Genomic DNA restriction fragment length polymorphisms at a highly polymorphic locus distinguish Old and New World subspe...

MISSING IMAGE

Material Information

Title:
Genomic DNA restriction fragment length polymorphisms at a highly polymorphic locus distinguish Old and New World subspecies of the honey bee, Apis mellifera L.
Physical Description:
vi, 111 leaves : ill., photos ; 29 cm.
Language:
English
Creator:
McMichael, Margaret Anne
Publication Date:

Subjects

Subjects / Keywords:
Honeybee -- Identification   ( lcsh )
Honeybee -- Genetics   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1994.
Bibliography:
Includes bibliographical references (leaves 103-110).
Statement of Responsibility:
by Margaret Anne McMichael.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001986587
notis - AKG3557
oclc - 32021354
System ID:
AA00004724:00001

Full Text










GENOMIC DNA RESTRICTION FRAGMENT LENGTH POLYMORPHISMS
AT A HIGHLY POLYMORPHIC LOCUS
DISTINGUISH OLD AND NEW WORLD SUBSPECIES
OF THE HONEY BEE, Apis mellifera L.














By

MARGARET ANNE MCMICHAEL


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

UNIVERSITY OF FLORIDA


1994














ACKNOWLEDGMENTS


I am grateful for the efforts of so many people, and I am bothered by the

omission of many names from these two pages. My thanks go to all who have

helped me.

First and foremost, Glenn Hall took a chance on me and has been patient

and generous in providing the opportunity, direction, and freedom I have

enjoyed in the course of obtaining my doctorate in his lab. The members of my

advisory committee provided tremendous assistance; Andy Cockburn, Jim

Maruniak, and Buffy Bondy. Don Campton generously donated his time in the

initial stages of my research, and Jan Conn reviewed my manuscripts and

postdoctoral research proposal. Chip Taylor provided a memorable week in

Linnares, Mexico, and taught me the zen of smoker-lighting.

Lois Lemmerman Myeroff, Case Western Reserve University, is

responsible for teaching me all I knew about molecular biology when I was in

the Mapstone/Goldthwait lab and as such was instrumental in my move to

Florida. Al6jandra Garcia and Raquel McTeirnan have been immeasurably

helpful to me here, and my way would have been rough without their tireless

and unselfish assistance. Marjorie Hoy's lab members--particularly Owain

Edwards (who also reviewed my manuscripts), Jim Presnail, Dr. Jey, and Greg









McDermott--have been instrumental in advancing my research efforts. Reg

Coler, Scott Yocom, and Owain Edwards helped me make slides for the

meetings and for my departmental seminar.

Dr. Strayer, Faith Oi, and Hugh Smith taught me a few things during one

of my best experiences here; assisting in Principles of Entomology. Don Hall,

Tom Dykstra, and Robin Goodson provided invaluable opportunities for me to

lead their classes and tours of the Bee Lab. Scott Yocom recruited me for the

Linnean Team, which was an exciting, not to mention a humbling, experience.

Deserving additional, special mention are John Strayer and the great students

he recruited for the department, with whom I have had the honor and pleasure

to work.













TABLE OF CONTENTS



ACKNOW LEDGMENTS ................................... ii

ABSTRACT .......................................... v

INTRODUCTION ........ ......... ...................... 1

IDENTIFICATION AND GEOGRAPHICAL DISTRIBUTION OF ALLELES AT
LOCUS 178 ...................................... 5


Introduction .......
Materials and Methods
Results ...........
Discussion ........


ALLELE FREQUENCIES AT LOCUS 178 REVEAL HYBRIDIZATION OF
EUROPEAN AND AFRICAN BEES IN THE NEOTROPICS ....... 36


Introduction ...................................
Materials and Methods ............................
Results ... ........... ...... ....... .. .... ......
Discussion ....................................

LOCALIZATION OF VARIATION AT LOCUS 178 IN Apis mellifera (L.) BY
RESTRICTION MAPPING ...........................

Introduction ...................................
Materials and Methods ............................
Results .............. ..... ....... .. .. ....... ..
Discussion ....................................

CONCLUDING REMARKS ...............................

LITERATURE CITED ..................................

BIOGRAPHICAL SKETCH ...............................


. 65

. 65
. 67
. 69
. 93

..99

.103

. 111


Illllllil


lijiil














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

GENOMIC DNA RESTRICTION FRAGMENT LENGTH POLYMORPHISMS
AT A HIGHLY POLYMORPHIC LOCUS
DISTINGUISH OLD AND NEW WORLD SUBSPECIES
OF THE HONEY BEE, Apis mellifera L.

By

Margaret Anne McMichael

April 1994

Chairperson: H. G. Hall
Major Department: Entomology and Nematology


Honey bees (Apis mellifera L.) of African and European ancestry were

distinguished by analysis of restriction fragment length polymorphisms (RFLPs)

defined by two enzymes at a highly polymorphic locus corresponding to

genomic probe pB178. Thirty-six Mspl restriction fragment patterns, or

variants, and thirty-three Ddel variants were identified. Sixty-five pairwise

combinations of the Mspl and Ddel variants, referred to as alleles, were found

among the individual drones tested. Variants and alleles were discontinuously

distributed in USA and South African drones: only one Ddel variant and no

Mspl variant (hence no allele) were common to these two sample populations.

The diversity in the variants and alleles found in the South African drones was









greater than in the USA drones. Mspl variants were discontinuously distributed

among workers bees of the subspecies A. m. mellifera (west European), A. m.

ligustica and A. m. caucasica (east European), and A. m. scutellata (South

African).

Ancestry in New World bees was inferred from variant and allele

frequencies at this locus. In USA bees, variants characteristic of east and west

European bees were found at frequencies consistent with previously identified

nuclear and mitochondrial DNA markers. In neotropical drones, European and

African origins were assumed for variants that were similar in fragment pattern

to variants in the USA and South Africa, respectively. The alleles identified in

the majority of neotropical colonies were African. There was little evidence of

hybridization of African and east European bees in neotropical bees: variants

specific to South African bees were detected at the highest frequencies, while

east European-specific variants were absent or detected at very low

frequencies. A variant that may be specific for A. m. mellifera was found

throughout the neotropics at frequencies that varied from 19% to 33% but did

not increase clinally in a northward direction.

Determination of the allelic relationships between Mspl fragments and

between Ddel fragments was initiated by mapping these restriction sites within

the probe, and correlating fragments on Southern blots to the locations of

restriction sites on the resulting map.














INTRODUCTION


Descendants of ancestral honey bees migrated from the site of their

origin, presumed to be northeast Africa or the Middle East, and became widely

distributed in the Old World (Garnery, Cornuet & Solignac 1992). Subsequent

adaptation to a wide variety of ecological conditions resulted in the evolution

of subspecies of the honey bee, Apis mellifera L. (Ruttner 1988). These

subspecies are distinguished on the basis of physical, behavioral, and ecological

criteria and are defined quantitatively by discriminant analysis of morphological

characters (Daly & Balling 1978; Ruttner 1988; Ruttner, Tassencourt &

Louveaux 1978).

Honey bees were first introduced to the New World by 17th and 18th

century European settlers. For over two hundred years, the genetic diversity

found in New World honey bees resulted from the importation of primarily four

European subspecies or races: west and north European A. m. mellifera

(referred to here as west European); north Mediterranean A. m. ligustica Spinola

and A. m. carnica Pollmann (which in this report will be called east European;

Ruttner 1988); and the east European bee A. m. caucasica Gorbachev (referred

to here as east European) (Kent 1988; Kerr, DeLeon & Dardo 1982).









2
Despite abundant nectar sources in the neotropics, self-sustaining feral

populations of European bees did not become established (Michener 1975).

The poor performance of European bees in the tropics has been attributed to

their failure to adapt to environmental conditions vastly different from those in

which they evolved (Ruttner 1988). To improve commercial honey production,

queens of a central and south African race, A. m. scutellata Lepeletier, were

imported into Brazil in 1956 for experimental breeding with European bees.

Swarms of African bees escaped into the wild (Kerr 1967) and subsequently

proliferated to form large feral populations that spread through tropical South

and Central America. African bees entered Texas in 1990 and Arizona in 1993.

The release and spread of African bees has been disruptive to the

behavior and ecology of the melliferous flora and fauna in the neotropics

(Cantwell 1974; Michener 1975; Roubik 1980, 1989; Spivak, Fletcher & Breed

1991; Taylor 1977; Winston, Taylor & Otis 1983). Explanations for the

migration pressure resulting from the introduction of the bees from Africa and

for the rapid colonization of New World territories previously unoccupied by

honey bees have focused on the extent to which neotropical bees represent an

admixture of African and European subspecies (Hall 1990, 1991, 1992a; Hall

& Muralidharan 1989; Rinderer 1986; Rinderer et al. 1985, 1991; Sheppard et

al. 1991; Smith, Taylor & Brown 1989; Taylor 1985). The ability to distinguish

honey bee subspecies is essential to ascertain the extent of hybridization. In

turn, the degree of hybridization will influence the northern limit of African bee









3
introgression in the USA, about which there is considerable debate (Dietz 1986;

Dietz, Krell & Pettis 1986; Rinderer 1986; Roubik 1986; Southwick, Roubik &

Williams 1990; Taylor 1977; Taylor & Spivak 1984; Villa, Gentry & Taylor

1987; Villa, Rinderer & Collins 1993).

The retention of African morphology and behavior in neotropical bees has

stimulated the quest for the identification of their ancestry, which will aid in

understanding the mechanisms contributing to their phenomenal success.

Efforts to ascertain the extent of interaction between the extant European bees

and the African bees in the neotropics have been hindered by a lack of genetic

markers specific for each Old World subspecies introduced. Intermediate values

of morphometric values and allozyme frequencies, taxonomic characters

traditionally used to identify honey bee subspecies and subspecies groups, can

only suggest that hybridization between races has occurred (Daly 1991; Del

Lama et al. 1988; Nunamaker et al. 1984; Spivak etal. 1988; Sylvester 1982,

1986). Thus, supplemental genetic characters subject to little environmental

modification have thus been sought.

Mitochondrial DNA (mtDNA) has been used effectively to identify honey

bee subspecies within populations in biogeographic studies (Arias, Soares &

Nobrega 1990; Cornuet & Garnery 1991; Garnery, Cornuet & Solignac 1992;

Hall & Smith 1991; Meixner, Sheppard & Poklukar 1993; Smith & Brown 1988,

1990; Smith 1991; Smith et al. 1991). Honey bee mtdna types are

discontinuously distributed among subspecies, reflecting the sustained temporal









4
separation and independent evolution of temperate and tropical lineages (Avise

et al. 1987; Cornuet & Garnery 1991). African mtDNA has been found in

almost all feral neotropical colonies. The small proportion of European mtDNA

that has persisted has remained largely confined to managed apiaries (Hall and

Muralidharan 1989; Hall & Smith 1991; Smith, Taylor & Brown 1989). These

findings supplied strong evidence that unbroken matrilines spreading as swarms

have been primarily responsible for the expansion of the African bee population

in the neotropics (Hall & Muralidharan 1989; Hall & Smith 1991; Smith, Taylor

& Brown 1989). Yet the features of mtDNA that make these molecules good

genetic markers (i.e. uniparental inheritance, no recombination) limit the

information that can be obtained from their analysis.

Genomic RFLPs at several loci have been identified that distinguish east

European from west European and African bees, but not west European from

African bees (Hall 1990, 1992b). Frequencies for the markers indicated limited

hybridization of African with European bees, but it was not possible to assess

hybridization with west European bees. Alleles at two loci, found at different

frequencies in African and west European bees, were found at intermediate

frequencies in neotropical bees, which could have resulted from hybridization

between these two groups of subspecies (Hall 1992b).

The objective of the research reported here was to identify genomic DNA

(molecular) markers or characters that distinguish among the east European,

west European, and African groups of subspecies.














IDENTIFICATION AND GEOGRAPHICAL DISTRIBUTION
OF ALLELES AT LOCUS 178


Introduction


Nuclear DNA restriction fragment length polymorphisms (RFLPs) are

biparentally inherited, codominantly expressed markers shown to be a valuable

source of genetic variability in honey bees (Hall 1986, 1990, 1992b). Relative

to the biochemical variability that can be discerned in allozyme frequencies,

RFLPs are more abundant and do not require gene expression for their

detection. East European- and African-specific RFLP markers have been

identified, as well as markers common to African and a west European species

(Hall 1986, 1990, 1992b). In neotropical colonies, the east European markers

have been found at very low frequencies in areas where African bees have

become established, indicating that there has been limited paternal introgression

from European colonies into the African bee population (Hall 1990). One

limitation to date in the use of RFLPs to study ancestry in neotropical bees has

been the lack of additional markers specific for other subspecies, particularly

A. m. mellifera.

Described in this chapter are nuclear DNA RFLPs, detected with a single

genomic probe, that distinguish bees of European and African ancestry. While









6
it is not possible to know the genotype of each queen bee introduced, nor

many details of the events contributing to the establishment of New World

honey bee populations, some reconstruction can be acheived by studying

contemporary bees in the Old and New World. Evidence is provided that

reflects the racial composition of Central and South American bees prior to the

introduction of African bees, and indicates the occurrence of some, albeit

limited, hybridization of European and African bees.



Materials and Methods


Sources of honey bees. Protocols described by Hall (1986, 1990) for the

collection, caste determination, and preservation of bees were followed without

modification. Drones were collected as larvae and pupae.

South African samples, from four locations in the Transvaal, were

collected in January 1990 by HGH. Brood samples from a colony in So Frango,

Brasilia, were obtained in 1990 by J. Maruniak, University of Florida. Honduran

samples included drones from feral colonies and from managed colonies

established from feral swarms, collected in November 1989 by HGH and A.

Suazo, Escuela Agricola Panamericana, Honduras. Samples from Tapachula,

Mexico, obtained in January 1988 by HGH, were from feral swarms captured

in bait hives maintained by the Mexican agency Secretariat of Agriculture and

Hydrologic Resources (SARH), and from two managed apiaries. Sources of

USA drones included a closed breeding population maintained in Arizona,









7
provided by J. Martin, G. Waller, G. Loper, and E. Erickson (Page, Erickson &

Laidlaw 1982; Severson, Page & Erickson 1986); managed colonies in Kansas,

provided by O. Taylor, University of Kansas; the University of Florida apiaries;

and the Kona Queen Company, Captain Hook, HI.

Electrophoretic analysis of honey bee DNA. The cloned probe pB178

came from a library of Pstl-digested European honey bee DNA ligated into the

plasmid vector pBR322 (Hall 1986). This clone was used as a radioactive

probe for detecting RFLPs in honey bee DNA digested with restriction

endonucleases and separated electrophoretically. Isolation of genomic DNA,

restriction endonuclease digestions, electrophoresis, probe preparation and

labeling, blotting, and hybridizations were conducted as previously described

or cited (Hall 1986, 1990) without further modification.

Initial detection of polymorphisms. In the initial search for

polymorphisms, DNA was isolated from a pool of ten workers from a New

World European (USA) colony and from a New World African (Costa Rica)

colony (Hall 1986, 1990). DNA from each pooled sample was digested

separately with nine restriction endonucleases, and separated in 2% agarose

gels. The restriction fragment profiles for the European and African samples

were compared following hybridization with cloned probe pB178. The sizes (in

kilobase pairs, kb) of the restriction fragments were estimated by comparison

to size standards using a HI-PAD digitizer (Houston Instruments).

Polymorphisms were detected in the Mspl- and Ddel-treated samples.









8
Additional individuals from many locations in the New and Old World were then

analyzed to ascertain the subspecies and/or geographic distribution of the

polymorphic fragments.

The polymorphisms generated by Mspl and Ddel are described here. For

the identification of the Mspl variants, 402 drone bees were analyzed. For the

identification of the Ddel variants, 508 drones were analyzed, including 401 of

those analyzed for the Mspl variants. The pairwise combinations of the Mspl

and Ddel variants found in these 401 drones are referred to as alleles.



Results


Allelic nature of the polymorphisms.

Initially, polymorphisms were detected with pB178 following Mspl and

Ddel digestion of DNA extracted from a sibling pool of New World European

and African worker (female) honey bees. In the Mspl digests, a 1.8kb fragment

was detected only in the European sample, and a unique 1.1kb fragment was

detected only in the African sample. In the Ddel digests, the European sample

contained a unique 1.25kb fragment, while the African sample contained

unique 3.4kb and 1.0kb fragments. RFLPs are codominantly expressed in

diploid workers, which can result in the comigration and masking of fragments.

Therefore, further characterization of the Mspl and Ddel RFLPs was

accomplished by analyzing drones (males), which are haploid offspring,

parthenogenetically derived from the queen.









9
The correspondence of the region of the genome detected by probe

pB178 to a contiguous, heritable segment of a chromosome, or locus (Fincham

1983; Hall 1990, 1992b), was demonstrated by analyzing at least six drone

progeny from a single colony (queen). For each pB178/restriction enzyme

combination, no more than two restriction fragment patterns were detected,

which reflected the segregation of the queen's alleles (not shown). Allelic

polymorphisms in individual drones were manifested as different restriction

fragment patterns. The different patterns generated by a single restriction

enzyme are referred to as variants. The pairwise combinations of the variants

produced by the two enzymes in individual drones are referred to as alleles.


Identification of Mspl variants.

Most of the 36 Mspl restriction patterns or variants identified are shown

diagrammatically in Figure 1. Each variant was composed of approximately 8

to 12 restriction fragments between 2.0kb and 0.3kb. All of the variants

included a 1.4kb fragment, one to four fragments between 1.3kb and 1.1kb,

two or three fragments of variable intensity between 0.7kb and 0.5kb, and one

to four fragments smaller than 0.5kb (Figures 1, 2, & 3).


Mspl variant pattern correlates with distribution.

Variants with subsets of common fragments were consistently detected

in individuals from populations shown previously to have similar genetic

backgrounds (Hall & Muralidharan 1989; Hall & Smith 1991). To illustrate this

















9Ls I I I I I I 5

OLSt I I I I I I II
0L91 i I | I I I I I


6091 |II I I I I I c
oo



L I I I I I I MY-
809lo II I I I II I o



O9A1 I I I I I Il I
SI/ I 1 I II
"-
e909Vt 1 1 1 1 1 1

9091v I I I I I I
O9l AI I I I II -I


90 Vl i I i I I i o
LOSw I I I I I I *s


,o0,V 1 .. I .-0

coy ,, i

LOVII III I I I II
O IA1 I I I I I I II I I>-


ZOWIIII I I III Il

02 0
O0Pll II I I I I I I
LO I I I I I I I I I I I I E.
ZOZIN I II I I I II I
LOZV. c w

z0 I I II I I I I I I I T





41 A

.Uj0 0 u < i
(n-y -- 66 6 o ZS















m o

* E
coE

o

E a
oE
0C
C
0.E

.0
o)
0)




0a

CC





(D -0
0-





a 0
a- a
S(0

(U
0 M


E
c =


0 0






.
aE
IIM




0 0



S.0




CC


h-a4



CM




h..


<8
H 0
2 4
> a)

a >


C
S- c0




04-
4- co
O




da
c



o m
0)C 0











co
m2
ECa-
01




0 W r
0.- a)
0





C >
41C







0o ,
-$5









CE a




.0) CD4.'
ao 'I-o
4D M
5 ( >
w t;
C (
Tl1 } (U
Lcc



cn


a
4-d
0
4-J






0
0)







C

10













4-
0)

oP







n(





. c
a C
-a




















c
a)C
















0.E
0Co


t O IAI


eotlim
L01I V































LOZIJ
lOSIAI













OLIAI







LOLVII


ZOLIV






sais









12
o


.0 N0 .0 0
* co


3.



1:


pt
..-..


.z


"~F~ ':~9~
- i ------














cN
~I-
a,4;

U)






E,

CO

, C

ow




40.-






m -1
0 0






ca)
a,0
I.-~










0'
"a,
(D







C
a,C







-C




4-J
S00
a,








ODL.
LO~
cti;

(D


8L9V'J







L M9V



OLSLAJ



609VAJ



809vy



L09VY



909Vy



9091VN





4-'
8091VN


E

0,9v



at
z



0



0










































-4*


I


00
O r0
o oo


CY C


''

" ;:"; .i



..;i
,.*~


Ih' '









15

correlation, variants consisting of patterns with common fragments and similar

distributions have been grouped together. For each population sampled, a

summary of the sample size, the number of variants in each Mspl variant group,

and the total number of variants detected are given in Table 1.

Variants M101-M104 (Figure 2, first panel; see also Figure 1) were

characterized by a pair of fragments of approximately 1.2kb and 0.85kb, and

an intense signal at approximately 0.4kb. These variants differed from each

other primarily in additional fragments, between 0.7kb and 0.5kb. Variants

M103 and M104 were difficult to distinguish and will be referred to together

as M103. Variants M201 and M202 (Figure 2, second panel) differed from

M101 and M102 only in the fragments smaller than 0.5kb. Variants M201 and

M202 were distinguished from each other by the fragments 0.7kb to 0.5kb in

size. The M100 and M200 variants were found in drones from the USA,

Mexico, and Honduras.

The Mspl variants M301-M304 (Figure 2, third panel; see also Figure 1)

featured 1.8kb and 1.2kb fragments. Variants M301, M303, and M304

differed very slightly from each other in the size of the fragments at

approximately 1.4kb and 1.2kb and in the fragments between 0.7kb and 0.5kb

(fragments not clear in Figure 2 can be seen in Figure 1). Variant M302 was

distinguished by a 0.75kb fragment and by fragments smaller than 0.5kb. At

least one of the M300 variants was found in each population.













Table 1. Summary of populations examined for variants and alleles at locus 178.

Sample origin South Africa Honduras Mexico USA

Mspl variant analysis:
n: individuals 94 106 56 146
colonies 23 21 12 58
different variants: 19 18 11 8
M100 total" 0 1 2 3
M200 total 0 1 0 3
M300 total 2 1 2 2
M400 total 3 3 3 0
M500 total 14 12 4 0
variants unique to
this population 10 2 0 3

Ddel variant analysis:
n: individuals 124 145 75 164
colonies 24 26 13 59
different variants 19 17 15 8
D100 total" 0 0 2 2
D200 total 0 3 1 1
D300 total 1 4 4 3
D400 total 12 8 5 2
D500 total 6 2 3 0
variants unique to
this population 10 3 1 1

Allele identification:
n: individuals 93 106 56 146
colonies 23 21 12 58
allelesb detected 28 25 14 14
alleles unique to this
population 26 15 3 8


a Total number of different
sample population.


variants from each variant group detected in each


b Pairwise combinations of Mspl and Ddel variants detected with pB178 in
individual drones.









17
Variants M401-M405 (Figure 2, fourth panel) were characterized by the

presence of a 1.1kb fragment and three fragments smaller than 0.5kb, and by

the absence of the 0.85kb fragment common to nearly all other variants. The

1.1kb fragment of the M400 group appeared to be allelic to the 0.85kb

fragment (Figures 1 & 2). The M400 variants were distinguished from each

other on the basis of restriction fragments between 1.4kb and 1.1kb

(fragments smaller than 0.7kb, not clear in Figure 2, can be seen in Figure 1).

The M400 variants were found in drones from South Africa, Honduras, and

Mexico.

Mspl variants M500 as a group (Figures 1 & 3)) were distinguished by

the presence, absence, or intensity of one or more unique restriction fragments.

For example, variant M501 contained a unique 1.5kb fragment and lacked the

0.85kb fragment. M502 contained one, and M510 and M511 contained two,

restriction fragments at approximately 1.3kb; these three differed from each

other in the restriction fragments smaller than 0.7kb. M503 contained two

fragments near 1.4kb, similar to variant M403, but contained the 0.85kb

fragment that M403 lacked. The M500 variants were found in drones from

South Africa, Honduras, and Mexico.


Identification of Ddel variants.

All 33 Ddel restriction fragment variants identified are shown

diagrammatically in Figure 4. Each pattern consisted of six or seven restriction

fragments. Most variants exhibited a 0.75kb fragment and one to three



















LO0g

9090







q L t



z LaG
o LaC]

60ita
80tra
LOt7O
90tOa








90C0
LOWO







coza





LOea
oea








LOLa
Noz


El I
* II
II
* II
* II
I II
I II

* I
I II
*I
I I
I I II
I II
I II
II
II
II
II
II
II
lI
II


El I
ml
ElI


*I







Ci n CI
Ca, cJ eCN


L-o
(,^ ,- 0


I
I
I
SI
)I
II
II
II
II
II
II
I'
)I
I
I

I
II I
II
I
II
II


C


> 4

(D
..




Co
a
C, (
N N
V- 5)
o0
0
(0


C
0)












Eo
4-' a,




a)





0) a,
E-




0 >
0
w





0
C,





a,-o
. 1-
1^,:
C% "0
CD (F}









19
fragments between 1.25kb and 1.4kb. The majority of variants contained a

pair of fragments of variable size and intensity between 2.0kb and 2.2kb: five

variants contained a 2.0kb fragment, and another fragment 2.3kb or larger.

Many of the Ddel variants were distinguished by fragments that differed

slightly in intensity and/or molecular weight. Similar to the Mspl restriction

fragment patterns, Ddel variant groups were identified with subsets of common

fragments, and variants in each group were generally restricted in distribution

to populations with similar genetic backgrounds.


Description and distribution of groups of Ddel variants.

Sample sizes, numbers of variants in each group, and total number of

variants detected in each population tested are given in Table 1.

Variants D101 and D102 (Figures 4 & 5) were characterized by a pair of

2.1 kb fragments and by 1.1 kb and 1.0kb fragments. They differed from each

other in the sizes of fragments near 2.0kb and 1.35kb. These two variants

were detected only in North American drones.

Variants D201-D203 (Figures 4 & 5) were identified by the presence of

a 1.25kb fragment and the absence of the 1.0kb fragment detected in D101

and D102. Variants D201-D203 were distinguished from each other by the

size and intensity of the fragments from 2.1 kb to 2.0kb and 1.35kb to 1.25kb

(Figure 4). D201 was found in one USA colony and in feral colonies in Mexico

and Honduras; D202 and D203 were found in feral colonies in Honduras.



















CV) U)

c 9 Lr
CO 4-'

> 0 0)

5- C 4
cnM 0)



o 2 0"

~E

EZ U)i




0
~ O 4-' '4 o'







G ) (D
o, C .0Cl






Z-0
r U















Im 0


-o

a) 0


a)
cn 0 c


















E 0 (4C
cm 0 .S0 CD 0I
0 't E
0 cn C)




~Bmnm
.C.) cn

4 cn~
C> E~d
C,4 c z
mICY



N Z)
-0 C-4


ir 0 :2 0













.0.0 ., .m,0 .0
g-K e -le -V .
,-o0 0 O- o 0




LOWo :
o o 1 1 i








|oga jo
zoia

ioia II



m 0 Lo OD r%




Loea 2












C lZ 0 1 4
Jl 0 0 0
iS -x -
to~ o in oo 1^









22
Similar to the D200 variants, D301-D306 (Figures 4, 5 & 6) contained

a 1.25kb fragment, lacked the 1.Okb fragment detected in D101 and D102 but

contained a 0.75kb fragment that the D100 and D200 variants lacked (Figures

4, 5 & 6). D301-D304 were individually distinguished by the combinations of

fragments between 2.1kb and 2.0kb and between 1.4kb and 1.25kb. D305

was further distinguished by a unique fragment of approximately 1.8kb. D306

contained a 3.4kb fragment and lacked a fragment near 1.1kb found in the

other D300 variants. Variants D301-D305 were found in North America and

Honduras, while D306 was found only in South Africa.

The largest group, D400, consisted of fifteen variants characterized by

1.Okb and 0.75kb fragments (Figures 4 & 7). The D400 alleles differed from

each other in the size and intensity of fragments of approximately 1.4kb to

1.3kb and in the absence, or slight variation in the size, of a fragment near

1.1kb. The majority of the D400 alleles were restricted in distribution to the

South African and Honduran populations, but two D400 variants were also

found in Mexico.

The D500 variants were characterized by 0.95kb and 0.75kb fragments

(Figures 4 & 5). Each variant in this group differed in its combination of two

fragments between 2.1 kb and 2.0kb, two fragments between 1.4kb and 1.3kb,

and a single fragment of variable size near 1.1kb. These variants were found

in Mexico, Honduras, and South Africa.
















a,



CV)

r C
u LO ~ aj ,


m c") 0) 0c
L- >-


0.0
CO 0*
E m 0 o






>j 0 4
0 0 4-



ta C


3.
o <'
C Lo U)









0 00 CM*

0 oD C4 '4
0 *-0 0
0) E


0o0 0'4y







4-1 0 C0
N





'0 cm0L
Qa, 0C

S C N





MCD
o a,
C C LN

.C, 0
M 0 (n > 0

CD

0 )
CV-) c




C3) 4- M C
m CD N~
> v U)










.0 24
0 3 3 .0 .0
A: ^^ Y in -: -; :;U)
-t rO 00 CM *- 10

A U ^ 1 1 1




















Logoa n .r"
9090








9060.



.. 0 3 CDi;i;







...... ~ ~ ~ ~ :"'F; : i'. ,'!i















(1) 4
aa)
a)

CO


.C 0
4-d
ja)


c



E r

0 0.
0. 0

C 0
OW

M 4-'


U)
oa





0>
h..






0
r
M





CD,
aa






4- U)
m >
0
C1.,













*0





CD
o co


>0






00
I- I
0C)

Oc
OLD


-a)



LL-b










Ci r C. o -o







L. S1..11LAIN6
EL17G t







60170

80170d 9

LOV(33; 5

90V1370





Etrotl


zova: :

0170 1









27

Distribution of alleles at locus 178 in Old and New World drone honey bees.

For each population, the sample size, the total number of alleles

detected, and the number of unique alleles detected are given in Table 1. The

identification and distribution of the 65 different pairwise combinations of

variants, corresponding to alleles, is summarized in Table 2.

In USA drones, 14 alleles were found that were composed of Mspl

variants from the M100, M200, and M300 groups, and Ddel variants from the

D100, D200, D300, and D400 groups. Mspl variants M101, M201, and

M203, and Ddel variant D401, and a total of eight alleles were unique to USA

drones (Table 2, distribution 1).

In the South African drones, a total of 19 Mspl variants from the M300,

M400, and M500 groups were detected; 19 Ddel variants from the D300,

D400, and D500 groups were detected. Ten Mspl and ten Ddel variants were

unique to the South African samples. A total of 28 different alleles were

identified, 26 of which were not found in any other population (Table 2,

distribution 8).

In samples from a single colony in Brazil, near the site of the introduction

and release of the African bees, the Ddel variants identified were D405 and

D504, which also had been found in South African drones (not shown). The

Mspl variants in this colony, and thus the alleles, were not determined.

In drones from Honduras, a total of 18 Mspl variants, 17 Ddel variants,

and 25 alleles were identified. Alleles in these samples were composed of











Table 2. Distribution of alleles detected at locus 178 in Old and New World drones.

o00000000000000000000o~--- 0000000
N N OON OOOODOODOODOODDOCD0OODDO


8 6
8


M101
M102
M103
M201
M202
M203
M301
M302
M303
M304
M401
M402
M403
M404
M405
M501
M502
M503
M504
M505
M506
M507
M508
M509
M510
M511
M512
M513
M514
M515
M516
M517
M518
M519


68


6 4


6 5


Distribution Key:


USA
USA
USA


Mexico
Mexico
Mexico
Mexico


+ Honduras

+ Honduras
Honduras
Honduras


+ South Africa
South Africa


3663


22


5
8 8 4


88








29

variants from all five Mspl groups and from all Ddel variant groups except

D100. Two Mspl (M202, M301) and two Ddel (D301 and D201) variants were

common to the Honduran and USA samples: three Mspl (M401, M512, M514)

and five Ddel variants (D403-D405, D407, D414) were common to the

Honduran and South African drones. Unique to Honduran drones were Mspl

variants M503 and M518, and Ddel variants D202, D203, and D305. Three

alleles were common to the USA and Honduras drones (M103/D303,

M301/D201 and M301/D301; Table 2, distribution 3). Two alleles were

common to Honduran and South African drones (M512/D403, and

M514/D404; Table 2, distribution 7). Fifteen of the 25 alleles identified were

found only in Honduran drones (Table 2, distribution 6).

In bees from Mexico, 11 Mspl variants, 15 Ddel variants, and a total of

14 alleles were identified. Of the Mspl variant groups, only M200 was not

represented, while variants from all five Ddel variant groups were identified.

Four Mspl (M102, M103, M301, M302) and seven Ddel variants (D101, D102,

D201, D301-D303, D404) were common to drones in Mexico and the USA.

Three Mspl (M405, M511, M515) and five Ddel variants (D404, D405, D408,

D414, D501) were common to drones in South Africa and Mexico. Three

alleles were found in both the USA and Mexico (M102/D303, M302/D102, and

M301/D301; Table 2, distribution 2). There were no alleles common to South

African and Mexican drones. Alleles M405/D502, M510/D304, and









30

M515/D408 (Table 2, distribution 4), and Ddel variant D304 were found only

in this population.

Eight of the 65 alleles contained exclusive combinations of Mspl and

Ddel variants: i.e., the Mspl variant was not associated with any other Ddel

variant. Fewer exclusive combinations were found in USA/European bees than

in the neotropical and South African bees: only M101/D401 was found in the

USA; M404/D501 and M502/D410 were found in Honduras and southern

Mexico; M513/D504 was detected in Honduras; and M304/D503,

M504/D415, M505/D507, and M519/D402 were found in South Africa.




Discussion


Common variants, uncommon alleles.

Several Mspl and Ddel variants were components of numerous alleles

and were found in more than one population (Table 2). The combination of

variants comprising each allele was indicative of the ancestry of the sample.

For example, Ddel variant D405 was found in seven alleles in South Africa and

the neotropics. The majority of samples with D405 were from South Africa,

and the associated Mspl variants in these individuals (M400s and M500s) were

only found in Old and New World colonies previously shown to have African

mtDNA (Hall & Muralidharan 1989; Hall & Smith 1991). In North America and









31
Honduras, Ddel variant D303, found in three alleles, was associated with Mspl

variants M102, M103, and M201, indicating European origin.

The presence of Ddel variant D404 in the USA, Honduras, and South

Africa was an exception to the correlation of fragment pattern with geographic

origin. The alleles with D404 were differentiated by Mspl variants, which were

consistent with the sample origin and what was known of its ancestry: M302

(USA), M514 (Honduras and South Africa), and M515 (South Africa) (Table 2).

Variants in the M300 group were found in all populations examined,

although individually they appeared to be limited in distribution. Variant M302,

which contained a 0.75kb fragment not present in M301, M303, and M304,

was found only in the USA and Mexico, with Ddel variants D101, D102, and

D404. Variant M301 was a component of two different alleles in the USA,

with Ddel variants D201 and D301. Variants M303 and M304, very similar to

M301 but found only in South Africa, were associated with Ddel variants D405

and D503 (Figures 1 and 2).

In neotropical samples, because of the small difference in fragment sizes

that distinguished M301, M303, and M304, the African or European origin of

the alleles containing the Mspl-1.8kb fragment was confirmed by the Ddel

variant with which they were associated. As in the USA, M301 was found

with D201 and D301 in drones in Mexico and Honduras, indicating that the

samples were European. In addition, variants D202, D203, and D305, found

only in drones from Honduras, had fragments common to variants found in USA









32
drones. Given that the fragment patterns of these Ddel variants were

consistent with other patterns detected in the USA, these three variants were

assumed to indicate European ancestry. Neotropical drones in which the Mspl-

1.8kb fragment was detected were concluded to be of European ancestry on

the basis of the Ddel variants (Table 2).


Greater diversity at locus 178 in tropical vs. temperate bees.

USA bees demonstrated the least variability, having the smallest number

of variants and alleles found, of the four populations analyzed. The greatest

diversity at locus 178 was found in the South African drones: twice as many

Mspl and Ddel variants comprising twice as many alleles were found in half as

many colonies, compared to the USA.

The neotropical samples were also more varied than the USA samples

and were collected from fewer colonies. The Honduran samples came from

colonies established from feral swarms several years after African bees had

arrived. The detection of both European and African variants and alleles was

expected. However, there was little indication of European ancestry in the

Honduran bees; the majority of alleles common to the USA and Honduras

contained the Mspl variant M301, and only two other alleles contained

USA/European-type variants (Table 2, distribution 3). More variants were

common to Honduran and South African drones. Like the South African

drones, the majority of the Mspl variants in the Honduran drones belonged to

the M400 and M500 group, and most of the Ddel variants detected were in the









33
D400 group. This finding was consistent with the feral origin of the colonies,

in which African mtDNA was found (Hall & Smith 1991).

Some of the drones from southern Mexico were collected from managed

colonies established from feral swarms, while the majority were collected from

feral swarms (Hall & Muralidharan 1989). It was expected that both

USA/European and South African markers would be found. The number of

variants common to Mexico and South Africa was far greater than those

common to the USA and Mexico, although there were no alleles common to the

Mexican and South African sample populations. The variants and alleles unique

to the Mexican samples resembled those in the South African samples and,

apparently, represented the recent establishment of the feral African population.

As in Honduras, most of the European-type variants and alleles common to the

USA and Mexican populations were in the M300 group (Table 2, distributions

2 and 3).

The greater number of variants and alleles identified in the neotropical

samples relative to the USA reflected the presence of African as well as

European variants. Variants unique to the neotropical drones, but similar to

those found in USA bees, probably reflect regional differences in European

races. Five alleles detected only in Honduras and Mexico (Table 2, distribution

5) were composed of variants that either were found, or were of the same

variant-type as those found, in the South African drones (Table 2, distribution









34
8). The detection of these alleles suggests that greater variation exists in the

parent South African population than has been revealed in this study.

Alleles were identified in Mexico (M510/D304) and Honduras

(M202/D414, and M510/D302) that were, according to the assignment of

variants to groups on the basis of fragment pattern and distribution, composed

of both European and African variants. The distributions determined in this

study may not be absolute, given the number of variants and alleles found, the

size of each sample, and the world-wide, human-assisted redistribution of bees.

Investigation with additional samples will likely be needed for a thorough

classification of variants at this highly polymorphic locus. However, alleles

composed of African and European variants may be the result of recombination

within the 178 locus.

Greater allele diversity has been detected in South African compared to

USA honey bees at another genomic locus (Hall 1992b). At least 14 variants

were detected in South African bees, two of which were detected in European

and USA bees. In Honduras nine variants were detected: the two common to

European and USA samples, as well as seven characteristic of South African

bees. A larger number of mtDNA size classes have been found in African bees

than in bees of European ancestry (Hall & Smith 1991; Smith et. al. 1991).

Greater diversity has been revealed in tropical relative to temperate populations

of Drosophila melanogaster (David & Capy 1988; Hale & Singh 1987),

Drosophila simulans (Hyytia et al. 1985), Ceratitis capitata (Gasperi et al.









35

1991), and Limulus polyphemus (Saunders, Kessler & Avise 1986). This

diversity has been attributed to the capacity of tropical populations to maintain

large population sizes over a long period of time (David & Capy 1988) and may

be consistent with the large reproductive capacity and population sizes of

tropical bees relative to temperate bees (Winston, Taylor & Otis 1983).














ALLELE FREQUENCIES AT LOCUS 178 REVEAL HYBRIDIZATION
OF EUROPEAN AND AFRICAN BEES IN THE NEOTROPICS


Introduction


Frequencies are reported here for the individual variants, for the groups

of variants with common fragments and distributions, and for the alleles found

at locus 178 in Old and New World populations of honey bees. Unlike previous

nuclear DNA markers (Hall 1990, 1992b), variants and alleles detected at locus

178 appeared to be specific to east European, west European, or African bees.

Ancestry in New World bees was inferred from the frequencies for the variants,

variant groups, and alleles. The results reveal a level and specificity of African-

European hybridization not observable in previous DNA studies.



Materials and Methods


Procedures for the isolation and electrophoretic analysis of honey bee

genomic DNA, identification of the source and size of probe pB178 used for the

detection of RFLPs, and the details of the initial detection of polymorphisms,

were given in the proceeding chapter. RFLPs were initially detected in pooled

sibling samples of worker bees. For determining the frequencies for the Mspl









37
and Ddel variants detected with pB178, individual Old and New World

European and African bees were analyzed.

Sources of honey bees. Frequencies of variants and alleles in drones

were determined for the samples used for the variant and allele identification.

Worker larvae and pupae were collected from these same colonies in South

Africa, Honduras, and southern Mexico.

Adults workers from Venezuela, collected between 1986 and 1988,

were provided by R. Hellmich Jr., J. Villa, A. Collins, and T. Rinderer, USDA-

ARS, Baton Rouge. Costa Rican samples were obtained in May 1989 from

apiaries at Cerro de la Muerte and San Isidro del General, by HGH with the help

of H. Arce and R. Dormond, National University, Heredia, Costa Rica. Five

colonies, maintained at 2200m to test resistance to African introgression at a

higher elevation had European mtDNA. Three colonies maintained at 700m had

African mtDNA. Samples from southern Mexico, near Tapachula, were

obtained in January 1988 by HGH from two managed apiaries and from feral

swarms captured in bait hives maintained by the Mexican agency Secretariat

of Agriculture and Hydrologic Resources (SARH). Feral worker samples from

northern Mexico, collected prior to the arrival of African bees, were provided

by W. Rubink and A. Collins, USDA-ARS, Weslaco. Some of the USA drones,

and all of the USA workers, were from a closed breeding colony in Arizona,

provided by J. Martin, G. Waller, G. Loper, and E. Erickson (Page, Erickson &

Laidlaw 1982; Severson, Page & Laidlaw 1986). Sources for additional USA









38
drones were reported in the previous chapter. Workers from Europe (larvae and

pupae) were provided by B. Vaissiere, Texas A&M, and J.-M. Cornuet, INRA

Monfavet, France. These included samples of west European A. m. mellifera

(8 colonies), and east European A. m. ligustica (3 colonies) and A. m.

caucasica (2 colonies).

For determining the frequencies of the Mspl variants, a total of 862 bees

were analyzed (402 drones, 460 workers), representing 128 Old and New

World colonies. For determining the Ddel variant frequencies, the 402 drones

analyzed for the Mspl variant frequencies and an additional 106 drones were

examined, from a total of 122 colonies. Allele frequencies were determined for

401 drones from 114 colonies, for which both the Mspl and Ddel variants had

been identified.

Frequencies of variants and alleles. In haploid drones, frequencies for

Mspl and Ddel variants and alleles were calculated as a fraction of the total

number of variants or alleles detected. The numbers of variants and alleles

detected were determined as described by Hall (1992b), using the data for

drones. At least two drones were analyzed per colony. When fewer than six

drones were analyzed per colony and only one variant or allele was found, the

variant or allele was counted once. When more than six drones were analyzed

and one variant or allele was detected, it was counted twice (greater than 95%

probability that the queen was homozygous). When two variants or alleles









39
were found heterozygouss queen), each was counted once. Frequencies for

groups of variants represent the sum of the individual variants within a group.

Drones from the same colony represent only the variants and alleles of

a single individual, the queen. Frequencies for variants identified from the

analysis of worker restriction fragment patterns reflect more of the variation

within a local population, due to the presence of multiple patrilines within

colonies.

In diploid worker bees (females), RFLPs are codominantly expressed.

When two restriction fragment variants contain common fragments, fragment

superimposition results as a consequence of comigration. Thus for workers,

frequencies were determined for groups of Mspl variants, since more than one

variant from the same group could account for the fragment patterns seen

(Figure 8), and since the variant group was correlated to the ancestry of the

sample (see previous chapter). If the identity of the variant group to which a

particular fragment belonged was uncertain, the frequency for the group was

reported as a range (Table 4).

In Ddel digests of the majority of worker samples, fragment comigration

precluded the identification of variants (Figure 11) and the determination of the

frequencies of both individual variants and of variant groups. Consequently,

frequencies of Ddel variants, and therefore of alleles, were determined only for

those populations for which drones were available (South Africa, Honduras,

southern Mexico, and the USA; Tables 5 & 6).










Results


Frequencies determined in drones for the individual Mspl restriction

fragment patterns (variants), and totals for the five groups of Mspl variants

with similar fragment patterns and distributions are presented in Table 3. The

fragments characteristic of each group of Mspl variants are shown in Figure 8.

Frequencies for groups of Mspl variants detected in workers are

summarized in Table 4. Examples of Mspl restriction fragment patterns are

shown in Figure 9. The pair of Mspl variant groups detected in a worker will

be referred to as the Mspl genotype. Three worker restriction fragment

patterns shown in Figure 10 demonstrate how the comigration of fragments

confounded the identification of individual Mspl variants in diploids.

Frequencies for Ddel variants detected in drones are summarized in Table

5, and the fragments characteristic of each of the five Ddel variant groups are

shown in Figure 11. Ddel fragment patterns were examined in selected

workers for evidence of fragments characteristic of particular variant groups

(not shown).

The distribution and frequencies for the alleles, the pairwise combinations

of the Mspl and Ddel variants, detected in drones are summarized in Table 6.


European bees.

A limited number of workers were available from European colonies of

known ancestry. Thus, the variants represent a subset of the variability in










Table 3. Frequency [%] of detection of Mspl variants at locus 178 in drones.
Variant South Africa Honduras Mexico USA
M101 3
M102 5 13
M103 3 19 51
M201 11
M202 3 3
M203 1
M301 26 24 9
M302 10 9
M303 3
M304 3
M401 3 3
M402 5
M403 8 10
M404 8 5
M405 5 5
M501 3
M502 5 5
M503 3
M504 3
M505 3
M506 5
M507 5
M508 27 3
M509 3 3
M510 5 5
M511 3 8 5
M512 11 10
M513 3 5
M514 5 5
M515 3 3 5
M516" 5 3
M517" 3
M518" 3
M519" 3
M100 total" 0 3 24 67
M200 total 0 3 0 14
M300 total 5 26 33 18
M400 total 14 18 19 0
M500 total 81 51 24 0
n: individuals 94 106 56 146
colonies 23 21 12 58
variants detected 37/44 39/42 21/24 76/112
ns Not shown in a figure.
a Sum of the frequencies of all variants detected in the respective variant groups.
b Ratio of the number of variants counted, to the total number of variants
possible = I no. drones/colony].














0 0 0 0 0
0 0 0 0 0
Size, kb 2E 2 2
3.4kb


2.3kb

2.0kb

1.8kb


1.4kb



1.0kb


0.85kb

0.75kb


0.6kb


0.4kb


Figure 8. Restriction fragment patterns of Mspl variants
detected at locus 178. Restriction fragments detected in
all variants within a group and characteristic of each
group are indicated by (-), and variable fragments
detected in at least one but not all variants within each
group of Mspl variants are designated by (0).


























.c
(1)

0.















*
0



.0
o ,)



2 X
*
0
0)


















o 0
Z o








*




























c
-o
0









>-c
0,

w





a,







C





c0
cos


cO 0


U) O ..
CV)
CM


'* U) 0) qt co
N I t




CV) )n C1
0)






S0 0 0
0 0 0 0 0
N- CC ) 0o U


Co











a) r- *-
M (

















o r- C -
- C C)





C) 0) C) r- 0M
CM CN C)



0 CO
6 CM CN UO
LU


C CO CM
C)











0co c







'- M r-


C-DCD CM
COC CM4






-C(D CM

mC CV


c- CM4





N N
CM0 0

















'a

a (D

0o -
3:yt


I-
C3

N
.1
0
E
E


m




o





.-
0





a
(0

















E




0
0
0
C








E


0
a,










4-
'
't



0
u,
03

0









au

0


u










*










Oc

- 0 c

.5 -. .9 .
4o-0



0
c o









0 0
TU) C
0 C D0
CL m
c .c S 0
o .2 D.0 U












V~CO
.-2 oc e
0 i m



4..J c 0


v m V $0 0
Co




a C U)
-0 V C0

.. 0 o




o C > r-
4 0) I- 0 ?



E E 45

o c








CLU- 0) 0C
0 "a C 0C
U) = 0
4-J 6
E 'D +o
0EOC
Q C 0O


009I N/009 VW











001V1AI/001AI




OOVlAI/OOZIAI
001lAJ/OOZIAJ


009N1J/0OLV4 :6








00LAIN00ILS :9
OOSIV/OOSIAI :9


00E/OOZLV :9


00oV/00L oo :v


OOZII/OOZW :s





00 LI/00 L IN :'t









00 MN LO 1
ge a- .






At



4- rrA If
*




8*- 115*


4 1












OO v% -O O vr ^-OO O
c 00 0 00 00 00
(kb) _2 M M __ __ 2 M


2.0kb

1.8kb



1.4kb

1.2kb
1.1kb

1.0kb


0.85kb

0.75kb


0.6kb






0.4kb


l


1--
m


mm



Smm
-


)--


Figure 10. Mspl patterns from three workers, demonstrating possible variant pairs.
#1: an example of an unequivocal identification of the variants which composed the
pattern of fragments detected. #2 and #3: examples demonstrating the difficulty
encountered in identifying variants in workers from neotropical areas where African
bees were established. Fragments in these patterns could be assigned to variants in
more than one variant group.


. .












Table 5. Frequency [%] of detection of Ddel variants at locus 178 in drones.
Variant South Africa Honduras So. Mexico USA
D101 4 1
D102 4 6
D201 2 4 7
D202 2
D203 2
D301 13 17 1
D302 4 4 28
D303 2 25 53
D304 4
D305 2
D306 2
D401 2
D402 2
D403 13 19
D404 2 4 1
D405 22 6-8 8
D406 2 4
D407 11 8-11
D408 9 4
D409 2 4
D410 4 4
D411 2
D412 2
D413 2
D414 11 13 4
D415 2
D501 2 6 4
D502 2 4
D503 2
D504 2 4
D505 4-7
D506 0-2
D507 2
D100 total" 0 0 8 7
D200 total 0 6 4 7
D300 total 2 21 44 82
D400 total 82 64 25 4
D500 total 16 8 12 0
n: individuals 124 145 75 164
colonies 24 26 13 59
variants detected 45/48 47/52 24/26 83/118
a Total number of different variants from each group detected in each population.
b Number of variants counted/total number of variants possible (two per queen
or colony).















0 0 0 0 0
0 0 0 0 0
Size, kb 0 a a a a
3.4kb n n


2.3kb

2.0kb

1.8kb


1.4kb



1.0kb


0.85kb

0.75kb


0.6kb




0.4kb


Figure 11. Restriction fragment patterns of groups of
Ddel variants detected at the locus corresponding to
probe pB178 in the honey bee genome. Restriction
fragments detected in all variants within a group and
characteristic of each group are indicated by (-), and
variable fragments detected in at least one but not all
variants within each group of Ddel variants are designated
by (0).

































L00 00








































00000

000- 0
- C0 CO CV C
00000


2Ssii


0)'-CY)

























Cv)


00 -0
C) CV I 0
0000
- 04 N. M
0000

2222C


00


00

22


miO c) Co) C)


- C- N M-CMCOMl' lC
00000000
00000000
NCO 2- VMC' t
00000000
~IC ') ')C' C) ')C


OIL)
V-











cCr) 0000


00000--0
00000000

00000000
lor.rio^'-t CM


i i-


CV C) CV CY













') C0 (0C')
0-0-00


000000
W I ) iL) W t 10 ic)
2 2 2 2 2 2IL


C)m' IC ) I )


- I s r% 00



































1O 1O un


Cr)LD 10cv)O(w) C~c)CIC'


CMMCV)C-)00 MC4
L6 6


Cm 00 C00m


cm


0-000000-0000000-00000000-


rrOOc oicnaoo') t MMo ttf imDoy
000000000m m- m- ,- m- -

22222222222222222222222222


co






r%J



C~lt-C4L
o 1















0))C




o),tj.
Ct V) 04







CY) 0







-I)
4-JO
C =

V-U) a
> 0 S
0)
16c75-5


Ui
0)
C
p
"0
L..




>
:6 C.


CQ)
c0c



000' a

0 a0rw
o o0


~oE
03.. 0



a)0 CL)
CO
Z E 0





)0.
(DC
aQ)>






wc Ea
*13






00)








C (D a

0'.- m
oOOg








0 0
4u-'
COa)












E.
a. z


S C


00'-0
0000
7- -.
(0 O O)O
102202









51
these east and west European populations. There was little variation in the

restriction fragment patterns within each European sample population.

Examples of Mspl fragment patterns detected in A. m. ligustica workers

are shown in Figure 9, lanes 1-3. The fragments detected corresponded to

M100 and M200 variants, with estimated group frequencies of 81% and 19%,

respectively (Table 4). The majority of A. m. ligustica workers appeared to

have two M100 variants (Figure 9, lane 1), while the A. m. caucasica workers

examined all appeared to have two M200 variants (Figure 9, lane 3). The Ddel

fragment patterns for the A. m. ligustica and A. m. caucasica workers

examined (not shown) contained 1.25kb and 0.75kb fragments characteristic

of variants in the D300 group (Figure 11). It is possible that D200 variants

were present but could not be detected due to comigration of D300 fragments

(see Figures 4 & 11).

Examples of Mspl fragment patterns detected in A. m. mellifera workers

are shown in Figure 9, lanes 4-7. All the A. m. mellifera workers examined

appeared to have at least one M300 variant. The frequency for the M300

variant group was 94% (Table 4). The majority of the workers appeared to

have two M300 variants (Figure 9, lanes 6 & 7). The remaining A. m. mellifera

workers appeared to have an M300 variant together with an M100 or M200

variant (Figure 9, lanes 4 & 5, respectively). All Ddel fragment patterns

detected in the A. m. mellifera workers contained a 1.25kb fragment

characteristic of the D200 and D300 groups (not shown). The majority









52

(-60%) of the patterns in the A. m. mellifera samples contained 1.25kb and

0.75kb fragments (the 1.25kb fragment was not detected in South African

drones), and lacked a fragment at approximately 1.0kb detected in the majority

of South African drones (Figures 4, 6, & 7). The A. m. mellifera Ddel patterns

lacking the 0.75kb fragment (- 25%) appeared to be composed of one or two

D200 variants. In a few (- 5%) workers the 1.25kb and 1.Okb fragments were

present, but the 0.75kb fragment was absent, indicative of D100 and D200

variants, which were not detected in South African drones. A combination of

D100, D200, and D300 variants was indicated in the workers (10%) in which

the 1.25kb, 1.0kb, and 0.75kb fragments were detected. No evidence was

found in the A. m. mellifera workers for Ddel patterns detected in South African

drones (Figures 4 & 11).


African bees.

In South African drones, the Mspl variant group with the highest

frequency (81%) was the M500 group (Table 3). Variants M508 and M512

had the highest individual frequencies (25% and 11%, respectively). Eleven

other M500 variants were found at lower frequencies, each in one or two

drones. The M400 group frequency was 14%: variants M401, M402, and

M405, were each detected at a low frequency. The frequency for the M300

group was 6%: variants M303 and M304 were each detected in a single drone

from colonies in different locations in the Transvaal (Table 3).









53
Examples of Mspl fragment patterns detected in South African workers

are shown in Figure 9, lanes 12, 14, 16, and 17. The estimated group

frequency for the M500 variants in South African workers was 92% (Table 4):

the majority of these workers appeared to carry only M500 variants (Figure 9,

lane 17). The M400 and M300 group frequencies were 5% and 3%,

respectively. Since no M200 variants were detected in South African drones,

it was assumed that M500 variants predominated in workers as well (Table 4:

M200 variants would be difficult to detect in A. m. scutellata workers due to

the comigration of fragments).

In South African drones, the majority of the Ddel variants detected were

in the D400 group (Table 5). Variants D405 and D403 were found at the

highest individual frequencies (22% and 13%, respectively). Only one D300

variant was detected (D306). The remaining variants detected were from the

D500 group. The ambiguity in the frequencies for D505 and D506 was due to

the similarity of these patterns, which were difficult to distinguish if the

samples were not adjacent on a blot.

Frequencies were low for the many alleles found in South African drones

(Table 6). Alleles M508/D408 and M512/D403 were detected at the highest

frequencies (11% and 8%, respectively): two-thirds of the alleles were each

detected in a single drone (equivalent to one allele in the queen from a single

colony, or one in forty-six).










North American bees.

United States. In USA drones, the variant group with the highest

frequency was M100 (67%), and the Mspl variant detected at the highest

frequency was M103 (51%; Table 3). Variant M201 was detected at the

highest frequency (11%) in the M200 group. Variants M301 and M302 were

each detected at a frequency of 9%.

In USA workers, Mspl fragment patterns characteristic of the M100,

M200, and M300 variants were detected at group frequencies of 80%, 5%,

and 14%, respectively (Table 4). The majority of workers appeared to have

two M100 variants (Figure 9, lane 1). All other Mspl fragment patterns in USA

workers corresponded to heterozygous genotypes (Figure 9, lanes 2, 4, and 5,

respectively).

In drones from the USA, Ddel variants D302 and D303, and the D300

group, were detected at the highest individual and group frequencies,

respectively (Table 5). Frequencies were low for all the other variants detected.

Allele M103/D303 was detected at the highest frequency (39%) in

drones from the USA, the highest frequency of all alleles in all Old and New

World samples examined. Frequencies for the other alleles were relatively low

(Table 6). Half as many alleles, in twice the number of colonies, were detected

in USA drones compared to South African drones.

Northern Mexico. Mspl fragment patterns detected in feral workers from

northern Mexico are shown in Figure 9, in lanes 2, 5, 6, and 7. The variants









55
detected were from the M100, M200, and M300 groups, at group frequencies

of 9%, 27%, and 64%, respectively. Only in the A. m. mellifera workers from

France was the M300 group frequency higher.


Neotropical bees.

Southern Mexico. Variants from four of the five Mspl variant groups

were detected in drones from southern Mexico, collected after African bees

were established. The group with the highest frequency was M300 (33%).

The variant with the highest individual frequency was M301 (24%). As in USA

drones, M302 was the only other M300 variant detected. Frequencies for the

M100, M400, and M500 variant groups were comparable to each other (Table

3). Of the M100 variants, M103 was detected at the highest frequency

(19%), as in the USA drones. Frequencies were low for the individual M400

and M500 variants detected (Table 3).

Examples of Mspl fragment patterns detected in workers from southern

Mexico are shown in Figure 9, in lane 2 and in lanes 4-17. In these samples,

fragments characteristic of variants from all five Mspl variant groups were

found (Table 4). Frequencies for the M300, M400, and M500 variant groups

were 32%, 21%, and 30%, respectively. The M100 and M200 variant groups

were detected at the lowest frequencies.

Variants from all five Ddel groups were detected in drones from southern

Mexico. The Ddel variants detected at the highest frequencies were D303 and

D301. Frequencies were low for all other variants detected (Table 5).









56
Allele frequencies in drones from southern Mexico were, in general,

intermediate to the frequencies in the USA and South African drones. The

notable exception was allele M301/D301 (19%), which was absent in South

African drones and detected in a single colony from the USA sample collection

(a colony from Florida).

Honduras. Variants from each Mspl group were detected in drones from

Honduras (Table 3). The majority of variants detected were in the M500 group

(Table 3). With the exception of variant M512, detected at a frequency of 11-

14%, the individual M500 variants were found at low frequencies. [An

ambiguity in the frequencies of variants M512 and M516 resulted from the

similarity in the sizes and intensities of fragments in these two variants: unless

the samples bearing these variants were adjacent in Southern blots, they were

difficult to distinguish.] Frequencies of variants in the M400 group (M401, 3%;

M403, 8%; M404, 8%) were comparable to the frequencies at which these

variants had been detected in South Africa and southern Mexico (Table 3). As

in southern Mexico and the USA, only M301 and M302 were detected from the

M300 group. Variants M103 and M202 were detected in single drones.

It appeared that all five Mspl variant groups were represented in workers

from Honduras (Figure 9, lanes 4, 6, 8, 11, 12, 14-17; see Table 4). As in

drones, the variant group detected at the highest frequency was M500 (39%).

The frequency for the M400 group, 27%, was the highest of all the New World

populations tested. The M300 group frequencies (26% in drones, 23% in









57

workers) were comparable to other neotropical colonies examined. The M100

and M200 groups were detected at the lowest frequencies (3% and 9%,

respectively; Table 4).

In Honduras, as in South Africa, the Ddel variant group with the highest

frequency (64%) in drones was D400 (Table 5). D403, D301, and D414, were

detected at the highest individual frequencies (19%, 13%, and 13%,

respectively). Frequencies were low for all other variants detected (Table 5).

Individual frequencies were low for the many alleles detected in drones

from Honduras, as in the South African drones in which similar alleles were

detected (Table 6). The allele detected at the highest frequency in Honduras

(15%) was M301/D301, the allele detected at the highest frequency in Mexico.

Costa Rica. Examples of the Mspl restriction fragment patterns detected

in the Costa Rican colonies maintained at 2200m are shown in Figure 9, lanes

1, 2, 4-9, and 12 (possibly lane 14 as well). In these colonies (Table 4, Costa

Rica, 'European'), the group of Mspl variants with the highest frequency was

M100 (38%), followed by M300 (35%). The detection of M400 and M500

variants indicated that African paternal introgression had occurred in these

colonies. The M500 group frequency may have been as low as 2%, but

fragment comigration confounded the differentiation of the M200 and M500

variants and the determination of their frequencies.

Examples of the restriction fragment patterns detected in the three Costa

Rican colonies maintained at 700m are shown in Figure 9, lanes 12, 14, 16,









58

and 17 (possibly lane 9 as well). In these colonies (Table 4, Costa Rica

'African') the Mspl variant group M500 was detected at the highest frequency,

54-58%. The M400 and M300 group frequencies were similar to each other

(Table 4).

Venezuela. Variants from all five Mspl groups were detected in workers

collected in Venezuela (Figure 9, lanes 4-6, 8, 9, 12, 14-17, and possibly 1, 3,

and 11). As in the South African samples, the variant group detected at the

highest frequency was M500, followed by the M300 and M400 groups. M100

and M200 variants were detected at the lowest frequencies (Table 4). The

M400/M500 genotype (Figure 9, lane 16) was detected most frequently,

followed by M500/M500 and M300/M500 genotypes; the other genotypes

were detected at much lower frequencies.




Discussion


Old World bees.

In the east European workers examined, Mspl variants from the M100

group appeared to be specific to the Italian subspecies A. m. ligustica, while in

A. m. causcasica, the Mspl restriction fragment patterns consisted only of

variants belonging to the M200 group. Mspl variants from the M300 group

were exclusive to the west European black or German bee, A. m. mellifera.

The Ddel restriction fragment patterns in A. m. ligustica and A. m. caucasica









59

appeared to consist only of variants from the D300 group. A. m. mellifera

samples examined contained Ddel variants from the D100, D200, and D300

groups. These distributions are not absolute, but accurately reflect current

sample availability.

A detailed accounting of the variation in A. m. scutellata was obtained

from the identification of alleles composed of Mspl and Ddel variants in drones

from South Africa, collected in the area from which the New World African

bees originated.


North American bees.

The genetic variation at locus 178 in New World European bees was

ascertained by examining drone bees in the USA and feral workers from

northern Mexico. The Mspl variant groups were the same as those found in

Old World European bees. The frequencies for the M100 restriction fragment

patterns in A. m. ligustica workers from Europe and workers from the USA

were nearly identical. In USA populations, the high frequency for the M100

variant group reflects the preferred use of A. m. ligustica for beekeeping.

These results concur with the frequencies in USA bees reported for east

European-specific nuclear and mitochondrial DNA markers (Hall 1986, 1990;

Hall & Smith 1991; Schiff & Sheppard 1993), as well as for malate

dehydrogenase (MDH) allozymes (Nunamaker, Wilson & Haley 1984; Sheppard

1988). The low frequency (9%) for the M100 variant group in bees collected









60
in northern Mexico may be a consequence of small sample size, but more likely

reflects the small east European contribution to the feral population in this area.

Frequencies for the Mspl variant group M300 in the USA (18%) and in

northern Mexico, prior to the arrival of African bees (64% the highest of all

the New World populations examined), are consistent with the history of the

importation of west European bees in North America, their limited commercial

use, and persistence as feral colonies. The presence of the west European

variant M301 is also in agreement with MDH allozyme frequencies in USA bees

(which indicate A. m. mellifera nuclear genes persist in feral colonies; Sheppard

1988, 1989). West European mitochondrial DNA was detected at frequencies

of 7% in USA managed colonies and 64% in feral colonies in northern Mexico

prior to the arrival of African bees (Hall & Smith 1991) and was detected in

21% of 422 feral colonies in the south-central and southeastern USA (Schiff

& Sheppard 1993). Thus, bees of west European ancestry continue to

contribute to the honey bee gene pool in feral colonies of North America.

Ddel variant group D300 was detected at the highest frequency in USA

bees, consistent with the detection of variants in this group in each of the three

European races examined, and at the highest frequency in A. m. ligustica.

Ddel variants D401 and D404 were found in two separate USA colonies; D401

was found in a feral colony near Tucson, Arizona, and D404 was found in a

managed colony in Kansas. This finding was noteworthy because the D400

variant group was found at high frequency in South African bees. D401 was









61
not found in the South African colonies, and D404 was detected in a single

South African drone. One explanation for the detection of these variants in the

USA is that the distribution of the Ddel variants may not be absolutely

discontinuous and correlated with fragment pattern. Alternatively, these

variants may indicate African ancestry at low frequency in USA bees. Evidence

for the persistence of A. m. lamarkii in the USA, which was introduced (and

soon abandoned) for beekeeping in 1866 (Schiff & Sheppard 1993; Sheppard

1989) was recently shown by the detection of non-scutellata African mtDNA

at a frequency of 1% in feral bees in the south-central and southeastern USA

(Schiff & Sheppard 1993). Variants D401 and D404 may reflect A. m. lamarkii

ancestry, but this possibility needs to be confirmed with samples of this race.

On the basis of the congruity in the Mspl and Ddel variants detected in

Europe and in North America, the alleles identified in drones in the USA

represented a subset of the variability in European bees. The allele detected at

the highest frequency (39%) in USA drones, M103/D303, may be specific for

A. m. ligustica ancestry. In USA bees, only variants from the D300 group were

detected with M100 and M200 variants, consistent with the restriction

fragment patterns in the east European samples. Mspl variants M301 and

M302 in USA bees were detected with Ddel variant D301 and variants in the

D100 and D200 groups, evidence of which was also seen in the A. m. mellifera

workers in which only M300 variants were present. Alleles M101/D401 and









62

M302/D404 were present at very low frequency; these Ddel variants, but not

these alleles, were also found in South African bees.


Neotropical bees.

Italian and German farmers brought A. m. ligustica and A. m. mellifera,

respectively, to their settlements in south and southeastern Brazil (Gongalves

1974; GonCalves, Stort & DeJong 1991; Lobo, Del Lama & Mestiner 1989) and

Argentina (Kerr, De Leon & Dardo 1982). A. m. caucasica and A. m. carnica

were also introduced (Kent 1988; Ruttner 1986). In contrast to the USA, the

contribution of A. m. ligustica to the total gene pool in the neotropics was

minor in spite of considerable importation (Kent 1988). The Italian bee has

been found concentrated in certain areas, particularly Argentina (Dietz, Krell &

Eischen 1985; Kerr, De Leon & Dardo 1982; Sheppard et al. 1991), Costa Rica

(Hall 1990; Kent 1988; Spivak 1991), and the Yucatan (Kent 1988; Rinderer

et al. 1991). There is little indication that significant introductions of east

European bees were made or were successful in central and northeastern Brazil

(Lobo, Del Lama & Mestriner 1989), the Guianas, Suriname (Taylor 1977), or

Panama (Boreham & Roubik 1987; Roubik 1982). Prior to the release of A. m.

scutellata in South America, A. m. mellifera remained dominant in the

neotropics, and it has been assumed that if early introductions of subspecies

originating outside Europe were made, they were unsuccessful (Kent 1988).

Frequencies for the variants and variant groups identified in east

European workers and in the USA were very low, or absent, in neotropical bees









63
collected in areas where African bees were established. The east European

variants were largely in managed colonies. In southern Mexico, the frequency

for the M100 group of variants was 24% in drones, and 10% in workers.

These frequencies are consistent with previously identified nuclear DNA RFLP

markers specific to east European bees (Hall 1986, 1990), and may have

reflected more recent colonization by African bees compared to Honduras and

Venezuela. Some of these managed colonies were known to be of east

European ancestry, based on mtDNA (Hall & Muralidharan 1989; Hall & Smith

1991; Smith, Taylor & Brown 1989).

The detection of the Mspl variant M301 in colonies established primarily

from feral swarms provided evidence for the persistence of A. m. mellifera in

the neotropics, as has been suggested by allozyme frequencies (Lobo, Del Lama

& Mestriner 1989). The M300 group frequency ranged from 19-33%,

including areas which have been occupied by African bees for over 20 years.

In workers from Honduras, the detection of an Mspl variant bearing the

1.8kb fragment could not be identified outright as M301 and thus attributed to

A. m. mellifera ancestry in the apiaries and feral populations from which the

samples were obtained, since variants M303 and M304, found in South African

drones at much lower frequencies, also contained this fragment (Figure 8;

Tables 3, 4 & 6). The possibility of African origin for the 1.8kb Mspl fragment

had to be considered, because other Mspl variants had been detected at higher

frequencies in Honduras than in South Africa (e.g., M400 variants). It was also









64
possible that a variant bearing the 1.8kb Mspl fragment detected in South

African bees may have originated from prior importation of bees to South Africa

from Europe or by way of North America (Fletcher 1973, 1978).

The west European origin of the 1.8kb fragment-bearing Mspl variants

(M300 group) in southern Mexico and Honduras was confirmed by the

identification of the associated Ddel variants at locus 178 (Table 6, Figure 11).

In South African drones the Mspl M303 and M304 variants were associated

with the Ddel variants D405 and D503, respectively. In neotropical drones,

M301 was identified as the Mspl variant with the 1.8kb fragment and was

associated with Ddel variants that were the same as (D101, D102, D201,

D301) or similar to (D202, D203, D305) Ddel variants identified in USA drones

and inferred from Ddel fragments detected in A. m. mellifera workers. Among

the M300 group variants, the detection of only M301 and M302 in Honduras

demonstrates that A. m. mellifera ancestry persists in feral colonies. Ddel

D200 variants detected with M301 in Honduras may indicate A. m. mellifera

ancestry, but fragment comigration prohibited the detection of these Ddel

variants in A. m. mellifera workers. There was no evidence for Ddel variants

D405 or D503 in A. m. mellifera or in the M300-containing samples in the

neotropics. The absence of variants M303 and M304 in the New World may

be due to sampling error; these variants may not have been present in the A.

m. scutellata queens imported, or were present at a low frequency not detected

in this study.














LOCALIZATION OF VARIATION AT LOCUS 178 IN Apis mellifera (L.)
BY RESTRICTION MAPPING



Introduction


African and European groups of honey bee subspecies have been shown

to differ at locus 178 in the number and location of the four- and five-

nucleotide base recognition sequences of the restriction enzymes Mspl and

Ddel, respectively. While the ultimate resolution of the variation at locus 178

occurs at the level of the DNA sequence, RFLP analysis has permitted an

assessment of the variation using a small portion of its nucleotides.

The five groups of variants identified at 178 for both Mspl and Ddel,

which have similar restriction fragments and geographic distributions, may

share a subset of restriction sites. The number of fragments present in all

variants within a group ranged from one in the M500 group to ten in the M200

group (Figure 8). Between any two groups of Mspl or Ddel variants, the

number of common fragments ranged from one to seven. One fragment was

common to all Mspl variants (Figure 8), and no fragments appeared to be

common to all Ddel variants (Figure 11). It is likely that a specific composite

of restriction sites exits for each subspecies or group of subspecies.









66

RFLP analysis has revealed greater subspecies-specific genetic variation

in honey bees than has any other method to date. However, for routine use

(e.g., identification for regulation, large scale population studies), a more rapid

method of analysis would be required. Limitations of RFLP analysis for routine

use include the isolation of sufficient genomic DNA for two restriction

digests/Southern blots, the (almost certain) use of radioisotope for labeling the

probe DNA, and the amount of time for obtaining results. In addition, a

different approach was needed to investigate the allelic nature of the restriction

fragments in each digest, and to identify subspecies-specific length

polymorphisms and/or polymorphic restriction sites. For such objectives, the

slight differences between variants in the sizes of restriction fragments, the

complex nature of the restriction fragment patterns, and the size of the locus

precluded the continued use of the RFLP/Southern blot technique.

Conversion of the RFLP analysis to a PCR (polymerase chain reaction)-

compatible format (Saiki et al., 1988) would expedite the identification of

individual bees, the determination of the allelic nature of the restriction

fragments, and permit the use of the polymorphic sites to initiate investigations

of the variation, organization, and genetic content of locus 178. The

localization of restriction site and length polymorphisms to specific regions of

the locus would identify the regions) for which PCR-primers could be made.

The conversion process was initiated by isolating smaller regions of probe

pB178, identifying the Mspl and Ddel restriction fragments to which each









67
region hybridized in Southern blots, and identifying Mspl and Ddel sites within

each region.




Materials and Methods

The cloned probe pB178 was replicated in Escherichia coli strain DH5a

(Gibco BRL) and isolated by the alkaline-lysis procedure (Sambrook, Fritsch &

Maniatis 1989).

Mapping restriction sites within pB178. Purified pB178 was digested

with Pstl to release the 9.45kb honey bee DNA insert (referred to as 178),

which was subsequently separated from the vector by electrophoresis in 0.8%

SeaPlaque agarose (FMC) and isolated by extraction with phenol and ether.

The relative positions of the restriction sites were determined by sequentially

and reciprocally digesting 178 and pB178 with up to four restriction enzymes.

pB178 was treated with one or more restriction enzymes, and one aliquot was

set aside for analysis while another was digested with Pstl to determine the

location of the endonuclease recognition sites relative to the Pstl ends of the

insert. The digested DNA was electrophoresed in 1% agarose (Kodak IBI), and

the fragments were visualized by staining with ethidium bromide. Fragment

sizes were estimated by comparison to molecular weight standards using a HI-

PAD digitizer. For comparison and confirmation, the fragment sizes were also

estimated using the National Center for Supercomputing Applications (NCSA)









68

GelReader 2.0 shareware package for the Macintosh (NCSA Software Tool

Group at the Center for Prokaryotic Genome Analysis, University of Illinois).

Mapping Mspl and Ddel sites. Smaller regions of 178 were purified from

SeaPlaque. These regions were digested with Mspl and Ddel in the presence

and absence of other restriction enzymes known to have recognition sites

within the region. One-half of each digest was electrophoresed in 3% agarose

and visualized by staining with ethidium bromide; the other half was end-

labeled (Sambrook, Fritsch & Maniatis 1989), electrophoresed in agarose, and

visualized by autoradiography. Fragment sizes were estimated using the

GelReader program.

The correspondence of the fragments composing the Mspl and Ddel

variants to the physical map was determined by probing Southern blots with

smaller regions of 178. The same regions of 178 used for mapping the Mspl

and Ddel sites were sequentially radiolabeled by random-priming (Feinberg &

Vogelstein 1983, 1984) and hybridized to the blots shown in Figures 2, 3, and

5-7.

Subcloning pB178. Most of the 9.45kb insert was subcloned into the

plasmid vector pGEM3Z (Promega). The subclones were given the prefix

pG178, and each was identified by a suffix specific for the region subcloned

(see Figure 12). The pG178 subclones were replicated and purified in the same

manner as pB178. The subcloned inserts were excised and extracted from

SeaPlaque. The ends of the subclones were sequenced, using the M 13 forward









69
and reverse sequencing primer sites in the vector, at the University of Florida

Interdisciplinary Center for Biotechnology Research sequencing facility. The

sequences were checked for Mspl and Ddel sites using the Seqaid II program

version 3.5 (D. Rhoads & D. Roufa, Kansas State University).



Results


Restriction sites in 178 used for subcloning, for mapping Mspl and Ddel

sites, and for Southern blot hybridizations are shown in Figure 12. Five non-

overlapping fragments, representing the entire 178 insert, were isolated and

used to probe Southern blots (Figure 12): 178PK, 178KH, 178HB, 178BE4,

and 178E4P2. The latter four regions were subcloned (Figure 12) and the

inserts were used to map Mspl and Ddel sites. The 2.6kb Pstl, to Kpnl

fragment (178P,K) was isolated from pB178 but was not subcloned. The first

1.1kb of 178P,K, from Pstl, to EcoRI, (178P,E,, Figure 12), was subcloned and

the Mspl and Ddel sites were determined, but it was not used to probe

Southern blots.

Identification of Mspl sites in 178.

The Mspl sites identified in 178 are shown in Figure 13 and are

presented in the same format as Figure 12. The results of the Southern blot

hybridizations, shown in Figures 14 19, are presented in the same format as

Figure 1.




















Figure 12. Map of restriction sites identified within the 9.45kb, Pstl-honey bee
genomic DNA insert of probe pB178. Numbers to the left of the continuous
verticle line of the map indicate the estimated position of the restriction site
(identified on the right) relative to one of the Pstl insertion sites (Pstl1).
Relative positions of sites were determined by digesting the intact clone pB178
with each enzyme in the presence and in the absence of Pstl, and in
combination with up to three additional enzymes. These sites were confirmed
by isolating the 9.45kb insert (digestion of pB178 with Pstl followed by
electrophoresis in and extraction from 0.8% Seaplaque) and repeating the
single and multiple enzyme digests in the absence of Pstl. These sites served
as landmarks for determining the locations of the Mspl sites (Figure 13) and for
the Ddel sites (Figure 20) within the probe.

The short verticle lines on the left side of the page, together with the
letters in bold, indicate the regions of pB178 subcloned into pGEM3Z. The
pGEM3Z subclones were identified according to the region contained in each:
pG178PE,: 1.1kb, Pstl1 site (top) to the first EcoRI site;
pG178KH: 0.9kb, Kpnl Hindlll;
pG178HB: 1.9kb, HindIll BgIIl;
pG178BE4: 2.0kb, Bg/Il EcoRI;
pG178S3E4: 0.75kb, third Sail site to the fourth EcoRI site;
pG178E4P2: 2.0kb, fourth EcoRI site to terminal Pstl2 site.

The dotted verticle line indicates the 2.6kb PsttI-Kpnl region which was
used for mapping and hybridized to Southern blots, but was not subcloned.









1 --- Pstl,


- P P






-E,


1898 BstXl EcoRI
1300 Sal, Sstl


Xbal
SaA

EcoRI
EcoRI


236 --- Kpnl


330o Hindll


o4830 EcoRV



6440 --- BgIll


= Spel


Sphl


3g8 I Sphl Nsil
6700 Sail


7440 -- EcoRI


Nsil
Xbal
EcoRV


8280 ---
8460
8710--


*460 PStI2


-K


K





-H


-B








-S,




-E,









72

Digestion of 178P,K with Mspl produced fragments of approximately

0.73kb, 0.62kb, 0.41kb, and 0.38kb (Figure 13). Double digestion of 178P,K

with Mspl and each restriction enzyme with a recognition site within 178PK

(Figure 12) revealed that there were two 0.38kb Mspl fragments. The

fragments to which 178P,K hybridized are shown in Figure 14. The intense

hybridization signal seen at 0.38kb in the M100 variants in the hybridization of

both pB178 and 178P,K was due to the presence of two fragments. Most of

the variants were identical in the fragments to which 178P,K hybridized. The

Mspl restriction fragment pattern of 178P1K resembled Mspl variants M101

and M103. Additional Mspl sites, shown in Figure 15, have been hypothesized

to account for the fragments seen in other variants.

Digestion of the 0.9kb Kpni HindIll region of 178 (178KH) with Mspl

produced a single visible fragment of approximately 0.65kb. Two Mspl sites

were identified near the Kpnl site in the sequence of the pG178KH subclone

(Figure 13: these fragments had not been detected following the digests due

to the small amount of DNA used in the gels). The Mspl fragments to which

178KH hybridized are shown in Figure 16. In the majority of variants, 178KH

hybridized to a 0.75kb fragment or to one just slightly larger or smaller. In

these variants a very faint hybridization signal was seen at 0.56kb (not shown

in Figure 16), which was also detected by 178P,K. In variants M102 and

M201, 178KH hybridized to a 0.75kb fragment, as well as to a 0.62kb

fragment to which 178PK appeared to have hybridized. As indicated in Figure




























Figure 13. Map of Mspl restriction sites identified in honey bee genomic DNA
insert of probe pB178. The Pstl sites marking the ends of the insert were
utilized for cloning in pBR322. Relative positions of sites were determined by
digesting smaller regions of the insert (Pstl, Kpnl, Kpnl Hindlll, HindIll Bg/ll,
Bg/I EcoRl4, and EcoRI4 -Pstl2) with Mspl in the absence and presence of
restriction enzymes within each region, which were indicated in Figure 12
(with the exception of the region from the Kpnl site to the Hindlll site). The
asterisks (*) indicate sites identified in the DNA sequences of the ends of the
respective subclones. Left: distances (in base pairs, bp) between Mspl sites.
















730


1110
1220


1800




2220


2835
279 -
2880




3630










6000 -


?8 =


5770 -
8020 -











7680



8140


Pstli





Mspl





Mspl



- Mspl


- KpnI





Hindlll

WfI


Mspi

~MspI *
BgfiI

MspI
Mspi









EcoRI





- MspI








MsPI
Pspti


1250bo


1400bp


1300bp



















cE -
90Q I I 1 1 |S


OL9Y m.
l60I I I I
rL9LAJ *



609yI Ioo 4 I

8099V1 I 1 =
i09 I | I

09v I I I I I



109'VM | I I
oo(D

LotA I I 1 0to





.>c
E >

80EsA I I I I

0 0-
IOV/M I I" I I



o
:O3i/ 1s) 4
,I, ,- -
NN O",O
LOI I I I I5 ,




LOLV I I .a

. C. .t cJ- Cor o oD of 0
'- ^- 6d d d LrCO 3













3: c : o
- .. 0- 0
0cE 4ot =-. 4-a-

.crBU)


0044' -

4-0 U)0 CV
%- o 0P~


0) 00
0 E
CL (M ~ (




to 0.

0 c c*

0~00
(D4'0 CL
.E Q.. 09 u,




C C 00.00
U)i )

C 0_ Vp 0)









0.05 N



4) U)


0Y 0 he
- 4 0., 0 U)
c > a SiP~o

o 0) r
















00 = a) 04-0
0 r +0 16. j
>~ 0 U) 0 0
or r0r

















4-0 -0
oQo C)In 0
m z,

CD tm,

t 0O
0.'i P Q
(D 'P
.0 Q5 )


M 4
< s- .;i


.l ci
C O > ;r


0 M CL :t)
,;: a C a
m 0;~C
0 0.- -
+j~-
!t 0 C
4- CL U)


CP u)

uij .2 c,4 c r
CT)g~ yjto
Ee
C 4-E~
7EC U)' F
o~a a en, ~!! )











; o o e -
1 2 1o o


I o A o
0 Q 0


i a liii1 I


a i


-s a
0,


0 I csII I
. 0 -,
So o o


o o 0 WN
d d d ds


' s .


.0 II !
a
O N
Lo o co





o


Ii
Lo


V4,
maI,
B g *m


0.
0
4


0 0 -
NtN
9


_


I










N


06


00


I I 1
Zd d



I I




l i j


m 10 0
t *0
C4N


1


I I I
go 1
no 0*


9L9V1

C LM

L LSI/M
0LSI4
6091f
8091I
L09kW

e09W

O09k

L091


9Ot:0'J

Ot'V/IAl
ZOVfI
LOV,







ZOZI/
IOZI

COLIAI

LOLW


N


co0 t c --
w- --


Lon L
00


.c
. 78




0.r


cia
CD
o wO






"0

S*s o *C
0 c x





E 50 a
e04

- C
. <= +_
+ c c
._ -< 0

^S4a

S 0 >

u,.n
E .-






- -W







*i C 0
.0S v
*C U,









i4 D 75
0-' E
1< 9.'f
zS^o
* o0 U

0).g


I
I I









79
16, the loss of an Mspl site in this region (at approximately 2785bp) could

explain the presence of the 0.62kb fragment in these two variants.

Digestion of the 1.9kb region from Hindlll to BgAl (178HB) with Mspl

produced fragments of approximately 1.3kb, 0.42kb, 100bp and 70bp. Mspl

sites were identified near the Hindlll and Bg/l sites in subclone pG178HB

(Figure 13). Mspl restriction fragments to which 178HB hybridized are shown

in Figure 17. The Mspl restriction fragment pattern of 178HB corresponded to

variant M103: two site polymorphisms are indicated in Figure 17 that could

account for the fragments detected by 178HB in two other variants.

Digestion of the 2.0kb Bgll EcoRI4 fragment (178BE4) with Mspl

produced fragments of approximately 1.4kb, 0.33kb, and 0.25kb (Figure 13).

The Mspl fragments to which 178BE4 hybridized in Southern blots are shown

in Figure 18: these included combinations of approximately 1.8kb, 1.4kb,

1.25kb, 0.77kb, 0.56kb, and 0.35kb fragments. The loss of an Mspl site in

the 178HB region (Figures 13 & 17, at approximately 5420bp) could account

for the 0.77kb fragment detected in variants M101 and M302, and for the

detection of this fragment with both 178HB and 178BE4. Since 178BE4

hybridized to more Mspl fragments than had been found in the digestion of the

insert and accounted for by overlapping with adjacent regions, the placement

of Mspl sites within 178BE4 was based on double digests with Spel, Sphl, Nsil,

and Sa/i, and the results of the hybridizations of 178HB, 178BE4, and 178E4P2

(following).









o 5 < 80



0 Z
S I I I o.





. o.0 0



II I i i s
.-C
c 31



ILSn I I ._
:a 0




L9I I M og



609V ( D"'

LO9V 0


9091II I o
909s1 0 Z

0 "r 0



II II
S90, It .. 0


* I 0 f I _. .
LOW LAg
U' S







01I .6
l,0ifti I | -.

N m DN ot0
Y,-







POlfI .

on 2C L t

0)l^ '- O 00 0 0






















9 L fI
tp L SVl

z L991


OLSVY
609IM
L091Y1

9091




ZOSIVI


s0141V









COEW



ZOZI^
LOZ



ZO L Vl
LOLVW

N0t
N


CL I

co.~c



Ca,


0 M



:z



LIU E
0
C >





o 0
Sc




N C
-c

r c. c

e r
-N C
0=0
10 0r-
i P (/I 4


o- C N-


c;o


0;



















0 c 0




0 .
O9W A2


EO 91/l I I
o. -c
6091VW c o
o
90 1 "Q..c
909W 0 f

Ogl/MI c I

tos071 I 2 <

S-c
cOWVJ I I ,-


LOt7lAJ I I ---
ZOSAJ II
IO W I -o
00






LOEvI I I
2.:-
oo




(D E
CD LO OOt A I-




ON 'It CI c00 0 0 a
Zo ,- :- -: ;-' do o U := .C_









83

Digestion of the 2.0kb EcoRI4 Pstl2 fragment (178E4P2) with Mspl

produced fragments of approximately 1.3kb and 0.56kb. Two additional Mspl

sites were found near the EcoRI site in subclone pG178E4P2 (Figure 13). The

fragments to which 178E4P2 hybridized, shown in Figure 19, appeared to

include the 1.8kb, 1.25kb, and 0.56kb fragments to which 178BE4 hybridized.

In the majority of variants, 178E4P2 hybridized to fragments of approximately

1.25kb and 0.56kb. The existence of the 1.8kb fragment in the M300 variants

could be explained by the loss of the Mspl site at 8140bp (Figure 13).

Since it is likely that the duplication in the fragments to which 178BE4

and 178E4P2 hybridized resulted from contamination of 178BE4 with 178E4P2,

the placement of the Mspl sites shown in Figure 13 was based on the digestion

of different aliquots of these two regions with Mspl, in the absence and

presence of other restriction enzymes which have sites in each region.



Identification of Ddel sites in 178.

The approximate locations of Ddel sites identified in the honey bee DNA

insert of pB178 are shown in Figure 20, in the same format as Figures 12 and

13. The results of Southern blot hybridizations are shown in Figures 21 25,

in the same format as Figure 4.

Digestion of 178P,K with Ddel produced fragments of approximately

1.3kb, 0.75kb, and 0.4kb. A single fragment of approximately 1.0kb was

detected when 178P,E, was digested with Ddel (Figure 20: refer to Figure 13





























Figure 20. Map of Ddel restriction sites identified in honey bee genomic DNA
insert of probe pB178. The Pstl sites marking the ends of the insert were
utilized for cloning in pBR322. Relative positions of sites were determined by
digesting smaller regions of the insert (Pstl, Kpnl, Kpnl Hindill, Hindill Bgill,
Bg/ll EcoRI4, and EcoRI4 Pstl2) with Ddel in the absence and presence of
restriction enzymes within each region, which were indicated in Figure 12 (with
the exception of the region between the Kpnl and HindIll sites). The asterisks
(*) indicate sites identified by sequencing the ends of the respective subclones.










- P,


1300bp


1 Pstl,
I.o -- Ddel *


1460 -- Ddel


750bp


2200


2636 -


2000bp


Ddel


Kpnl


330 -- Hindill


428o Ddel


1250bp


440 -
6630 -

5780-


1270bp


7060 -


7440 -


1200bp


- de

-Ddel








SDdel


EcoRI


8300o Ddel


> 1150bp


- p2


- Pst,2


9460









86
for 178P,E1 location and size). Analysis of the sequences of the ends of

subclone pG178PE, revealed a Ddel site at 150bp from the Pstl site (Figures

20 & 21). The Ddel fragments to which 178PK hybridized are shown in Figure

21. In the majority of variants, 178P,K hybridized to a fragment of

approximately 0.75kb, and another fragment of approximately 1.3kb. In the

D100 and D200 variants, a single fragment of approximately 2.1kb was

detected. The loss of the Ddel site at approximately 1450bp would account

for the 2.1kb fragment in the D100 and D200 variants. The variation in the

size of the fragment of approximately 1.3kb may be the result of length

polymorphism(s) closer to the Pstl1 site.

There were no Ddel sites within 178KH. A single fragment in each

variant, approximately 2.0kb, was detected when 178KH was used to probe

this region, shown in Figure 22. On the basis of the Ddel sites identified in

178P,K, it was concluded that the same 2.0kb fragment had been detected by

178P,K and 178KH, and would be detected by 178HB as well. The placement

of the Ddel sites on either side of 178KH is shown in Figure 22.

Digestion of 178HB with Ddel resulted in fragments of approximately

1.15kb and 0.75kb. The Ddel fragments to which 178HB hybridized are

shown in Figure 23. In each variant, 178HB hybridized to the 2.0kb fragment

detected by 178P,K and 178KH. One other fragment was detected by 178HB

in each variant: approximately 1.0kb in the D100s and D400s, 1.25kb in the

D200s and D300s, and 0.98kb in the D500s. Two possible locations for a site











I I-
N

c.0


LO90
9090
sosa

coga

z090
t 090

9 Lt'C
t7 LtFQ
S L,1a




DL tO

60tp0
80to
LOtO
90ta
901a
gotva

1OW

zotla
9o0w


90ca











zo La
LO tO


050






L0VJ





a0,0


0. 0
+ .C 9C.




o o


(D oco



--
00- M-






0 ) ) (0
0 0



o a 0o




.-D o 0








4.c > EE
0 0- -0
0. w C 0


0D Ca 0,-








0) 0 0 0
I*E c o-
=rI .-I1 5
WOc :i E^
irl5^


Cit Cn C


CD
tQ N ,-0




















I I I I


9t cm C
C C4 C-


UO


LOgO
9090
9090
tF090

zos090



9 L4'O




L Ltn
0 L t'

60k(0
L0170

90tr


t otla





90g
90g
0






ZOLO




LOLE


=- 0
EQ
(U D0 0L 00

o .c o.


cf4C 0





S0 o,



-Z_0 0C

0C 0
o 0 0.








1.0 0 -0
Co C -o
o 0 *53





C .S CT0







0 0E

. a>.









.L :(5
U. $+ 0
.9s >.,T













I I
0N0



I I a I

o N



N 1


LOSO
9090
9090

vosa
Cosa
0osa

9 L 170
V' Li'o


L LPa

0 LtpO
60ta
801'
LOWt'
901a




90S1a



90t0
coea

zosa










0
in S


--0


LO


O0



Z0 o .X C
, 0 .=0
00 +- C) (A
O C RI $

r- 2 -- Z.
oo C (a









0*- N -
c a-' .0
, a0Eo .



0 5 Cu L











c *0 ) 0 C
SE... + a
. *5 (0 .-
z 0 K







il: .0 = 4- c


(V) (0
CV) cjn C









90

which could account for the 1.0kb fragment seen in the D100 and D400

variants are shown in Figure 23.

Digestion of 178BE4 with Ddel resulted in fragments of approximately

1.27kb, 0.39kb, and 0.25kb. The Ddel fragments to which 178BE4 hybridized

are shown in Figure 24. In the majority of variants 178BE4 hybridized to

fragments of approximately 2.1kb, 1.3kb, and 1.05kb. A Ddel site was

identified near the Bglll site of subclone pG178BE4 (Figure 20). Since 178BE4

hybridized to more Ddel fragments than it appeared to contain, which also

occurred when 178BE4 was used to probe Southern blots containing the Mspl

variants, the placement of Mspl sites within 178BE4 was based on double

digests with Spel, Sphl, Nsil, and Sail, and the results of the hybridizations of

178HB, 178BE4, and 178E4P2 (following).

Digestion of 178E4P2 with Ddel produced fragments of approximately

1.2kb and 0.85kb (Figure 20). The fragments to which 178E4P2 hybridized are

shown in Figure 25, and include what appeared to be the same 2.1kb and

1.05kb fragments to which 178BE4 had hybridized. Sites corresponding to the

fragments greater than 2.3kb in the D400 variants lie outside the probe region.

The duplication in the Ddel fragments to which 178BE4 and 178E4P2

hybridized was concluded to be due to contamination of 178BE4 with 178E4P2.

Therefore the placement of the Ddel sites shown in Figure 20 was based on the

digestion of different aliquots of these two regions with Ddel, in the absence

and presence of other restriction enzymes with sites in each region.





















LOWC
9090
9090


L09a
Losa
L090


9 LtPa



Z L C]
L Lt7G

60'G

LOt?0
90t f
L0t G







ZOtJ'0


90M0
VOW







90G
9oea


0 G



coza
L 0oza



O La
LOLO





N
C/3 -M


2u I
I I
I I
I
I I
I I
I I

I
I
I I
I
I
I I
I I
I
I I
I I
I I
*
I I
El
I


CY CN C4


4- >
0)00
0 o
ca"
(D 0
z r
4 -4
:3 E
C

00

0
crc
u

00 >
N 0
0
co

F..
- 0


o E;
0I




c mo
C

olr
o c
a)00

4'l
*0)a
acr to
CCPC
^g
* TO


0

(D


C2-


0
osi





c w
0 Z 00



00
N c-




.20 0
uO o 0
. >ic




eq 7 C
saSC
(D -ro
L-~

3 4C
U0


LC)
"tC1


,-o

















u( -o
LOQa A ? .
9090



Z090 |n



9L170 I I 0
eLtpa i w Q
LL170 I I
OLeO I I
60t, O E
Lo ta Z I a'

90L 0 I (ai
90170 'I

EO7I I i S -
ZOI7 I I aS
O00 I IE
.0 4, M
gooa 0 .0




coeo ( D



Voca I I :,
zoso '0'

ozoea C4 U
ioza m n Z o

*ota m n-





O 1 l1 o a
6--5 ..6 L w
ln 1 3 .









93

The sizes of fragments detected in Southern blots and by ethidium

bromide-staining or end-labeling did not always correspond. The migration of

DNA molecules in agarose, and ultimately the determination of the sizes of the

DNA fragments so separated, depends on the buffer system, the voltage

gradient, and the concentration of the agarose (Southern 1979). While the

same buffer was used for all electrophoretic separations reported here, the

voltage gradient and agarose concentration were not always the same. All

Southern blots involved 2% agarose, but the fragments produced from the

digestion of 178 and smaller regions of 178 were separated in anywhere from

1% 3% agarose. Although more than 135 Southern blots were run, some of

the fragment sizes from the Mspl single and double digestions were only

determined in one or two gels. A more precise estimation of the sizes of the

Mspl and Ddel fragments within each region of 178 could be obtained by

repeating the digests, gels, and measuring the fragment migration distances,

followed by linear regression analysis. However, the purpose of the

hybridizations and digestions performed was to localize polymorphisms, which

was accomplished. Further details can be obtained from the proposed PCR

work and by sequencing.



Discussion


The Mspl fragment pattern resulting from the digestion of 178 with Mspl

resembled the pattern of M103, which was found at high frequency in USA









94
bees and belonged to a group of variants found at very high frequency in bees

sampled from east Europe and North America. This was not surprising, as the

library from which the original clone pB178 was obtained was constructed with

DNA from bees of east European ancestry (Hall 1986).

A few polymorphic Mspl sites were identified in 178. Additional Mspl

sites in each region of 178, not detected in Mspl digested of the probe DNA,

have been hypothesized to account for the fragments detected in other

variants, some of which are indicated in Figures 14 19. For example, the

M400 variants as a group were characterized by a unique 1.1kb fragment, but

lacked a 0.85kb fragment present in all of the other Mspl variants (Figure 1).

The results of the hybridization with 178P,K (Figure 15) demonstrated the

allelic relationship of the 1.1kb and 0.85kb fragments.

M101, M103, and the M400s did not have a fragment around 0.27kb

which was observed in the other variants (smallest fragment shown in Figure

1). The Mspl sites which could account for the 1.1kb and 0.85kb fragments

were concluded to lie outside the probe region (Figure 15, sites a and b), in

which case the Mspl fragment representing the difference between these two

sites would not be detected by the probe. The site that accounts for the

0.27kb fragment seen in all variants except M101, M103, and the M400s

could then be located between sites c and d, or between sites e and h. The

pattern detected for M503 provided the answer; it contained 0.49kb and

0.27kb fragments, and did not contain a 0.38kb fragment. Site f, between