Citation
Biochemical, Molecular, and Physiological Aspects of Fluridone Herbicide Resistance in Hydrilla (Hydrilla verticillata)

Material Information

Title:
Biochemical, Molecular, and Physiological Aspects of Fluridone Herbicide Resistance in Hydrilla (Hydrilla verticillata)
Creator:
PURI, ATUL ( Author, Primary )
Copyright Date:
2008

Subjects

Subjects / Keywords:
Diploidy ( jstor )
Dosage ( jstor )
Genetic mutation ( jstor )
Herbicides ( jstor )
Plants ( jstor )
Ploidies ( jstor )
Population growth ( jstor )
Population growth rate ( jstor )
Species ( jstor )
Triploidy ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Atul Puri. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
5/31/2008
Resource Identifier:
496613304 ( OCLC )

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

BIOCHEMICAL, MOLECULAR AND PHYSIOLOGICAL ASPECTS OF FLURIDONE HERBICIDE RESISTANCE IN HYDRILLA ( Hydrilla verticillata ) By ATUL PURI A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2006

PAGE 2

Copyright 2006 by Atul Puri

PAGE 3

This document is dedicated to my beloved parents..

PAGE 4

iv ACKNOWLEDGMENTS I express my sincere appreciation for my me ntor and major advisor, Dr. Gregory E. MacDonald, for his encouragement, guidance, constructive criticism and intellectual stimulation throughout my tenure at the University of Florida. His expertise and advice in this endeavor have been indispensable to my success. I express my sincere appreciation to Dr. Megh Singh, for nurturing my curiosity in weed science, and for his support and unfailing courtesy throughout my tenure at the University of Florida. I thank my committee members: Dr. William T. Haller, Dr . Fredy Altpeter, Dr. George Bowes, and Dr. Donn Shilling. Their advice a nd direction were crucial to my success in this research. I acknowledge the advice, support and friendship of Dr. Frank Dayan. I extend my special thanks to Robert Quer ns and Richard Fethiere for their support and technical help in pursuing my experime nts. In addition, I thank Amit Sethi, Dr. Victoria James, Raman Kaur, Gabriela Lucian i, Sunil Joshi, and Christopher Mudge for their valuable time and assistance in this research. I express my deepest gratitude to my gi rlfriend Disha Sharma, for her support and encouragement. I am eternally indebted to my parents who have been a constant source of encouragement and support th roughout this work. I thank my brother, Ankur Puri, and my sister, Neha Puri, for their inspiration. During my tenure at the University of Flor ida, I have had the opportunity to build lasting friendships. The experiences I had will always be cherished.

PAGE 5

v TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES...........................................................................................................viii LIST OF FIGURES.............................................................................................................x ABSTRACT......................................................................................................................x ii CHAPTER 1 INTRODUCTION........................................................................................................1 Review of Pertinent Literature......................................................................................6 Background Information and Distribution of Hydrilla in the United States.........6 Hydrilla Biology and Growth................................................................................7 Reproduction and Turion Production....................................................................9 Genetic Diversity in Hydrilla..............................................................................11 Ploidy...................................................................................................................13 Hydrilla Management in Florida Lakes...............................................................14 Hydrilla and Fluridone........................................................................................16 Development of Fluridone Resistance in Hydrilla..............................................17 Molecular Aspects of Fluridone Resistance........................................................17 2 BIOCHEMICAL DIFFERENCES AMONG DIFFERENT HYDRILLA POPULATIONS BY COMPARING CHANGES IN PHYTOENE AND CAROTENE LEVELS AS A FUNCTI ON OF HYDRILLA POPULATION (RESISTANT VS SUSCEPTIBLE) AND FLURIDONE DOSES OVER TIME.....20 Introduction.................................................................................................................20 Materials and Methods...............................................................................................25 Collection and Maintenance of Plant Material....................................................25 Laboratory Evaluations.......................................................................................25 Pigment Analysis.................................................................................................26 Statistical Procedures...........................................................................................27 Results and Discussion...............................................................................................27

PAGE 6

vi 3 GROWTH AND REPRODUCTIVE PHYSIOLOGY OF FLURIDONE SUSCEPTIBLE AND RESISTANT HYDRILLA POPULATIONS........................40 Introduction.................................................................................................................40 Materials and Methods...............................................................................................47 Results and Discussion...............................................................................................49 Shoot Length.......................................................................................................49 Shoot and Root Biomass.....................................................................................50 Turion Production................................................................................................51 Production of Axillary Branches.........................................................................53 Flower Production...............................................................................................53 Total Biomass and Relative Growth Rate...........................................................54 4 MOLECULAR CHARACTERIZATION OF FLURIDONE RESISTANT HYDRILLA IN FLORIDA........................................................................................65 Introduction.................................................................................................................65 Materials and Methods...............................................................................................68 Collection of Plant Material................................................................................68 Chemicals and Reagents......................................................................................68 Genomic RNA Extraction and cDNA Synthesis.................................................68 PCR Amplification and Sequenci ng of Phytoene Desaturase ( pds ) Gene..........69 Results and Discussion...............................................................................................71 5 HYDRILLA PLOIDY IN RELATION TO FLURIDONE RESISTANCE...............91 Introduction.................................................................................................................91 Materials and Methods...............................................................................................96 Plant Materials.....................................................................................................96 Cytological Preparation.......................................................................................96 Flow Cytometry Procedure and Stan dardization of Flow Cytometry.................96 Results and Discussion...............................................................................................98 6 SUMMARY AND CONCLUSIONS.......................................................................108 APPENDIX A PHYTOENE, CAROTENE AND PHYTOENE: CAROTENE RATIO IN S AFFECTED BY FLURIDONE DOSES AT VARIOUS GROWTH STAGES......114 B PHYTOENE, CAROTENE AND PHYTOENE: CAROTENE RATIO IN R1 AFFECTED BY FLURIDONE DOSES AT VARIOUS GROWTH STAGES......115 C PHYTOENE, CAROTENE AND PHYTOENE: CAROTENE RATIO IN R2 AFFECTED BY FLURIDONE DOSES AT VARIOUS GROWTH STAGES......116 D PHYTOENE, CAROTENE AND PHYTOENE: CAROTENE RATIO IN R3 AFFECTED BY FLURIDONE DOSES AT VARIOUS GROWTH STAGES......117

PAGE 7

vii E PHYTOENE, CAROTENE AND PHYTOENE: CAROTENE RATIO IN R4 AFFECTED BY FLURIDONE DOSES AT VARIOUS GROWTH STAGES......118 F PHYTOENE, CAROTENE AND PHYTOENE: CAROTENE RATIO IN R5 AFFECTED BY FLURIDONE DOSES AT VARIOUS GROWTH STAGES......119 LIST OF REFERENCES.................................................................................................120 BIOGRAPHICAL SKETCH...........................................................................................135

PAGE 8

viii LIST OF TABLES Table page 2-1 EC50 phytoene (Fluridone dose; g L-1) concentrations for fluridone susceptible (S) and resistant (R1-R5) hydrilla pop ulations as a function of time.......................38 2-2 EC50 carotene (Fluridone dose; g L-1) concentrations for fluridone susceptible (S) and resistant (R1-R5) hydrilla pop ulations as a function of time.......................39 3-1 Shoot length (cm) in fluridone susc eptible (S) and resistant (R1-R5) hydrilla populations...............................................................................................................57 3-2 Shoot biomass (g per plant) in flurid one susceptible (S) a nd resistant (R1-R5) hydrilla populations..................................................................................................58 3-3 Root biomass (g per plant) in fluri done susceptible (S) and resistant (R1-R5) hydrilla populations..................................................................................................59 3-4 Subterranean turions produced (per plan t) in fluridone fluridone susceptible (S) and resistant (R1-R5) hydrilla populations..............................................................60 3-5 Axillary branches (per plant) in fluridone susceptible (S) and resistant (R1-R5) hydrilla populations..................................................................................................61 3-6 Flower number (per plant) in fluri done susceptible (S) and resistant (R1-R5) hydrilla populations..................................................................................................62 3-7 Total biomass (g per plant) in flurid one susceptible (S) and resistant (R1-R5) hydrilla populations..................................................................................................63 4-1 Nucleotide differences, along with amino acid changes, in pds sequences in resistant hydrilla populations comp ared with fluridone-sensitive pds sequence at planting (Sept 2004). Corresponding amino acid positions in pds where available, are in super-script.....................................................................................88 4-2 Nucleotide differences, along with amino acid changes, in pds sequences in resistant hydrilla populations comp ared with fluridone-sensitive pds sequence at 12 MAP (Sept 2005). Corresponding amino acid positions in pds where available, are in super-script.....................................................................................89

PAGE 9

ix 5-1 Ploidy levels in hydrilla plants in fluridone susceptib le (S) and resistant (R1-R5) hydrilla populations................................................................................................105 5-2 Mean nuclear DNA contents (pg) in plants with di fferent ploidy levels among fluridone susceptible (S) and resistant (R1-R5) hydrilla popul ations determined by flow cytometry..................................................................................................106

PAGE 10

x LIST OF FIGURES Figure page 1-1 Geographical distribution of the invasive aquatic weed hydrilla in the USA............5 2-1 Regression analysis of fluridone sus ceptible and resistan t hydrilla populations between phytoene or carotene content and fluridone doses at planting (Sept 20, 2004); ) phytoene content; ) carotene content; Data represents mean of two experiments with 3 replications; Means values (n= 6) are presented with 95% confidence interval bars (Standard error x 1.96)......................................................33 2-2 Regression analysis of fluridone sus ceptible and resistan t hydrilla populations between phytoene or carotene content and fluridone doses 3 MAP (Dec 20, 2004); ) phytoene content; ) carotene content; Data represents mean of two experiments with 3 replications; Means values (n=6) are presented with 95% confidence interval bars (Standard error x 1.96)......................................................34 2-3 Regression analysis of fluridone sus ceptible and resistan t hydrilla populations between phytoene or carotene content and fluridone doses 6 MAP (Mar 20, 2005); ) phytoene content; ) carotene content; Data represents mean of two experiments with 3 replications; Means values (n=6) are presented with 95% confidence interval bars (Standard error x 1.96)......................................................35 2-4 Regression analysis of fluridone sus ceptible and resistan t hydrilla populations between phytoene or carotene content and fluridone doses 9 MAP (Jun 20, 2005); ) phytoene content; ) carotene content; Data represents mean of two experiments with 3 replications; Means values (n=6) are presented with 95% confidence interval bars (Standard error x 1.96)......................................................36 2-5 Regression analysis of fluridone sus ceptible and resistan t hydrilla populations between phytoene or carotene content and fluridone doses 12 MAP (Sep 20, 2005); ) phytoene content; ) carotene content; Data represents mean of two experiments with 3 replications; Means values (n=6) are presented with 95% confidence interval bars (Standard error x 1.96)......................................................37 3-1 Relative growth rate in fluridone sus ceptible (S) and resist ant (R1-R5) hydrilla populations during the growth season......................................................................64

PAGE 11

xi 4-1 Agarose gel electrophoresis of PCR amplified pds gene sequences from different hydrilla populations; Lane M: 1 Kb molecula r ladder; Lane 1, 2: Susceptible (S); Lane 3, 4: R1; Lane 5, 6: R2; Lane 7, 8: R3; Lane 9, 10: R4; Lane 11, 12: R5; Lane 13: negative control; 1.9 Kb is the size of amplified pds gene sequence........79 4-2 Agarose gel electrophoresis of restric tion digest of TOPO vector containing pds alleles from hydrilla populations with Eco R1 enzyme amplified; Lane M: 1 Kb molecular ladder; Lane 1, 2: Susceptible (S); Lane 3, 4: R1; Lane 5, 6: R2; Lane 7, 8: R3; Lane 9, 10: R4; Lane 11, 12: R5 ; Lane 13: negative control (uncut plasmid); 1.9 Kb is the size of amplified pds gene sequence...................................80 4-3 Aligned sequences of pds alleles from different hydrilla populations. Start, stop codons, and mutations in re sistant hydrilla populati ons are in bold letters..............81 4-4 Aligned sequences of PDS protein fr om different hydrilla populations. Amino acid changes in resistant populations are denoted in bold letters.............................86 4-5 Alignment of partial phytoene desa turase gene sequences from various organisms showing arginine codon at the amino acid position 304 of hydrilla pds gene..........................................................................................................................9 0 5-1 Histograms of flow cytometric nuclear analysis of standard (triploid) (5.1a), diploid (5.1b), triploid ( 5.1c) and tetraploid (5.1d) shoot apical meristematic tissue of hydrilla.....................................................................................................104 5-2 Histograms of endoreduplication patterns of diploid hydrilla plants from hydrilla populations R1 (5.2a) and R2 (5.2b and 5.2c).......................................................107

PAGE 12

xii Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy BIOCHEMICAL, MOLECULAR AND PHYSIOLOGICAL ASPECTS OF FLURIDONE HERBICIDE RESISTANCE IN HYDRILLA ( Hydrilla verticillata ) By Atul Puri May 2006 Chair: G.E. MacDonald Major Department: Agronomy Hydrilla is one of the most serious aqua tic weed problems in the United States. Fluridone is the only United States Environment Protection Agency (USEPA) approved herbicide that provides syst emic control of hydrilla and recently, there has been a decrease in fluridone efficacy for hydrilla co ntrol in many Florida lakes. To characterize fluridone resistance, hydrilla populations were co llected from different Florida lakes with varied histories of fluridone use and grow n under controlled conditions for a period of one year in the absence of fluridone. During this one year period, phenotypic measurements were performed to monitor differences in growth and reproductive physiology. In addition, shoot tissue was colle cted from each population at 0, 3, 6, 9, and 12 months after planting and exposed to 5, 10, 15, 20, 30 and 50 g L-1 fluridone to assess changes in fluridone susceptibility over time. Regression analysis was performed to calculate EC50 values for phytoene and carotene. EC50 carotene values of 9 and 63 g L-1 fluridone were found in the suscepti ble and the most resistant population,

PAGE 13

xiii respectively. There was no significant cha nge in fluridone resi stance in hydrilla populations over time. Molecular characteriza tion was performed by cloning the gene for phytoene desaturase ( pds ) from fluridone susceptible a nd resistant hydrilla plants. Two independent somatic mutations at the Arg304 codon of pds were observed. The codon usage for Arg304 is CGT and a single point mutation yielding either Ser304 (AGT) or His304 (CAT) was identified in different re sistant hydrilla popul ations. Resistant populations were significantly supe rior to or at par with su sceptible hydrilla in growth and reproductive parameters, indicating no dele terious effects of mu tations. To correlate varying levels of fluridone resistance to pl oidy in hydrilla, flow cytometric analysis was performed. Differential ploidy levels (diplo id 2n= 2x= 16; triploid 2n= 3x= 24; and tetraploid 2n= 4x= 32), along with endore duplication patterns were observed among different hydrilla populations, and plants w ithin each population. Aggressive spread of fluridone resistant dioecious hydrilla in aquatic ecosystems can severely impact hydrilla management, and consequently cause subs tantial and long-las ting ecological and economic problems throughout the Southern USA.

PAGE 14

1 CHAPTER 1 INTRODUCTION Hydrilla ( Hydrilla verticillata ) is an invasive, submersed, aquatic vascular plant that causes major problems in fresh water eco systems of the United States. Discovered in Florida in 1959 (Blackburn et al., 1969), it has spread throughout the south-eastern United States and is found as far west as Ca lifornia and as far north as Washington state and Connecticut (Lazor, 1978). This monocot yledonous species has become the most abundant submersed aquatic plant in Florida a nd is one of the most serious aquatic weed problems in the southern and western US A (Figure 1-1) (USGS, 2002). Recently its infestations are also recorded in different la kes in Maine. Problems associated with its excessive plant growth are well documented and include both econom ical losses due to interference with irri gation, flood control, navigation, a nd recreational activities as well as ecological consequences resulting in th e displacement of nativ e plant communities. Consequently, much of the aquatic plant re search in the past 30 years in Florida has concentrated on the development of ma nagement programs for hydrilla due to its dominance in aquatic ecosystems. Nonethel ess, hydrilla has continued to expand, and control of this exotic weed sp ecies with either sterile grass carp or herbicidal applications remains both controversial and costly. Dioecious hydrilla regrows in the spring fr om the root crowns and/or subterranean turions in the hydrosoil as the water temperat ure begins to increase (Haller et al., 1976). During the initial phases of growth, the shoot s quickly reach the water surface forming a dense surface canopy. This mat of hydrilla shoo ts effectively absorbs nearly 95% of the

PAGE 15

2 available light, limiting competition from su bmersed native species. Rapid hydrilla growth and expansion are favor ed by its low light and CO2 compensation points, reduced photorespiration due to a C4-like photosynthetic mechanis m and prolific reproductive capacity. Hydrilla is extremely difficult to manage and control because spread can occur through a variety of mechanisms including fr agmentation, root crowns, and specialized dormant buds called turions (Langeland, 1996). Seed production has been reported for the monoecious biotype but does not occur in dioecious hydrilla in the United States. Turions can be formed in the axils of leaves, or at the ends of positively geotropic rhizomes which extend into the hydrosoil (Haller, 1976; Yeo et al., 1984). Left unmanaged, this invasive plant can rapidly expand over thousands of contiguous hectares, displacing native plant communities and causing significant damage to the ecosystems (Colle and Shireman, 1980; Schmitz and Osborne, 1984; Bates and Smith, 1994). Fluridone herbicide has been the primary means of hydrilla control in Florida over the past 20 to 25 years (Michel et al., 2004). The widespread use of fluridone is due to several factors including low use rates, favor able native plant selectivity, slow activity (reduced oxygen depletion) and often provides one to two years of control of hydrilla. Hydrilla is controlled in large water bodies (> 100 000 ha) by sustaining between 12 and 36 nm (4 g/L) concentrations of fl uridone in lake water for several weeks (Netherland and Getsinger, 1995; Fox et al., 1996). Fluridone is a non-competitive inhibitor of the enzyme phytoene desaturase (PDS) which constitutes a rate limiting step in carotenoid synthesis in plants (Cham ovitz et al., 1993). PDS is a nuclear-encoded protein and has activity in the ch loroplasts, the site of carotenoid synthesis (Bartley et al.,

PAGE 16

3 1991). Under high light intensities, carotenoid s stabilize the photos ynthetic apparatus by quenching excess excitation energy; therefor e, inhibition of PDS decreases colored carotenoid concentration and causes photobl eaching of green ti ssues (Bger and Sandmann, 1998). Only the dioecious form of hydrilla is curre ntly found in Florida, with spread and reproduction limited to asexual means (subterran ean turions, axillary turions, fragments, and root crowns). Therefore, the developm ent of resistance to herbicides was not expected. However in recent years, hydrilla populations in some Florida lakes have become resistant to this herb icide, although the exact mechanis m of resistance is still not fully known. Most biochemical studies investig ating the potency of PDS inhibitors have been made either in vivo with intact cells or in vitro with cyanobacterium cell extracts, with limited data available on higher plants. Recently it has been determined that the resistance of fluridone is due to a point mutation in the phytoene desaturase gene in hydrilla, though the difference in levels of resistance in different biotypes is not explained. Therefore, studies should be made to determine the presence and influence of these point mutations on the development of resistance as target-site resistance mechanism, thereby providing a tool for predic ting susceptibility of hydrilla infestations prior to herbicide applications. Ploidy level in hydrilla is highly variable with cells of diploid, triploid and tetraploid plants within the same population (Langeland et al ., 1992). Recent research has demonstrated that cytological changes from diploid to triplo id plants occur naturally in various hydrilla populations in Japan. Such diffe rences in the level of genetic variation in different weed species suggest that the populat ion genetic structure of successful exotic

PAGE 17

4 weed species like hydrilla may be an importa nt factor influencing the adaptability of plants under different ecological conditions which could include control measures. Observations by field researchers suggested that hydrilla from fluridone resistant and fluridone sensitive populations varied in their ability or capacity to produce subterranean turions, with re sistant populations having a lo wer growth rate and producing a fewer number of turions than the susceptible hydrilla popul ations (Puri et al., 2004). Therefore, the possibility exists that the once resistant hydrilla popul ation may revert to susceptible in the absence of fluridone. Theref ore, there is also a need to investigate possible phenotypic differences in growth and turion physiology between susceptible and resistant hydrilla biotypes. This will increas e our understanding of fluridone resistance and its long-term persistence in Florida Lakes. There are several economic implications of these findings that could dramatically impact hydrilla management in Florida. At >30 ppb fluridone use, the cost per hectare triples from current prices to $250-300/ha. According to recent Florida Department of Environmental Protection budget requests, hydr illa infests over 11, 500 hectares with $17,906,098 needed for control (Florida DEP, 2004) . However, application rates of 30 ppb or higher will also result in a loss of se lectivity (i.e., damage or loss of some emergent and many or all submersed species). Furthermore, if additional resistance occurs, the cost and environmental impact of fluridone may preclude its further use in Florida. Therefore, to prevent the spread of fluridone resistance development in hydrilla, other herbicidal programs need to be employed in alteration with fluridone.

PAGE 18

5 Figure 1-1. Geographical distribu tion of the invasive aquatic weed hydrilla in the USA. The dioecious form (red) was first re ported in Tampa and Miami, FL and spread to the southern and wester n United States. The monoecious form (green) was introduced in the east coast and California. Areas in purple have both dioecious and monoecious forms pres ent. The lighter colored States do not have records whereas the darker colored States do have records of hydrilla. The map is from the United Stat es Geological Services and was last updated September 2002 (USGS, 2002). Arrow designates points of appearance of fluridone resistance in hydrilla.

PAGE 19

6 Review of Pertinent Literature Despite extensive control efforts during the last 4-5 decades, the exotic invasive weed species hydrilla remains a dominant weed problem in the southeastern United States. Hydrilla has spread th roughout the country’s waterw ays, clogging irrigation and drainage canals, degrading wa ter quality, reducing productivity of recreational fisheries, and impeding navigation. Despite an expend iture of $50 million for control during the 1980’s, the percentage of Florida waters i nvaded by hydrilla increased from 37 to 41% (USEPA, 2001). The cost of using contact herbicides to control hydrilla is $500 per hectare (UCS, 2005) while the mechanical cont rol for hydrilla can cost as much as $2500 per hectare (SE-EPPC, 2005). Background Information and Distribution of Hydrilla in the United States Hydrilla is a monotypic plant belonging to the family Hydrocharitaceae , with the centre of origin thought to be tropical Asia (Cook and Lnd, 1982). However, it is distributed throughout the tropi cal and subtropical parts of the world including Germany, England, Poland, the upper Nile of Africa, the Southeast Asia , Australia, India, China, Japan and the United States (Lazor, 1978). Due to its specialized growth habit, physiological characteristics, and various means of asexual reproduction, hydrilla has been described as “the perfect aquatic weed” (Langeland, 1996). Both monoecious (staminate and pistill ate flowers on the same plant) and dioecious (staminate and pistillate flower s on different plants) biotypes have been described, and both are present in the Unite d States. Cook and Lnd (1982) reported that on a worldwide basis, the monoecious forms la rgely dominate in the tropical regions, whereas the dioecious forms are largely temper ate. However, the current distribution for both monoecious and dioecious biotypes in Nort h America is contrary to this observation.

PAGE 20

7 The dioecious form was discovered in Flor ida in the 1950’s while the monoecious form was first reported in the Potomac River in the mid 1980’s (Steward et al., 1984). The dioecious female plant spread aggressively throughout Florida, then west to Texas and California (Yeo and McHenry, 1977). Its range also extends north into Georgia, Tennessee, South Carolina and North Carolin a (Yeo et al., 1984) . A second, separate introduction of a monoecious plant was reporte d in Delaware in 1976 and quickly spread into Virginia, Maryland, and North Carolina. A monoecious plant has also been reported in California (Ryan and Hommberg, 1994). The tw o biotypes have several differences in terms of vegetative growth habit and ase xual propagule production (Steward and Van, 1987; Ames et al., 1986). Hydrilla Biology and Growth Anatomically hydrilla has reduced vasculat ure. The xylem tissue is vestigial and the phloem is greatly reduced (Yeo et al., 1984) . Hydrilla leaves are only two cells thick with upper epidermis much larger than the lower cells (Pendland, 1979). An interesting feature of hydrilla is that it is the only known plan t to operate a C4 photosynthetic CO2 concentration mechanism without possessing Kr anz anatomy (Bowes and Slavucci, 1984; Magnin et al., 1997). Hydrilla has developed an inducible C4-acid cycle to combat adverse conditions such as limiting CO2, high O2 concentrations, high temperature and irradiance. Therefore, h ydrilla can shift between C3 and C4-type photosynthesis depending on environmental conditions (Magnin et al., 1997). It is believed that the hydrilla system represents an archetypal form of C4 photosynthesis among angiosperms, and that this process may have occurred in water before its appearance on land (Magnin et al., 1997).

PAGE 21

8 Hydrilla regrows in the spring from the root crowns and/or subt erranean turions in the hydrosoil as the water temperature begins to increase (Haller et al., 1976). During the initial phases of growth, the shoots quickly reach the water surface forming a dense green canopy which effectively absorbs nearly 95% of the available light, limiting competition from other native species. Ra pid hydrilla growth and expansion are favored by its low light and CO2 compensation points, reduced photorespiration due to a C4 like photosynthetic mechanism and its prolific reproductive capacity (Van et al., 1975; Holaday et al., 1983; Magnin et al., 1997). In addition, Kulushreshtha and Gopal (1983) reported allelopathic properties of hydrilla on a species of Ceratophyllum , thereby further increasing the competitive advantage of this invasive species. A hydrilla colony originating from a singl e shoot can expand radially via stolon growth at a rate of 4 cm d-1, with an average production of 1 new ramet m-2 d-1 (Madsen and Smith, 1999). For the dioecious hydrilla, most of the colony expansion (99.9%) is by stoloniferous growth, while the spread from fragmentation is only 0.02 ramets m-2 d-1 (Madsen and Smith, 1999). There have been reports of 1,250 to 1,976 tubers (subterranean turions) per m-2 being produced by the dioeciou s hydrilla within a period of four months in various Florida Lake sedime nts (Sutton and Portier, 1995). However, in other areas, up to 2,812 tubers m-2 were produced by the dioecious hydrilla biotype during winter, while for the monoecious hydr illa the production of tubers during the summer was 5,366 tubers m-2 and 2,740 tubers m-2 during winter (Sutton et al., 1992). Hydrilla can grow from the s ubstrate to the water surface a nd reach up to 15 m in length (Langeland, 1990). From there, the stems bran ch with leaf whorls (nodes) every 11-12

PAGE 22

9 mm in the dioecious hydrilla and every 16 mm in the monoecious (Ryan et al., 1995). For each biotype, individual nodes can rege nerate a new plant (Haller, 1976). Biomass allocation studies of the dio ecious hydrilla show this plant can accumulate up to 1200 g m-2 dry weight of aboveground shoot tissue (Madsen and Owens, 1998). In shallow ponds in Texas, dioecious hydrilla produced 200 g m-2 dry shoot weight during the wint er months, and about 600 g m-2d-1 during the summer months (Madsen and Owens, 1998). Hydrilla has several organs for storage of carbohydrates, e.g., tubers, turions, stolons, st ems and root crowns (Madsen and Owens, 1998). Of these, the upper and lower stem s contain the largest amount of total nonstructural carbohydrates, ranging from 100 to 700 g m-2 in the upper and lower stems respectively. Stolons and root crowns are th e main source for regrowth in spring, rather than from tubers. Tubers can remain viable in the sediment for at least four years. Reproduction and Turion Production Hydrilla reproduces in nature through a variety of means including fragmentation, seeds, stolons, and rhizomes. Stem fragme nts containing as few as one node (Langeland and Sutton, 1980) can form a mature plant a nd will form as much biomass as 16 shoot tips (Sutton et al., 1992 ). The main reason for its persis tence and longevity in natural environments is through production of speciali zed dormant buds called “turions”, and is the greatest restrain for control of this aqua tic weed species. Turions can be formed in the axils of the leaves i.e., axill ary turions, or at the end of positively geotropic rhizomes, which extend into the hydrosoil i.e., subt erranean turions (tub ers) (Sculthorpe, 1967; Haller, 1976; Yeo et al., 1984). Subterranean tu rions can remain dormant for as long as 5 years (Van and Steward, 1990) a nd are thought to maintain hydr illa within a given area, possibly through periods of drought. Axillary turions are smaller than subterranean

PAGE 23

10 turions and are thought to f unction in dispersal (Thullen, 1990). These structures are generally formed on detached floating mats of hydrilla (Thullen, 1990; Miller at al, 1993) and the majority only last for about 1 year on the hydrosoil (Van and Steward, 1990), presumably due to their smaller size. Subterra nean turions outnumber axillary turions by ten times in a given area of hydrilla (Mitr a, 1955). There are two biotypes of hydrilla (monoecious and dioecious) throughout the U.S. and different parts of the world. These biotypes appear very similar, but differ in their reproductive cycle. Research has shown far more production of subterra nean turions than axillary in the dioecious biotype (Haller and Sutton, 1975; Spencer et al., 1987), though monoecious plants produce 20 to 30 percent of their tota l number of turions aboveground i. e axillary turions (Anderson and Spencer, 1986). Turion formation has been sh own to be a photoperiodic effect by various researchers (Haller et al., 1976; Steward, 1993) and most pr oduction in dioecious hydrilla occurs under short days (Steward an d Van, 1987; Spencer and Anderson, 1986). Monoecious biotypes produce turions throughout year while dioeciou s biotypes produce turions only under short-day condition s (Haller et al., 1976; Steward, 1997). Seed production has been reported for the monoecious biotype in different parts of the world including India (M itra, 1955), Australia (Sainty and Jacobs, 1981) and in the U.S. (Langeland and Smith, 1984), but has not reportedly occurred in the dioecious biotype in the U.S. Flowering may not be an a ccurate indicator of sexuality in this species as sex conversion has been reported in dioecious forms depending on the external environmental factors (cultivation), leading to doubts on the truly dioecious nature of some races. However European plants that have been observed in the wild and in cultivation since ca. 1830 have never de veloped male flowers (Cook and Lnd, 1982).

PAGE 24

11 Research has shown the sexual compatibility of both monoecious and dioecious biotypes including the production of vi able seeds by dioecious female plants from the crosses between dioecious pollen donor s and monoecious pollen donors under culture conditions (Steward, 1993). The offspring of such a cross is likely to be geneti cally variable, thereby increasing the genetic variation betw een plants at a community level. Genetic Diversity in Hydrilla Genetic diversity in aquatic macrophytes is generally lower than in terrestrial plants as revealed by compara tive isozyme studies (Barret et al., 1993) which is primarily due to the predominance of vegetative reproduct ion in aquatic plants. Hydrilla however is regarded as a species of high genetic varia tion (Triest, 1991). High genetic variation in hydrilla strains from the U.S. (Verkleij et al., 1983b), Africa (Pieterse et al., 1985), and elsewhere have been shown by comparative is ozyme studies (Verkle ij and Pieterse, 1986; Nakamura et al., 1998), and random pol ymorphic DNA (RAPD) (Les et al., 1997; Madeira et al., 1997; Hofstra et al., 2000). This analysis wa s performed to identify new hydrilla strains and infestations, and to determine the genetic relationship between geographically diverse hydrilla populations. Ryan et al. (1991) also i nvestigated isozymic variability in subterranean turions of monoecious and di oecious hydrilla and reported different isozymic bands for alchohol dehydr ogenase (ADH), aspartat e aminotransferase (AAT) and NADP-malic enzyme (NADP-ME). Pi eterse et al. (1985) studied genetic variation within hydrilla plants of th e same population by isozymatic comparison. However, only glutamate oxaloacetate transami nase (GOT) showed variation in isozyme phenotypes, whereas other enzymes such as phosphoglucomutase (PGM), peroxidase (PO), shikimate dehydrogenase (SDH), superoxide dismutase (SOD), NADH

PAGE 25

12 dehydrogenase (NADH-DH), malic enzyme (M E), and alchohol dehydrogenase (ADH) showed no isozymatic vari ation within the species. Isozyme patterns of African plants are not very distinctive probably due to localized infestations and the non-weed nature of hydrilla on the continent. Large levels of variations from isozyme patterns have b een observed in hydrilla collections from different parts of Southeast Asia and also be tween hydrilla plants w ithin the same lake (Verkleij and Pieterse, 1986). This geographi cal area is reported to be the center of differentiation of this noxious weed species (Cook and Lnd, 1982). The genetic closeness of African and a group of plants fr om the Indian subcontinent have also been reported (Verkleij and Pieterse, 1991). The U.S. dioecious and monoecious biot ypes seem to have different isozyme patterns (Ryan et al., 1991). The dioecious hydr illa populations in the United States are more closely related to hydrilla strains from Bangalore, India, while monoecious forms are genetically closer to hydr illa strains from Seoul, Korea (Madiera et al., 1997). Nakamura et al. (1998) tested isozymatic patterns of 226 accessions of hydrilla collected from different lakes in Japan. Of these, 17 and 23 electrophoretic phenotypes, based on polymorphism in banding patterns of enzyme PGM, were identifiable in diploid and triploid plants, respectively. To the contrar y, monoecious plants showed no variation of the banding patterns. However, monoecious plants were distinct from dioecious in having specific alleles which were not shared by dioecious plants (Nakamura et al., 1998). Monoecious accessions in Japan were all triploid (Nakamura and Kadono, 1983) and their reproduction was almost exclusively vegetative. In monoecious plants, fruiting occurred by selfing only when they were diploid (Cook and Lnd, 1982). On the other

PAGE 26

13 hand, dioecious plants showed diverse ge netic variation. The chances of sexual reproduction in nature are very limited in hydrilla because co-occurrence of the female diploid plants and a pollen donor is required (Nakamura et al., 1998). However, limited chances of sexual reproduction must have brought about new geneti c recombination, and increased genetic diversity of hydrilla. Ploidy Hydrilla does not sexually re produce in Florida and south east U.S.; i. e. seeds are not produced (Haller, 1976). In general, se xually reproducing speci es are characterized by high genetic diversity as compared with vegetatively propagat ed species (Loveless and Hamrick, 1984). However, it has been documen ted that hydrilla is a polyploid plant, with chromosome counts varying widely within a vegetative population (Cook and Lnd, 1982; Verkleij et al., 1983a). Plants in As ia, India and Europe are either diploid (2n = 2x = 16) or triploid (2n = 3x = 24) (Cook and Lnd, 1982). Presence of tetraploid plants (2n = 4x = 32) has been reported in Alabama (Davenport, 1980). Both diploid and triploid plants have been collected from different parts of Washington DC, Maryland, and Texas (Verkleij et al., 1983a; Langeland, 1989) ; and those collected from California, Florida, Texas, and Connecticut have been recorded as triploid (Harlan et al., 1985; Langeland, 1989). Hydrilla plants in various dioecious populat ions in Japan are either diploid or triploid, whereas monoecious forms in Japan are always triploid (Nakamura and Kadono, 1993) and mostly reproduce by vegetative means. However in different parts of the U.S., the presence of diploid and tr iploid plants has been reported within the monoecious form (Harlan et al., 1985). It has been suggested that hydrilla may be an endopolyploid plant due to the occurrence of different ploidy leve ls from plants within the same population

PAGE 27

14 (Langeland, 1989). Some species apparently have no consistent ploidy level in the developing root tissue and the presence of cells of differe nt chromosome numbers is common in roots of those species (Wardlaw, 1968). This appears to be the case with hydrilla. Research has shown the endopolypl oid nature of hydrilla (Sharma and Bhattacharya, 1956; Chaudri and Sharma, 1978) , and various combinations of diploid, triploid, and tetraploid cells has been repor ted in the same root tips (Langeland et al., 1992). There have also been reports of cytologi cal changes from diploid to triploid plants occurring many times independently in various hydrilla strains in Ja pan (Nakamura et al., 1998), and in the U.S. (Langeland, 1989). The more frequent occurrence of triploid plants than diploid in a hydrilla population may suggest an ecological advantage for triploid plants (Nakamura et al., 1998). Hydrilla Management in Florida Lakes Several methods have been investigated for the control of hydrilla, including mechanical, biological and chemical control. Management of hydrilla is difficult due to its rapid growth rate and pr olific turion production. Cultura l management schemes such as drawdowns to deplete tuber populations ha ve had limited success (Haller et al., 1976) and biocontrol agents, such as grass carp are unpredictable forms of control (Martyn, 1985). Drawdowns and desiccation have been shown to reduce tuber viability. However, the efficacy of the desiccation treatment depends of several factors. Although a reduction of 90% in the number of tubers was achi eved by this method, it did not completely eliminate the tubers from the sediments, a nd in some cases a drawdown of 12 months was not sufficient to reduce the number of tubers or to reduce viability (Doyle and Smart, 2001). It has been reported that if the dr awdown is not sufficient to kill the shoot

PAGE 28

15 biomass of hydrilla, it may stimulate tuber pr oduction, which translates into more plants sprouting and colonizing a gi ven area (Poovey and Kay, 1998). Several biological agents have been intr oduced for the control of hydrilla. The fungal pathogen Mycoleptodiscus terrestris , an endemic pathogen that causes a short duration disease on hydrilla without persistence in plant debris or plant tissue, has been used as a mycoherbicide (Shearer, 1998). In mesocosm studies this pathogen reduces hydrilla biomass in 80 days by up to 40% when applied alone, or by 93% when used in combination with fluridone treatments (Nelson et al., 1998). The use of this mycoherbicide in combination with the system ic treatment of lakes with the herbicide fluridone seemed to increase the susceptibil ity of hydrilla to the herbicide (Netherland and Shearer, 1996). A complete li st of insects evaluated and introduced into the U.S. as candidates for biological cont rol has been surveyed and published (Balciunas et al., 2002). Mechanical control of hydril la involves the removal of the above soil vegetative tissues and/or the tubers/turions by dredgi ng. Improvements in mechanical harvesting have resulted in bigger mach ines, capable of harvesting 1 hectare of hydrilla in approximately three hours at a cost of approximately $500/ha. However, the rapid growth of hydrilla requires 23 harvests per year. Thus mechanical removal is not practical for large lakes and is too expensive, costing as much as $2500 per hectare per year (SE-EPPC, 2005). Although much is known about hydrilla population biology, there are limited methods and technical developments to control this invasive weed species by mechanical or biological means (Simberloff, 2003).

PAGE 29

16 Hydrilla and Fluridone Due to high costs associated with the m echanical harvesting, the insufficiency of biocontrol agents, hydrilla management in aq uatic ecosystems relies heavily on chemical methods. Use of fluridone over time has become the basis of hydrilla control programs in Florida over the past 15-20 years. Discovere d in 1974 by Eli Lily Laboratories (Elanco) in Indiana (Banks and Merkle, 1978), fluridone received full EPA registration for aquatic use in 1985, with a maximum labeled rate of 150 ppb. Dosages of 8-12 ppb were commonly applied, although research has shown that fluridone concentrations of 4-7 ppb can control hydrilla if the dos age is maintained for several weeks (Van and Steward, 1985). By the late 1980's, hydrilla management plans were developed to treat large lakes at between 6 and 10 ppb fluridone, effectively co ntrolling large areas of hydrilla at costs of $250/hectare or less (Haller et al., 1990). Hydrilla was particularly susceptible to fluridone and could essentially be controlled with little or no non-target damage. The herbicidal activity of fluridone on hydrilla was very slow but lasted for 12 months or more with minimal impact to water qualit y. Consequently, by the year 2000, 80-90% of hydrilla control programs ($10+ million/year) in Florida utilized fluridone due to its environmental characteristics, longevity of control and relatively low costs. Fluridone affects plant tissue by decreas ing pigment levels (Maas and Dunlap, 1989). This is due to a decrease in carotenoid levels followed by a subsequent decrease in chlorophyll content due to ex cess photo-oxidative stress (Dev lin et al., 1978). Bartels and Watson (1978) demonstrated the mechanism of action of fluridone to be carotenoid synthesis inhibition with the actual site of action being the inhi bition of the enzyme phytoene desaturase. This enzyme catalyzes the conversion of phytoene to phytofluene in the terpenoid biosynthetic pathway (Mayer et al., 1989). Fluridone causes similar effects

PAGE 30

17 on hydrilla tissue, decreasing the chlorophy ll and carotenoid cont ent, but increasing anthocyanin content (Doong et al., 1993). Fluridone is absorbed by hydrilla from the surrounding water and apparently little or no tr anslocation occurs within the plant. Thus, this herbicide has to remain in contact with the plant for a relatively long time period (4060 days) in order to deplete the carbohydrate su pply within the plant. When plant death does finally occur, the entire pl ant including roots, rhizomes, et c. succumb to the lack of carbohydrates and reinfestation from tubers/turions usually takes one to two years to reach problematic levels. Development of Fluridone Resistance in Hydrilla In approximately 2000, aquatic plant mana gers in Florida began to observe populations of hydrilla that were no longer co ntrolled by previously lethal doses of fluridone (5 to 15 ppb). In some cases the plants appeared stunted but remained green and not bleached. They also observed much qui cker recovery of these populations from fluridone applications which suggested that certain populations of hydrilla were resistant to fluridone. Subsequent mana gement efforts showed that rates of 30 ppb or higher were needed to control these populat ions and resulted in a loss of selectivity to non-target desired species. Molecular Aspects of Fluridone Resistance Enzyme kinetics studies w ith recombinant phytoene desaturase revealed a noncompetitive inhibition with respect to th e substrate phytoene (Ogawa et al., 2001). NADP+ is a necessary cofactor for phytoene desaturase (Schne ider et al., 1997). A competition against the inhibitor was shown by the cofactor NADP+, suggesting an interaction of pyrrolidinones at the cofactor-binding site of phytoene desaturase (Ogawa et al., 2001). Most of the bi ochemical studies investiga ting the potency of phytoene

PAGE 31

18 desaturase inhibitors (fluridone, no rflurazon) have been made either in vivo with intact cells or in vitro with cyanobacterium cell extracts (May er et al., 1989; Ba bczinski et al., 1995; Sandmann, 2001; Sandmann and Mitchell, 2001). Therefore, only limited data on the inhibition of PDS from higher plan ts are available (L aber et al., 1999). However, recent advancements in molecular methods has allowed additional study on the resistance to bleach ing herbicides affecting PDS in vitro involving both cyanobacterial cell extracts (Navarro et al., 1995; Windhve l et al., 1997) and higher plants (Misawa et al., 1994; Albr echt et al., 1995, Ar ias et al., 2004). The genes encoding phytoene desaturases from cyanobacteria (Chamovitz et al., 1991), green algae and plants (Bartley et al ., 1991; Pecker et al ., 1992; Hugueney et al., 1992) have been isolated ( pds genes). Cyanobacteria and pl ants possess a PDS enzyme which catalyzes the first two dehydrogenati on reactions in the carotenoid synthesis pathway, producing carotene from phytoene as an end product (Sandmann, 1994). However, there is little homology between pl ant and bacterial PDS (Misawa et al., 1990). Albrecht et al. (1995) cloned cDNA from Capsicum annuum encoding an enzyme mediating desaturation of carotene to lycopene. Sequence comparison revealed 33-35% similarity with a previously cloned plant or cyanobacterial PDS. Wi ndhvel et al. (1994) showed that the strain Synechococcus PC 742-PIM8-BG1, which contained bacterial PDS (CRT-1) was resistant to the bleaching herbic ides inhibiting plant type PDS. A tobacco transformant, expressing a foreign bacter ial phytoene desaturase gene which is structurally completely unrelated to the plant-type enzyme, lacks the common binding site for many bleaching herbicides (Misawa et al., 1994). As such, this transformant was totally resistant to both fluridone and norflurazon.

PAGE 32

19 Recently, Michel et al. (2004) have been able to determine the whole hydrilla PDS gene sequence and demonstrated that the re sistance to fluridone in hydrilla is due to point mutations occurring at a single codon in the enzyme phytoene desaturase. So far, only three point mutations have been identi fied, each with potential for conferring a different resistance level. Th erefore, future studies shoul d be made to determine the presence and influence of each of these point mutations on the development of resistance. These studies may also provide a tool fo r predicting the susceptibility of hydrilla populations prior to herbicide appl ications (Netherland et al., 2002). Therefore, the rationale of this resear ch was to elucidate the cytological, molecular and physiological aspe cts of fluridone resistance, explore methods to prevent resistance in currently susceptible populat ions, and strategies to overcome currently existing fluridone resistance in hydrilla biot ypes. To accomplish these goals, the specific objectives of this research were: (1) to evaluate the biochemical differences among different hydrilla populations by co mparing changes in phytoene and -carotene levels as a function of hydrilla populati on (resistant vs. susceptible) and fluridone doses over time; (2) to determine growth and turion physiol ogy of fluridone resistant and susceptible hydrilla populations; (3) To ch aracterize different fluridone resistant hydrilla populations at the molecular level by is olating the phytoene desaturase (PDS) gene and determining the possible point mutations in the gene; and (4) to conduct cytological studies to determine the correlation between ploidy vari ation in different hydr illa populations and fluridone resistance.

PAGE 33

20 CHAPTER 2 BIOCHEMICAL DIFFERENCES AMONG DI FFERENT HYDRILLA POPULATIONS BY COMPARING CHANGES IN PHYTOENE AND -CAROTENE LEVELS AS A FUNCTION OF HYDRILLA POPULATION (RESISTANT VS SUSCEPTIBLE) AND FLURIDONE DOSES OVER TIME Introduction Hydrilla ( Hydrilla verticillata ) is an exotic, submersed, aquatic vascular plant that causes major problems in fresh water ecosyst ems of the United States. Discovered in Florida in 1959 (Blackburn et al., 1969), it has spread throughout the south-eastern United States and can be found as far west as California and as far north as Washington state and Maine. This monocotyledonous sp ecies is now the most abundant submersed aquatic plant in Florida and is one of the most serious aquatic weed problems in the Southern and Western U.S. Hydrilla is very competitive in Florida lakes due to the ability to grow under lower light conditions than na tive species (Van et al., 1978). Hydrilla has developed an inducible C4-acid cycle to allow growth in a dverse conditions, such as limiting CO2, high O2 concentration, high temperature and irradi ance (Bowes and Slavucci, 1984; Magnin et al., 1997). Therefore, hydril la can shift between C3 and C4-type photosynthetic mechanisms depending on environmental conditi ons (Magnin et al., 1997). This allows hydrilla to effectively compete and dominate over native species in shallow water and colonize deeper waters where native submersed species have been unable to survive due to low light conditions (Van et al., 1978) . Hydrilla possesses numerous means of vegetative reproduction including fragmentati on, root crowns, and specialized dormant

PAGE 34

21 buds called turions, (Langeland, 1996), which en ables it to spread very rapidly. Left unmanaged, this invasive plant can rapidly infest thousands of contiguous hectares depending on size and water depth, displacing native plant communities, fisheries, and interfering with human use (n avigation, irrigation, flood control, recreation, etc.), causing significant damage to the ecosystems (Colle and Shireman, 1980; Schmitz and Osborne, 1984; Bates and Smith, 1994). Consequently, much of the aquatic plan t research in the past 30 years in Florida has concentrated on the development of management programs for hydrilla. Several methods have been investigated for the control of hydrilla populations, including mechanical, biological and chemical control. Drawdowns and desiccation have been shown to reduce the tuber viability. However, the efficacy of the desiccation treatment depends on several factors. A lthough a reduction of 90% in the number of tubers was achieved by this method, it did not completely eliminate the tubers from the sediments, and in some cases a drawdown of 12 months were not sufficient to reduce the number of tubers or to reduce viability (D oyle and Smart, 2001). It has been reported that if the drawdown is not sufficient to kill the shoot biomass of hydrilla, it may stimulate tuber production, which translates into more plants sprouting and colonizing a given area (Poovey and Kay, 1998). Several biol ogical agents have been studied, but have failed to provide sufficient hydrilla cont rol. Improvements in mechanical harvesting have resulted in bigger machines , capable of harvesting 1 hectar e of hydrilla in an hour at a cost of approximately $500/hectare. Howe ver, the rapid regrowth of hydrilla often requires 2 to 3 harvests per year. Biocontrol of hydrilla with sterile grass carp in open

PAGE 35

22 systems is not practiced due to movement of the fish and the inability to remove grass carp after weeds are controlled (Martyn, 1985). Due to the high costs associated w ith mechanical harvesting and the unpredictability and insufficiency of biocontro l, hydrilla management in Florida has relied heavily on chemical methods (i.e., herbic ides). The contact herbicides (short-lived in the environment and kill plant tissue on contact) used for hydrilla control include diquat and endothall (Vandiver, 1999). These herbicides pr ovide good initial control but require two to three applications per year to keep hydrilla below nui sance levels. Costs associated with using these herbicides are between $500 and $1000 per hectare per application. In addition, contac t herbicides kill hydrilla ve ry quickly and treating large areas can result in reduced water quality, al beit, usually only for a short time period. Nevertheless, water quality problems such as algae blooms and lowered dissolved oxygen levels have to be considered when treating la rge areas with these contact-type herbicides (Westerdahl and Getsinger, 1988). Fluridone (1-methyl-3-phenyl-5-[3-(trifl uoromethyl) phenyl]-4 -(1 H)-pyridinone) is the only USEPA approved systemic compou nd registered for treatment of large water bodies. Fluridone was approved its use in aqua tic systems in 1986 and is highly effective for selective control of hydrilla (Doong et al., 1993). The wide spread use of fluridone is due to several factors including : low use rates, favorable nati ve plant selectivity, and slow activity (reduced oxygen depletion) . Hydrilla is controlled in large water bodies (> 100– 12 000 ha) by sustaining between 12 and 36 nm (4 g L-1) concentrations of the chemical fluridone (Sonar) in lake water fo r several weeks (Nethe rland and Getsinger, 1995; Fox et al., 1996). It also provides contro l of several other submersed aquatic weed

PAGE 36

23 species with application rates of 5 to 15 g L-1, although the USEPA registered maximum labeled rate is 150 g fluridone L-1 in aquatic system.1 The herbicide fluridone was discovered in 1974 by Eli Lily Laboratories (Elanco) (Banks and Merkle, 1978). It belongs to th e substituted tetrahydropyrimidinones class of herbicides which affect carotenoid synthesi s in plants. Carotenoids are polyunsaturated antioxidants that play an essential role in photosynthetic organisms (plants, algae, and cyanobacteria). In chloroplasts, carotenoids ac cumulate in the thylakoids, in association with the photosynthetic appara tus. These molecules dissipate excess light energy trapped by the antenna pigments, while participating in trapping light for photosynthesis, and protecting chlorophylls from photodegrada tion under high light intensities. The carotenoid biosynthetic pathway is an excellent target for herbicides because it is essential for plant development, but abse nt in animals. Severa l chemical classes of phytoene desaturase (PDS) inhibitors have been discovered, including pyridazinones, pyridinecarboxamides, phenoxy-butanamides (Dayan and Duke, 2003). Of the many enzymes involved in the formation of carote noids, PDS is the primary herbicide target site. PDS is a nuclear-encoded protein and has activity in the chloroplasts, the site of carotenoid synthesis (Bartley et al., 1991). Inhibition of these en zymes stops the synthesis of carotenoids in developing tissues and results in the di sappearance of chlorophylls, producing characteristically white foliage (Bger and Sandman, 1998). Consequently, these herbicides are often refe rred to as bleacher or bleach ing herbicides. However, only a few inhibitors of this pathway have been commercialized because most compounds lack sufficient crop selectivity. 1 SePRO Corp.1994. Sonar A.S. and Sona r Specimen Label. SePRO, Carmel, IN.

PAGE 37

24 Fluridone is a non-competitive inhibitor of the PDS enzyme in plants (Ogawa et al., 2001). Under normal conditio ns, phytoene does not accumula te in plant cells but is rapidly converted to the colore d carotenoids phytofluene and carotene by PDS. Fluridone blocks the synthesis of carotenoi ds, more specifically the formation of phytofluene from phytoene. The catalytic activit y of PDS appears to be rate-limiting and controls the remainder of the carot enoid pathway (Chamovitz et al., 1993). Mechanistically, PDS catalyzes the removal of tw o pairs of electrons (4 electrons total) to convert phytoene to carotene. Upon inhibition of th is enzyme, phytoene (a colorless carotenoid) accumulates. Many kinetic studies have shown that the inhibitors do not compete for the binding of phytoene on PDS (Sandmann and Mitchell, 2001). Only the dioecious form of hydrilla is found in Florida, with spread and reproduction limited to asexual means (subterran ean turions, axillary turions, fragments, and root crowns), the development of resist ance to herbicides was unexpected. However, in approximately 2000, aquatic pl ant managers in Florida bega n to observe populations of hydrilla that were no longer controlled by previously lethal doses (4 g L-1) of fluridone. In some cases the plants app eared stunted but remained green and not bleached. They also observed much quicker reco very of these populations from fluridone. Subsequent field and greenhouse studies c onfirmed that these populations of hydrilla required much higher rates of fl uridone to achieve control. In vitro assays by Michel et al. (2004) reported that mutations at the amino acid 304 of the PDS of hydrilla rendered this enzyme insensitive to the herbicide fluridone. Once resistance had been confirmed, field observations suggested that there was variability in the levels of resistance, and resistance was dynamic and would change as a function of fluridone pressure. Therefore, the

PAGE 38

25 rationale for this study was to further inve stigate this phenomenon to better understand the biology of resistance in hydr illa. Specifically the objectives were to 1) determine the differential response of hydrilla populations to fluridone; and to 2) determine the stability (in terms of fluridone response) of each resi stant population over time . This was achieved by monitoring pigment (phytoene and carotene) response to fluridone in different hydrilla populations over a period of one year. Materials and Methods Collection and Maintenance of Plant Material Five confirmed fluridone resistant hydrilla populations were collected from different Florida public lakes with a known fluridone a pplication history . These populations were arbitrarily designated as R1, R2, R3, R4, and R5 based on general response to field results. One population, term ed fluridone susceptible (S) was collected from a private pond in north central Florida th at has never been treated with fluridone. Plant samples were cleaned thoroughly and brought to the Weed Science Building in University of Florida, Gainesville. E ach population was grown in a separate 900 L concrete vault under controlled conditions w ith ambient sunlight in the absence of fluridone from Sept 20, 2004 to th e last fortnight of Sept 2005. Laboratory Evaluations Pigment content experiments were conduc ted on Sept 20, 2004 (at planting), Dec 20, 2004 (3 months after planting (MAP)), March 20, 2005 (6 MAP), June 20, 2005 (9 MAP), and Sept 20, 2005 (12 MAP). Hydrilla sh oot tips of 4 cm length were excised randomly from plants in each population and placed in 300 mL volumetric flasks containing 1% Murashige and Skoog (MS) medi a. Each flask contained three shoot tips from a hydrilla population Flasks were then placed on a rotary shaker with 200 mol cm-2

PAGE 39

26 s-1 light intensity at a14 hr light/10 hr dark cycle at 25C . Fluridone treatments were applied at the beginning of the experi ment as technical grade fluridone2 at a concentration of 0, 5, 10, 15, 20, 30, and 50 g L-1. After 14 days, tissues from all new growth were excised from the original shoot tip and analyzed for phytoene, carotene, and phytofluene content. At each time interval , the experiment was conducted twice with three replications in each run. Pigment Analysis Phytoene, and carotene pigment extraction an d quantification were conducted according to the protocol reported by Sprech er et al. (1998). Appr oximately 0.25-0.5 g of shoot tissue was ground in a pestle and mortar in liquid nitrogen and 5 mL of a freshly made solution of 6% (w/v) KOH in MeOH, and transferred to a 15 mL disposable glass test tube. The homogenate was then centrif uged 5 min at 2000 g. The supernatant was decanted into a fresh tube and mixed with 2 mL of petroleum benzin. The remaining pellet was discarded. The tube was then ca pped and shaken vigorously to thoroughly mix the samples. Samples were covered to avoid light and allowed to sit for approximately 30 min to yield a completely clear solution. Th e petroleum epiphase was then placed in a cuvette and sample absorbance (A) was measur ed spectrophotometrically at wavelengths of 287 nm ( cis phytoene), 347 nm (phytofluene), and 445 nm ( carotene), using petroleum benzin as a blank. Pigments were calculated based on the fresh weight of the tissue and the 2-mL epiphase. The ex tinction coefficients (E) of 1108 for cis phytoene or 2500 for -carotene (Sandmann and Bger, 1983) we re used in the following formula: 2 Supplied by SePRO Corp, Carmel, IN.

PAGE 40

27 g g-1 FW = ((A/(1 cm E) 2 mL/100 mL)/g FW) 106 Statistical Procedures Population and rate response experiments we re repeated twice in time (i.e., two runs) during each time interval (i.e., at pl anting, 3 MAP, 6 MAP, 9 MAP, and 12 MAP). Treatments were arranged in a completely randomized block (CRB) design with three replications. No significant differences were found between the two runs at all time intervals; therefore data were pooled (n= 6) within each interval. ANOVA (P< 0 . 05) was used to separate means, and results are pres ented as means with 95% confidence intervals (C.I.). Regression analysis was performed to fit phytoene and carotene response data for each population to fluridone dose. EC50 values for phytoene/ carotene (effective fluridone concentration to in crease/decrease the phytoene/ carotene content in hydrilla plant tissue by 50% over the untr eated control) were calculat ed for the different hydrilla populations. Data for phytoene, carotene, and phytoene: carotene ratio for each population as a function of fluridone dose ove r time is presented in Appendix A-F. Results and Discussion All hydrilla populations showed differ ential response to Fluridone. Initial response levels are presented in Figure 21. The susceptible (S) population showed an increase in phytoene and a decrease in carotene content when treated with 5 g L-1 fluridone. Higher concentrations of fluridone were needed in all the resistant populations to affect both these pigments and differen ces existed among all th e resistant populations in their response. R1 showed an increase in phytoene content with the application of 10 g L-1 fluridone, whereas a fluridone dose of 20 g L-1 was required to increase phytoene content in resistant populations R2, R3, and R4. Resistant population R5 was the least sensitive population, and only demonstrated an increase in phytoene content when treated

PAGE 41

28 with fluridone dose of 30 g L-1 or greater. carotene response followed a similar trend to phytoene in response to fluridone concentr ations for all populations (Figure 2-1). Application of fluridone in wheat and morningglory ( Ipomoea lacunosa L) increased phytoene to 72 g g-1 fresh weight (FW) and phytofluene to 5 g g-1 FW, from undetectable contents in untreated pl ants (Bartels and Watson, 1978; Duke et al., 1985). These changes in phytoene contents in hydrilla are co mparable to those reported for the terrestrial plants. P hytoene typically does not accumulate in plant cells but is rapidly converted to the colore d carotenoids phytofluene and carotene by PDS (Chamovitz et al., 1993). Inhibition of PDS enzyme by application of bleaching herbicides such as fluridone results in rapi d accumulation of this pigment in susceptible plants. A much lower level of increas e in phytoene was observed in all fluridone resistant populations compared the susceptibl e population. Lesser reductions in carotene content were also observed in all resistant popul ations. Doong et al. ( 1993) reported a carotene concentration of 207 g g-1 FW in susceptible hydrilla, with a reduction of 78% with 2week exposure to 5 g L-1 fluridone over untreated hydrilla plants. Response of hydrilla populations to flur idone at 3 month intervals (Dec 20, 2004 (3 MAP), March 20, 2005 (6 MAP), June 20, 2005 (9 MAP), and Sept 20, 2005 (12 MAP) are presented in Figures 2-2, 2-3, 2-4, and 2-5. Susceptible (S) hydrilla recorded increased/decreased phytoene/ carotene content when treated with 5 g L-1 fluridone at all time intervals. Although all fluridone resistant hydr illa populations behaved differently in response to fluridone concentrations over time, phytoene and carotene

PAGE 42

29 content followed almost similar trends in res ponse to varying fluri done concentrations in all populations (Figures 23, 2-4, 2-5, and 2-6). To better understand these relationships, we also calculated and statistically analyzed phytoene/ carotene ratio in all hydrilla populat ions during all the time intervals to determine whether this ratio was more i ndicative of their differential response to fluridone over time (App endix A-F). It was obser ved that this phytoene/ carotene ratio followed similar trends as phytoene response during all the time intervals and for all hydrilla populations (Tables 2-1 and 2-6). Regression analysis was performed in each population to quantify the relationship between fluridone concen tration and phytoene or carotene content at each time interval. EC50 values for phytoene (effective fluridone concentration required to increase phytoene content by 50% over control), and carotene (effective fluridone concentration required to decrease carotene content by 50% over control) we re calculated and are presented in Tables 2-1 and 2-2. The susceptible hydrilla population recorded no change in EC50 phytoene over time, with an effective fluridone concentration of 7.47 g L-1 and 7.56 g L-1 required to increase phytoene content by 50% over the un treated control at planting (Sept 20, 04) and 12 MAP (Sept 20, 05), respectively. R1, R 2, R3, R4, and R5 plants required 16.78, 20.39, 22.39, 22.5, 36.62, and 61.25 g L-1 fluridone to increase phytoene content by 50% over control at planting (Sept 20, 04), and 13.86, 17.39, 20.16, 30.57, and 57.29 g L-1 fluridone to have the same affect on phytoene content at 12 MAP (Sept 20, 05). Although all fluridone resistant hydrilla populations recorded a general decreasing trends in EC50 phytoene over time, the response was not significant (Table 2-1).

PAGE 43

30 EC50 carotene values were similar to EC50 phytoene in all the hydrilla populations and the response was also sim ilar. The susceptible hydrilla population showed no change in EC50 carotene over time, with a fluridone concentration of 8.89 g L-1 and 9.37 g L-1 required to decrease carotene content by 50% over control plants at planting (Sept 20, 04) and 12 MAP (Sept 20, 05), respectively. Resistant populations showed a response similar to EC50 phytoene. R1 and R2 whic h were less resistant to fluridone recorded reducti ons of 15% and 18.3% in EC50 carotene at 12 MAP than at planting. Conversely, the highly re sistant populations R3, R4, a nd R5 recorded reductions of only 6.1%, 2.6% and 6.7%, respectively in fl uridone concentration to elicit the same affect at 12 MAP than at planting (Table 2-8). Though the response in EC50 carotene was also failed to reach statistically signifi cant levels for R2, R3, R4, and R5 populations, there was reduction in EC50 carotene in R1 at 12 MAP than at planting. Differential response was observed in all the hydrilla popul ations, and there appeared to be a threshold value for flur idone concentration for each hydrilla population at which there was sharp increase/decrease in phytoene/ carotene. Susceptible hydrilla populations showed an e ffect on both phytoene and carotene content at exposure of 5 to10 g L-1 fluridone, whereas much higher doses were required in fluridone resistant hydrilla populations. These data confirmed th e differential resistan ce level to fluridone herbicide among hydrilla populatio ns in different lakes. Weed resistance is not unique. Herbicide-resistant weeds were first discovered in the United States in the late 1960s in a pine nursery where triazine herbicides had been used repeatedly. Factors that accelerate th e selection of resistant biotypes include repeated use of a single herbicid e in large areas, the lack of alternating different modes of

PAGE 44

31 action, high efficacy of the herbicide on the sens itive biotype at the rate used, and long residual herbicide activity (Maxwell and Mortimer, 1994; Volenberg et al., 2002). Fluridone herbicide has been the primary herb icide used in hydrilla control programs for over two decades. Fluridone is widely used to treat aquatic weeds in large areas of lakes and reservoirs. Concentrations of 4 g L-1 fluridone are able to control hydrilla when exposure is maintained for 60-90 or more days (Netherland et al., 1993; Sprecher et al., 1998). The asexual nature of hydrilla growth and reproduction should have precluded the resistance development. However, imposing a sustained selection pressure on a plant species like hydrilla which can grow 10 cm pe r day and produce an entire plant from a single node (Langeland, 1996), coupl ed with the sub-lethal do ses of fluridone over the past several years may have favored th e development of fluridone resistance. Although resistance has been reported and conferred in PDS from cyanobacteria (Chamovitz et al., 1993; Windhvel et al., 1994), hydrilla is the first higher plant to have developed resistance from a PDS-inhibiting herbicide (Mic hel et al., 2004). Recently, Walsh et al. (2004) reported a terrestrial weed Raphanus raphanistrum to have developed resistance to the PDS inhibitor diflufen ican. This weed deve loped resistance to diflufenican after only four a pplications of this herbicide with 16% of these populations surviving four-fold times the commercial application rate of diflufenican. The mechanism of fluridone resistance in hydrilla was reported recently as a result of one of three independent somatic mutations at the ar ginine 304 codon of the gene encoding phytoene desaturase, the molecula r target site of flur idone (Michel et al., 2004). A high rate of somatic mutations has b een described in hydrilla (Michel et al., 2004). These somatic mutations can perpetuate as hydrilla has a growth rate of as much

PAGE 45

32 as 100 mm per day (Langeland, 1996), and se veral means of vege tative reproduction, thereby further increasing the spread of fluridone resistant hydrilla populations.

PAGE 46

33 Figure 2-1. Regression analysis of fluridone susceptible and resistant hydrilla populations between phytoene or carotene content and fluridone doses at planting (Sept 20, 2004); ) phytoene content; ) carotene content; Data represents mean of two experiments with 3 replications ; Means values (n= 6) are presented with 95% confidence interval bars (Standard error x 1.96).

PAGE 47

34 Figure 2-2. Regression analysis of fluridone susceptible and resistant hydrilla populations between phytoene or carotene content and fluridone doses 3 MAP (Dec 20, 2004); ) phytoene content; ) carotene content; Data represents mean of two experiments with 3 replications; M eans values (n=6) are presented with 95% confidence interval bars (Standard error x 1.96).

PAGE 48

35 Figure 2-3. Regression analysis of fluridone susceptible and resistant hydrilla populations between phytoene or carotene content and fluridone doses 6 MAP (Mar 20, 2005); ) phytoene content; ) carotene content; Data represents mean of two experiments with 3 replications; M eans values (n=6) are presented with 95% confidence interval bars (Standard error x 1.96).

PAGE 49

36 Figure 2-4. Regression analysis of fluridone susceptible and resistant hydrilla populations between phytoene or carotene content and fluridone doses 9 MAP (Jun 20, 2005); ) phytoene content; ) carotene content; Data represents mean of two experiments with 3 replications; M eans values (n=6) are presented with 95% confidence interval bars (Standard error x 1.96).

PAGE 50

37 Figure 2-5. Regression analysis of fluridone susceptible and resistant hydrilla populations between phytoene or carotene content and fluridone doses 12 MAP (Sep 20, 2005); ) phytoene content; ) carotene content; Data represents mean of two experiments with 3 replications; M eans values (n=6) are presented with 95% confidence interval bars (Standard error x 1.96).

PAGE 51

38 Table 2-1. EC50 phytoene (Fluridone dose; g L-1) concentrations for fluridone susceptible (S) and resistant (R1-R5) hydrilla populations as a function of time. Population Sept 20, 2004 (At planting) Dec 20, 2004 (3 MAP) March 20, 2005 (6 MAP) June 20, 2005 (9 MAP) Sept 20, 2005 (12 MAP) S 7.47.331 8.49.43 8.86.76 8.6.76 7.56.90 R1 16.78.29 15.03.02 16.05.82 13.02.72 13.86.88 R2 20.39.78 19.6.35 19.54.88 17.41.02 17.39.08 R3 22.50.99 22.33.86 21.72.35 21.16.12 20.16.96 R4 36.62.11 33.48.25 35.52.2 30.23.78 30.57.84 R5 61.25.31* 64.49.84*62.60.8* 62.26.9* 57.29.7* MAPMonths after planting; 1Mean values followed by 95% confidence interval values (Standard error x 1.96); * values calc ulated by extrapolating the curve.

PAGE 52

39 Table 2-2. EC50 carotene (Fluridone dose; g L-1) concentrations for fluridone susceptible (S) and resistant (R1-R5) hydrilla populations as a function of time. Population Sept 20, 2004 (At planting) Dec 20, 2004 (3 MAP) March 20, 2005 (6 MAP) June 20, 2005 (9 MAP) Sept 20, 2005 (12 MAP) S 8.89.531 9.52.45 9.26.21 9.85.23 9.37.52 R1 18.01.47 17.36.58 16.35.57 16.69.58 15.31.12 R2 21.19.76 19.8.90 19.32.22 17.6.99 17.31.09 R3 25.26.47 24.44.13 26.24.98 23.9.25 23.72.20 R4 42.24.31 45.39.45 42.75.88 42.64.64 41.10.10 R5 62.70.58* 64.27.9*64.84.8* 57.14.55* 58.50.70* MAPMonths after planting; 1Mean values followed by 95% confidence interval values (Standard error x 1.96); * values calc ulated by extrapolating the curve.

PAGE 53

40 CHAPTER 3 GROWTH AND REPRODUCTIVE PHYSIOLOGY OF FLURIDONE SUSCEPTIBLE AND RESISTANT HYDRILLA POPULATIONS Introduction Hydrilla is a submersed vascular plant which has become one of the most troublesome aquatic weeds in Florida. Desp ite extensive management practices during the last three decades, this exotic macrophyte remains a dominant weed species throughout the Southeastern United States. Th ere are dioecious and monoecious forms of this species and both are present in the United States (Madeira et al ., 1997; Madeira et al., 1999). In the late 1950s, the female form of the dioecious hydrilla was brought from Sri Lanka to Missouri, from where it was sent to Tampa Bay, Florida (Schmitz, 1991). This dioecious female plant spread aggressively throughout Florida, then west to Texas and California (Yeo and McHenry, 1977). Its range also extends north into Georgia, Tennessee, south and North Carolina (Yeo et al., 1984). The monoecious form of hydrilla was first identified in North Carolina in 1980 (Langeland and Schill er, 1983). A second, separate introduction of a m onoecious plant was reported in Delaware in 1976 and it spread into Virginia, Maryland, and North Ca rolina. A monoecious plant has also been reported in California (Ryan a nd Hommberg, 1994). In the overa ll distribution of hydrilla in the world, the monoecious form is more pr evalent in warmer tropical areas, while the dioecious form occupies lower temperat ure areas (Cook and Lnd, 1982). However, the current distribution for both monoecious and dioecious biotypes in North America is contrary to this observation, with dioecious biotype dominating the Southeastern United

PAGE 54

41 States, and monoecious hydrilla present mostly in central and northern U.S. The two biotypes have several differences in terms of vegetative growth habit and asexual propagule production. Research has shown great er production of subterranean turions (tubers) than axillary turions in the dioecious biotype (H aller and Sutton, 1975; Spencer et al., 1987), and monoecious plants produce 20 to 30% of their tota l number of turions aboveground i.e axillary turi ons (Anderson, 1986). McFarland and Barko (1999) reported that monoecious hydrilla might be well adap ted to high temperatures, making it highly feasible of both biotypes to overlap in the Southern states of the U.S. Once hydrilla becomes established in an aquatic ecosystem, it spreads rapidly through various means of vege tative reproduction such as fr agmentation, stolon growth, or production of subterranean turions (tube rs) and axillary turions. Viable seeds are produced by the monoecious biotype. Dio ecious hydrilla, which is the only biotype present in Florida, develops only pistill ate flowers (Langeland and Smith, 1984), and there is no report of seed production by hydrilla in Florida. Hydrilla regrows in the spring from the root crowns and/or subterranean turions in or on the hydrosoil as the water temperature begins to increas e (Haller et al., 1976). During the initial phases of growth, the shoots quickly reach the water surface fo rming a dense canopy. This surface mat of hydrilla shoots effectively absorbs nearly 95% of the available light, limiting competition from other native species. Rapi d hydrilla growth and expansion is favored by its low light and CO2 compensation points, reduced photorespiration due to a C4-like photosynthetic mechanism and prolific reproductive capacity (Van et al., 1975; Holaday et al., 1983; Magnin et al., 1997).

PAGE 55

42 Hydrilla produces two distinct types of turions i.e., ax illary turions (occasionally the tip of a stem swells and forms this struct ure) and subterranean tu rions (at the tips of the positively geotropic stems or rhizomes which extend into the hydrosoil). These stems generally arise from the crown or rooted por tion of the hydrilla plan t and penetrate 5 to 15 cm deep into the hydrosoil before the ti ps begin to swell an d subterranean turion formation occurs (Spencer et al., 2001) . The ability of hydrilla to reproduce by subterranean turions (tubers) is the greatest restraint to the control of this species. Subterranean turions are simila r structures to tubers but are morphologically less complex and almost exclusively found in aqua tic macrophytes such as hydrilla ( Hydrilla verticillata ) (Thullen, 1990) a nd giant duckweed ( Spirodella polyrhiza ) (Smart and Fleming, 1993). Axillary turions are comprised of several densely packed nodes (Yeo et al., 1984). The outer portion of an axillary turion is made of leaf tissue, which swells and envelops the terminal apex, giving the scal e like appearance. These leaves accumulate carbohydrates and an abscission layer forms at the base of the lowermost leaves, thereby detaching the axillary turion from the mo ther plant (Appenroth and Bergfeld, 1993). Hydrilla forms subterranean and axillary turions during the late summer and early fall in Florida (Haller, 1976; Miller et al., 1993). Van et al. (1 978) reported that in dioecious hydrilla, tu rion production is photoperiodic in nature, with turions produced only under short day conditions. Monoecious hydri lla produces subterranean and axillary turions throughout the year (Steward, 1997). Influence of photoperiod on vegetative propagule production in hydrilla has been widely studied (S teward, 1993; Steward, 1997; Steward, 2000). The classic phytochrome-me diated and photoreversible system is involved in the initiation of turion production in hydril la, with red light (660 nm)

PAGE 56

43 stimulation and far-red (750 nm) repressi on (Thakore et al., 1997). Thullen (1990) reported that dioecious turion production from floating plant fragments was influenced by daily temperature ranges, the source of the pl ants, the length of time the plants were in the study, and aeration. Pieterse et al. (1984) suggested that tu rion formation is stimulated by low levels of nitrogen and phosphorous in th e water. Free floati ng plants would be much more subject to this nutrient stress as compared to rooted plants. However, Thullen (1990) concluded that turion production was not stimulated solely by low levels of nitrogen and phosphorous, but required an adeq uate daily temperature range (17 to 27C) and photoperiod. Production of plant propagul es has been shown to increase the competitive abilities of a pl ant (Grace, 1985). Spencer et al. (1987) reported that subterranean and axillary turi ons represent different survival strategies, with axillary turions better suited for dispersal and po ssible occupation of non-vegetated areas where they are likely to face little competition. In c ontrast, subterranean turions are primarily responsible for long term pe rsistence in a given area. Hydrilla has a very rapid growth rate. A hydrilla colony, originating from a single turion can expand radially at a rate of 4 cm d-1, with an average production of 1 new ramet m-2 d-1 (Madsen and Smith, 1999). Root crowns develop stolons (horizontal aboveground shoots) that extend into the area surrounding the pare nt plant and establish new plants (Madsen and Smith, 1999). For dioecious hydrilla, most of the colony expansion is by stoloniferous growth, while the spread from fragmentation is only 0.02 ramets m-2 d-1 (Madsen and Smith, 1999). There have been reports of 1,250 to 1,976 tubers per m-2 being produced by dioecious hydrilla within a period of four months on various Florida Lake sediments (Sutton and Portier, 1995). H ydrilla can grow from the substrate to the

PAGE 57

44 water surface and reach up to 15 m in length (Langeland, 1990), the stems can branch and have leaf whorls (nodes) every 11-12 mm in dioecious hydrilla (Ryan et al., 1995), and each node can regenerate a new plant (Haller, 1976). Hydrilla has several organs for storage of carbohydrates, e.g., turions, stol ons, stems and root crowns (Madsen and Owens, 1998). Collectively, these growth and specia lized reproductive characteristics make hydrilla a formidable aquatic weed. This sp ecies has rapidly spr ead over thousands of hectares, displacing native plant communities and causing significant ecosystem damage. Uncontrolled, hydrilla forms dense monoc ultures reducing native plant diversity, negatively affecting other plan t and fisheries populations a nd interfering with human use (navigation, irrigation, flood cont rol, recreation, etc.). Conseq uently, much of the aquatic plant research in the past 30 years in Flor ida has concentrated on the development of management programs for hydrilla due to it s rapid spread and dominance of aquatic ecosystems. Several biological agents have been studied for hydrilla control; however, these have failed to provide sufficient contro l. Improvements in mechanical harvesting have resulted in bigger machines , capable of harvesting 1 hectar e of hydrilla in an hour at a cost of approximately $500/hectare. Howe ver, the rapid regrowth of hydrilla requires two to three harvests per year . Biocontrol of hydrilla with sterile grass carp in open systems (Kissimmee Lakes, etc.) is not practi ced due to movement of the fish and the inability to remove grass carp when overstock ed or plants are controlled (Martyn, 1985). Due to the high costs associated with mechan ical harvesting and the unpredictability and insufficiency of bio control, hydrilla mana gement in Florida has relied heavily on chemical methods (i.e., herbicides). Contact aquatic herbicides such as complexed

PAGE 58

45 copper, diquat, and endothall can be used to control hydrilla; however, fluridone is the only USEPA approved systemic herbicide that can be used for treatment of large water bodies. Fluridone herbicide has been the prim ary means of hydrilla control programs in Florida over the past 20 to 25 years. Hydrilla is controlled in larg e water bodies (> 100– 12 000 ha) by sustaining between 12 and 36 nm (4 g L-1) concentrations of fluridone in lake water for several weeks (Netherland and Getsinger, 1995; Fox et al., 1996). Only the dioecious form of hydrilla is found in Florida, with spread and reproduction limited to asexual means (subterran ean turions, axillary turions, fragments, and root crowns). Therefore, the developm ent of resistance to herbicides was not expected. During last three to four years, aquatic plant managers in Florida began to observe populations of hydrilla that were not being effec tively controlled by previously lethal doses of fluridone (4 g L-1). In some cases the plants appeared stunted, but remained green and not bleached and these pl ants exhibited a much quicker recovery from fluridone treatments. These observations ultimately lead to the discovery of fluridone resistant hydrilla. W ith the continuous spread of resistant hydrilla populations in many different lakes in Flor ida during last three to four years, much of the current research has concentrated on understanding th e mechanism of fluridone resistance in hydrilla and to develop alte rnative control practices for this noxious weed. Recently it has been determined that the resistance of fluridone herbicide is due to a point mutation in hydrilla (Michel et al ., 2004), though the difference in levels of resistance in different hydrilla populations having the same muta tion is not understood.

PAGE 59

46 We also reported that resistan ce factors in resistant biotypes varied from two-fold in R1, to as high as seven-fold in R5, indicating a wide range of resi stance (Chapter 2). Many studies, largely with terrestrial speci es have shown decrease in fitness in resistant plants compared to susceptible plant in the absence of the selective agent such as herbicide. Fitness measures describe the pot ential evolutionary su ccess of a genotype. It may be defined as the reproductive success or the proportion of genes an individual leaves in the gene pool of a population, with the most fit leaving the greatest number of offspring. Bergelson and Purri ngton (1996) reported that 50% of the published studies on resistance to herbicides and diseases result ed in significant reduction of the fitness of resistant plants. Fitness measures normally involve evaluation of vegetative and reproductive growth. Bergelson (1996) reporte d inferior in reproductive capacity in acetolactate synthase (ALS) resistant Arabidopsis thaliana plants, resulting in 34% reduction in seed production compared to sus ceptible plants. Nume rous studies have reported that triazine herbicide resistant biotypes have lower phot osynthetic rates, CO2 fixation, and lower levels of electron transport than suscep tible biotypes (Radosevich and Holt, 1982; Ahrens and Stoller, 1983). Re duced competitiveness and productivity in triazine herbicide resistant bi otypes has been reported for Amaranthus powelli (Radosevich, and Holt, 1982), Poa annua (Bulcke et al., 1985) and Brassica rapa (Mappelback et al., 1982). Observations by re searchers and aquatic plant managers also suggest that hydrilla from fluridone resistan t and fluridone sensitive populations vary in their ability or capacity to produce turions, with resistant populations having a lower growth rate and producing a fewer number of turions than the susceptible hydrilla populations (Puri et al., 2004).

PAGE 60

47 Knowledge of growth and physiology ( phenology) is a basis of developing successful management practices for a weed species. Physiological studies have been improved the control of cattail ( Typha latifolia ) (Linde, 1976). Similarly, research on hydrilla growth and reproductive physiology may lead to development of improved means to manage this noxious weed speci es in the southeastern United States. Therefore, the objective of this experime nt was to investigat e possible phenotypic differences in growth and turion physiology be tween fluridone susceptible and resistance hydrilla populations. Materials and Methods Five confirmed fluridone resistant hydrilla populations were collected from different Florida lakes with known fluridone application histories . These populations were designated as R1, R2, R3, R4, and R 5. One population, termed fluridone susceptible (S) was collected from a private pond in north central Florida that ha s never been treated with fluridone. Plant samples were cleaned thoroughly and brought to the Weed Science Building in University of Florida, Gainesville. Eight shoot apices (10-12 cm length) we re excised from each hydrilla population, and planted in 3 L pots filled with commercial potting soil1, amended with 5 g of the slow-release fertilizer2 . A 2 cm-deep sand layer was placed over the potting soil to prevent floating and turbidity. Pots were then placed in 950 L vaults (219 cm long x 76 cm wide x 64 cm deep), with a separate vault for each different hydrilla population. Screens were fitted over the tanks to preven t contamination. Hydrilla populations were 1 Greenleaf Product Inc., Haines City, FL. 2 Osmocote; 15-10-15.

PAGE 61

48 grown under natural atmospheric conditions with ambient sunlight and day length in the absence of fluridone herbicide from Sept 20, 2004 to the last fortnight of Sept, 2005. An immersion heater with a thermostat to k eep water temperature falling below 0C was placed in each vault during winter mont hs (Nov 2004 to Feb 2005) to prevent hydrilla tissue injury. Constant aeration was provide d in each vault to maintain adequate CO2 and oxygen and to prevent algal growth. Four pots were harvested randomly from each vault (i.e., population) at one month intervals from Oct 20, 2004 (1 month after planting) through Aug 20, 2005, with a total of 11 harvests during the entire study. Plan ts were separated into shoot tissue, root crowns, rhizomes, subterranean turions and axillary turions. Hydrilla shoot length was recorded by measuring the length of the longest shoot from hydr illa plants from each of the four pots. Total plant biomass wa s determined along with the number of inflorescences, rhizomes and subterranean turions/plant. Shoo t elongation rate was calculated as increase in shoot length per day. Relative growth rate (RGR) was calculated using the following equation (Evans, 1972): RGR = (ln TW2 ln TW1)/(T2 T1) Where T1 = time 1, T2 = time 2; TW2 and TW1 are total biomass per plant during time T1 and T2, and ln is the natural logarithm. All data were statistically analyzed usi ng standard procedures (ANOVA) to test for differences in hydrilla populations for all growth and reproductive parameters. Means were separated using Fisher’s least significant differences (LSD) at the 5% probability level.

PAGE 62

49 Results and Discussion Shoot Length Hydrilla populations differed from one a nother in shoot length for nine of the eleven harvests (Table 3-1) . All hydrilla populations grew quickly following planting with populations R1, R2, R4, a nd R5 recording greater shoot length than populations S and R3 one month after planting (Oct 2004). As expected, hydrilla gr ew slowly from Nov 2004 to Mar 2005, due to the lower temperatur es and shorter day lengths during these months. No differences were observed in shoot length between the populations during Nov and Dec 2004, but from Jan 2005 to Apr 2005 the resistant hydrilla populations R1, R2, and R3 were growing at the same rate as the susceptible hydrilla (S). During this same time frame, resistant populations R4 and R5 showed significantly greater shoot length than the susceptible population. Maximu m rate of increase in shoot length for hydrilla populations except R5 was reco rded from Apr 2005 to May 2005 which was likely due to an increase in temperatur e and photoperiod (McFarland and Barko, 1999). During this one month period, hydr illa populations S, R1, R2, a nd R4 grew at a rate of 0.5, 0.33, 0.33, and 0.4 cm d-1, respectively. The R3 popul ation recorded a shoot elongation rate of 0.83 cm d-1 during this period. These data are contradicting to previous shoot length data, where initia lly this population was statisti cally inferior to all other hydrilla populations. R3 population was also significantly higher than hydrilla populations S, R1, and R2 during the follo wing growth period from May 2005 to Aug 2005 (Table 3-1). R5 recorded maximum increase during a period from May 2005 to June 2005 with a rate of 0.7 cm d-1. Shoot length is an important index of plan t growth, particularly in aquatic species such as hydrilla, and its measurement is ofte n used to monitor the effect of different

PAGE 63

50 biotic and abiotic parameters on growth of a plant species (Evans, 1972). Although hydrilla populations differed in shoot elongation during majority of the harvest periods, all fluridone resistant populations were growing either at the same rate or significantly better rate than the susceptible hydrilla, indi cating no decrease in growth rate due to development of herbicide resi stance. Moreover, populations R4, and R5 which were the higher in fluridone resistance le vels than other hydrilla popula tions (Chapter 2), recorded higher shoot length than the su sceptible (S) and less resist ant hydrilla populations (R1 and R2) during most of the growth periods (Table 3-1). Shoot and Root Biomass Significant differences were observed am ong the different hydrilla populations for shoot biomass (Table 3-2). All hydrilla popula tions showed a steady increase in shoot biomass over time, although from Oct 2004 through Apr 2005, the shoot biomass increased at much slower rate. During this winter period, the resi stant populations R1, R2, R4, and R5 were at par with the sus ceptible hydrilla S in production of shoot biomass. Resistant population R3 was lowe r in production of shoot biomass than susceptible and other resistant populations. The lower shoot biomass R3 can directly correlated to the lower shoot elongation ra te evident from Table 3-1. During the remaining growth period (Apr 2005 to Aug 2005), no differences were observed in the biomass of fluridone resist ant and susceptible hydrilla populations (Table 3-2). All hydrilla populations reco rded a steady increase in production of root biomass from Oct 2004 to Mar 2005, and no differen ce was observed among different populations throughout the study (Table 3-3). During the warmer months, hydrilla shoot s rapidly elongated to the surface to maximize light absorption, resulting in an increase in shoot biomass. From Apr 2005

PAGE 64

51 through Aug 2005, a rapid increase in shoot biomass was observed for all populations. Populations S, R1, R2, R3, R4, and R5 produced 102.8, 97.7, 94, 112.8, 104.6, and 118 g shoot biomass per plant during the period from Apr 2005 to Aug 2005, which was approximately 60% of the total shoot bioma ss production during the entire growth period by these populations (Table 3-2). Biomass a llocation studies of the dioecious hydrilla show this plant can a ccumulate up to 1200 g m-2 dry weight of a boveground shoot tissue between July and October (Madsen and Owe ns, 1998). These researchers also reported accumulation of 200 g m-2 shoot dry weight during winter months and as high as 600 g m-2 shoot dry weight during summer months. R oot crowns are major storage organs for the hydrilla plant to over-wint er or survive during unfavorable environment conditions. Root biomass production in all hydrilla popul ations remained consistent during the warmer months from Apr 2005 to Aug 2005, a nd showed a steady increase during winter months after planting from Oct 2004 to Ma r 2005. Data showed resistant hydrilla populations were either significan tly better or at par with su sceptible hydrilla in growth rate (evident from shoot and root biomass production). This indicates there is no fitness penalty with respect to growth for fluridone resistance. Turion Production Hydrilla populations, except R3, initiate d subterranean turion production within one month of planting and there were diffe rences among populations in turion production throughout the study (Table 3-4). Populations R4 and R5 produced greater numbers of subterranean turions than susceptible hydr illa (S) during Oct and Nov 2004. During the remainder of the study period, there were no significant differences among fluridone susceptible and resistant hydrilla populations except R3, which produced lower subterranean turions (Table 3-4). As expect ed, populations in this study formed turions

PAGE 65

52 short day lengths. Subterra nean turion production under s hort day conditions has been well documented for dioecious hydr illa (Haller at al, 1976; Va n et al., 1978; Spencer and Anderson, 1996; Steward and Van, 1987; Steward, 1997; Steward, 2000). Many aquatic plants produce specialized pr opagules in order to survive conditions that are unfavorable for growth and to ensu re vegetative reproduc tion (Sculthorpe, 1967). Subterranean turions provide a major surviv al strategy for hydrilla. There have been reports of 1,250 to 1,976 tubers per m-2 being produced by the dioecious hydrilla within a period of four months during winter on va rious Florida Lake sediments (Sutton and Portier, 1995). However, in other areas, up to 2,812 tubers m-2 were produced by the dioecious hydrilla during winter (Sutton et al., 1992). In this study, maximum turion production was observed during months of Jan 2005 through Mar 2005 for all hydrilla populations and turion production ce ased practically thereafter (Table 3-4). Miller et al. (1993) also reported that in dioecious hydrilla, turion pr oduction began under short day conditions in September, decreased during cold months of winter a nd increased again in spring, and ceased during June through August. Inhibition of turion production has also been reported in hydrilla at 35C, presumably due to meta bolic losses associated with increased respiration rate s (Barko and Smart, 1981). Although we examined the plants for axillary turions, none were observed throughout this experiment. Axilla ry turion production is often considered to be quite low on rooted dioecious plants, and similar results have been reported by Madsen and Smith (1999). Miller (1993) also showed detach ed plants produce axillary turions and no subterranean turions.

PAGE 66

53 Production of Axillary Branches Production of axillary branches in this study followed a similar trend as all other growth parameters and turion productivity in the different hydrilla populations. Resistant hydrilla R3 produced significantly lower number of axillary branches per plant than other hydrilla populations (S, R1, R2, R4, and R5) during the ea rly growth phase from Oct 2004 through Mar 2005. During the remainder of study period, no significant differences were observed among R3 and other hydrilla populations except R5, which produced significantly greater number of axillary branches pe r plant (Table 3-5). In our study, there was no disturbance to cause breakage of hydrilla stems (e.g., boat traffic, wave action, or other physical in juries to the plants growing in lakes under natural conditions). This might also have resu lted in reduced branchi ng rate in hydrilla. Flower Production Flowering in hydrilla was observed in all hydrilla populations starting from Oct 2004 and during this period, there was no di fference among hydrilla populations in flower production (Table 3-6). Little research has been conducted on the flowering aspects of hydrilla. Madsen and Owens (1998) observed flowering in hydri lla in late September prior to turion formation, while Yeo et al. (1984) also repor ted flowering in hydril la from September through November. Although all populations responded similarly initially, R3 was producing higher number of flowers per plant than all other hydril la populations during Nov 2004. Flower production ceas ed in all hydrilla populations (except R3) in Dec 2004 and no flower was observed during the rema ining growth period until Aug 2005 when they resumed flowering. On the other hand, R3 produced flowering from Oct 2004

PAGE 67

54 through Jan 2005, ceased during warmer months and resumed during Aug 2005 (Table 36). Total Biomass and Relative Growth Rate Hydrilla populations recorded a steady increase in total biomass production over time. There were no differences among th e populations during Oct 2004. Thereafter, during Oct 2004, Nov 2004, Dec 2004, Jan 2005, Feb 2005, Mar 2005 and Apr 2005, resistant population R3 had lower total biom ass production than susceptible hydrilla (S). However, other fluridone resi stant hydrilla populations were at par with susceptible hydrilla in biomass production (Table 3-7) . Maximum biomass production was observed in R5 during later growth stages from Mar 2005 through Aug 2005 which was the most resistant hydrilla population in our study (Chapter 2). During May 2005 through Aug 2005, there were no differences in biom ass production among susceptible and the fluridone resistant hydrilla populations. Maxi mum biomass was produced during months of Apr 2005 through Aug 2005 with populations S, R1, R2, R3, R4 and R5 producing 111, 104, 126, 138, 129, and 150 g total biomass per plant between Apr 2005 to Aug 2005. This reflected 50%, 50%, 60%, 65%, 60%, and 63% of the total biomass production for the entire growth period by th ese populations, respectiv ely (Table 3-7). These results are similar to previ ous research (Van et al., 1978). We also calculated relative growth ra tes of hydrilla populations for all time intervals during our study (Figure 3-1). Relati ve growth rate (RGR) is defined as the amount of new plant biomass that is produced during an interval of time by an initial amount of biomass that serves as the invest ment "capital" at th e beginning of the time period. It is considered a measure of grow th efficiency. Maximu m RGR was observed in susceptible hydrilla duri ng the initial growth pe riod of Oct through Nov 2004.

PAGE 68

55 Susceptible hydrilla grew at a rate of 11.7 mg g-1 d-1 during this period, while a RGR of 7.3, 8.9, 3.7, 1.2 and 3.8 mg g-1 d-1 was recorded in resistant hydrilla populations R1, R2, R3, R4 and R5, respectively (Figure 3-1). Higher RGR during this period was due to rapid shoot elongation after planting in Sept 2004. However, R4 was much slower in growth as evident from significantly lower RGR during this period than other hydrilla populations. Thereafter, duri ng the winter months from Nov 2004 through Feb 2005, the RGR of all hydrilla populations decreased and no differences were observed among susceptible and resistant hydrilla populations. RGR in creased again during March coinciding with increased ambient temperatur e and day length in all hydrilla populations with maximum RGR of 7.8, 6.1, 8.6, 7.0 mg g-1 d-1 between Apr 2005 and May 2005 for hydrilla populations S, R1, R 2, and R4, respectively. Resistant hydrilla R3 and R5 grew at much higher rates during this partic ular period with RGR of 14.0 and 10.4 mg g-1 d-1 were recorded for R3 and R5, respectively (Fig ure 3-1). Thereafter, re lative growth rate decreased from June through Augus t for all hydrilla populations. Phenology of hydrilla growth in the Southern region of the U.S. exhibits an active growth period during the summer months, w ith a period of senescence during winter (Madsen and Owens, 1998). This was eviden t from the reduced RGR during winter months and a much higher RGR during warm er months of Marc h through May during our study. A decrease in RGR thereafter from May through August, 05 could be attributed to the much higher mean ambient te mperatures of 35C, due to metabolic losses associated with increased respiration rates. Collectively, all the fluridone resistan t hydrilla populations showed comparable rates of growth and reproduction to fluri done susceptible hydrill a which indicates no

PAGE 69

56 deleterious effect due to development of flur idone herbicide resistance. However, hydrilla population R3 generally showed more differe nces in growth compared to all other hydrilla populations. Its early growth rate was slow as evident from lower shoot elongation rates, and lower shoots biom ass production from Oct 2004 to Mar 2005. However, from Apr 2005 through Aug 2005, it reco rded a very high relative growth rate and was growing at par with all other hydri lla populations. The population R3 was also lower in the production of subterranean tu rions and axillary branches. In addition, flowering extended two months longer in R3 compared to the other hydrilla populations, including the susceptible populat ion. While growth of R3 was different than the other hydrilla populations, overall those differences were not large or consistent among the resistant populations. These data suggested that there are no deleterious effects on growth and reproductive physiology due to deve lopment of fluridone resistance. Michel et al. (2004) suggeste d that fluridone resistant biotypes might be equally competitive against the wild-type hydrilla and may persist as the dominant forms in lakes. Aggressive spread of hydrilla in aquatic ecosy stems, and the evolution of resistance to fluridone in Florida may forecast signifi cant and long-lasting eco logical and economic problems throughout the Southern states of the USA.

PAGE 70

57Table 3-1. Shoot length (cm) in fluridone suscep tible (S) and resistant (R 1-R5) hydrilla populations. Hydrilla Population Oct 2004 Nov 2004 Dec 2004 Jan 2005 Feb 2005 Mar 2005 Apr 2005 May 2005 Jun 2005 Jul 2005 Aug 2005 S 42b 52 54 55ab 57ab 58ab 61b 76b 84bc 89b 93ab R1 48a 50 51 56ab 57ab 58ab 62b 72b 77c 85b 88b R2 50a 51 51 56ab 57ab 58ab 62b 72b 77c 85b 88b R3 42b 46 50 52b 55b 56b 60b 85a 93a 96a 99a R4 51a 52 53 57a 60a 62a 66ab 78b 83bc 84b 85b R5 50a 51 55 58a 61a 64a 70a 72b 93a 94a 95a LSD (0.05) 5.9 NS NS 4.3 4.5 7.8 6.0 6.3 8.7 7.5 8.7

PAGE 71

58Table 3-2. Shoot biomass (g per plant) in fluridone su sceptible (S) and resistant (R1-R5) hydrilla populations. Hydrilla Population Oct 2004 Nov 2004 Dec 2004 Jan 2005 Feb 2005 Mar 2005 Apr 2005 May 2005 Jun 2005 Jul 2005 Aug 2005 S 7.8ab 11.2a 12.1a 12.3a 12.6a 13.3ab 13.9ab 18.8 21 23.5 25.6 R1 8.6a 10.8a 11.7a 12a 12ab 12.5ab 13.7ab 17.3 19.8 22.9 24.3 R2 8ab 10.9a 10.5ab 10.9ab 11ab 12ab 12.8b 17 18.5 21.9 23.5 R3 6.1b 6.9b 8.1b 9.2b 9.7b 11.6b 12.5b 21.4 24.6 26.4 27.8 R4 9.2a 9.6a 11ab 12.4a 12.8a 13.3ab 15ab 19 21.9 23.7 24.9 R5 8.8a 9.8a 12.5a 13.3a 13.8a 14.5a 16a 20.8 24.9 27.8 28.6 LSD (0.05) 2.3 3.89 3.97 2.44 3.8 2.58 3.06 NS NS NS NS

PAGE 72

59Table 3-3. Root biomass8 (g per plant) in fluridone susceptible (S ) and resistant (R1-R5 ) hydrilla populations. Hydrilla Population Oct 2004 Nov 2004 Dec 2004 Jan 2005 Feb 2005 Mar 2005 Apr 2005 May 2005 Jun 2005 Jul 2005 Aug 2005 S 1.60 2.61 2.80 2.93 3.00 3.20 3.18 3.25 3.30 3.40 3.40 R1 1.55 1.68 2.38 2.40 2.75 2.63 2.95 3.05 3.08 3.18 3.23 R2 2.15 2.20 2.95 3.10 3.20 3.25 3.30 3.33 3.33 3.38 3.40 R3 2.38 2.48 3.15 3.25 3.45 3.63 3.63 3.65 3.68 3.63 3.70 R4 1.53 1.58 1.88 2.05 2.55 2.75 2.88 2.90 2.88 2.92 2.90 R5 1.48 1.50 2.61 2.75 2.55 2.90 2.95 2.90 2.88 2.92 2.90 LSD (0.05) NS NS NS NS NS NS NS NS NS NS NS 8 Root biomass consists of root crowns and rhizome weight.

PAGE 73

60Table 3-4. Subterranean turions produced (p er plant) in fluridone flur idone susceptible (S) and re sistant (R1-R5) hydrilla popu lations. Hydrilla Population Oct 2004 Nov 2004 Dec 2004 Jan 2005 Feb 2005 Mar 2005 Apr 2005 May 2005 Jun 2005 Jul 2005 Aug 2005 S 0.5b 1.9b 3.1ab 6.6a 8.9a 9.9a 10a 10.5a 9.8a 10a 10a R1 0.8ab 1.8b 3.0ab 6.1a 7.5a 8.8a 9.5a 9.5a 9.7a 10.2a 9a R2 0.9ab 2.0b 2.8ab 5.3a 7.0a 9.0a 9.2ab 9.0a 8.5ab 9.2a 9a R3 0c 1.1b 2.0b 2.4b 3.7b 6.7b 6.7b 6.5b 6.2b 6.3b 6b R4 0.6a 2.4ab 3.1ab 5.6a 7.2a 9.6a 9.8a 10a 10.5a 10a 10a R5 1.5a 3.4a 4.3a 5.0a 11a 12.3a 12a 10a 10.2a 10a 10a LSD (0.05) 0.85 1.1 2.24 1.68 2.86 2.96 2.76 2.2 2.94 3.54 3.42

PAGE 74

61Table 3-5. Axillary branches (per plant) in fluridone susceptible (S) and resistan t (R1-R5) hydrilla populations. Hydrilla Population Oct 2004 Nov 2004 Dec 2004 Jan 2005 Feb 2005 Mar 2005 Apr 2005 May 2005 Jun 2005 Jul 2005 Aug 2005 S 7.1a 7.7a 9.2a 8.7a 9.5a 9.1a 8.3ab 8.5ab 7.8ab 7.8ab 8ab R1 6.4a 8.1a 9.0a 9.5a 8.5a 9.3a 9.3ab 9.8ab 9.5ab 9.0ab 9.8ab R2 7.0a 8.8a 9.9a 10.9a 9.6a 9.2a 8.8ab 8.3ab 8.8ab 8.3ab 8.7ab R3 4.8b 4.0b 4.4b 5.0b 5.8b 5.9b 6.9b 6.8b 7b 6.9b 6.3b R4 8.0a 8.5a 10a 10.3a 10a 9.9b 10.3ab 10.1ab 10ab 9.9ab 10ab R5 7.5a 8.6a 9.3a 9.6a 10a 9.9b 11.1a 10.9a 11.3a 10.9a 10.6a LSD (0.05) 2.67 2.36 3.15 3.38 3.44 3.2 4.03 4 4.1 3.4 3.97

PAGE 75

62Table 3-6. Flower number (per plant) in fluridone sus ceptible (S) and resistant (R 1-R5) hydrilla populations. Hydrilla Population Oct 2004 Nov 2004 Dec 2004 Jan 2005 Feb 2005 Mar 2005 Apr 2005 May 2005 Jun 2005 Jul 2005 Aug 2005 S 5.9 5.4b 0b 0b 0 0 0 0 0 0 5.3a R1 3.7 4.3b 0b 0b 0 0 0 0 0 0 3.7ab R2 5.6 5.1b 0b 0b 0 0 0 0 0 0 4.8a R3 4 8.3a 4a 2.3a 0 0 0 0 0 0 1.7b R4 3.1 4.5b 0b 0b 0 0 0 0 0 0 4ab R5 7 5.3b 0b 0b 0 0 0 0 0 0 6.3a LSD (0.05) NS 2.61 1.49 0.42 ------3.07

PAGE 76

63Table 3-7. Total biomass (g per plant) in fluridone su sceptible (S) and resistant (R1-R5) hydrilla populations. Hydrilla Population Oct 2004 Nov 2004 Dec 2004 Jan 2005 Feb 2005 Mar 2005 Apr 2005 May 2005 Jun 2005 Jul 2005 Aug 2005 S 9.8 13.9a 14.8a 16.1a 17.2a 18.6ab 19.1ab 24.1 26.3 29.2 31.7 R1 10.2 12.7ab 14.4a 15.3a 16.1ab 16.7b 18.5ab 22.2 24.8 28.0 29.4 R2 10.2 13.3ab 13.7ab 15.1a 15.6ab 17.1ab 17.7b 22.9 26.1 29.0 30.3 R3 8.5 9.5b 11.4b 12.2b 13.8b 16.5b 17.4b 26.5 29.7 31.4 32.9 R4 10.8 11.2ab 13.2ab 15a 15.9ab 17.8ab 19.4ab 23.9 26.9 28.8 30.0 R5 10.8 12.1ab 13.5ab 15.3a 16.1ab 19.7a 20.8a 28.4 31.4 34.6 35.6 LSD (0.05) NS 3.6 2.62 2.4 2.21 3.07 3.16 NS NS NS NS

PAGE 77

64 Month O ct , 0 4 N o v , 0 4 D e c, 0 4 Ja n , 0 5 F e b , 0 5 M a r , 0 5 A p r , 0 5 M a y , 0 5 J u n , 0 5 J u l , 0 5 A u g , 0 5 Relative growth rate (mg g-1 d-1) 0 2 4 6 8 10 12 14 16 S R1 R2 R3 R4 R5 Figure 3-1. Relative growth rate in fluri done susceptible (S) and resistant (R1-R5) hydrilla populations during the growth season.

PAGE 78

65 CHAPTER 4 MOLECULAR CHARACTERIZATION OF FLURIDONE RESISTANT HYDRILLA IN FLORIDA Introduction Invasive plants are considered to be the greatest threat to biodiversity in all ecosystems, after habitat de struction (Pimm and Gilpin, 1989; Randall, 1996). Invasive, exotic weeds have been serious problems in many fresh water ecosystems in the United States for more than a century. Species including water hyacinth ( Eichchornia crasipes ), eurasian watermilfoil ( Myriophyllum spicatum ), purple loosestrife ( Lythrum salicaria ), parrotfeather ( Myriophyllum aquaticum ), Brazilian elodea ( Egeria densa ) and hydrilla ( Hydrilla verticillata ), largely have been a result of either intentional introductions for ornamental use or the aquarium trad e (Countryman, 1970; Couch and Nelson, 1985; Sutton 1985; Schmitz, 1990). Escape of these species from cultivation and subsequent spread has caused the expenditure of millions of dollars annually for their control and management. Herbicides are usually the most last effective choice for aquatic weed control (Bottrell, 1979). Herbicides offer longer lasting control than mechanical methods, minimal expenditures of labor and equipm ent, and offer greater flexibility and predictability which ultimately leads to reduced costs. However, health and environmental concerns restrict the use of chemicals in aquatic ecosystems (Brooker and Edwards, 1975). Consequently, the number of chemicals registered by the USEPA for aquatic use is limited (Way and Chance llor, 1976). Nevertheless, most of the

PAGE 79

66 management programs for invasive aquatic weed control in United States rely on the use of the few registered herbicides. Hydrilla is one of the most serious invasive aquatic weed problems in the United States, and has now become the most abunda nt, nonindigenous submersed aquatic plant in Florida. Fluridone is the only USAEPA a pproved herbicide for the treatment of large water bodies that efficiently controls hydrilla. As a result, fluridone has been widely used in hydrilla control programs for over three decades. Fluridone targets the carotenoid biosynthetic pathway in plants , thereby, inhibiting synthesis of carote noids in developing tissues. The lack of carotenoids cause s photooxidation and the disappearance of chlorophylls and results in white foliage. Flur idone acts as a non-competitive inhibitor of the enzyme phytoene desaturase (PDS). PDS is encoded by the pds gene, a member of a low copy number nuclear gene family (Bartley et al., 1991). The protei n is then imported into the chloroplasts where mature PDS prot ein is detected in soluble and membrane fractions of chloroplasts (Bar tley et al., 1991; Bonk et al., 19 97). The catalyt ic activity of phytoene desaturase appears to be ratelimiting and controls the remainded of the carotenoid pathway (Chamovitz et al., 1993). The genes encoding phytoene desaturase fr om cyanobacteria (Chamovitz et al., 1991), green algae and plants (Bartley et al ., 1991; Pecker et al ., 1992; Hugueney et al., 1992) have been isolated ( pds genes). Cyanobacteria and plants possess a PDS which catalyzes the first two dehydrogenation reactio ns in the carotenoid synthesis pathway, producing carotene from phytoene as an e nd product (Sandmann, 1994). However, there is little homology between plant and b acterial PDS (Misawa et al., 1990). Albrecht et al. (1995) cl oned cDNA from Capsicum annuum encoding an enzyme mediating

PAGE 80

67 desaturation of carotene to lycopene. Sequence comp arison revealed 33-35% similarity with a previously cloned pl ant or cyanobacterial PDS. Cy anobacteria as model systems have the advantage of combining the presence of a photosynthetic appa ratus with the fast growth rate of prokaryotes. Most of the bioc hemical studies investig ating the potency of phytoene desaturase inhibito rs (fluridone, norflurazon) have been made either in vivo with intact cells or in vitro with cyanobacterium cell ex tracts (Mayer et al., 1989; Babczinski et al., 1995; Sa ndmann, 2001; Sandmann and Mitchell, 2001). There is limited data on chemical inhibition of PDS from higher plants (Laber et al., 1999). With recent advancements in molecular tools, research has been ongoing to study the resistance to bleaching herbicides affecting PDS in vitro involving both cyanobacterial cell extracts (Navarro et al., 1995; Windhvel et al., 1997) and higher plants (Misawa et al., 1994; Albrecht et al., 1995, Arias et al., 2004). Although resistance has been reported and conferred in PDS from cyanobacteria (Chamovitz et al., 1993; Windhvel et al., 1994), hydrilla is the first higher plant to have developed resistance from a PDS inhibiting he rbicide due to somatic mutations (Michel et al., 2004). Hydrilla reprodu ces only through vegetative mean s in Florida, therefore the development of herbicide resistance was cons idered unlikely (Hill, 1982). Nevertheless, fluridone resistant hydrilla biotypes have been confirmed in public and private water bodies throughout Florida. More over, different levels of fluridone resistance have been demonstrated (Chapter 2). Therefore, the obj ective of this experiment was to determine the molecular basis for fluridone herbicide resistance, including cloning of the nuclearencoded pds gene, and detection of possible mu tations within the gene. This was

PAGE 81

68 performed in hydrilla populations displaying vari ous levels of resistance in an attempt to correlate resistance with molecular analysis. Materials and Methods Collection of Plant Material Five confirmed fluridone resistant hydrilla populations were collected from different Florida lakes with known fluridone application histories . These populations were designated as R1, R2, R3, R4, and R 5. One population, termed fluridone susceptible (S) was collected from a private pond in north central Florida that ha s never been treated with fluridone. Plant samples were clean ed thoroughly and grown in separate 900 L concrete vaults under natural conditions in the absence of fluridone from Sept 2004 to Sept 2005. Chemicals and Reagents RNeasy Plant kit (for total RNA extract ion), Oligotex mRNA kit (for mRNA purification), and QIAprep Spin Miniprep kit (for plasmid DNA purification) were obtained from Qiagen (Valencia, California, USA). Platinum Taq DNA polymerase High Flidelity (for polymerase chain reaction) and TOPO TA Cloning kit (for vector cloning of gene) were purchased from Invitrogen (Carls bad, California, USA) . Restriction enzymes and dNTPs were ordered from Promega (Mad ison, Wisconsin, USA). All other chemicals were highest grade available. Genomic RNA Extraction and cDNA Synthesis Total RNA was isolated from hydrilla leaf tissue using the Qiagen RNeasy plant mini kit according to the manufacturer’s pr otocol. Immature leaf tissue from actively growing hydrilla shoots was taken and two plants per population were used for RNA extraction. To check integrity of RNA, 3-5 g total RNA were loaded on a 1% agarose

PAGE 82

69 gel containing 2.2 M of formaldehyde in the pr esence of ethidium bromide (Maniatis et al. 1982). Integrity was judged by the presence and intensity of rRNA bands. Purification of poly A+ mRNA was performed using a Qi agen Oligotex mRNA kit for isolating mRNA from total RNA following the manuf acturer’s protocol. Poly A+ mRNA RNA was reverse transcribed usi ng ThermoScript RT-PCR system1. The reaction mixture contained 1 L of 50 M Oligo (dT)20, 8 L of mRNA, 2 L of 10 mM dNTP mix, 4 L of cDNA synthesis buffer, I L of 0.1 M DTT, 1 L of RNase OUT (40 U/L), and 3 L of DEPC-treated water and the resulting single-stranded cDNA product was treated with RNase H at 37C for 30 min. RNA a nd cDNA were stored at -20C. PCR Amplification and Sequencing of Phytoene Desaturase ( pds ) Gene Phytoene desaturase ( pds ) is a nuclear-encoded gene expressed in chloroplasts. Synthesized cDNA was used to clone pds genes from individual hydrilla specimens. Forward and reverse primers were designed on the basis of the pds gene sequence obtained from Dr. F. E. Dayan2. Alleles for the pds gene were amplified via PCR using Platinum Taq Polymerase High Fidelity and the primers PDS-start (5’ATG ACT GTT GCT AGG TCG GTC GTT 3’) and AtPD S-1849 (5’ TAC CCTT TGC TTG CTG ATG 3’) on cDNA. The sequence of hydrilla pds gene is available fr om GenBank (Accession number AY639658). The PCR cocktail consisted of 2 L of cDNA, 1 L of each primer (10 pmol), 5 mL of 10X High Fidelity PCR buffer, 2 L of 50mM MgSO4, 1 L of 10 mM dNTPs, 0.2 L of Platinum Taq Polymerase High Fidelity and 37.8 L of DEPCtreated water. In order to determine optimal amplification conditions, PCR was 1 Invitrogen Corp (Catalog no. 11146-024), Carlsbad, California, USA. 2 Research Plant Physiologist, USDA/ARS, Natural Products Utilization Research Unit, PO Box 8048, University of Mississippi, Ox ford, Mississippi 38677, USA.

PAGE 83

70 conducted at the annealing temperature of 65, 62, 60, 59 and 55C, respectively. This determined the appropriate annealing temp erature of the PCR. The optimum reaction protocol consisted of 10 cycles of 95C for 2 min, 94C for 30 s, 65C for 30 s, and 68C for 1 min, followed by 10 cycles of 94C for 30 s, 65C for 30 s, and 62C for 1 min, followed by another 15 cycles of 94C for 30 s, 65C for 30 s, and 59C for 1 min. These 35 amplification cycles were followed by one cycle of 68C for 10 mi n. PCR products were then resolved on a 1% Tris-b orate-EDTA buffer and stained with ethidium bromide. PCR fragments were cloned into pCR4-TOPO vector according to manufacturer’s instructions. Escherichia coli strain TOP 10 was transformed with the plasmid vector harboring the PCR products and plated on Luria-Bertani brot h (LB) plus agar supplemented with 100 mg mL-1 kanamycin. Isolation of recombinant clones was carried out using standard procedures3 . Plates were incubated for 24 h at 37 C, after which individual colonies were selected and grown overnight in liqui d LB medium amended with 100 mg mL-1 kanamycin under constant shaking. Plasmid DNA was purified using Qiagen plasmid isolation kit. The presence of pds allele was verified using restriction digest with Eco R1 enzyme. Double stranded PCR products were sent to MWG-Biotech (High Point, North Carolina, USA) for sequencing. Homology search for sequences of selected clones was performed using basic local ali gnment sequence tool (BLAST) at http://www.ncbi.nlm.nih.gov/blast. Furthermor e, the sequences were aligned using CLUSTALW 1.82 at http://www.ebi.ac.uk/clustalw. 3 TOPO TA Clonig Kit for Sequencing (Catalogue no. K 4575), Invitrogen Corp., Carlsbad, California, USA

PAGE 84

71 The molecular characterization were done at planting (Sept 2004) and 12 months after planting (Sept 2005) to detect any possible ch ange in mutations in the pds gene sequences from resistant hydrilla populations over time in the absence of fluridone. Results and Discussion PCR amplification of pds gene segments using c DNA resulted in a single DNA fragment from each amplification reaction with the expected size (Figure 4-1). The resulting pds gene was cloned into TOPO vector a nd digested with restriction enzyme Eco R1 to confirm accuracy (Figure 4-2). The resulting full length pds gene sequences from fluridone susceptible and resistant hydr illa populations are show n in Figure 4-3. The total length of coding region of pds gene was 1743 nucleotides. PDS protein consisted of 580 amino acids and is presented in Figure 4-4. Nucleotide sequences from different flur idone resistant hydri lla biotypes were compared with one another and with fluridone susceptible hydrilla . Overall, there was 99% identity at the nucleotide leve l in the coding region for all pds sequences. This indicates that the phytoene desaturase seque nce is highly conserved among the different hydrilla populations. The pds genes from R2 and R4 had identical nucleotide sequences (Figure 4-3). Comparisons of the sequences of different resistant populations showed mutations at different nucleotide positions in different hydrilla populations compared to susceptible hydrilla (Tables 4-1 and 4-2). R1 showed three different mutations at nucleotide positions 306, 910, and 1055. One of these was a silent mutation i.e., AGC to GGC at 306 (Ser102) but the mutations at nucleotid e position 910 and 1055 resulted in Arg to Ser304 and Asn to Ser352 amino acid changes in R1 compared to S, respectively (Table 4-1). Resistant hydr illa populations R2, R4 and R5 showed mutation only at

PAGE 85

72 nucleotide position 910, resulting in change of amino acid Arg304 to Ser304 in R2 and R4 and Arg to His304 amino acid change in R5. Resistant hydrilla R3 recorded 13 different mutations in the pds gene compared to susceptible hydrilla. Nine of these mutations were silent mutations i.e., CCC to TCC at 60 (Tyr20), TGG to CGG at 150 (Phe50), GTG to TTG at 447 (Ala149), GCC to ACC at 597 (Lys199), ATG to TTG at 657 (Ile219), CCT to ACT at 894 (Ile298), GAC to AAC at 1425 (Ala475), ACA to GCA at 1710 (Val570), and TGC to CGC at 1754 (+16 downstream coding region). Four mutations resulted in amino acid changes from susceptible hydrilla and were recorded at nucleotide positions 260, 910, 992, and 1021 resulting in change in Ala to Val87, Arg to Ser304, Asp to Ala330, and Ile to Val341 at these positions, respectively (Tables 4-1 and 4-2). Alignment of pds gene sequences from Arabidopsis , cyanobacteria and higher plants revealed conserved Arg codons co rresponding to nucleo tide position 304 in hydrilla (Figure 4-5). Two separate and inde pendent single-point mutations of the codon 304 encoding for Arg (Arg304) in pds were recorded. The codon usage for Arg304 in the fluridone susceptible hydrilla is CGT and single-point mutations yielding either Ser (AGT) substitution in R1, R2, R3, and R 4, or His (CAT) substitution in R5 were recorded. Similar substitutions to those descri bed for hydrilla (Arg to Ser, Cys and His) were observed at different positions within the pds sequences of the cyanobacteria Synechococcus and Synechocystis when the cu ltures were grown on selection media with various PDS inhibitors (Linden et al ., 1989; Martinez-Ferez and Vioque, 1992). We reported the presence of susceptible and resistant pds alleles in resistant hydrilla populations (data not shown). This show ed that the resistant hydrilla plants were

PAGE 86

73 heterozygous. Similar results have been reported previous ly (Michel et al., 2004). The pds gene was cloned from the herbicide-susceptible as well as from all herbicide-resistant populations at 12 months after planting, to determine if any differences in the pds gene sequences in the resistant hydrilla occurred in the absence of fluridone over time. No differences were observed at codon 304 in the PDS protein of any hydrilla population (Tables 4-1 and 4-2). Weed resistance is not unique. Herbicide-resistant weeds were first discovered in the United States in the late 1960s in a pine nursery where triazine herbicides had been used repeatedly. Factors that accelerate th e selection of resistant biotypes include repeated use of a single herbicid e in large areas, the lack of alternating different modes of action, high efficacy of the herbicide on the sens itive biotype at the rate used, and long residual herbicide activity (Maxwell and Mortimer 1994; Volenberg et al., 2002). Examples of plants that have developed re sistance to herbicide via mutations can be found for many herbicides (Tan et al ., 2005; Kaundun and Windass, 2006). Several mutations resulting in amino-acid substitutions of ACCase and ALS proteins have been reported in several plant species (Tan et al., 2005; Kaundun and Windass, 2006). By 2002, eight different amino acid substitutions fo r Pro197 have been reported to confer herbicide (ALS-inhibitors) resistance in w eeds, and 17 amino acid substitutions that conferred resistance in various organisms such as plants, yeast, bacteria and green algae (Tranel and Wright, 2002). Fluridone has been used exclusively for c ontrolling hydrilla for last three decades. The widespread use of fluridone is due to several factors including: low use rates, favorable native plant selectivity, slow activ ity (reduced oxygen depletion) and extreme

PAGE 87

74 effectiveness on hydrilla. Hydr illa is controlled in large water bodies (> 100 000 ha) by sustaining between 12 and 36 nm (4 g/L) concentrations of the chemical fluridone (Sonar) in lake water for 60-90 da ys (Netherland and Getsinger, 1995; Fox et al., 1996). The asexual nature of hydrilla grow th and reproduction s hould have precluded the resistance development. However, imposi ng a sustained selection pressure on a plant species like hydrilla which can grow 10 cm per day and produce entire plant from a single node (Langeland, 1996), coupled with th e sustained low doses of fluridone over the past several years likely have favored the selection for herbicide resistance. In prokaryotes such as cyanobact eria, there have been reports of several mutations on the pds gene that conferred resistance to PD S-inhibiting herbicides, i.e., Val403Gly, Leu320Pro, Arg195Pro and Leu436Arg, and a 20 nucleotide deletion in the transit peptide (Chamovitz et al., 1993). Windhvel et al. (1994) utilized a cyanobacteria based model to design a screening method to discove r resistance to bleaching herbicides. This consisted of transferring the herbicide insensitive pds genes from the bacterium Erwinia uredovora into the cyanobacterium Synechococcus . They showed that the strain Synechococcus PC 742-PIM8-BG1, which contained bacterial PDS (CRT-1), was resistant to the bleaching herbicides that inhi bited the plant type PDS. In particular, the point mutation resulting in an amino acid subs titution (valine to glyc ine) in the phytoene desaturase gene of Synechococcus PCC 7942 was responsible for herbicide resistance in the mutant NFZ 4 (Chamovitz et al., 1991). A tobacco transformant, expressing a foreign bacterial phytoene desaturase gene which is structurally unrelate d to the plant-type enzyme and thus lacks the common binding si te for many bleaching herbicides (Misawa et al., 1994), was totally resistant to both fluri done and norflurazon. Several

PAGE 88

75 cyanobacterial mutants having resistance ag ainst norflurazon also exhibited cross resistance to fluridone and ot her bleaching herbicides (Babcz inski et al., 1995; Laber et al., 1999; Sandmann, 2001). Wagner et al . (2002) determined that the pds gene of Synechococcus conferred resistance to norflurazon and fluridone, 58 and 3 fold higher than the wild type controls, respectivel y. Recently, Walsh et al. (2004) reported a terrestrial weed Raphanus raphanistrum to have developed resistance to the PDS inhibitor diflufenican. This weed developed resistance to diflufenican after only four applications of this herbicid e with 16% of these populations surviving four-fold times the commercial application rate of diflufeni can. (Walsh et al., 2004) The molecular mechanisms related to the development of resistance in R. raphanistrum are not known at this time. Resistant hydrilla populations collected from different Florida lakes recorded two different mutations resulting in an amino acid change at codon 304 (Table 4-1). These results were consistent with recent findings of Michel et al. (2004). They also reported three independent base pair substitutions at the amino acid 304 codon (Arg to Ser, Cys and His) of PDS protein, in relation to flur idone resistance. However, they did not find mutations at other codons, especially those observed in the R3 population. PDS is a nuclear-encoded protein with ac tivity in the chloroplasts, the site of carotenoid synthesis (Bartley et al., 1991). Despite caroteno ids being synthesized in all types of photosynthetic tissues, plants usually have very low levels of pds transcript (Bartley et al., 1991; Pecker et al., 1992). Low levels of transcription of a single copy gene have been negatively correlated w ith its mutation frequency. Kovalchuk et al. (2000) determined that the frequency of so matic mutations for a single copy gene in

PAGE 89

76 plants is two to three orders of magnitude hi gher than in animals, yeast or even bacteria. The predicted frequency of forward mutations for a single copy gene with low level of transcription was 10-6 to 10-7 events per base pair in A. thaliana (Kovalchuk et al., 2000). According to these predicted values, the 1.7 kb pds gene of hydrilla could have mutations in the order of 1.7 x 10-3 to 1.7 x 10-4. If genes with low levels of transcription have higher levels of mutation, then the low level of transcription pds gene could be prone to present mutations and this could be an addi tional factor in the development of hydrilla biotypes with herbicide resistance. In different resistant hydrilla populations from six different Florida lakes, we found a total of a total of eight transversi ons and eleven transition mutations on the pds gene. In general, transitions (changes from purine to purine or pyrimidine to pyrimidine) are more common than transversions (change s from purine to pyrimidine and vice versa) (Muse, 2000). In hydrilla, the fr equency of transversions was C A (5x), A T (1x) G T (1x), and A C (1x) while for transitions it was A G (4x), G A (3x), C T (2x), and T C (2x). This frequency was much higher than would have been expected in an asexually propagated species. Hydrilla is a fast growing plant with multiple means of propagation; able to generate new individuals from a single node; and produces profuse axillary branching (Langeland, 1996). A high ra te of somatic mutations can perpetuate since this plant species u ndergoes only asexual propaga tion. In general, asexually propagated plant species are under strong unipare ntal constraints which limit their ability to respond to environmental changes (Holsinge r, 2000). The reduced ge netic variability is also thought to lead to accumulation of deleterious mutations called mutational meltdown that reduce the survival chances of these plants (Kle kowski, 2003). Charlesworth et

PAGE 90

77 al. (1993) reported that a selec tion against deleterious alleles is facilitated or sustained by mutations and results in the reduced genetic dive rsity of highly inbreeding and asexual populations. We reported presence of both susceptible and resistance pds alleles in various resistant hydrilla populat ions. Therefore, the deleteri ous effect of mutations in pds gene could have been on the susceptible alle les, rather than the resistant alleles of gene. This could explain the high growth rate of resistant hydrilla populations in different Florida lakes similar to susceptible hydrilla. Resistant hydrilla populations R1, R2, and R4 showed the Arg to Ser mutation at the 304 codon in their respective fluridone resistant PDS protein, and the resistance factors observed were two-fold (in R1 and R2), and five-fold in R4 (Chapter 2). R5 showed Arg to His mutation at the 304 codon, a nd it resulted in sevenfold resistance at the plant level to fluridone. Michel et al. ( 2004) also reported a hi gher resistance level in hydrilla as a result of His mutation compar ed to the Ser mutation. R3 hydrilla also has Arg to Ser mutation at codon 304 in PDS protein, but also ha d a Asp to Ala mutation and Ile to Val mutation at codons 330 and 341, resp ectively. Currently, only the mutation at codon 304 has been reported to be a cause of herbicide resistance in hydrilla (Michel et al., 2004). The presence of these two mutati ons at codons 330 and 341 close to codon 304 may contribute to resistance in the R3 population, but this remains unconfirmed. There are several economic implicati ons of these findings regarding the development of fluridone re sistance and how it impacts hydrill a management in Florida. According to recent Florida Department of Environmental Protection budget requests, hydrilla infests over 11,500 hectares w ith $17,906,098 per year needed for control (Florida DEP, 2004). Over 90% of hydrilla control programs ($10+ million/year) in

PAGE 91

78 Florida utilized fluridone due to its envir onmental characteristics, longevity of control and relatively low cost. Furthermore, if additional resistance occurs, the cost and environmental impact of fluridone might precl ude its further use in Florida. The lack of sustainable alternatives to manage hydrilla ha s further worsened the problem. Due to the aggressive spreading nature of hydrilla in aquatic ecosystems, the development and spread of fluridone resistant hydrilla biot ypes may forecast significant and long-lasting ecological and economic problems throughout the Southern states of the U.S.

PAGE 92

79 Figure 4-1. Agarose gel electr ophoresis of PCR amplified pds gene sequences from different hydrilla populations; Lane M: 1 Kb molecular ladder; Lane 1, 2: Susceptible (S); Lane 3, 4: R1; Lane 5, 6: R2; Lane 7, 8: R3; Lane 9, 10: R4; Lane 11, 12: R5; Lane 13: negative cont rol; 1.9 Kb is the size of amplified pds gene sequence. 1.9Kb

PAGE 93

80 Figure 4-2. Agarose gel electrophor esis of restriction digest of TOPO vector containing pds alleles from hydrilla populations with Eco R1 enzyme amplified; Lane M: 1 Kb molecular ladder; Lane 1, 2: Suscep tible (S); Lane 3, 4: R1; Lane 5, 6: R2; Lane 7, 8: R3; Lane 9, 10: R4; Lane 11, 12: R5; Lane 13: negative control (uncut plasmid); 1.9 Kb is the size of amplified pds gene sequence. 1.9Kb

PAGE 94

81 PDS-R2 -------AATTGAATTTAGCGGCCGCGAATTCGCCCTTATGACTGTTGCTAGGTCGGTCG PDS-R5 ---GGCGAATTGAATTTAGCGGCCGCGAATTCGCCCTTATGACTGTTGCTAGGTCGGTCG PDS-R1 ----GCGAATTGAATTTAGCGGCCGCGAATTCGCCCTTATGACTGTTGCTAGGTCGGTCG PDS-S -AGGGCGAATTGAATTTAGCGGCCGCGAATTCGCCCTTATGACTGTTGCTAGGTCGGTCG PDS-R4 TAGGGCGAATTGAATTTAGCGGCCGCGAATTCGCCCTTATGACTGTTGCTAGGTCGGTCG PDS-R3 TAAGGCGAATTGAATTTAGCGGCCGCGAATTCGCCCTTATGACTGTTGCTAGGTCGGTCG ***************************************************** PDS-R2 TTGCAGTCAATCTAAGTGGTTCCCTTCAAAACAGATACCCAGCCAGTTCATCAGTCAGCT PDS-R5 TTGCAGTCAATCTAAGTGGTTCCCTTCAAAACAGATACCCAGCCAGTTCATCAGTCAGCT PDS-R1 TTGCAGTCAATCTAAGTGGTTCCCTTCAAAACAGATACCCAGCCAGTTCATCAGTCAGCT PDS-S TTGCAGTCAATCTAAGTGGTTCCCTTCAAAACAGATACCCAGCCAGTTCATCAGTCAGCT PDS-R4 TTGCAGTCAATCTAAGTGGTTCCCTTCAAAACAGATACCCAGCCAGTTCATCAGTCAGCT PDS-R3 TTGCAGTCAATCTAAGTGGTTCCCTTCAAAACAGATATCCAGCCAGTTCATCAGTCAGCT ************************************* ********************** PDS-R2 GCTTCCTTGGCAAAGAGTACAGATGCAACAGTATGTTAGGATTCTGCGGTAGTGGAAAAT PDS-R5 GCTTCCTTGGCAAAGAGTACAGATGCAACAGTATGTTAGGATTCTGCGGTAGTGGAAAAT PDS-R1 GCTTCCTTGGCAAAGAGTACAGATGCAACAGTATGTTAGGATTCTGCGGTAGTGGAAAAT PDS-S GCTTCCTTGGCAAAGAGTACAGATGCAACAGTATGTTAGGATTCTGCGGTAGTGGAAAAT PDS-R4 GCTTCCTTGGCAAAGAGTACAGATGCAACAGTATGTTAGGATTCTGCGGTAGTGGAAAAT PDS-R3 GCTTCCTTGGCAAAGAGTACAGATGCAACAGTATGTTAGGATTCTGCGGTAGTGGAAAAT ************************************************************ PDS-R2 TGGCTTTTGGCGCAAATGCACCCTATTCTAAGATTGCAGCTACCAAACCAAAGCCCAAAC PDS-R5 TGGCTTTTGGCGCAAATGCACCCTATTCTAAGATTGCAGCTACCAAACCAAAGCCCAAAC PDS-R1 TGGCTTT T GGCGCAAATGCACCCTATTCTAAGATTGCAGCTACCAAACCAAAGCCCAAAC PDS-S TGGCTTTTGGCGCAAATGCACCCTATTCTAAGATTGCAGCTACCAAACCAAAGCCCAAAC PDS-R4 TGGCTTTTGGCGCAAATGCACCCTATTCTAAGATTGCAGCTACCAAACCAAAGCCCAAAC PDS-R3 TGGCTTTCGGCGCAAATGCACCCTATTCTAAGATTGCAGCTACCAAACCAAAGCCCAAAC ******* **************************************************** PDS-R2 TTCGCCCTTTGAAGGTCAACTGCATGGATTTCCCAAGACCTGATATAGATAACACTGCTA PDS-R5 TTCGCCCTTTGAAGGTCAACTGCATGGATTTCCCAAGACCTGATATAGATAACACTGCTA PDS-R1 TTCGCCCTTTGAAGGTCAACTGCATGGATTTCCCAAGACCTGATATAGATAACACTGCTA PDS-S TTCGCCCTTTGAAGGTCAACTGCATGGATTTCCCAAGACCTGATATAGATAACACTGCTA PDS-R4 TTCGCCCTTTGAAGGTCAACTGCATGGATTTCCCAAGACCTGATATAGATAACACTGCTA PDS-R3 TTCGCCCTTTGAAGGTCAACTGCATGGATTTCCCAAGACCTGATATAGATAACACTGTTA ********************************************************* ** PDS-R2 ATTTCTTGGAAGCTGCTGCTCTTTCTTCCTCTTTTCGCAATTCAGCAAGACCAAGTAAAC PDS-R5 ATTTCTTGGAAGCTGCTGCTCTTTCTTCCTCTTTTCGCAATTCAGCAAGACCAAGTAAAC PDS-R1 ATTTCTTGGAAGCTGCTGCTCTTTCTTCCTCTTTTCGCAATTCGGCAAGACCAAGTAAAC PDS-S ATTTCTTGGAAGCTGCTGCTCTTTCTTCCTCTTTTCGCAATTCAGCAAGACCAAGTAAAC PDS-R4 ATTTCTTGGAAGCTGCTGCTCTTTCTTCCTCTTTTCGCAATTCAGCAAGACCAAGTAAAC PDS-R3 ATTTCTTGGAAGCTGCTGCTCTTTCTTCCTCTTTTCGCAATTCAGCAAGACCAAGTAAAC *******************************************.**************** PDS-R2 CTCTTCAAGTTGTAATTGCTGGTGCAGGTTTGGCTGGTCTTTCAACAGCAAAGTATCTCG PDS-R5 CTCTTCAAGTTGTAATTGCTGGTGCAGGTTTGGCTGGTCTTTCAACAGCAAAGTATCTCG PDS-R1 CTCTTCAAGTTGTAATTGCTGGTGCAGGTTTGGCTGGTCTTTCAACAGCAAAGTATCTCG PDS-S CTCTTCAAGTTGTAATTGCTGGTGCAGGTTTGGCTGGTCTTTCAACAGCAAAGTATCTCG PDS-R4 CTCTTCAAGTTGTAATTGCTGGTGCAGGTTTGGCTGGTCTTTCAACAGCAAAGTATCTCG PDS-R3 CTCTTCAAGTTGTAATTGCTGGTGCAGGTTTGGCTGGTCTTTCAACAGCAAAGTATCTCG ************************************************************ Figure 4-3. Aligned sequences of pds alleles from different hydrilla populations. Start, stop codons, and mutations in resistant hydri lla populations are in bold letters.

PAGE 95

82 PDS-R2 CAGATGCAGGGCACATACCCATACTACTGGAGGCTAGAGATGTATTGGGTGGCAAGGTGG PDS-R5 CAGATGCAGGGCACATACCCATACTACTGGAGGCTAGAGATGTATTGGGTGGCAAGGTGG PDS-R1 CAGATGCAGGGCACATACCCATACTACTGGAGGCTAGAGATGTATTGGGTGGCAAGGTGG PDS-S CAGATGCAGGGCACATACCCATACTACTGGAGGCTAGAGATGTATTGGGTGGCAAGGTGG PDS-R4 CAGATGCAGGGCACATACCCATACTACTGGAGGCTAGAGATGTATTGGGTGGCAAGGTGG PDS-R3 CAGATGCAGGGCACATACCCATACTACTGGAGGCTAGAGATGTATTGGGTGGCAAGGTGG ************************************************************ PDS-R2 CAGC G TGGAAAGATGATGATGGAGACTGGTATGAGACAGGCCTGCATATATTTTTTGGTG PDS-R5 CAGC G TGGAAAGATGATGATGGAGACTGGTATGAGACAGGCCTGCATATATTTTTTGGTG PDS-R1 CAGC G TGGAAAGATGATGATGGAGACTGGTATGAGACAGGCCTGCATATATTTTTTGGTG PDS-S CAGC G TGGAAAGATGATGATGGAGACTGGTATGAGACAGGCCTGCATATATTTTTTGGTG PDS-R4 CAGC G TGGAAAGATGATGATGGAGACTGGTATGAGACAGGCCTGCATATATTTTTTGGTG PDS-R3 CAGC T TGGAAAGATGATGATGGAGACTGGTATGAGACAGGCCTGCATATATTTTTTGGTG **** ******************************************************* PDS-R2 CATATCCCAATGTGCAGAATTTATTTGGTGAACTTGGCATAAATGATCGTCTACAATGGA PDS-R5 CATATCCCAATGTGCAGAATTTATTTGGTGAACTTGGCATAAATGATCGTCTACAATGGA PDS-R1 CATATCCCAATGTGCAGAATTTATTTGGTGAACTTGGCATAAATGATCGTCTACAATGGA PDS-S CATATCCCAATGTGCAGAATTTATTTGGTGAACTTGGCATAAATGATCGTCTACAATGGA PDS-R4 CATATCCCAATGTGCAGAATTTATTTGGTGAACTTGGCATAAATGATCGTCTACAATGGA PDS-R3 CATATCCCAATGTGCAGAATTTATTTGGTGAACTTGGCATAAATGATCGTCTACAATGGA ************************************************************ PDS-R2 AAGAGCATTCAATGATTTTTGCGATGCCAAACAAGCCAGGGGAATTTAGTCGCTTTGATT PDS-R5 AAGAGCATTCAATGATTTTTGCGATGCCAAACAAGCCAGGGGAATTTAGTCGCTTTGATT PDS-R1 AAGAGCATTCAATGATTTTTGCGATGCCAAACAAGCCAGGGGAATTTAGTCGCTTTGATT PDS-S AAGAGCATTCAATGATTTTTGCGATGCCAAACAAGCCAGGGGAATTTAGTCGCTTTGATT PDS-R4 AAGAGCATTCAATGATTTTTGCGATGCCAAACAAGCCAGGGGAATTTAGTCGCTTTGATT PDS-R3 AAGAGCATTCAATGATTTTTGCGATGCCAAACAAACCAGGGGAATTTAGTCGCTTTGATT **********************************.************************* PDS-R2 TTCCAGAAGTACTTCCTGCTCCACTAAATGGAATATGGGCAATCCTTAAAAACAATGAAA PDS-R5 TTCCAGAAGTACTTCCTGCTCCACTAAATGGAATATGGGCAATCCTTAAAAACAATGAAA PDS-R1 TTCCAGAAGTACTTCCTGCTCCACTAAATGGAATATGGGCAATCCTTAAAAACAATGAAA PDS-S TTCCAGAAGTACTTCCTGCTCCACTAAATGGAATATGGGCAATCCTTAAAAACAATGAAA PDS-R4 TTCCAGAAGTACTTCCTGCTCCACTAAATGGAATATGGGCAATCCTTAAAAACAATGAAA PDS-R3 TTCCAGAAGTACTTCCTGCTCCACTAAATGGAATTTGGGCAATCCTTAAAAACAATGAAA **********************************:************************* PDS-R2 TGCTCACTTGGCCAGAGAAAGTGCAATTTGCTATTGGACTACTACCTGCAATGATTGGGG PDS-R5 TGCTCACTTGGCCAGAGAAAGTGCAATTTGCTATTGGACTACTACCTGCAATGATTGGGG PDS-R1 TGCTCACTTGGCCAGAGAAAGTGCAATTTGCTATTGGACTACTACCTGCAATGATTGGGG PDS-S TGCTCACTTGGCCAGAGAAAGTGCAATTTGCTATTGGACTACTACCTGCAATGATTGGGG PDS-R4 TGCTCACTTGGCCAGAGAAAGTGCAATTTGCTATTGGACTACTACCTGCAATGATTGGGG PDS-R3 TGCTCACTTGGCCAGAGAAAGTGCAATTTGCTATTGGACTACTACCTGCAATGATTGGGG ************************************************************ PDS-R2 GGCAGCCATATGTTGAAGCTCAGGATGGCTTAACAGTTCAAGAGTGGATGAGAAAACAGG PDS-R5 GGCAGCCATATGTTGAAGCTCAGGATGGCTTAACAGTTCAAGAGTGGATGAGAAAACAGG PDS-R1 GGCAGCCATATGTTGAAGCTCAGGATGGCTTAACAGTTCAAGAGTGGATGAGAAAACAGG PDS-S GGCAGCCATATGTTGAAGCTCAGGATGGCTTAACAGTTCAAGAGTGGATGAGAAAACAGG PDS-R4 GGCAGCCATATGTTGAAGCTCAGGATGGCTTAACAGTTCAAGAGTGGATGAGAAAACAGG PDS-R3 GGCAGCCATATGTTGAAGCTCAGGATGGCTTAACAGTTCAAGAGTGGATGAGAAAACAGG ************************************************************ Figure 4.3: Continued.

PAGE 96

83 PDS-R2 GTGTGCCGGATCGAGTCAATGACGAGGTTTTCATTGCAATGTCAAAGGCTCTTAACTTCA PDS-R5 GTGTGCCGGATCGAGTCAATGACGAGGTTTTCATTGCAATGTCAAAGGCTCTTAACTTCA PDS-R1 GTGTGCCGGATCGAGTCAATGACGAGGTTTTCATTGCAATGTCAAAGGCTCTTAACTTCA PDS-S GTGTGCCGGATCGAGTCAATGACGAGGTTTTCATTGCAATGTCAAAGGCTCTTAACTTCA PDS-R4 GTGTGCCGGATCGAGTCAATGACGAGGTTTTCATTGCAATGTCAAAGGCTCTTAACTTCA PDS-R3 GTGTGCCGGATCGAGTCAATGACGAGGTTTTCATTGCAATGTCAAAGGCTCTTAACTTCA ************************************************************ PDS-R2 TAAACCCTGATGAACTTTCCATGCAATGCATCCTGATTGCCTTAAACAGTTTCCTTCAGG PDS-R5 TAAACCCTGATGAACTTTCCATGCAATGCATCCTGATTGCCTTAAACCATTTCCTTCAGG PDS-R1 TAAACCCTGATGAACTTTCCATGCAATGCATCCTGATTGCCTTAAACAGTTTCCTTCAGG PDS-S TAAACCCTGATGAACTTTCCATGCAATGCATCCTGATTGCCTTAAACCGTTTCCTTCAGG PDS-R4 TAAACCCTGATGAACTTTCCATGCAATGCATCCTGATTGCCTTAAACAGTTTCCTTCAGG PDS-R3 TAAACCCTGATGAACTTTCCATGCAATGCATACTGATTGCCTTAAACCGTTTCCTTCAGG *******************************.***************..*********** PDS-R2 AAAAGCATGGGTCGAAGATGGCCTTTTTAGATGGTAATCCACCTGAAAGATTATGTAAGC PDS-R5 AAAAGCATGGGTCGAAGATGGCCTTTTTAGATGGTAATCCACCTGAAAGATTATGTAAGC PDS-R1 AAAAGCATGGGTCGAAGATGGCCTTTTTAGATGGTAATCCACCTGAAAGATTATGTAAGC PDS-S AAAAGCATGGGTCGAAGATGGCCTTTTTAGATGGTAATCCACCTGAAAGATTATGTAAGC PDS-R4 AAAAGCATGGGTCGAAGATGGCCTTTTTAGATGGTAATCCACCTGAAAGATTATGTAAGC PDS-R3 AAAAGCATGGGTCGAAGATGGCCTTTTTAGATGGTAATCCACCTGAAAGATTATGTAAGC ************************************************************ PDS-R2 CAATTGCTGATCACATCGAGTCATTGGGTGGCCAAGTCATCCTTAATTCCCGAATACAGA PDS-R5 CAATTGCTGATCACATCGAGTCATTGGGTGGCCAAGTCATCCTTAATTCCCGAATACAGA PDS-R1 CAATTGCTGATCACATCGAGTCATTGGGTGGCCAAGTCATCCTTAATTCCCGAATACAGA PDS-S CAATTGCTGATCACATCGAGTCATTGGGTGGCCAAGTCATCCTTAATTCCCGAATACAGA PDS-R4 CAATTGCTGATCACATCGAGTCATTGGGTGGCCAAGTCATCCTTAATTCCCGAATACAGA PDS-R3 CAATTGCTGATCACATCGAGTCATTGGGTGGCCAAGTCGTCCTTAATTCCCGAATACAGA **************************************.********************* PDS-R2 AGATTGAGCTGAATGCAGACAAATCCGTCAAGCATTTTGTGCTCACCAATGGAAATATAA PDS-R5 AGATTGAGCTGAATGCAGACAAATCCGTCAAGCATTTTGTGCTCACCAATGGAAATATAA PDS-R1 AGATTGAGCTGAGTGCAGACAAATCCGTCAAGCATTTTGTGCTCACCAATGGAAATATAA PDS-S AGATTGAGCTGAATGCAGACAAATCCGTCAAGCATTTTGTGCTCACCAATGGAAATATAA PDS-R4 AGATTGAGCTGAATGCAGACAAATCCGTCAAGCATTTTGTGCTCACCAATGGAAATATAA PDS-R3 AGATTGAGCTGAATGCAGACAAATCCGTCAAGCATTTTGTGCTCACCAATGGAAATATAA ************.*********************************************** PDS-R2 TAACAGGAGATGCATATGTATTTGCAACACCTGTTGATATCTTGAAGCTTCTGTTACCTG PDS-R5 TAACAGGAGATGCATATGTATTTGCAACACCTGTTGATATCTTGAAGCTTCTGTTACCTG PDS-R1 TAACAGGAGATGCATATGTATTTGCAACACCTGTTGATATCTTGAAGCTTCTGTTACCTG PDS-S TAACAGGAGATGCATATGTATTTGCAACACCTGTTGATATCTTGAAGCTTCTGTTACCTG PDS-R4 TAACAGGAGATGCATATGTATTTGCAACACCTGTTGATATCTTGAAGCTTCTGTTACCTG PDS-R3 TAACAGGAGATGCATATGTATTTGCAACACCTGTTGATATCTTGAAGCTTCTGTTACCTG ************************************************************ PDS-R2 AAGATTGGAAGGAGATTTCATATTTCAAAAAATTGGACAAGTTGGTTGGCGTACCTGTGA PDS-R5 AAGATTGGAAGGAGATTTCATATTTCAAAAAATTGGACAAGTTGGTTGGCGTACCTGTGA PDS-R1 AAGATTGGAAGGAGATTTCATATTTCAAAAAATTGGACAAGTTGGTTGGCGTACCTGTGA PDS-S AAGATTGGAAGGAGATTTCATATTTCAAAAAATTGGACAAGTTGGTTGGCGTACCTGTGA PDS-R4 AAGATTGGAAGGAGATTTCATATTTCAAAAAATTGGACAAGTTGGTTGGCGTACCTGTGA PDS-R3 AAGATTGGAAGGAGATTTCATATTTCAAAAAATTGGACAAGTTGGTTGGCGTACCTGTGA ************************************************************ Figure 4.3: Continued.

PAGE 97

84 PDS-R2 TAAATGTACACATATGGTTTGATAGGAAGTTGAAGAACACATACGATCATCTTCTTTTCA PDS-R5 TAAATGTACACATATGGTTTGATAGGAAGTTGAAGAACACATACGATCATCTTCTTTTCA PDS-R1 TAAATGTACACATATGGTTTGATAGGAAGTTGAAGAACACATACGATCATCTTCTTTTCA PDS-S TAAATGTACACATATGGTTTGATAGGAAGTTGAAGAACACATACGATCATCTTCTTTTCA PDS-R4 TAAATGTACACATATGGTTTGATAGGAAGTTGAAGAACACATACGATCATCTTCTTTTCA PDS-R3 TAAATGTACACATATGGTTTGATAGGAAGTTGAAGAACACATACGATCATCTTCTTTTCA ************************************************************ PDS-R2 GCAGGAGTCCACTGTTGAGCGTTTATGCAGACATGTCTGTTACATGCAAGGAATACTACA PDS-R5 GCAGGAGTCCACTGTTGAGCGTTTATGCAGACATGTCTGTTACATGCAAGGAATACTACA PDS-R1 GCAGGAGTCCACTGTTGAGCGTTTATGCAGACATGTCTGTTACATGCAAGGAATACTACA PDS-S GCAGGAGTCCACTGTTGAGCGTTTATGCAGACATGTCTGTTACATGCAAGGAATACTACA PDS-R4 GCAGGAGTCCACTGTTGAGCGTTTATGCAGACATGTCTGTTACATGCAAGGAATACTACA PDS-R3 GCAGGAGTCCACTGTTGAGCGTTTATGCAGACATGTCTGTTACATGCAAGGAATACTACA ************************************************************ PDS-R2 ATCCAAATCAATCCATGCTTGAGCTAGTATTTGCACCAGCAGAGAAATGGATTTCATGCA PDS-R5 ATCCAAATCAATCCATGCTTGAGCTAGTATTTGCACCAGCAGAGAAATGGATTTCATGCA PDS-R1 ATCCAAATCAATCCATGCTTGAGCTAGTATTTGCACCAGCAGAGAAATGGATTTCATGCA PDS-S ATCCAAATCAATCCATGCTTGAGCTAGTATTTGCACCAGCAGAGAAATGGATTTCATGCA PDS-R4 ATCCAAATCAATCCATGCTTGAGCTAGTATTTGCACCAGCAGAGAAATGGATTTCATGCA PDS-R3 ATCCAAATCAATCCATGCTTGAGCTAGTATTTGCACCAGCAGAGAAATGGATTTCATGCA ************************************************************ PDS-R2 GTGACAGTGAAATCATTAACGCGACTATGCAAGAGCTTGCTAAACTCTTTCCAGATGAGA PDS-R5 GTGACAGTGAAATCATTAACGCGACTATGCAAGAGCTTGCTAAACTCTTTCCAGATGAGA PDS-R1 GTGACAGTGAAATCATTAACGCGACTATGCAAGAGCTTGCTAAACTCTTTCCAGATGAGA PDS-S GTGACAGTGAAATCATTAACGCGACTATGCAAGAGCTTGCTAAACTCTTTCCAGATGAGA PDS-R4 GTGACAGTGAAATCATTAACGCGACTATGCAAGAGCTTGCTAAACTCTTTCCAGATGAGA PDS-R3 GTGACAGTGAAATCATTAACGCAACTATGCAAGAGCTTGCTAAACTCTTTCCAGATGAGA **********************.************************************* PDS-R2 TTTCTGCTGATCAAAGCAAGGCCAAAATTTTGAAATATCATGTTGTAAAGACCCCGAGGT PDS-R5 TTTCTGCTGATCAAAGCAAGGCCAAAATTTTGAAATATCATGTTGTAAAGACCCCGAGGT PDS-R1 TTTCTGCTGATCAAAGCAAGGCCAAAATTTTGAAATATCATGTTGTAAAGACCCCGAGGT PDS-S TTTCTGCTGATCAAAGCAAGGCCAAAATTTTGAAATATCATGTTGTAAAGACCCCGAGGT PDS-R4 TTTCTGCTGATCAAAGCAAGGCCAAAATTTTGAAATATCATGTTGTAAAGACCCCGAGGT PDS-R3 TTTCTGCTGATCAAAGCAAGGCCAAAATTTTGAAATATCATGTTGTAAAGACCCCGAGGT ************************************************************ PDS-R2 CAGTTTACAAGACGGTCCCTGATTGTGAACCATGCCGGCCTTTGCAAAGATCTCCAATTG PDS-R5 CAGTTTACAAGACGGTCCCTGATTGTGAACCATGCCGGCCTTTGCAAAGATCTCCAATTG PDS-R1 CAGTTTACAAGACGGTCCCTGATTGTGAACCATGCCGGCCTTTGCAAAGATCTCCAATTG PDS-S CAGTTTACAAGACGGTCCCTGATTGTGAACCATGCCGGCCTTTGCAAAGATCTCCAATTG PDS-R4 CAGTTTACAAGACGGTCCCTGATTGTGAACCATGCCGGCCTTTGCAAAGATCTCCAATTG PDS-R3 CAGTTTACAAGACGGTCCCTGATTGTGAACCATGCCGGCCTTTGCAAAGATCTCCAATTG ************************************************************ PDS-R2 AAGGGTTCTACTTGGCTGGTGACTACACAAAGCAGAAGTATTTGGCCTCAATGGAAGGTG PDS-R5 AAGGGTTCTACTTGGCTGGTGACTACACAAAGCAGAAGTATTTGGCCTCAATGGAAGGTG PDS-R1 AAGGGTTCTACTTGGCTGGTGACTACACAAAGCAGAAGTATTTGGCCTCAATGGAAGGTG PDS-S AAGGGTTCTACTTGGCTGGTGACTACACAAAGCAGAAGTATTTGGCCTCAATGGAAGGTG PDS-R4 AAGGGTTCTACTTGGCTGGTGACTACACAAAGCAGAAGTATTTGGCCTCAATGGAAGGTG PDS-R3 AAGGGTTCTACTTGGCTGGTGACTACACAAAGCAGAAGTATTTGGCCTCAATGGAAGGTG ************************************************************ Figure 4.3: Continued.

PAGE 98

85 PDS-R2 CCGTGTTATCTGGGAAGCTATGTGCTCAGGCAATTGTGCAGGACTGCAGCTTGTTGGCTT PDS-R5 CCGTGTTATCTGGGAAGCTATGTGCTCAGGCAATTGTGCAGGACTGCAGCTTGTTGGCTT PDS-R1 CCGTGTTATCTGGGAAGCTATGTGCTCAGGCAATTGTGCAGGACTGCAGCTTGTTGGCTT PDS-S CCGTGTTATCTGGGAAGCTATGTGCTCAGGCAATTGTGCAGGACTGCAGCTTGTTGGCTT PDS-R4 CCGTGTTATCTGGGAAGCTATGTGCTCAGGCAATTGTGCAGGACTGCAGCTTGTTGGCTT PDS-R3 CCGTGTTATCTGGGAAGCTATGTGCTCAGGCAATTGTGCAGGACTGCAGCTTGTTGGCTT ************************************************************ PDS-R2 CTAGGGTACAGAAAAGCCCACAGACGTTGACGATTGCCTGATTCAGGAAACTTTTATGCA PDS-R5 CTAGGGTACAGAAAAGCCCACAGACGTTGACGATTGCCTGATTCAGGAAACTTTTATGCA PDS-R1 CTAGGGTACAGAAAAGCCCACAGACGTTGACGATTGCCTGATTCAGGAAACTTTTATGCA PDS-S CTAGGGTACAGAAAAGCCCACAGACGTTGACGATTGCCTGATTCAGGAAACTTTTATGCA PDS-R4 CTAGGGTACAGAAAAGCCCACAGACGTTGACGATTGCCTGATTCAGGAAACTTTTATGCA PDS-R3 CTAGGGTGCAGAAAAGCCCACAGACGTTGACGATTGCCTGATTCAGGAAACTTTTACGCA *******.************************************************ *** PDS-R2 GGTTCAGTTTGTAGGGGGAATATTTCTGGTTTTGTTTCATTCAGATGTTTTTCTTTTAGA PDS-R5 GGTTCAGTTTGTAGGGGGAATATTTCTGGTTTTGTTTCATTCAGATGTTTTTCTTTTAGA PDS-R1 GGTTCAGTTTGTAGGGGGAATATTTCTGGTTTTGTTTCATTCAGATGTTTTTCTTTTAGA PDS-S GGTTCAGTTTGTAGGGGGAATATTTCTGGTTTTGTTTCATTCAGATGTTTTTCTTTTAGA PDS-R4 GGTTCAGTTTGTAGGGGGAATATTTCTGGTTTTGTTTCATTCAGATGTTTTTCTTTTAGA PDS-R3 -----------------------------------------------------------PDS-R2 GCATATGTCTTTATAGTAAAAACTCCCACCTCTTTCTCATGTATAGCTACATCAGCAAGC PDS-R5 GCATATGTCTTTATAGTAAAAACTCCCACCTCTTTCTCATGTATAGCTACATCAGCAAGC PDS-R1 GCATATGTCTTTATAGTAAAAACTCCCACCTCTTTCTCATGTATAGCTACATCAGCAAGC PDS-S GCATATGTCTTTATAGTAAAAACTCCCACCTCTTTCTCATGTATAGCTACATCAGCAAGC PDS-R4 GCATATGTCTTTATAGTAAAAACTCCCACCTCTTTCTCATGTATAGCTACATCAGCAAGC PDS-R3 -----------------------------------------------------------PDS-R2 AAAGGGGGTAAAGGGC-------------------------------------------PDS-R5 AAAGGGGGTAAAGGGCGAAT---------------------------------------PDS-R1 AAAGGGGGTAAAGGGCGAATTCGTTTAAACCTGCAGGACTAGTCCCTTTAGTGAGGGTAA PDS-S AAA--------------------------------------------------------PDS-R4 AAAGGGGGTAAAGGGCGAATTCGTTT---------------------------------PDS-R3 -----------------------------------------------------------Figure 4.3: Continued.

PAGE 99

86 PDS-R2 MTVARSVVAVNLSGSLQNRYPASSSVSCFLGKEYRCNSMLGFCGSGKLAFGANAPYSKIA PDS-R4 MTVARSVVAVNLSGSLQNRYPASSSVSCFLGKEYRCNSMLGFCGSGKLAFGANAPYSKIA PDS-R1 MTVARSVVAVNLSGSLQNRYPASSSVSCFLGKEYRCNSMLGFCGSGKLAFGANAPYSKIA PDS-R5 MTVARSVVAVNLSGSLQNRYPASSSVSCFLGKEYRCNSMLGFCGSGKLAFGANAPYSKIA PDS-S MTVARSVVAVNLSGSLQNRYPASSSVSCFLGKEYRCNSMLGFCGSGKLAFGANAPYSKIA PDS-R3 MTVARSVVAVNLSGSLQNRYPASSSVSCFLGKEYRCNSMLGFCGSGKLAFGANAPYSKIA ************************************************************ PDS-R2 ATKPKPKLRPLKVNCMDFPRPDIDNTANFLEAAALSSSFRNSARPSKPLQVVIAGAGLAG PDS-R4 ATKPKPKLRPLKVNCMDFPRPDIDNTANFLEAAALSSSFRNSARPSKPLQVVIAGAGLAG PDS-R1 ATKPKPKLRPLKVNCMDFPRPDIDNTANFLEAAALSSSFRNSARPSKPLQVVIAGAGLAG PDS-R5 ATKPKPKLRPLKVNCMDFPRPDIDNTANFLEAAALSSSFRNSARPSKPLQVVIAGAGLAG PDS-S ATKPKPKLRPLKVNCMDFPRPDIDNTANFLEAAALSSSFRNSARPSKPLQVVIAGAGLAG PDS-R3 ATKPKPKLRPLKVNCMDFPRPDIDNTVNFLEAAALSSSFRNSARPSKPLQVVIAGAGLAG **************************.********************************* PDS-R2 LSTAKYLADAGHIPILLEARDVLGGKVAAWKDDDGDWYETGLHIFFGAYPNVQNLFGELG PDS-R4 LSTAKYLADAGHIPILLEARDVLGGKVAAWKDDDGDWYETGLHIFFGAYPNVQNLFGELG PDS-R1 LSTAKYLADAGHIPILLEARDVLGGKVAAWKDDDGDWYETGLHIFFGAYPNVQNLFGELG PDS-R5 LSTAKYLADAGHIPILLEARDVLGGKVAAWKDDDGDWYETGLHIFFGAYPNVQNLFGELG PDS-S LSTAKYLADAGHIPILLEARDVLGGKVAAWKDDDGDWYETGLHIFFGAYPNVQNLFGELG PDS-R3 LSTAKYLADAGHIPILLEARDVLGGKVAAWKDDDGDWYETGLHIFFGAYPNVQNLFGELG ************************************************************ PDS-R2 INDRLQWKEHSMIFAMPNKPGEFSRFDFPEVLPAPLNGIWAILKNNEMLTWPEKVQFAIG PDS-R4 INDRLQWKEHSMIFAMPNKPGEFSRFDFPEVLPAPLNGIWAILKNNEMLTWPEKVQFAIG PDS-R1 INDRLQWKEHSMIFAMPNKPGEFSRFDFPEVLPAPLNGIWAILKNNEMLTWPEKVQFAIG PDS-R5 INDRLQWKEHSMIFAMPNKPGEFSRFDFPEVLPAPLNGIWAILKNNEMLTWPEKVQFAIG PDS-S INDRLQWKEHSMIFAMPNKPGEFSRFDFPEVLPAPLNGIWAILKNNEMLTWPEKVQFAIG PDS-R3 INDRLQWKEHSMIFAMPNKPGEFSRFDFPEVLPAPLNGIWAILKNNEMLTWPEKVQFAIG ************************************************************ PDS-R2 LLPAMIGGQPYVEAQDGLTVQEWMRKQGVPDRVNDEVFIAMSKALNFINPDELSMQCILI PDS-R4 LLPAMIGGQPYVEAQDGLTVQEWMRKQGVPDRVNDEVFIAMSKALNFINPDELSMQCILI PDS-R1 LLPAMIGGQPYVEAQDGLTVQEWMRKQGVPDRVNDEVFIAMSKALNFINPDELSMQCILI PDS-R5 LLPAMIGGQPYVEAQDGLTVQEWMRKQGVPDRVNDEVFIAMSKALNFINPDELSMQCILI PDS-S LLPAMIGGQPYVEAQDGLTVQEWMRKQGVPDRVNDEVFIAMSKALNFINPDELSMQCILI PDS-R3 LLPAMIGGQPYVEAQDGLTVQEWMRKQGVPDRVNDEVFIAMSKALNFINPDELSMQCILI ************************************************************ PDS-R2 ALNSFLQEKHGSKMAFLDGNPPERLCKPIADHIESLGGQVILNSRIQKIELNADKSVKHF PDS-R4 ALNSFLQEKHGSKMAFLDGNPPERLCKPIADHIESLGGQVILNSRIQKIELNADKSVKHF PDS-R1 ALNSFLQEKHGSKMAFLDGNPPERLCKPIADHIESLGGQVILNSRIQKIELSADKSVKHF PDS-R5 ALNHFLQEKHGSKMAFLDGNPPERLCKPIADHIESLGGQVILNSRIQKIELNADKSVKHF PDS-S ALNRFLQEKHGSKMAFLDGNPPERLCKPIADHIESLGGQVILNSRIQKIELNADKSVKHF PDS-R3 ALNRFLQEKHGSKMAFLDGNPPERLCKPIAAHIESLGGQVVLNSRIQKIELNADKSVKHF *** ************************** *********:**********.******** PDS-R2 VLTNGNIITGDAYVFATPVDILKLLLPEDWKEISYFKKLDKLVGVPVINVHIWFDRKLKN PDS-R4 VLTNGNIITGDAYVFATPVDILKLLLPEDWKEISYFKKLDKLVGVPVINVHIWFDRKLKN PDS-R1 VLTNGNIITGDAYVFATPVDILKLLLPEDWKEISYFKKLDKLVGVPVINVHIWFDRKLKN PDS-R5 VLTNGNIITGDAYVFATPVDILKLLLPEDWKEISYFKKLDKLVGVPVINVHIWFDRKLKN PDS-S VLTNGNIITGDAYVFATPVDILKLLLPEDWKEISYFKKLDKLVGVPVINVHIWFDRKLKN PDS-R3 VLTNGNIITGDAYVFATPVDILKLLLPEDWKEISYFKKLDKLVGVPVINVHIWFDRKLKN ************************************************************ Figure 4-4. Aligned sequences of PDS protein from different hydrilla populations. Amino acid changes in resistant populations are denoted in bold letters.

PAGE 100

87 PDS-R2 TYDHLLFSRSPLLSVYADMSVTCKEYYNPNQSMLELVFAPAEKWISCSDSEIINATMQEL PDS-R4 TYDHLLFSRSPLLSVYADMSVTCKEYYNPNQSMLELVFAPAEKWISCSDSEIINATMQEL PDS-R1 TYDHLLFSRSPLLSVYADMSVTCKEYYNPNQSMLELVFAPAEKWISCSDSEIINATMQEL PDS-R5 TYDHLLFSRSPLLSVYADMSVTCKEYYNPNQSMLELVFAPAEKWISCSDSEIINATMQEL PDS-S TYDHLLFSRSPLLSVYADMSVTCKEYYNPNQSMLELVFAPAEKWISCSDSEIINATMQEL PDS-R3 TYDHLLFSRSPLLSVYADMSVTCKEYYNPNQSMLELVFAPAEKWISCSDSEIINATMQEL ************************************************************ PDS-R2 AKLFPDEISADQSKAKILKYHVVKTPRSVYKTVPDCEPCRPLQRSPIEGFYLAGDYTKQK PDS-R4 AKLFPDEISADQSKAKILKYHVVKTPRSVYKTVPDCEPCRPLQRSPIEGFYLAGDYTKQK PDS-R1 AKLFPDEISADQSKAKILKYHVVKTPRSVYKTVPDCEPCRPLQRSPIEGFYLAGDYTKQK PDS-R5 AKLFPDEISADQSKAKILKYHVVKTPRSVYKTVPDCEPCRPLQRSPIEGFYLAGDYTKQK PDS-S AKLFPDEISADQSKAKILKYHVVKTPRSVYKTVPDCEPCRPLQRSPIEGFYLAGDYTKQK PDS-R3 AKLFPDEISADQSKAKILKYHVVKTPRSVYKTVPDCEPCRPLQRSPIEGFYLAGDYTKQK ************************************************************ PDS-R2 YLASMEGAVLSGKLCAQAIVQDCSLLASRVQKSPQTLTIA PDS-R4 YLASMEGAVLSGKLCAQAIVQDCSLLASRVQKSPQTLTIA PDS-R1 YLASMEGAVLSGKLCAQAIVQDCSLLASRVQKSPQTLTIA PDS-R5 YLASMEGAVLSGKLCAQAIVQDCSLLASRVQKSPQTLTIA PDS-S YLASMEGAVLSGKLCAQAIVQDCSLLASRVQKSPQTLTIA PDS-R3 YLASMEGAVLSGKLCAQAIVQDCSLLASRVQKSPQTLTIA **************************************** Figure 4.4: Continued.

PAGE 101

88 Table 4-1. Nucleotide differences, along with amino acid changes, in pds sequences in resistant hydrilla populations comp ared with fluridone-sensitive pds sequence at planting (Sept 2004). Corres ponding amino acid positions in pds where available, are in super-script. 1 Nucleotide Change; 2 Altered Amino Acid.

PAGE 102

89 Table 4-2. Nucleotide differences, along with amino acid changes, in pds sequences in resistant hydrilla populations comp ared with fluridone-sensitive pds sequence at 12 MAP (Sept 2005). Correspondi ng amino acid positions in pds where available, are in super-script 1 Nucleotide Change; 2 Altered Amino Acid.

PAGE 103

90 Arg Hydrilla 875 tcca-tgcaatgcatcctgattgccttaaaccgtttccttcaggaaaagcatg Arabidopsis 955 tcaatgcaatg-cattttgatagctttgaaccggtttcttcaggaaaaacatg Pisum 319 tcaatgcaatg-tattttgattgcattaaaccgatttcttcaggagaagcatg Glycine 1060 tcaatgcaatgta-tattgattgctttaaaccgatttcttcaggagaaacatg Dunaliella 762 tcta-tgaccgttgtgctaacagcactgaaccgtttcctgcaagagcgacatg Synechocystis 648 tccg-ccacggtcgtcctaacggcactcaaccgcttcttgcaagagaagaaag Lycopersicon 1278 tcaatgcagtg-cattttgatcgcattgaacaggtttcttcaggagaaacatg Papaver 39 tcgatgcagtg-cattttgatagctttgaaccgtttccttcaggaaaagcatg Figure 4-5. Alignment of par tial phytoene desaturase gene sequences from various organisms showing arginine codon at the amino acid position 304 of hydrilla pds gene.

PAGE 104

91 CHAPTER 5 HYDRILLA PLOIDY IN RELATION TO FLURIDONE RESISTANCE Introduction The submersed aquatic weed, Hydrilla verticillata belonging to the monocot family Hydrocharitaceae, is highly poly morphic. Cook and Lnd (1982) provided a taxonomic revision of this single species ge nus providing general information its ecology, floral biology, anatomy, chromo somes, genetics, and variatio n. Intraspecific variation of H. verticillata has been reported in leaf and stem morphology, sex expression and chromosome numbers (Verkleij et al., 1983a ; Pieterse et al., 1985). Both monoecious (staminate and pistillate flowers on the same plant) and dioecious (staminate and pistillate flowers on separate plants) biotypes have been described (Cook and Lnd, 1982), and both are present in the US. Cook and Lnd ( 1982) report that on a worldwide basis, the monoecious form dominates in climatically tropical regions, whereas the dioecious forms are largely temperate. However, the current distribution and estimates of potential distribution for both monoecious and dioecious biotypes in North America are contrary to this observation. The dioecious form was first re ported in the US in south Florida in the late 1950’s, while the monoecious form was fi rst reported in the Po tomac River in the mid 1980’s (Steward et al., 1984). Genetic diversity within a plant speci es and the genetic structure of plant populations depend on life history and ecol ogical factors. Genetic diversity on the population level is usually higher in terrestria l than in aquatic plan ts. Hydrilla does not sexually reproduce in Florida and the southeas tern U.S.; presumably due to presence of

PAGE 105

92 only the dioecious female form (Haller, 1976). In general, sexually reproducing species are characterized by high genetic diversit y compared with vegetatively propagated species (Loveless and Hamrick, 1984). However, hydrilla is regarded as a species of high genetic variation (Triest, 1991). Genetic vari ation in hydrilla stra ins from the U.S. (Verkleij et al., 1983b), Africa (P ieterse et al., 1985), and ot her geographic regions have been compared using isozyme studies (Ver kleij and Pieterse, 1986; Nakamura et al., 1998), and random polymorphic DNA (RAPD) assa ys (Les et al., 1997; Madeira et al., 1997; Hofstra et al., 2000). The purpose wa s to identify new hydrilla strains and infestations, and to determine genetic re lationship between geographically diverse hydrilla populations. Variation within populations tends to be low for aquatic plants, but strong differentiation among populations s eems common (Hofstra et al., 1995 ) . The low levels of within-population varia tion and the high population diffe rentiation are most likely related to widespread clonal multiplication. Cl onal multiplication is like ly to result in low genotypic diversity, but does not necessarily imply low gene tic diversity, and genotypic diversity values of aquatic a nd terrestrial asexually propagate d plants seems to be similar (Lokker et al., 1994). Furthermore, differe nces among populations may result from adaptive responses to local diffe rences in selection pressure . Population differentiation in morphological and physiological traits has been reported for a number of aquatic species in response to various external growth a nd reproductive factors (Barrett et al., 1993). Various hydrilla populations from Europe show ed distinct isozyme patterns compared to other hydrilla strains (Pieters e et al., 1985), which may be due to differential ecological adaptation and genetic drift. Isozyme patterns of African plants are not distinctive

PAGE 106

93 presumably due to local infestations and the non-aggressive nature of hydrilla on the continent. Large variation from isozyme patterns has been observed in hydrilla collections from different parts of Southeast Asia and also between hydrilla plants within the same lake Curug (Verkleij and Pieterse, 1986). This study produced distinct clusters of plants from the island of Java (Indonesia), which is su spected to be the center of differentiation of this noxious weed species (Cook and Lnd, 1982). High levels of seed production have been reported in Penang Is land, indicating sexual re production is more important in Southeast Asia (M adeira et al., 1997). The gene tic closeness of African and a group of plants from the Indian subcontin ent have also been reported (Verkleij and Pieterse, 1991). The U.S. dioecious and m onoecious biotypes have different isozyme patterns (Ryan et al., 1991). The dioecious hydr illa populations in the United States are more closely related to hydrilla strains from Bangalore, India, while monoecious forms are genetically closer to hydr illa strains from Seoul, Korea (Madiera et al., 1997). Comparative studies have al so been conducted to determ ine the differentiation of ecological and physiological tra its of various biotypes of H. verticillata within the United States (Spencer and Anders on, 1986; Steward and Van, 1987). It has been documented that hydrilla is a polyploid plant (Cook and Lnd, 1982; Verkleij et al., 1983a). Polyploidy is a prominent process and has been important in the evolutionary history of plants . Recent estimates suggest that 70% of angiosperms have experienced one or more episodes of polyploidization (Masterson, 1994). Hydrilla p lants in Asia, India and Europe are either diploid (2n= 2x= 16) or triplo id (2n= 3x= 24) (Cook and Lnd, 1982). Presence of tetraploid plan ts (2n= 4x= 32) has been reported in Alabama (Davenport, 1980). Both diploid and tr iploid plants have been collected from

PAGE 107

94 different parts of Washington DC, Mary land, and Texas (Verkleij et al., 1983a; Langeland, 1989); and those collected from Calif ornia, Florida, Texas, and Connecticut have been recorded as triploid (Harlan et al., 1985; Langeland, 1989). Dioecious female populations have been observed in the south eastern USA, California, Texas, and Poland while male populations have been reporte d in Malaysia, Indonesia, and Panama. Monoecious hydrilla populations occur in India, Indonesia, Virginia, Maryland, and North Carolina. Occurrence of the monoecious plants in the USA with reported viable seed production might increase genetic variability (Langeland and Smith, 1984). The relationship between hydrilla morphol ogy and karyotype has been previously studied (Verkleij et al., 1983a; Chaudhuri and Sharma, 1978). Chaudhuri and Sharma (1978) observed differences in the chromosome structure of eight populations within a range of 50 Km in India, but the difference s were not related to whole plant morphology. Langeland (1989) recorded no morphological diff erences in the karyotype of monoecious and dioecious hydrilla populat ions. Various dioecious hydri lla populations in Japan are either diploid or triploid, whereas monoeci ous forms in Japan are always triploid (Nakamura and Kadono, 1993), and mostly repro duce by vegetative means. However in different parts of the U.S., the presence of di ploid and triploid plan ts has been reported within the monoecious form (Harlan et al., 1985 ). It has been sugge sted that hydrilla may be an endopolyploid plant, wh ich explains the occurrence of different ploidy levels from plants within the same population (Langeland, 1989). Some species apparently have no consistent ploidy level in de veloping root tissue and the pr esence of cells of different chromosome numbers is common in root tissu e (Wardlaw, 1968). This appears to be the case with hydrilla. Research has shown the endopolyploid nature of hydrilla (Sharma and

PAGE 108

95 Bhattacharya, 1956; Chaudhuri and Sharma, 1978), and various combinations of diploid, triploid, and tetraploid cells were observed in the same root tips (Langeland et al., 1992). While the lack of sexual recombination in hydrilla suggests that development of resistance to an herbicide is unlikely, th e existence of endopolyploid populations offers the species an opportunity to have various clones that may react differently to environmental conditions. These large geneti c variations within a population along with endopolyploidy may imply differences in surv ival strategy, and offer the species an opportunity to have various clones that may react differently to environmental stress (Hofstra et al., 2000). One such stress is the presence of an herbicide, particularly long term exposure as in the case of fluridone. In addition to the assessm ent of local hydrilla populations for determining the effectiveness of control efforts, genetic information about the populations of invasive species such as hydr illa can also play an important role in understanding the adaptive potential of the species. Flow cytometry is one of the most powerful and specific methods for the integrated study of molecular and morphol ogical events occurri ng during cell growth. This technique is used widely for determ ining amounts of nuclear DNA (Nandini et al., 1997; Lysk et al., 2000) and can also be us ed to determine (DNA) ploidy (Emshwiller, 2002), although cytological studies are requir ed for confirmation (Bennett et al., 2000). Because the nuclear DNA content of the G1 nucleus reflects the ploidy of a cell, estimation of DNA content is an indicator of the ploidy level. The purpose of this experiment was to determine the possible ploidy variations (o r the presence of endopolyploidy) among different hydr illa populations with varying levels of resistance to the herbicide fluridone by using flow cytometry.

PAGE 109

96 Materials and Methods Plant Materials Hydrilla populations collected from differe nt lakes in Florida were cultured in outdoor vaults at the Weed Science Building, Un iversity of Florida in Gainesville, FL and served as stock plants for this study. Cytological Preparation Rapidly growing root tips were taken from growing hydrilla plants between 7.00 and 9.00 A.M. to obtain the highest percentage of dividing cells in metaphase. Harvested root tips were placed directly into 2 mM 8-hydroxyquinoline and maintained at 10C for 4-5 hours. Roots were then rinsed once with deionized water and then fixed with 3:1 ethanol/acetic acid solution fo r 24-28 h. After fixation, root s were rinsed with 70% ethanol three times, and then hydrolyzed w ith 1N HCl for 15-18 min. After hydrolysis, roots were rinsed with dei onized water three times to remove HCl, and stained with Feulgen stain in the dark at 25C for a minimum of 30 h. These were immediately rinsed with deionized water before microscopic ev aluation. A drop of Orcein stain was applied on the slide before squashing fo r optimal chromosomal staining. Flow Cytometry Procedure and Sta ndardization of Flow Cytometry The ploidy level of the hydrilla plants was determined by flow cytrometric analysis with a Partec PAS flow cytometer (Partec, http://www.partec.de/) equipped with a mercury lamp and filter combinatio ns of KG1, BG38, UG1, TK420, and GG435. Young growing hydrilla shoot tip s (apical meristematic tissu e) were collected from different hydrilla populations and prep ared for flow cytometry analysis. D etermination of nuclear DNA content utilized 4, 6-diamino-2-phenylindole dihydrochloride (DAPI) as the fluorescent stain. Analyses were performed on 25 hydrilla plants of each of six hydrilla

PAGE 110

97 populations. Samples of apical meristematic tissue of hydrilla shoot tips (0.5 cm) were used for analysis. Leaf material was chopped with a sharp razor bl ade for at least 60 s with 0.4 mL DAPI. The resulting extract was passed through a 20 m filter into a 3.5 mL plastic tube. Additional DAPI (1.6 mL) was then added to the tube. Samples were kept at room temperature for 6-8 min before analys is by flow cytometry. For each sample, at least 10,000 nuclei were analyzed. The precision and linearity of the flow cytometer were checked on a daily basis using the sta ndard procedure of the manufacturer1 . The results were displayed as one-parameter DNA histogr ams. Peak positions of DNA histograms were compared with standard peaks derived fr om leaf tissue of a control triploid plant. Ploidy of control triploid plants was determined usi ng cytological preparations. Histograms were analyzed using a DPAC v.2.2 computer program (Partec Mnster, Germany) and the percentage of nuclei of particular DNA contents was calculated. The ploidy level of a triploid plan t (2n= 3x= 24) was determined using cytological preparation and this plant was used as the standard in flow cytometry analysis. Chicken blood cell nuclei were used as an internal standard. Nucleated chicken red blood cells (CRBC) have a known genome size of 2.33 picograms (pg), a nd were used to calculate genome size of hydrilla plants. The absolute DNA content of a hydrilla plant sample was calculated based on the values of the G1 peak means: Sample 2C DNA content = [(Sample G1 peak mean)/(CRBC G1 peak mean)] x CRBC 2C DNA content (2.33 pg DNA) 1 Partec Mnster , Genmany (http://www.partec.de/).

PAGE 111

98 Results and Discussion Chromosome counting and flow cytometr y are both considered to be accurate methods for ploidy analysis ( Bennett et al., 2000; Emshwiller, 2002) . Flow cytometric analysis revealed histograms of DAPI-stain ed nuclei that showed one dominant peak, which indicated the ploidy level of the plant tissue sample (Figure 5-1). We adjusted the G1 peak of our standard triploid on channel 300 in flow cytometric analysis to aid in analysis of tissue samples of unknown ploidy. Representative histograms of flow cytometric analysis of tissue samples from a diploid, triploid, and te traploid hydrilla plant are presented in Figure 5-1. Various combinations of diploi d, triploid and tetraploid plants were recorded in diffe rent hydrilla populations (Tab le 5-1). In hydrilla population S, only triploid (28%) and tetraploid plan ts (72%) were observe d and no diploid plant was found while R2, and R3 recorded presen ce of only diploid and triploid hydrilla plants. There were 20% and 38% diploid plants and 80% and 62% triploid plants in R1 and R2, respectively (Table 5-2) . In hydrilla populations R1, R4, and R5, plants with all the three known levels of pl oidy (2x, 3x, and 4x) were obser ved. Hydrilla populations R1, R4, and R5 contained 20%, 20%, and 36% di ploid plants, 64%, 76%, and 56% triploid plants and 16%, 4% and 8% tetraploid hydr illa plants, respectively (Table 5-2). The coefficient of variation of cytometry values , as a parameter of the reliability of the measurement, varied between 2% and 8% and was rarely <1%. These values agree with the observations of Galbraith (1990) who re ported CV values below 10% were fully acceptable for difficult species such as hydrilla. Hydrilla populations in Flor ida lakes are reported to be dioecious, with only female form present (Langeland, 1990) and di ploid and triploid plants in dioecious hydrilla populations have been previously reported (Langeland et al., 1992; Nakamura

PAGE 112

99 and Kodono, 1993). Davenport (1980) reported th e presence of dioe cious tetraploid hydrilla in Alabama. Langeland et al. (1992) re ported occurrence of di ploid, triploid and tetraploid cells in root tissue of hydrilla in hydrilla populations from different parts of the world. This suggests discrepancie s in the ploidy of different hydrilla plants in the same hydrilla population or plants from different populations. In our study, triploid plants were dominant in all populations. Diploid plants were observed in all the hydril la populations except S and tetraploid plants were rare within populations. There were no obvious morphological differences among plants of di fferent ploidy levels within a population and among different hydrilla popul ations (personal observation). Table 5-2 shows the estimations of nucl ear DNA content for the diploid, triploid and tetraploid plants in different hydr illa populations. There were no significant differences in nuclear DNA content among plants with the same ploidy levels ( P 0.05) in different populations, but hi gher nuclear contents were obs erved with increased ploidy from 2x (diploidy) to 4x (tetraploidy). The diploid, triploid and te traploid plants had nuclear DNA content of 2.43.73 pg, 3.44.71 pg, and 4.64.90 pg, respectively (Table 5-2). The presence of cells of different pl oidy levels in somatic tissue is called endoreduplication. In contrast to dividing cel ls, endoreduplicating cells are not believed to undergo mitosis, and in such cells nucle ar DNA content successively doubles from 2x to 4x to 8x, etc. Endoreduplication patterns we re observed only in diploid plants of populations S and R3. In the susceptible populati on, two plants out of twenty-five, and in R3, six plants out of twenty-five plants te sted recorded endore duplication patterns and cells with 4x nuclear DNA content were obser ved (Table 5-1 and Fi gure 5-2). None of

PAGE 113

100 the plants with higher ploidy levels (3x or 4x) in any hydrilla populations showed patterns of endoreduplication. One pg of DNA equals 965 million base pairs (Mb) (Bennet and Smith, 1976). Therefore, th e genome size of hydrilla might be 2345 Mb in diploid, 3320 Mb in triploid, and 4480 Mb in te traploid hydrilla plants, respectively. Polyploidy is a prominent process and ha s been important in the evolutionary history of plants. The chromosome number in crement and polyploidization are important in the evolution of plants (Yang, 2001). During differentiation of plant tissues, cells of various ploidy levels occur in one and th e same organ, which is called polysomaty (Smulders et al., 1994). Cell polyp loidization is mainly due to either endomitosis or endoreduplication, and it occurs in over 90% of angiosperms. During the process of polyploidization in nature, sexual polyploidi zation is the principle mode, in which 2 n gametes play the pivotal role. Polysomaty in different organs of diploid, triploid and tetraploid sugarbeet dry seed s and in seedlings during thei r early development in tissue culture (starting from radicle protrusion up to expansion of the first pair of leaves) was studied using flow cytometry (Sliwinsk a and Lukaszewska, 2005). The endopolyploidy patterns in diploid and tetrap loid plants of maize (Birad ar et al., 1993), and tomato (Smulders et al., 1995) were also previous ly established using flow cytometry. There are, however, no reports comparing polysomaty on the three different ploidy levels of an asexually propagated pl ant species. We reported for the first time evidence of differential ploidy among different plants within an asexually propagated clonal species.

PAGE 114

101 Some species apparently have no consistent ploidy level in the developing root tissue and this was also observed with hydril la where presence of diploid, triploid, and tetraploid cells was detected. There have also been reports of cyto logical changes from diploid to triploid plants occurring many times independently in vari ous hydrilla strains in Japan (Nakamura et al., 1998), and in the U.S. (Langeland, 1989); but no valid explanation has been given. The presence of similar isozyme patterns and alleles in diploid and triploid plants from various hydrilla accessions (Madeira et al., 1997; Nakamura et al., 1998) might explain these ch anges. The advanced understanding of the nuclear genome and its components may e xplain the mechanisms which could be responsible for genome increase, including the activation of Class I retrotransposons (Bennetzen et al., 2005). Research has shown the sexual compatib ility of both monoecious and dioecious biotypes and production of viab le seeds by dioecious female plants from the crosses between dioecious pollen donor s and monoecious pollen donors under culture conditions (Steward, 1993). The chances of sexual repr oduction in hydrilla are reduced because of limited co-occurrence of female and male pl ants. However, coexistence, and sexual compatibility between different hydrilla strain s (with varying ploidy) also might increase genetic variation in hydrilla populations, ther efore having a signifi cant influence on the adaptability of this noxious weed under di fferent environmental conditions including control measures, such as the presence of an herbicide. The only other possible explanation of the presence of diploid and triploid plants within dioecious hydrilla populations could be due to the separate in troduction of diploid and triploid hydrilla plants in Florida.

PAGE 115

102 The more frequent occurrence of triplo id plants also suggests an ecological advantage (Nakamura et al., 1998). Presence of fewer tetraploid plants in different hydrilla populations (Langeland et al., 1992; Table 5-2) suggested a possible evolution from diploid hydrilla plants over a period of time due to endoreduplication. Nagl (1976) suggested that DNA endoredupli cation represents an evolutionary strategy which substitutes for a lack of phylogenetic increase in the nuclear DNA. Despite many studies on endopolyploidy, it is still not clear why the cell enters the endoreduplication cycle inst ead of continuing proliferat ion. However, the relationship between endopolyploidy, aging and differen tiation of cells has been well known for several decades (Amato, 1952). It has been hypothesized that polysomaty has a functional significance, including the need to coordinate gene expression required for the interaction of nuclear and organelle genomes (Nagl, 1976; Galbraith, 1991). It has been suggested that the developmental program determines the number of endoreduplication cycles, independently of initial DNA content. S liwinska and Lukaszewska (2005) reported differences in the patterns of endopolyploidy between diploid, triplo id and tetraploid seedlings in sugarbeet. It is suggested that polysomaty is most prev alent in plants with small genomes (Galbraith, 1991; Rocher et al ., 1990). Our results also showed diploids expressing a higher polysomaty, with no endor eduplication cycles obs erved in triploid and tetraploid plants, confirming a nega tive correlation betw een genome size and endoreduplication levels. Polyploidy evolution, as a molecular pr ocess, is the primary mechanism to generate genomic redundancy (Wendel, 2000) . With endoreduplication resulting in polysomaty, genome size duplicates, with more a lleles of a particular gene present, and

PAGE 116

103 its significance is still unclear. Usually in young (100-8,000 yrs) polyp loids, both copies of a duplicated gene retain expression (We ndel, 2000), and some authors have observed a direct relationship between the polyploidy a nd proliferation of tr ansposable elements (Wendel, 2000). Genome duplication could allo w for gene function of duplicated genes to diversify and bring about evolutionary innovation in gene ral (Meyer and Van de Peer, 2003). Although the relative impor tance of genome duplic ation in evolution is not clear, it is possible that the variable ploidy of hydr illa could contribute to its adaptation and development of fluridone resi stance (Meyer and Van de Peer , 2003). These large genetic variations within a population, along w ith endopolyploidy, may offer hydrilla an opportunity to have clones that may react di fferently to varying environmental conditions and may have contributed to the rapid development of herbicide resistance.

PAGE 117

104 Figure 5-1. Histograms of flow cy tometric nuclear analysis of standard (triploid) (5.1a), diploid (5.1b), triploid ( 5.1c) and tetraploid (5.1d) shoot apical meristematic tissue of hydrilla.

PAGE 118

105 Table 5-1. Ploidy levels in hydri lla plants in fluridone sus ceptible (S) and resistant (R1R5) hydrilla populations. Hydrilla Population S R1 R2 R3 R4 R5 Plant Number G1 P2 G P G P G P G P G P 1 309 3x 308 3x 289 3x 276 3x 318 3x 323 3x 2 312 3x 290 3x 280 3x 312 3x 298 3x 321 3x 3 333 3x 320 3x 291 3x 302 3x 302 3x 321 3x 4 312 3x 319 3x 295 3x 305 3x 317 3x 301 3x 5 298 3x 309 3x 282 3x 300 3x 312 3x 312 3x 6 288 3x 284 3x 277 3x 310 3x 295 3x 326 3x 7 343 3x 322 3x 303 3x 303 3x 293 3x 342 3x 8 324 3x 312 3x 289 3x 312 3x 291 3x 343 3x 9 355 3x 315 3x 305 3x 308 3x 270 3x 289 3x 10 310 3x 313 3x 289 3x 317 3x 346 3x 299 3x 11 304 3x 318 3x 282 3x 315 3x 342 3x 307 3x 12 302 3x 280 3x 317 3x 312 3x 298 3x 326 3x 13 309 3x 296 3x 253 3x 323 3x 287 3x 327 3x 14 343 3x 312 3x 264 3x 330 3x 296 3x 329 3x 15 313 3x 309 3x 291 3x 312 3x 339 3x 202 2x 16 345 3x 311 3x 273 3x 287 3x 287 3x 212 2x 17 312 3x 197,378 2x,4x 287 3x 334 3x 299 3x 201 2x 18 301 3x 208,352 2x,4x 292 3x 343 3x 362 3x 198 2x 19 445 4x 241 2x 282 3x 192 2x 311 3x 187 2x 20 434 4x 243 2x 311 3x 248,350 2x,4x 202 2x 212 2x 21 4x 214 2x 231 2x 209,357 2x,4x 199 2x 231 2x 22 413 4x 388 4x 200 2x 238,363 2x,4x 180 2x 224 2x 23 405 4x 390 4x 222 2x 190,368 2x,4x 238 2x 236 2x 24 409 4x 407 4x 201 2x 208,402 2x,4x 180 2x 411 4x 25 411 4x 406 4x 211 2x 234,389 2x,4x 399 4x 389 4x 1 G1 Peak Mean Value in flow cytometry analysis; 2 Presumed ploidy level

PAGE 119

106 Table 5-2. Mean nuclear DNA cont ents (pg) in plants with different ploidy levels among fluridone susceptible (S) and resi stant (R1-R5) hydr illa populations determined by flow cytometry. Hydrilla Population 2x 3x 4x S -3.70 0.22 (72%) 4.90 0.18 (28%) R1 2.73 0.21a (20%) 3.58 0.16 (64%) 4.64 0.12 (16%) R2 2.47 0.16 (20%) 3.44 0.17 (80%) -R3 2.54 0.35 (38%) 3.60 0.16 (62%) -R4 2.43 0.25 (20%) 3.58 0.26 (76%) 4.65 (4%) R5 2.45 0.18 (36%) 3.71 0.18 (56%) 4.66 0.18 (8%) a mean values followed by standard deviation; values in parenthesis indicate percent plants of a given ploidy level in a hydrilla population; n= 25.

PAGE 120

107 Figure 5-2. Histograms of endoreduplication pa tterns of diploid hydr illa plants from hydrilla populations R1 (5.2a) and R2 (5.2b and 5.2c).

PAGE 121

108 CHAPTER 6 SUMMARY AND CONCLUSIONS Despite extensive control efforts during the last 4-5 decades, the exotic invasive weed species Hydrilla verticillata remains a dominant weed pr oblem in the southeastern United States. Hydrilla has spread thr oughout the country’s waterways, clogging irrigation and drainage cana ls, degrading water quality, reducing productivity of recreational fisheries, and impeding naviga tion. Despite an expenditure of $50 million during the 1980’s, the percentage of Florida waters infested by hydrilla increased from 37 to 41% (USEPA, 2001).It is listed on the Fede ral List of Noxious Weeds and considered a Category I species on the Florida Exotic Pest Plant Council List. Management of hydrilla for the past 30 years has centered on th e use of the herbicid e fluridone. Fluridone disrupts the carotenoid biosynthetic pathway by non-competitive inhibition of the enzyme phytoene desaturase (PDS). As only the dioeci ous form of hydrilla is found in Florida, with spread and reproduction limited to ase xual means (subterranean turions, axillary turions, fragments, and root crowns), th e development of herbicide resistance was considered unlikely. Nevertheless, fluri done resistant hydrilla biotypes have been confirmed in several public and priv ate water bodies th roughout Florida. This study documented differential resi stance to fluridone herbicide among several hydrilla populations. The re sistance factors in resistant biotypes varied from twofold in R1, to as high as seven-fold in R5, indicating a wide range of resistance. Susceptible hydrilla populat ions showed a significant effect on both phytoene and carotene content at exposure to 5-10 g L-1 fluridone, whereas much higher doses were

PAGE 122

109 required in the fluridone resistant hydrilla populations . Regression analysis was performed in each population to quantif y the relationship between fluridone concentration and phytoene or carotene content at each time interval. EC50 values for phytoene (effective fluridone concentration required to increase phytoene content by 50% over control), and carotene (effective fluridone c oncentration required to decrease carotene content by 50% over c ontrol) were calculated. EC50 phytoene and EC50 carotene values of 7.5 g L-1 and 8.9 g L-1 fluridone were reported for the susceptible population. For resistant hydrilla populations, EC50 phytoene values varied from 16.8 g L-1 for R1 to as high as 61.3 g L-1 for R5. EC50 carotene values followed a trend similar to phytoene, and values of 18 and 63 g L-1 were recorded for the R1 and R5 populations, respectively. We performed molecular analysis by cl oning the gene for phytoene desaturase ( pds ) to correlate variation in the herbicide resistance in di fferent hydrilla populations at the molecular level from fluridone suscepti ble and resistant hydril la populations. Two separate and independent single-point muta tions of the codon 304 encoding for Arg in pds were found. The codon usage for Arg304 in the fluridone suscep tible hydrilla is CGT. Resistant hydrilla populations R1, R2, and R4 showed the Arg to Ser mutation at the 304 codon, and the resistance factors observed were two-fold (in R1 and R2), and five-fold in R4. R5 showed Arg to His mutation at th e 304 codon, and it resulted in seven-fold resistance to fluridone. Several other mutations were also found in resistant pds alleles, though their possible role in herb icide resistance is unclear. We also performed molecular analysis at one year after initial planting to detect any possible changes in the pds gene, when resistant populations were grown in the

PAGE 123

110 absence of fluridone. No differences were observed at codon 304 in the PDS protein of any hydrilla population. In diffe rent resistant hydrilla populations from six different Florida lakes, we found a tota l of a total of eight transv ersions and eleven transition mutations on the pds gene. This frequency was much higher than would have been expected in an asexually propagated species . Hydrilla is a fast growing plant with multiple means of propagation; able to genera te new individuals fr om a single node; and produces profuse axillary branching. A high ra te of somatic mutations can perpetuate since this plant species u ndergoes only asexual propa gation in Florida and the Southeastern U.S. Phenotypic measurements were performed to monitor differences in growth and reproductive physiology of different fluridon e resistant populations. Collectively, the resistant populations were supe rior to or at par with sus ceptible hydrilla in growth and reproductive parameters. This indicated no de leterious effect due to development of fluridone herbicide resistance. The hydrilla population R3 showed significant differences in growth compared to all other hydrilla populations. This population had slow early growth rate as evident from lower shoot elongation rates, and lower shoot biomass production from Oct 2004 to Mar 2005. Howe ver, from Apr 2005 through Aug 2005, it recorded a very high relative growth rate a nd was growing statistica lly at par with all other hydrilla populations. The population R3 wa s also inferior in the production of subterranean turions and axil lary branches. In addition, the flowering period extended two months longer in R3 compared to th e other hydrilla popu lations, including the susceptible population.

PAGE 124

111 The other possible reason of differential resistance level in different hydrilla populations in spite of the same mutation might be due to th e ploidy variation in hydrilla. There have been no conclusive studies on the variation in ploidy le vel in hydrilla in Florida and the Southeastern U.S. Although it ha s been mentioned in European literature, no data have validated the results. Ther efore, we determined the possible ploidy variations (or the presence of endopolyploidy) among different hydrilla populations with varying level of resistance to the he rbicide fluridone using flow cytometry. Differential ploidy levels (diploid 2n= 2x= 16; triploid 2n= 3x= 24; and tetraploid 2n= 4x= 32) were reported among different hydri lla biotypes, plants with in each biotype and within shoot tissues of the same plant. Triploid plants were the most predominant in all the populations. Diploid plants were observe d in all the hydrilla populations except S. Plants with tetraploidy were rare within popul ations. The diploid, trip loid and tetraploid plants had nuclear DNA contents of 2.43.73 pg, 3.44.71 pg, and 4.64.90 pg, respectively, and no differences were obs erved among plants with same ploidy for nuclear DNA content in different populations. We also observed endoreduplication pattern s in diploid plants of populations S and R3. However, no plant with higher pl oidy levels (3x or 4x) in any hydrilla populations showed endoreduplication. With endoreduplication resulting in polysomaty, genome size duplicates, with more alleles of a particular gene present, and its significance is still unclear in herbicide re sistance. Genome duplication could allow for gene function of duplicated gene s to diversify and bring abou t evolutionary innovation in general, and it is possible that the variable ploidy of hydrilla could contribute to its adaptation. These large geneti c variations within a population along with endopolyploidy

PAGE 125

112 may imply differences in survival strategy, a nd offer the species an opportunity to have various clones that may react differently to varying environmental conditions. One such environment condition is the presence of an he rbicide, particularly large term exposure as in the case of fluridone. The development of fluridone resistance in hydrilla can be attributed to several factors. Hydrilla is a fast growing plan t with multiple means of propagation, able to generate new individuals from a single node, and each plant can be several meters long. The ability of hydrilla to grow at low light intensities and to alternate between C3 and C4 metabolism, assures nutrient suppo rt within the plant at distan t areas of tissue that could be growing in adverse conditions. Also, leaves in hydrilla are only two cells thick which may explain its susceptibility to fluridone at concentrations that ar e sub-lethal for other submersed vegetation. However, the same l eaf anatomy, may contribute to higher rates of mutation due to UV radiation on hydrilla biomass growing on the surface of water bodies. A high rate of somatic mutations has been described in di oecious hydrilla, and these mutations can perpetuate since this biotype undergoes only asexual propagation. Hydrilla has variable ploidy which can result in gene duplication and duplication of gene function, favoring the presence of various alle les to adapt to variable environmental conditions. In addition, the use of a single he rbicide at low doses being the best cost effective method to control this weed, may in itself have contributed to the selection of resistant biotypes, given the adaptive potential of hydrilla. There are several economic implications of these findings that could dramatically impact hydrilla management in Florida. W eed management in large water bodies relies and depends heavily on fluridone, the onl y USEPA approved herbicide available for

PAGE 126

113 hydrilla control in large wate r-bodies. At >30 ppb fluridone use, the cost per hectare triples from current prices to $250-300/hectare. According to recent Florida Department of Environmental Protection budget requests, hydrilla infests ove r 11,500 hectares with $17,906,098 needed for control (Florida DEP, 2004) . However, application rates of 30 ppb or higher will also result in a loss of se lectivity (i.e., damage or loss of some emergent and many or all submersed species). Furthermore, if additional resistance occurs, the cost and environmental impact of fluridone may preclude its further use in Florida. The lack of sustainable alternatives to fluridone has led to a concerted effort between industry, state and federal agencies to devise contingent management plans including looking into new chemistries as well as biological agents for hydrilla control. Aggressive spread of hydrilla in aquatic ecosy stems, and the evolution of resistance to fluridone in Florida may forecast signifi cant and long-lasting eco logical and economic problems throughout the Southern states of the U.S.

PAGE 127

114 APPENDIX A PHYTOENE, CAROTENE AND PHYTOENE: CAROTENE RATIO IN S AFFECTED BY FLURIDONE DOSES AT VARIOUS GROWTH STAGES Sept 20, 2004 (At planting) Dec 20, 2004 (3 MAP) March 20, 2005 (6 MAP) June 20, 2005 (9 MAP) Sept 20, 2005 (12 MAP) Fluridone dose (g L-1) P B P/B P B P/B P B P/B P B P/B P B P/B 0 71 1.41 21.2 1.27 3.35 0.38 70 0.9 22.8 2.14 3.17 0.59 76 3.1 24.4 1.24 3.33 1.05 78 4.4 25.9 1.9 3.09 0.6 78 3.6 26.6 2.37 2.98 0.37 5 82 4.2* 17.8 1.28* 4.65 0.72* 81 1.6* 19.1 1.21* 4.35 0.85* 87 3.1* 19.9 1.33* 4.43 0.6* 91 3.5* 21.8 0.95* 4.17 0.25* 91 4.0* 22.3 1.1* 4.11 0.29* 10 109 5.4 11.3 0.94 9.75 0.83 108 3.2 12.1 3.16 9.31 1.99 111 6.6 12.4 1.76 9.07 1.07 114 3.4 13.6 4.1 8.99 2.44 127 5.6 14.5 4.05 9.36 2.63 15 112 6.9 7.9 2.65 15.2 4.13 109 3.6 9.6 2.27 11.9 3.04 128 3.3 11.9 3.67 11.5 3.7 131 3.9 11.9 3.15 11.5 2.47 129 2.1 9.4 1.0 13.7 1.33 20 120 8.7 7.6 0.93 15.9 2.18 113 8.8 8.6 1.23 13.2 2.21 132 4.3 7.1 1.53 19.2 3.96 127 9.7 8.3 1.49 15.4 2.07 133 2.7 9.0 2.7 16.0 4.49 30 122 6.3 4.7 0.61 26.5 4.13 124 2.3 4.9 1.59 27.0 7.23 135 4.9 6.2 1.97 23.4 4.79 136 9.2 5.7 2.01 26.3 8.96 144 3.6 5.2 1.7 29.4 7.02 50 123 5.1 4.4 0.38 27.6 2.94 125 4.3 3.9 1.40 36.2 9.39 139 3.2 5.0 1.28 29.3 8.83 142 9.89 4.9 2.11 34.7 10.72 145 4.5 4.7 1.57 32.8 7.82 P : Phytoene content (g per g fresh we ight), value rounded to whole number; B : carotene content (g per g fresh weight); P/B : Phytoene/ carotene ratio; 1Mean values followed by 95% confidence in terval values (Standard error x 1.96); * denotes the fluridone concentration at whic h there is significan t affect on phytoene, carotene, and Phytoene: carotene ratio over control at each time period.

PAGE 128

115 APPENDIX B PHYTOENE, CAROTENE AND PHYTOENE: CAROTENE RATIO IN R1 AFFECTED BY FLURIDONE DOSES AT VARIOUS GROWTH STAGES Sept 20, 2004 (At planting) Dec 20, 2004 (3 MAP) March 20, 2005 (6 MAP) June 20, 2005 (9 MAP) Sept 20, 2005 (12 MAP) Fluridone dose (g L-1) P B P/B P B P/B P B P/B P B P/B P B P/B 0 64 5.4 21.2 2.28 3.07 0.47 67 1.3 22.9 2.05 2.94 0.23 70 4.4 26.5 2.19 2.64 0.30 70 2.5 28.1 1.39 2.50 0.14 73 2.3 28.8 2.01 2.53 0.24 5 67 7.3 20.3 2.38 3.35 0.61 68 1.7 22.4 1.32 3.05 0.25 71 3.2 25.6 2.27 2.77 0.34 73 3.1 27.2 1.71 2.68 0.23 74 5.8 27.8 1.75 2.66 0.24 10 75 5.4 19.0 1.07 3.97 0.47 75 4.1* 18.3 1.71* 4.13 0.59* 76 4.0 22.2 1.05* 3.40 0.16* 81 6.4* 23.9 3.69* 3.42 0.44* 82 5.5* 24.3 2.69* 3.40 0.38* 15 90 6.0* 16.7 2.52* 5.48 0.67* 92 4.4 17.0 3.37 5.62 1.26 93 7.3* 18.9 2.18 5.01 0.80 90 3.0 19.6 0.99 4.62 0.34 109 3.8 19.9 2.53 5.58 1.00 20 104 7.6 12.8 2.91 8.44 1.93 105 4.9 14.2 2.93 7.69 1.85 115 4.8 16.3 1.31 7.11 0.55 101 4.9 17.7 1.41 5.74 0.47 113 16.1 15.0 2.14 7.56 0.89 30 106 5.0 9.5 1.0 11.3 1.52 115 7.4 9.2 1.97 12.9 2.02 124 3.7 11.4 1.96 11.2 2.24 103 4.1 11.2 1.94 9.43 1.47 121 7.7 11.2 2.36 11.2 2.35 50 122 6.4 8.0 2.23 16.7 4.50 116 5.8 8.7 2.41 14.8 6.92 130 4.1 10.6 2.36 12.7 2.63 132 6.5 10.4 1.16 12.8 1.18 131 5.0 10.2 1.68 13.1 1.96 P : Phytoene content (g per g fresh we ight), value rounded to whole number; B : carotene content (g per g fresh weight); P/B : Phytoene/ carotene ratio; 1Mean values followed by 95% confidence in terval values (Standard error x 1.96); * denotes the fluridone concentration at whic h there is significan t affect on phytoene, carotene, and Phytoene: carotene ratio over control at each time period.

PAGE 129

116 APPENDIX C PHYTOENE, CAROTENE AND PHYTOENE: CAROTENE RATIO IN R2 AFFECTED BY FLURIDONE DOSES AT VARIOUS GROWTH STAGES Sept 20, 2004 (At planting) Dec 20, 2004 (3 MAP) March 20, 2005 (6 MAP) June 20, 2005 (9 MAP) Sept 20, 2005 (12 MAP) Fluridone dose (g L-1) P B P/B P B P/B P B P/B P B P/B P B P/B 0 56 4.5 17.4 0.88 3.26 0.40 60 0.9 23.5 2.10 2.56 0.25 67 1.8 27.9 1.48 2.41 0.19 70 1.8 29.0 2.77 2.43 0.26 72 2.6 29.9 2.00 2.42 0.22 5 62 7.8 17.4 1.88 3.59 0.40 61 5.1 23.3 1.95 2.64 0.32 71 5.6 27.5 2.18 2.59 0.33 72 3.3 28.6 1.48 2.53 0.17 77 3.2 28.8 0.77 2.68 0.10 10 65 3.7* 15.0 2.75 4.48 0.95 67 5.1* 20.4 1.76 3.29 0.35* 78 7.1* 26.1 2.47 2.99 0.24* 77 6.8 27.1 1.56 2.83 0.30 82 5.1* 27.2 4.14 3.07 0.48 15 67 2.7 13.3 0.96* 5.02 0.33* 72 4.7 18.4 2.19* 3.96 0.57 80 4.9 22.8 2.01* 3.54 0.29 85 3.8* 23.5 2.86* 3.68 0.58* 88 5.0 23.7 2.90* 3.77 0.47* 20 82 8.5 11.2 0.99 7.30 1.00 87 4.2 15.5 1.58 5.61 0.48 98 2.3 16.8 1.38 5.84 0.58 106 5.9 14.9 2.47 7.26 1.27 110 5.5 14.3 2.86 7.85 1.15 30 101 6.8 8.5 1.43 12.2 2.28 104 4.4 9.4 1.16 11.2 1.11 114 4.8 8.0 1.96 14.9 4.21 117 2.9 9.6 1.62 12.4 1.97 121 6.2 9.3 2.10 13.6 3.32 50 104 3.9 5.9 1.26 18.4 4.83 105 4.0 8.0 1.06 13.4 2.26 119 5.6 7.8 1.84 16.0 4.00 12 5.9 8.9 1.26 13.7 2.43 123 4.5 9.2 1.51 13.6 2.57 P : Phytoene content (g per g fresh we ight), value rounded to whole number; B : carotene content (g per g fresh weight); P/B : Phytoene/ carotene ratio; 1Mean values followed by 95% confidence in terval values (Standard error x 1.96); * denotes the fluridone concentration at whic h there is significan t affect on phytoene, carotene, and Phytoene: carotene ratio over control at each time period.

PAGE 130

117 APPENDIX D PHYTOENE, CAROTENE AND PHYTOENE: CAROTENE RATIO IN R3 AFFECTED BY FLURIDONE DOSES AT VARIOUS GROWTH STAGES Sept 20, 2004 (At planting) Dec 20, 2004 (3 MAP) March 20, 2005 (6 MAP) June 20, 2005 (9 MAP) Sept 20, 2005 (12 MAP) Fluridone dose (g L-1) P B P/B P B P/B P B P/B P B P/B P B P/B 0 66 6.1 21.7 2.97 3.1 0.33 69 3.9 27 4.22 2.61 0.51 73 3.4 32.5 5.26 2.30 0.33 75 3.0 33.1 3.27 2.3 0.27 78 2.4 34.7 3.47 2.25 0.2 5 66 11.5 21.4 2.61 3.13 0.45 69 3.3 26.9 4.8 2.65 0.61 74 2.9 31.7 3.06 2.36 0.31 77 4.1 32.1 1.69 2.4 0.2 78 1.8 33.4 1.68 2.35 0.11 10 67 4.1 20.6 5.03 3.41 0.79 70 5.7 25.8 3.99 2.78 0.59 76 3.6 31.4 2.45 2.42 0.28 78 3.0 31.6 3.34 2.48 0.32 79 5.2 32.8 1.46 2.42 0.09 15 69 14.3 20.4 1.09 3.39 0.74 70 4.4 25.6 2.78 2.77 0.32 76 4.3 29.7 2.06 2.58 0.27 79 4.0 30.1 1.3 2.62 0.25 81 5.8 31.2 1.53 2.61 0.2 20 81 4.7* 17.8 2.72 4.63 0.81* 85 4.5* 22.3 1.4* 3.81 0.28* 93 3.9* 26 3.38* 3.64 0.49* 98 5.1* 26.1 2.02* 3.77 0.39* 101 3.4* 26.2 2.65* 3.88 0.42* 30 102 6.0 11.1 3.34* 9.83 2.85 110 8.9 11.5 1.95 9.69 1.12 114 7.1 16.5 2.61 7.02 1.21 119 1.8 13.0 1.78 9.29 1.39 125 8.7 13.6 1.63 9.23 0.9 50 109 8.1 7.2 1.10 15.2 4.93 117 9.8 7.7 0.76 15.3 2.08 122 3.5 9.2 2.43 14 4.11 126 6.3 8.4 0.79 15.2 1.61 125 11.0 9.2 2.54 14.8 5.87 P : Phytoene content (g per g fresh we ight), value rounded to whole number; B : carotene content (g per g fresh weight); P/B : Phytoene/ carotene ratio; 1Mean values followed by 95% confidence in terval values (Standard error x 1.96); * denotes the fluridone concentration at whic h there is significan t affect on phytoene, carotene, and Phytoene: carotene ratio over control at each time period.

PAGE 131

118 APPENDIX E PHYTOENE, CAROTENE AND PHYTOENE: CAROTENE RATIO IN R4 AFFECTED BY FLURIDONE DOSES AT VARIOUS GROWTH STAGES Sept 20, 2004 (At planting) Dec 20, 2004 (3 MAP) March 20, 2005 (6 MAP) June 20, 2005 (9 MAP) Sept 20, 2005 (12 MAP) Fluridone dose (g L-1) P B P/B P B P/B P B P/B P B P/B P B P/B 0 71 5.1 26.9 1.9 2.67 0.32 74 2.9 26.1 4.21 2.90 0.64 77 5.2 29.8 1.61 2.59 0.29 84 3.0 30.9 2.37 2.74 0.21 86 2.4 32.0 5.04 2.74 0.42 5 73 4.0 25.6 2.7 2.88 0.4 75 3.0 25.9 3.70 2.95 0.42 79 2.6 29.1 1.11 2.70 0.13 86 4.0 29.7 3.36 2.93 0.37 89 4.3 30.8 2.20 2.88 0.08 10 73 4.2 24.3 2.04 3.02 0.31 77 3.6 25.5 3.89 3.09 0.60 81 2.5 28.4 2.54 2.85 0.29 87 1.9 29.7 5.46 3.03 0.75 89 5.5 30.7 1.53 2.91 0.28 15 79 6.6 23.0 1.2 3.44 0.33 80 3.5 23.9 2.67 3.39 0.41 81 2.0 26.3 1.61 3.07 0.22 95 2.8 27.1 1.73 3.52 0.16 98 2.4* 28.0 2.63 3.52 0.31* 20 90 9.0* 21.0 3.26* 4.35 0.62* 100 3.2* 22.9 2.78 4.41 0.46* 84 2.9 26.2 2.60 3.23 0.41 100 7.6* 25.3 3.27 4.01 0.56* 111 5.9 25.3 2.90 4.44 0.60 30 107 8.2 19.0 1.55 5.64 0.25 93 9.4 21.0 3.26* 4.48 0.59 101 7.3* 22.0 3.26* 4.68 0.90* 128 4.9 23.0 1.20* 5.60 0.44 132 2.9 22.0 3.97* 6.13 1.11 50 139 11.0 10.0 2.19 14.5 3.87 131 8.3 10.6 1.63 12.6 2.13 132 4.8 10.9 1.36 12.2 1.27 158 7.2 11.3 3.12 14.6 3.46 153 6.3 11.3 2.13 14.0 3.12 P : Phytoene content (g per g fresh we ight), value rounded to whole number; B : carotene content (g per g fresh weight); P/B : Phytoene/ carotene ratio; 1Mean values followed by 95% confidence in terval values (Standard error x 1.96); * denotes the fluridone concentration at whic h there is significan t affect on phytoene, carotene, and Phytoene: carotene ratio over control at each time period.

PAGE 132

119 APPENDIX F PHYTOENE, CAROTENE AND PHYTOENE: CAROTENE RATIO IN R5 AFFECTED BY FLURIDONE DOSES AT VARIOUS GROWTH STAGES Sept 20, 2004 (At planting) Dec 20, 2004 (3 MAP) March 20, 2005 (6 MAP) June 20, 2005 (9 MAP) Sept 20, 2005 (12 MAP) Fluridone dose (g L-1) P B P/B P B P/B P B P/B P B P/B P B P/B 0 76 5.3 32.3 2.15 2.37 0.17 76 3.3 33.2 3.18 2.31 0.22 78 2.4 34.7 2.64 2.26 0.21 78 2.4 34.7 2.64 2.26 0.21 79 1.7 36.9 3.61 2.16 0.23 5 77 3.2 32.0 2.22 2.41 0.18 77 3.0 31.0 1.21 2.48 0.13 78 2.2 34.5 2.97 2.28 0.15 78 2.2 34.5 2.97 2.28 0.15 80 2.1 36.5 2.37 2.21 0.12 10 78 1.8 30.2 1.01 2.58 0.14 78 2.0 30.3 1.06 2.57 0.11 79 1.3 32.0 2.04 2.49 0.20 79 1.3 32.0 2.04 2.49 0.20 81 1.3 35.7 2.96 2.29 0.19 15 80 1.3 29.0 0.93 2.75 0.10 80 3.4 28.4 2.46 2.83 0.34 83 1.9 32.4 2.36 2.57 0.21 82 1.9 32.4 2.36 2.57 0.21 83 2.7 33.2 1.65 2.51 0.16 20 82 3.4 28.0 2.34 2.94 0.26 82 1.9 27.2 1.84 3.02 0.26 85 2.5 30.8 2.14 2.77 0.24 85 2.5* 30.8 2.14 2.77 0.24* 88 1.4* 32.5 2.23 2.70 0.16* 30 88 2.1* 24.3 2.46* 3.65 0.43* 88 3.2* 23.8 1.83* 3.70 0.32* 88 2.4* 27.8 1.51* 3.19 0.15* 88 2.4 27.8 1.51* 3.19 0.15 90 2.0 28.3 2.63* 3.21 0.36 50 101 5.8 19.3 2.45 5.29 0.64 102 2.0 19.1 1.24 5.36 0.30 97 1.8 21.0 1.66 4.66 0.40 97 1.8 21.0 1.66 4.66 0.40 98 1.0 21.0 1.25 4.70 0.30 P : Phytoene content (g per g fresh we ight), value rounded to whole number; B : carotene content (g per g fresh weight); P/B : Phytoene/ carotene ratio; 1Mean values followed by 95% confidence in terval values (Standard error x 1.96); * denotes the fluridone concentration at whic h there is significan t affect on phytoene, carotene, and Phytoene: carotene ratio over control at each time period.

PAGE 133

120 LIST OF REFERENCES Ahrens WH, and Stoller EW, Co mpetition, growth rate and CO2 fixation in triazinesusceptible and resistant smooth pigweed ( Amaranthus hybridus ). Weed Sci 31 : 438-444 (1983). Albrecht M, Klein A, Hugueney P, Sandmann G, and Kuntz M, Molecular-cloning and functional expression in Escherichia coli of a novel plant en zyme mediating zetacarotene desaturation. FEBS Letters 372 : 199-202 (1995). Aman R, Schieber A, and Carle R, Eff ects of heating and illumination on trans-cis isomerization and degradation of beta-car otene and lutein in isolated spinach chloroplasts. J Agri Food Chem 54 : 9512-9518 (2005) Amato FD, Polyploidy in the differentiation a nd function of tissues and cells in plants. Caryologia 4 : 311-358 (1952). Ames MD, Anderson LWJ, and Spencer DF , Growth of monoecious and dioecious hydrilla: effects of temperature and light. Res Prog Rep WSWS, pp 298-299 (1986). Anderson LWJ, and Spencer DF, Photointrupt ion of long nights i nhibits production of turions and tuber in monoecious hydrilla. Res Progr Rep WSWS , pp 301-303 (1986). Appenroth KJ, and Bergfeld R, Photophysiology of turion germination in Spirodelapolyrhiza (L.) Schleiden .11. Structur al-changes during red-light induced responses. J Plant Physiol 141 : 583-588 (1993). Arias RS, Dayan FE, Scheffer BE, Duke SO, Michel A, and Netherland MD, Bioengineering resistance to phytoe ne desaturase inhibitors in Arabidopsis thaliana. Proc WSSA , Kansas City, Missouri. pp 62 (2004). Babczinski P, Sandmann G, Schmidt RR, Shiokawa K, and Yasui K, Substituted Tetrahydropyrimidinones: A new herbicidal class of compounds inducing chlorosis by inhibition of phytoe ne desaturation. Pest Biochem Physiol 52 : 33-44 (1995). Balciunas JK, Grodowitz MJ, Cofrances co AF, and Shearer JF, Hydrilla. In : Biological Control of Invasive Plants in the Eastern United Stat es (eds.R Van Driesche, S Lyon, B

PAGE 134

121 Blossey, M Hoddle, and R Reardon), USDA Forest Service Publication FHTET-2002-04, pp 413 Web page http://www.invasive.org/easter n/biocontrol/7Hydrilla.html (Last accessed, 11 December 2002). Banks PA, and Merkle MG, Effectiven ess and persistence of fluridone. Proc SWSS 31 : 74 (1978). Barko JW, and Smart RM, Comparative infl uences of light and temperature on the growth and metabolism of selected submersed fresh water macrophytes. Ecol Monogr 51 : 219-235 (1981). Barret SCH, Eckert CG, and Husband BC, E volutionary processes in aquatic plant populations. Aquatic Botany 44 : 105-145 (1993). Bartels PG, and Watson CW, Inhibition of carotenoid synthesis by fluridone and norflurazon. Weed Sci 26 : 198-203 (1978). Bartley GE, Viitanen PV, Pecker I, Chamov itz D, Hirschberg J, and Scolnik PA, Molecular-cloning and expre ssion in photosynthetic bact eria of a soybean cDNA coding for phytoene desaturase, an enzyme of the carotenoid biosynthesis pathway. PNAS 88 : 6532-6536 (1991). Bates AL, and Smith CS, Submersed plant inva sions and declines in the south-eastern United States. Lake Reservoir Manage 10 : 53 (1994). Bennet MD, and Smith JB, Nuclear DNA amounts in angiosperms. Phil Trans R Soc London B274 : 227-274 (1976). Bennett MD, Bhandol P, and Leitch IJ, Nuclear DNA amounts in angiosperms and their modern uses807 new estimates. Ann Botany 86 : 859-909 (2000). Bennetzen J, Ma J, and Devos KM, Mechan isms of recent genome size variation in flowering plants. Ann Botany 95 : 127-132 (2005). Bergelson J, and Purrlington CB, Surveying patterns in the cost s of resistance in plants. Amer Nat 148 : 536-558 (1996). Bergelson J, Purrlington CB, Palm CJ, and Lopez-Gutierezz JC, Costs of resistance: a test using transgenic Arabiopsis thaliana . Proc Royal Soc London, Biological Sciences 263 : 1659-1663 (1996). Biradar DP, Rayburn AL, and Bullock DG, E ndopolyploidy in diploi d and tetraploid maize (Zea mays L.). Ann Botany 71 : 417-421 (1993). Blackburn RD, Weldon LW, Yeo RR, and Taylor TM, Identification and distribution of similar appearing aquatic weeds in Florida. J Aquat Plant Manage 8 : 11-21 (1969).

PAGE 135

122 Böger P, and Sandmann G, Carotenoid biosynthes is inhibitor herbicides mode of action and resistance mechanisms. Pesticide Outlook 9 : 29–35 (1998). Bonk M, Hoffmann B, Von Lintig J, Schledz M, Al-Babili S, Hobeika E, Kleinig H, and Beyer P, Chloroplast import of four carotenoid biosynthetic enzymes in vitro reveals differential fates prior to memb rane and oligomeric assembly. Eur J Biochem 247 : 942-950 (1997). Bottrell DR, Integrated Pest Management . US Govt. Printing Office, pp 120 (1979). Bowes G, and Salvucci ME , Hydrilla: inducible C4-type photosynthesis without Kranz anatomy. In : Advances in Photosynthesis Re search III (eds. C Sybesma), pp 829832, Martinus Nijhoff/Dr W Junk, The Hague, The Netherlands (1984). Bowes GA, Holaday S, and Haller WT, Seasonal variation in the biomass, tuber density, and photosynthetic metabolism of hy drilla in three Florida lakes. J Aquat Plant Manage 17 : 61-65 (1979). Brooker MP, and Edwards RW, Aquatic herbic ides and the control of water weeds. Water Res 9 : 1-15 (1975). Bulcke R, De Praeter H, Van Himme M, and Strycker H, Resistant of annual meadowgrass, Poa annua L. to 2-chloro-1,3,5-triazines. Meded Rijksfac Landbouwwet Genetics 47 : 211-220 (1985). Chamovitz D, Pecker I, and Hirshberg J, The molecular basis of resistance to the herbicide norflurazon. Plant Mol Biol 16 : 967-974 (1991). Chamovitz D, Sandmann G, and Hirsc hberg J, Molecular and biochemical characterization of herbicide-resistant mutants of cyanobacteria reveals that phytoene desaturation is a rate-limiting step in ca rotenoid biosynthesis. J Biol Chem 268 : 17348-17353 (1993). Charlesworth B, Morgan MT, and Charlesworth D, The effect of deleterious mutations on neutral molecular variation. Genetics 134 : 1289–1303 (1993). Chaudhuri JB, and Sharma A, Cytological studies on three aquatic members of Hydrocharitaceae in relation to their morphological and eco logical characteristics. Cytologia (Japan) 43 : 1-20 (1978). Colle DE, and Shireman JV, Coefficients of condition for largemouth bass, bluegill, and red ear sunfish in Hydrilla -infested lakes. Transactions of the American Fisheries Society , 109 : 521–531 (1980). Cook CDK, and Lüönd R, A revision of the genus Hydrilla (Hydrocharitaceae). Aquatic Botany 13 : 485-504 (1982).

PAGE 136

123 Couch R, and Nelson E, Myriophyllum spicatum in North America. Pro c APMS , Vicksburg, MS, pp. 8-18 (1985). Countryman WD, The history, spread and presen t distribution of some immigrant aquatic weeds in New England. Hyacinth J 8 : 50-52 (1970). Davenport LJ, Chromosome number reports. LXVII. Taxon 29 : 351 (1980). Dayan FE, and Duke SO, Herbicides: Caro tenoid biosynthesis i nhibitors, Vol. 2. In : Encyclopedia of Agrochemicals (eds. JR Plimmer, DW Gammon, NN Ragsdale, and T Roberts), pp 744-749, John Wile y and Sons, New York, NY (2003). Devlin RM, Saras CN, Kisiel MJ, and Ko stusiak AS, Influence of fluridone on chlorophyll content of wheat ( Triticum aestivum ) and corn ( Zea mays ). Weed Sci 26 : 432-433 (1978). Doong RL, MacDonald GE, and Shilling DG, Effect of fluridone on chlorophyll, carotenoids and anthocyanin content of hydrilla. J Aquat Plant Manage 31 : 55-59 (1993). Doyle RD, and Smart RM, Effects of drawdowns and dessication on tubers of hydrilla, an exotic aquatic weed. Weed Sci 49 : 135-140 (2001). Duke SO, Kenyon WH, and Paul RN, FMC 57020 effects on chloroplast development in pitted morningglory ( Ipomoea lacunosa ) cotyledons. Weed Sci 33 : 786-794 (1985). Emshwiller E, Ploidy levels among species in the “ Oxalis tuberosa alliance” as inferred by flow cytometry. Ann Botany 89 : 741-753 (2002). Evans GC , The quantitative analysis of plant growth, pp 300325, Blackwell Scientific Publications, University of Calif ornia Press, Berkeley, CA (1972). Florida Department of Environment Protec tion (Florida DEP), Appendix II-Key Funds Needed to Manage Invasive Aquatic Plants in Florida’s Public Waters, Web page http://www.dep.state.fl.us (Last accessed, 20 July 2004) Fox AM, Haller WT, and Shilling DG, Hydrilla co ntrol with split treatments of fluridone in Lake Harris, Florida. Hydrobiologia 340 : 235-239 (1996). Galbraith DW, Flow cytometric analysis of plant genomes. Methods Cell Biol 33 : 549561 (1990). Galbraith DW, Harkins KR, and Kn app S, Systemic endopolyploidy in Arabidopsis thaliana . Plant Physiol 96 : 985-989 (1991). Grace JB, Juvenile vs. adult competitive abilit ies in plants: size dependence in cattails ( Typha ). Ecology 66: 1630-1638 (1985).

PAGE 137

124 Haller WT, Hydrilla: A new and rapidly sp reading aquatic weed problem. I.F.A.S., University of Florida, Circul ar S-245, Gainesville, FL (1976). Haller WT, Fox AM, and Shilling DG, Hydrilla control programs in the upper St. Johns River, Florida, USA. Proc EWRS Symposium on Aquatic Plants 8 : 111-116 (1990). Haller WT, Miller JL, and Garrard LA, Seas onal production and germination of hydrilla vegetative propagules. J Aquat Plant Manage 14 : 26-29 (1976). Haller WT, and Sutton DL, Community struct ure and competition between hydrilla and vallisneria. Hyacinth Cont J 13 : 48-50 (1975). Haller WT, Sutton DI, and Barlowe WC, Effects of salinity on growth of several aquatic macrophytes. Ecology 55 : 891-894 (1974). Harlan SM, Davis GJ, and Pe sacreta GJ, Hydrilla in three North Carolina lakes. J Aquat Plant Manage 23 : 68-71 (1985). Hejazi MA, and Wijffels RH, Effect of light intensity on beta-carotene production and extraction by Dunaliella salina in two-phase bioreactors. Biomol Engineering 20 : 171-175 (2003). Hill RJ, Taxonomy and biological considerations of herbicide-resistant and herbicidetolerant biotypes. In : HerbicideResistance in Plants (eds. HM Lebaron, and J. Gressel), pp 81-98, Wiley and S ons, Inc, New York NY (1982). Hofstra DE, Adam KD, and Clayton JS, Isoz yme variation in New Zealand populations of Myriophyllum and Potamogeton species. Aquatic Botany 52 : 121-131 (1995). Hofstra DE, and Clayton JS, Evaluation of se lected herbicides for control of exotic submerged weeds in New Zealand: 1. The use of endothall, triclopyr and dichlobenil. J Aquat Plant Manage 39 : 20-24 (2001b). Hofstra DE, Clayton J, Green JD, and Adam KD, RAPD profiling a nd isozyme analysis of New Zealand Hydrilla verticillata . Aquatic Botany 66 : 153-166 (2000). Holaday AS, Salvucci ME, and Bowes G, Variable photosynthesis photo-respiration ratios in hydrilla and other subm ersed aquatic macrophyte species. Can JB 61 : 229236 (1983). Holsinger KE , Reproductive systems a nd evolution in vascular plants. PNAS 97: 7037– 7042 (2000). Hugueney P, Romer S, Kuntz M, and Camara B, Characterization and molecular-cloning of a flavoprotein catalyzing the synthe sis of phytofluene and zeta-carotene in Capsicum chromoplasts. Eur J Biochem 209 : 399-407 (1992).

PAGE 138

125 Kaundun SS, and Windass JD, Derived cleaved amplified polymorphic sequence, a simple method to detect a key point mutation conferring acetyl CoA carboxylase inhibitor herbicide resi stance in grass weeds. Weed Res 46 : 34-39 (2006). Klekowski EJ, Plant clonality, mutation, di plontic selection a nd mutational meltdown. Biol J Linnean Soc 79 : 61–67 (2003). Kovalchuk I, Kovalchuk O, and Hohn B, Geno me-wide variation of the somatic mutation frequency in transgenic plants. EMBO J 19 : 4432-4438 (2000). Kulshreshtha M, and Gopal B, Allelopathic influence of Hydrilla verticillata (L.F.) Royle on the distribution of Ceratophyllum species. Aquatic Botany 16 : 207-209 (1983). Laber B, Usunow G, Wiecko E, Franke W, Franke H, and Kohn A, Inhibition of Narcissus pseudonarcissus phytoene desaturase by herb icidal 3-trifluoromethyl-1,1 '-biphenyl derivatives. Pest Biochem Physiol 63 : 173-184 (1999). Langeland KA, Karyotypes of hydrilla (hydroc haritaceae) populations in the UnitedStates. J Aquat Plant Manage 27 : 111-115 (1989). Langeland KA, Hydrilla (Hydrilla verticillata (L.f.) Ro yle): a continuing problem in Florida waters . University of Florida Coop. Ex tension Service Circular No. 884. University of Florida, Gainesville, Florida, U.S.A. (1990). Langeland KA, Hydrilla verticillata (L.F.) Royle (Hydrocharitaceae), “The Perfect Aquatic Weed”. Castanea 61 : 293-304 (1996). Langeland KA, Hill ON, Koschnick TJ, and Ha ller WT, Evaluation of a new formulation of Reward Landscape and aquatic herbic ide for control of duckweed, water hyacinth, water lettuce and hydrilla. J Aquat Plant Manage 40 : 51-53 (2002). Langeland KA, and Schiller DL, Hydrilla in North Carolina. Aquatics 5 : 8-14 (1983). Langeland KA, and Smith CB, Hydrilla produce s viable seed in No rth Carolina lakesmechanism for long term disposal. Aquatics 6 : 20-21 (1984). Langeland KA, and Sutton DL, Regrowth of hydrilla from axillary buds. J Aquat Plant Manage 18 : 27-29 (1980). Langeland KA, Shilling D, Carter J, Laro che F, Steward KK, and Madiera P, Chromosome morphology and numbe r in various populations of Hydrilla verticillata (L.f.) Royle. Aquatic Botany 42 : 253-263 (1992). Lazor RL, The ecology, nomenclature and distribution of hydrilla ( Hydrilla verticillata Casp.) and Brazilian elodea ( Egeria densa Planch.). Proc SWSS 28 : 269-273 (1978).

PAGE 139

126 Les DH, Mehrhoff LJ, Cleland MA, and Gabel JD, Hydrilla verticillata (Hydrocharitaceae) in Connecticut. J Aquat Plant Manage 35 : 10-14 (1997). Linde AF, Janisch T, and Smith D, Cattail —The significance of its growth, phenology and carbohydrate storage to its control and management. Technical Bulletin #94, Wisconsin Department of Natural Resource, Madison, WI (1976). Linden H, Sandmann G, Chamovitz D, Hirs chberg J, and Böger P, Biochemical characterization of herbicide resistant Anacystis mutants selected against the bleaching herbicide norflurazon. Pest Biochem Physiol 36: 46-51 (1989). Lokker C, Susko D, Lovett-Doust L, and Love tt-Doust J, Population genetic structure of Vallisneria americana , a dioecious clonal macrophyte. Am J Bot 81 : 1004-1012 (1994). Loveless MD, and Hamrick JL, Ecological dete rminants of genetic-structure in plantpopulations. Ann Rev Ecol and Systematics 15 : 65-95 (1984). Ly ák MA, Rostková A, Dixon JM, Rossi G, and Dole el J, Limited genome size variation in Sesleria albicans . Ann Botany 86 : 399-403 (2000). Maas SJ, and Dunlap JR, Reflectance, transmittance, and absorptance of light by normal, etiolated, and albi no corn leaves. Agron J 81: 105-110 (1989). Madeira PT, Van TK, Steward KK, and Raymond JS, Random amplified polymorphic DNA analysis of the phenetic relationsh ips among world-wide accessions of Hydrilla verticillata. Aquatic Botany 59 : 217-236 (1997). Madsen JD, and Owens CS, Seasonal bi omass and carbohydrate allocation in the dioecious hydrilla. J Aquat Plant Manage 36 : 138-145 (1998). Madsen JD, and Smith DH, Vegetative spread of dioecious hydr illa colonies in experimental ponds. J Aquat Plant Manage 37 : 25-29 (1999). Magnin NC, Cooley BA, Reiskind JB, and Bowe s G, Regulation and localization of key enzymes during the induction of Kranz-Less, C4-type photosynthesis in Hydrilla verticillata . Plant Physiol 115 : 1681-1689 (1997). Maniatis T, Fritsch EF, and Sambro ok J, Gel electrophoresis of RNA. In : Molecular Cloning: A Laboratory Manua l, pp 199-203, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY (1982). Mappleback LR, Souza Machado V, and Grodzinski B, Seed germination and seedling growth characteristics of atrazine-s usceptible and resi stant biotypes of Brassica compestris . Can J Plant Sci 62 : 733-739 (1982).

PAGE 140

127 Martinez-Ferez IM, and Vioque A, Nucleotide sequence of the phytoene desaturase gene from Synechocystis sp. PCC 6803 and characterizati on of a new mutation which confers resistance to th e herbicide norflurazon. Plant Mol Biol 18 : 981-983 (1992). Martyn RD, Color infrared photography for dete rmining the efficacy of grass carp in aquatic weed control. Proc SWSS 38 : 381-390 (1985). Masterson J, Stomatal size in fossil plan ts: evidence for polyploi dy in majority of angiosperms. Science 264 : 421-424 (1994). Maxwell BD, and Mortimer MM, Sele ction for herbicide resistance. In :Herbicide Resistance in Plants: Biology and Bioc hemistry (eds. SB Powles, and JAM Holtum), pp 1-25. CRC Press, Boca Raton, FL (1994). Mayer MP, Bartlett DL, Beye r P, and Kleinig H, The in vitro mode of action of bleaching herbicides on the desaturation of 15cis -phytoene and cis -carotene in isolated daffodil chromoplasts. Pest Biochem Physiol 34 : 11-117 (1989). McFarland DG, and Barko JW, High temp erature effects on growth and propagule formation in hydrilla biotypes. J Aquat Plant Manage 37 : 17-25 (1999). Meyer A, and Van de Peer Y, ‘Natural selection merely modified while redundancy created’Susumu Ohno’s idea of the evolutionary impor tance of gene and genome duplications. J Struct Funct Genomics 3:vii-ix (2003). Michel A, Scheffler BE, Arias RS, Duke SO, Netherland M, and Dayan FE, Somatic mutation-mediated evolution of herbicide resistance in the nonindigenous invasive plant hydrilla ( Hydrilla verticillata ). Mol Ecol 13 : 3229-3237 (2004). Miller JD, Haller WT, and Glenn MS, Turion production by dioecious hydrilla in north Florida. J Aquat Plant Manage 31 : 101-105 (1993). Misawa N, Masamoto K, Hori T, Ohtani T, Boger P, and Sandmann G, Expression of an Erwinia phytoene desaturase gene not only confers multiple resistance to herbicides interfering with carotenoid biosynthesis but also alters xanthophyll metabolism in transgenic plants. Plant J 6 : 481-489 (1994). Misawa N, Nakagawa M, Kobayashi K, Yamano S, Izawa Y, Nakamura K, and Harashima K, Elucidation of the Erwinia uredovora carotenoid biosyntheticpathway by functional-analysis of gene-products expressed in Escherichia coli. J Bacteriol 172 : 6704-6712 (1990). Mitra E, Contributions to our knowledge of Indian freshwater plants. 1. On some aspects of the structure and life history of Hydrilla verticillata Presl. With notes on its autecology. J Asiatic Soc Sci 21 :1-17 (1955). Muse SV, Examining rates and patterns of nucleotide substitution in plants. Plant Mol Biol 42: 25–43 (2000).

PAGE 141

128 Nagl W, DNA endoreduplication and polyteny understood as evolut ionary Strategies. Nature 261 : 614-615 (1976). Nakamura T, and Kodono Y, Chromosome nu mber and geographical distribution of monoecious and dioecious Hydrilla verticillata (L.f.) Royle (Hydrocharitaceae) in Japan. Acta Phytotax Geobot 44: 123-140 (1993). Nakamura T, Suzuki T, and Kadono Y, A comparative study of isozymes of Hydrilla verticillata (L.f) Royle in Japan. J Plant Res 111 : 581-585 (1998). Nandini AV, Murray BG, O’ Brien IEW, a nd Hammett KRW, Intraand interspecific variation in genome size in Lathyrus (Leguminosae). Bot J Linnean Soc 125 : 359366 (1997). Navarro E, Sandmann G, and Santiago TM, Mu tants of carotenoid biosynthetic pathway of Mucor Circinelloides . Exp Mycol 19 : 186-190 (1995). Nelson LS, Shearer JF, and Netherland MD , Mesocosm evaluation of integrated fluridone-fungal pathogen treatment on four submersed plants. J Aquat Plant Manage 36 : 73-77 (1998). Netherland MD, and Getsinger KD, Laborat ory evaluation of threshold fluridone concentrations under static conditions for controlling hydrilla and Eurasian watermilfoil. J Aquat Plant Manage 33 : 33-36 (1995). Netherland MD, and Shearer JF, Integrated use of fluridone and a fungal pathogen for control of hydrilla. J Aquat Plant Manage 34 : 4-8 (1996). Netherland MD, Dayan FE, Scheffler B, and Cockreham S, Three and a half years of field monitoring of fluridone-toleran t hydrilla: what have we learned? Abstract, APMS Annual Meeting , Keystone, CO (2002). Netherland MD, Getsinger KD, and Skogerboe JD, Mesocosm evalua tion of the speciesselective potential of fluridone. J Aquat Plant Manage 35 : 41-50 (1997). Netherland MD, Getsinger KD, and Turner EG, Fluridone concentration and exposure time requirements for control of Eura sian water milfoil and hydrilla. J Aquat Plant Manage 31 : 189-194 (1993). Office of Technology Assessment, Two case stud ies: nonindigenous plants in Hawaii and Florida. In : Harmful non-indigenous species in the United States, pp 233–266, OTA-F565, US Government Printing Offi ce, US Congress, Washington, DC (1993). Ogawa H, Yamada I, Arai K, Hirase K, Mori yasu K, Schneider C, Sandmann G, Boger P, and Wakabayashi K, Mode of bleach ing phytotoxicity of herbicidal diphenylpyrrolidinones. Pest Manage Sci 57 : 33-40 (2001).

PAGE 142

129 Pecker I, Chamovitz D, Linden H, Sandmann G, and Hirschberg J, A single polypeptide catalyzing the conversion of phytoene to -carotene is transcriptionally regulated during tomato fruit ripening. PNAS 89 : 4962-4966 (1992). Pendland J, Ultrastructural characteristics of Hydrilla leaf tissue. Tissue Cell 11 : 79-88 (1979). Pennington TG, Skogerboe JG, and Getsinger KD, Herbicide/copper combinations for improved control of Hydrilla verticillata. J Aquat Plant Manage 39 : 56-58 (2001). Pieterse AH, Staphorst HPM, and Verkli ej JAC, Some effects of nitrogen and phosphorous concentration on the phenology of Hydrilla verticillata (L.f.) Royle. J Aquat Plant Manage 22 : 62-63 (1984). Pieterse AH, Verkleij JAC, and Staphorst HPM, A comparative study of isoenzyme patterns, morphology and chromosome number of Hydrilla verticillata (L.F.) Royle in Africa. J Aquat Plant Manage 23 : 72-76 (1985). Pimentel D, Lach L, Zuniga R, and Morri son D, Environmental and economic costs of nonindigenous species in the United States. Biosci 50 : 53-65 (2000). Pimm S, and Galpin M, Theoretical issues in conservation biology. In : Perspectives in Ecological Theory (eds.J Roughgard en, R May, and S Leven), pp 287-305, Princeton University Press, Princeton, NJ (1989). Poovey AG, and Kay SH, The potential of a summer drawdown to manage monoecious hydrilla. J Aquat Plant Manage 36 : 127-130 (1998). Puri A, MacDonald GE, and Haller WT, Grow th and physiological differences between fluridone tolerant and sus ceptible hydrilla populations. Proc FWSS , Ft. Pierce, Florida. pp 8 (2004). Radosevich SR, and Holt JS, Physiological re sponses and fitness of susceptible and resistant weed biotypes of atrazine herbicides. In : Herbicide Resistance in Plants (eds. HM LeBaron and J Gressel), pp 163-183, John Wiley and Sons, New York, NY (1982). Randall JM, Weed control for the pr eservation of biological diversity. Weed Tech 10 : 370-383 (1996). Rocher EJD, Harkins KR, Galbraith DW, a nd Bohnert HJ, Developmentally regulated systemic endopolyploidy in succulents with small genomes. Science 250 : 99-101 (1990). Ryan FJ, and Hommberg DL, Keeping track of hydrilla. Aquatics 16: 14–20 (1994). Ryan FJ, Coley CR, and Kay SH, Coexistenc e of monoecious and di oecious hydrilla in lake Gaston, North Carolina and Virginia. J Aquat Plant Manage 33 : 8-12 (1995).

PAGE 143

130 Ryan FJ, Thullen JS, and Holmberg DL, Non-ge netic origin of isoenz ymic variability in subterranean turions of monoecious and dioecious Hydrilla. J Aquat Plant Manage 29 : 3-6 (1991). Sainty GR, and Jacobs SWL, Water plants of New South Wales. Water Resources Commission, NSW, pp 239-241 (1981). Sandmann G, and Böger P, Comparison of the bleaching activity of norflurazon and oxyfluorfen. Weed Sci 31 : 338-341 (1983). Sandmann G, and Mitchell G, In vitro inhibition studies of phytoene desaturase by bleaching ketomorpholine derivatives. J Agr Food Chem 49 : 138-141 (2001). Sandmann G , Bleaching activities of substituted pyrimidines and structure activity comparison to related heterocyclic derivatives. Pest Biochem Physiol 70 : 86-91 (2001). Sandmann G, Carotenoid biosynthesis in microorganisms and plants. Eur J of Biochemistry 223: 7-24 (1994). Schmitz DC, The invasion of exotic aquatic a nd wetland plants into Florida: history and effects to prevent new introductions. Aquatics 12 : 6-24 (1990). Schmitz DC, and Osborne JA, Zooplankton densities in a Hydrilla infested lake. Hydrobiologia 111 : 127–132 (1984). Schmitz DC, Nelson BV, Nall LE, and Schardt JD, Exotic aquatic plants in Florida: a historical perspective and review of th e present aquatic plant regulation program. Proc Symposium on Exotic Pest Plants, Nov 2-4 1998, Miami FL, pp. 303-323 (1991). Schneider C, Boger P, and Sandmann G, Phytoene desaturase: Heterologous expression in an active state, purification, and biochemical properties. Protein Express Purif 10: 175-179 (1997). Sculthorpe CD, The biology of aquatic vascular plants, pp 610, Arnold Ltd, London, England (1967). Sharma AK, and Bhattacharyya B, A study of Hydrocharitaceae as an aid to trace the lines of evolution. Python (Buenos Aires) 6 : 121-132 (1956). Shearer JF, Biological cont rol of hydrilla using an endemic fungal pathogen. J Aquat Plant Manage 36 : 54-56 (1998). Siefermannharms D, Light and temperature co ntrol of season-dependent changes in the alpha-carotene and beta-carotene content of spruce needles. J Plant Physiol 143 : 488-494 (1994 )

PAGE 144

131 Simberloff D, How much information on popul ation biology is needed to manage introduced species? Conserv Biol 17 : 83-92 (2003). Skogerboe JG, and Getsinger KD, Endothall species selectivity evaluation: Southern latitude aquatic plant community. J Aquat Plant Manage 39 : 129-135 (2001). Sliwinska E, and Lukaszewska E, Polyso maty in growing in vitro sugar-beet ( Beta vulgaris L.) seedlings of different ploidy level. Plant Sci 168 : 1067-1074 (2005). Smart CC, and Fleming AJ, A plant gene w ith homology to D-myo-inositol-3-phosphate synthase is rapidly and spatially up-re gulated during an abscisic-acid-induced morphogenic response in Spirodela polyrrhiza . Plant J 4 : 279-293 (1993). Smith CS, and Pullman GD, Experiences using Sonar A.S. aquatic herbicide in Michigan. J Lake Res Manage 13 : 338-346 (1997). Smulders MJM, Rus-Kortekaas W, and G illissen LJW, Development of polysomaty during differentiation in diploi d and tetraploid tomato ( Lycopersicon esculentum ) plants. Plant Sci 97 : 53-60 (1994). Smulders MJM, Rus-Kortekaas W, and Gillisse n LJW, Natural variation in patterns of polysomaty among individual tomato plan ts and their regenerated progeny. Plant Sci 106 : 129-139 (1995). Southeast Exotic Pest Plant Council (SE-EPPC), Web page http://www.seepcc.org/manual/HYVE.html (Last accessed, 15 April 2005) Spencer D, Ryan FJ, Aung L, and Ksander G, Soluble sugar concentrations associated with tuber and winter bud sprouting. J Aquat Plant Manage 39 : 45-47 (2001). Spencer DF, and Anderson LWJ, Photoperiod responses in monoecious and dioecious Hydrilla verticillata. Weed Sci 34 : 551-557 (1986). Spencer DF, Anderson LWJ, Ames MD, and Ryan FJ, Variation in Hydrilla verticillata (L.f.) Royle propagule weight. J Aquat Plant Manage 25 : 11-14 (1987). Sprecher SL, Netherland MD, and Stewart AB, Phytoene and carotene response of aquatic plants to fluridone under laboratory conditions. J Aquat Plant Manage 36 : 112-120 (1998). Steward KK, Seed production in monoecious and dioecious popul ations of hydrilla. Aquatic Botany 46: 169-183 (1993). Steward KK, Influence of photoperiod on t uber production in vari ous races of hydrilla ( Hydrilla verticillata ). Hydrobiologia 354 : 57-62 (1997).

PAGE 145

132 Steward KK, Influence of photoperiod on ve getative propagule production in three turion-producing races of Hydrilla verticillata (L.f.) Royle. Hydrobiologia 432 : 1-8 (2000). Steward KK, and Van TK, Comparative studies of monoecious and dioecious Hydrilla ( Hydrilla verticillata ) biotypes. Weed Sci 35 : 204-210 (1987). Steward KK, Van TK, Carter V, and Pieterse AH, Hydrilla invades Washington D.C. and the Potomac. Amer J Bot 71 : 162-163 (1984). Sutton DL, Biology and ecology of Myriophyllum spicatum . In : Ist International Symposium Watermilfoil and related Halo ragaceae Species (eds. LW Anderson), pp 59-71 (1985). Sutton DL, and Portier KM, Growth of dioeci ous hydrilla in sediments from six Florida lakes. J Aquat Plant Manage 33 : 3-7 (1995). Sutton DL, Van TK, and Portier KM, Growth of dioecious and monoecious Hydrilla from single tubers. J Aquat Plant Manage 30 : 15-20 (1992). Tan S, Evans RR, Dahmer ML, Singh BK, and Shaner DL, "Imidazolinone-tolerant crops: history, current status and future" Pest Manage Sci 61 : 246-257 (2005). Thakore JN, Haller WT, and Shilling DG, S hort day exposure period for subterranean turion formation in dioceous hydrilla. J Aquat Plant Manage 35 : 60-63 (1997). Thullen JS, Production of axillary turions by the dioecious Hydrilla verticillata. J Aquat Plant Manage 28 : 11-15 (1990). Tranel PJ, and Wright TR, Resi stance of weeds to ALS-inhibiting herbicides: what have we learned? Weed Sci 50 :700-712 (2002). Triest L, Conservation of gene tic diversity in water plants. In : Isozymes in Water Plants (eds. L Triest ), pp 201-240, Opera Botanica Belgica 4, National Botanical Garden of Belgium, Meise (1991). Union of Concerned Scientists (UCS), Web page http://www.ucsusa.org/global_ environment/invasive_species , (Last accessed, 19 December 2005). United States Environmental Protection Agency (USEPA), Region/ORD nonindigenous species—An emerging issue for the EPA, Volume 2: A landscape in transition effects of invasive species on ecosyst ems, human health and EPA goals, EPA report, May 2001, prepared by Lee II H and Chapman JW (2001). United States Geographical Survey (USGS), Web page http://nas.er.usgs.gov/taxgr oup/plants/maps/hy_vert_bio.jpg (Last accessed, 15 September 2002)

PAGE 146

133 Van TK, Haller WT, and Bowes G, Comparison of the photosynthetic characteristics of three submersed aquatic plants. Plant Physiol 58: 761-768 (1976). Van TK, Haller WT, and Garrard LA, The e ffect of day length and temperature on hydrilla growth and tuber production. J Aquat Plant Manage 16 : 57-59 (1978). Van TK, and Steward KK, The use of controlled -release fluridone fibers for control of hydrilla ( Hydrilla verticillata ). Weed Sci 34 : 70-76 (1985). Van TK, and Steward KK, Longevity of monoecious Hydrilla propagules. J Aquat Plant Manage 28 : 74-76 (1990). Vandiver VV, 1999 Florida Aquatic Weed Mana gement Guide. Special Publication # 55 (1999). Verkleij JAC, and Pieterse AH, Identification of Hydrilla verticillata (L.f.) Royle strains by means of isoenzyme patterns. Proc EWRS, Loughborough, UK, pp 381-388 (1986). Verkleij JAC, and Pieterse AH, Isoenzyme patterns in leaves of Hydrilla verticillata (Hydrocharitaceae). In : Isozymes in Water Plants (eds. L Triest ), pp 49-57, Opera Botanica Belgica 4, National Botanical Garden of Belgium, Meise (1991). Verkleij JAC, Pieterse AH, Horneman GJT, and Torenbeek , A comparative-study of the morphology and isoenzyme patterns of Hydrilla verticillata (L.f.) Royle. Aquatic Botany 17 : 43-59 (1983a). Verkleij JAC, Pieterse AH, Staphorst HPN, and Steward KK, Identification of two different genotypes of Hydrilla verticillata (L.f.) Royle in the USA by means of isonzyme studies. Proc Int Symp Aquatic Macrophytes , The Netherlands, September 18-23, 1983. pp 251-261 (1983b). Volenberg DS, Stoltenberg DE, and Boerboom CM, Green foxtail ( Setaria viridis ) resistance to acetolactate synthase inhibitors. Phytoprotection 83 : 9-109 (2002). Walsh MJ, Powles SB, Beard BR, Parkin BT, and Porter SA, Multiple-herbicide resistance across four modes of action in wild radish ( Raphanus raphanistrum ). Weed Sci 52 : 8-13 (2004). Wardlaw CW, Morphogenesis in Plants, pp 451, Methuen Publications, London (1968). Way JM, and Chancellor RJ, Herbicides and higher plant ecology. In : Herbicides (eds. LJ Audus), pp 345-372, Academic Press, London (1976). Wendel JF, Genome evolution in polyploids. Plant Mol Biol 42 : 225-249 (2000).

PAGE 147

134 Westerdahl HE, and Getsinger KD, Aquatic pl ant identification and herbicide use guide. Vol. I: Aquatic herbicides and applic ation equipment. US Army Corps of Engineers, Technical Report A-88-9, Vicksburg, Mississippi pp 145 (1988). Windhövel U, Geiges B, Sandmann G, and Böger P, Expression of Erwinia uredovora phytoene desaturase in Synechococcus PCC7942 leading to resistance against a bleaching herbicide. Plant Physiol 104 : 119-125 (1994). Windhövel U, Sandmann G, and Boger P, Gene tic engineering of re sistance to bleaching herbicides affecting phytoene desaturase and lycopene cyclase in cyanobacteria l carotenogenesis. Pest Biochem Physiol 57 : 68-78 (1997). Yang J, The formation and evolution of polyploid genomes in plants. Acta Phytol Sin 39 : 357-371 (2001). Yeo RR, and McHenry WB, Hydrilla, A ne w noxious aquatic weed in California. Calif Agric 31 : 4-5 (1977). Yeo RR, Falk RH, and Thurston JR, The morphology of hydrilla ( Hydrilla verticillata (L.f.) Royle). J Aquat Plant Manage 22 : 1-17 (1984).

PAGE 148

135 BIOGRAPHICAL SKETCH Atul Puri was born on January 7, 1979, at Ja landhar (Punjab) in India. He did his primary education in Jammu, India, and develope d interest in science at an early age. He graduated from Model Academy High School in 1995 and received a Bachelor of Science degree in agriculture from P unjab Agricultural University, Ludhiana (Punjab), India, in July1999. In August 1999, he entered graduate sc hool at the same institution, and did his graduate work on the effects of growth reta rdants and environment factors in American cotton. He received a Master of Science degree in agronomy in December 2001. Upon graduating, he accepted a job of Assistant Crop P hysiologist with the Punjab Agricultural University, Ludhiana. His work focused on evaluating salinity, sodicity, and waterlogging resistant wheat and rice genotypes In August 2002, he enrolled at the Univers ity of Florida to pursue a Doctor of Philosophy degree under Dr. G. E. MacDonald. He has won many academic awards and scholarships. He was awarded the University Gold Medal for the best student in the bachelorÂ’s program, and Labh Singh Gold Me dal for the best student in Agronomy Department at PAU, Ludhiana. Recently he was awarded the Outstanding Graduate Student award from the Florida Weed Science Society for the year 2006. He has presented many talks and posters at the Weed Science Society of America, the Southern Weed Science Society, the A quatic Plant Management Society, and the Florida Weed Science Society, and is a member of the Gamma Sigma Delta honor society.