions 24 and 25 from Q-sepharose
fraction 32. RM is the rainbow marker 104
5-3. Comparison of amino acid sequence of Spodoptera frugiperda (Lepidoptera:
Noctuidae) p-glucosidase with mass spectroscopy identified peptides from H.
himera. The letters in bold indicate the homologous amino acids sequencies
from H. himera that match with S. frugiperda. A. Mr 55.000 and B. Mr 65,000. 105
xiv


68
Cyanogen profile of Passiflora punctata
This plant is currently being analyzed by Dr J. Jaroszewsky (personal
communication). However, from the HPLC/MS analysis, we know that this plant
contains passibiflorin with a MW of 433 and a retention time of 34.00 min.
Cyanogenic Glycosides of Butterflies Fed Natural Host Plants
A (3-glucosidase test indicated that there were no differences in cyanogenic
glycoside types between the sexes in either of the two races. All butterflies possessed
aliphatic cyanogenic glycoside, except for H. e. favorinus that tested acyanogenic when
fed on P. rubra and P trifasciata. Based on the results presented in Chapter 3, it is
known that H. e. favorinus contains low amounts of cyanogenic glycosides. Aliphatic
and Type I glycosides were present when males and females from the two races were fed
P. auriculata. Cyanogenesis occurred within 30 minutes in extracts of H. e. favorinus fed
P. punctata and P. auriculata, while cyanogenesis reaction occurred within one hour in
almost all tested butterflies samples (Table 4-6).
No glycosides were found in larval fecal pellets, suggesting that the ingested
glycosides were absorbed by the larvae, and, if excreted, metabolized prior to excretion.
[3-glucosidase classification tests were not able to detect small quantities of
aliphatic and Type II glycosides in butterfly samples, but these were later detected by the
HPLC/MS method. This suggests that the P-glucosidase test is not conclusive and it
should be used only to detect high concentrations of cyanogenic glycosides.
De novo synthesis
HPLC/MS showed that all butterflies contained aliphatic cyanogenic glycosides
(linamarin and lotaustralin), even the ones that tested acyanogenic with the P-glucosidase
enzymatic test (Table 4-6, 4-7 and 4.8). There were variations among the aliphatic


105
A
1 MKLLVVLSLV AVACNASIVR QQRRFPDDFL FGTATASYQI EGAWDEDGKG ENIWDYMVHN
61 TPEVIRDLSN GDIAADSYHN YKRDVEMMRE LGLDAYRFSL SWARILPTGM ANEVNPAGIA
121 FYNNYIDEML KYNITPLITL YHWDLPQKLQ ELGGFANPLI SDWFEDYARV VFENFGDRVK
181 MFITFNEPRE ICFEGYGSAT KAPILNATAM GAYLCAKNLV TAHAKAYYLY DREFRPVQGG
241 QCGITISVNW FGPATPTPED EMAAELRRQG EWGIYAHPIF SAEGGFPKEL SDKIAEKSAQ
301 QGYPWSRLPE FTEEEKAFVR GTSDLIGVNH YTAFLVSATE RKGPYPVPSL LDDVDTGSWA
361 DDSWLKSASA WLTLAPNSIH TALTHLNNLY NKPVFYITEN GWSTDESREN SLIDDDRIQY
421 YRASMESLLN CLDDGINLKG YMAWSLMDNF EWMEGYIERF GLYEVDFSDP ARTRTPRKAA
481 FVYKHIIKHR VVDYEYEPET MVMTIDEGH
B
1 MKLLVVLSLV AVACNASIVR QQRRFPDDFL FGTATASYQI EGAWDEDGKG ENIWDYMVHN
61 TPEVIRDLSN GDIAADSYHN YKRDVEMMRE LGLDAYRFSL SWARILPTGM ANEVNPAGIA
121 FYNNYIDEML KYNITPLITL YHWDLPQKLQ ELGGFANPLI SDWFEDYARV VFENFGDRVK
181 MFITFNEPRE ICFEGYGSAT KAPILNATAM GAYLCAKNLV TAHAKAYYLY DREFRPVQGG
241 QCGITISVNW FGPATPTPED EMAAELRRQG EWGIYAHPIF SAEGGFPKEL SDKIAEKSAQ
301 QGYPWSRLPE FTEEEKAFVR GTSDLIGVNH YTAFLVSATE RKGPYPVPSL LDDVDTGSWA
361 DDSWLKSASA WLTLAPNSIH TALTHLNNLY NKPVFYITEN GWSTDESREN SLIDDDRIQY
421 YRASMESLLN CLDDGINLKG YMAWSLMDNF EWMEGYIERF GLYEVDFSDP ARTRTPRKAA
481 FVYKHIIKHR VVDYEYEPET MVMTIDEGH
Figure 5-3. Comparison of amino acid sequence of Spodoptera frugiperda (Lepidoptera:
Noctuidae) P-glucosidase with mass spectroscopy identified peptides from H.
himera. The letters in bold indicate the homologous amino acids sequencies
from H. himera that match with S. frugiperda. A. Mr 55,000 and B. Mr
65,000.


54
b. Metabolism: Toxic compounds may be converted into less harmful derivatives
through solubilization or conjugation. The metabolites are then reused in the
intermediary metabolic pathway of the organism, or they can be converted to a
metabolic product that can be stored or excreted. The first alternative is common
for toxic compounds that are highly water soluble such as toxic amino acids,
whereas lipid soluble toxins are converted to water soluble metabolites and then
eliminated as such, rather than being recycled (Brattsten 1992).
c. Sequestration: Secondary compounds are deposited intact in particular sites of the
body. The substance to be sequestered is absorbed through the gut membrane,
transported by the hemolymph and stored in body tissues or special glands (Nishida
2002). Sequestration of plant secondary compounds, in general, is used as a
defense against predators and it has been described in many herbivorous insects
such as the Lepidoptera, Coleptera, Hemiptera and Hymenoptera (Brower 1984,
Bowers 1992, Rothschild 1972). Most of the insects found to sequester toxic
compounds specialize on one or a very few plants, from which they extract their
defensive compounds (Bowers 1992).
Cyanogenic Glycosides in the Family Passifloraceae
The chemistry of Passiflora plants has been studied since the 1960s. Some
Passiflora have been found to contain simple harmane alkaloids (Copeland and Slaytor
1974), flavonoids (McCormick and Mabrick 1981, Ulubelen et al. 1982), a distinctive
triterpene glycoside (Bombardelli et al. 1975), and tannins (Smiley and Wisdom 1985).
However, the Passiflora are unusual in producing cyanogenic glycosides with a
cyclopentene moiety (Spencer and Seigler 1984, 1985a, 1985b, Olafsdottir et al. 1989a,
Jaroszewski et al. 2002).
Chemically, the cyanogenic glycosides in the Passifloraceae have been divided into
four distinct structural classes (Jarozewsky et al. 2002)
a. Type I (called by Spencer 1988, simple cyclopentenoid). These glycosidic
molecules are peculiar because they usually (if not always) occur as pairs of p -D
glucopyranosides having enantiomeric aglycones, i.e., tetraphyllin B co-occurs
with volkenin; epivolkenin co-occurs with taraktophyllin; and tetraphyllin A co
occurs with deidaclin (Figure 4-1 A).
b. Type II (includes the sulfated cyanogens and complex cyclopentenoid (Spencer
1988)). These glycosides are similar in structure to Type I, but they either have an
unusual sugar molecule, sulfate group, or additional oxygenation of the


91
/
OH
l-D-glucopyranosyl
Passicapsin (1S.4R)
Molecular formula = C18 H27 N 01C
Molecular Weight = 417.1635
2,6-dideoxy-p-D-xy/o-hexoyranosyl
.N
HO
III
1..
OH H3£
1 \
4" 1.
\,
L
MW 255 y '"'2
* 11
\3
-O
\ ,,a
\ A
\^1
H
///
H0\_
OH
\ -O
I
MW 417
< "1
OH H3C
MW 239 c/ 'jj
I \ \ ,..3
VA /\
I \
OH
OH h3C
I \
V \ /\
V j-tv \| .n h
H
OH
OH
HO^ \ 5'
ho:
OH
H0-"
H\.
OH
MW 271
4.
OH
MW 287
,in
'X
1.,
\ J
4 *
/\
HO H
Figure 4-7. Potential enzymatic hydrolysis products of passicapsin.


44
summer and fall seasons. The result of the distribution of cyanogens within the plant is
consistent with the optimal allocation theory (McKey 1974) that states that vulnerable
plant tissues and organs are defended more than older senescing ones. It has already been
demonstrated elsewhere that for several cyanogenic species, young developing tissues
contain the highest amounts of cyanogenic glycosides (Conn 1981a).
Implications of Cyanide Concentration for Palatability
It seems that H. e. favorinus is better adapted than H. e. cyrbia to handle different
compounds because its larval survival rate is higher, whether feeding on natural or
alternative host plants (as has been demonstrated in Chapter 2). H. e. favorinus' superior
ability to handle different chemical compounds has both benefits and costs in terms of
survival. The survival benefit is reduced larval mortality due to plant toxicity as the
larvae is less likely to be killed by the plants chemical defenses. The survival cost is
increased palatability (the larvae may be less toxic), which increases the potential for
larval mortality from predation (e.g., by Hyla frogs). H. e. cyrbia, on the other hand, has
the advantage, perhaps, of being more toxic, implying better protection against predation,
but at the cost of high mortality during larval development.
Implications of Cyanide Concentration for Mimicry
Different concentrations of cyanide within a butterflys body produce a palatability
spectrum (Brower et al. 1967). The palatability spectrum theory states that if members of
a Mllerian mimicry complex are not equally distasteful, then the presence of individuals
that are less distasteful would lead to an increase in predation on the models, and this will
result in an overall weakening of the Mllerian mimicry.
H. erato and its co-mimic H. melpomene are members of a Mllerian mimicry ring
of races in the South and Central America rainforests. There is some disagreement as to


CHAPTER 6
CONCLUSIONS
Through their respective coevolutionary races with their host plants, Heliconius
erato and H. himera have developed different strategies to deal with toxic compounds
contained in their cyanogenic host plants in the genus Passiflora. Subspecies of H. erato
from the eastern and the western sides of the Andes differ in body size and this difference
is not due to nutritional differences, but is genetically based. When two subspecies, one
from the eastern side of the Andes (//. e. favorinus) and a second from the western side of
the Andes (//. e. cyrbia), were subjected to feeding on different host plants (natural and
alternative hosts), the larvae showed plastic growth trajectories. Feeding on one of the
host plants, P. rubra, resulted in lower growth rates in both subspecies and higher
mortality in H. e. cyrbia, but not in H. e. favorinus. As a result, the few individuals from
the former subspecies that survived while feeding on this highly toxic cyanogenic plant
were probably better protected against predators because they accumulated higher
concentrations of cyanogenic glycosides in their bodies, a condition which in turn
generates higher concentrations of hydrogen cyanide. H. e. favorinus, on the other hand,
turned out to be less cyanogenic. As a result, individuals of the subspecies that were less
i
cyanogenic were the target of predation by frogs of the genus Hyla. The results of this
study showed that two different geographical races of H. erato utilize different strategies
while feeding on similar host plants.
By studying the chemical composition of cyanogenic plants and butterflies, I
learned that H. erato not only synthesized de novo aliphatic cyanogenic glycosides, but
106


42
naive predator will rapidly learn to avoid feeding on that particular pattern of wing
coloration (Chai 1988), and flight (Srygley and Chai 1990) and the members that
constitute a Mllerian ring will benefit from it. The selection for belonging to the rings
common pattern will be reinforced. A similar mechanism of avoidance is likely used for
the larval stage, however a definitive test was not run. Hyla frogs were more consistently
found in the cage of H. e. favorinus, than in the cages of H. e. cyrbia and H. himera,
suggesting that larvae of H. e. favorinus were more palatable. On the other hand, being
highly unpalatable has a negative aspect as well, as we observed with the members of H.
e. cyrbia fed with P. rubra. Members that fed on this particular host plant gained the
benefit of being distasteful, but at the same time, there was a high cost paid for this
benefit, because individuals feeding on P. rubra spent a longer period feeding on the
host, which increases the time they are exposed to parasitoids. On average, individuals
feeding on such plants are slightly smaller in size than comparable insects feeding on
other host plants.
H. e. favorinus. seems to be better adapted to detoxifying and tolerating different
chemicals in the plants, probably due to more efficient degradative enzymatic activity, or
higher tolerance of chemicals, for example by having a target-site insensitivity (Brattsten
1979, Lindroth and Weisbrod 1991) because excreta was not found to be cyanogenic.
The differences between strategies for coping with cyanogenic compounds probably
reflect genetic variation between the races of H. erato.
Lastly, it has been demonstrated that, at least for H. e. cyrbia, there is a limit range
of toxicity because of high mortality at the larval stage, implying that there is a level of
effectiveness of cyanogenic defenses in the plants in relation to a herbivore. However,


112
Dunlap-Pianka H. L., C. L. Boggs, and L. E. Gilbert. 1977. Ovarian dynamics in
heliconiine butterflies: programmed senescence versus eternal youth. Science
197:487-490.
Ehrlich, P. R. and P. H. Raven. 1964. Butterflies and plants: a study of coevolution.
Evolution 18:586-608.
Ellis, W. M., R. J. Keymer, and D. A. Jones. 1977. The effect of temperature on the
polymorphism of cyanogenesis in Lotus corniculatus L. Heredity 38: 339-347.
Eltringham, H. 1916. On specific and mimetic relationships in the genus Heliconius.
Trans. Ent. Soc. Lond. 1916: 101-148.
Emsley, M. G. 1965. Speciation in Heliconius (Lep. Nymphalidae): morphology and
geographic distribution. Zoolgica, New York 50:191-254.
Engler, H. S., K. C. Spencer, and L.E. Gilbert. 2000. Preventing cyanide release from
leaves. Nature 406: 144-145.
f alconer, D. S. 1989. Introduction to Quantitative Genetics. 3rd ed. Longman. London.
Leeny, P. 1992. The evolution of chemical ecology: Contributions from the study of
herbivorous insects, pp. 1-44, in G.A. Rosenthal and M.R. Berenbaum (Eds),
Herbivores, Their Interactions with Secondary Plant Metabolites Vol II, Academic
Press, San Diego.
Feigl, F. and V. Anger. 1966. Replacement of benzidine by copper ethylacetoacetate and
tetra base as spot-test reagent for hydrogen cyanide and cyanogen. Analyst 91:
282-284.
Ferreira, C., B. B. Torres, and W. R. Terra. 1998. Substrate specificities of midgut (3-
glycosidases from different insect orders. Comp. Biochem. Physiol. 119B (1): 219-
225.
Ferreira, C., J., R. P. Parra, and W. R. Terra. 1997. The effect of dietary plant glycosides
on larval midgut P-glucosidases from Spodoptera frugiperda and Diatraea
saccharalis. Insect Biochem. Mol. Biol. 27(1): 55-59.
Fisher, R. A. 1930. The Genetical Theory of Natural Selection. 2nd ed. Dover.
1 isher, F. C., S. Y. Fung, and P. P. Lankhorst. 1982. Cyanogenesis in Passifloraceae.
Cyanogenic glycosides from Passiflora capsularis, P. warmingii and P. perfoliata.
Planta Medica 45: 42-45.
Foulds, W. and J. P. Grime. 1972. The influence of soil moisture on the frequency of
cyanogenic plants in populations of Trifolium repens L. and Lotus corniculatus L.
Heredity 28: 143-146.


114
Higgins, L. E. and M. A. Rankin. 1996. Different pathways in arthropod postembryonic
development. Evolution 50(2): 573-582.
Hosel, W. 1981. The enzymatic hydrolysis of cyanogenic glycosides, pp. 217-232, in B.
Vennesland, E. E. Conn, C. Knowles, J. Westley, and F. Wissing (Eds.), Cyanide in
Biology, Academic Press, London.
Hruska, A. J. 1988. Cyanogenic glvcosides as defense compounds. J. Chem. Ecol.
14(12): 2213-2217.
Hughes, M. A. 1981. The genetic control of plant cyanogenesis, pp. 495-508, in B.
Vennesland, E.E. Conn, C. J. Knowles, J. Westly, and F. Wissing (Eds.). Cyanide
in Biology. Academic Press, London.
jaroszewsky, J. W.. E. S. Olafstottir, P. Wellendorph, J. Christensen, H. Franzyk, B.
Somanadhan, B. A. Budnik, L. B. Jorgensen, and V. Clausen. 2002. Cyanohydrin
glycosides of Passiflora: distribution pattern, a saturated cyclopentene derivative
from P. guaiemalensis, and formation of pseudocyanogenic a-hydroxyamides as
isolation artifacts. Phytochemistry 59: 501-511.
jaroszewsky, J. W., P. S. Jenzen, C. Cornett, and J. R. Biberg. 1988. Occurrence of
lotaustralin in Berberidopsis beckleri and its relation to the chemical evolution of
Flacourtiaceae. Biochem. Syst. Ecol. 16: 23-28.
Jaroszewsky, J. W., E. S. Olafstottir, C. Cornett, and K. Schaumburg. 1987a.
Cyanogenesis of Adenia volkensii Harms and Tetrapathaea tetranda Cheeseman
(Passifloraceae) revisited: tetraphyllin B and volkenin. Optical rotary power of
cyclopentenoid cyanohydrin glucosides. Acta Chem. Scand. 41B: 410-421.
Jaroszewsky, J. W., J. V. Anderson, and I. Billeskov. 1987b. Plants as a source of chiral
cyclopentenes: taraktophyllin and epivolkenin, new cyclopentenoid cyanohydrin
glucosides from Flacourtiaceae. Tetrahedron 43: 2349-2354.
Jiggms, C. D., W. O. McMillan, and J. Mallet. 1997. Host plant adaptation has not played
a role in the recent speciation of Heliconius himera and Heliconius erato. Ecol.
Entomol. 22: 361-365.
Jones, D. A. 1998. Why are so many food plants cyanogenic? Phytochemistry 47(2): 155-
162.
Jones, D. A. 1988. Cyanogenesis in animal-plant interactions, pp. 151-176, in D. Evered
and S. Harnett (Eds), Cyanide Compounds in Biology, Ciba Foundation
Symposium, Wiley, Chichester.
Jones, D. A. and A. D. Rammani. 1985. Altruism and movement of plants. Evol. Theor.
7: 143-148.


65
either (a) 0.5% formic acid in methanol (Burdick & Jackson (Muskegon, MI), B&J Brand
High Purity Solvent) or (b) methanol without any modifiers. The column used was a
Phenomenex (Torrance, CA) Synergi Hydro-RP column (2 x 150 mm; 4p; 80 A) with the
equivalent guard column (2x4 mm). A number of different gradients were used to
achieve the best separation of all components of the butterfly extracts. The following
gradient was used at a mobile phase flow of 0.15 ml/min:
A:B(min) = 100:0(0-5) => 60:40(45) => 0:100(75-120) => 100:0(130-165)
UV: During some of the initial analyses, an Applied Biosystems (Foster City, CA)
Model 785A, programmable absorbance detector was interfaced between the column
effluent and the mass spectrometer. The absorbance at 254 nm was monitored. After the
initial analyses, the UV was no longer used and the column effluent was connected
directly to the mass spectrometer.
Post-column mobile phase modification: The efficiency of electrospray
ionization was low when spraying 100% aqueous mobile phase. Often, methanol or 0.5%
formic acid in methanol was added to the column effluent via a PEEK tee-union. The
post-column modifier was either provided by Applied Biosystems model 400 solvent
delivery system at ~ 0.1 ml/min or via the LCQs syringe pump at 20 pl/min. The latter
was found to be more reproducible, as the Applied Biosystems pump did not provide
steady flow at 0.1 ml/min and under.
Amino Acid Analysis
The following standard amino acids were used for the amino acid analysis:
isoleucine, leucine, phenylalanine, tryptophan, tyrosine and valine. The amino acids
were combined with linamarin, which was used as a marker, because the retention time


36
The tubes were vortexed vigorously and the solution was allowed to sit for 10 min.
The absorbance was then measured at 580 nm in a spectrophotometer (Spectro 22;
Labomed, Culver City, CA).
Standard curve
For the standard curve, a 0.02 M solution of NaCN in 0.1 N NaOH (0.980g
NaCN/1) was prepared. The exact amount of CN in solution was determined by a
modified Liebig titration method as follows:
A standard 0.1 M AgNC>3 solution was prepared by diluting dry AgNC>3 (at 100C
for 3 h) in 1 L deionized water in a volumetric flask. The solution was stored in the dark.
Twenty five ml of the NaCN solution was placed in a 50 ml flask equipped with a stirring
bar. Into this solution, 1.5 ml 6 N NH4OH and 0.5 ml 10% KI were added. This mixture
was then titrated with the standard AgNC>3 solution, until the mixture became turbid,
which occurred at approximately 2.4 ml of AgNC>3. The endpoint was easier to see
against a black background.
To determine the amount of CN in the 25 ml sample, the amount of AgNC>3 used
to reach the endpoint was multiplied by 0.005204 g, then divided by 25. The result
reflected the g CN /ml.
Aliquots (1 ml) of the NaCN solution prepared above (close to 500 pg CN_/ml)
were then diluted in a volumetric flask with 0.1 N NaOH to 100 ml to reach a final
concentration of approximately 5 pg CN /ml.
The standard curve was then prepared using aliquots of the standardized NaCN
solution. Between 0 and 0.2 ml of NaCN are increased to 1 ml with 0.1 M NaOH. This
gives CN concentrations ranging from 0 to approximately 1 pg CN~/ml.


87
Figure 4-3. Passiflora plants used in this study (A) P. auriculata, (B) P. punctata, (C) P.
rubra, (D) P. biflora, (E) P. trifasciata.


LIST OF TABLES
Table E^ge
2-1. Average wing length and standard error of geographical races of Heliconius eruto
from museum specimens in different regions 20
2-2. Repeated measurements ANOVA of head capsule size with race, host and family
as factors 20
2-3. Repeated measurements ANOVA of larval developmental time with race and host
as factors 21
2-4. Sexual differences in larval developmental time (LDT), growth rate (GR) and
wing length (WL) in H. e. favorinus and H. e. cyrbia feeding on five different
host plants 21
2-5. Mortality in different stages of the life history of H. e. favorinus and H. e. cyrbia
feeding on natural and suitable host plants 22
3-1. Heliconius eruto fed natural and alternative Passifloru host plant and sample size
used in the study 46
3-2. Passifloru plants tested for cyanogenesis in this study. All plants are used by
Heliconius eruto in different geographical areas 46
4-1. Qualitative analysis of the extracts of Passifloru plants. Number 0=no cyanide
release. l=light blue, slightly cyanogenic, 2=blue, cyanogen and 3=dark blue,
very cyanogenic 75
4-2. id NMR(a) spectrum of a cyanogen, identified as passicapsin, from P. rubra in
CD3OD. Chemical shifts (ppm) and coupling constants (Hz). Degree of splitting
is identified by: (s) single, (d) doublet, (q) quartet, (m) multiplet 76
4-3. 1 C NMR,a> spectrum of a cyanogen, identified as passicapsin, from P. rubra in
CD3OD 77
4-4. 'H NMR(a) spectrum of a cyanogen, identified as passitrifasciatin, from
P trifasciata in CD3OD. Chemical shifts (ppm) and coupling constants (Hz).
Degree of splitting is identified by: (d) doublet, (t) triplet, (m) multiplet 78
xi


LIST OF FIGURES
Figure page
2-1. Life history of Heliconius butterflies 23
2-2. Museum specimens of H. erato from the eastern (left column) and western sides
(right column) of the Andes that were included in the analysis. (A) H. e. phyllis,
(B) H. e. favor inns, (C) H. e. magnificus, (D) H. e. colombina, (E) hydara,
(F) H. e. cyrbia, (G) H. e. petiveranus 24
2-3. Study organisms. (A) H. e. favor inns from Tingo Maria. Huanuco, Per and
(B) H. e. cyrbia from Pichincha, Ecuador 25
2-4. Growth trajectories of (A) //. e.favorinus and (B) H. e. cyrbia feeding on five
different host plants. Each point represents the mean value of larvae from the
first to fifth instar. Bars on each point indicate standard error 26
2-5. Growth trajectories of H. e.favorinus and H. e. cyrbia fed their natural host plants
(P. trifasciata and P. rubra, respectively). Each point represents the mean value
of larvae from the first instar to the fifth. Bars on each point indicate standard
error 27
2-6. Pupal weight of H. e.favorinus (A) and //. e. cyrbia (B) fed on natural and
alternative host plants. Different letters indicate significant differences between
means (Tukey HSD-test). Bars in each point indicate standard error 28
2-7. Average wing length of H. e. favorinus and H. e. cyrbia fed natural and alternative
host plants. Different letters indicate significant differences between means
(Tukey HSD-test). Bar in each point indicate standard error 28
3-1. General process of enzymatic hydrolysis of cyanogenic glycosides 47
3-2. Cyanide quantification of Passiflora plants in the (A) summer and (B) fall seasons.
Bars indicate standard error 48
3-3. Cyanide quantification of male and female H. e. cyrbia fed P. rubra. Different
letters indicate significant differences between means (Tukey HSD-test). Bars
indicate standard error 49
3-4. Cyanide quantification of male and female H. e. cyrbia fed P. punctata. Different
letters indicate significant differences between means (Tukey HSD-test). Bars
indicate standard error 49
3-5. Cyanide quantification of male and female H. e.favorinus fed P. trifasciata.
Different letters indicate significant differences between means (Tukey HSD-test).
Bars indicate standard error 50


13
all cases, family was nested within race to observe either maternal or genetic variation.
Also, sex was used as factor and if the interaction was not statistically significant it was
excluded from the analysis. A Tukey test was used for multiple comparisons.
Results
Analysis of Specimens from the Museum Collection
The ANOVA of the measurements taken from the museum specimens supports the
hypothesis that subspecies from the eastern side of the Andes tend to be larger than the
subspecies from the western side of the Andes (Table 2-1). Regions alone had a large
effect on size (F=254.2; df= 1, 179; pO.OOl), and different races of erato within the
regions are also different (F=14.4; df= 5, 179; pO.OOl).
The subspecies from the eastern region are relatively constant with respect to their
body size. In contrast, the subspecies from the western side of the Andes show more
variability in their sizes.
Life History Studies in H. e.favorinus and H. e. cyrbia
Data collected in the field (by myself and museum specimens) indicated that H. e.
favorinus is larger than H. e. cyrbia (ANOVA for wing length F= 283.8, pO.OOl and
body length F=45.3, pO.OOl) (Figure 2-3).
Eggs
Heliconius erato egg mass was not significantly different between races.
However, egg mass was strongly influenced by family within each race (F= 5.69,
pO.OOl). This result indicates maternal effects for egg mass. Both races laid eggs
singly, mostly on young growth (tendrils, and the tips of developing natural Passiflora
host plants). The eggs are cylindrical, yellow, with a pointed apex resembling a crown. I
did not find any difference in egg morphology between races. First instars hatch from the


CHAPTER 1
INTRODUCTION
Heliconius erato is a brightly colored, highly variably patterned butterfly that is
distributed from Central to South America. H. erato is adapted to altitudes that range
from sea level to 1,600 m. It is frequently found flying in open pastures in disturbed
forest and secondary growth (DeVries 1987). This species is unpalatable to predators and
shows strong Mllerian mimicry with other distasteful species. Geographical races have
different patterns of coloration that range from orange and black, to red, yellow and
black, to iridescent blue and pink (this pattern of coloration is shared with the co-mimic
H. melpomene). The races have diversified broadly and there are now approximately 30
parapatric races that are able to mate randomly in narrow hybridization zones (Emsley
1964. Mallet 1986).
In this study, three races of erato were observed: H. e. cyrbia from Pichincha,
Ecuador; H. e. favorinus from Tingo Maria, Huanuco, Peru; and H. e. demophon from
Panama. Heliconius himera (used in experiments described in Chapter 5) constitutes an
intermediate step in the transition from a race to a species (McMillan et al. 1997). It had
previously been considered a race of erato (Lamas 1976, Brown 1979, Sheppard et al.
1985), but recent mtDNA studies suggest a divergence approximately 1 million years ago
(Brower 1994; see also Emsley 1965 for arguments that it is actually a separate species).
Phylogenetic studies of H. erato group the races into two main clades. One clade
groups races that are located on the eastern side of the Andes, while the second clade can
be found on the western side of the Andes (Brower 1994). The ecology (Brown 1979,
1


95
approximately 1 million years ago. Also, hybridization studies (Jiggins et at., 1997,
McMillan et al. 1997) suggest that genomic incompatibilities appeared in the earliest
stages of speciation.
Insect Rearing Procedure
Larvae of H. himera were reared under controlled environmental conditions at
28C, 75% humidity and 14L:10D photoperiod. Each individual larva was fed P. biflora.
No artificial diet is available for Heliconius butterfly.
Tissue Preparation
Midguts of 60-75 4th instar larvae were dissected in physiological saline. The
peritrophic membrane was carefully removed without disturbing gut contents and the
interior gut lumen and exterior were thoroughly rinsed with a stream of physiological
saline. The guts were then homogenized in a glass grinder in 500 pi ice cold buffer
consisting of 50mM K2HPO4 and 250 mM sucrose, pH 7.5. The crude homogenate was
centrifuged in 1 ml phosphate sucrose buffer at 7,500g for 15 minutes and the recovered
supernatant was then centrifuged at 32,000g for 90 minutes in an ultracentrifuge
(Beckman L8-70M, SW 60 rotor) cooled to 4C. The supernatant that contains the
cytosolic fraction was separated for further analysis. The microsomal fraction (the pellet)
was re-suspended in 200 pi of the phosphate sucrose buffer. Both cytosolic and
microsomal fractions were then stored at 4C until purification.
Protein Purification
The protein sample was loaded in a high performance Q-Sepharose column
(Amersham Bioscience Corp. Piscataway, New Jersey) equilibrated with 2.5 mM
imidazole, 10% glycerol buffer, pH 7.5 and an increasing salt gradient of 100-1500 mM
NaCl. The flow rate was 30 drops/min and 0.5 ml fractions were collected. Fractions 31-


60
(Spectra/Por Regenerated Cellulose) against pH 6.8 phosphate buffer every 12 h. The
buffer was changed 5 to 6 times until the HCN had been removed, as determined by the
Feigl-Anger test strip. In order to dehydrate and reduce salt content, the final product
was concentrated to 50 ml with Aquacide I (Calbiochem) and the hydrolytic activity was
tested with the appropriate standard cyanogens and the Feigl-Anger test.
Because P-glucosidases are specific to cyanogenic glycoside types, the following P-
glucosidases were prepared as explained above and used for the detection and general
identification of the cyanogens in extracts, fractions of the first purification method and
larval pellets: p-glucosidase from P. coricea was used for identified Type I glycosides,
p-glucosidase from Passiflora biflora to test Type II glycosides, and linamarase from
linseed for identified Type III glycosides. Two commercially available P-glucosidases
were also used: sulfatase and emulsin from almonds (Sigma Chemical Co., St. Louis,
Missouri) to test sulfated cyanogens and Type IV glycosides, respectively. However,
emulsin is not a very specific enzyme as it can react with other cyanogenic glycoside
types (Brimer et al. 1983). For the samples purified by TLC and HPLC, the enzyme P-
glucuronidase Type H-l from Helix pomatia (Sigma Chemical Co., St. Louis, Missouri)
was used because this enzyme is not specific and can detect all cyanogenic glycoside
types.
Purification of Cyanogenic Glycoside from Plants and Butterflies
Purification by liquid chromatography
The thick syrup obtained from the extraction procedure was applied on top of a 3:1
Whatman CF11 and microcrystalline cellulose column (25 cm x 3 cm) packed with
isopropanol-butanol-water (6:3:1). The column was eluted with the same solvent
mixture. The column was connected to a peristaltic pump and a fraction collector.


73
ecologists recognize the importance that pollen feeding has for Heliconius butterflies
(Gilbert 1991). Pollen feeding is correlated not only with unpalatability (Cardoso,
umpublished), but also with nutrition for adult somatic maintenance and extended
reproductive longevity (Dunlap-Pianka et al. 1977, Dunlap-Pianka 1979), and nuptial
gifts (Boggs and Gilbert 1979). Thus, in the Heliconius system, fourteen days are spent
searching for Psiguria flowers after emergence from the puparium. Adult butterflies
keep synthesizing the aliphatic glycosides from the amino acids gained during the larval
stage, until they are able to collect new amino acids via pollen feeding. Therefore, the
ability of teneral adults to invest in reproduction and defense is constrained by the
resources that the individual butterfly accumulates as a larva.
Phenylalanine was found in high percentages in adult H. erato fed P. biflora or P.
trifasciata as larvae and in H. himera fed P punctata and P. rubra. Tyrosine, on the
other hand, was found in males of H. e. cyrbia fed P. auriculata as larvae. Tyrosine and
phenylalanine are aromatic amino acids which are important for insect development and
reproduction. Phenylalanine is an essential amino acid for all insects, whereas tyrosine is
synthesized from phenylalanine by hydroxylation (Gilmour 1965). Tyrosine plays an
important role in morphogenetic processes and protein synthesis and, with phenylalanine,
it is involved in the hardening and tanning process of newly formed cuticle (Zografou et
al. 2001). Furthermore, tyrosine and tryptophan are precursors of the pigments
ommochrome and melanin, respectively (Nijhout 1991).
In plants, there is evidence suggesting that cyanogenic compounds may also
function in the metabolism and transport of nitrogen to form proteins (Selmar et al. 1988,
Nahrstedt 1992). For example, cyanide is not released from the seeds of the rubber tree,


33
For this chapter, the concentrations of cyanogenic glycosides in various body parts
of races of Heliconius erato fed natural host plants were examined. The same analysis
was performed on separate parts of Passiflora plants (young leaves, mature leaves and
stems). Roots, fruits and flowers were not included in the analysis because Heliconius
larvae rarely feed upon these tissues. Cyanide quantification analysis of the plants was
performed over two seasons (summer and fall). Finally, cyanide concentrations in the
plant were correlated with the concentrations in H. erato which fed upon them. The
discussion addresses the costs and benefits of allocating cyanogens as a defensive
mechanism in the different body parts of this species and the implications of this
variation on the palatability of this butterfly and on its mimetic relationships.
Materials and Methods
Plant Growth and Insect Rearing
This study was conducted at the University of Florida in Gainesville with butterfly
colonies that were raised in Lord & Burnham glass houses (5x8 m). During the first two
years of experimentation the colonies were unable to survive the winter, but improvement
in the glass houses environmental conditions made breeding several generations possible
in the last years of research. H. e.favorinus was collected in Tingo Maria, Huanuco,
Peru, while H. e. cyrbia was collected in Pichincha, Ecuador, and H. e. demophon was
collected in Soberania National Park, Panama.
Most of the Passiflora plants began as cuttings brought from the collection of Dr.
L. E. Gilbert at the University of Texas at Austin; a few cuttings were also donated by R.
Boender, Butterfly World, Fort Lauderdale. P. trifasciata, the natural host plant of H. e.
favorinus, was collected in Tingo Maria by the author of this research. All vouchers of
Passiflora plants used for this study are deposited at the Missouri Botanical Garden, St.


58
Materials and Methods
Butterflies and Plants
This study was conducted at the University of Florida at Gainesville with butterfly
colonies that were maintained in Lord & Burnham glass houses (5x8 m). Each
geographical race was confined in separate cages to avoid hybridization. Wild butterflies
were collected and brought from the field, and stock was replenished with new live
specimens every six to twelve months. During the first two years of experimentation, the
colonies were unable to survive the winters, but improvement in greenhouse
environmental settings made winter breeding possible in the last several years of
research.
Most of the butterfly stocks were raised and propagated in environmental chambers
under controlled temperatures (27C), humidity (78%) and photoperiods (14L:10D).
Adult butterfly colonies were provided with natural host species Passiflora plants
for oviposition, potted Pentas lanceolata and Lantana camara flowers, sugar water and
Psiguria flowers. When Psiguria flowers were not available, a supplemental amino acid
solution was sprayed on Tuffy plastic scrubbing pads, whose orange and yellow color
attracted the butterflies.
Two races of Heliconius erato were used in this experiment: H. e.favorinus from
Peru and PI. e. cyrbia from Ecuador. H. himera, a recently named species which has been
considered a race of erato, was also included in this study (Figure 4-2). Five Passiflora
plants were used in the analysis: P. punctata, P. rubra, P. auriculata, P. trifasciata and P.
biflora.


96
33 had the highest activity (refer to the next section). The fractions were stored
separately at -20C. Aliquots of the eluates from the Q-Sepharose column (fractions 31-
33) were then applied to a column of Mono Q HR 5/5 (Amersham Bioscience Corp.
Piscataway, New Jersey) each fraction separately. The Mono Q column contained a
mobile phase A with the detergent buffer described above and phase B contained 1M
NaCl plus the detergent buffer. The flow rate was 0.5 ml/min, and 0.5 ml fractions were
collected. Fractions 23-25 had the highest activity. The fractions were stored separately
at -20C. In all cases the protein elutes at approximately 400 mM NaCl.
Hydrophobic interaction chromatography with phenyl-sepharose was also
attempted; however, the proteins were not retained in the column.
Enzymatic activity of p-glucosidase
The cytosolic and the microsomal fractions were tested for activity. The synthetic
substrate used to estimate P-glucosidase was 4-nitrophenyl P-D-glucopyranoside in 1 ml
of 0.1 M NaOH citrate buffer (pH 6.0). Upon action of a P-glucosidase, this substrate is
hydrolyzed and forms p-nitrophenol, which turns yellow at alkaline pHs.
A 250-pl quantity of the cytosolic fraction was used to test activity, while 50 pi
was used from the microsomal fraction. The fractions were then incubated at 30C for 2
h. The reaction was halted by immersing the incubation tubes in boiling water for 10
min. For the control, an identical mixture was boiled for 10 min. before the incubation.
All tubes were centrifuged at 10,000 g for 10 min. after incubation. The concentration of
the reaction product p-nitrophenol was determined in a spectrophotometer (Spectronic
Genesys 5, Spectronic Instruments, Rochester, New York) at 400 nm. The molar
extinction coefficient used was 18.13 mM'1. (One unit is defined as the amount of
enzyme hydrolyzing 1 pmol of substrate per min at 30C.)


47
p-glucosidase
Cyanogenic glycoside
Toxic
a-hydroxynitrile lyase
T
+ HCN
aldehyde/ketone
Toxic Toxic
a-hydroxynitrile
Unstable
Figure 3-1. General process of enzymatic hydrolysis of cyanogenic glycosides.


69
compounds found in butterfly samples that were related to host plant feeding. When
butterflies were fed plants containing monoglycoside cyclopentenoids, the percentage of
lotaustralin was very low. Also, H. e. cyrbia and H. e. favorinus have a similar percent
ratio of linamarimlotaustralin, while H. himera generally have a higher ratio of linamarin
to lotaustralin.
Sequestration of cyanogenic glycosides
Butterflies developing from larvae fed P. auriculata sequestered monoglycoside
cyclopentenoids. !H NMR analysis showed that the female H. e. cyrbia contained
epivolkenin (Table 4-9), linamarin, and another monoglycoside (perhaps taraktophyllin,
based on HI and H4). The male sample also contained epivolkenin, linamarin and
another glycoside (perhaps taraktophyllin, based on HI and H4). Sarauriculatin was not
found in the samples. The results of the HPLC/MS (Table 4-7 and 4-8) showed the
percent values of each compound. However, taraktophyllin was not distinguishable from
epivolkenin. under the conditions in which this analysis was conducted.
Small percentages of the cyanogen epivolkenin, passicapsin and MW 433
(passibiflorin or passitrifasciatin) were found when butterflies were fed diglycoside
cyclopentenoids, demonstrating both sequestration and metabolism of this type of
cyanogenic compound (Table 4-7 and 4-8).
The entire butterfly extracts for the cyanogenic and amino acids analysis had a
blank in between each sample in order to avoid carry-over.
Amino Acids Analysis in Butterfly Extracts
The amino acids valine, isoleucine/leucine, tyrosine, phenylalanine, and
tryptophan, were measured in butterfly samples (Table 4-10 and 4-11).


Table 4-9. 'H NMRU> spectrum in CD3OD at 27C of a cyanogen, identified as epivolkenin, from female H. e. cyrbia fed on P.
auriculata. Chemical shifts (ppm), coupling constants (Hz) and multiplicities. Multiplicity is identified by: (d) doublet,
(m) multiplet.
Cyanogen
H2
H3
H4
H5A
H5B
HI'
H2'
H3'-H5'
H6A'
H6B'
Multiplicity
dd
dd
m
dd
dd
d
dd
m,m,m
dd
dd
Epivolkenin(b)
8 ppm
6.14
6.27
4.80
2.25
3.02
4.62
3.22
3.3-3.4
3.68
3.86
J values
5.5,1.3
5.5,2.0
14.6,4.8
14.6,7.1
7.7
9.1,7.8
c
12,5.5
12,2
H. e. cyrbia
5 ppm
6.13
(6.14)
6.26
(6.26)
. w
(4.81)
2.25
(2.25)
3.01 (3.02)
4.62
(4.62)
3.22
3.3-3.4
(3.85)
J values
5.5,1.3
5.5,2.0
14.6,4.8
14.6,7.1
7.7
9.1,7.8
c
-
12,2
600.13 MHz; see Figure 4-1 for a numbered structure.
(b) From Jaroszewski et al. 1987a, b.
chemical shift at 17 C relative to the proton impurity in methanol-D4 (CD3OD) i.e. 'HCD20D (H) assigned 3.3 ppm.


14
egg on the third day following oviposition. All first instars in both races ate their egg
shells.
Larvae
Both subspecies underwent five instars of development. No supernumerary instars
were detected during the experiment.
Three different analyses were carried out with head capsule measurements. The
first was the comparison of head capsule size and larval developmental time between and
within geographical races feeding on natural vs. alternative host plants.
Repeated measurements ANOVA for head capsule size showed a significant
variation in size between the two races (Table 2-2). Analysis within H. e.favorinus
showed that host plants directly affect head capsule size (Table 2-2). Tukey comparison
test indicates that individuals fed on P. auriculata tend to be bigger than individuals fed
P. trifasciata and P. rubra during the second, third and fourth instars of larval
development. No significant differences in head capsule size were found among
members of H. e. cyrbia. There was no size variation among full-sib within races.
Differences in larval developmental time between and within races were
statistically significant between family and between individuals fed different host plants
(Table 2-3). On average, H. e. favorinus larvae fed longer on P. auriculata and P. rubra
(fable 2-4). Both H. e. favorinus, and H. e. cyrbia also showed significant differences
between larval developmental time between full-sibs, indicating genetic variation within
them and between individuals fed different hosts (Table 2-3). Longer larval
developmental time was experienced by both male and female cyrbia fed on rubra, as
well as females fed auriculata and males fed trifasciata (Table 2-4).


Head capsule height (mm) Head capsule height (mm)
26
Larval developmental time (days)
Larval developmental time (days)
Figure 2-4. Growth trajectories of (A) H. e. favorimis and (B) H. e. cyrbia feeding on five
different host plants. Each point represents the mean value of larvae from the
first to fifth instar. Bars on each point indicate standard error.


COMPARATIVE PROCESSING OF CYANOGENIC GLYCOSIDES AND A NOVEL
CYANIDE INHIBITORY ENZYME IN HELICONIUS BUTTERFLIES
(LEPIDOPTERA: NYMPHALIDAE: HELICONIINAE)
By
MIRIAN MEDINA HAY-ROE
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
2004

Copyright 2004
by
Mirian Medina Hay-Roe

To all the people that love science and helped me through the development of this
research.

ACKNOWLEDGMENTS
I would like to thank all of the members of my committee, especially my chairman.
Dr. T. C. Emmel, for financial support for a field trip to Per and Panama, for providing
laboratory' space and two greenhouses where I kept plants and butterfly colonies,and for
his friendship and trust. Dr. J. L. Nation was kind enough to guide me throughout many
of the experiments performed for this study. I would also like to thank him for his
laboratory space, which became my "home' for many years. Thanks go to Dr. H. J.
MacAuslane, for allowing me to use an entire compartment of her ultrafreezer. I am very
grateful for her friendship and her support. Thanks go to Dr. D. A. Jones and Dr. F.
Slansky for providing me with enzymes, and some cyanogens used as standards. Dr.
Jones kindly allowed me to use his extensive library containing the most important
literature on passionvines.
I extend my special thanks to several people who were not members of my
committee, but who provided equipment for the chemical analyses and whose comments
were crucial in completing my dissertation. My deepest gratitude is expressed to Dr. J.
Tumlinson and Dr. P. Teal from the Center for Medical, Agricultural and Veterinary
Entomology (CMAVE), U. S. Department of Agriculture (USDA), without whose help
my dissertation would not have been possible. I thank them for their trust and support.
Special thanks as well are due to Dr. C. Lait, with whom I spent a summer through a
USDA Student Intership Program during which I had the opportunity to practice various
IV

> l ¡t
techniques for protein purification. I thank Dr. S. Yu and Dr. D. Boucias for discussing
enzymology work with me and for the use of their laboratory equipment.
Special thanks to the National Science Foundation (NSF) and the National High
Magnetic Field Laboratory External User Program via Dr. P. Teal, CMAVE, USDA for
the use of the Nuclear Magnetic Resonance (NMR) at the Advanced Magnetic Resonance
imaging and Spectroscopy (AMRIS) facility in the McKnight Brain Institute of the
University of Florida. I am very grateful to Mr. Jim Rocca from for his help in the
chemical elucidation of my compounds through (NMR) analysis. I thank Dr. J. Johnson,
who runs the HPLC/Mass Spectroscopy unit in the Chemistry Department. His reports
on the analysis performed are the best! I also appreciate his friendship and long
discussions on the HPLC/MS analysis. Dr. I. Ghiviriva kindly ran the NMR analysis to
corroborate the chemical composition of two plants: P. auriculata and P coricea. Dr.
H. Albom from the USDA generously ran certain samples in the mass spectrometry and
also gave me advice on some of the techniques I used. Special thanks are due to Dr. B.
Torto, for his friendship.
I owe special thanks to Dr. J. Jarozewsky from the Royal Danish School of
Pharmacy in Copenhagen, Denmark, for sending me several milligrams of purified
gynocardin and tetraphyllin B.
Thanks are due to Dr. R. Ferl, Dr. P. Sehnke, and B. Laughter from the
Horticultural Science Department, who allowed me to use the lab ultracentrifuge and
electrophoresis equipment. Beth taught me all of the tricks for getting the best
electrophoretic gel. Working with her was very enjoyable.
v

My former M. A. supervisor and professor, Dr. L. Gilbert from the University of
Texas at Austin, allowed me to take all the cuttings I needed from his great collection of
identified Passiflora. My deepest gratitude is also given to Mr. R. Boender from
Butterfly World for providing me with some Passiflora cuttings, adult plants, and
butterflies, and for financial support at key times. I also thank Dr. J. McDougall for the
identification of the Passiflora plants.
I extend thanks to Jacob Olander in Quito, Ecuador, for providing me with H. e.
cyrbia and H. himera.
The Institute of Natural Resources (INRENA) in Lima, Peru, provided me with
collection and export permits for two seasons of work in Tingo Maria, Huanuco, where I
collected Heliconius erato favorinus. Thanks are also due to the Smithsonian Tropical
Research Institute in Panama for providing me with permits to collect H. e. demophon.
I am very grateful for the financial support provided by the Delores A. Auzenne
Scholarship, the minority programs at the University of Florida, the Mulrennan
Scholarship, and for the various semesters on teaching assistantship, which I received in
the Entomology and Nematology Department.
I also extend heartfelt thanks to Christine Eliazar, who was always there to listen
and give me comfort during stressful times, and to Arif, Steve, David, Alfredo, Matthew
and Chris for helping me with the greenhouse maintenance chores. Special thanks are
due to Alfredo Rios for scientific discussions on butterfly issues, and for raising
caterpillars to maintain the greenhouse colonies. Steve and Jim Schlachta and Jerry
Wenzel were always there when I had electrical problems in the greenhouses and
environmental chambers and fixed them quickly to help prevent any harm to my cultures.
vi

And finally, my thanks go to my husband, Keith, for his love and willingness in
helping me with some of this research and especially for being the outstanding editor that
he is, and for being the father of my daughter, Kylie, who shares my love for butterflies
and enjoys being my best helper.
vii

TABLE OF CONTENTS
gage
ACKNOWLEDGMENTS iv
LIST OF TABLES xi
LIST OF FIGURES xiii
ABSTRACT xv
CHAPTER
1 INTRODUCTION 1
2 LIFE HISTORIES AND NUTRITIONAL EFFECTS ON SIZE OF HELICONIUS
ERATO FEEDING ON DIFFERENT PASSIFLORA HOST PLANTS 6
Materials and Methods 8
Museum Specimens 8
Living Insects and Host Plants 9
Feeding Experiments 9
Life History Studies 10
Egg measurements 11
Larval development 11
Pupae and adult offspring measurement 12
Statistical Analysis 12
Results 13
Analysis of Specimens from the Museum Collection 13
Life History Studies in H. e. favor mus and H. e. cyrbia 13
Eggs 13
Larvae 14
Pupae 15
Adult 15
Mortality 16
Discussion and Conclusions 16
3 C YANOGENESIS IN HELICONIUS ERA TO AND PASSIFLORA HOST
PLANTS 29
Occurrence of Cyanogenesis in Plants 29
Occurrence of Cyanogenesis in Lepidoptera 30
viii

Cyanogenic Glycosides as a Defensive Mechanism 31
Materials and Methods 33
Plant Growth and Insect Rearing 33
Experimental Design and Treatments 34
Butterflies 34
Temporal cyanide quantification of Pass ¡flora plants 34
Quantitative Determination of Cyanide 35
Colorimetric analysis 35
Standard curve 36
Predator-Prey Interaction 37
Statistical Analysis 37
Results 37
Temporal Variation of Cyanide Release in Passijlora Plants 37
Cyanide Quantification in Heliconius erato Fed Their Natural Host Plants 38
H. e. cyrbia fed on P. rubra and P. punctata 38
H. e.favorinus fed P. trifasciata 39
H. e. demophon fed on natural and alternative hosts plants 39
Cyanide Correlation Between Heliconius erato Fed Natural Host Plants 40
Predator-Prey Interaction 40
Discussion and Conclusions 40
Cyanide Concentrations in Heliconius erato fed Pass ¡flora host plants 40
Temporal Variation in Cyanide in Passiflora Plants 43
Implications of Cyanide Concentration for Palatability 44
Implications of Cyanide Concentration for Mimicry 44
4 CYANOGENIC GLYCOSIDES IN PASSIFLORA PLANTS AND
BUTTERFLIES 52
Mechanisms Used by Insects to Process Toxic Compounds from Plants 53
Cyanogenic Glycosides in the Family Passifloraceae 54
Cyanogenic Glycosides in Butterflies 55
Synthesis De Novo 55
Sequestration of Cyanogenic Glycosides 56
Enzymes 57
Materials and Methods 58
Butterflies and Plants 58
Plant and Butterfly Cyanogen Extraction 59
P-Glucosidase Preparations 59
Purification of Cyanogenic Glycoside from Plants and Butterflies 60
Purification by liquid chromatography 60
Purification by thin layer chromatography 61
Purification by high performance liquid chromatography (HPLC) 62
Identification of Cyanogenic Glycosides from Passijlora Plants and
Butterflies 62
Nuclear magnetic resonance (NMR) 62
HPLC/ (+) ESI-MS experimental 64
Amino Acid Analysis 65
IX

Results 66
Cyanogenic Glycosides in Passiflora Plants 66
Identification of cyclopentanoid cyanogenic in Passiflora rubra 66
Identification of cyclopententanoid cyanogenic in Passiflora trifasciata ...67
Cyanogen profile of Passiflora auriculata 67
Cyanogen profile of Passiflora biflora 67
Cyanogen profile of Passiflora punctata 68
Cyanogenic Glycosides of Butterflies Fed Natural Host Plants 68
De novo synthesis 68
Sequestration of cyanogenic glycosides 69
Amino Acids Analysis in Butterfly Extracts 69
Thiol Analysis 70
Discussion and Conclusions 70
Sequestration. De Novo Synthesis and Metabolism of Cyanogenic
Glycosides 70
Amino Acids and their Importance For Heliconius Biology 72
5 STUDIES OF CYANIDE INHIBITION BY p-GLUCOSIDASE FROM
HELICONIUS LARVAL GUTS 92
Materials and Methods 94
Study Organism 94
Insect Rearing Procedure 95
Tissue Preparation 95
Protein Purification 95
Enzymatic activity of P-glucosidase 96
Protein concentration measurements 97
In vitro Cyanide Quantification of P-Glycosidase from H. himera Larval
Midgut and Various Cyanogenic Glycosides as Substrates 97
Sodium Dodecyl Sulfate (Sds)-Polyacrylamide Gel Electrophoresis (PAGE)...97
Protein Identification 98
Results 98
In vitro Cyanide Quantification of P-Glycosidase from H. himera Larval
Midgut and Various Cyanogenic Glycosides as Substrates 98
P-glucosidase Activity in the Midgut of H. himera 98
Protein Purification and Protein Identification 99
Discussion and Conclusions 99
6 CONCLUSIONS 106
LIST OF REFERENCES 108
BIOGRAPHICAL SKETCH 122
x

LIST OF TABLES
Table E^ge
2-1. Average wing length and standard error of geographical races of Heliconius eruto
from museum specimens in different regions 20
2-2. Repeated measurements ANOVA of head capsule size with race, host and family
as factors 20
2-3. Repeated measurements ANOVA of larval developmental time with race and host
as factors 21
2-4. Sexual differences in larval developmental time (LDT), growth rate (GR) and
wing length (WL) in H. e. favorinus and H. e. cyrbia feeding on five different
host plants 21
2-5. Mortality in different stages of the life history of H. e. favorinus and H. e. cyrbia
feeding on natural and suitable host plants 22
3-1. Heliconius eruto fed natural and alternative Passifloru host plant and sample size
used in the study 46
3-2. Passifloru plants tested for cyanogenesis in this study. All plants are used by
Heliconius eruto in different geographical areas 46
4-1. Qualitative analysis of the extracts of Passifloru plants. Number 0=no cyanide
release. l=light blue, slightly cyanogenic, 2=blue, cyanogen and 3=dark blue,
very cyanogenic 75
4-2. id NMR(a) spectrum of a cyanogen, identified as passicapsin, from P. rubra in
CD3OD. Chemical shifts (ppm) and coupling constants (Hz). Degree of splitting
is identified by: (s) single, (d) doublet, (q) quartet, (m) multiplet 76
4-3. 1 C NMR,a> spectrum of a cyanogen, identified as passicapsin, from P. rubra in
CD3OD 77
4-4. 'H NMR(a) spectrum of a cyanogen, identified as passitrifasciatin, from
P trifasciata in CD3OD. Chemical shifts (ppm) and coupling constants (Hz).
Degree of splitting is identified by: (d) doublet, (t) triplet, (m) multiplet 78
xi

4-5. 13C NMR(a) spectrum of a cyanogen, identified as passitrifasciatin from P.
trifasciata in CD3OD
79
4-6. Qualitative analysis of cyanide release of whole body extracts tested with different
P-glucosidases by Feigl-Anger strips monitored over time in H. e.favorinus and
H. e. cyrbia. Numbers indicate, 0=no cyanide release, l=light blue, slightly
cyanogenic, 2=blue, cyanogen and 3=dark blue, very cyanogenic 80
4-7. Peak areas counts of cyanogenic glycosides detected by HPLC/MS in H. erato and
H. himera butterflies 81
4-8. Percent of cyanogenic glycosides detected by HPLC/MS in H. erato and H. himera
butterflies 81
4-9. 'H NMR(a) spectra of cyanogen, epivolkenin, in an extract from female H. e. cyrbia
fed on P. auriculata in CD3OD at 27 C. Chemical shifts (ppm), coupling
constants (Hz) and multiplicities. Multiplicity is identified by: (d) doublet, (m)
multiplet 82
4-10. Peak areas of measured free amino acids in butterfly samples 83
4-11. Percent total peak areas of measured free amino acids in butterfly samples 84
5-1. Experimental treatments performed with P-glycosidase 102
5-2. Cyanide concentration when larval P-glycosidase from Heliconius himera were
tested against different cyanogenic substrates 102
xii

LIST OF FIGURES
Figure page
2-1. Life history of Heliconius butterflies 23
2-2. Museum specimens of H. erato from the eastern (left column) and western sides
(right column) of the Andes that were included in the analysis. (A) H. e. phyllis,
(B) H. e. favor inns, (C) H. e. magnificus, (D) H. e. colombina, (E) hydara,
(F) H. e. cyrbia, (G) H. e. petiveranus 24
2-3. Study organisms. (A) H. e. favor inns from Tingo Maria. Huanuco, Per and
(B) H. e. cyrbia from Pichincha, Ecuador 25
2-4. Growth trajectories of (A) //. e.favorinus and (B) H. e. cyrbia feeding on five
different host plants. Each point represents the mean value of larvae from the
first to fifth instar. Bars on each point indicate standard error 26
2-5. Growth trajectories of H. e.favorinus and H. e. cyrbia fed their natural host plants
(P. trifasciata and P. rubra, respectively). Each point represents the mean value
of larvae from the first instar to the fifth. Bars on each point indicate standard
error 27
2-6. Pupal weight of H. e.favorinus (A) and //. e. cyrbia (B) fed on natural and
alternative host plants. Different letters indicate significant differences between
means (Tukey HSD-test). Bars in each point indicate standard error 28
2-7. Average wing length of H. e. favorinus and H. e. cyrbia fed natural and alternative
host plants. Different letters indicate significant differences between means
(Tukey HSD-test). Bar in each point indicate standard error 28
3-1. General process of enzymatic hydrolysis of cyanogenic glycosides 47
3-2. Cyanide quantification of Passiflora plants in the (A) summer and (B) fall seasons.
Bars indicate standard error 48
3-3. Cyanide quantification of male and female H. e. cyrbia fed P. rubra. Different
letters indicate significant differences between means (Tukey HSD-test). Bars
indicate standard error 49
3-4. Cyanide quantification of male and female H. e. cyrbia fed P. punctata. Different
letters indicate significant differences between means (Tukey HSD-test). Bars
indicate standard error 49
3-5. Cyanide quantification of male and female H. e.favorinus fed P. trifasciata.
Different letters indicate significant differences between means (Tukey HSD-test).
Bars indicate standard error 50

3-6. Cyanide quantification of male and female H. e. clemophon ted natural (P. biflora)
and alternative host plants. Bars indicate standard error 50
3-7. Regression of the mean cyanide concentration of H. erato over the mean cyanide
concentration of the natural Passiflora host plants upon which they fed. (The
means were used in the figure to enhance the clarity of presentation.) 51
4-1. Cyanogenic glycosides in Passifloraceae plants (A) Type I. (B) Type II.
(C) Type III and (D) Type IV (Jaroszewski et al. 2002) 85
4-2. Heliconius used in this study (A) H. e. cyrbia, (B) H. e.favorinus, and (C) H.
himera 86
4-3. Passiflora plants used in this study (A) P. auriculata, (B) P. punctata, (C) P. rubra.
(D) P. biflora, (E) P. trifasciata 87
4-4. Proton/Carbon heteronuclear multiple quantum coherence (HMQC) spectrum of
cyanogen passicapsin in CD3OD from P. rubra 88
4-5. Proton/Carbon heteronuclear multiple quantum coherence (HMQC) spectrum of
cyanogen passitrifasciatin in CD3OD from Passiflora sp 89
4-6. Potential enzymatic hydrolysis products of passibiflorin 90
4-7. Potential enzymatic hydrolysis products of passicapsin 91
5-1. Heliconius himera from Vilcabamba, Loja Province, Ecuador 103
5-2. Electrophoresis in SDS-10% gel slab. The gel shows different levels of purification
of the protein studied. CE: is the cytosolic crude extract, Q sepharose elutes:
Q1-Q3 fractions 31-33, Mono-Q elutes: Ml-M2, fractions 23 and 24 from Q-
sepharose fraction 31, and M3-M4, fractions 24 and 25 from Q-sepharose
fraction 32. RM is the rainbow marker 104
5-3. Comparison of amino acid sequence of Spodoptera frugiperda (Lepidoptera:
Noctuidae) p-glucosidase with mass spectroscopy identified peptides from H.
himera. The letters in bold indicate the homologous amino acids sequencies
from H. himera that match with S. frugiperda. A. Mr 55.000 and B. Mr 65,000. 105
xiv

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
COMPARATIVE PROCESSING OF CYANOGENIC GLYCOSIDES AND A NOVEL
CYANIDE INHIBITORY ENZYME IN HELICONIUS BUTTERFLIES
(LEPIDOPTERA: NYMPHALIDAE: HELICONIINAE)
By
Mirian Medina Hay-Roe
May 2004
Chair: Thomas C. Emmel
Major Department: Entomology and Nematology
Heliconius erato, a neotropical butterfly known to specialize in larval feeding on
cyanide-producing Passiflora plants, has evolved different strategies for dealing with
cyanogenic compounds. By combining ecological and biochemical data, I investigated
some of these mechanisms. Subspecies of Heliconius erato from the eastern and the
western side of the Andes differ in body size. The subspecies from the eastern side of the
Andes proved to be consistently larger than their relatives from the western side of the
Andes. These differences in body size are genetically based. When two subspecies (H.
e. cyrbia and H. e. favorinus) were fed different host plants (natural and alternative host),
plastic growth trajectories resulted which were related to the host plant fed on by the
larvae. Feeding on one of the host plants, P. rubra, resulted in lower growth rates in both
subspecies and higher mortality in H. e. cyrbia, but not in H. e. favorinus. As a result, the
few individuals from the former subspecies that survived while feeding on this highly
toxic cyanogenic plant were better protected against frog predators because they
xv

accumulated higher concentrations of cyanogenic glycosides in their bodies, a condition
which in turn generates higher concentrations of hydrogen cyanide. H. e. favorinus, on
the other hand, turned out to be less cyanogenic and therefore less well-protected against
predation by frogs, demonstrating that chemical defenses in H. erato vary between
geographical races and that lower cyanide concentrations may lead to higher predation
rates.
H erato not only synthesizes de novo aliphatic cyanogenic glycosides, but also
sequesters both simple and complex cyclopentenoid glycosides from their host plants. In
addition, during the larval stage H. erato metabolized the complex cyclopentenoid
glycosides into simple cyclopentenoid glycosides. Analysis of amino acids in the adults
revealed that these butterflies are storing not only cyanogenic compounds for defense, but
also essential amino acids from metabolized cyclopentenoids gathered during the larval
stage.
Finally, a novel cyanide detoxification enzyme was tested in vitro. Two (3-
glucosidases were isolated from larval midgut; they had some homologies with
Spodoptera frugiperda P-glucosidase. The enzymes are specific to simple
cyclopentenoid glycosides and proved to inhibit the production of cyanide when
cyanogenic glycosides, plant P-glucosidase and larval midgut P-glucosidase were tested.
This discovery leads to the conclusion that this detoxification mechanism is related to the
ability of Heliconius to metabolize different Passiflora cyanogen types. However, more
experiments are warranted to describe the characteristics of the enzymes. The
importance of this new mechanism of cyanide detoxification is discussed.
xvi

CHAPTER 1
INTRODUCTION
Heliconius erato is a brightly colored, highly variably patterned butterfly that is
distributed from Central to South America. H. erato is adapted to altitudes that range
from sea level to 1,600 m. It is frequently found flying in open pastures in disturbed
forest and secondary growth (DeVries 1987). This species is unpalatable to predators and
shows strong Mllerian mimicry with other distasteful species. Geographical races have
different patterns of coloration that range from orange and black, to red, yellow and
black, to iridescent blue and pink (this pattern of coloration is shared with the co-mimic
H. melpomene). The races have diversified broadly and there are now approximately 30
parapatric races that are able to mate randomly in narrow hybridization zones (Emsley
1964. Mallet 1986).
In this study, three races of erato were observed: H. e. cyrbia from Pichincha,
Ecuador; H. e. favorinus from Tingo Maria, Huanuco, Peru; and H. e. demophon from
Panama. Heliconius himera (used in experiments described in Chapter 5) constitutes an
intermediate step in the transition from a race to a species (McMillan et al. 1997). It had
previously been considered a race of erato (Lamas 1976, Brown 1979, Sheppard et al.
1985), but recent mtDNA studies suggest a divergence approximately 1 million years ago
(Brower 1994; see also Emsley 1965 for arguments that it is actually a separate species).
Phylogenetic studies of H. erato group the races into two main clades. One clade
groups races that are located on the eastern side of the Andes, while the second clade can
be found on the western side of the Andes (Brower 1994). The ecology (Brown 1979,
1

2
1981, Gilbert 1991, Benson et al. 1976), behavior (Chai 1986, 1990, 1996), and
physiology (Dunlap-Pianka et al. 1977, Boggs 1987) of this species have been studied
extensively. However, very little is known about the chemical ecology of this species.
Heliconius erato is known to specialize on Passiflora plants (Passifloraceae). The
females lay eggs individually on new growth or tendrils. In many areas of South
America, this species has been considered an oligophagous species, as larvae feed on
many different Passiflora species, but only from the subgenus Decaloba (Benson et al.
1976, Menna-Barreto and Araujo 1985. Hay-Roe 1996). In Central America, however,
this species tends to be monophagous, selecting only one or two host plant species at any
one location (Gilbert 1991, Benson et al. 1976). Diet breadth has been associated with
competition for resources within members of different species of Heliconius that are
distributed in each geographical region. This drives the partitioning of the resources
among competing species.
With respect to its adult diet, Heliconius has developed a special adaptation
allowing it to feed on pollen from Psiguria and Gurania flowers (Cucurbitaceae) by
externally regurgitated digestive enzymes (Gilbert 1972). Pollen plays an important role
in reproduction because butterflies assimilate the amino acids released from the pollen for
egg development and long-term survival (Dunlap-Pianka 1979). Adult H. erato also feed
on liquid nectar from Lantana flowers which provide carbohydrates in their nectar as
well as in pollen (Boggs 1981).
Ehrlich and Raven (1964) have determined that secondary plant compounds play a
critical role in determining patterns of plant utilization. Insects appear to follow their
host plants adaptive radiation, dispersal, and elaboration of secondary chemical

3
constituents. Host ranges diverge and insects specialize with respect to their host plants
as they radiate. In the Heliconius-Passiflora system, coevolution is thought to be the
process that has generated this tight insect-plant interaction (Benson et al. 1976, Futuyma
and Keese 1992).
Passiflora plants form a large genus which has been reviewed by many botanists
(e.g., Harms 1924, Killip 1938, De Wilde 1971, 1974). The most recent review, by
McDougal and Feuillet (unpublished), divided the genus into four subgenera and various
super sections comprising a total of 521 species. Characters such as plant chemistry
(Spencer 1984) and ecological data nicely fit the taxonomic data, specifically within each
of the sections of Passiflora.
Passiflora plants have developed a series of morphological characteristics to avoid
predation by herbivores such as Heliconius larvae (Benson et al. 1976). The
development of a great diversity of leaf shapes by the same plant (a phenomenon called
heteroblastia) is notable in the genus. Gilbert (1983) speculated that the difference in leaf
appearance assists the plant in escaping detection by female Heliconiinae butterflies as
they search for appropriate oviposition substrates. Mechanical defense in Passiflora
species may occur in the form of protective hairs on the leaves, especially in the section
Pseudodysosmia. The hooks of Passiflora adenopoda penetrate the cuticle of Heliconius
larvae, causing immobilization and death from desiccation (Gilbert 1971). Passiflora
also develop deciduous filiform stipules which resemble tendrils, and which may function
to stimulate egg placement by Heliconius, later causing the eggs to fall off with the
deciduous stipules. Various Passiflora produce novel structures on the stipules, leaves or
new shoots which resemble the eggs of Heliconius butterflies in shape and color, and

4
cause female butterflies to avoid further oviposition on such plants (Williams and
Gilbert 1981). For chemical protection against herbivores, several strategies are used by
the Passiflora. Most of the Passiflora produce sugary secretions from extrafloral
nectaries on petioles, leaf margins, surface, tips, bracts and stipules, which attract ants
that actively defend the plants. Ants then eat heliconian eggs, and attack and carry off
small larvae (Lanza 1988). But most importantly, Passiflora plants are protected from
most herbivores by their production of toxic cyanide-releasing secondary compounds
called cyanogenic glycosides (Spencer 1988). The Passifloraceae are among the few
plants producing cyanogenic glycosides with a cyclopentene moiety (Spencer 1988).
1his characteristic is shared only by members of five other closely related families:
Flacortiaceae, Malesherbiaceae, Tumeraceae, Achariaceae and Caricaceae.
Few Heliconius species have been investigated with regard to their chemical
interaction with Passiflora cyanogens. It is known that the larvae, pupae, and adults of
Heliconius and other related genera biosynthesize de novo two simple aliphatic
cyanogens, linamarin and lotaustralin (Nahrstedt and Davis 1983). Also, it has recently
been demonstrated that Heliconius sara, a specialist on P. auriculata, sequesters
cyanogens (epivolkenin) from its host plant and metabolizes most of this compound into
the corresponding thiol derivative (sarauriculatin) by replacing the nitrile group which
prevents cyanide release (Engler et al. 2000). Adult butterflies of the species Acraea
horta (Nymphalidae: Acraeinae) store the cyclopentyl cyanogen, gynocardin, when they
develop as larvae on Passiflora plants (Raubenheimer 1989).
In order to understand the phenomena of synthesis and sequestration, of chemical
compounds in Lepidoptera, it is important to understand the life histories of these species

5
(Nishida 2002). Vital clues to the biochemistry of defensive and other chemicals may be
revealed by ecological information gathered in the field or in live captive cultures.
In this dissertation I will combine ecological and biochemical data in order to
understand the patterns of host utilization and detoxification in geographical races of
Heiiconius erato. In Chapter 2,1 compare the life histories of two races of Heliconius
erato (one from the eastern and one from the western side of the Andes) that differ in
body size and describe the effects of natural host plants vs. alternate Pass ¡flora plants on
their growth, nutrition and mortality. In Chapter 3,1 analyze cyanide concentrations in
butterflies and host plants, in order to understand whether there are any differences in
defense mechanisms between races. In Chapter 4,1 describe the isolation and
identification of cyanogenic glycosides in both butterflies and plants in order to learn
how the butterflies process these toxic compounds. In Chapter 5,1 present preliminary
results on the analysis of P-glucosidase from larval guts in order to test the hypothesis
that this enzyme will inhibit cyanide release by Passiflora plants.

CHAPTER 2
LIFE HISTORIES AND NUTRITIONAL EFFECTS ON SIZE OF HELICONIUS
ERATO FEEDING ON DIFFERENT PASSIFLORA HOST PLANTS
In the last two decades, many studies of life history evolution and evolution of
body size have been published. These studies present a retrospective analysis of patterns
of variation in size and age at maturity in an attempt to understand the mechanisms
responsible for generating phenotypic variation in maturation. However, there are
important aspects that these studies did not take into account, such as invariant age and/or
size thresholds that occur during instar transitions in larval development (Nijhout 1974,
1975). To study the life history of a population and/or species, one must, first of all,
study the egg because it constitutes potential maternal (Dunlap-Pianka et al. 1977,
Sinervo et al. 1992) and paternal (Boggs and Gilbert 1979, Wiklund and Kaitala 1993)
contributions to the offsprings fitness. The second stage, juvenile development, is
probably the most important stage in the life history of an arthropod. At this stage, three
different characters may be responsible for the age and size at maturity: the number of
instars, the larval developmental time or molting interval, and the change in size at
ecdysis. These characters can be plastic or canalized (Higgins and Rankin 1996) and
might differ, depending on the different post-embryonic pathways of the arthropod. In
the end, the variation among these stages may affect the final adult size in the species or
population (Figure 2-1).
Recent models are focusing not only on the analysis of growth but also on age, and
other complex and combined characters, which reflect more accurately how adaptive
6

7
maturation phenotypes are achieved by different organisms (Bernardo 1993, Higgins and
Rankin, 1996, Kirkpatrick and Lofsrold 1992, Nylin 1992, Nylin et al. 1989, Steam and
Koella 1986). Thus, age and size at maturity are the results of a growth and
developmental history which reflect both genetic and environmental factors. These
factors act throughout the juvenile developmental time. The relative contribution of
genetic and environmental effects to age and size are likely to be different for individuals
within and between populations and species.
Plastic juvenile development may play an important role in the interaction between
the life cycle and the habitat. The phenotypic value in this context is best explained using
the quantitative genetics formula from Falconer (1989):
P = G + E + M
where P is the phenotypic value; G is the genotypic value; E is the environmental
deviation; and M is the contribution from maternal effects (non-genetic effects).
The latest phylogeny of H. erato (Brower 1994) grouped different subspecies of H.
erato (geographical races) into two main clades. One of the clades grouped races that are
geographically distributed on the eastern side of the Andes (the eastern clade), while the
second grouped races distributed on the western side of the Andes (the western clade).
In a previous study (Hay-Roe 1996), differences in body size between two
subspecies of H. erato were investigated. One subspecies from the western side of the
Andes (H. e. petiveranus) is very specific in its host plant feeding, while a second
subspecies (H. e. phyllis) from the eastern side of the Andes is a broad specialist on
various subgenera of the Passiflora. Differences in body size were genetically based, H.
e. phyllis is bigger than H. e. petiveranus. Differences in growth trajectories between the

8
races were plastic in H. e. petiveranus (due to variation in larval developmental time), but
canalized in H. e. phyllis. The results of this study led me to hypothesize that there may
be a difference in patterns of growth within subspecies from the western vs. the eastern
side of the Andes. 1 also wondered whether the members of the subspecies located on the
eastern side of the Andes were always bigger in size than the members of the western
side of the Andes.
In this chapter of the dissertation, I have first examined museum specimens of the
most representative members of each region to see whether there is a common variation
in size between different races of erato from the eastern and the western sides of the
Andes. Second, I have studied the life histories of two geographical races, H. e.
favorinus from the eastern side of the Andes and H. e. cyrbia from the western side of the
Andes. For each race, phenotypic effects were analyzed based on environmental,
genotypic, and maternal effects. Growth trajectories, final size and mortality were
compared between groups fed natural host plant vs. alternative host plants.
Materials and Methods
Museum Specimens
The Lepidoptera collection of the Allyn Museum of Lepidoptera, Sarasota, Florida,
was used to record the wing lengths of different subspecies of H. erato (Figure 2-2).
Wing length was measured from the base of the wing to the apex. A Vernier caliper (Spi
2000) graduated to 0.1 mm was used to make the measurements.
Four subspecies, H. e. colombina, H. e. hydara, H. e. cyrbia, and H. e. petiveranus,
are representatives from the western side of the Andes. These subspecies were compared
with H. e. phyllis, H. e. favorinus, H. e. magnificus from the eastern side of the Andes.

9
Living Insects and Host Plants
Two subspecies were maintained in temperature controlled Lord & Burnham glass
houses (5 x 8 m) at the University of Florida, Department of Entomology and
Nematology. H. e. favorinus were the descendents of 60 individuals collected at Tingo
Maria, Huanuco, Peru, which is located in the valley of the upper Huallaga river, at the
base of the western slopes of the eastern chain of the Andes (locally known as "Cordillera
Azul"). H. e cyrbia were the descendents of 50 individuals collected at Pichincha,
Ecuador, located on the western chain of the Andes. The two subspecies were kept in
separate cages to avoid hybridization. Butterflies in each colony were provided with their
natural host plants. Pass (flora rubra, P. punctata, and P. aur ¡culata from Ecuador are
the natural host plants of H. e. cyrbia, while P. trifasciata is one of the natural host plants
of H. e. favorinus. Although this Passiflora species is reported here as P. trifasciata, it is
actually an undescribed sister species of P. tricuspis and P. trifasciata (McDougal
personal communication).
P. biflora from Costa Rica is an alternative host plant for both subspecies. Lantana
and Pentas plants were used as adult nectar sources, while Psiguria spp. flowers were
used for pollen feeding.
Cups containing 20% sugar-water and red and orange Tuffy" sponges sprayed
with sugar water and an adult artificial diet of an amino acid solution that simulates the
amino acid content of Psiguria flowers (Gilbert unpublished) were also provided to
ensure an adequate food supply to the adults.
Feeding Experiments
Seventeen H. e. favorinus females were used for the host plant feeding experiment.
Iwelve H. e. cyrbia females were used to represent the Ecuadorian race. The colony was

10
obtained from a company in Ecuador called Heliconius Works Quito, Ecuador. Some H.
e. cyrbia were also obtained from other colonies in the USA. Since I was not sure
whether this race was fed on its natural host plant as larvae. I raised H. e. cyrbia on their
natural host plants P. punctata and P. rubra for at least three generations prior to starting
my experiments.
The offspring from each female are referred to as family and were so tracked
through the experiments. This identification allowed me to recognize any life history
differences between offspring from different female butterflies. Such differences would
indicate genetic variations within populations for the trait in question. Analyzing the egg
mass reveals potential maternal effects.
Isolated females were allowed to lay eggs on their natural host plants. The eggs
were then collected and taken to environmental chambers for the hatching larvae to grow
under constant laboratory conditions (27C, 75% humidity and a 14L:10D photoperiod).
Life History Studies
Different aspects of the life histories were observed and contrasted between races
from the time the egg was laid until adult emergence from the pupae.
As the larvae were raised, they were checked at least three times a day (morning,
afternoon, night) for the presence of head capsule apolysis, or for evidence of ecdysis.
Larval developmental time in each instar was recorded. When larvae in the last instar
were observed in the position typical of pupation, that day was recorded as the first day
of pupation. Pupal developmental time was recorded to have ended when the adult
butterfly emerged from the pupal case.

11
Egg measurements
Eggs were collected every day from the experimental cages and placed in a small
SOLO cup. Each egg was weighed on an electro balance graduated to 0.01 mg (Mettler
AC 100, Mettler Instruments Corp. Hightstown, New Jersey). Each egg was placed in an
individual cup with the family identification number (ID), offspring ID, host plant and
oviposition date. Fresh shoots were placed into the cups, usually on the second and third
day when I expected the larvae to hatch from the egg. A small wet ball of Kimwipe
tissue was also placed in the cup to provide eggs with adequate moisture.
Larval development
Upon hatching, larvae were removed from the cups with a fine artists brush and
placed on shoots of the experimental host plants (P. auriculata, P. punctata, P. rubra. P.
trifasciata and P. biflora).
The base of each plant cutting (approximately 15 cm in length) was placed in a
glass vial with distilled water. The bottle with the plant was then placed in a plastic
aquarium (one gallon) that was labeled with the following data: family ID, host plant
used, geographical race of erato, and the offspring ID.
As a larva develops, it periodically molts, which is first signaled by apolysis of the
head capsule. The head capsule is a strongly sclerotized part of the larval body and is
therefore a good indicator of growth. Consequently, I measured head capsule
dimensions. Head capsule width and height were measured. Measurements were made
with an Olympus microscope and an ocular micrometer. All measurements were done at
20X magnification. Relying on head capsule width has one drawback. In the last instar,
the epicranial suture splits the head capsule during the molt, so these measurements are
not accurate indicators of growth at this instar. Head capsule height, however, is not

12
affected by the split, so one can use the height of the last instar head capsule to indicate
growth during the last instar. The following formula was used to calculate growth rate:
Growth rate = (In (Hf) In (H¡))/t
where Hf is the measurement of the fifth instar head capsule height; H¡ is the
measurement of the first instar head capsule height; and t is the larval developmental
time. A natural logarithm transformation was performed as recommended by LaBarbera
(1989).
Pupae and adult offspring measurement
Each pupa was weighed two days after pupation. Measurements of pupa length
and width also were taken.
For the adult offspring, I recorded the following data: sex, wing length, total body
length (measured from the head to the tip of the abdomen), and fresh weight.
Measurements were made in the afternoon to allow time for the new adults to dry their
wings after emergence. A Vernier caliper (Spi 2000) graduated to 0.1 mm was used for
those measurements. Mass measurements were made with an electro balance. Teneral
butterflies were then stored at -85C for further analysis.
Statistical Analysis
Systat program V 10 was used for the statistical analysis. Since the variation in the
head capsule size between instars w'as large, the measurements were converted to a
logarithmic scale to do the statistical analysis. Repeated measures ANOVA was used for
analyzing growth trajectories between races. The statistical test was performed on both
head capsule width (using the measurement from the first to the fourth instar) and head
capsule height. Since both yielded the same results, I present here the results of the
analysis of head capsule height. Finally ANOVA was used for the rest of the analysis. In

13
all cases, family was nested within race to observe either maternal or genetic variation.
Also, sex was used as factor and if the interaction was not statistically significant it was
excluded from the analysis. A Tukey test was used for multiple comparisons.
Results
Analysis of Specimens from the Museum Collection
The ANOVA of the measurements taken from the museum specimens supports the
hypothesis that subspecies from the eastern side of the Andes tend to be larger than the
subspecies from the western side of the Andes (Table 2-1). Regions alone had a large
effect on size (F=254.2; df= 1, 179; pO.OOl), and different races of erato within the
regions are also different (F=14.4; df= 5, 179; pO.OOl).
The subspecies from the eastern region are relatively constant with respect to their
body size. In contrast, the subspecies from the western side of the Andes show more
variability in their sizes.
Life History Studies in H. e.favorinus and H. e. cyrbia
Data collected in the field (by myself and museum specimens) indicated that H. e.
favorinus is larger than H. e. cyrbia (ANOVA for wing length F= 283.8, pO.OOl and
body length F=45.3, pO.OOl) (Figure 2-3).
Eggs
Heliconius erato egg mass was not significantly different between races.
However, egg mass was strongly influenced by family within each race (F= 5.69,
pO.OOl). This result indicates maternal effects for egg mass. Both races laid eggs
singly, mostly on young growth (tendrils, and the tips of developing natural Passiflora
host plants). The eggs are cylindrical, yellow, with a pointed apex resembling a crown. I
did not find any difference in egg morphology between races. First instars hatch from the

14
egg on the third day following oviposition. All first instars in both races ate their egg
shells.
Larvae
Both subspecies underwent five instars of development. No supernumerary instars
were detected during the experiment.
Three different analyses were carried out with head capsule measurements. The
first was the comparison of head capsule size and larval developmental time between and
within geographical races feeding on natural vs. alternative host plants.
Repeated measurements ANOVA for head capsule size showed a significant
variation in size between the two races (Table 2-2). Analysis within H. e.favorinus
showed that host plants directly affect head capsule size (Table 2-2). Tukey comparison
test indicates that individuals fed on P. auriculata tend to be bigger than individuals fed
P. trifasciata and P. rubra during the second, third and fourth instars of larval
development. No significant differences in head capsule size were found among
members of H. e. cyrbia. There was no size variation among full-sib within races.
Differences in larval developmental time between and within races were
statistically significant between family and between individuals fed different host plants
(Table 2-3). On average, H. e. favorinus larvae fed longer on P. auriculata and P. rubra
(fable 2-4). Both H. e. favorinus, and H. e. cyrbia also showed significant differences
between larval developmental time between full-sibs, indicating genetic variation within
them and between individuals fed different hosts (Table 2-3). Longer larval
developmental time was experienced by both male and female cyrbia fed on rubra, as
well as females fed auriculata and males fed trifasciata (Table 2-4).

15
Both H. e. favorinus and H. e. cyrbia had plastic growth trajectories (Figure 2-4
AB) (Higgins and Rankin 1996). However, as indicated above in H e. favorinus the
plasticity is due to variation of head capsule size and developmental time, while in H. e.
cyrbia it is only due to variation in larval developmental time. Differences in growth
trajectories for the two races on their natural host plants can be observed in Figure 2-5.
The last part of the analysis was to calculate growth rates and compare them
between races. Growth rates were significantly different between races (F=6.18; df=l,
227; p<0.01). Host plants also affected growth rate (F=12.11; df=4, 227; pO.OOl)
between races. Lower growth rates in female H. e. favorinus fed P. auriculata, and in
males and females fed P. rubra, were due to prolonged development. Within H e.
cyrbia, low' growth rates occurred when males and females were fed P. rubra (Table 2-4).
Pupae
The average pupal developmental time for both races was eight days. When H. e.
favorinus fed on different host plants, it resulted in a significant variation in pupal mass
(F-4.57; df=4, 145; pO.OOl). Multiple comparison test indicated that individuals fed P.
trifasciata were significantly heavier than individuals fed P. punctata, P. auriculata and
P rubra (Figure 2-6).
Adult
Final size under laboratory conditions corresponds to the data gathered in the field
and from museum specimens. Regardless of the host plant. H. e. favorinus was greater in
wing length than H. e cyrbia (F= 271.63; df=l, 230; pO.OOl) (Table 2-4).
Analysis within H. e. favorinus indicates that host plant selection caused a slight
but significant variation in wing length (F= 2.34; df=4, 145; p=0.05). Multiple
comparison showed that individuals that fed on P. punctata were smaller in size than the

16
ones fed P. trifasciata. No other host caused a significant variation in wing length. On
the other hand, H. e. cyrbia wing length did not differ significantly between individuals
fed on different hosts. There was no size difference between sexes (Figure 2-7).
Mortality'
In general, higher mortality occurred throughout the life cycle in H. e. cyrbia. I
started with 148 individuals from H. e. cyrbia and 143 individuals from H. e.favorinus.
At the end of the experiment I had a total life cycle mortality of 38.74% in H. e. cyrbia
and 13.26% in H. e. favorinus. During the experiments, 8 eggs from H. e.favorinus and
10 from H. e. cyrbia did not hatch; some were dried, others were bitten by ants. High
mortality during the larval stage was observed in H. e. cyrbia fed its host plant P. rubra.
In all cases, larvae died during the early instars (Table 2-5).
Discussion and Conclusions
The results presented here support the preliminary hypothesis that there are
differences in body sizes between races of Heliconius erato from the western and eastern
sides of the Andes. While the group of races from the eastern side of the Andes reveal
low variations in body size, the group of races from the western side of the Andes present
much greater variation. In the eastern region, stabilizing selection probably acts against
extreme phenotypes and favors the more common variant. This mode of selection may
reduce variation and maintains the status of this particular large phenotypic character.
The races of the western side of the Andes are perhaps more exposed to ecological
pressures, such as competitive interactions, for example, as discussed by Benson et al.
(1976). As Brower (1994) has explained, the vicariant separation of the races could have
occurred during the Cenozoic era, specifically the Pleistocene, when the formation of
large mountains around the world (including the Andes) occurred.

17
In this chapter, I demonstrated through life history studies of the two races of
Heliconius erato that differences in body size seem to be genetically based and size
differences will remain, regardless of the host plant provided. Within each population,
there were maternal effects confirmed by egg mass variation between full-sibs. Also,
genotypic variation occurred in each stage of the life history within each race.
//. e.favorinus proved to have more plastic growth trajectories, due to variation in
their average larval developmental time and variation in size when the larvae were fed
different host plants. This result demonstrates that both developmental pattern and
growth are influenced by host plants (chemical/nutritional). It also indicates that
maturation is not simply a consequence of growth, and that environments can have
correlated effects on both growth and developmental rates. Within this race, I also
proved that when favorinus fed on their natural host plant, P. trifasciata, as well as the
alternative host P. punctata, they generate earlier-born progeny. Earlier-bom progeny
might have a significant impact on population growth (Fisher 1930 and Charlesworth
1980) because they might increase reproductive success and fitness in stable populations.
This particularly occurred in H. e. favorinus fed its natural host P. trifasciata, which
produced larger individuals of both sexes. Larger individuals will again have a direct
effects on fecundity and competitive ability (mating success), especially in competitions
for pupal mating (Deinert et al. 1994). However, the alternative host plants (P. punctata)
produced smaller females, on average, which directly affects fitness, because smaller
females will tend to produce fewer eggs on average (Dunlap-Pianka 1979 and Boggs
1979). Interestingly, fast growing individuals become heavier than individuals that grow
more slowly. This is probably due to faster food consumption (not measured here). Final

18
adult size indicates that there is only a slight difference (p=0.05) in size offavorinus
when fed its natural host plant compared to other host plants. This result indicates that
there is a strong selection for being the right size" as well as having a short
developmental time.
On the other hand, H. e. cyrbia also presented plastic growth across host plants.
However, the most striking result was the high mortality and low growth rate when
individuals fed on their natural host plant, P. rubra. Although not significant, there was
also a lower growth rate when female larvae fed on P auriculata and they generated
smaller females. This particular race has been investigated (Jiggins et al. 1997) in
relationship to their hybridization zones and the recent speciation of H. himera. In
contrast to their results, my greenhouse colony of cyrbia displayed a marked preference
for oviposition on P. punctata. H. e. cyrbia was exposed to both natural host plants for
more than 10 generations. Two important observations regarding host plant preferences
were made during these studies (Hay-Roe and Rios, personal observations),
a. On average, we collected 3 to 5 eggs per day on P. rubra and 10 to 12 eggs per day
on P. punctata.
b. On various occasions, P. punctata were defoliated when we allowed the larvae to
grow on the plants inside the cage. P. rubra were never defoliated.
Regardless of the observations of oviposition behavior, the results of this study
suggest that there is a difference in survival of larvae fed P. rubra vs. P. punctata, a fact
that Jiggins et al. (1997) did not examine explicitly (also, Mallet personal
communications). Perhaps, one ought to revive the hypothesis that differences in host
plant usage between the sister species H. e. cyrbia and H. himera could be one cause that
induced the recent speciation of H. himera.

19
Several factors might be operating during the life history of these subspecies,
creating selective pressures affecting larval growth rates and trajectories (e.g. temporal
differences in host availability, chemical differences among hosts that affect survival
ability, etc.) The results of this study suggest that H. e. favorinus employed different
strategies than H. e. cvrbia did to overcome the intake of different chemicals and/or
nutrients from the host plants. This is suggested by differences in larval performance and
survival. In the following chapters 1 shall explore the importance of plant chemistry as
related to the coevolutionary interaction between H. erato and Passiflora host plants.

20
Table 2-1. Average wing length and standard error of geographical races of Heliconius
erato from museum specimens in different regions.
Geographical races
Regions
N
WL (cm)
X se
H e. magnificus
Eastern
15
3.85 0.06
i H. e.favorinus
Eastern
57
3.94 0.03
H. e. phyllis
Eastern
32
3.89 0.04
H. e. hydara
Western
15
3.30 0.06
H. e. petiveranus
Western
25
3.01 0.08
H. e. cyrbia
Western
31
3.55 0.03
j H. e. colombina
Western
15
3.24 0.03
Table 2-2. Repeated measurements ANOVA of head capsule size with race, host and
familv as factors.
Source
Df
F
P
Between races
Race
1
751.96
0.001
Host
4
3.86
0.005
Rac x Host
4
1.18
0.323
Error
223
Within races
H. e. favorinus
Family
16
1.47
0.120
Host
4
3.58
0.008
Error
129
H. e cyrbia
family
9
1.52
0.159
Host
4
1.32
0.272
Error
69

21
Table 2-3. Repeated measurements ANOVA of larval developmental time with race and
host as factors.
Source
Df
F
P
Between races
j Race
1
9.68
0.002
j Host
4
15.75
0.001
| Race x Host
4
0.62
0.649
Error
223
Within races
H. e. favorinus
Family
16
2.68
0.001
Host
4
33.89
0.001
Error
129
H. e. cyrbia
Family
9
3.27
0.002
Host
4
3.41
0.013
Error
69
fable 2-4. Sexual differences in larval developmental time (LDT). growth rate (GR) and
wing length (WL) in H. e. favorinus and H. e. cyrbia feeding on five different
host plants.
Race
Host Plants
N
Both
sexes
Sex
LDT (days)
^ise
GR
(mm/days)
Jise
WL (cm)
X1 se
H. e.
auriculata
24
8
9.1810.26
0.19710.006
3.95 1 0.044
favorinus
9
9.5410.31
0.18610.007
3.91 10.040
biflora
25
0
8.3710.18
0.21710.006
3.9810.027
$
8.6410.10
0.205 1 0.003
3.92 1 0.034
punctata
26
e
8.0710.12
0.21910.004
3.9410.033
9
8.3610.15
0.208 1 0.004
3.81 10.090
rubra
23
8
10.4610.28
0.17410.007
3.9010.060
9
9.7510.25
0.18810.006
3.92 1 0.046
trifasciata
26
cJ
8.84 10.21
0.205 1 0.006
4.0010.034
9
8.7910.18
0.208 1 0.005
4.0010.030
H. e cyrbia
auriculata
19
8
9.1310.30
0.211 10.007
3.58 10.060
9
9.7010.73
0.203 10.012
3.4810.085
biflora
19
8
8.3810.26
0.2261 0.006
3.68 1 0.047
9-
9.0010.88
0.22010.014
3.55 10.036
punctata
17
6
9.3810.99
0.21210.016
3.53 1 0.067
9
8.22 1 0.22
0.23010 008
3.58 1 0.036
rubra
17
8
9.8010.33
0.19010.007
3.5410.051
9
10.0010.26
0.1881 0.006
3.54 1 0.032
trifasciata
18
8
9.63 1 0.60
0.19810.014
3.51 10.063
.
1
?
9.0010.31
0.21510.007
3.61 10.029

22
Table 2-5. Mortality in different stages of the life history of H. e favorinus and H. e
Race
Egj
y
Host plant
Larvae
Pupae
Adult
N
%
dead
N
%
dead
N
%
dead
N
%
dead
H. e.
favorinus
8
5.6
auriculata
1
0.7
0
0
0
0
biflora
2
1.4
2
1.4
0
0
punctata
1
0.7
0
0
0
0
rubra
5
3.5
0
0
0
0
trifasciata
0
0
0
0
0
0
H. e.
cyrbia
10
6.75
auriculata
2
1.4
2
1.35
1
1.1
biflora
5
3.4
2
1.4
0
0
punctata
1
0.7
3
2.0
0
0
rubra
20
13.5
2
1.4
2
1.4
trifasciata
7
4.7
0
0
1
0.7

LARVAL DEVELOPMENT
Figure 2-1. Life history of Heliconius butterflies.
PUPA
\? \
ri:.
ca. 8
Days"
ADULT
ca. 3
Months

24
Figure 2-2. Museum specimens of H. erato from the eastern (left column) and western
sides (right column) of the Andes that were included in the analysis. (A) H. e.
phyllis, (B )H. e. favorinus, (C) H. e. magnificus, (D) H. e.colombina, (E)
hydara, (F) H. e. cyrbia, (G) H. e. petiveratms.

25
Figure 2-3. Study organisms. (A) H. e.favoritms from Tingo Maria, Huanuco, Per and
(B) H. e. cyrbia from Pichincha, Ecuador.

Head capsule height (mm) Head capsule height (mm)
26
Larval developmental time (days)
Larval developmental time (days)
Figure 2-4. Growth trajectories of (A) H. e. favorimis and (B) H. e. cyrbia feeding on five
different host plants. Each point represents the mean value of larvae from the
first to fifth instar. Bars on each point indicate standard error.

27
Figure 2-5. Growth trajectories of H. e.favorinus and H. e. cyrbia fed their natural host
plants (P. trifasciata and P. rubra, respectively). Each point represents the
mean value of larvae from the first instar to the fifth. Bars on each point
indicate standard error.

Pupal weight (g)
28
auriculata biflora punctata rubra
Host plant
A female
A male
B female
B male
trifasciata
Figure 2-6. Pupal weight of H. e.favorinus (A) and H. e. cyrbia (B) fed on natural and
alternative host plants. Different letters indicate significant differences
between means (Tukey HSD-test). Bars in each point indicate standard error.
auriculata biflora punctata rubra trifasciata
Host Plants
Figure 2-7. Average wing length of H. e.favorinus and H. e. cyrbia fed natural and
alternative host plants. Different letters indicate significant differences
between means (Tukey HSD-test). Bar in each point indicate standard error.

CHAPTER 3
CYANOGENESIS IN HELICONIUS ERATO AND PASS1FLORA HOST PLANTS
Cyanogenic glycosides are secondary metabolites derived from amino acids.
Insects or plants that contain cyanogenic glycosides undergo a process called
cyanogenesis, which occurs during tissue disruption by feeding herbivores, predators or
pathogens. The process occurs in two steps: first, the enzyme P-glucosidase hydrolyzes
the cyanogenic glycosides, generating a cyanohydrin (a-hydroxynitrile) and a sugar
moiety (usually D-glucose, but the products may include other sugars, e. g., gentibiose).
The second step involves the degradation of the cyanohydrin and occurs rapidly at
normal cell pH. This is thermodynamically favored, and in many instances, but not
always, a second type of enzyme (a-hydroxynitrile lyase) catalyzes the dissociation of the
cyanohydrin to a carbonyl compound and hydrogen cyanide (Hosel 1981) (Figure 3.1).
Occurrence of Cyanogenesis in Plants
Normally, the cyanogenic glycosides and the enzymes that hydrolyze them to
produce cyanide are spatially separated within the plant tissues to prevent autotoxicity
(Hosel 1981, Poulton 1988). In plants, the cyanogenic glycosides are found in the
epidermal layer of the plant tissues, whereas both hydrolyzing enzymes are found in the
mesophyll cells (Kojima et al. 1979, Saunders et al. 1977, Poulton 1988).
Cyanogenic plants show high variability in hydrogen cyanide (HCN) production,
which has been related to low moisture stress (Foulds and Grime 1972, Abbot 1977),
frost or low temperatures (Ellis et al. 1977), or presence of herbivores (Cooper-Driver
and Swain 1976). This variation reflects phenotypic and genotypic diversity in both the
29

30
production of cyanogenic glycosides and the existence or absence of their degrading
enzymes (Vetter 2000. Jones 1977, Hughes 1981. Schappert and Shore 1995, 2000).
Plants with such genetic variation constitute an important source of information on the
direct effects of this type of compound in defense against herbivores that prey upon the
plants (Jones 1971).
Occurrence of Cyanogenesis in Lepidoptera
Cyanogenic glycosides have been found in all genera of Heliconiinae, including
Cethosia biblis and C. hypsea (Heliconiinae: Nymphalidae) (Nahrstedt and Davis 1983).
Acraea horta (Acraeinae: Nymphalidae) (Raubenheimer 1989), the lycaenids Brephidium
exillis and Polyommatus icarus (Nahrstedt 1987), and in moths of the family Zygaenidae
(Franzl and Naumann 1985, Muhtasib and Evans 1987). The location within the body of
cyanogenic glycosides has been studied in Zygaena moths and Heliconius butterflies. In
both, the cyanogenic compounds were found in the cuticular part of the integument,
although in zygaenids, 30% of the cyanogenic compounds were also found in the
hemolymph (Davis and Nahrstedt 1985, Nahrstedt and Davis 1983, 1985). The enzyme
jj-glucosidase, on the other hand, is mainly stored in the hemolymph (Nahrstedt 1988).
All stages in the life history of Heliconiinae have been tested and the highest
concentrations of cyanide have been found at the egg stage. The concentrations in the
iarval and pupal stages do not vary greatly, but cyanide concentration drops in adults
(Nahrstedt and Davis 1985).
The variation of the cyanogenic glycoside concentrations within body parts of adult
Heliconius, as related to their host plants, has not been explored. Within insects that are
chemically protected, it is expected that the greater concentrations of the toxic
compounds will be found in areas of the body with which the predators come in contact

31
first. For example, toxic compounds are more concentrated in the wings of monarch
butterflies (Brower and Glazier 1975). Chrysomeiid beetles (of the genus Oreina)
rapidly distribute toxic compounds stored in pronotal and elytral glands onto the surface
of the cuticle (Pasteis et al. 1995, 1979) when disturbed. Variations in the concentration
of defensive compounds among body parts may be due to selection by predators for more
effective protection and/or may depend on the mode of action or properties of the toxic
compounds (Bowers 1988).
Cyanogenic Glycosides as a Defensive Mechanism
It is generally accepted that cyanogenic glycosides act as a defense mechanism
against many potential herbivores and predators (Jones 1988, Seigler 1991, Nahrstedt
1992, 1988, Gleadow and Woodrow 2002). In the last decade, there has been some
controversy over the effectiveness of cyanogenesis as a herbivore defense (Kassarov
1999, 2001, Hruska 1988), but the utilization of cyanogenic glycoside in plant defense
has recently been proven by transferring the pathway for cyanogenic glycoside
biosynthesis into the acyanogenic Arabidopsis thaliana (Tattersall et al. 2001). There is
also some doubt about whether HCN or the aglycone (aldehydes or ketones), which are
simultaneously released upon cyanogenesis, are the source of toxicity (Conn 1979, Jones
1988, Spencer 1987). For example, regurgitate of enteric fluids by the eastern tent
caterpillar, Malacosoma americana fed black cherry contained both HCN and the
carbonyl compound, benzaldehyde. When these compounds were tested on ants, they
tolerated the HCN but were repelled by the benzaldehyde (Peterson et al. 1987). Also,
Jones (1988) revealed that the carbonyl compounds, acetone and 2-butanone, acted as
deterrents to feeding on Lotus corniculatus and Trifolium repens by mollusks. Despite

32
this controversy, what is clear is that feeding is required in order to provoke tissue
disruption and enzymatic hydrolysis.
In Lepidoptera, defensive compounds have been classified in relation to how
predators react to them (Brower 1984). Class 1 are compounds that are noxious due to
their ability to irritate or poison; they may or may not stimulate the olfactory or gustatory
receptors of predators. Class II are harmless chemicals that stimulate the predators
olfactory and/or gustatory receptors (Brower 1984). Thus cyanogenic glycosides are
considered a Class I compound; however, Heliconius also possess Class II pungent odors
emitted through their abdominal glands, (Gilbert 1976), and pyrazines produced in some
unknown region of the body (Hay-Roe and McAuslane unpublished).
Laboratory experiments have demonstrated that insect predators such as birds,
frogs, and lizards can learn to avoid a particular class of prey, depending on the
unpalatability of that prey (Brower 1958, Brower et al. 1963, 1971; Chai 1990, 1996).
Cha) (personal communication) described how inexperienced birds will attack almost
anything. However, after tasting the unpalatable butterflies, they quickly become
conditioned to the pattern of coloration (Chai 1988) and flying styles of distasteful prey
(Srygley and Chai 1990). In the Costa Rican rainforest, Heliconius erato was rejected
41% of the time when tested with jacamar birds (Glbula ruficauda) (Chai 1986).
Various experiments have shown that some butterflies and moths that are generally
considered to be highly toxic can be consumed by birds without any rejection behavior
(Collins and Watson 1983, Chai, 1986). Within a single species, unpalatability cannot be
absolute; the amount and types of chemical defense can vary among individuals,
producing a palatability spectrum (Brower et al. 1967).

33
For this chapter, the concentrations of cyanogenic glycosides in various body parts
of races of Heliconius erato fed natural host plants were examined. The same analysis
was performed on separate parts of Passiflora plants (young leaves, mature leaves and
stems). Roots, fruits and flowers were not included in the analysis because Heliconius
larvae rarely feed upon these tissues. Cyanide quantification analysis of the plants was
performed over two seasons (summer and fall). Finally, cyanide concentrations in the
plant were correlated with the concentrations in H. erato which fed upon them. The
discussion addresses the costs and benefits of allocating cyanogens as a defensive
mechanism in the different body parts of this species and the implications of this
variation on the palatability of this butterfly and on its mimetic relationships.
Materials and Methods
Plant Growth and Insect Rearing
This study was conducted at the University of Florida in Gainesville with butterfly
colonies that were raised in Lord & Burnham glass houses (5x8 m). During the first two
years of experimentation the colonies were unable to survive the winter, but improvement
in the glass houses environmental conditions made breeding several generations possible
in the last years of research. H. e.favorinus was collected in Tingo Maria, Huanuco,
Peru, while H. e. cyrbia was collected in Pichincha, Ecuador, and H. e. demophon was
collected in Soberania National Park, Panama.
Most of the Passiflora plants began as cuttings brought from the collection of Dr.
L. E. Gilbert at the University of Texas at Austin; a few cuttings were also donated by R.
Boender, Butterfly World, Fort Lauderdale. P. trifasciata, the natural host plant of H. e.
favorinus, was collected in Tingo Maria by the author of this research. All vouchers of
Passiflora plants used for this study are deposited at the Missouri Botanical Garden, St.

34
Louis. Missouri and the identification of the plants has been confirmed by Dr. John
MacDougal, an expert in the systematics of the family Passifloraceae.
Experimental Design and Treatments
Butterflies
H. e favorinus, H. e. cyrbia and H. e. demophon larvae were raised in
environmental chambers under controlled temperatures (27C), humidity (78%) and
.photoperiod (14L:10D).
The first two races were fed only natural host plants, Passiflora trifasciata for H. e.
favorinus, and P. punctata and P. rubra for H. e. cyrbia. H. e. demophon, on the other
hand, was fed both natural host plants (P. biflora) and alternative host plants (Table 3.1).
The experimental procedure used was as follows:
a. Specimens were selected from both H. e. cyrbia and H. e. favorinus.
b. Each individual male and female was tested separately.
. For each specimen, the head, thorax, abdomen and wings were tested separately.
The procedure was slightly altered for the testing of H. e. demophon (the full
procedure had not been designed at the time that this testing took place). For this
subspecies the head, thorax and abdomen were not tested separately, only the wings were.
Temporal cyanide quantification of Passiflora plants
Various Passiflora plants were tested during the summer (July-August) and fall
(September-October) of 1999 (Table 3.2). All the plants tested were potential host plants
of Heliconius erato in different geographical areas, such as Brazil, Costa Rica, and
Ecuador.
The analysis was performed on three different plant parts: young leaves, mature
leaves, and stems. Plant parts not normally eaten by Heliconius larvae were not tested.

35
All Passiflora plants were subjected to the same environmental conditions
(humidity, temperature and natural photoperiod) in the greenhouses. None of the plants
tested was exposed to herbivores.
Quantitative Determination of Cyanide
Colorimetric analysis
Cyanogenesis was quantified by the Lambert procedure (Lambert et al. 1975)
which was modified by Brinkler and Seigler (1989, 1992).
The plant or butterfly samples were ground to a fine powder in liquid nitrogen.
The powdered plant or butterfly tissue was poured into 10 ml vials containing 0.1 M
phosphate buffer (pH 6.8). The volume of liquid was kept to a minimum, as cyanide is
very soluble in water. A 5 ml vial containing fresh 1M NaOH solution was then placed
inside the 10 ml vial and the 10 ml vial was stoppered tightly. The samples were
prepared quickly to avoid loss of HCN. The vials were then incubated at 37C overnight.
During the incubation, HCN released from tissue is trapped in the NaOH solution, and
forms NaCN.
After the incubation, aliquots of the solution from the small vial were transferred to
test tubes and adjusted to 0.1 M NaOH with distilled water. Typically, 0.1 ml aliquots of
1 M NaOH solution are removed and diluted to 1 ml; if the total sample volume is less,
0.1 M NaOH is added to bring the volume up to 1 ml.
The assay involved sequential addition of the following reagents: 1.0 M acetic acid
to neutralize the NaOH; a succinimide/N-chlorosuccinimide oxidizing reagent, and a
barbituric acid/pyridine coupling reagent (Lambert et al. 1975; Brinker and Seigler 1989).

36
The tubes were vortexed vigorously and the solution was allowed to sit for 10 min.
The absorbance was then measured at 580 nm in a spectrophotometer (Spectro 22;
Labomed, Culver City, CA).
Standard curve
For the standard curve, a 0.02 M solution of NaCN in 0.1 N NaOH (0.980g
NaCN/1) was prepared. The exact amount of CN in solution was determined by a
modified Liebig titration method as follows:
A standard 0.1 M AgNC>3 solution was prepared by diluting dry AgNC>3 (at 100C
for 3 h) in 1 L deionized water in a volumetric flask. The solution was stored in the dark.
Twenty five ml of the NaCN solution was placed in a 50 ml flask equipped with a stirring
bar. Into this solution, 1.5 ml 6 N NH4OH and 0.5 ml 10% KI were added. This mixture
was then titrated with the standard AgNC>3 solution, until the mixture became turbid,
which occurred at approximately 2.4 ml of AgNC>3. The endpoint was easier to see
against a black background.
To determine the amount of CN in the 25 ml sample, the amount of AgNC>3 used
to reach the endpoint was multiplied by 0.005204 g, then divided by 25. The result
reflected the g CN /ml.
Aliquots (1 ml) of the NaCN solution prepared above (close to 500 pg CN_/ml)
were then diluted in a volumetric flask with 0.1 N NaOH to 100 ml to reach a final
concentration of approximately 5 pg CN /ml.
The standard curve was then prepared using aliquots of the standardized NaCN
solution. Between 0 and 0.2 ml of NaCN are increased to 1 ml with 0.1 M NaOH. This
gives CN concentrations ranging from 0 to approximately 1 pg CN~/ml.

37
After the analysis, the ground plant or insect tissue was dried in crucibles for
several days in a drying oven at 65C, until they reached a constant dry weight (approx. 6
days). 1 recorded the pg CN /g dry weight of plant tissue or butterfly body tissue.
Predator-Prey Interaction
During the present study I consistently observed the presence of tree-frogs (Hyla
sp.) in the cage containing the colony of H. e. favorinus. In order to learn whether this
vertebrate was preying on the larvae, I collected frogs and confined them in a separate 1
gallon empty aquaria until the frog had excreted. No permit by the Institutional Animal
Care and Use Committee (UF) was required to conduct this experiment.
Statistical Analysis
Two-way analysis of variance (ANOVA) was performed using the statistical
program SYSTAT 10. Some of the data was not normally distributed and the data was
subjected to square root transformation to compensate for increasing variation associated
with larger mean values (Zar 1996). However, the results using transformed data did not
vary significantly from the raw data, so the results of the raw data are reported here. A
Tukey test was used for multiple comparisons.
In order to learn whether cyanide concentration in butterflies is correlated with
cyanide concentration in Passiflora host plants, a linear regression was done on the mean
cyanide concentration of Passiflora host plant with mean cyanide concentration of H.
erato fed natural host plants.
Results
t emporal Variation of Cyanide Release in Passiflora Plants
Cyanide concentrations varied significantly between the two analyzed seasons
(F=l29.95; df=l, 388; p<0.001). In the fall, cyanide concentrations increased to

38
approximately two times the amount released in the summer. Also, there was a
significant difference in cyanide concentrations between host plants tested (F=45.06;
df=6, 388; p<0.001) and plant parts (F=l31.83; df=2, 388; p<0.001). The concentration
of cyanide was always higher in the new leaves, and this value differed significantly from
the cyanide concentration in both mature leaves and stems. Cyanide concentration in
mature leaves and stems did not follow a constant pattern, and varied according to the
species (Figure 3.2).
Cyanide Quantification in Heliconius erato Fed Their Natural and Alternative Host
Plants
Cyanide concentrations varied significantly between the races (F=29.90; df=l,
139; p<0.001). H. e. cyrbia released greater amounts of cyanide than H. e.favorinus.
There were also differences in cyanide concentration between the sexes (F=8.26; df=l,
139; p<0.005) and also between body parts (F=8.14; df=3, 139; pO.OOl).
H. e. cyrbia fed on P. rubra and P. punctata
Differences in host plant utilization resulted in different concentrations of cyanide
being released from individuals of this race (F=10.02; df=l, 97 p<0.002). When H. e.
cyrbia fed on P. rubra (Table 3-3), the butterflies released higher concentrations of
cyanide, as compared to the butterflies that fed on P. punctata (Figure 3.4). Also, males
and females of this subspecies differed significantly in the quantity of cyanide released
(F=5.68; df=l, 97; p<0.025). The females were able to concentrate higher amounts of
cyanide in their body when fed P. rubra. Furthermore, the concentrations of cyanide
were stored differently within the body parts (F=7.32; df=3, 97; p<0.001). The thorax
had higher concentrations of cyanide, which differed significantly from the amounts
stored in the head, the abdomen, and the wings. When H. e. cyrbia was fed P. punctata,

39
the cyanide concentrations did not differ between the sexes. However, there was a
significant difference in concentrations between body parts (F=12.31; df=3, 52; p<0.001).
The thorax and the head had higher cyanide concentrations, as compared to the abdomen
and wings (Figure 3.4).
H. e. favorinus fed P. trifasciata
Low concentrations of cyanide were found within the body of this race. Although
the female thorax and head seemed to have higher cyanide concentration than the males,
the statistical analysis did not show any significant differences between these values.
Wings and abdomen had the lowest concentrations of cyanide. More males than females
were tested because more than 50% of the males tested negative. Addition of exogenous
p-glucosidase did not favor the release of cyanide, indicating that cyanide release is not
due to the lack of the enzyme, but to the lower concentrations of cyanogenic glycosides
within the butterflys body (Figure 3.5).
H. e. demophon fed on natural and alternative hosts plants
Host plant selection had a significant effect on the cyanide concentration found in
//. e. demophon (F=10.15; df=3, 50; pO.OOl). Butterflies fed P. rubra and P. biflora
exhibited significantly lower concentrations of cyanide than butterflies fed P. auriculata
and P. trifasciata. Also, the interaction between host plant and sex was statistically
significant (F=7.34; df=3, 52; pO.OOl). The natural host plant, P. biflora, is associated
with lower cyanide concentrations in the male butterflies. Furthermore, cyanide
concentrations varied between body parts (F=4.96; df=l, 52; p=0.03). In general, the
body (which in these samples includes the head, thorax and abdomen) possessed higher
cyanide concentrations than the wings (Figure 3.6).

40
Cyanide Correlation Between Heliconius erato Fed Natural Host Plants
The regression line in Figure 3.7 indicates that 68% of the variation in cyanide
concentration in Heliconius butterflies is due to differences in the cyanide concentration
in the host plants.
Predator-Prey Interaction
At least 10 frogs were collected within the captive H. e.favorinus butterflies
colonies and the feces was analyzed during this study. I found, on average, 4 head
capsules and many spines of H. erato in each frogs excreta. The Heliconius head
capsules corresponded to the third and fourth instar of development. Usually, I found the
frogs camouflaged on the main stem of the plant or on leaves. Occasionally, I found
frogs on the cage walls. Frog feces also contained body parts of other insects such as
cockroach legs and ant heads.
Discussion and Conclusions
Cyanide Concentrations in Heliconius erato fed Passifiora host plants
Host plant selection was the main reason for the variation in cyanide concentrations
between the different races of Heliconius erato. In natural habitats, resource availability
and competitive interactions are among the factors that play important roles in this
variation. Within each race, genetic variation is probably responsible for differences in
the ability to sequester cyanogenic glycosides.
The results of the analysis of the different body parts of H. erato are related to the
allocation of the defensive compounds in the butterflies and the mode of action of the
chemicals involved in defense. The thorax and the head were the body parts with higher
cyanide concentrations. These results did not follow the patterns of cardenolide toxicity
reported for monarch butterflies (Brower and Glazier 1975). Since the cyanogenic

41
defensive mechanism requires hydrolysis to function, we expected to find higher
concentrations of cyanide release at the interface of the hemolymph containing (3-
glucosidase and the cuticle. The thorax is a heavily sclerotized body part, filled with
hemolymph (Nation 2002). These two properties allow a heavy deposition of the
cyanogenic glycosides on the cuticle. By mixing these glycosides with hemolymph
containing (3-glucosidase, the butterfly is able to generate high concentrations of cyanide.
A similar rationale also enables us to understand finding high concentrations of cyanide
in the head. The head is small in mass and releases smaller amounts of cyanide, as
compared to the thorax, but since the head has a large surface area to volume ratio the
volatile cyanide will be released faster. Low cyanide concentrations within the wings and
abdomen might be the result of low levels of hemolymph in the former, and low levels of
sclerotized area in the latter, which implies a low level of hemolymph containing P-
glucosidase in the w'ings and low levels of cyanogenic glycosides in the cuticle of the
abdomen.
Brower et al. (1988) hypothesized that higher cuticular concentrations of
cardenolides may have evolved due to pre-adaptation from biochemical deposition in the
developing cuticle, rather than having evolved due to selection from predation or
parasitism.
Cyanogenesis in Heliconius butterflies probably functions as an alert signal to
reinforce memory of toxicity, also called recall-to-mind mimicry (Rothschild 1984).
Under this assumption it is not necessary for selection to favor highly cyanogenic wings,
because low levels of HCN will be enough to trigger recall of experiences of the bad taste
and smell of HCN and carbonyl compounds, which are formed during cyanogenesis. A

42
naive predator will rapidly learn to avoid feeding on that particular pattern of wing
coloration (Chai 1988), and flight (Srygley and Chai 1990) and the members that
constitute a Mllerian ring will benefit from it. The selection for belonging to the rings
common pattern will be reinforced. A similar mechanism of avoidance is likely used for
the larval stage, however a definitive test was not run. Hyla frogs were more consistently
found in the cage of H. e. favorinus, than in the cages of H. e. cyrbia and H. himera,
suggesting that larvae of H. e. favorinus were more palatable. On the other hand, being
highly unpalatable has a negative aspect as well, as we observed with the members of H.
e. cyrbia fed with P. rubra. Members that fed on this particular host plant gained the
benefit of being distasteful, but at the same time, there was a high cost paid for this
benefit, because individuals feeding on P. rubra spent a longer period feeding on the
host, which increases the time they are exposed to parasitoids. On average, individuals
feeding on such plants are slightly smaller in size than comparable insects feeding on
other host plants.
H. e. favorinus. seems to be better adapted to detoxifying and tolerating different
chemicals in the plants, probably due to more efficient degradative enzymatic activity, or
higher tolerance of chemicals, for example by having a target-site insensitivity (Brattsten
1979, Lindroth and Weisbrod 1991) because excreta was not found to be cyanogenic.
The differences between strategies for coping with cyanogenic compounds probably
reflect genetic variation between the races of H. erato.
Lastly, it has been demonstrated that, at least for H. e. cyrbia, there is a limit range
of toxicity because of high mortality at the larval stage, implying that there is a level of
effectiveness of cyanogenic defenses in the plants in relation to a herbivore. However,

43
there are other factors such as specialization, mode of feeding, etc, that herbivores could
also use to cope with plant cyanogenesis.
Temporal Variation in Cyanide in Passiflora Plants
Considerable variation in cyanide concentration was found among Passiflora tested
over two seasons (summer and fall). The concentration of cyanide in the fall was almost
double the amount found in the summer season. I could not find any literature related to
seasonal variation in cyanogenic glycosides. Most of the studies are linked to herbivore
presence (Cooper-Driver et al. 1977), or to altered climatic conditions such as
temperature (Ellis et al. 1977, Rammani and Jones 1985, and Jones and Rammani 1985),
which, it is suggested, contribute to seasonal changes in cyanogenic glycosides levels.
However, in this study, none of the above factors is a contributing factor in the
differences in cyanide concentration, since the biotic and abiotic conditions were similar.
The only environmental factor that might be responsible for the variation is the difference
in photoperiod due to seasonal change. A study of seasonal variation in flavonoid
concentration performed on temperate-zone Alliaria plants (Haribal and Renwick 2001)
produced results similar to those obtained in this study. (Flavonoid concentrations were
also lower in the summer and higher in the fall.) Because all of these Passiflora are
tropical species where photoperiod varies only 30 minutes during a year from a 12
light: 12 dark cycle, this is a puzzling finding that deserves further investigation. Also, it
w'ould be interesting to see how this variation in cyanide concentration correlates to the
survival rates of larvae and/or adults of Heliconius species.
Finally, different Passiflora parts were found to possess different cyanide
concentrations. Higher concentrations of cyanide are found in the new growth. The
concentration within the plant parts followed the same pattern of alteration between the

44
summer and fall seasons. The result of the distribution of cyanogens within the plant is
consistent with the optimal allocation theory (McKey 1974) that states that vulnerable
plant tissues and organs are defended more than older senescing ones. It has already been
demonstrated elsewhere that for several cyanogenic species, young developing tissues
contain the highest amounts of cyanogenic glycosides (Conn 1981a).
Implications of Cyanide Concentration for Palatability
It seems that H. e. favorinus is better adapted than H. e. cyrbia to handle different
compounds because its larval survival rate is higher, whether feeding on natural or
alternative host plants (as has been demonstrated in Chapter 2). H. e. favorinus' superior
ability to handle different chemical compounds has both benefits and costs in terms of
survival. The survival benefit is reduced larval mortality due to plant toxicity as the
larvae is less likely to be killed by the plants chemical defenses. The survival cost is
increased palatability (the larvae may be less toxic), which increases the potential for
larval mortality from predation (e.g., by Hyla frogs). H. e. cyrbia, on the other hand, has
the advantage, perhaps, of being more toxic, implying better protection against predation,
but at the cost of high mortality during larval development.
Implications of Cyanide Concentration for Mimicry
Different concentrations of cyanide within a butterflys body produce a palatability
spectrum (Brower et al. 1967). The palatability spectrum theory states that if members of
a Mllerian mimicry complex are not equally distasteful, then the presence of individuals
that are less distasteful would lead to an increase in predation on the models, and this will
result in an overall weakening of the Mllerian mimicry.
H. erato and its co-mimic H. melpomene are members of a Mllerian mimicry ring
of races in the South and Central America rainforests. There is some disagreement as to

45
whether erato or melpomene is the model for this ring, but Mallet (1999) listed three
different reasons that indicate that erato is the model. The first is that erato is more
abundant in the areas where it coexists with H. melpomene; second, erato has a broader
geographical distribution, and third, erato has a broader habitat than melpomene. If H. e.
favorinus in Tingo Maria, Huanuco is less palatable, then this species is in fact a weak
model. We should then expect that this will directly affect the presence of the less
common H. melpomene. This would be true if P. trifasciata were the only host plant
used by H. e. favorinus in the area. However, P. auriculata (which is also the host plant
of Heliconius sara), P. laurifolia, P. tricuspis, P. rubra, and P. vespertilio have been
recorded in Huanuco (Brako and Zarucchi 1993) and can be potential host plants for H. e.
favorinus.
In the last four years, doubts have been raised with regards to the existence of the
classical Mllerian mimic. Instead, some have proposed a new term quasi-Batesian
mimicry (Speed and Turner 1999) as an alternative explanation for some types of
polymorphism associated with mimicry. Quasi-Batesian mimicry describes Mllerian
mimics as weakly unpalatable, which benefit from the presence of more unpalatable
conspecifics.
In this particular case, members of H. e. favorinus that fed on P. trifasciata would
be considered quasi-Batesian mimics, within their race. The males get protection from
the presence of more unpalatable conspecifics (females or other conspecifics that did not
iced on P. trifasciata). However, the females will be affected, since this sex is probably
more exposed to the attack of predators, particularly during oviposition and while
searching for host plants.

46
Table 3-1. Heliconius erato fed natural and alternative Passiflora host plant and sample
size used in t
e study.
Geographical race
Host plant
N
Male
Female
H. e. cyrbia
P. rubra
6
6
P. punctata
7
7
H. e. favorinus
P. trifasciata
5
4
H. e. demophon
P. biflora
5
5
P. trifasciata1
3
3
P. aur¡culata1
5
5
P. rubra1
2
2
1 Alternative host plants from other regions
Table 3-2. Passiflora plants tested for cyanogenesis in this study. All plants are used by
Heliconius erato in different geographical areas.
Subgenus
Section
Passiflora
species
Country of
Origin

Summer
Fall
Decaloba
(Plectostemma)
Decaloba
P. biflora
Costa Rica
22
12
Decaloba
P. misera
Brazil
12
6
Decaloba
P. punctata
Ecuador
6
0
Decaloba
P. trifasciata
Peru
17
6
Auriculata
P. auriculata
Ecuador
22
10
Xerogona
P. capsularis
Brazil
6
14
Xerogona
P. rubra
Ecuador
12
0

47
p-glucosidase
Cyanogenic glycoside
Toxic
a-hydroxynitrile lyase
T
+ HCN
aldehyde/ketone
Toxic Toxic
a-hydroxynitrile
Unstable
Figure 3-1. General process of enzymatic hydrolysis of cyanogenic glycosides.

pg CN/g of dry tissue pg CN/g of dry tissue
48
Passiflora plants
auri culata biflora capsularis misera trifasciata
Passiflora plants
Figure 3-2. Cyanide quantification o Passiflora plants in the (A) summer and (B) fall
seasons. Bars indicate standard error.

49
Head Thorax Abdomen Wings
Body parts
Figure 3-3. Cyanide quantification of male and female H. e. cyrbia fed P. rubra.
Different letters indicate significant differences between means (Tukey HSD-
test). Bars indicate standard error.
Female
Male
Body parts
Figure 3-4. Cyanide quantification of male and female H. e. cyrbia fed P. punctata.
Different letters indicate significant differences between means (Tukey HSD-
test). Bars indicate standard error.

50
O)
Vi
600
500
400
300
200
>>
la
o
M
O
D£
z
u
M 100
B
L
B
B
Head
Thorax Abdomen Wings
Body parts
Female
Male
Figure 3-5. Cyanide quantification of male and female H. e.favorirtus fed P. trifasciata.
Different letters indicate significant differences between means (Tukey HSD-
test). Bars indicate standard error.
3
o
,W)
z
u
or
Body
Wings
Figure 3-6. Cyanide quantification of male and female H. e. demophon fed natural (P.
biflora) and alternative host plants. Bars indicate standard error.

51
c
o
1000 -1
9
900
k.
c
800
g i
700 -
o w
600
500
t
400 -
ih O)
ss 3
300
-O
c
200 -
(V
o
100 -
s
0 -
y = 0.8726x- 356.67
R2 = 0.6808
H. e. cyrbia
& P. rubra
600 800 1000 1200
Mean plant CN concentration (pg CN/g tissue)
1400
Figure 3-7. Regression of the mean cyanide concentration of H. erato over the mean
cyanide concentration of the natural Passiflora host plants upon which they
fed. (The means were used in the figure to enhance the clarity of
presentation.)

CHAPTER 4
IDENTIFICATION OF CYANOGENIC GLYCOSIDES IN PASSIFLORA AND THEIR
FATE WHEN INGESTED BY HELICONIUS LARVAE
In the tropical rainforest, the evolutionary relationships between insects and their
host plants have driven a variety of interactions and adaptations. Insect pressure has led
to the evolution of mechanical and chemical defenses in plants. Herbivores, in turn, have
evolved to overcome all of these defenses and are a strong selective force in the evolution
and diversification of plant secondary compounds. This evolutionary process, called
coevolution, was first introduced by Mode (1958) and was developed by Ehrlich and
Raven (1964), who proposed the theory of radiation and escape between plants and
butterflies.
A few life history studies have examined the correlations between offspring size,
maintenance, and growth, but usually did not relate the resulting integrated life history
strategies to toxic secondary compounds in plants. In chapter 2, the life histories of H.
erato fed natural host plants vs. alternative host plants were investigated and in chapter 3
it has been shown that butterflies feeding on different host plants generate different
concentrations of toxic compounds which are used for their own defense. The
compounds found in insect bodies are synthesized de novo, and/or sequestered from
plants.
In plants, from which butterflies and moths often sequester toxic secondary
compounds, there are three principal building blocks for these compounds: (1) acetate,
(2) amino acids, and (3) shikimic acids (Trigo 2000). Passiflora plants containing
52

53
cyanogenic glycosides use amino acids as building blocks. These toxic substances take
part in a plants chemical defense against insects that feed upon it.
Cyanogenic glycosides are secondary metabolites that possess intermediate polarity
and are water soluble. Chemically, they are defined as O-P-glycosides of a-
hydroxynitriles (cyanohydrins), biosynthetically derived from amino acids. Cyanogenic
glycosides generally co-occur with P-glycosidases, which are specific to the type of
cyanogenic glycosides and are spatially separated in the plant or animal tissue to avoid
autotoxicity. The enzymatic cleavage that occurs upon disruption of the tissue releases
HCN plus sugar, and ketones or aldehydes. It is thought that the high degree of
specificity of the glycosidase is due to the structure of the aglycone (Hosel 1981).
Subspecies of Heliconius erato (Nymphalidae: Heliconiinae) are aposematic
butterflies that have a tight relationship with Passiflora plants. They specialize in the
group of Passiflora plants of the subgenus Decaloba. Heliconius erato evidently
managed to counteract the effects of the cyanogens in their specific Passiflora host
plants, but their mechanism for overcoming these cyanogens are unknown. (See,
however the new mechanism of detoxification in Heliconius sara described below.)
Toxic plant compounds would set off many reactions in the insects digestive
system. The mechanisms employed by insects to counteract these chemicals vary
between species and can be used solely or in combinations as described below.
Mechanisms Used by Insects to Process Toxic Compounds from Plants
Blum (1983) suggests three ways in which an insect can cope with host allelochemicals:
a. Excretion and egestion: Compounds may be eliminated from the insects body, i.e.,
nicotine is eliminated without any transformation from the body of Trichoplusia ni
and Heliothis virescens (Self et al. 1964).

54
b. Metabolism: Toxic compounds may be converted into less harmful derivatives
through solubilization or conjugation. The metabolites are then reused in the
intermediary metabolic pathway of the organism, or they can be converted to a
metabolic product that can be stored or excreted. The first alternative is common
for toxic compounds that are highly water soluble such as toxic amino acids,
whereas lipid soluble toxins are converted to water soluble metabolites and then
eliminated as such, rather than being recycled (Brattsten 1992).
c. Sequestration: Secondary compounds are deposited intact in particular sites of the
body. The substance to be sequestered is absorbed through the gut membrane,
transported by the hemolymph and stored in body tissues or special glands (Nishida
2002). Sequestration of plant secondary compounds, in general, is used as a
defense against predators and it has been described in many herbivorous insects
such as the Lepidoptera, Coleptera, Hemiptera and Hymenoptera (Brower 1984,
Bowers 1992, Rothschild 1972). Most of the insects found to sequester toxic
compounds specialize on one or a very few plants, from which they extract their
defensive compounds (Bowers 1992).
Cyanogenic Glycosides in the Family Passifloraceae
The chemistry of Passiflora plants has been studied since the 1960s. Some
Passiflora have been found to contain simple harmane alkaloids (Copeland and Slaytor
1974), flavonoids (McCormick and Mabrick 1981, Ulubelen et al. 1982), a distinctive
triterpene glycoside (Bombardelli et al. 1975), and tannins (Smiley and Wisdom 1985).
However, the Passiflora are unusual in producing cyanogenic glycosides with a
cyclopentene moiety (Spencer and Seigler 1984, 1985a, 1985b, Olafsdottir et al. 1989a,
Jaroszewski et al. 2002).
Chemically, the cyanogenic glycosides in the Passifloraceae have been divided into
four distinct structural classes (Jarozewsky et al. 2002)
a. Type I (called by Spencer 1988, simple cyclopentenoid). These glycosidic
molecules are peculiar because they usually (if not always) occur as pairs of p -D
glucopyranosides having enantiomeric aglycones, i.e., tetraphyllin B co-occurs
with volkenin; epivolkenin co-occurs with taraktophyllin; and tetraphyllin A co
occurs with deidaclin (Figure 4-1 A).
b. Type II (includes the sulfated cyanogens and complex cyclopentenoid (Spencer
1988)). These glycosides are similar in structure to Type I, but they either have an
unusual sugar molecule, sulfate group, or additional oxygenation of the

55
cyclopentene ring. None of the glycosides in Type II co-occur with another
enantiomer, as it occurs in glycosides of Type I. Also, all the glycosides in Type II
have the same stereochemistry in position C-l. The cyanogenic glycoside grouped
within Type I and Type II are believed to be biosynthesized from the amino acid 2-
cyclopentenyl glycine via oxidation, oxidative decarboxylation, dehydration, and/or
glycosylation (Conn 1981b). However, Olafsdottir et al. (1992) do not exclude the
possibility that they originated from oxygenated amino acids (Figure 4-1B).
c. Type III (called aliphatic cyanogenic glycoside by Spencer 1988). These
glycosides do not contain a cyclopentene ring. They are derived from the amino
acids valine and isoleucine. This group includes the glycosides linamarin and
lotaustralin, the cyanogens found most often in nature (Figure 4-1C).
d. Type IV. This group of glycosides include the aromatic cyanogens, for example
amygdalin, which is present in seeds of bitter almonds and prunasin found in
Prunus species. The glycosides within this type originated from the amino acid
phenylalanine (See Figure 4-ID).
No other plant group is as diversified in its types of cyanogenic glycosides as the
genus Passiflora (Spencer 1988, Jaroszewski et al. 2002).
Cyanogenic Glycosides in Lepidoptera
Unpalatable insects can acquire their defensive toxins either by sequestering plant
derived substances or by synthesis de novo (Blum 1981, Rothschild et al. 1970). In
butterflies and moths feeding on cyanogenic plants, both acquisition methods have been
described.
Synthesis De Novo
De novo synthesis has been described in many species of the family Zygaenidae
(Jones et al. 1962, Davis and Nahrstedt 1979), in three subfamilies of the Nymphalidae
(Heliconiinae, Nymphalinae and Acraeinae), and in one subfamily of the Lycaenidae
(Nahrstedt and Davis 1985).
Based on radiolabeled amino acids, Nahrstedt and Davis (1983, 1985)
demonstrated that larvae and adult Heliconius butterflies synthesize the cyanogenic
glycoside themselves, linamarin and lotaustralin from valine and isoleucine, respectively.

56
However, there was an unresolved issue in their studies. The cyanogenic glycoside of
larvae and adults increased proportionally to the administration of labeled amino acids;
however, within the control, the adult butterflies were also able to synthesize these same
cyanogenic glycosides within 14 days after emergence. Nahrstedt and Davis
hypothesized that substrates for cyanogenic glycosides biosynthesis can be derived from
catabolism of tissue proteins in the absence of dietary supplies of amino acids. During
the adult stage, the acquisition of additional amino acids through pollen feeding would
provide relevant precursors for the production of additional linamarin and lotaustralin
(Gilbert 1991).
The biosynthesis of linamarin and lotaustralin from the amino acids valine and
isoleucine in both plants and insects seems to follow a common process. This process is
achieved using a glucosyltransferase enzyme system (Jaroszewski et al. 1988; Olafsdottir
et al. 1992).
Sequestration of Cyanogenic Glycosides
Intake of cyanogenic glycosides from a plant has been described for the larvae of
the African Acraea horta (Nymphalidae: Acraeinae) which sequesters the cyclopentyl
cyanogenic glycoside (gynocardin) from its host plant Kiggelaria africana
(Flacortiaceae) (Raubenheimer 1987). Also, Parnassius spp. (Papilionidae), the cotton
moth Abraxas spp. (Geometridae), Pryeria snica (Zygaenidae), and Yponomeuta
hexabolous (Yponomeutidae) sequester the cyanoglucoside sarmentosin when fed on
their host plants in the families Crassulaceae and Celastraceae (Nishida 1994, Nishida
and Rothschild 1995, Nishida et al. 1994). However, the species in the genus Abraxas
synthesize sarmentosin even if this compound is not present in any of the plants they fed

57
upon. This fact suggests that these species synthesize their own cyanogenic glycoside,
and also sequester sarmentosin (Nishida 1994).
Moreover, it has recently been demonstrated that a cyclopentyl monoglycoside
(epivolkenin) was sequestered by Heliconius sara when it fed upon its host plant P.
auriculata (Engler et al. 2000). This species converts most of this compound into the
corresponding thiol derivative (sarauriculatin), suggesting that the replacement of the
nitrile group by a thiol would prevent cyanide release from the host plant and release
valuable nitrogen into the insects primary metabolism. This finding suggested that a
unique enzymatic mechanism for dealing with plant cyanogenic glycoside exists in this
species (Engler et al. 2000).
Enzymes
Plant cyanogens co-occur with hydrolyzing P-glucosidase enzymes, which are very
specific in the type of compound with which they can co-occur (Robinson 1930, Spencer
1987). The aglycone, together with the sugar part of the substrate, play an important role
in this specificity, although the aglycone seems to be the main determinant of the
specificity (Hosel 1981).
In this chapter, the fate of cyanogenic glycosides from Passiflora plants after
ingestion by Heliconius erato larvae was studied. The chemical compounds found in
Passiflora plant will be discussed first, and then the cyanogens found in larval and adult
samples. The mechanisms used by H. erato and H. himera species to counteract host
ailelochemicals will be described. Finally, since amino acids are the precursors of
cyanogenic glycosides, I determined the essential amino acids composition derived from
larval feeding.

58
Materials and Methods
Butterflies and Plants
This study was conducted at the University of Florida at Gainesville with butterfly
colonies that were maintained in Lord & Burnham glass houses (5x8 m). Each
geographical race was confined in separate cages to avoid hybridization. Wild butterflies
were collected and brought from the field, and stock was replenished with new live
specimens every six to twelve months. During the first two years of experimentation, the
colonies were unable to survive the winters, but improvement in greenhouse
environmental settings made winter breeding possible in the last several years of
research.
Most of the butterfly stocks were raised and propagated in environmental chambers
under controlled temperatures (27C), humidity (78%) and photoperiods (14L:10D).
Adult butterfly colonies were provided with natural host species Passiflora plants
for oviposition, potted Pentas lanceolata and Lantana camara flowers, sugar water and
Psiguria flowers. When Psiguria flowers were not available, a supplemental amino acid
solution was sprayed on Tuffy plastic scrubbing pads, whose orange and yellow color
attracted the butterflies.
Two races of Heliconius erato were used in this experiment: H. e.favorinus from
Peru and PI. e. cyrbia from Ecuador. H. himera, a recently named species which has been
considered a race of erato, was also included in this study (Figure 4-2). Five Passiflora
plants were used in the analysis: P. punctata, P. rubra, P. auriculata, P. trifasciata and P.
biflora.

59
Plant and Butterfly Cyanogen Extraction
Fresh plant (approximately 50 mg) and insect samples (in all cases six individuals
of each sex were used, except for H. e. favorinus fed P. trifasciata in which twelve
individuals of each sex were used) were ground separately in liquid nitrogen. The
powdery ground material was extracted with 80% cold aqueous methanol (HPLC grade,
Fisher Scientific) and allowed to sit in the solvent overnight in the refrigerator. The
extract was then filtered under vacuum through a Bchner funnel, and dried on a rotary
evaporator at 40C. The syrup was partitioned between chloroform and water several
times to remove lipophilic substances.
A 50-pl aliquot of the aqueous cyanogenic fraction was combined with an equal
amount of various P-glucosidase enzymes (see below) in 4.5 x 1.3 cm glass vials. Each
vial was corked with a (5 x 1 cm) freshly prepared Feigl-Anger cyanide test strip (Feigl
and Anger 1966) and checked for cyanogenesis reaction for 24 h. Color changes in the
test strips were ranked as 0, 1,2 or 3, according to the intensity of blue. Exogenous (3-
glucosidase was used when the sample was acyanogenic (blue intensity of 0). The
aqueous layer was then concentrated to a thick syrup on a rotary evaporator and stored in
an ultra freezer at -85C, until further purification could be done.
P-Glucosidase Preparations
Fresh plant material was ground to a fine powder in liquid nitrogen in a pestle and
mortar and extracted with acetone (HPLC grade, Fisher Scientific). The suspension was
filtered with Whatman #1 filter paper under vacuum through a Bchner funnel and rinsed
with acetone until the solid residue had lost its color and had dried in the filter. The dry
residue was resuspended in pH 6.8 phosphate buffer (0.02M, 500 ml) and stirred at 4C
for 1 h and filtered with Whatman #1. The filtrate was dialyzed using a membrane tubing

60
(Spectra/Por Regenerated Cellulose) against pH 6.8 phosphate buffer every 12 h. The
buffer was changed 5 to 6 times until the HCN had been removed, as determined by the
Feigl-Anger test strip. In order to dehydrate and reduce salt content, the final product
was concentrated to 50 ml with Aquacide I (Calbiochem) and the hydrolytic activity was
tested with the appropriate standard cyanogens and the Feigl-Anger test.
Because P-glucosidases are specific to cyanogenic glycoside types, the following P-
glucosidases were prepared as explained above and used for the detection and general
identification of the cyanogens in extracts, fractions of the first purification method and
larval pellets: p-glucosidase from P. coricea was used for identified Type I glycosides,
p-glucosidase from Passiflora biflora to test Type II glycosides, and linamarase from
linseed for identified Type III glycosides. Two commercially available P-glucosidases
were also used: sulfatase and emulsin from almonds (Sigma Chemical Co., St. Louis,
Missouri) to test sulfated cyanogens and Type IV glycosides, respectively. However,
emulsin is not a very specific enzyme as it can react with other cyanogenic glycoside
types (Brimer et al. 1983). For the samples purified by TLC and HPLC, the enzyme P-
glucuronidase Type H-l from Helix pomatia (Sigma Chemical Co., St. Louis, Missouri)
was used because this enzyme is not specific and can detect all cyanogenic glycoside
types.
Purification of Cyanogenic Glycoside from Plants and Butterflies
Purification by liquid chromatography
The thick syrup obtained from the extraction procedure was applied on top of a 3:1
Whatman CF11 and microcrystalline cellulose column (25 cm x 3 cm) packed with
isopropanol-butanol-water (6:3:1). The column was eluted with the same solvent
mixture. The column was connected to a peristaltic pump and a fraction collector.

61
Seventy 8-ml fractions were collected. Fractions were monitored by testing 50-pl of each
fraction with P-glucosidase enzymatic test with the Feigl-Anger strip to test for
cyanogenesis. Appropriate fractions were combined and evaporated in a speed vacuum
(Savant radiant cover PC210B, Savant Instruments Inc. Holbrook, NY)).
Purification by thin layer chromatography
A second purification method, thin layer chromatography (TLC), was then used. In
this case, the right edge of the TLC plate was used as a control and the material to be
purified was loaded on the rest of the plate. Merck silica gel 60 TLC plate (60-200 pm)
was the stationary phase and ethyl acetate-acetone-chloroform-methanol-water
(40:30:12:10:8) constituted the mobile phase. Cyanogenic compounds were monitored
by the sandwich method (Brimer et al. 1983) as follows. After development, the portion
of the plate containing the cyanogenic glycosides was covered with glass, and the control
side of the TLC plate was sprayed with an enzyme solution (aqueous solution of 13-
glucuronidase) obtained from Sigma. A pre-coated Polygram ion exchange sheet
(Polygram ionex 25-SB-Ac) impregnated with three different solutions (a saturated
solution of picric acid in water, a 1 M aqueous sodium carbonate solution and 2 % (w/v)
ethanolic 1 -hexadecanol solution) was used to capture the cyanogenic band overnight.
Location of the cyanogenic spot in the control side of the plate allowed me to scrape the
purified material on the covered side of the plate from the silica gel. The scraped area
was placed in a test tube and the cyanogen deabsorbed with methanol-water mixtures
(50%), vortexed and centrifuged for 20-30 min. The supernatant was collected, filtered
and evaporated in a speed vacuum. Rf values from the control cyanogenic band were
calculated.

62
Purification by high performance liquid chromatography (HPLC)
The dry sample from the TLC purification was diluted in 1 ml 20% methanol in
water. From this mix, 60 pi was further mixed with 40 pi methanol and injected into a
Cl8 reverse phase column (Econosil, 10 pm, 250 x 10 mm. Alltech). The column was
maintained at room temperature with a flow rate of 3.5 ml/min. The HPLC column was
eluted with 20% methanol in water and the separations were monitored with a differential
refractometer connected to an integrator. The individual components were collected and
tested for cyanogenesis (Brimer et al. 1983).
Identification of Cyanogenic Glycosides from Passiflora Plants and Butterflies
Two methods were used for the identification of the cyanogenic glycosides in
plants and butterflies.
Nuclear magnetic resonance (NMR) experimental
All NMR spectra were acquired at the Advanced Magnetic Resonance Imaging and
Spectroscopy (AMRIS) facility in the McKnight Brain Institute of the University of
Florida. Proton ('H) NMR spectra, as well as two-dimensional !H /*H -correlation
(COSY) spectra and 'H/l3C-heteronuclear multiple quantum coherence (HMQC) spectra,
were acquired on Bruker Avance spectrometers equipped with 2.5 mm or 5 mm inverse
detection (TXI) probes operating at 500 or 600 MHz using standard pulse programs and
techniques. Carbon (13C) spectra were acquired with continuous, composite (Waltz-16),
proton decoupling on the Bruker Avance-500 spectrometer equipped with a 5 mm
broadband observe (BBO) probe operating at 125.8 MHz.
NMR analysis of Passiflora rubra. The cyanogen (4.3 mg) was dissolved in
approximately 0.22 ml of methanol-D4, CD3OD (Acros Organics, 99.96% atom %-D)
and placed in a Wilmad 520-1 A, 2.5 mm NMR tube for analysis. The H-NMR spectrum

63
(500.4 MHz) of this sample was recorded at 17C, and the chemical shift axis was
referenced to internal, residual 'HCDaOD which was assigned to 3.3 ppm. Subsequent
addition of tetramethylsilane (TMS) to this solution verified that the assignment of the
residual methanol signal at 3.31ppm was valid. The 2D-COSY and 2D-HMQC spectra
were acquired with the same sample under the same conditions. The UC-NMR spectrum
(125.8 MHz) of the glycoside was recorded at 27C on a larger sample (7.5 mg), obtained
by the addition of more chromatographic material to the original sample, and this
required a larger solution volume of 0.60 ml of CD3OD, as well as a larger 5mm NMR
tube. The chemical shift axis was referenced to the nitrile carbon (CN), which was
assigned 120.4 ppm for comparison to the data of Olafsdottir et al. 1989b.
NMR analysis of Pass ¡flora sp. The cyanogenic glycoside (1.2 mg) was dissolved
in approximately 0.14 ml of methanol-D4, CD3OD (Aldrich, 99.95% atom %-D) and
placed in a Wilmad 520-1A 2.5 mm NMR tube for analysis. The 'H-NMR spectrum
(500.4 MHz) of the glycoside was recorded at 22C, and the chemical shift axis was
referenced to internal, residual 'HCDjOD which was assigned to 3.3 ppm, as above. The
2D-COSY and 2D-HMQC spectra were acquired with the same sample under the same
conditions. The 13C-NMR spectrum (125.8 MHz) of the glycoside was also recorded on
the same sample, but at a temperature of 27C. The chemical shift axis was referenced to
the nitrile carbon (CN), which was assigned 120.5 ppm for comparison to the data of
Olafsdottir et al. 1991.
NMR analysis of Heliconius erato cyrbia fed on Passiflora auriculata. The
partially purified cyanogen (about 2.4 mg) was dissolved in approximately 0.4 ml of
methanol-D4, CD3OD (Acros Organics, 99.96% atom %-D) and placed in a 5 mm NMR

64
tube for analysis. The H-NMR spectrum (600.13 MHz) of the glycoside was originally
recorded at 27C and then subsequently at 17C, in order to move the OH resonance of
the solvent away from the H-4 resonance of the glycoside. The chemical shift axis was
referenced to internal, residual 'HCD20D which was assigned to 3.3 ppm. 2D-COSY
spectra were acquired with the same sample at both 27C and 17C.
HPLC/ (+) ESI-MS experimental
MS: All mass spectrometric data were acquired with a Thermo-Finnigan (San Jose,
CA) LCQ Classic quadrupole ion trap mass spectrometer operated in the electrospray
ionization (ESI) mode, typically under the following conditions: sheath gas (N2) = 60;
aux gas (N2) = 5; spray voltage = 3.3 kV; capillary temperature = 250C; capillary
voltage = 15 V; tube lens offset = 0 V. In order to enhance the sensitivity and obtain
dependent MS" (tandem mass spectrometry) data on different components, the (+) ESI-
normal mass spectra were normally acquired over two ranges: for example, m/z 75-210
for expected amino acids and m/z 205-1200 for higher molecular weight components,
including cyanogenic glycosides. Tandem mass spectrometry (MSn, n=2, 3,..) was
normally performed in a dependent fashion (i.e., under software control) on the most
intense ion from the preceding normal mass spectrum. The MSn spectra were normally
acquired with a parent ion isolation of 4-6 u (u=atomic mass unit) and a normalized %
CID (collision induced dissociation) energy of 35-40%.
HPLC: High performance liquid chromatography (HPLC) was conducted with an
Agilent HP 1100 series binary pump system. Mobile phase A was 0.5% formic acid
(Fisher Scientific, 88%; Certified, A.C.S.) and 5 mM ammonium formate (Fisher
Scientific, Certified) in water (Fisher Scientific, HPLC grade). Mobile phase B was

65
either (a) 0.5% formic acid in methanol (Burdick & Jackson (Muskegon, MI), B&J Brand
High Purity Solvent) or (b) methanol without any modifiers. The column used was a
Phenomenex (Torrance, CA) Synergi Hydro-RP column (2 x 150 mm; 4p; 80 A) with the
equivalent guard column (2x4 mm). A number of different gradients were used to
achieve the best separation of all components of the butterfly extracts. The following
gradient was used at a mobile phase flow of 0.15 ml/min:
A:B(min) = 100:0(0-5) => 60:40(45) => 0:100(75-120) => 100:0(130-165)
UV: During some of the initial analyses, an Applied Biosystems (Foster City, CA)
Model 785A, programmable absorbance detector was interfaced between the column
effluent and the mass spectrometer. The absorbance at 254 nm was monitored. After the
initial analyses, the UV was no longer used and the column effluent was connected
directly to the mass spectrometer.
Post-column mobile phase modification: The efficiency of electrospray
ionization was low when spraying 100% aqueous mobile phase. Often, methanol or 0.5%
formic acid in methanol was added to the column effluent via a PEEK tee-union. The
post-column modifier was either provided by Applied Biosystems model 400 solvent
delivery system at ~ 0.1 ml/min or via the LCQs syringe pump at 20 pl/min. The latter
was found to be more reproducible, as the Applied Biosystems pump did not provide
steady flow at 0.1 ml/min and under.
Amino Acid Analysis
The following standard amino acids were used for the amino acid analysis:
isoleucine, leucine, phenylalanine, tryptophan, tyrosine and valine. The amino acids
were combined with linamarin, which was used as a marker, because the retention time

66
was known. The analysis was performed by HPLC/MS under the same conditions
described above.
Results
Cyanogenic Glycosides in Passiflora Plants
The (3-glucosidases classification test of plant extracts revealed that all Passiflora
species tested possess cyanogenic glycoside Type II, with the exception of P. auriculata,
which possess cyanogenic glycoside Type I. Interestingly, P. rubra reacted with all p-
glucosidases. This means that the aglycone in P. rubra was not specific to any P-
glucosidase. Emulsin was also able to hydrolyze the cyanogenic glycoside from P.
trifasciata and P. auriculata. Cyanogenesis occurred very rapidly in the extracts of P.
rubra with high release of cyanide within 30 minutes of being tested, followed by P.
punctata and P. auriculata (Table 4-1). Cyanogenesis occurred within one hour in P.
biflora and within two hours in P. trifasciata. Cyanide release then increased over time.
Identification of cyclopentanoid cyanogenic glycoside in Passiflora rubra
The chemistry of this plant has not been analyzed before. One peak was isolated
from Passiflora rubra by HPLC and tested for cyanogenesis by TLC (Rf = 0.2-0.37).
The molecular weight was verified by HPLC-MS (MW 417).
'id NMR spectral data of the cyanogen in P. rubra showed chemical shifts and
signal splitting similar to those reported for passicapsin in CD3OD (Table 4-2). I3C NMR
(Table 4-3) also showed characteristic peaks for passicapsin (Olafsdottir et al. 1989b). It
was not clear whether the resonances for H2, H3, C2, and C3 had been assigned in
Olafsdottirs data. Those 'H assignments are given in Table 4-2 by analogy to the data in
Table-4-4, and the coupling constants for these resonances appear consistent, too. The

67
corresponding l3C resonances (Table 4-3) were assigned by way of a 2D-HMQC
spectrum (Figure 4-4).
Identification of cyclopentanoid cyanogenic glycoside in Passiflora sp.
Although this Passiflora species is reported throughout this dissertation as P.
trifasciata, it is actually an undescribed sister species of P. tricuspis and P. trifasciata
(McDougal personal communication).
A peak was isolated from Passiflora sp. by HPLC and tested for cyanogenesis by
TLC (Rf = 0.16-0.24). The molecular weight was verified by HPLC-MS (MW 433). 'H
NMR spectral of this species showed the coupling constants (Hz) and chemical shifts
(ppm) listed in Table 4-4. I3C NMR also showed characteristic peaks for passitrifasciatin
(Table 4-5) (Olafsdottir et al. 1991). The resonances for C2, C3, CL, and Cl" had been
assigned in Table 4-5 by correlations with their corresponding ]H resonances (Table 4-4),
which were assigned by Olafsdottir et al. 1991; this was accomplished using the 2D-
HMQC spectrum (Figure 4-5).
Cyanogen profile of Passiflora auriculata
Two peaks were isolated from P. auriculata by HPLC and tested for
cyanogenesis by TLC, which resulted in a Rf (0.11, 0.2) and Rf2 (0.28, 0.37). The
molecular weight for both was verified by HPLC-MS (MW 287). Fraction #1 was
identified by HPLC/MS as epivolkenin MW=287, eluting at 17.76 min. Fraction #2 was
identified as taraktophyllin MW=287 and eluted at 2.73 min.
Cyanogen profile of Passiflora biflora
P. biflora contained passibiflorin as described by Olafsdottir et al. (1989b). The Rf
values obtained by TLC were 0.16-0.23 and the MW=433.

68
Cyanogen profile of Passiflora punctata
This plant is currently being analyzed by Dr J. Jaroszewsky (personal
communication). However, from the HPLC/MS analysis, we know that this plant
contains passibiflorin with a MW of 433 and a retention time of 34.00 min.
Cyanogenic Glycosides of Butterflies Fed Natural Host Plants
A (3-glucosidase test indicated that there were no differences in cyanogenic
glycoside types between the sexes in either of the two races. All butterflies possessed
aliphatic cyanogenic glycoside, except for H. e. favorinus that tested acyanogenic when
fed on P. rubra and P trifasciata. Based on the results presented in Chapter 3, it is
known that H. e. favorinus contains low amounts of cyanogenic glycosides. Aliphatic
and Type I glycosides were present when males and females from the two races were fed
P. auriculata. Cyanogenesis occurred within 30 minutes in extracts of H. e. favorinus fed
P. punctata and P. auriculata, while cyanogenesis reaction occurred within one hour in
almost all tested butterflies samples (Table 4-6).
No glycosides were found in larval fecal pellets, suggesting that the ingested
glycosides were absorbed by the larvae, and, if excreted, metabolized prior to excretion.
[3-glucosidase classification tests were not able to detect small quantities of
aliphatic and Type II glycosides in butterfly samples, but these were later detected by the
HPLC/MS method. This suggests that the P-glucosidase test is not conclusive and it
should be used only to detect high concentrations of cyanogenic glycosides.
De novo synthesis
HPLC/MS showed that all butterflies contained aliphatic cyanogenic glycosides
(linamarin and lotaustralin), even the ones that tested acyanogenic with the P-glucosidase
enzymatic test (Table 4-6, 4-7 and 4.8). There were variations among the aliphatic

69
compounds found in butterfly samples that were related to host plant feeding. When
butterflies were fed plants containing monoglycoside cyclopentenoids, the percentage of
lotaustralin was very low. Also, H. e. cyrbia and H. e. favorinus have a similar percent
ratio of linamarimlotaustralin, while H. himera generally have a higher ratio of linamarin
to lotaustralin.
Sequestration of cyanogenic glycosides
Butterflies developing from larvae fed P. auriculata sequestered monoglycoside
cyclopentenoids. !H NMR analysis showed that the female H. e. cyrbia contained
epivolkenin (Table 4-9), linamarin, and another monoglycoside (perhaps taraktophyllin,
based on HI and H4). The male sample also contained epivolkenin, linamarin and
another glycoside (perhaps taraktophyllin, based on HI and H4). Sarauriculatin was not
found in the samples. The results of the HPLC/MS (Table 4-7 and 4-8) showed the
percent values of each compound. However, taraktophyllin was not distinguishable from
epivolkenin. under the conditions in which this analysis was conducted.
Small percentages of the cyanogen epivolkenin, passicapsin and MW 433
(passibiflorin or passitrifasciatin) were found when butterflies were fed diglycoside
cyclopentenoids, demonstrating both sequestration and metabolism of this type of
cyanogenic compound (Table 4-7 and 4-8).
The entire butterfly extracts for the cyanogenic and amino acids analysis had a
blank in between each sample in order to avoid carry-over.
Amino Acids Analysis in Butterfly Extracts
The amino acids valine, isoleucine/leucine, tyrosine, phenylalanine, and
tryptophan, were measured in butterfly samples (Table 4-10 and 4-11).

70
All butterflies fed plants containing diglycoside cyclopentenoids (Type II) had
levels of valine greater than 13 %, except for H. e. cyrbia fed P. rubra. Unfortunately,
given the conditions used in the HPLC/MS analysis, leucine and isoleucine could not be
separated. However, the combined percentages of these amino acids were always very
high, especially when H. e. cyrbia was fed P. rubra. A further amino acid analysis will
be run under different conditions in order to separate leucine and isoleucine.
Another interesting result from the amino acid analysis is the low percentage of
valine in H. e. cyrbia fed on auriculata (in both male and female samples), and the
presence of high percentages of tyrosine and tryptophan in the female, but not in the
male. Phenylalanine, on the other hand, was present in the males but not in the female
samples. Also, high percentages of phenylalanine were found in H. e. cyrbia fed biflora,
H. e. favorims fed trifasciata and H. himera fed punctata and rubra.
Thiol Analysis
As mentioned above, sarauriculatin was not found in samples of H. erato fed P.
auriculata. Also, thiol and sulfates were not found in these samples, demonstrating that
these may not be major paths for metabolism of cyanogenic glycoside by H. erato fed P.
auriculata.
Discussion and Conclusions
Sequestration, De Novo Synthesis and Metabolism of Cyanogenic Glycosides
Although sequestration of simple cyclopentenoid cyanogens by H. erato fed P.
auriculata was reported based on a P-glycosidase enzymatic test (Engler et al. 2000), the
compounds were not identified. H. e. cyrbia fed P. auriculata possessed linamarin,
lotaustralin, epivolkenin and, possibly, taraktophylin. This butterfly did not metabolize
sarauriculatin as reported for H. sara by Engler et al. (2000).

71
Also, H. erato and H. himera butterflies sequestered small amounts of complex
cyclopentenoids (Type II) from host plants and contained various products of metabolism
of these cyanogenic glycosides. The presence of epivolkenin can be explained as a
hydrolysis product of both passibiflorin and passicapsin, as shown in Figures 4-6 and 4-7.
Since passibiflorin and passicapsin contain the epivolkenin structure, the two former
compounds can be hydrolyzed to form the latter (Fisher et al. 1982, Olafsdottir et al.
1989b).
Passicapsin was also found in butterflies fed on plants containing passibiflorin and
passitrifasciatin. Perhaps the MW 433 cyanogenic glycoside underwent a loss of an O to
yield this compound as follows: R-OH ====> RH, difference = 16 units. Notice that this
is the only difference between passibiflorin and passicapsin.
Sequestration of secondary compounds has been described in many insects,
including the Lepidoptera. For example, Junonia coenia (Nymphalidae) sequesters
iridoid glycosides from Plantago lanceolata (Bowers and Stamp 1997), Ideopsis similes
(Nymphalidae: Danainae) sequesters phenanthroindolizidine alkaloids from Tilophora
tanakae, an asclepiad plant (Abe et al. 2001), and the lycaenids Eumaeus and Taenaris
spp. take in cycasin from cycads (Bowers and Larin 1989, also see Nishida 2002). On
the other hand, the cases of mixed strategies combining de novo synthesis and
sequestration are scarce, as mentioned in the introduction. In this study, it was found that
when simple and complex cyclopentenoid glycosides are sequestered by the butterflies,
linamarin and lotaustralin were always present. However, lotaustralin was present in
small amounts when the larvae were fed host plant containing simple cyclopentenoids. In
general, a mixed strategy (de novo synthesis and sequestration) may increase

72
distastefulness of Heliconius to predators. This discovery could have important
consequences for the theory of defense mechanisms in these insects.
As was discussed in the introduction, sequestration has been associated with
defense against natural predators (Brower 1984). If an organism both sequesters and
synthesizes its own toxic compound, it is expected that the individual will be better
protected against predation, but only if the quantities of the stored and metabolized
compound are high.
Although metabolic processes similar to those in H. sara may occur in H. erato, I
discard the possiblility of the formation of sarauriculatin because no evidence of thiols
and/or sulfates was found in H. erato fed P. auriculata.
Amino Acids and their Importance For Heliconius Biology
Since none of the butterflies in the experiment was ever fed on Psiguria or other
pollen containing flowers as an adult, I conclude that all of the reported amino acids are
essential amino acids that are generated from larval feeding. The percentage of amino
acids found were related to the larval host plant. For example, butterflies that fed as
larvae on diglycoside cyclopentenoid-containing plants gathered higher percentages of
valine and leucine/isoleucine, while butterflies fed as larvae on plants containing
monoglycoside cyclopentenoids produced higher percentages of leucine/isoleucine and
tyrosine in the males. Nahrstedt and Davis (1985) illustrated the proportional increases
of linamarin and lotaustralin related to their amino acid precursors valine and isoleucine;
however, they were unable to explain how their control butterflies were able to synthesize
linamarin within 14 days after emergence, despite the absence of dietary valine. Here,
we demonstrate that butterflies are potentially able to synthesize linamarin from amino
acids stored in their bodies during their larval development. Furthermore, butterfly

73
ecologists recognize the importance that pollen feeding has for Heliconius butterflies
(Gilbert 1991). Pollen feeding is correlated not only with unpalatability (Cardoso,
umpublished), but also with nutrition for adult somatic maintenance and extended
reproductive longevity (Dunlap-Pianka et al. 1977, Dunlap-Pianka 1979), and nuptial
gifts (Boggs and Gilbert 1979). Thus, in the Heliconius system, fourteen days are spent
searching for Psiguria flowers after emergence from the puparium. Adult butterflies
keep synthesizing the aliphatic glycosides from the amino acids gained during the larval
stage, until they are able to collect new amino acids via pollen feeding. Therefore, the
ability of teneral adults to invest in reproduction and defense is constrained by the
resources that the individual butterfly accumulates as a larva.
Phenylalanine was found in high percentages in adult H. erato fed P. biflora or P.
trifasciata as larvae and in H. himera fed P punctata and P. rubra. Tyrosine, on the
other hand, was found in males of H. e. cyrbia fed P. auriculata as larvae. Tyrosine and
phenylalanine are aromatic amino acids which are important for insect development and
reproduction. Phenylalanine is an essential amino acid for all insects, whereas tyrosine is
synthesized from phenylalanine by hydroxylation (Gilmour 1965). Tyrosine plays an
important role in morphogenetic processes and protein synthesis and, with phenylalanine,
it is involved in the hardening and tanning process of newly formed cuticle (Zografou et
al. 2001). Furthermore, tyrosine and tryptophan are precursors of the pigments
ommochrome and melanin, respectively (Nijhout 1991).
In plants, there is evidence suggesting that cyanogenic compounds may also
function in the metabolism and transport of nitrogen to form proteins (Selmar et al. 1988,
Nahrstedt 1992). For example, cyanide is not released from the seeds of the rubber tree,

74
Hevea brasiliensis, during germination due to the cyanide detoxification enzyme P-
cyanoalanine synthase and hydrase, which cause the transport of nitrogen to be used in
protein synthesis (Selmar et al. 1998). If Heliconius is able to sequester cyanogenic
glycoside, this means that there must be a way to cope with cyanide release. This could
be achieved with an enzymatic mechanism. In H. sara, for example, the cyanogenic
glycoside is metabolized within the insect, by means of an unknown mechanism, and the
nitrogen is recovered and used in protein synthesis (Engler et al. 2000). If metabolism of
sequestered compounds is useful for nitrogen metabolism and reallocation for nutritional
needs, Heliconius is gaining in the coevolutionary race, because butterflies are able to
gain nitrogen for their primary metabolic needs via nutrition from the plant, and can also
use it for defense. A similar mechanism, with other compounds, has been described in
the leaf beetle, Chrysomela confluens (Kearsley and Witman 1992).
This study investigated the basis of the importance of cyanogenic glycosides in
relation to amino acids. Future experiments could proceed in two ways, the first using
radiolabeled amino acids to reveal the pathway of cyanide derived from sequestered
cyanogens in Heliconius butterflies, and the second using HPLC/MS to learn the
contribution of different cyanogenic glycosides towards amino acid composition in the
adult butterfly.

75
Table 4-1. Qualitative analysis of the extracts of Passiflora plants. Number 0=no cyanide
release, l=light blue, slightly cyanogenic, 2=blue, cyanogen and 3=dark blue,
very cyanogenic. .
Passiflora
Less
than 30
min
30min
1 h
2 h
3 h
4 h
5 h
12 h
24 h
auriculata
0
1
1
1
1
2
3
2
2
biflora
0
0
1
1
1
1
2
2
2
punctata
0
1
1
1
2
2
3
2
2
rubra
2
2
2
3
3
3
3
3
3
trifasciata
lo-
0
0
1
1
1
2
2
2

Table 4-2. !H NMR'2 spectrum of a cyanogen, identified as passicapsin, from P. rubra in CD3OD. Chemical shifts (ppm) and
coupling constants (Hz).
Degree of splitting is identified
?y: (d) dou
?let, (q) quartet, (m) multiplet, (br) broad.
Cyanogen
H2
H3
H4
H5A &
H5B
Hr
H2'-H5'
H6'A &
H6'B
HI"
H2"A & H2"B
H3"
H4"
H5"
ch3
Multiplicity
dd
dd
m
dd
d
m,m,m,m
dd, dd
dd
dt, ddd
br, q
dlL
d
Passicapsin(b)
8 ppm
6.12
6.36
4.87
2.34,3.04
4.59
3.33-3.51
3.67
3.86
4.89
1.70
1.83
3.96
3.20
4.00
1.25
J values
1.3,5.5
2.1,5.5
4.7.14.8
7.1.14.8
7.7
c
5.2,12.0
2.0,12.0
2.5,9.5
2.5.13.5
3.2.9.5.13.5
3.5
1.5,6.6
6.6
P. rubra
8 ppm
6.14
6.37
4.88
2.34,3.05
4.58
3.22-3.39
3.67
3.86
4.89
1.71
1.82
3.96
3.20
3.99
1.25
J values
1.0,5.5
2.0,5.5
4.7.14.8
7.1.14.8
7.8
c
5.3,12.0
2.0,12.0
2 ,9.5
2.1,13.6
3.1,9.9,13.3
3.2
1.1,6.6
6.6
(c) Too complex for J-analysis.
On

Cyanogen
Cl
C2
j
C3
C4
C5
CN
Cl'
Cl"
Remaining resonances
Passicapsim
82.1
132.7
142.4
81.7
46.9
120.4
101.0
99.7
17.2, 34.9, 62.9, 70.6 (two), 71.3, 71.7. 74.9, 78.3, 78.4
P. rubra(c)
82.0
132.6
142.3
81.6
46.5
120.4
101.0
99.6
17.2, 34.9, 62.8, 70.6, 70.5, 71.3, 71.6. 74.9, 78.2, 78.4
Ta' 125.8 MHz; see Figure 4-1 for a numbered structure.
ic) Nitrile carbon (CN) assigned 120.4 ppm for comparison to spectrum (b).

Table 4-4. *H NMR(a) spectrum of a cyanogen, identified as passitrifasciatin, from Passijlora sp. (referred throughout this dissertation
as P. trifasciata) in CD3OD. Chemical shifts (ppm) and coupling constants (Hz). Degree of splitting is identified by: (d)
Cyanogen
H2
H3
H4
H5A&H5B
HI'
H2'
H3'- H5'
H6'A&
H6'B
HI"
H2"
H3"
H4"
H5"
CHj
Multiplicity
dd
dd
dd, dd
d
dd
m,m,m
dd, dd
d
m
t
dd
dq
d
Passitrifasciatin(b)
8 ppm
6.16
6.36
4.86
2.49,3.0
4.61
3.21
3.25-3.40
3.66,3.85
4.69
3.25-
3.40
4.0
3.15
3.75
1.25
J values
5.5,1.0
5.5,2.1
15.0,4.0
15.0, 7.0
7.7
7.7,9.0
c
12.1,5.1
12.1,1.9
8.0
c
-
9.5,2.8
9.5,6.3
6.3
Passijlora sp.
8 ppm
6.17
6.36
e
2.50,3.01
4.62
3.22
3.25-3.40
3.67,3.85
4.7
3.25-
3.40
4.01
3.16
3.76
1.26
J values
5.5,1.0
5.5,2.1
"
15.0,4.1
15.0, 7.1
7.8
7.9,9.0
c
12.0,5.1
12.1,1.7
8.0
c
2.9
9.5,2.8
9.5,6.2
6.2
M Too complex for J-analysis.

Table 4-5. nC NMR,a> spectrum of a cyanogen, identified as passitrifasciatin, from Passiflora .sp. (referred throughout this dissertation
as P. trifasciata) in CD3O.
Cyanogen
Cl
C2
C3
C4
C5
CN
Cl'
Cl"
C6'
ch3
Remaining resonances
Passitrifasciatin '
82.1
133
142
82.4
46.2
120.5
100.8
101.5
62.8
18.5
71.0, 71.6, 72.7, 73.0, 74.5, 74.9, 78.2, 78.3
Passiflora spM
82.1
133
142
82.4
46.2
120.5
100.8
101.5
62.8
18.4
71.0, 71.7, 72.7, 73.1, 74.6, 74.9, 78.2, 78.4
(c) Nitrile carbon (CN) assigned 120.5 ppm for comparison to spectrum (b).

80
Table 4-6. Qualitative analysis of cyanide release of whole body extracts tested with
different P-glucosidases by Feigl-Anger strips monitored over time in H. e.
favorinus and H. e. cyrbia. Numbers indicate, 0=no cyanide release, l=light
blue, slightly cyanogenic, 2=blue, cyanogen and 3=dark blue, very
N
30min
1 h
2 h
3 h
4 h
5 h
12 h
24 h
H. e.
favorinus
S punctata
6
0
0
1
1
1
1
0
0
9 punctata
6
1
2
2
2
2
1
0
0
0 rubra
6
0
0
0
0
0
0
0
0
$ rubra
6
0
0
0
0
0
0
0
0
S auriculata
6
0
1
1
2
2
2
1
1
9 auriculata
6
1
1
1
1
1
1
1
1
S trifasciata
6
0
0
0
0
0
0
0
0
2 trifasciata
6
0
0
1
0
0
0
0
0
0 biflora
6
0
0
2
2
2
2
1
1
9 biflora
6
0
0
1
1
1
1
0
0
H. e. cyrbia
S punctata
6
0
3
3
3
3
2
2
1
2 punctata
6
0
3
3
3
3
2
2
1
S rubra
6
0
3
3
3
3
3
2
1
9 rubra
6
0
1
2
2
2
2
2
1
S auriculata
6
0
1
2
3
3
2
1
1
P auriculata
6
0
1
2
2
2
3
2
1
S trifasciata
6
0
1
2
3
3
3
2
1
9 trifasciata
6
0
1
2
3
3
3
2
1
S biflora
6
0
1
2
2
2
2
2
1
9 biflora
6
0
1
2
2
2
2
2
1

Table 4-7. Peak areas counts of cyanogenic glycosides detected by HPLC/MS in H. erato and H. himera butterflies.
Butterfly sample
Peak area of cyanogenic glycosides x 10b
Linamarin
Lotaustralin
Epivolkenin
Passicapsin
MW 433
Total
Cyanogenic
glycoside
9 H. e. cyrbia fed auriculata
307.70
4.59
681.24
0
0
993.54
S H. e. cyrbia fed auriculata
396.37
3.25
650.00
0.10
0
182.91
9 H. e. cyrbia fed punctata
560.98
573.63
4.41
0.67
1.83
1141.51
9 H. e. cyrbia fed rubra
69.62
66.64
12.21
0.49
0.54
149.51
9 H. e. cyrbia fed biflora
766.52
693.98
2.60
0.26
0.23
1463.59
9 H. e. favorinus fed trifasciata
117.53
110.10
12.12
0.74
5.90
246.39
9 H. himera fed punctata
1106.17
536.34
18.85
0.60
0.40
1662.35
9 H. himera fed rubra
271.24
198.94
22.47
0.45
0
493.11
Butterfly sample
~~~ j -
Percent total peak area of cyanogenic glycosides
Linamarin
Lotaustralin
Epivolkenin
Passicapsin
MW 433
Total
Cyanogenic
glycoside
9 H. e. cyrbia fed auriculata
30.97
0.46
68.57
0.00
0.00
100.00
S H. e. cyrbia fed auriculata
37.76
0.31
61.92
0.01
0.00
100.00
9 H e. cyrbia fed punctata
49.14
50.25
0.39
0.06
0.16
100.00
9 H. e. cyrbia fed rubra
46.57
44.57
8.17
0.31
0.36
100.00
9 H. e. cyrbia fed biflora
52.37
47.42
0.18
0.02
0.02
100.00
9 H. e. favor inus fed trifasciata
47.70
44.68
4.92
0.30
2.40
100.00
9 H. himera fed punctata
66.54
32.26
1.13
0.04
0.02
100.00
9/7. himera fed rubra
55.01
40.34
4.56
0.09
0.00
100.00

Table 4-9. 'H NMRU> spectrum in CD3OD at 27C of a cyanogen, identified as epivolkenin, from female H. e. cyrbia fed on P.
auriculata. Chemical shifts (ppm), coupling constants (Hz) and multiplicities. Multiplicity is identified by: (d) doublet,
(m) multiplet.
Cyanogen
H2
H3
H4
H5A
H5B
HI'
H2'
H3'-H5'
H6A'
H6B'
Multiplicity
dd
dd
m
dd
dd
d
dd
m,m,m
dd
dd
Epivolkenin(b)
8 ppm
6.14
6.27
4.80
2.25
3.02
4.62
3.22
3.3-3.4
3.68
3.86
J values
5.5,1.3
5.5,2.0
14.6,4.8
14.6,7.1
7.7
9.1,7.8
c
12,5.5
12,2
H. e. cyrbia
5 ppm
6.13
(6.14)
6.26
(6.26)
. w
(4.81)
2.25
(2.25)
3.01 (3.02)
4.62
(4.62)
3.22
3.3-3.4
(3.85)
J values
5.5,1.3
5.5,2.0
14.6,4.8
14.6,7.1
7.7
9.1,7.8
c
-
12,2
600.13 MHz; see Figure 4-1 for a numbered structure.
(b) From Jaroszewski et al. 1987a, b.
chemical shift at 17 C relative to the proton impurity in methanol-D4 (CD3OD) i.e. 'HCD20D (H) assigned 3.3 ppm.

Table 4-10. Peak areas of measured free amino acids in butterfly samples.
Butterfly sample
Areas of detected amino acids
Valine
Leu/IsoLeu(,)
Tyrosine
Phenylalanine
Tryptophan
Total amino
acids
$ H. e. cyrbia fed auriculata
2079517
39047401
15421636
298840
9648436
66495830
$ H. e. cyrbia fed auriculata
208897
24267233
0
2623200
0
27099330
$ H. e. cyrbia fed punctata
117372033
166137488
7019235
50572236
29184233
370285225
$ H. e. cyrbia fed rubra
8207693
190182200
218414
3928763
0
202537070
$ H. e. cyrbia fed biflora
62722048
266334406
18532600
142270940
8758244
498618238
$//. e. favorinus fed trifasciata
27204475
91750748
2622139
35121022
13495640
170194024
$//. hintera fed punctata
256841130
517108138
7361421
535650216
10175322
1327136227
QH. himera fed rubra
215165754
456333019
27780824
206786183
14364120
920429900
Cannot be distinguished under the given conditions.

Table 4-i 1. Percent total peak areas of measured free amino acids in butterfly samples.
Butterfly sample
% Total peak area of measured amino acids
Valine
Leu/IsoLeu*1
Tyrosine
Phenylalanine
Tryptophan
Total amino
acids
$ H. e. cyrbia fed auriculata
3.13
58.72
23.19
0.45
14.51
100.00
S H. e. cyrbia fed auriculata
0.77
89.55
0.00
9.68
0.00
100.00
$ H. e. cyrbia fed punctata
31.70
44.87
1.90
13.66
7.88
100.00
9 H. e. cyrbia fed rubra
4.05
93.90
0.11
1.94
0.00
100.00
9 H. e. cyrbia fed biflora
12.58
53.41
3.72
28.53
1.76
100.00
9H. e.favorinus fed trifasciata
15.98
53.91
1.54
20.64
7.93
100.00
9H. himera fed punctata
19.35
38.96
0.55
40.36
0.77
100.00
9H. himera fed rubra
23.38
49.58
3.02
22.47
1.56
100.00
Cannot be distinguished under the given conditions.
OO

85
A. TYPE I
.OH HO^
3 4
Tetraphyllin A
Oeidaclin
Tetraphyllin B
(S)
(R)
(1S.4S)
MW 271.1055
MW 271.1055
MW 287.1004
Volkenin
(1R, 4R)
MW 287.1004
Epivolkenin
(1S.4R)
MW 287.1004
Taraktophyllin
(1R.4S)
MW 287.1004
B. TYPE II
p-O-glucopyranosy! /
OH
OH
HO'
\ \ N
Wmi
i-D-glucopyranosyl 6
,4 V
HO'
HO
\
OH
6-deoxy-p-D-gulopyranosyl / 26-dideoxy-IS-D-xy/ohaxoyranosyl /
11 ohH3c V I
I \
V \ /\
OHH3?
\
\
v \ A
\
OH
OH
Passibiflorin (1S.4R)
Molecular formula C18H27N(
Molecular Weight 433.1584
OH
Passicapsin (1S.4R)
Molecular formula C1S H27 N 0
Molecular Weight 417.1635
Passitrifasciatin (1S,4R)
Molecular formula C,8H27NO,,
Molecular Weight 433.1584
C. TYPE III
D. TYPE IV
Linamarin Lotaustralln(R)
MW 247.1 MW 261.1
Glc = p-D-glucopyranosyl, C^H^Oj MW 163
Figure 4-1. Cyanogenic glycosides in Passifloraceae plants (A) Type I, (B) Type II, (C)
Type III and (D) Type IV (Jaroszewski et al. 2002).

86
B
Figure 4-2. Heliconius used in this study (A) H. e. cyrbia, (B) H. e.favorinus, and (C) H.
himera.

87
Figure 4-3. Passiflora plants used in this study (A) P. auriculata, (B) P. punctata, (C) P.
rubra, (D) P. biflora, (E) P. trifasciata.

88
Figure 4-4. Proton/Carbon heteronuclear multiple quantum coherence (HMQC) spectrum
of cyanogen passicapsin in CD3OD from P. rubra.

89
ppm
20
40
60
80
100
120
140
Figure 4-5. Proton/Carbon heteronuclear multiple quantum coherence (HMQC) spectrum
of cyanogen passitrifasciatin in CD3OD from Passiflora sp.

90
OH
6'
H0^V 5 \
mW/, ///
OH
NIW271 5
/
\ J
.OH
6
h0^4\ '5' \ .N
f
/'
\ J
OH
MW 287 5
/ "|
4'
/\
HO H
Figure 4-6. Potential enzymatic hydrolysis products of passibiflorin.

91
/
OH
l-D-glucopyranosyl
Passicapsin (1S.4R)
Molecular formula = C18 H27 N 01C
Molecular Weight = 417.1635
2,6-dideoxy-p-D-xy/o-hexoyranosyl
.N
HO
III
1..
OH H3£
1 \
4" 1.
\,
L
MW 255 y '"'2
* 11
\3
-O
\ ,,a
\ A
\^1
H
///
H0\_
OH
\ -O
I
MW 417
< "1
OH H3C
MW 239 c/ 'jj
I \ \ ,..3
VA /\
I \
OH
OH h3C
I \
V \ /\
V j-tv \| .n h
H
OH
OH
HO^ \ 5'
ho:
OH
H0-"
H\.
OH
MW 271
4.
OH
MW 287
,in
'X
1.,
\ J
4 *
/\
HO H
Figure 4-7. Potential enzymatic hydrolysis products of passicapsin.

CHAPTER 5
STUDIES OF CYANIDE INHIBITION BY P-GLUCOSIDASE FROM HELICONIUS
LARVAL GUTS
In the course of evolutionary time, herbivores have adapted to different degrees to
plant hosts in their struggle to obtain the nutrients required for growth, development,
reproduction and other biological processes (Fraenkel 1959, Feeny 1976, Rhoades and
Cates 1983, Futuyma 1983). Plants, in turn, protect themselves by either physical or
chemical means. Plant defensive compounds are diverse but are not always toxic to all
insects (Schoonhaven 1972, Slansky 1992). Obviously, many phytophagous insects are
able to overcome these otherwise toxic allelochemicals by using different enzymes to
transform them (Brattsten 1992).
For herbivores feeding on plants containing cyanogenic glycosides, two different
detoxification pathways for hydrogen cyanide have been demonstrated (Witthoun and
Naumann 1987). One way involves the thiosulfate sulfur transferase, rhodanase, which
converts hydrogen cyanide to thiocyanate. The other pathway is based on two enzymes,
P-cyanoalanine synthase and p-cyanoalanine hydrase, which convert hydrogen cyanide to
P-cyanoalanine (Conn 1981b, Narshtedt 1988, Jones 1988). Some data suggest that
Heliconius butterflies may have evolved a tolerance to cyanide through the possession of
P-cyanoalanine synthase and p-cyanoalanine hydrase (Conn 1981b, Narshtedt 1988,
Witthoum and Naumann 1987). Some Heliconius may prevent cyanogenesis altogether
in their host plant via larval gut P-glucosidase, which may inhibit Passiflora P-
glucosidase (Spencer 1988). Spencer (1988) observed that crude extracts of larval
92

93
digestive enzymes, which have limited glycosidase activity, prevent expected
cyanogenesis in reaction mixtures of Passiflora cyanogenic glycoside/p-glycosidase.
Unfortunately, no data on these observations have ever been published. However,
Spencer (1984) published interesting data regarding P-glycosidase inhibition, when
combined two P-glucosidases (from plant or commercial origin) and a range of different
cyanogenic substrates. In all cases, cyanide release (measured by the Fiegl-Anger
method) was inhibited. Based on all of these data, Spencer (1987) proposed that P-
glycosidases of insect origin bind with a plant substrate-enzyme complex, either during
or after complex formation, in a competitive manner. The one-substrate, two-enzyme
complex then precipitates from solution (Spencer 1987).
P-glycosidases (E. C. 3.2.1) are hydrolytic enzymes specialized to break glycosidic
bonds to O-, N-, or S- groups on the aglycone. These exoenzymes hydrolyze the terminal
p-linked monosaccharide from the corresponding glycoside (di and/or oligosaccharides).
Depending on the monosaccharide that is removed, the P-glycosidase is named P*
glucosidase (glucose), P-galactosidase (galactose), and so on. Some insects have three or
four digestive P-glycosidases with different substrate specificities. In this case, P-
glucosidase is used as the generic name for all enzymes which remove glucose efficiently
(Terra and Ferreira 1994). Among the P-glycosidase substrates found in insects are toxic
p-glycosides, which are produced by plants and function to avoid or reduce herbivore
attack. Some insects are able to feed on these toxic plants without any apparent harm,
whereas others will show slower development and growth. These differences in
performance may be due to a detoxification mechanism acting after glycoside hydrolysis
or may be based on differential P- glycosidase specificity (Ferreira et al. 1997). However

94
it is unknown whether the specificity is due to the sugar or the aglycone components of
the glycosides.
Insect P-glucosidases may be divided into three classes (Ferreira et al. 1998):
a. Class 1 enzymes with glycosyl P-glucosidase and aryl P-glucosidases activity, e.g.,
the sphingid moth, Erinnyis ello (Santos and Terra 1985).
b. Class 2 enzymes with only glycosyl P-glucosidase activity, particularly active
against disaccharides or oligosaccharides, e.g., the migratory locust, Locusta
migratoria (Morgan 1975).
c. Class 3 enzymes with only aryl (or alkyl) P-glucosidases activity, mainly active
against monosaccharides e.g., p-nitrophenyl p-glucosidase (NPpGluc), p-
nitrophenyl P-galactosidase (NPpGal) and salicin, e.g., Thaumetopoea pytiocampa
(Praviel-Sosa et al. 1987).
This chapter constitutes a preliminary study of the P-glycosidase found in
Heliconius himera. The objective of this chapter was to test whether or not larval P-
glycosidase per se inhibits cyanogenesis. Also, I wanted to document whether inhibition
was specific to one type of cyanogenic glycoside, or common to all types. Finally, I
wished to see whether there was only one or many glycosidases in the larval gut.
Materials and Methods
Study Organism
H. himera (Figure 5-1) is an aposematic butterfly that is distributed over
southwestern Ecuador, northwestern Peru and the Maraon Valley on the eastern side of
the Andes. The status of this species has been controversial for many years. Some
authors considered this species to be a race of Heliconius erato (Eltringham 1916, Lamas
1976, Brown 1979, Sheppard et al. 1985), while others consider this a separate species
(Kaye 1916, Emsley 1965). In the last few years, there has been a tendency to believe
that H. himera is a separate species, based on mitochondria DNA phylogenetic studies
(Brower 1994), which suggest a divergence between H. erato and H. himera

95
approximately 1 million years ago. Also, hybridization studies (Jiggins et at., 1997,
McMillan et al. 1997) suggest that genomic incompatibilities appeared in the earliest
stages of speciation.
Insect Rearing Procedure
Larvae of H. himera were reared under controlled environmental conditions at
28C, 75% humidity and 14L:10D photoperiod. Each individual larva was fed P. biflora.
No artificial diet is available for Heliconius butterfly.
Tissue Preparation
Midguts of 60-75 4th instar larvae were dissected in physiological saline. The
peritrophic membrane was carefully removed without disturbing gut contents and the
interior gut lumen and exterior were thoroughly rinsed with a stream of physiological
saline. The guts were then homogenized in a glass grinder in 500 pi ice cold buffer
consisting of 50mM K2HPO4 and 250 mM sucrose, pH 7.5. The crude homogenate was
centrifuged in 1 ml phosphate sucrose buffer at 7,500g for 15 minutes and the recovered
supernatant was then centrifuged at 32,000g for 90 minutes in an ultracentrifuge
(Beckman L8-70M, SW 60 rotor) cooled to 4C. The supernatant that contains the
cytosolic fraction was separated for further analysis. The microsomal fraction (the pellet)
was re-suspended in 200 pi of the phosphate sucrose buffer. Both cytosolic and
microsomal fractions were then stored at 4C until purification.
Protein Purification
The protein sample was loaded in a high performance Q-Sepharose column
(Amersham Bioscience Corp. Piscataway, New Jersey) equilibrated with 2.5 mM
imidazole, 10% glycerol buffer, pH 7.5 and an increasing salt gradient of 100-1500 mM
NaCl. The flow rate was 30 drops/min and 0.5 ml fractions were collected. Fractions 31-

96
33 had the highest activity (refer to the next section). The fractions were stored
separately at -20C. Aliquots of the eluates from the Q-Sepharose column (fractions 31-
33) were then applied to a column of Mono Q HR 5/5 (Amersham Bioscience Corp.
Piscataway, New Jersey) each fraction separately. The Mono Q column contained a
mobile phase A with the detergent buffer described above and phase B contained 1M
NaCl plus the detergent buffer. The flow rate was 0.5 ml/min, and 0.5 ml fractions were
collected. Fractions 23-25 had the highest activity. The fractions were stored separately
at -20C. In all cases the protein elutes at approximately 400 mM NaCl.
Hydrophobic interaction chromatography with phenyl-sepharose was also
attempted; however, the proteins were not retained in the column.
Enzymatic activity of p-glucosidase
The cytosolic and the microsomal fractions were tested for activity. The synthetic
substrate used to estimate P-glucosidase was 4-nitrophenyl P-D-glucopyranoside in 1 ml
of 0.1 M NaOH citrate buffer (pH 6.0). Upon action of a P-glucosidase, this substrate is
hydrolyzed and forms p-nitrophenol, which turns yellow at alkaline pHs.
A 250-pl quantity of the cytosolic fraction was used to test activity, while 50 pi
was used from the microsomal fraction. The fractions were then incubated at 30C for 2
h. The reaction was halted by immersing the incubation tubes in boiling water for 10
min. For the control, an identical mixture was boiled for 10 min. before the incubation.
All tubes were centrifuged at 10,000 g for 10 min. after incubation. The concentration of
the reaction product p-nitrophenol was determined in a spectrophotometer (Spectronic
Genesys 5, Spectronic Instruments, Rochester, New York) at 400 nm. The molar
extinction coefficient used was 18.13 mM'1. (One unit is defined as the amount of
enzyme hydrolyzing 1 pmol of substrate per min at 30C.)

97
Protein concentration measurements
The protein concentration of the cytosolic and microsomal extracts and fractions
from each purification were determined with the Bio-Rad Bradford protein assay (Bio-
Rad Laboratories. Hercules, CA). A standard curve was prepared from different dilutions
of bovine serum albumin.
In vitro Cyanide Quantification of P-Glycosidase from H. himera Larval Midgut and
Various Cyanogenic Glycosides as Substrates
Amounts of cyanide released for various combinations of plant P-glycosidase, plant
cyanogenic glycosides and larval P-glycosidase enzyme preparations were measured in
vitro, as indicated in Table 5-1. A 0.5-mg quantity of Passiflora species cyanogens was
diluted in 150 pi of distilled water, and for linamarin and amygdalin (Sigma
Laboratories), 0.1 mg of the glycoside was diluted in 150 pi of distilled water.
The reaction mixtures were prepared quickly to avoid cyanide loss. The vials were
then tightly closed with a stopper, and incubated overnight (as described in Chapter 3).
Cyanogenesis was quantified by the Lambert method (described in Chapter 3). A
cyanide standard curve was also calculated from different dilutions of NaCN.
The control for each experimental treatment contained an identical amount of plant
enzyme and cyanogen, but lacked the insect p-glycosidase. Each treatment was
performed twice.
Sodium Dodecyl Sulfate (SDS)-Polyacrylamide Gel Electrophoresis (PAGE)
Samples containing approximately 3 pg protein were combined with 10% (v/v)
glycerol, 50mM TRIS, pH 6.8, 2% SDS, 0.1% bromophenol blue. The samples were
heated for 3 min at 100C in a water bath, before being loaded onto a 10% (w/v)
polyacrylamide gel slab containing 0.1 % SDS (Laemmli 1970). The gel was run at a

98
constant voltage of 175 V and stained for proteins with Coomassie blue. A rainbow
marker (Amersham) was used to calculate molecular weights (Mr).
Protein Identification
A new gel prepared with fresh solvents and reagents containing only fraction 23
from the Mono Q purification was tested, as described above. The resolving peptides in
the gel from this fraction were stained with Coomassie blue and the two prominent bands
were separated by a sterilized scalpel. Each band was placed in a sterilized
microcentrifuge tube with approximately 500 pi of destain and then taken to the Protein
Core Lab at the University of Florida for protein identification. After trypsin digestion
and LC-MS analysis, the SEQUEST database search was performed at the Protein Core
Lab.
Results
In vitro Cyanide Quantification of p-Glycosidase from H. himera Larval Midgut and
Various Cyanogenic Glycosides as Substrates
Larval P-glycosidase alone greatly inhibited cyanide release from all substrates,
except for the Type II glycoside from P. biflora. Also, apart from the Type II and Type
IV glycoside, in most cases the addition of P-glycosidase from larval guts decreased
cyanogenesis in each reaction mixture as compared to the control. The resulting
inhibition of cyanogenesis for Type I glycosides was dramatic (Table 5-2).
P-glucosidase Activity in the Midgut of H. himera
Comparing cytosolic and the microsomal fractions with active P-glycosidase in the
cytosolic fraction showed that the cytosolic fraction was consistently more active
throughout all the steps of purification. The recovery and enrichment of P-glycosidase
increased throughout the purification steps, having its highest enzymatic activity at 36.62

99
nmol/ mg/min protein in the cytosolic fraction, whereas it was only 10.88 nmol/ mg
protein/min in the microsomal fraction. The cytosolic fraction was chosen to use in
sequencing and bioassays.
Protein Purification and Protein Identification
Two different molecular weight P-glucosidases were purified. The molecular
weight (Mr) of the (3-glucosidases obtained by SDS-PAGE were 55,000 and 65,000
(Figure 5-2). Both P-glucosidases obtained from H. himera have some homologies with
the amino acid sequences of (3-glucosidase from a (3-glucosidases in the fall armyworm,
Spodoptera frugiperda Hiibner (Figure 5-3).
Discussion and Conclusions
A good yield was obtained from the purification of the (3-glucosidases of H.
himera, as was proved upon examination of the enzymatic activity and the bands formed
during electrophoresis. Because these proteins have some homologies with the (3-
glucosidase of the fall armyworm, it is possible to say that the purified proteins are
indeed (3-glucosidases.
The (3-glucosidases from larval midgut inhibit the release of cyanide, especially
when the substrate is a monoglycoside cyclopentenoid (Type I). In this case only,
cyanide release dropped drastically, not only when the P-glucosidases from larval midgut
were tested against the substrate, but also with the addition of the plant P-glycosidase.
Because Type I glycosides only possess one sugar (P-D glucopyranoside), it is
hypothesized that at least one of the P-glucosidases is specific to hydrolyze the glycosidic
bond of this cyanogenic glycoside type. However, the possibility that the P-glycosidase
is specific to the aglycone has not been completely discarded. This, in particular, may
explain why butterflies larvae are able to sequester CG Type I (Chapter 4) so efficiently.

100
Since P-glucosidase may have this new function, it is suggested that |3-glucosidases is
likely to be a new inhibitory enzyme for cyanide as Spencer (1987) hypothesized.
Although it looks like this protein did not target Type II glycosides, it will be
necessary to re-test this substrate because of the low cyanide released by the control.
Also, since epivolkenin was a metabolic product found in larvae fed plants containing
passicapsin and passibiflorin (see Figures 4.6 and 4.7) the action of P-glucosidases could
explain conversion of diglycosides to monoglycosides.
Future studies should attempt to separate these P-glucosidases and test them
independently against different cyanogenic glycosides substrates to see their effects on
cyanide release upon hydrolysis.
It is known that plants containing cyanogenic glycosides also contain P-
glycosidases. Cyanogenesis occurs upon herbivore feeding, which causes the release of
cyanide and carbonyl products, particularly alkylating cyclopentenones which are more
toxic than cyanide (Yu 1989). The aglycone can therefore be detrimental to the plant as
much as a defense against herbivores, which must also have enzymes that can detoxify
the aglycone. An example of such a situation occurs in the caterpillars of the tiger
swallowtail that feed upon quaking aspen. After the phenolic aglycone is released, it is
quickly detoxified by an esterase (Lindroth 1988). In this paper it has been demonstrated
that P-glucosidases from H. himera inhibit hydrogen cyanide release from the hydrolysis
of Type I glycosides. Thus, a second enzyme can detoxify cyanide in addition to p-
cyanoalanine synthase and hydrase, which transfer the nitrile group back into the amino
acid cycle (Nahrstedt 1992). The discovery of this novel mechanism of inhibition of
cyanide production is important because it can lead to an effective way of dealing with
t

101
cyanogenic crops that are consumed by humans all around the world. For example, in the
Neotropical region and in Africa cassava consumption is the main source of
carbohydrates in the diet of the inhabitants. Also, important economic plants such as
lima beans, almonds, peanuts, pineapples, barley, sorghum, etc., are cyanogenic (Jones
1998). Ruminants fed on cyanogenic plants showed signs of difficulty in breathing,
tremors, tachycardia, anorexia, and weakness. In some cases, such as buffalos that fed on
Mimosa invisa, the animal died (Vetter 2000). The result of this study on Heliconius
butterflies suggest a potential option for bioengineering herbivores for an enzyme system
that removes cyanide from crop plants with cyanogenic glycosides.

102
Table 5-1. Experimental treatments performed with (3-glycosidase
Experiments 1 and 2 with P. coricea and P. biflora to test Type I and Type II CG
Plant CG (50 pi) + plant (3-glycosidase (50 pi): CONTROL
Plant CG (50 pi) + insect |3-glucosidase (50 pi)
Plant CG (50 pi) + plant [3-glycosidase (50 pi) + insect P-glucosidase (50 pi)
Experiments 3 and 4 with linamarin and amygdalin to test Type III and Type IV CG
CG (50 pi) + [3-glycosidase (50 pi): CONTROL
CG (50 pl)+ [3-glucosidase larval gut (50 pi)
CG (50 pi) + P-glycosidase (50 pi) + [3-glucosidase larval gut (50 pi)
Table 5-2. Cyanide concentration when larval P-glycosidase from Heliconius himera
were tested against different cyanogenic substrates.
CYANOGEN TYPE
pgCN /gof
substrate
Yse
Type I
P. coricea
Plant CG + plant P-glycosidase
Plant CG + larval P-glucosidase
Plant CG + plant P-glycosidase + larval P-glucosidase
60.09 4.92
6.69 0.65
6.49 0.32
Type II
P. biflora
Plant CG + plant P-glycosidase
Plant CG + larval P-glucosidase
Plant CG + plant P-glycosidase + larval P-glucosidase
0.60 0.09
0.81 0.02
0.57 0.02
Type III
Linamarin
CG + plant p-glycosidase
CG + larval P-glucosidase
CG + plant P-glycosidase + larval P-glucosidase
53.82 2.22
3.08 0.10
37.29 1.30
Type IV
Amygdalin
CG + plant P-glycosidase
CG + larval P-glucosidase
CG + plant P-glycosidase + larval P-glucosidase
51.04 10.02
3.28 0.01
53.80 0.01

103
Figure 5-1. Heliconius himera from Vilcabamba, Loja Province, Ecuador.

104
Figure 5-2. Electrophoresis in SDS-10% gel slab. The gel shows different levels of
purification of the protein studied. CE: is the cytosolic crude extract, Q
sepharose elutes: Q1-Q3 fractions 31-33, Mono-Q elutes: Ml-M2, fractions
23 and 24 from Q-sepharose fraction 31, and M3-M4, fractions 24 and 25
from Q-sepharose fraction 32. RM is the rainbow marker.

105
A
1 MKLLVVLSLV AVACNASIVR QQRRFPDDFL FGTATASYQI EGAWDEDGKG ENIWDYMVHN
61 TPEVIRDLSN GDIAADSYHN YKRDVEMMRE LGLDAYRFSL SWARILPTGM ANEVNPAGIA
121 FYNNYIDEML KYNITPLITL YHWDLPQKLQ ELGGFANPLI SDWFEDYARV VFENFGDRVK
181 MFITFNEPRE ICFEGYGSAT KAPILNATAM GAYLCAKNLV TAHAKAYYLY DREFRPVQGG
241 QCGITISVNW FGPATPTPED EMAAELRRQG EWGIYAHPIF SAEGGFPKEL SDKIAEKSAQ
301 QGYPWSRLPE FTEEEKAFVR GTSDLIGVNH YTAFLVSATE RKGPYPVPSL LDDVDTGSWA
361 DDSWLKSASA WLTLAPNSIH TALTHLNNLY NKPVFYITEN GWSTDESREN SLIDDDRIQY
421 YRASMESLLN CLDDGINLKG YMAWSLMDNF EWMEGYIERF GLYEVDFSDP ARTRTPRKAA
481 FVYKHIIKHR VVDYEYEPET MVMTIDEGH
B
1 MKLLVVLSLV AVACNASIVR QQRRFPDDFL FGTATASYQI EGAWDEDGKG ENIWDYMVHN
61 TPEVIRDLSN GDIAADSYHN YKRDVEMMRE LGLDAYRFSL SWARILPTGM ANEVNPAGIA
121 FYNNYIDEML KYNITPLITL YHWDLPQKLQ ELGGFANPLI SDWFEDYARV VFENFGDRVK
181 MFITFNEPRE ICFEGYGSAT KAPILNATAM GAYLCAKNLV TAHAKAYYLY DREFRPVQGG
241 QCGITISVNW FGPATPTPED EMAAELRRQG EWGIYAHPIF SAEGGFPKEL SDKIAEKSAQ
301 QGYPWSRLPE FTEEEKAFVR GTSDLIGVNH YTAFLVSATE RKGPYPVPSL LDDVDTGSWA
361 DDSWLKSASA WLTLAPNSIH TALTHLNNLY NKPVFYITEN GWSTDESREN SLIDDDRIQY
421 YRASMESLLN CLDDGINLKG YMAWSLMDNF EWMEGYIERF GLYEVDFSDP ARTRTPRKAA
481 FVYKHIIKHR VVDYEYEPET MVMTIDEGH
Figure 5-3. Comparison of amino acid sequence of Spodoptera frugiperda (Lepidoptera:
Noctuidae) P-glucosidase with mass spectroscopy identified peptides from H.
himera. The letters in bold indicate the homologous amino acids sequencies
from H. himera that match with S. frugiperda. A. Mr 55,000 and B. Mr
65,000.

CHAPTER 6
CONCLUSIONS
Through their respective coevolutionary races with their host plants, Heliconius
erato and H. himera have developed different strategies to deal with toxic compounds
contained in their cyanogenic host plants in the genus Passiflora. Subspecies of H. erato
from the eastern and the western sides of the Andes differ in body size and this difference
is not due to nutritional differences, but is genetically based. When two subspecies, one
from the eastern side of the Andes (//. e. favorinus) and a second from the western side of
the Andes (//. e. cyrbia), were subjected to feeding on different host plants (natural and
alternative hosts), the larvae showed plastic growth trajectories. Feeding on one of the
host plants, P. rubra, resulted in lower growth rates in both subspecies and higher
mortality in H. e. cyrbia, but not in H. e. favorinus. As a result, the few individuals from
the former subspecies that survived while feeding on this highly toxic cyanogenic plant
were probably better protected against predators because they accumulated higher
concentrations of cyanogenic glycosides in their bodies, a condition which in turn
generates higher concentrations of hydrogen cyanide. H. e. favorinus, on the other hand,
turned out to be less cyanogenic. As a result, individuals of the subspecies that were less
i
cyanogenic were the target of predation by frogs of the genus Hyla. The results of this
study showed that two different geographical races of H. erato utilize different strategies
while feeding on similar host plants.
By studying the chemical composition of cyanogenic plants and butterflies, I
learned that H. erato not only synthesized de novo aliphatic cyanogenic glycosides, but
106

107
also sequestered both simple cyclopentenoid (Type I) efficiently, and small amounts of
complex cyclopentenoid (Type II) from their host plants. In addition, during the larval
stage the butterfly metabolized the complex cyanogenic glycosides into simple
cyclopentenoids which then were stored in the butterfly body. Because H. erato were
able to sequester simple cyclopentenoid glycosides efficiently, these butterflies might
have a way to deal with these toxic compounds, perhaps by an enzymatic mechanism.
Amino acid analysis of the adults showed that these butterflies are not only storing
cyanogenic compounds for defense but also essential amino acids gathered during the
larval stage that are probably used for its primary metabolism, nutrition and possibly for
the further synthesis of aliphatic glycosides during its adult stage.
Finally, a novel enzyme inhibiting cyanide release was discovered which is specific
for simple cyclopentenoid glycosides. The enzyme was purified, tested in vitro, and
isolated for amino acid sequencing. Two P-glucosidases were isolated from larval
midgut. The sequencing of these proteins produced some homologies with the complete
sequence for P-glucosidase of the moth Spodoptera frugiperda (Noctuidae). In vitro
testing of P-glucosidase from the plant, and cyanogenic glycoside and P-glucosidase from
the larval gut showed that the protein was specific to simple cyclopentenoid glycosides
and that it inhibited the production of cyanide. This discovery leads to the conclusion
that this detoxification mechanism is related to the ability of Heliconius to inhibit the
metabolism of different Passiflora cyanogen types. However, more experiments are
warranted to describe the characteristics of the enzymes. The possible genetic
manipulation of this new mechanism of cyanide detoxification through bioengineering
may have important ramifications for agricultural crop improvement.

LIST OF REFERENCES
Abbot, R. J. 1977. A quantitative association between soil moisture content and the
frequency of the cyanogenic forms of Lotus corniculatus L. at Birsay, Orkney.
Heredity 38: 397-400.
Abe, F., T. Yamauchi, K. Honda, H. mura, and N. Hayashi. 2001. Sequestration of
phenanthroindolizidine alkaloids by an Asclepiadaceae-feeding danaid butterfly,
Ideopsis similes. Phytochemistry 56: 697-701.
Benson. W. W., K. S. Brown, and L. E. Gilbert. 1976. Coevolution of plant and
herbivores: passion flower butterflies. Evolution 29(4):659-680.
Bernardo, J. 1993. Determinants of maturation in animals. Trends Ecol. Evolut. 8(5):
166-173.
Blum, M. S. 1981. Chemical Defense of Arthropods. Academic Press, New York. 562 pp.
Blum. M. S. 1983. Detoxification, deactivation, and utilization of plant compounds by
insects, pp. 265-275, in P. Hedin (Ed.) Plant Resistance to Insects. American
Chemical Society. Washington D.C.
Boggs, C. L. 1987. Ecology of nectar and pollen feeding in Lepidoptera, pp. 369-391, in
F. Slansky Jr. and J. G. Rodriguez (Eds.) Nutritional Ecology of Insects, Mites and
Spider, John Wiley and Sons. New York.
Boggs. C. L. 1981. Nutritional and life-history determinants of resource allocation in
holometabolous insects. Am. Nat. 117: 692-709.
Boggs. C. L. 1979. Resource Allocation and Reproductive Strategies in Several
Heliconiine Butterfly Species. Ph. D. Dissertation, University of Texas at Austin.
Boggs, C. L. and L. E. Gilbert. 1979. Male contribution to egg production in butterflies:
evidence for transfer of nutrients at mating. Science 206: 83-84.
Boinbardelli, E., A. Bonati. B. Gabetta, E. Martinelli, G. Mustich, and B. Danieli. 1975.
Passiflorine, a new glycoside from Passiflora edulis. Phytochemistry 14: 2661-
2665.
Bowers, M. D. 1988. Chemistry and coevolution: Iridoid glycosides, plants and
herbivorous insects, pp. 133-165, in K. Spencer (Ed.) Chemical Mediation of
Coevolution. Academic Press, London.
108

109
Bowers, M. D. 1992. The evolution of unpalatability and the cost of chemical defense in
insects, pp. 216-244, in B. D. Roitberg and M. D. Isman (Eds.), Insect Chemical
Ecology: An Evolutionary Approach. Chapman and Hall, New York.
Bowers, M. D. and N. E. Stamp. 1997. Fate of host-plant iridoid glycosides in
lepidopteran larvae of Nymphalidae and Arctiidae. J. Chem. Ecol. 23: 2955-1965.
Bowers, M. D. and Z. Larin. 1989. Acquired chemical defense in the lycaenid butterfly,
Eumaeus atala. J. Chem. Ecol. 15: 1133-1146.
Brako, L. and J. L. Zarucchi. 1993. Catalogue of the flowering plants and gymnosperms
of Peru. Syst. Bot. Missouri Bot. Garcl. 45: I-XI, 1-1286.
Brattsten, L. B. 1979. Biochemical defense mechanisms in herbivores against plant
allelochemicals, pp. 199-270, in G. A. Rosenthal and D. H. Janzen (Eds).
Herbivores, Their Interactions with Secondary Plant Metabolites. Academic Press,
New York.
Brattsten, L. B. 1992. Metabolic defenses against plant allelochemicals, pp. 175-242, in
G.A. Rosenthal and M.R. Berenbaum (Eds.) Herbivores, Their Interactions with
Secondary Plant Metabolites Vol II, Academic Press, San Diego.
Brimer, L., S. B. Christensen, P. Molgaard, and F. Nartey. 1983. Determination of
cyanogenic compounds by thin layer chromatography. 1. A densitometric method
for quantification of cyanogenic glycosides, employing enzymes preparations (P-
glucuronidase) from Helix pomada and picrate-impregnated ion-exchange sheets. J.
Agrie. Food Chem. 31: 789-793.
Brinkler, A. M. and D. S. Seigler. 1989. Methods for the detection and quantitative
determination of cyanide in plant material. Phytochem. Bull. 21: 24-31.
Brinkler. A. M. and D. S. Seigler. 1992. Determination of cyanide and cyanogenic
glycosides from plants, pp. 360-381, in H. F. Linskens and J. F. Jackson (Eds.).
Plant Toxin Analysis. Springler-Verlag, Berlin.
Brower, A. V. Z. 1994. Rapid morphological radiation and convergence among races of
the butterfly Heliconius erato inferred from patterns of mitochondrial DNA
evolution. Proc. Nat. Acad. Sci. 91 (14):6491-6495.
Brower, J. V. Z. 1958. Experimental studies in mimicry in some North American
butterflies I. The monarch, Danaus plexippus, and viceroy, Limenids archippus
archippus. Evolution 12: 32-47.
Brower, L. P. 1984. Chemical defense in butterflies, pp. 109-134, in R. I. Vane-Wright
and P. R. Ackery (Eds.) The Biology of Butterflies, Princeton University Press,
Princeton.

110
Brower, L. P., C. J. Nelson, J. N. Seiber, L.S. Fink, and C. Bond. 1988. Exaptation as an
alternative to coevolution in the cardenolide-based chemical defense of monarch
butterflies (Danaus plexippus L.) against avian predators, pp. 447-475, in K.
Spencer (Ed.). Chemical Mediation of Coevolution. Academic Press. London.
Brower, L. P. and S. C. Glazier. 1975. Localization of heart poisons in the monarch
butterfly. Science 188: 19-25.
Brower, L. P., J. Alcock, and J. V. Z. Brower. 1971. Avian feeding behavior and the
selective advantage of incipient mimicry, pp. 261-274, in E. R. Creed (Ed.)
Ecological Genetics and Evolution, Blackwell Scientific, Oxford.
Brower, L. P., J. V. Z. Brower, and Corvino J. M. 1967. Plant poisons in a terrestrial food
chain. Proc. Nat. Acad. Sci. USA 57: 893-898.
Brower, L. P., J. V. Z. Brower, and C. T. Collins. 1963. Experimental studies in
mimicry. 7. Relative palatability and Mllerian mimicry among Neotropical
butterflies of the subfamily Heliconiinae. Zoolgica (New York) 48: 65-84.
Brown, K. 1981. The biology of Heliconius and related genera Ann. Rev. Entomol. 26:
427-456.
Brown, K. 1979. Ecologa Geogrfica e Evolugo as Florestas Neotropicais (Livre de
Docencia) Campinas, Brazil: Universidade Stadual de Campinas.
Chai, P. 1996. Butterfly visual characteristics and ontogeny of responses to butterflies by
a specialized tropical bird. Biol. J. Linn. Soc. 59: 166-189.
Chai, P. 1990. Relationships between visual characteristics of rainforest butterflies and
responses of a specialized insectivorous bird, pp. 31-60. In M. Wicksten (ed.),
Adaptive Coloration in Invertebrates. Proc. Symp. Spon. Am. Soc. Zool., Texas
A&M University, Galveston, Texas.
Chai, P. 1988. Wing coloration of free-flying Neotropical butterflies as a signal learned
by specialized avian predator. Biotropica 20: 20-30.
Chai, P. 1986. Field observations and feeding experiments on the responses of rufous
tailed jacamars (Galkula ruficauda) to free-flying butterflies in a tropical rainforest.
Biol. J. Linn. Soc. 29: 161-189.
Charlesworth, B. 1980. Evolution in Age Structured Populations. University Press.
Cambridge.
Collins, C. T. and A. Watson. 1983. Field observations of bird predation on Neotropical
moths. Biotropica 15(1): 53-60.

Ill
Conn. E. E. 1979. Cyanide and cyanogenic glycosides, pp. 387-412, in G.A. Rosenthal
and M.R. Berenbaum (Eds.), Herbivores, Their Interactions with Secondary Plant
Metabolites Vol II, Academic Press, San Diego.
Conn, E. E. 1981a. Cyanogenic glycosides, pp. 479-499, in E. E. Conn (Ed.). The
Biochemistry of Plants. A Comprehensive Treatise, Vol 7, Secondary Plant
Products. Academic Press, New York.
Conn, E. E. 1981b. Biosynthesis of cyanogenic glycosides, pp. 183-196, in B.
Vennesland, E. E. Conn, C. Knowles, J. Westley, and F. Wissing (Eds.), Cyanide in
Biology, Academic Press, London.
Cooper-Driver, G. and T. Swain. 1976. Cyanogenic polymorphism in bracken in relation
to herbivore predation. Nature 260: 604.
Cooper-Driver, G., S. Finch, T. Swain, and E. Bemays. 1977. Seasonal variation in
secondary plants compounds in relation to the palatability of Pteridium aquilinum.
Biochem, Syst. Ecol. 5: 177-183.
Copeland, L. and M. Slaytor. 1974. The excretion of a P-carboline alkaloid harmane in
passion fruit. Physiological Plantarum 31:327-329.
Davis, R. H. and A. Nahrstedt. 1985. Cyanogenesis in insects, pp. 635-654, in G.A.
Kerkut, L.I. Gilbert (Ed.), Comprehensive Insect Physiology, Biochemistry and
Pharmacology II. Pharmacology, Pergamon Press. Oxford.
Davis, R. H. and A. Nahrstedt. 1979. Linamarin and lotaustralin as the source of cyanide
in Zygaena filipendulae L. (Lepidoptera). Comp. Biochem. Physiol. 64B: 395-397.
De Wilde, W. J. J. O. 1971. The systematic position of tribe Paropsiaeae, in particular the
genus Ancistrothyrsus, and a key to the genera Passifloraceae. Blumea 19(1): 99-
104.
De Wilde. W. J. J. O. 1974. The genera of tribe Passifloreae (Passifloraceae), with special
reference to flower morphology. Blumea 22: 37-50.
Deinert, E. I.. J. T. Longino, and L. E. Gilbert. 1994. Male competition in butterflies.
Science 370:23-24.
DeVries, P. 1987. Butterflies of Costa Rica and Their Natural History. Princeton
University Press. Princeton.
Dunlap-Pianka, H. L. 1979. Ovarian dynamics in Heliconius butterflies: Correlations
among daily oviposition rates, egg weight, and quantitative aspects of oogenesis. J.
Insect Physiol. 25: 741-749.

112
Dunlap-Pianka H. L., C. L. Boggs, and L. E. Gilbert. 1977. Ovarian dynamics in
heliconiine butterflies: programmed senescence versus eternal youth. Science
197:487-490.
Ehrlich, P. R. and P. H. Raven. 1964. Butterflies and plants: a study of coevolution.
Evolution 18:586-608.
Ellis, W. M., R. J. Keymer, and D. A. Jones. 1977. The effect of temperature on the
polymorphism of cyanogenesis in Lotus corniculatus L. Heredity 38: 339-347.
Eltringham, H. 1916. On specific and mimetic relationships in the genus Heliconius.
Trans. Ent. Soc. Lond. 1916: 101-148.
Emsley, M. G. 1965. Speciation in Heliconius (Lep. Nymphalidae): morphology and
geographic distribution. Zoolgica, New York 50:191-254.
Engler, H. S., K. C. Spencer, and L.E. Gilbert. 2000. Preventing cyanide release from
leaves. Nature 406: 144-145.
f alconer, D. S. 1989. Introduction to Quantitative Genetics. 3rd ed. Longman. London.
Leeny, P. 1992. The evolution of chemical ecology: Contributions from the study of
herbivorous insects, pp. 1-44, in G.A. Rosenthal and M.R. Berenbaum (Eds),
Herbivores, Their Interactions with Secondary Plant Metabolites Vol II, Academic
Press, San Diego.
Feigl, F. and V. Anger. 1966. Replacement of benzidine by copper ethylacetoacetate and
tetra base as spot-test reagent for hydrogen cyanide and cyanogen. Analyst 91:
282-284.
Ferreira, C., B. B. Torres, and W. R. Terra. 1998. Substrate specificities of midgut (3-
glycosidases from different insect orders. Comp. Biochem. Physiol. 119B (1): 219-
225.
Ferreira, C., J., R. P. Parra, and W. R. Terra. 1997. The effect of dietary plant glycosides
on larval midgut P-glucosidases from Spodoptera frugiperda and Diatraea
saccharalis. Insect Biochem. Mol. Biol. 27(1): 55-59.
Fisher, R. A. 1930. The Genetical Theory of Natural Selection. 2nd ed. Dover.
1 isher, F. C., S. Y. Fung, and P. P. Lankhorst. 1982. Cyanogenesis in Passifloraceae.
Cyanogenic glycosides from Passiflora capsularis, P. warmingii and P. perfoliata.
Planta Medica 45: 42-45.
Foulds, W. and J. P. Grime. 1972. The influence of soil moisture on the frequency of
cyanogenic plants in populations of Trifolium repens L. and Lotus corniculatus L.
Heredity 28: 143-146.

113
Fraenkel, G. 1959. The raison detre of secondary plant substances. Science 129:1466-
1470.
Franzl, S. and C. M. Naumann. 1985. Cuticular cavities: storage chambers for
cyanoglucoside-containing defensive secretions in larvae of zygaenid moth. Tissue
and Cell 17(2): 267-278.
Futuyma, D. J. 1983. Evolutionary interactions among herbivorous insects and plants, pp.
207-231, in D. J. Futuyma and M. Slatkin (Eds.). Coevolution. Sinauer Associates.
Massachussets.
Futuyma, D. J. and M. C. Keese. 1992. Evolution and coevolution of plants and
phytophagous arthropods, pp. 439-475, in G. A. Rosenthal and M. R. Berenbaum
(Eds.), Herbivores, Their Interactions With Secondary Plant Metabolites Vol II,
Academic Press, San Diego.
Gilbert, L. E. 1991. Biodiversity of a Central American Heliconius community: Patterns,
process, and problems, pp. 403-427, in P. W. Price, T. M. Lewinsohn. G. W.
Fernandes, and W. W. Benson (Eds.), Plant-Animal Interactions: Evolutionary
Ecology in Tropical and Temperate Regions, John Wiley and Sons, New York.
Gilbert, L. E. 1983. Coevolution and mimicry, pp. 263-285, in D. Futuyma and D. M.
Slatkin (Eds), Coevolution, Sinauer, Sunderland, MA.
Gilbert, L. E. 1976. Post mating female odor in Heliconius butterflies: a male-contributed
antiphrodisiac. Science 193: 419-420.
Gilbert, L. E. 1972. Pollen feeding and the reproductive biology of Heliconius butterflies.
Proc. Nat. Acad Sci. USA. 69: 1403-1407.
Gilbert, L. E. 1971. Butterfly-plant coevolution: Has Passiflora adenopoda won
selectional race with Heliconiinae butterflies? Science 172: 585-586.
Gilmour, D. 1965. The Metabolism of Insects. W. H. Freeman and Company. San
Francisco.
Gleadow. R. M. and I. E. Woodrow. 2002. Constraints on effectiveness of cyanogenic
glycosides in herbivore defense. J. Chem. Ecol. 28(7): 1301-1313.
Haribal, M. and J. A. A. Renwick. 2001. Seasonal and population variation in flavonoid
and alliarinoside content of Alliariapetiolata. J. Chem. Ecol. 27(8): 1585-1594.
Harms, H. 1924. Passifloraceae, pp. 69-94, in A. Engler & K. Prantyl (Eds.), Die
naturlichen Pflanzenfamilien III (6A).
Hay-Roe, M. M. 1996. Growth rate plasticity in two races of Heliconius erato that differ
in body size. M. A. Thesis, University of Texas at Austin.

114
Higgins, L. E. and M. A. Rankin. 1996. Different pathways in arthropod postembryonic
development. Evolution 50(2): 573-582.
Hosel, W. 1981. The enzymatic hydrolysis of cyanogenic glycosides, pp. 217-232, in B.
Vennesland, E. E. Conn, C. Knowles, J. Westley, and F. Wissing (Eds.), Cyanide in
Biology, Academic Press, London.
Hruska, A. J. 1988. Cyanogenic glvcosides as defense compounds. J. Chem. Ecol.
14(12): 2213-2217.
Hughes, M. A. 1981. The genetic control of plant cyanogenesis, pp. 495-508, in B.
Vennesland, E.E. Conn, C. J. Knowles, J. Westly, and F. Wissing (Eds.). Cyanide
in Biology. Academic Press, London.
jaroszewsky, J. W.. E. S. Olafstottir, P. Wellendorph, J. Christensen, H. Franzyk, B.
Somanadhan, B. A. Budnik, L. B. Jorgensen, and V. Clausen. 2002. Cyanohydrin
glycosides of Passiflora: distribution pattern, a saturated cyclopentene derivative
from P. guaiemalensis, and formation of pseudocyanogenic a-hydroxyamides as
isolation artifacts. Phytochemistry 59: 501-511.
jaroszewsky, J. W., P. S. Jenzen, C. Cornett, and J. R. Biberg. 1988. Occurrence of
lotaustralin in Berberidopsis beckleri and its relation to the chemical evolution of
Flacourtiaceae. Biochem. Syst. Ecol. 16: 23-28.
Jaroszewsky, J. W., E. S. Olafstottir, C. Cornett, and K. Schaumburg. 1987a.
Cyanogenesis of Adenia volkensii Harms and Tetrapathaea tetranda Cheeseman
(Passifloraceae) revisited: tetraphyllin B and volkenin. Optical rotary power of
cyclopentenoid cyanohydrin glucosides. Acta Chem. Scand. 41B: 410-421.
Jaroszewsky, J. W., J. V. Anderson, and I. Billeskov. 1987b. Plants as a source of chiral
cyclopentenes: taraktophyllin and epivolkenin, new cyclopentenoid cyanohydrin
glucosides from Flacourtiaceae. Tetrahedron 43: 2349-2354.
Jiggms, C. D., W. O. McMillan, and J. Mallet. 1997. Host plant adaptation has not played
a role in the recent speciation of Heliconius himera and Heliconius erato. Ecol.
Entomol. 22: 361-365.
Jones, D. A. 1998. Why are so many food plants cyanogenic? Phytochemistry 47(2): 155-
162.
Jones, D. A. 1988. Cyanogenesis in animal-plant interactions, pp. 151-176, in D. Evered
and S. Harnett (Eds), Cyanide Compounds in Biology, Ciba Foundation
Symposium, Wiley, Chichester.
Jones, D. A. and A. D. Rammani. 1985. Altruism and movement of plants. Evol. Theor.
7: 143-148.

115
Jones, D. A. 1977. On the polymorphism of cyanogenesis in Lotus corniculatus L. VII.
The distribution of the cyanogenic form in Western Europe. Heredity 39: 27-44.
Jones. D. A. 1971. Chemical defense mechanisms and genetic polymorphism. Science
173: 945.
Jones, D. A., J. Parsons, and M. Rothschild. 1962. Release of hydrocyanic acid from
crushed tissues of all stages in the life-cycle of species of the Zygaeninae
(Lepidoptera). Nature 193(4819): 52-53.
Kaye, W. J. 1916. A reply to Dr. Eltringham's paper on the genus Heliconius. Trans. Ent.
Soc. London. 1916:149-155.
Kassarov, L. 2001. Do cyanogenic glycosides and pyrrolizidine alkaloids provide some
butterflies with a chemical defense against their bird predators? A different point of
view. Behavior 138:45-67.
Kassarov, L. 1999. Are birds able to taste and reject butterflies based on beak mark
tasting? A different point of view. Behavior 136: 965-981.
Kearsley, M. J. C. and T. G. Whitham. 1992. Guns and butter: a no cost defense against
predation for Chrysomela confluens. Oecologia 92: 556-562.
Killip, E. P. 1938. The American species of Passifloraceae. Field Mus. Nat. Hist. Bot.
Ser. 19: 1-613.
Kirkpatrick, K. and D. Lofsvold. 1992. Measuring selection and constraint in the
evolution of growth. Evolution 46(4): 954-971.
Kojima, M., J. E. Poulton, S. S. Thayer, and E. E. Conn. 1979. Tissue distributions of
dhurrin and of enzymes involved in its metabolism in leaves of Sorghum bicolor.
Planta Physiol. 63: 1022-1028.
Lamas, G. 1976. Notes on Peruvian butterflies (Lepidoptera). II New Heliconius from
Cusco and Madre de Dios. Revista Peruana de Entomologa 19: 1-7.
Lambert, J. L.. J. Ramasamy, and J. V. Paukstells. 1975. Stable reagents for the
colorimetric determination of cyanide by modified Konig reactions. Anal. Chem.
47(6): 916-918.
Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of
the bacteriophage T4. Nature 227: 680-685.
l anza, J. 1988. Ant preference for Passiflora nectar mimics that contain amino acids.
Biotropica 20(4): 341-344.
Lindroth. R. L. 1988. Hydrolysis of phenolic glycosides by midgut P-glucosidase in
Papilio glaucus subspecies. Insect Biochem. 18: 789-792.

116
Lindroth, R. L. and A. V. Weisbrod. 1991. Genetic variation in response of the gypsy
moth to aspen phenolic glycosides. Biochem. Syst. Ecol. 19(2): 97-103.
Mallet, J. 1999. Causes and consequences of a lack of coevolution in Mullerian mimicry.
Evol. Ecol. 13: 777-806.
Mallet, J. 1986. Hybrid zones in Heliconius butterflies in Panama, and the stability and
movement of warming colour dines. Heredity 56: 191-202.
McCormick, S. and T. J. Mabry. 1981. Flavonoids of Passiflorapavonis. J. Nat. Prod.
44:623-624.
McKey, D. 1974. Adaptive patterns in alkaloid physiology. Am. Nat. 108:305-320.
McMillan, W. O., C. D. Jiggins, and J. Mallet 1997. What initiates speciation in passion-
vine butterflies? Proc. Nat. Acad. Sci. USA. 94: 8628-8633.
Menna-Barreto, Y. and A. M. Araujo 1985. Evidence for host plant preference in
Heliconius erato phyllis from southern Brazil (Nymphalidae). J. Res. Lepid. 24(1):
41-46.
Mode, C. J. 1958. A mathematical model for the co-evolution of obligate parasites and
their hosts. Evolution 12(2): 158-165.
Morgan, M. R. J. 1975. Relationships between gut cellobiase, lactase, aryl [3-glucosidase
and aryl P-galactosidase activities of Locusta migratoria. Insect Biochem. 5: 609-
617.
Muhtasib, H. and D. L. Evans. 1987. Linamarin and histamine in the defense of adult
Zygaena filipndula. J. Chem. Ecol. 13(1): 133-142.
Nahrstedt, A. 1992. The biology of the cyanogenic glycosides: new developments, pp.
249-269, in K. Mengel and D. J. Pilbeam (Eds.) Nitrogen Metabolism of Plants.
Clarendon Press, Oxford.
Nahrstedt, A. 1988. Cyanogenesis and the role of cyanogenic compounds in insects, pp.
131-150, in D. Evered and S. Harnett (Eds), Cyanide Compounds in Biology, Ciba
Foundation Symposium, Wiley, Chichester.
Nahrstedt, A. 1987. Recent developments in chemistry, distribution and biology of the
cyanogenic glycosides, pp. 213-234, in K. Hostettmann and P. J. Lea (Eds.),
Biochemically Active Natural Products. Clarenton Press. Oxford.
Nahrstedt, A. and R. H. Davis. 1985. Biosynthesis and quantitative relationships of the
cyanogenic glycosides, linamarin and lotaustralin, in genera of the Heliconiini
(Insecta: Lepidoptera). Comp. Biochem. Physiol. 82B: 745-749.

117
Nahrstedt, A. and R. H. Davis. 1983. Occurrence, variation, and biosynthesis of the
cyanogenic glucosides linamarin and lotaustralin in the species of Heliconiini
(Insects: Lepidoptera). Comp. Biochem. Physiol. 75B: 65-73.
Nation, J. L. 2002. Insect Physiology and Biochemistry. CRC Press. Washington D.C.
Nijhout, H. F. 1991. The Development and Evolution of Butterfly Wing Patterns.
Smithsonian Institution Press. Washington, DC.
Nijhout, H. F. 1975. A threshold size for metamorphosis in the tobacco hornworm,
Manduca sexta (L). Biol. Bull. 149: 214-225.
Nijhout, H. F. 1974. Control of moulting and metamorphosis in the tobacco hornworm,
Manduca sexta (L): Growth of the last instar larva and the decision to pupate. J.
Exp. Biol. 61:481-491.
Nishida, R. 2002. Sequestration of defensive substances from plants by Lepidoptera.
Annu. Rev. Entomol. 47: 57-92.
Nishida, R. 1994. Sequestration of plant secondary compounds by butterflies and moths.
Chemoecology 5/6: 127-138.
Nishida, R. and M. Rothschild. 1995. A cyanoglucoside stored by a Sedum-feeding
Apollo butterfly, Parnassius phoebus. Experientia 51: 267-269.
Nishida. R., M. Rothschild, and R. Mummery. 1994. A cyanoglucoside, sarmentosin,
from the magpie moth. Abraxas grossularia, Geometridae: Lepidoptera.
Phytochemistry 36: 37-38.
Nylin, S. 1992. Seasonal plasticity in life history traits growth and development in
Polygonia c-album (Lepidoptera: Nymphalidae). Biol. J. Linn. Soc. 47: 301-323.
Nylin, S., P.O. Wickman, and C. Wiklund. 1989. Seasonal plasticity in growth and
development of the speckled wood butterfly, Pararge aegeria (Satyrinae). Biol. J.
Linn. Soc. 38: 155-171.
Olafsdottir, E. S., L. B. Jorgensen, and J. W. Jaroszewsky. 1992. Substrate specificity in
the biosynthesis of cyclopentanoid cyanohydrin glucosides. Phytochemistry 31:
4129-4134.
Olafsdottir, E. S., J. Jaroszewski, and D. Seigler 1991. Cyanohydrin glycosides with
unusual sugar residues: revised structure of passitrifasciatin. Phytochemistry 30(3):
867-869.
Olafsdottir, E. S., J. V. Andersen, and J. W. Jaroszewski. 1989a. Cyanohydrin glycosides
of Passifloraceae. Phytochemistry 28 (1): 127-132.

118
Olafsdottir, E. S., C. Cornett, and J. W. Jaroszewski. 1989b. Cyclopentenoid cyanohydrin
glycosides with unusual sugar residues. Acta Chem. Scand. 43: 51-55.
Pasteis, J. M., S. Dobler, M. Rowell-Rahier, A. Ehmke, and T. Hartmann. 1995.
Distribution of autogenous and host derived chemical defense in Oreina leaf
beetles (Coleptera: Chrysomelidae) J. Chem. Ecol. 21(8): 1163-1179.
Pasteis, J. M., D. Daloze, W. van Dorsser, and J. Roba. 1979. Cardiac glycosides in the
defensive secretion of Chrysolina herbcea. Identification, biological role and
pharmacological activity. Comp. Biochem. Physiol. 63C: 117-121.
Peterson, S. C., N. D. Johnson, and J. L. LeGuyader. 1987. Defensive regurgitation of
allelochemicals derived from host cyanogenesis by eastern tent caterpillars.
Ecology 68(5): 1268-1272.
Poulton, J. E. 1988. Localizations and catabolism of cyanogenic glycosides, pp. 67-91, in
D. Evered and S. Harnett (Eds.), Cyanide in Biology, Academic Press, London.
Praviel-Sosa, F., S. Clermont, F. Percheron, and C. Chararas. 1987. Studies of
glycosidases and glucanases in Thaumetopoea pytiocampa larvae. II. Purification
and some properties of a broad specific P-D glucosidase. Comp. Biochem. Physiol.
86B: 173-178.
Rammani, A. D. and D. A. Jones. 1985. Flexibility in cyanogenic phenotype of Lotus
corniculatus L. in response to low fluctuating temperatures. Pak. J. Bot. 17(1): 9-
23.
Raubenheimer, D. 1989. Cyanoglycoside gynocardin from Acraea horta (L)
(Lepidoptera: Acraeinae): possible implications for evolution of Acraeine host
choice. J. Chem Ecol. 15(8): 2177-2189.
Rhoades, D. F. and R. G. Cates. 1976. Toward a general theory of plant anti-herbivore
chemistry. Recent Adv. Phytochem. 10: 168-213.
Robinson, M. E. 1930. Cyanogenesis in plants. Biol. Rev. 5: 126-142.
Rothschild, M. 1972. Secondary plant substances and warning colouration in insects.
Symp. Roy. Entomol. Soc. London 6: 59-83.
Rothschild, M. 1984. Aide mmoire mimicry. Ecol. Entomol. 9: 311-319.
Rothschild, M., R. T. Reichstein, J. V. Euw, R. T. Aplin, and R. R. M. Harman. 1970.
Toxic Lepidoptera. Toxicon 8: 293-299.
Santos, C. D. and W. R. Terra. 1985. Physical properties, substrate specificities and a
probable mechanism for a P-D glucosidase (cellobiase) from midgut cells of the
cassava homworm (Erinnys ello). Biochim. Biophys. Acta 831: 179-185.

119
Saunders, J. A., E. E. Conn, H. L. Chin, and C. R. Stocking. 1977. Subcellular location of
the cyanogenic glycoside of Sorghum by autoradiography. Plant Physiol. 59: 647-
652.
Schappert, P. J and J. S. Shore. 2000. Cyanogenesis in Turnera ulmifolia L.
(Tumeraceae): II. Developmental expression, heritability and cost of cyanogenesis.
Evolutionary Ecology Research 2: 337-352.
Schappert, P. J and J. S. Shore. 1995. Cyanogenesis in Turnera ulmifolia L.
(Tumeraceae). I. Phenotypic distribution and genetic variation for cyanogenesis on
Jamaica. Heredity 74: 392-404.
Schoonhaven, L. M. 1972. Secondary plant substances and insects. Rec. Adv. Phytochem.
5: 197-224.
Seigler, D. S. 1991. Cyanide and cyanogenic glycosides, pp. 35-77, in G. A. Rosenthal,
M. R. Berenbaum (Eds.), Herbivores: Their Interactions with Secondary' Plant
Metabolites, Vol. I: The Chemical Participants, Academic Press, San Diego.
Self, L. S., F. E. Guthrie and E. Hodgson. 1964. Metabolism of nicotine by tobacco
feeding insects. Nature 204: 300-302.
Selmar, D., R. Lieberei, and B. Biehl. 1988. Mobilization and utilization of cyanogenic
glycosides. Plant Physiol. 86: 711-716.
Sheppard, P. M., J. R. G. Turner, K. Brown, W. W. Benson, and M. C. Singer. 1985.
Genetic and the evolution of muellerian mimicry in Heliconius butterflies.
Phil.Trans. Roy. Soc. Lond. 308B: 433-613.
Sinervo, B., P. Doughty, R. B. Huey, and K. Zamudio. 1992. Allometric engineering: a
causal analysis of natural selection on offspring size. Science 258: 1927-1930.
Slansky, F. 1992. Allelochemical-nutrient interactions in herbivore nutritional ecology,
pp. 135-174, in G.A. Rosenthal and M. R. Berenbaum (Eds.) Herbivores, Their
interactions with Secondary Plant Metabolites Vol II, Academic Press, San Diego.
Smiley, J. T. and C. S. Wisdom. 1985. Determinants of growth rate on chemical
heterogeneous host plants by specialist insects. Biochem. Syst. Ecol. 13:305-312.
Speed, M. P. and J. R. G. Turner. 1999. Learning and memory in mimicry: II. Do we
understand the mimicry spectrum? Biol. J. Linn. Soc. 67: 281-312.
Spencer, K. C. 1988. Chemical mediation of coevolution in the Passiflora-Heliconius
interaction, pp. 167-240, in K. C. Spencer (Ed.), Chemical Mediation of
Coevolution, Academic Press, New York.
Spencer, K. C. 1987. Specificity of action of allelochemicals: diversification of
glycosides. ACS Symp. Series 330: 275-288.

120
Spencer, K. C. 1984. Cyclopentenoid cyanogens: their chemistry, importance in plant
systematics, and their role in the coevolution of plants and insects. Ph. D. thesis,
University of Illinois, Urbana.
Spencer, K. C. and D. S. Seigler. 1985a. Passicoccin, a novel cyclopentenoid cyanogen
from Passiflora coccnea. Phytochemistry 24: 2615-2617.
Spencer, K. C. and D. S. Seigler. 1985b. Passibiflorin, epipassibiflorin and
passitrifasciatin: Novel cyclopentenoid cyanogen from Passiflora. Phytochemistry
24- 981-986.
Spencer, K. C. and D. S. Seigler. 1984. Isolation and identification of cyclopentenoid
cyanogens. Phytochem. Bull. 16: 13-21.
Srygley, R. B. and P. Chai. 1990. Flight morphology of Neotropical butterflies:
palatability and distribution of mass to the thorax and abdomen. Oecologia 84: 491 -
499.
Steam, S. C. and J. Koella. 1986. The evolution of phenotypic plasticity in life history
traits: predictions for norms of reaction for age and size at maturity. Evolution
40 893-913.
Tattersall, D. B., S. Bak. P. R. Jones, C. E. Olsen, J. K. Nielsen, M. L. Hansen, P. V. Hoj,
and B. L. Moller. 2001. Resistance to an herbivore through engineered cyanogenic
glucoside synthesis. Science 293: 1826-1828.
Terra, W. R. and C. Ferreira. 1994. Insect digestive enzymes: properties,
compartamentalization and function. Comp. Biochem. Physiol. 109B: 1-62.
Trigo, J. R. 2000. The chemistry of antipredator defense by secondary compounds in
neotropical Lepidoptera: Facts, perspectives and caveats. J. Braz. Chem. Soc.
11(6): 551-561.
Ulubelen, A., R. R. Kerr, and T. J. Mabry. 1982. Two new neoflavonoids and C-
glycosylflavones from Passiflora serratodigitata. Phytochemistry 21:1145-1147.
Vetter, J. 2000. Plant cyanogenic glycosides. Toxicon 38: 11-36
Wiklund, C. and A. Kaitala. 1993. Sexual selection for large male size in a polyandrous
butterfly: the effect of body size on male versus female reproductive success in
Pieris napi. Behav. Ecol. 6(1): 6-13.
Williams, K. S. and L. E. Gilbert. 1981. Insects as selective agents on plants vegetative
morphology: Egg mimicry reduces egg laying by butterflies. Science 212: 467-469.
Witthoun, K. and C. M. Naumann. 1987. Cyanogenesis. A general phenomenon in the
Lepidoptera? J. Chem. Ecol. 138: 1789-1809.

121
Yu, S. J. 1989. p-glucosidase in four phytophagous Lepidoptera. Insect Biochem. 19(1):
103-108.
Zar, J. H. 1996. Biostatistical Analysis, 3rd edn. Prentice Hall, Upper Saddle River, New
Jersey.
Zografou, E. N., G. J. Tsiropoulos and L. H. Margaritis. 2001. Effect of phenylalanine
and tyrosine analogues on Bactrocera oleae Gmelin (Dipt., Tephritidae)
reproduction. J. Appl. Entomol. 125: 365-369.

BIOGRAPHICAL SKETCH
Mirian Medina Hay-Roe was born in Lima, Peru, the daughter of Yolanda Polanco
de Medina and Jose Medina Vigil. She completed her high school studies in 1977. She
had a double major in economics and biological science. As an economist she worked
for about three years as a consultant on an international project with the Andean Common
Market and the FAO, Lima. In her spare time she worked as a volunteer at the Museo de
Historia Natural of the Universidad Nacional Mayor de San Marcos, Lima, where she
became involved in butterfly ecology and taxonomy. Shortly thereafter, she decided to
leave economics and dedicate herself to the biological sciences. She conducted research
in the Peruvian rainforest and gained financial support from the Smithsonian Institution
to conduct her own project on butterfly flight stratification in Manu National Park. She
was then accepted at the University of Texas at Austin to do her masters degree, where
she was a teaching assistant in genetics, ecology, general biology, and physiology. After
graduating she taught invertebrate zoology at Southwest Texas State University, San
Marcos, Texas. In 1997, she was accepted at the University of Florida to pursue her
doctoral degree. At the University of Florida, Mirian has taught principles in entomology
and a general biology course in the Zoology Department. She is happily married to Keith
A. Hay-Roe, who holds a doctoral degree in philosophy and works as a computer
programmer at UF. She also has a daughter, who was born during the studies that
became this doctoral dissertation.
122

I certify that I have read this study and that in my opinion it conforms to acceptable
standards of scholarly presentation and is fully adequate, in scope and quality, as a
dissertation for the degree of Doctor of Philosophy.
Thomas C. Emmel, Chair
Professor of Entomology and Nematology
I certify that 1 have read this study and that in my opinion it conforms to acceptable
standards of scholarly presentation and is fully adequate, in scope and quality, as a
dissertation for the degree of Doctor of Phil
tames L. Nation
Professor of Entomology and Nematology
1 certify that I have read this study and that in my opinion it conforms to acceptable
standards of scholarly presentation and is fully adequate, in scope and quality, as a
dissertation for the degree of Doctor of Philosophy.
Heather J. McAuslane
Associate Professor of Entomology and
Nematology
I certify that I have read this study and that in my opinion it conforms to acceptable
standards of scholarly presentation and is fully adequate, in scope and quality, as a
dissertation for the degree of Doctor of Philosophy.
Frank Slansky, Jr.
Professor of Entomology and Nematology
I certify that I have read this study and that in my opinion it conforms to acceptable
standards of scholarly presentation and is fully adequate, in scope and quality, as a
dissertation for the degree of Doctor of Philosophy.
David A. Jones
Professor of Botany

This dissertation was submitted to the Graduate Faculty of the College of
Agricultural and Life Sciences and to the Graduate School and was accepted as partial
fulfillment of the requirements for the degree of Doctor^Ehosophy^/,
May 2004
7- T\ /
Dean, College of Agricultural and Lim'^~~
Sciences
Dean, Graduate School



51
c
o
1000 -1
9
900
k.
c
800
g i
700 -
o w
600
500
t
400 -
ih O)
ss 3
300
-O
c
200 -
(V
o
100 -
s
0 -
y = 0.8726x- 356.67
R2 = 0.6808
H. e. cyrbia
& P. rubra
600 800 1000 1200
Mean plant CN concentration (pg CN/g tissue)
1400
Figure 3-7. Regression of the mean cyanide concentration of H. erato over the mean
cyanide concentration of the natural Passiflora host plants upon which they
fed. (The means were used in the figure to enhance the clarity of
presentation.)


7
maturation phenotypes are achieved by different organisms (Bernardo 1993, Higgins and
Rankin, 1996, Kirkpatrick and Lofsrold 1992, Nylin 1992, Nylin et al. 1989, Steam and
Koella 1986). Thus, age and size at maturity are the results of a growth and
developmental history which reflect both genetic and environmental factors. These
factors act throughout the juvenile developmental time. The relative contribution of
genetic and environmental effects to age and size are likely to be different for individuals
within and between populations and species.
Plastic juvenile development may play an important role in the interaction between
the life cycle and the habitat. The phenotypic value in this context is best explained using
the quantitative genetics formula from Falconer (1989):
P = G + E + M
where P is the phenotypic value; G is the genotypic value; E is the environmental
deviation; and M is the contribution from maternal effects (non-genetic effects).
The latest phylogeny of H. erato (Brower 1994) grouped different subspecies of H.
erato (geographical races) into two main clades. One of the clades grouped races that are
geographically distributed on the eastern side of the Andes (the eastern clade), while the
second grouped races distributed on the western side of the Andes (the western clade).
In a previous study (Hay-Roe 1996), differences in body size between two
subspecies of H. erato were investigated. One subspecies from the western side of the
Andes (H. e. petiveranus) is very specific in its host plant feeding, while a second
subspecies (H. e. phyllis) from the eastern side of the Andes is a broad specialist on
various subgenera of the Passiflora. Differences in body size were genetically based, H.
e. phyllis is bigger than H. e. petiveranus. Differences in growth trajectories between the


101
cyanogenic crops that are consumed by humans all around the world. For example, in the
Neotropical region and in Africa cassava consumption is the main source of
carbohydrates in the diet of the inhabitants. Also, important economic plants such as
lima beans, almonds, peanuts, pineapples, barley, sorghum, etc., are cyanogenic (Jones
1998). Ruminants fed on cyanogenic plants showed signs of difficulty in breathing,
tremors, tachycardia, anorexia, and weakness. In some cases, such as buffalos that fed on
Mimosa invisa, the animal died (Vetter 2000). The result of this study on Heliconius
butterflies suggest a potential option for bioengineering herbivores for an enzyme system
that removes cyanide from crop plants with cyanogenic glycosides.


121
Yu, S. J. 1989. p-glucosidase in four phytophagous Lepidoptera. Insect Biochem. 19(1):
103-108.
Zar, J. H. 1996. Biostatistical Analysis, 3rd edn. Prentice Hall, Upper Saddle River, New
Jersey.
Zografou, E. N., G. J. Tsiropoulos and L. H. Margaritis. 2001. Effect of phenylalanine
and tyrosine analogues on Bactrocera oleae Gmelin (Dipt., Tephritidae)
reproduction. J. Appl. Entomol. 125: 365-369.


70
All butterflies fed plants containing diglycoside cyclopentenoids (Type II) had
levels of valine greater than 13 %, except for H. e. cyrbia fed P. rubra. Unfortunately,
given the conditions used in the HPLC/MS analysis, leucine and isoleucine could not be
separated. However, the combined percentages of these amino acids were always very
high, especially when H. e. cyrbia was fed P. rubra. A further amino acid analysis will
be run under different conditions in order to separate leucine and isoleucine.
Another interesting result from the amino acid analysis is the low percentage of
valine in H. e. cyrbia fed on auriculata (in both male and female samples), and the
presence of high percentages of tyrosine and tryptophan in the female, but not in the
male. Phenylalanine, on the other hand, was present in the males but not in the female
samples. Also, high percentages of phenylalanine were found in H. e. cyrbia fed biflora,
H. e. favorims fed trifasciata and H. himera fed punctata and rubra.
Thiol Analysis
As mentioned above, sarauriculatin was not found in samples of H. erato fed P.
auriculata. Also, thiol and sulfates were not found in these samples, demonstrating that
these may not be major paths for metabolism of cyanogenic glycoside by H. erato fed P.
auriculata.
Discussion and Conclusions
Sequestration, De Novo Synthesis and Metabolism of Cyanogenic Glycosides
Although sequestration of simple cyclopentenoid cyanogens by H. erato fed P.
auriculata was reported based on a P-glycosidase enzymatic test (Engler et al. 2000), the
compounds were not identified. H. e. cyrbia fed P. auriculata possessed linamarin,
lotaustralin, epivolkenin and, possibly, taraktophylin. This butterfly did not metabolize
sarauriculatin as reported for H. sara by Engler et al. (2000).


59
Plant and Butterfly Cyanogen Extraction
Fresh plant (approximately 50 mg) and insect samples (in all cases six individuals
of each sex were used, except for H. e. favorinus fed P. trifasciata in which twelve
individuals of each sex were used) were ground separately in liquid nitrogen. The
powdery ground material was extracted with 80% cold aqueous methanol (HPLC grade,
Fisher Scientific) and allowed to sit in the solvent overnight in the refrigerator. The
extract was then filtered under vacuum through a Bchner funnel, and dried on a rotary
evaporator at 40C. The syrup was partitioned between chloroform and water several
times to remove lipophilic substances.
A 50-pl aliquot of the aqueous cyanogenic fraction was combined with an equal
amount of various P-glucosidase enzymes (see below) in 4.5 x 1.3 cm glass vials. Each
vial was corked with a (5 x 1 cm) freshly prepared Feigl-Anger cyanide test strip (Feigl
and Anger 1966) and checked for cyanogenesis reaction for 24 h. Color changes in the
test strips were ranked as 0, 1,2 or 3, according to the intensity of blue. Exogenous (3-
glucosidase was used when the sample was acyanogenic (blue intensity of 0). The
aqueous layer was then concentrated to a thick syrup on a rotary evaporator and stored in
an ultra freezer at -85C, until further purification could be done.
P-Glucosidase Preparations
Fresh plant material was ground to a fine powder in liquid nitrogen in a pestle and
mortar and extracted with acetone (HPLC grade, Fisher Scientific). The suspension was
filtered with Whatman #1 filter paper under vacuum through a Bchner funnel and rinsed
with acetone until the solid residue had lost its color and had dried in the filter. The dry
residue was resuspended in pH 6.8 phosphate buffer (0.02M, 500 ml) and stirred at 4C
for 1 h and filtered with Whatman #1. The filtrate was dialyzed using a membrane tubing


37
After the analysis, the ground plant or insect tissue was dried in crucibles for
several days in a drying oven at 65C, until they reached a constant dry weight (approx. 6
days). 1 recorded the pg CN /g dry weight of plant tissue or butterfly body tissue.
Predator-Prey Interaction
During the present study I consistently observed the presence of tree-frogs (Hyla
sp.) in the cage containing the colony of H. e. favorinus. In order to learn whether this
vertebrate was preying on the larvae, I collected frogs and confined them in a separate 1
gallon empty aquaria until the frog had excreted. No permit by the Institutional Animal
Care and Use Committee (UF) was required to conduct this experiment.
Statistical Analysis
Two-way analysis of variance (ANOVA) was performed using the statistical
program SYSTAT 10. Some of the data was not normally distributed and the data was
subjected to square root transformation to compensate for increasing variation associated
with larger mean values (Zar 1996). However, the results using transformed data did not
vary significantly from the raw data, so the results of the raw data are reported here. A
Tukey test was used for multiple comparisons.
In order to learn whether cyanide concentration in butterflies is correlated with
cyanide concentration in Passiflora host plants, a linear regression was done on the mean
cyanide concentration of Passiflora host plant with mean cyanide concentration of H.
erato fed natural host plants.
Results
t emporal Variation of Cyanide Release in Passiflora Plants
Cyanide concentrations varied significantly between the two analyzed seasons
(F=l29.95; df=l, 388; p<0.001). In the fall, cyanide concentrations increased to


99
nmol/ mg/min protein in the cytosolic fraction, whereas it was only 10.88 nmol/ mg
protein/min in the microsomal fraction. The cytosolic fraction was chosen to use in
sequencing and bioassays.
Protein Purification and Protein Identification
Two different molecular weight P-glucosidases were purified. The molecular
weight (Mr) of the (3-glucosidases obtained by SDS-PAGE were 55,000 and 65,000
(Figure 5-2). Both P-glucosidases obtained from H. himera have some homologies with
the amino acid sequences of (3-glucosidase from a (3-glucosidases in the fall armyworm,
Spodoptera frugiperda Hiibner (Figure 5-3).
Discussion and Conclusions
A good yield was obtained from the purification of the (3-glucosidases of H.
himera, as was proved upon examination of the enzymatic activity and the bands formed
during electrophoresis. Because these proteins have some homologies with the (3-
glucosidase of the fall armyworm, it is possible to say that the purified proteins are
indeed (3-glucosidases.
The (3-glucosidases from larval midgut inhibit the release of cyanide, especially
when the substrate is a monoglycoside cyclopentenoid (Type I). In this case only,
cyanide release dropped drastically, not only when the P-glucosidases from larval midgut
were tested against the substrate, but also with the addition of the plant P-glycosidase.
Because Type I glycosides only possess one sugar (P-D glucopyranoside), it is
hypothesized that at least one of the P-glucosidases is specific to hydrolyze the glycosidic
bond of this cyanogenic glycoside type. However, the possibility that the P-glycosidase
is specific to the aglycone has not been completely discarded. This, in particular, may
explain why butterflies larvae are able to sequester CG Type I (Chapter 4) so efficiently.


53
cyanogenic glycosides use amino acids as building blocks. These toxic substances take
part in a plants chemical defense against insects that feed upon it.
Cyanogenic glycosides are secondary metabolites that possess intermediate polarity
and are water soluble. Chemically, they are defined as O-P-glycosides of a-
hydroxynitriles (cyanohydrins), biosynthetically derived from amino acids. Cyanogenic
glycosides generally co-occur with P-glycosidases, which are specific to the type of
cyanogenic glycosides and are spatially separated in the plant or animal tissue to avoid
autotoxicity. The enzymatic cleavage that occurs upon disruption of the tissue releases
HCN plus sugar, and ketones or aldehydes. It is thought that the high degree of
specificity of the glycosidase is due to the structure of the aglycone (Hosel 1981).
Subspecies of Heliconius erato (Nymphalidae: Heliconiinae) are aposematic
butterflies that have a tight relationship with Passiflora plants. They specialize in the
group of Passiflora plants of the subgenus Decaloba. Heliconius erato evidently
managed to counteract the effects of the cyanogens in their specific Passiflora host
plants, but their mechanism for overcoming these cyanogens are unknown. (See,
however the new mechanism of detoxification in Heliconius sara described below.)
Toxic plant compounds would set off many reactions in the insects digestive
system. The mechanisms employed by insects to counteract these chemicals vary
between species and can be used solely or in combinations as described below.
Mechanisms Used by Insects to Process Toxic Compounds from Plants
Blum (1983) suggests three ways in which an insect can cope with host allelochemicals:
a. Excretion and egestion: Compounds may be eliminated from the insects body, i.e.,
nicotine is eliminated without any transformation from the body of Trichoplusia ni
and Heliothis virescens (Self et al. 1964).


BIOGRAPHICAL SKETCH
Mirian Medina Hay-Roe was born in Lima, Peru, the daughter of Yolanda Polanco
de Medina and Jose Medina Vigil. She completed her high school studies in 1977. She
had a double major in economics and biological science. As an economist she worked
for about three years as a consultant on an international project with the Andean Common
Market and the FAO, Lima. In her spare time she worked as a volunteer at the Museo de
Historia Natural of the Universidad Nacional Mayor de San Marcos, Lima, where she
became involved in butterfly ecology and taxonomy. Shortly thereafter, she decided to
leave economics and dedicate herself to the biological sciences. She conducted research
in the Peruvian rainforest and gained financial support from the Smithsonian Institution
to conduct her own project on butterfly flight stratification in Manu National Park. She
was then accepted at the University of Texas at Austin to do her masters degree, where
she was a teaching assistant in genetics, ecology, general biology, and physiology. After
graduating she taught invertebrate zoology at Southwest Texas State University, San
Marcos, Texas. In 1997, she was accepted at the University of Florida to pursue her
doctoral degree. At the University of Florida, Mirian has taught principles in entomology
and a general biology course in the Zoology Department. She is happily married to Keith
A. Hay-Roe, who holds a doctoral degree in philosophy and works as a computer
programmer at UF. She also has a daughter, who was born during the studies that
became this doctoral dissertation.
122


4-5. 13C NMR(a) spectrum of a cyanogen, identified as passitrifasciatin from P.
trifasciata in CD3OD
79
4-6. Qualitative analysis of cyanide release of whole body extracts tested with different
P-glucosidases by Feigl-Anger strips monitored over time in H. e.favorinus and
H. e. cyrbia. Numbers indicate, 0=no cyanide release, l=light blue, slightly
cyanogenic, 2=blue, cyanogen and 3=dark blue, very cyanogenic 80
4-7. Peak areas counts of cyanogenic glycosides detected by HPLC/MS in H. erato and
H. himera butterflies 81
4-8. Percent of cyanogenic glycosides detected by HPLC/MS in H. erato and H. himera
butterflies 81
4-9. 'H NMR(a) spectra of cyanogen, epivolkenin, in an extract from female H. e. cyrbia
fed on P. auriculata in CD3OD at 27 C. Chemical shifts (ppm), coupling
constants (Hz) and multiplicities. Multiplicity is identified by: (d) doublet, (m)
multiplet 82
4-10. Peak areas of measured free amino acids in butterfly samples 83
4-11. Percent total peak areas of measured free amino acids in butterfly samples 84
5-1. Experimental treatments performed with P-glycosidase 102
5-2. Cyanide concentration when larval P-glycosidase from Heliconius himera were
tested against different cyanogenic substrates 102
xii


90
OH
6'
H0^V 5 \
mW/, ///
OH
NIW271 5
/
\ J
.OH
6
h0^4\ '5' \ .N
f
/'
\ J
OH
MW 287 5
/ "|
4'
/\
HO H
Figure 4-6. Potential enzymatic hydrolysis products of passibiflorin.


72
distastefulness of Heliconius to predators. This discovery could have important
consequences for the theory of defense mechanisms in these insects.
As was discussed in the introduction, sequestration has been associated with
defense against natural predators (Brower 1984). If an organism both sequesters and
synthesizes its own toxic compound, it is expected that the individual will be better
protected against predation, but only if the quantities of the stored and metabolized
compound are high.
Although metabolic processes similar to those in H. sara may occur in H. erato, I
discard the possiblility of the formation of sarauriculatin because no evidence of thiols
and/or sulfates was found in H. erato fed P. auriculata.
Amino Acids and their Importance For Heliconius Biology
Since none of the butterflies in the experiment was ever fed on Psiguria or other
pollen containing flowers as an adult, I conclude that all of the reported amino acids are
essential amino acids that are generated from larval feeding. The percentage of amino
acids found were related to the larval host plant. For example, butterflies that fed as
larvae on diglycoside cyclopentenoid-containing plants gathered higher percentages of
valine and leucine/isoleucine, while butterflies fed as larvae on plants containing
monoglycoside cyclopentenoids produced higher percentages of leucine/isoleucine and
tyrosine in the males. Nahrstedt and Davis (1985) illustrated the proportional increases
of linamarin and lotaustralin related to their amino acid precursors valine and isoleucine;
however, they were unable to explain how their control butterflies were able to synthesize
linamarin within 14 days after emergence, despite the absence of dietary valine. Here,
we demonstrate that butterflies are potentially able to synthesize linamarin from amino
acids stored in their bodies during their larval development. Furthermore, butterfly


110
Brower, L. P., C. J. Nelson, J. N. Seiber, L.S. Fink, and C. Bond. 1988. Exaptation as an
alternative to coevolution in the cardenolide-based chemical defense of monarch
butterflies (Danaus plexippus L.) against avian predators, pp. 447-475, in K.
Spencer (Ed.). Chemical Mediation of Coevolution. Academic Press. London.
Brower, L. P. and S. C. Glazier. 1975. Localization of heart poisons in the monarch
butterfly. Science 188: 19-25.
Brower, L. P., J. Alcock, and J. V. Z. Brower. 1971. Avian feeding behavior and the
selective advantage of incipient mimicry, pp. 261-274, in E. R. Creed (Ed.)
Ecological Genetics and Evolution, Blackwell Scientific, Oxford.
Brower, L. P., J. V. Z. Brower, and Corvino J. M. 1967. Plant poisons in a terrestrial food
chain. Proc. Nat. Acad. Sci. USA 57: 893-898.
Brower, L. P., J. V. Z. Brower, and C. T. Collins. 1963. Experimental studies in
mimicry. 7. Relative palatability and Mllerian mimicry among Neotropical
butterflies of the subfamily Heliconiinae. Zoolgica (New York) 48: 65-84.
Brown, K. 1981. The biology of Heliconius and related genera Ann. Rev. Entomol. 26:
427-456.
Brown, K. 1979. Ecologa Geogrfica e Evolugo as Florestas Neotropicais (Livre de
Docencia) Campinas, Brazil: Universidade Stadual de Campinas.
Chai, P. 1996. Butterfly visual characteristics and ontogeny of responses to butterflies by
a specialized tropical bird. Biol. J. Linn. Soc. 59: 166-189.
Chai, P. 1990. Relationships between visual characteristics of rainforest butterflies and
responses of a specialized insectivorous bird, pp. 31-60. In M. Wicksten (ed.),
Adaptive Coloration in Invertebrates. Proc. Symp. Spon. Am. Soc. Zool., Texas
A&M University, Galveston, Texas.
Chai, P. 1988. Wing coloration of free-flying Neotropical butterflies as a signal learned
by specialized avian predator. Biotropica 20: 20-30.
Chai, P. 1986. Field observations and feeding experiments on the responses of rufous
tailed jacamars (Galkula ruficauda) to free-flying butterflies in a tropical rainforest.
Biol. J. Linn. Soc. 29: 161-189.
Charlesworth, B. 1980. Evolution in Age Structured Populations. University Press.
Cambridge.
Collins, C. T. and A. Watson. 1983. Field observations of bird predation on Neotropical
moths. Biotropica 15(1): 53-60.


98
constant voltage of 175 V and stained for proteins with Coomassie blue. A rainbow
marker (Amersham) was used to calculate molecular weights (Mr).
Protein Identification
A new gel prepared with fresh solvents and reagents containing only fraction 23
from the Mono Q purification was tested, as described above. The resolving peptides in
the gel from this fraction were stained with Coomassie blue and the two prominent bands
were separated by a sterilized scalpel. Each band was placed in a sterilized
microcentrifuge tube with approximately 500 pi of destain and then taken to the Protein
Core Lab at the University of Florida for protein identification. After trypsin digestion
and LC-MS analysis, the SEQUEST database search was performed at the Protein Core
Lab.
Results
In vitro Cyanide Quantification of p-Glycosidase from H. himera Larval Midgut and
Various Cyanogenic Glycosides as Substrates
Larval P-glycosidase alone greatly inhibited cyanide release from all substrates,
except for the Type II glycoside from P. biflora. Also, apart from the Type II and Type
IV glycoside, in most cases the addition of P-glycosidase from larval guts decreased
cyanogenesis in each reaction mixture as compared to the control. The resulting
inhibition of cyanogenesis for Type I glycosides was dramatic (Table 5-2).
P-glucosidase Activity in the Midgut of H. himera
Comparing cytosolic and the microsomal fractions with active P-glycosidase in the
cytosolic fraction showed that the cytosolic fraction was consistently more active
throughout all the steps of purification. The recovery and enrichment of P-glycosidase
increased throughout the purification steps, having its highest enzymatic activity at 36.62


19
Several factors might be operating during the life history of these subspecies,
creating selective pressures affecting larval growth rates and trajectories (e.g. temporal
differences in host availability, chemical differences among hosts that affect survival
ability, etc.) The results of this study suggest that H. e. favorinus employed different
strategies than H. e. cvrbia did to overcome the intake of different chemicals and/or
nutrients from the host plants. This is suggested by differences in larval performance and
survival. In the following chapters 1 shall explore the importance of plant chemistry as
related to the coevolutionary interaction between H. erato and Passiflora host plants.


40
Cyanide Correlation Between Heliconius erato Fed Natural Host Plants
The regression line in Figure 3.7 indicates that 68% of the variation in cyanide
concentration in Heliconius butterflies is due to differences in the cyanide concentration
in the host plants.
Predator-Prey Interaction
At least 10 frogs were collected within the captive H. e.favorinus butterflies
colonies and the feces was analyzed during this study. I found, on average, 4 head
capsules and many spines of H. erato in each frogs excreta. The Heliconius head
capsules corresponded to the third and fourth instar of development. Usually, I found the
frogs camouflaged on the main stem of the plant or on leaves. Occasionally, I found
frogs on the cage walls. Frog feces also contained body parts of other insects such as
cockroach legs and ant heads.
Discussion and Conclusions
Cyanide Concentrations in Heliconius erato fed Passifiora host plants
Host plant selection was the main reason for the variation in cyanide concentrations
between the different races of Heliconius erato. In natural habitats, resource availability
and competitive interactions are among the factors that play important roles in this
variation. Within each race, genetic variation is probably responsible for differences in
the ability to sequester cyanogenic glycosides.
The results of the analysis of the different body parts of H. erato are related to the
allocation of the defensive compounds in the butterflies and the mode of action of the
chemicals involved in defense. The thorax and the head were the body parts with higher
cyanide concentrations. These results did not follow the patterns of cardenolide toxicity
reported for monarch butterflies (Brower and Glazier 1975). Since the cyanogenic


62
Purification by high performance liquid chromatography (HPLC)
The dry sample from the TLC purification was diluted in 1 ml 20% methanol in
water. From this mix, 60 pi was further mixed with 40 pi methanol and injected into a
Cl8 reverse phase column (Econosil, 10 pm, 250 x 10 mm. Alltech). The column was
maintained at room temperature with a flow rate of 3.5 ml/min. The HPLC column was
eluted with 20% methanol in water and the separations were monitored with a differential
refractometer connected to an integrator. The individual components were collected and
tested for cyanogenesis (Brimer et al. 1983).
Identification of Cyanogenic Glycosides from Passiflora Plants and Butterflies
Two methods were used for the identification of the cyanogenic glycosides in
plants and butterflies.
Nuclear magnetic resonance (NMR) experimental
All NMR spectra were acquired at the Advanced Magnetic Resonance Imaging and
Spectroscopy (AMRIS) facility in the McKnight Brain Institute of the University of
Florida. Proton ('H) NMR spectra, as well as two-dimensional !H /*H -correlation
(COSY) spectra and 'H/l3C-heteronuclear multiple quantum coherence (HMQC) spectra,
were acquired on Bruker Avance spectrometers equipped with 2.5 mm or 5 mm inverse
detection (TXI) probes operating at 500 or 600 MHz using standard pulse programs and
techniques. Carbon (13C) spectra were acquired with continuous, composite (Waltz-16),
proton decoupling on the Bruker Avance-500 spectrometer equipped with a 5 mm
broadband observe (BBO) probe operating at 125.8 MHz.
NMR analysis of Passiflora rubra. The cyanogen (4.3 mg) was dissolved in
approximately 0.22 ml of methanol-D4, CD3OD (Acros Organics, 99.96% atom %-D)
and placed in a Wilmad 520-1 A, 2.5 mm NMR tube for analysis. The H-NMR spectrum


Results 66
Cyanogenic Glycosides in Passiflora Plants 66
Identification of cyclopentanoid cyanogenic in Passiflora rubra 66
Identification of cyclopententanoid cyanogenic in Passiflora trifasciata ...67
Cyanogen profile of Passiflora auriculata 67
Cyanogen profile of Passiflora biflora 67
Cyanogen profile of Passiflora punctata 68
Cyanogenic Glycosides of Butterflies Fed Natural Host Plants 68
De novo synthesis 68
Sequestration of cyanogenic glycosides 69
Amino Acids Analysis in Butterfly Extracts 69
Thiol Analysis 70
Discussion and Conclusions 70
Sequestration. De Novo Synthesis and Metabolism of Cyanogenic
Glycosides 70
Amino Acids and their Importance For Heliconius Biology 72
5 STUDIES OF CYANIDE INHIBITION BY p-GLUCOSIDASE FROM
HELICONIUS LARVAL GUTS 92
Materials and Methods 94
Study Organism 94
Insect Rearing Procedure 95
Tissue Preparation 95
Protein Purification 95
Enzymatic activity of P-glucosidase 96
Protein concentration measurements 97
In vitro Cyanide Quantification of P-Glycosidase from H. himera Larval
Midgut and Various Cyanogenic Glycosides as Substrates 97
Sodium Dodecyl Sulfate (Sds)-Polyacrylamide Gel Electrophoresis (PAGE)...97
Protein Identification 98
Results 98
In vitro Cyanide Quantification of P-Glycosidase from H. himera Larval
Midgut and Various Cyanogenic Glycosides as Substrates 98
P-glucosidase Activity in the Midgut of H. himera 98
Protein Purification and Protein Identification 99
Discussion and Conclusions 99
6 CONCLUSIONS 106
LIST OF REFERENCES 108
BIOGRAPHICAL SKETCH 122
x


This dissertation was submitted to the Graduate Faculty of the College of
Agricultural and Life Sciences and to the Graduate School and was accepted as partial
fulfillment of the requirements for the degree of Doctor^Ehosophy^/,
May 2004
7- T\ /
Dean, College of Agricultural and Lim'^~~
Sciences
Dean, Graduate School


109
Bowers, M. D. 1992. The evolution of unpalatability and the cost of chemical defense in
insects, pp. 216-244, in B. D. Roitberg and M. D. Isman (Eds.), Insect Chemical
Ecology: An Evolutionary Approach. Chapman and Hall, New York.
Bowers, M. D. and N. E. Stamp. 1997. Fate of host-plant iridoid glycosides in
lepidopteran larvae of Nymphalidae and Arctiidae. J. Chem. Ecol. 23: 2955-1965.
Bowers, M. D. and Z. Larin. 1989. Acquired chemical defense in the lycaenid butterfly,
Eumaeus atala. J. Chem. Ecol. 15: 1133-1146.
Brako, L. and J. L. Zarucchi. 1993. Catalogue of the flowering plants and gymnosperms
of Peru. Syst. Bot. Missouri Bot. Garcl. 45: I-XI, 1-1286.
Brattsten, L. B. 1979. Biochemical defense mechanisms in herbivores against plant
allelochemicals, pp. 199-270, in G. A. Rosenthal and D. H. Janzen (Eds).
Herbivores, Their Interactions with Secondary Plant Metabolites. Academic Press,
New York.
Brattsten, L. B. 1992. Metabolic defenses against plant allelochemicals, pp. 175-242, in
G.A. Rosenthal and M.R. Berenbaum (Eds.) Herbivores, Their Interactions with
Secondary Plant Metabolites Vol II, Academic Press, San Diego.
Brimer, L., S. B. Christensen, P. Molgaard, and F. Nartey. 1983. Determination of
cyanogenic compounds by thin layer chromatography. 1. A densitometric method
for quantification of cyanogenic glycosides, employing enzymes preparations (P-
glucuronidase) from Helix pomada and picrate-impregnated ion-exchange sheets. J.
Agrie. Food Chem. 31: 789-793.
Brinkler, A. M. and D. S. Seigler. 1989. Methods for the detection and quantitative
determination of cyanide in plant material. Phytochem. Bull. 21: 24-31.
Brinkler. A. M. and D. S. Seigler. 1992. Determination of cyanide and cyanogenic
glycosides from plants, pp. 360-381, in H. F. Linskens and J. F. Jackson (Eds.).
Plant Toxin Analysis. Springler-Verlag, Berlin.
Brower, A. V. Z. 1994. Rapid morphological radiation and convergence among races of
the butterfly Heliconius erato inferred from patterns of mitochondrial DNA
evolution. Proc. Nat. Acad. Sci. 91 (14):6491-6495.
Brower, J. V. Z. 1958. Experimental studies in mimicry in some North American
butterflies I. The monarch, Danaus plexippus, and viceroy, Limenids archippus
archippus. Evolution 12: 32-47.
Brower, L. P. 1984. Chemical defense in butterflies, pp. 109-134, in R. I. Vane-Wright
and P. R. Ackery (Eds.) The Biology of Butterflies, Princeton University Press,
Princeton.


Table 4-10. Peak areas of measured free amino acids in butterfly samples.
Butterfly sample
Areas of detected amino acids
Valine
Leu/IsoLeu(,)
Tyrosine
Phenylalanine
Tryptophan
Total amino
acids
$ H. e. cyrbia fed auriculata
2079517
39047401
15421636
298840
9648436
66495830
$ H. e. cyrbia fed auriculata
208897
24267233
0
2623200
0
27099330
$ H. e. cyrbia fed punctata
117372033
166137488
7019235
50572236
29184233
370285225
$ H. e. cyrbia fed rubra
8207693
190182200
218414
3928763
0
202537070
$ H. e. cyrbia fed biflora
62722048
266334406
18532600
142270940
8758244
498618238
$//. e. favorinus fed trifasciata
27204475
91750748
2622139
35121022
13495640
170194024
$//. hintera fed punctata
256841130
517108138
7361421
535650216
10175322
1327136227
QH. himera fed rubra
215165754
456333019
27780824
206786183
14364120
920429900
Cannot be distinguished under the given conditions.


Table 4-5. nC NMR,a> spectrum of a cyanogen, identified as passitrifasciatin, from Passiflora .sp. (referred throughout this dissertation
as P. trifasciata) in CD3O.
Cyanogen
Cl
C2
C3
C4
C5
CN
Cl'
Cl"
C6'
ch3
Remaining resonances
Passitrifasciatin '
82.1
133
142
82.4
46.2
120.5
100.8
101.5
62.8
18.5
71.0, 71.6, 72.7, 73.0, 74.5, 74.9, 78.2, 78.3
Passiflora spM
82.1
133
142
82.4
46.2
120.5
100.8
101.5
62.8
18.4
71.0, 71.7, 72.7, 73.1, 74.6, 74.9, 78.2, 78.4
(c) Nitrile carbon (CN) assigned 120.5 ppm for comparison to spectrum (b).


71
Also, H. erato and H. himera butterflies sequestered small amounts of complex
cyclopentenoids (Type II) from host plants and contained various products of metabolism
of these cyanogenic glycosides. The presence of epivolkenin can be explained as a
hydrolysis product of both passibiflorin and passicapsin, as shown in Figures 4-6 and 4-7.
Since passibiflorin and passicapsin contain the epivolkenin structure, the two former
compounds can be hydrolyzed to form the latter (Fisher et al. 1982, Olafsdottir et al.
1989b).
Passicapsin was also found in butterflies fed on plants containing passibiflorin and
passitrifasciatin. Perhaps the MW 433 cyanogenic glycoside underwent a loss of an O to
yield this compound as follows: R-OH ====> RH, difference = 16 units. Notice that this
is the only difference between passibiflorin and passicapsin.
Sequestration of secondary compounds has been described in many insects,
including the Lepidoptera. For example, Junonia coenia (Nymphalidae) sequesters
iridoid glycosides from Plantago lanceolata (Bowers and Stamp 1997), Ideopsis similes
(Nymphalidae: Danainae) sequesters phenanthroindolizidine alkaloids from Tilophora
tanakae, an asclepiad plant (Abe et al. 2001), and the lycaenids Eumaeus and Taenaris
spp. take in cycasin from cycads (Bowers and Larin 1989, also see Nishida 2002). On
the other hand, the cases of mixed strategies combining de novo synthesis and
sequestration are scarce, as mentioned in the introduction. In this study, it was found that
when simple and complex cyclopentenoid glycosides are sequestered by the butterflies,
linamarin and lotaustralin were always present. However, lotaustralin was present in
small amounts when the larvae were fed host plant containing simple cyclopentenoids. In
general, a mixed strategy (de novo synthesis and sequestration) may increase


Ill
Conn. E. E. 1979. Cyanide and cyanogenic glycosides, pp. 387-412, in G.A. Rosenthal
and M.R. Berenbaum (Eds.), Herbivores, Their Interactions with Secondary Plant
Metabolites Vol II, Academic Press, San Diego.
Conn, E. E. 1981a. Cyanogenic glycosides, pp. 479-499, in E. E. Conn (Ed.). The
Biochemistry of Plants. A Comprehensive Treatise, Vol 7, Secondary Plant
Products. Academic Press, New York.
Conn, E. E. 1981b. Biosynthesis of cyanogenic glycosides, pp. 183-196, in B.
Vennesland, E. E. Conn, C. Knowles, J. Westley, and F. Wissing (Eds.), Cyanide in
Biology, Academic Press, London.
Cooper-Driver, G. and T. Swain. 1976. Cyanogenic polymorphism in bracken in relation
to herbivore predation. Nature 260: 604.
Cooper-Driver, G., S. Finch, T. Swain, and E. Bemays. 1977. Seasonal variation in
secondary plants compounds in relation to the palatability of Pteridium aquilinum.
Biochem, Syst. Ecol. 5: 177-183.
Copeland, L. and M. Slaytor. 1974. The excretion of a P-carboline alkaloid harmane in
passion fruit. Physiological Plantarum 31:327-329.
Davis, R. H. and A. Nahrstedt. 1985. Cyanogenesis in insects, pp. 635-654, in G.A.
Kerkut, L.I. Gilbert (Ed.), Comprehensive Insect Physiology, Biochemistry and
Pharmacology II. Pharmacology, Pergamon Press. Oxford.
Davis, R. H. and A. Nahrstedt. 1979. Linamarin and lotaustralin as the source of cyanide
in Zygaena filipendulae L. (Lepidoptera). Comp. Biochem. Physiol. 64B: 395-397.
De Wilde, W. J. J. O. 1971. The systematic position of tribe Paropsiaeae, in particular the
genus Ancistrothyrsus, and a key to the genera Passifloraceae. Blumea 19(1): 99-
104.
De Wilde. W. J. J. O. 1974. The genera of tribe Passifloreae (Passifloraceae), with special
reference to flower morphology. Blumea 22: 37-50.
Deinert, E. I.. J. T. Longino, and L. E. Gilbert. 1994. Male competition in butterflies.
Science 370:23-24.
DeVries, P. 1987. Butterflies of Costa Rica and Their Natural History. Princeton
University Press. Princeton.
Dunlap-Pianka, H. L. 1979. Ovarian dynamics in Heliconius butterflies: Correlations
among daily oviposition rates, egg weight, and quantitative aspects of oogenesis. J.
Insect Physiol. 25: 741-749.


49
Head Thorax Abdomen Wings
Body parts
Figure 3-3. Cyanide quantification of male and female H. e. cyrbia fed P. rubra.
Different letters indicate significant differences between means (Tukey HSD-
test). Bars indicate standard error.
Female
Male
Body parts
Figure 3-4. Cyanide quantification of male and female H. e. cyrbia fed P. punctata.
Different letters indicate significant differences between means (Tukey HSD-
test). Bars indicate standard error.


85
A. TYPE I
.OH HO^
3 4
Tetraphyllin A
Oeidaclin
Tetraphyllin B
(S)
(R)
(1S.4S)
MW 271.1055
MW 271.1055
MW 287.1004
Volkenin
(1R, 4R)
MW 287.1004
Epivolkenin
(1S.4R)
MW 287.1004
Taraktophyllin
(1R.4S)
MW 287.1004
B. TYPE II
p-O-glucopyranosy! /
OH
OH
HO'
\ \ N
Wmi
i-D-glucopyranosyl 6
,4 V
HO'
HO
\
OH
6-deoxy-p-D-gulopyranosyl / 26-dideoxy-IS-D-xy/ohaxoyranosyl /
11 ohH3c V I
I \
V \ /\
OHH3?
\
\
v \ A
\
OH
OH
Passibiflorin (1S.4R)
Molecular formula C18H27N(
Molecular Weight 433.1584
OH
Passicapsin (1S.4R)
Molecular formula C1S H27 N 0
Molecular Weight 417.1635
Passitrifasciatin (1S,4R)
Molecular formula C,8H27NO,,
Molecular Weight 433.1584
C. TYPE III
D. TYPE IV
Linamarin Lotaustralln(R)
MW 247.1 MW 261.1
Glc = p-D-glucopyranosyl, C^H^Oj MW 163
Figure 4-1. Cyanogenic glycosides in Passifloraceae plants (A) Type I, (B) Type II, (C)
Type III and (D) Type IV (Jaroszewski et al. 2002).


43
there are other factors such as specialization, mode of feeding, etc, that herbivores could
also use to cope with plant cyanogenesis.
Temporal Variation in Cyanide in Passiflora Plants
Considerable variation in cyanide concentration was found among Passiflora tested
over two seasons (summer and fall). The concentration of cyanide in the fall was almost
double the amount found in the summer season. I could not find any literature related to
seasonal variation in cyanogenic glycosides. Most of the studies are linked to herbivore
presence (Cooper-Driver et al. 1977), or to altered climatic conditions such as
temperature (Ellis et al. 1977, Rammani and Jones 1985, and Jones and Rammani 1985),
which, it is suggested, contribute to seasonal changes in cyanogenic glycosides levels.
However, in this study, none of the above factors is a contributing factor in the
differences in cyanide concentration, since the biotic and abiotic conditions were similar.
The only environmental factor that might be responsible for the variation is the difference
in photoperiod due to seasonal change. A study of seasonal variation in flavonoid
concentration performed on temperate-zone Alliaria plants (Haribal and Renwick 2001)
produced results similar to those obtained in this study. (Flavonoid concentrations were
also lower in the summer and higher in the fall.) Because all of these Passiflora are
tropical species where photoperiod varies only 30 minutes during a year from a 12
light: 12 dark cycle, this is a puzzling finding that deserves further investigation. Also, it
w'ould be interesting to see how this variation in cyanide concentration correlates to the
survival rates of larvae and/or adults of Heliconius species.
Finally, different Passiflora parts were found to possess different cyanide
concentrations. Higher concentrations of cyanide are found in the new growth. The
concentration within the plant parts followed the same pattern of alteration between the


32
this controversy, what is clear is that feeding is required in order to provoke tissue
disruption and enzymatic hydrolysis.
In Lepidoptera, defensive compounds have been classified in relation to how
predators react to them (Brower 1984). Class 1 are compounds that are noxious due to
their ability to irritate or poison; they may or may not stimulate the olfactory or gustatory
receptors of predators. Class II are harmless chemicals that stimulate the predators
olfactory and/or gustatory receptors (Brower 1984). Thus cyanogenic glycosides are
considered a Class I compound; however, Heliconius also possess Class II pungent odors
emitted through their abdominal glands, (Gilbert 1976), and pyrazines produced in some
unknown region of the body (Hay-Roe and McAuslane unpublished).
Laboratory experiments have demonstrated that insect predators such as birds,
frogs, and lizards can learn to avoid a particular class of prey, depending on the
unpalatability of that prey (Brower 1958, Brower et al. 1963, 1971; Chai 1990, 1996).
Cha) (personal communication) described how inexperienced birds will attack almost
anything. However, after tasting the unpalatable butterflies, they quickly become
conditioned to the pattern of coloration (Chai 1988) and flying styles of distasteful prey
(Srygley and Chai 1990). In the Costa Rican rainforest, Heliconius erato was rejected
41% of the time when tested with jacamar birds (Glbula ruficauda) (Chai 1986).
Various experiments have shown that some butterflies and moths that are generally
considered to be highly toxic can be consumed by birds without any rejection behavior
(Collins and Watson 1983, Chai, 1986). Within a single species, unpalatability cannot be
absolute; the amount and types of chemical defense can vary among individuals,
producing a palatability spectrum (Brower et al. 1967).


2
1981, Gilbert 1991, Benson et al. 1976), behavior (Chai 1986, 1990, 1996), and
physiology (Dunlap-Pianka et al. 1977, Boggs 1987) of this species have been studied
extensively. However, very little is known about the chemical ecology of this species.
Heliconius erato is known to specialize on Passiflora plants (Passifloraceae). The
females lay eggs individually on new growth or tendrils. In many areas of South
America, this species has been considered an oligophagous species, as larvae feed on
many different Passiflora species, but only from the subgenus Decaloba (Benson et al.
1976, Menna-Barreto and Araujo 1985. Hay-Roe 1996). In Central America, however,
this species tends to be monophagous, selecting only one or two host plant species at any
one location (Gilbert 1991, Benson et al. 1976). Diet breadth has been associated with
competition for resources within members of different species of Heliconius that are
distributed in each geographical region. This drives the partitioning of the resources
among competing species.
With respect to its adult diet, Heliconius has developed a special adaptation
allowing it to feed on pollen from Psiguria and Gurania flowers (Cucurbitaceae) by
externally regurgitated digestive enzymes (Gilbert 1972). Pollen plays an important role
in reproduction because butterflies assimilate the amino acids released from the pollen for
egg development and long-term survival (Dunlap-Pianka 1979). Adult H. erato also feed
on liquid nectar from Lantana flowers which provide carbohydrates in their nectar as
well as in pollen (Boggs 1981).
Ehrlich and Raven (1964) have determined that secondary plant compounds play a
critical role in determining patterns of plant utilization. Insects appear to follow their
host plants adaptive radiation, dispersal, and elaboration of secondary chemical


102
Table 5-1. Experimental treatments performed with (3-glycosidase
Experiments 1 and 2 with P. coricea and P. biflora to test Type I and Type II CG
Plant CG (50 pi) + plant (3-glycosidase (50 pi): CONTROL
Plant CG (50 pi) + insect |3-glucosidase (50 pi)
Plant CG (50 pi) + plant [3-glycosidase (50 pi) + insect P-glucosidase (50 pi)
Experiments 3 and 4 with linamarin and amygdalin to test Type III and Type IV CG
CG (50 pi) + [3-glycosidase (50 pi): CONTROL
CG (50 pl)+ [3-glucosidase larval gut (50 pi)
CG (50 pi) + P-glycosidase (50 pi) + [3-glucosidase larval gut (50 pi)
Table 5-2. Cyanide concentration when larval P-glycosidase from Heliconius himera
were tested against different cyanogenic substrates.
CYANOGEN TYPE
pgCN /gof
substrate
Yse
Type I
P. coricea
Plant CG + plant P-glycosidase
Plant CG + larval P-glucosidase
Plant CG + plant P-glycosidase + larval P-glucosidase
60.09 4.92
6.69 0.65
6.49 0.32
Type II
P. biflora
Plant CG + plant P-glycosidase
Plant CG + larval P-glucosidase
Plant CG + plant P-glycosidase + larval P-glucosidase
0.60 0.09
0.81 0.02
0.57 0.02
Type III
Linamarin
CG + plant p-glycosidase
CG + larval P-glucosidase
CG + plant P-glycosidase + larval P-glucosidase
53.82 2.22
3.08 0.10
37.29 1.30
Type IV
Amygdalin
CG + plant P-glycosidase
CG + larval P-glucosidase
CG + plant P-glycosidase + larval P-glucosidase
51.04 10.02
3.28 0.01
53.80 0.01


15
Both H. e. favorinus and H. e. cyrbia had plastic growth trajectories (Figure 2-4
AB) (Higgins and Rankin 1996). However, as indicated above in H e. favorinus the
plasticity is due to variation of head capsule size and developmental time, while in H. e.
cyrbia it is only due to variation in larval developmental time. Differences in growth
trajectories for the two races on their natural host plants can be observed in Figure 2-5.
The last part of the analysis was to calculate growth rates and compare them
between races. Growth rates were significantly different between races (F=6.18; df=l,
227; p<0.01). Host plants also affected growth rate (F=12.11; df=4, 227; pO.OOl)
between races. Lower growth rates in female H. e. favorinus fed P. auriculata, and in
males and females fed P. rubra, were due to prolonged development. Within H e.
cyrbia, low' growth rates occurred when males and females were fed P. rubra (Table 2-4).
Pupae
The average pupal developmental time for both races was eight days. When H. e.
favorinus fed on different host plants, it resulted in a significant variation in pupal mass
(F-4.57; df=4, 145; pO.OOl). Multiple comparison test indicated that individuals fed P.
trifasciata were significantly heavier than individuals fed P. punctata, P. auriculata and
P rubra (Figure 2-6).
Adult
Final size under laboratory conditions corresponds to the data gathered in the field
and from museum specimens. Regardless of the host plant. H. e. favorinus was greater in
wing length than H. e cyrbia (F= 271.63; df=l, 230; pO.OOl) (Table 2-4).
Analysis within H. e. favorinus indicates that host plant selection caused a slight
but significant variation in wing length (F= 2.34; df=4, 145; p=0.05). Multiple
comparison showed that individuals that fed on P. punctata were smaller in size than the


16
ones fed P. trifasciata. No other host caused a significant variation in wing length. On
the other hand, H. e. cyrbia wing length did not differ significantly between individuals
fed on different hosts. There was no size difference between sexes (Figure 2-7).
Mortality'
In general, higher mortality occurred throughout the life cycle in H. e. cyrbia. I
started with 148 individuals from H. e. cyrbia and 143 individuals from H. e.favorinus.
At the end of the experiment I had a total life cycle mortality of 38.74% in H. e. cyrbia
and 13.26% in H. e. favorinus. During the experiments, 8 eggs from H. e.favorinus and
10 from H. e. cyrbia did not hatch; some were dried, others were bitten by ants. High
mortality during the larval stage was observed in H. e. cyrbia fed its host plant P. rubra.
In all cases, larvae died during the early instars (Table 2-5).
Discussion and Conclusions
The results presented here support the preliminary hypothesis that there are
differences in body sizes between races of Heliconius erato from the western and eastern
sides of the Andes. While the group of races from the eastern side of the Andes reveal
low variations in body size, the group of races from the western side of the Andes present
much greater variation. In the eastern region, stabilizing selection probably acts against
extreme phenotypes and favors the more common variant. This mode of selection may
reduce variation and maintains the status of this particular large phenotypic character.
The races of the western side of the Andes are perhaps more exposed to ecological
pressures, such as competitive interactions, for example, as discussed by Benson et al.
(1976). As Brower (1994) has explained, the vicariant separation of the races could have
occurred during the Cenozoic era, specifically the Pleistocene, when the formation of
large mountains around the world (including the Andes) occurred.


45
whether erato or melpomene is the model for this ring, but Mallet (1999) listed three
different reasons that indicate that erato is the model. The first is that erato is more
abundant in the areas where it coexists with H. melpomene; second, erato has a broader
geographical distribution, and third, erato has a broader habitat than melpomene. If H. e.
favorinus in Tingo Maria, Huanuco is less palatable, then this species is in fact a weak
model. We should then expect that this will directly affect the presence of the less
common H. melpomene. This would be true if P. trifasciata were the only host plant
used by H. e. favorinus in the area. However, P. auriculata (which is also the host plant
of Heliconius sara), P. laurifolia, P. tricuspis, P. rubra, and P. vespertilio have been
recorded in Huanuco (Brako and Zarucchi 1993) and can be potential host plants for H. e.
favorinus.
In the last four years, doubts have been raised with regards to the existence of the
classical Mllerian mimic. Instead, some have proposed a new term quasi-Batesian
mimicry (Speed and Turner 1999) as an alternative explanation for some types of
polymorphism associated with mimicry. Quasi-Batesian mimicry describes Mllerian
mimics as weakly unpalatable, which benefit from the presence of more unpalatable
conspecifics.
In this particular case, members of H. e. favorinus that fed on P. trifasciata would
be considered quasi-Batesian mimics, within their race. The males get protection from
the presence of more unpalatable conspecifics (females or other conspecifics that did not
iced on P. trifasciata). However, the females will be affected, since this sex is probably
more exposed to the attack of predators, particularly during oviposition and while
searching for host plants.


8
races were plastic in H. e. petiveranus (due to variation in larval developmental time), but
canalized in H. e. phyllis. The results of this study led me to hypothesize that there may
be a difference in patterns of growth within subspecies from the western vs. the eastern
side of the Andes. 1 also wondered whether the members of the subspecies located on the
eastern side of the Andes were always bigger in size than the members of the western
side of the Andes.
In this chapter of the dissertation, I have first examined museum specimens of the
most representative members of each region to see whether there is a common variation
in size between different races of erato from the eastern and the western sides of the
Andes. Second, I have studied the life histories of two geographical races, H. e.
favorinus from the eastern side of the Andes and H. e. cyrbia from the western side of the
Andes. For each race, phenotypic effects were analyzed based on environmental,
genotypic, and maternal effects. Growth trajectories, final size and mortality were
compared between groups fed natural host plant vs. alternative host plants.
Materials and Methods
Museum Specimens
The Lepidoptera collection of the Allyn Museum of Lepidoptera, Sarasota, Florida,
was used to record the wing lengths of different subspecies of H. erato (Figure 2-2).
Wing length was measured from the base of the wing to the apex. A Vernier caliper (Spi
2000) graduated to 0.1 mm was used to make the measurements.
Four subspecies, H. e. colombina, H. e. hydara, H. e. cyrbia, and H. e. petiveranus,
are representatives from the western side of the Andes. These subspecies were compared
with H. e. phyllis, H. e. favorinus, H. e. magnificus from the eastern side of the Andes.


63
(500.4 MHz) of this sample was recorded at 17C, and the chemical shift axis was
referenced to internal, residual 'HCDaOD which was assigned to 3.3 ppm. Subsequent
addition of tetramethylsilane (TMS) to this solution verified that the assignment of the
residual methanol signal at 3.31ppm was valid. The 2D-COSY and 2D-HMQC spectra
were acquired with the same sample under the same conditions. The UC-NMR spectrum
(125.8 MHz) of the glycoside was recorded at 27C on a larger sample (7.5 mg), obtained
by the addition of more chromatographic material to the original sample, and this
required a larger solution volume of 0.60 ml of CD3OD, as well as a larger 5mm NMR
tube. The chemical shift axis was referenced to the nitrile carbon (CN), which was
assigned 120.4 ppm for comparison to the data of Olafsdottir et al. 1989b.
NMR analysis of Pass ¡flora sp. The cyanogenic glycoside (1.2 mg) was dissolved
in approximately 0.14 ml of methanol-D4, CD3OD (Aldrich, 99.95% atom %-D) and
placed in a Wilmad 520-1A 2.5 mm NMR tube for analysis. The 'H-NMR spectrum
(500.4 MHz) of the glycoside was recorded at 22C, and the chemical shift axis was
referenced to internal, residual 'HCDjOD which was assigned to 3.3 ppm, as above. The
2D-COSY and 2D-HMQC spectra were acquired with the same sample under the same
conditions. The 13C-NMR spectrum (125.8 MHz) of the glycoside was also recorded on
the same sample, but at a temperature of 27C. The chemical shift axis was referenced to
the nitrile carbon (CN), which was assigned 120.5 ppm for comparison to the data of
Olafsdottir et al. 1991.
NMR analysis of Heliconius erato cyrbia fed on Passiflora auriculata. The
partially purified cyanogen (about 2.4 mg) was dissolved in approximately 0.4 ml of
methanol-D4, CD3OD (Acros Organics, 99.96% atom %-D) and placed in a 5 mm NMR


56
However, there was an unresolved issue in their studies. The cyanogenic glycoside of
larvae and adults increased proportionally to the administration of labeled amino acids;
however, within the control, the adult butterflies were also able to synthesize these same
cyanogenic glycosides within 14 days after emergence. Nahrstedt and Davis
hypothesized that substrates for cyanogenic glycosides biosynthesis can be derived from
catabolism of tissue proteins in the absence of dietary supplies of amino acids. During
the adult stage, the acquisition of additional amino acids through pollen feeding would
provide relevant precursors for the production of additional linamarin and lotaustralin
(Gilbert 1991).
The biosynthesis of linamarin and lotaustralin from the amino acids valine and
isoleucine in both plants and insects seems to follow a common process. This process is
achieved using a glucosyltransferase enzyme system (Jaroszewski et al. 1988; Olafsdottir
et al. 1992).
Sequestration of Cyanogenic Glycosides
Intake of cyanogenic glycosides from a plant has been described for the larvae of
the African Acraea horta (Nymphalidae: Acraeinae) which sequesters the cyclopentyl
cyanogenic glycoside (gynocardin) from its host plant Kiggelaria africana
(Flacortiaceae) (Raubenheimer 1987). Also, Parnassius spp. (Papilionidae), the cotton
moth Abraxas spp. (Geometridae), Pryeria snica (Zygaenidae), and Yponomeuta
hexabolous (Yponomeutidae) sequester the cyanoglucoside sarmentosin when fed on
their host plants in the families Crassulaceae and Celastraceae (Nishida 1994, Nishida
and Rothschild 1995, Nishida et al. 1994). However, the species in the genus Abraxas
synthesize sarmentosin even if this compound is not present in any of the plants they fed


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EY32K9BI9_H9HZJR INGEST_TIME 2015-04-01T19:19:23Z PACKAGE AA00029850_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


24
Figure 2-2. Museum specimens of H. erato from the eastern (left column) and western
sides (right column) of the Andes that were included in the analysis. (A) H. e.
phyllis, (B )H. e. favorinus, (C) H. e. magnificus, (D) H. e.colombina, (E)
hydara, (F) H. e. cyrbia, (G) H. e. petiveratms.


38
approximately two times the amount released in the summer. Also, there was a
significant difference in cyanide concentrations between host plants tested (F=45.06;
df=6, 388; p<0.001) and plant parts (F=l31.83; df=2, 388; p<0.001). The concentration
of cyanide was always higher in the new leaves, and this value differed significantly from
the cyanide concentration in both mature leaves and stems. Cyanide concentration in
mature leaves and stems did not follow a constant pattern, and varied according to the
species (Figure 3.2).
Cyanide Quantification in Heliconius erato Fed Their Natural and Alternative Host
Plants
Cyanide concentrations varied significantly between the races (F=29.90; df=l,
139; p<0.001). H. e. cyrbia released greater amounts of cyanide than H. e.favorinus.
There were also differences in cyanide concentration between the sexes (F=8.26; df=l,
139; p<0.005) and also between body parts (F=8.14; df=3, 139; pO.OOl).
H. e. cyrbia fed on P. rubra and P. punctata
Differences in host plant utilization resulted in different concentrations of cyanide
being released from individuals of this race (F=10.02; df=l, 97 p<0.002). When H. e.
cyrbia fed on P. rubra (Table 3-3), the butterflies released higher concentrations of
cyanide, as compared to the butterflies that fed on P. punctata (Figure 3.4). Also, males
and females of this subspecies differed significantly in the quantity of cyanide released
(F=5.68; df=l, 97; p<0.025). The females were able to concentrate higher amounts of
cyanide in their body when fed P. rubra. Furthermore, the concentrations of cyanide
were stored differently within the body parts (F=7.32; df=3, 97; p<0.001). The thorax
had higher concentrations of cyanide, which differed significantly from the amounts
stored in the head, the abdomen, and the wings. When H. e. cyrbia was fed P. punctata,


46
Table 3-1. Heliconius erato fed natural and alternative Passiflora host plant and sample
size used in t
e study.
Geographical race
Host plant
N
Male
Female
H. e. cyrbia
P. rubra
6
6
P. punctata
7
7
H. e. favorinus
P. trifasciata
5
4
H. e. demophon
P. biflora
5
5
P. trifasciata1
3
3
P. aur¡culata1
5
5
P. rubra1
2
2
1 Alternative host plants from other regions
Table 3-2. Passiflora plants tested for cyanogenesis in this study. All plants are used by
Heliconius erato in different geographical areas.
Subgenus
Section
Passiflora
species
Country of
Origin

Summer
Fall
Decaloba
(Plectostemma)
Decaloba
P. biflora
Costa Rica
22
12
Decaloba
P. misera
Brazil
12
6
Decaloba
P. punctata
Ecuador
6
0
Decaloba
P. trifasciata
Peru
17
6
Auriculata
P. auriculata
Ecuador
22
10
Xerogona
P. capsularis
Brazil
6
14
Xerogona
P. rubra
Ecuador
12
0


27
Figure 2-5. Growth trajectories of H. e.favorinus and H. e. cyrbia fed their natural host
plants (P. trifasciata and P. rubra, respectively). Each point represents the
mean value of larvae from the first instar to the fifth. Bars on each point
indicate standard error.


Table 4-7. Peak areas counts of cyanogenic glycosides detected by HPLC/MS in H. erato and H. himera butterflies.
Butterfly sample
Peak area of cyanogenic glycosides x 10b
Linamarin
Lotaustralin
Epivolkenin
Passicapsin
MW 433
Total
Cyanogenic
glycoside
9 H. e. cyrbia fed auriculata
307.70
4.59
681.24
0
0
993.54
S H. e. cyrbia fed auriculata
396.37
3.25
650.00
0.10
0
182.91
9 H. e. cyrbia fed punctata
560.98
573.63
4.41
0.67
1.83
1141.51
9 H. e. cyrbia fed rubra
69.62
66.64
12.21
0.49
0.54
149.51
9 H. e. cyrbia fed biflora
766.52
693.98
2.60
0.26
0.23
1463.59
9 H. e. favorinus fed trifasciata
117.53
110.10
12.12
0.74
5.90
246.39
9 H. himera fed punctata
1106.17
536.34
18.85
0.60
0.40
1662.35
9 H. himera fed rubra
271.24
198.94
22.47
0.45
0
493.11
Butterfly sample
~~~ j -
Percent total peak area of cyanogenic glycosides
Linamarin
Lotaustralin
Epivolkenin
Passicapsin
MW 433
Total
Cyanogenic
glycoside
9 H. e. cyrbia fed auriculata
30.97
0.46
68.57
0.00
0.00
100.00
S H. e. cyrbia fed auriculata
37.76
0.31
61.92
0.01
0.00
100.00
9 H e. cyrbia fed punctata
49.14
50.25
0.39
0.06
0.16
100.00
9 H. e. cyrbia fed rubra
46.57
44.57
8.17
0.31
0.36
100.00
9 H. e. cyrbia fed biflora
52.37
47.42
0.18
0.02
0.02
100.00
9 H. e. favor inus fed trifasciata
47.70
44.68
4.92
0.30
2.40
100.00
9 H. himera fed punctata
66.54
32.26
1.13
0.04
0.02
100.00
9/7. himera fed rubra
55.01
40.34
4.56
0.09
0.00
100.00


> l ¡t
techniques for protein purification. I thank Dr. S. Yu and Dr. D. Boucias for discussing
enzymology work with me and for the use of their laboratory equipment.
Special thanks to the National Science Foundation (NSF) and the National High
Magnetic Field Laboratory External User Program via Dr. P. Teal, CMAVE, USDA for
the use of the Nuclear Magnetic Resonance (NMR) at the Advanced Magnetic Resonance
imaging and Spectroscopy (AMRIS) facility in the McKnight Brain Institute of the
University of Florida. I am very grateful to Mr. Jim Rocca from for his help in the
chemical elucidation of my compounds through (NMR) analysis. I thank Dr. J. Johnson,
who runs the HPLC/Mass Spectroscopy unit in the Chemistry Department. His reports
on the analysis performed are the best! I also appreciate his friendship and long
discussions on the HPLC/MS analysis. Dr. I. Ghiviriva kindly ran the NMR analysis to
corroborate the chemical composition of two plants: P. auriculata and P coricea. Dr.
H. Albom from the USDA generously ran certain samples in the mass spectrometry and
also gave me advice on some of the techniques I used. Special thanks are due to Dr. B.
Torto, for his friendship.
I owe special thanks to Dr. J. Jarozewsky from the Royal Danish School of
Pharmacy in Copenhagen, Denmark, for sending me several milligrams of purified
gynocardin and tetraphyllin B.
Thanks are due to Dr. R. Ferl, Dr. P. Sehnke, and B. Laughter from the
Horticultural Science Department, who allowed me to use the lab ultracentrifuge and
electrophoresis equipment. Beth taught me all of the tricks for getting the best
electrophoretic gel. Working with her was very enjoyable.
v


97
Protein concentration measurements
The protein concentration of the cytosolic and microsomal extracts and fractions
from each purification were determined with the Bio-Rad Bradford protein assay (Bio-
Rad Laboratories. Hercules, CA). A standard curve was prepared from different dilutions
of bovine serum albumin.
In vitro Cyanide Quantification of P-Glycosidase from H. himera Larval Midgut and
Various Cyanogenic Glycosides as Substrates
Amounts of cyanide released for various combinations of plant P-glycosidase, plant
cyanogenic glycosides and larval P-glycosidase enzyme preparations were measured in
vitro, as indicated in Table 5-1. A 0.5-mg quantity of Passiflora species cyanogens was
diluted in 150 pi of distilled water, and for linamarin and amygdalin (Sigma
Laboratories), 0.1 mg of the glycoside was diluted in 150 pi of distilled water.
The reaction mixtures were prepared quickly to avoid cyanide loss. The vials were
then tightly closed with a stopper, and incubated overnight (as described in Chapter 3).
Cyanogenesis was quantified by the Lambert method (described in Chapter 3). A
cyanide standard curve was also calculated from different dilutions of NaCN.
The control for each experimental treatment contained an identical amount of plant
enzyme and cyanogen, but lacked the insect p-glycosidase. Each treatment was
performed twice.
Sodium Dodecyl Sulfate (SDS)-Polyacrylamide Gel Electrophoresis (PAGE)
Samples containing approximately 3 pg protein were combined with 10% (v/v)
glycerol, 50mM TRIS, pH 6.8, 2% SDS, 0.1% bromophenol blue. The samples were
heated for 3 min at 100C in a water bath, before being loaded onto a 10% (w/v)
polyacrylamide gel slab containing 0.1 % SDS (Laemmli 1970). The gel was run at a


11
Egg measurements
Eggs were collected every day from the experimental cages and placed in a small
SOLO cup. Each egg was weighed on an electro balance graduated to 0.01 mg (Mettler
AC 100, Mettler Instruments Corp. Hightstown, New Jersey). Each egg was placed in an
individual cup with the family identification number (ID), offspring ID, host plant and
oviposition date. Fresh shoots were placed into the cups, usually on the second and third
day when I expected the larvae to hatch from the egg. A small wet ball of Kimwipe
tissue was also placed in the cup to provide eggs with adequate moisture.
Larval development
Upon hatching, larvae were removed from the cups with a fine artists brush and
placed on shoots of the experimental host plants (P. auriculata, P. punctata, P. rubra. P.
trifasciata and P. biflora).
The base of each plant cutting (approximately 15 cm in length) was placed in a
glass vial with distilled water. The bottle with the plant was then placed in a plastic
aquarium (one gallon) that was labeled with the following data: family ID, host plant
used, geographical race of erato, and the offspring ID.
As a larva develops, it periodically molts, which is first signaled by apolysis of the
head capsule. The head capsule is a strongly sclerotized part of the larval body and is
therefore a good indicator of growth. Consequently, I measured head capsule
dimensions. Head capsule width and height were measured. Measurements were made
with an Olympus microscope and an ocular micrometer. All measurements were done at
20X magnification. Relying on head capsule width has one drawback. In the last instar,
the epicranial suture splits the head capsule during the molt, so these measurements are
not accurate indicators of growth at this instar. Head capsule height, however, is not


ACKNOWLEDGMENTS
I would like to thank all of the members of my committee, especially my chairman.
Dr. T. C. Emmel, for financial support for a field trip to Per and Panama, for providing
laboratory' space and two greenhouses where I kept plants and butterfly colonies,and for
his friendship and trust. Dr. J. L. Nation was kind enough to guide me throughout many
of the experiments performed for this study. I would also like to thank him for his
laboratory space, which became my "home' for many years. Thanks go to Dr. H. J.
MacAuslane, for allowing me to use an entire compartment of her ultrafreezer. I am very
grateful for her friendship and her support. Thanks go to Dr. D. A. Jones and Dr. F.
Slansky for providing me with enzymes, and some cyanogens used as standards. Dr.
Jones kindly allowed me to use his extensive library containing the most important
literature on passionvines.
I extend my special thanks to several people who were not members of my
committee, but who provided equipment for the chemical analyses and whose comments
were crucial in completing my dissertation. My deepest gratitude is expressed to Dr. J.
Tumlinson and Dr. P. Teal from the Center for Medical, Agricultural and Veterinary
Entomology (CMAVE), U. S. Department of Agriculture (USDA), without whose help
my dissertation would not have been possible. I thank them for their trust and support.
Special thanks as well are due to Dr. C. Lait, with whom I spent a summer through a
USDA Student Intership Program during which I had the opportunity to practice various
IV


TABLE OF CONTENTS
gage
ACKNOWLEDGMENTS iv
LIST OF TABLES xi
LIST OF FIGURES xiii
ABSTRACT xv
CHAPTER
1 INTRODUCTION 1
2 LIFE HISTORIES AND NUTRITIONAL EFFECTS ON SIZE OF HELICONIUS
ERATO FEEDING ON DIFFERENT PASSIFLORA HOST PLANTS 6
Materials and Methods 8
Museum Specimens 8
Living Insects and Host Plants 9
Feeding Experiments 9
Life History Studies 10
Egg measurements 11
Larval development 11
Pupae and adult offspring measurement 12
Statistical Analysis 12
Results 13
Analysis of Specimens from the Museum Collection 13
Life History Studies in H. e. favor mus and H. e. cyrbia 13
Eggs 13
Larvae 14
Pupae 15
Adult 15
Mortality 16
Discussion and Conclusions 16
3 C YANOGENESIS IN HELICONIUS ERA TO AND PASSIFLORA HOST
PLANTS 29
Occurrence of Cyanogenesis in Plants 29
Occurrence of Cyanogenesis in Lepidoptera 30
viii


100
Since P-glucosidase may have this new function, it is suggested that |3-glucosidases is
likely to be a new inhibitory enzyme for cyanide as Spencer (1987) hypothesized.
Although it looks like this protein did not target Type II glycosides, it will be
necessary to re-test this substrate because of the low cyanide released by the control.
Also, since epivolkenin was a metabolic product found in larvae fed plants containing
passicapsin and passibiflorin (see Figures 4.6 and 4.7) the action of P-glucosidases could
explain conversion of diglycosides to monoglycosides.
Future studies should attempt to separate these P-glucosidases and test them
independently against different cyanogenic glycosides substrates to see their effects on
cyanide release upon hydrolysis.
It is known that plants containing cyanogenic glycosides also contain P-
glycosidases. Cyanogenesis occurs upon herbivore feeding, which causes the release of
cyanide and carbonyl products, particularly alkylating cyclopentenones which are more
toxic than cyanide (Yu 1989). The aglycone can therefore be detrimental to the plant as
much as a defense against herbivores, which must also have enzymes that can detoxify
the aglycone. An example of such a situation occurs in the caterpillars of the tiger
swallowtail that feed upon quaking aspen. After the phenolic aglycone is released, it is
quickly detoxified by an esterase (Lindroth 1988). In this paper it has been demonstrated
that P-glucosidases from H. himera inhibit hydrogen cyanide release from the hydrolysis
of Type I glycosides. Thus, a second enzyme can detoxify cyanide in addition to p-
cyanoalanine synthase and hydrase, which transfer the nitrile group back into the amino
acid cycle (Nahrstedt 1992). The discovery of this novel mechanism of inhibition of
cyanide production is important because it can lead to an effective way of dealing with
t


5
(Nishida 2002). Vital clues to the biochemistry of defensive and other chemicals may be
revealed by ecological information gathered in the field or in live captive cultures.
In this dissertation I will combine ecological and biochemical data in order to
understand the patterns of host utilization and detoxification in geographical races of
Heiiconius erato. In Chapter 2,1 compare the life histories of two races of Heliconius
erato (one from the eastern and one from the western side of the Andes) that differ in
body size and describe the effects of natural host plants vs. alternate Pass ¡flora plants on
their growth, nutrition and mortality. In Chapter 3,1 analyze cyanide concentrations in
butterflies and host plants, in order to understand whether there are any differences in
defense mechanisms between races. In Chapter 4,1 describe the isolation and
identification of cyanogenic glycosides in both butterflies and plants in order to learn
how the butterflies process these toxic compounds. In Chapter 5,1 present preliminary
results on the analysis of P-glucosidase from larval guts in order to test the hypothesis
that this enzyme will inhibit cyanide release by Passiflora plants.


3
constituents. Host ranges diverge and insects specialize with respect to their host plants
as they radiate. In the Heliconius-Passiflora system, coevolution is thought to be the
process that has generated this tight insect-plant interaction (Benson et al. 1976, Futuyma
and Keese 1992).
Passiflora plants form a large genus which has been reviewed by many botanists
(e.g., Harms 1924, Killip 1938, De Wilde 1971, 1974). The most recent review, by
McDougal and Feuillet (unpublished), divided the genus into four subgenera and various
super sections comprising a total of 521 species. Characters such as plant chemistry
(Spencer 1984) and ecological data nicely fit the taxonomic data, specifically within each
of the sections of Passiflora.
Passiflora plants have developed a series of morphological characteristics to avoid
predation by herbivores such as Heliconius larvae (Benson et al. 1976). The
development of a great diversity of leaf shapes by the same plant (a phenomenon called
heteroblastia) is notable in the genus. Gilbert (1983) speculated that the difference in leaf
appearance assists the plant in escaping detection by female Heliconiinae butterflies as
they search for appropriate oviposition substrates. Mechanical defense in Passiflora
species may occur in the form of protective hairs on the leaves, especially in the section
Pseudodysosmia. The hooks of Passiflora adenopoda penetrate the cuticle of Heliconius
larvae, causing immobilization and death from desiccation (Gilbert 1971). Passiflora
also develop deciduous filiform stipules which resemble tendrils, and which may function
to stimulate egg placement by Heliconius, later causing the eggs to fall off with the
deciduous stipules. Various Passiflora produce novel structures on the stipules, leaves or
new shoots which resemble the eggs of Heliconius butterflies in shape and color, and


My former M. A. supervisor and professor, Dr. L. Gilbert from the University of
Texas at Austin, allowed me to take all the cuttings I needed from his great collection of
identified Passiflora. My deepest gratitude is also given to Mr. R. Boender from
Butterfly World for providing me with some Passiflora cuttings, adult plants, and
butterflies, and for financial support at key times. I also thank Dr. J. McDougall for the
identification of the Passiflora plants.
I extend thanks to Jacob Olander in Quito, Ecuador, for providing me with H. e.
cyrbia and H. himera.
The Institute of Natural Resources (INRENA) in Lima, Peru, provided me with
collection and export permits for two seasons of work in Tingo Maria, Huanuco, where I
collected Heliconius erato favorinus. Thanks are also due to the Smithsonian Tropical
Research Institute in Panama for providing me with permits to collect H. e. demophon.
I am very grateful for the financial support provided by the Delores A. Auzenne
Scholarship, the minority programs at the University of Florida, the Mulrennan
Scholarship, and for the various semesters on teaching assistantship, which I received in
the Entomology and Nematology Department.
I also extend heartfelt thanks to Christine Eliazar, who was always there to listen
and give me comfort during stressful times, and to Arif, Steve, David, Alfredo, Matthew
and Chris for helping me with the greenhouse maintenance chores. Special thanks are
due to Alfredo Rios for scientific discussions on butterfly issues, and for raising
caterpillars to maintain the greenhouse colonies. Steve and Jim Schlachta and Jerry
Wenzel were always there when I had electrical problems in the greenhouses and
environmental chambers and fixed them quickly to help prevent any harm to my cultures.
vi


120
Spencer, K. C. 1984. Cyclopentenoid cyanogens: their chemistry, importance in plant
systematics, and their role in the coevolution of plants and insects. Ph. D. thesis,
University of Illinois, Urbana.
Spencer, K. C. and D. S. Seigler. 1985a. Passicoccin, a novel cyclopentenoid cyanogen
from Passiflora coccnea. Phytochemistry 24: 2615-2617.
Spencer, K. C. and D. S. Seigler. 1985b. Passibiflorin, epipassibiflorin and
passitrifasciatin: Novel cyclopentenoid cyanogen from Passiflora. Phytochemistry
24- 981-986.
Spencer, K. C. and D. S. Seigler. 1984. Isolation and identification of cyclopentenoid
cyanogens. Phytochem. Bull. 16: 13-21.
Srygley, R. B. and P. Chai. 1990. Flight morphology of Neotropical butterflies:
palatability and distribution of mass to the thorax and abdomen. Oecologia 84: 491 -
499.
Steam, S. C. and J. Koella. 1986. The evolution of phenotypic plasticity in life history
traits: predictions for norms of reaction for age and size at maturity. Evolution
40 893-913.
Tattersall, D. B., S. Bak. P. R. Jones, C. E. Olsen, J. K. Nielsen, M. L. Hansen, P. V. Hoj,
and B. L. Moller. 2001. Resistance to an herbivore through engineered cyanogenic
glucoside synthesis. Science 293: 1826-1828.
Terra, W. R. and C. Ferreira. 1994. Insect digestive enzymes: properties,
compartamentalization and function. Comp. Biochem. Physiol. 109B: 1-62.
Trigo, J. R. 2000. The chemistry of antipredator defense by secondary compounds in
neotropical Lepidoptera: Facts, perspectives and caveats. J. Braz. Chem. Soc.
11(6): 551-561.
Ulubelen, A., R. R. Kerr, and T. J. Mabry. 1982. Two new neoflavonoids and C-
glycosylflavones from Passiflora serratodigitata. Phytochemistry 21:1145-1147.
Vetter, J. 2000. Plant cyanogenic glycosides. Toxicon 38: 11-36
Wiklund, C. and A. Kaitala. 1993. Sexual selection for large male size in a polyandrous
butterfly: the effect of body size on male versus female reproductive success in
Pieris napi. Behav. Ecol. 6(1): 6-13.
Williams, K. S. and L. E. Gilbert. 1981. Insects as selective agents on plants vegetative
morphology: Egg mimicry reduces egg laying by butterflies. Science 212: 467-469.
Witthoun, K. and C. M. Naumann. 1987. Cyanogenesis. A general phenomenon in the
Lepidoptera? J. Chem. Ecol. 138: 1789-1809.


Cyanogenic Glycosides as a Defensive Mechanism 31
Materials and Methods 33
Plant Growth and Insect Rearing 33
Experimental Design and Treatments 34
Butterflies 34
Temporal cyanide quantification of Pass ¡flora plants 34
Quantitative Determination of Cyanide 35
Colorimetric analysis 35
Standard curve 36
Predator-Prey Interaction 37
Statistical Analysis 37
Results 37
Temporal Variation of Cyanide Release in Passijlora Plants 37
Cyanide Quantification in Heliconius erato Fed Their Natural Host Plants 38
H. e. cyrbia fed on P. rubra and P. punctata 38
H. e.favorinus fed P. trifasciata 39
H. e. demophon fed on natural and alternative hosts plants 39
Cyanide Correlation Between Heliconius erato Fed Natural Host Plants 40
Predator-Prey Interaction 40
Discussion and Conclusions 40
Cyanide Concentrations in Heliconius erato fed Pass ¡flora host plants 40
Temporal Variation in Cyanide in Passiflora Plants 43
Implications of Cyanide Concentration for Palatability 44
Implications of Cyanide Concentration for Mimicry 44
4 CYANOGENIC GLYCOSIDES IN PASSIFLORA PLANTS AND
BUTTERFLIES 52
Mechanisms Used by Insects to Process Toxic Compounds from Plants 53
Cyanogenic Glycosides in the Family Passifloraceae 54
Cyanogenic Glycosides in Butterflies 55
Synthesis De Novo 55
Sequestration of Cyanogenic Glycosides 56
Enzymes 57
Materials and Methods 58
Butterflies and Plants 58
Plant and Butterfly Cyanogen Extraction 59
P-Glucosidase Preparations 59
Purification of Cyanogenic Glycoside from Plants and Butterflies 60
Purification by liquid chromatography 60
Purification by thin layer chromatography 61
Purification by high performance liquid chromatography (HPLC) 62
Identification of Cyanogenic Glycosides from Passijlora Plants and
Butterflies 62
Nuclear magnetic resonance (NMR) 62
HPLC/ (+) ESI-MS experimental 64
Amino Acid Analysis 65
IX


86
B
Figure 4-2. Heliconius used in this study (A) H. e. cyrbia, (B) H. e.favorinus, and (C) H.
himera.