Biochemical Genetics, Vol. 14, Nos.

9[10, 1976

Biology of a Duplicate Gene System with Glucose

6-Phosphate Dehydrogenase Activity in
Drosophila melanogaster: Genetic Analysis and
Differences in Fitness Components and Reaction
to Environmental Parameters Among Zw Genotypes
J a m e s T. Giesel I

Received 27 May 1975--Final5 Apr. 1976

There are two structural forms of glucose 6-phosphate dehydrogenase activity

in Drosophila melanogaster. Whether one or the other or both show in vitro
(and probably in vivo) activity depends on the genotype of a sex-linked locus
(Zw). In this article, the relative fitnesses of heterozygotes (with both electro-
morphs active) and homozygotes (with activity demonstrable for only one or the
other electromorph) for the Zw locus are described. It is shown that the relative
fitness of heterozygotes increases with increase in population density, or degree
of crowding and trophic stress, and that the mean development times of Zw
heterozygotes are lower than those of the Zw homozygotes. In addition, and
perhaps accounting for the fitness and viability excess of the heterozygotes, one
set of evidence strongly suggests that they are better buffered against trophic
stress than the homozygotes.

KEY WORDS: fitness; Zw genotypes; genetics.

INTRODUCTION
Duplicate gene systems are quite common in poikilotherms, where their
fitness value is thought to be related to the ability of their possessors to
acclimate and maintain near-optimum levels of metabolic function in spite of
changes in physical parameters of the environment. For example, there appear
to be two distinct gene-enzymes with acetylcholinesterase activity in the steel-
t Department of Zoology, University of Florida, Gainesville, Florida.
823
© I976 Plenum Publishing Corporation, 227 West 17th Street, New York, N.Y. 1001 I. N o part o f this publica-
tion may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic,
mechanical, photocopying, microfilming, recording, or otherwise, without written permission o f the publisher°
824 Giesel

head trout. One of these, with a high temperature optimum, is thought to be

induced or activated when the fish enter warm water; the other, with a lower
optimum temperature, replaces the high-temperature gene-enzyme in cold
environments (see Somero and Hochachka, 1973). It appears that all members
of the population have this ability to switch from one gene-enzyme form to
the other. The phenomenon is not limited to this enzyme or species but
appears to be quite general among poikilotherms.
In Drosophila melanogaster there exist two different forms of glucose
6-phosphate dehydrogenase. One of these migrates slowly upon electro-
phoresis, has a molecular weight of 317,000, a relatively low Km G6P, and a
relatively high K,, NADP. The other migrates much faster, has a molecular
weight of 150,000, a higher Km G6P, and a lower/£,, NADP (Steel et al., 1968;
but see also Komma, 1968). These two forms are the allelic products of Zw,
an X-linked gene. Zw has been defined as the structural gene for G6PD
activity (Young et al., 1964). However, when electrophoresis is carried out
for a longer time than usual and under a variety of conditions, all designed
to minimize subunit dissociation (see Steel et al., 1968), one finds that both
forms segregate in their own right for several electromorphs.
It is the purpose of this article to present genetic data which demonstrate
that the slow- and fast-migrating forms of G6PD are the products of different
(duplicate) structural genes and that Zw must be a modifier locus which acts
to determine the relative concentrations of the two structural gene products.
In addition data will be presented which speak to the relative fitness of Zw
homozygotes and heterozygotes and possible reasons for the results obtained
will be discussed.

MATERIALS AND M E T H O D S

Genetic Inferences

In order to investigate the genetics of this system, a mapping experiment was

done on the slow form (G6PDs), segregation analysis was done on the fast
form (G6PD0, and genetic distinctness of the loci was investigated by using
over 100 flies to find whether electromorphic phenotypes of the two loci are
independent. Mapping of G6PDs was done by crossing h-, ri-, es-, I D H s,
AO ~, G6PD~ with Oregon-r (h +, ri +, e~+, I D H ~, AO y, G6PDZ~), backcrossing
their progeny to h-, ri-, e~-, I D H ~, AO ~, G6PD], and analyzing gametes for
recombination frequency between the known marker loci and G6PD~.
Segregation analysis was done on G6PDf by crossing putatively homozygous
females (1/1) with putatively heterozygous males (1/3) and investigating the
progeny genotype frequencies for fit to Mendelian expectation by X~ dr.
Biology of Duplicate Gene System 825

Genetic independence of the loci was investigated by Z2 analysis of over 100

flies taken from a cage population in which both were segregating for two
alleles.
In the mapping and segregation analysis experiments, electrophoresis
was done on 11% starch gels made with 0.025 M histidine-HC1 (pH 7.5) to
which 0.005 M TPN had been added. Electrophoresis was carried out in a
0.47 M Na-citrate (pH 8.0) bridge buffer for 18 hr at 25 mA/gel at 5 C. Gels
were split and stained for G6PD, isocitrate dehydrogenase, and aldehyde
oxidase following the methods explained by Brewer (1970). The flies used in
the independence analysis were treated as explained below.

Fitness Analysis
To collect the data presented here, individual flies were homogenized in 1.5/~1
of a 20% solution of sucrose in 0.1 M TBEDTA (pH 8.6). The samples were
centrifuged at high speed in a Misco refrigerated centrifuge, after which
supernatants were loaded into the pockets of vertical 5 ~ polyacrylamide gels
made with 0.3 M TBEDTA (pH 8.6) to which TPN had been added. The gels
were electrophoresed in the same bridge at 300 V for 5 hr. Phenotypes were
resolved by the staining method of Brewer (1970) for G6PD in Drosophila.
G6PDs migrated under these conditions at about half the rate of G6PDf (see
Fig. 1), but both forms stained with the same time constant. Phenotypes were
read after 45 rain of staining at room temperature (about 25 C). Both forms
were polymorphic.
The data presented here are from three sets of experiments, only the last
of which was designed specifically to investigate differences between genotypes
of the Zw locus in the components of fitness as related to levels of temperature
and food.
In the first experiment, three large cage populations were kept at constant
food and temperature and censused at monthly intervals and simultaneously
assayed for genotypes of several gene-enzymes, including the two G6PD loci.
Using these data, relationships between population density and the allele
and genotype frequencies of the Zw locus could be statistically investigated.
In the second set of experiments, again done for another purpose, flies
were collected from a large, long-standing cage population as eggs by allowing
the population's females to oviposit on fresh cups of medium for 1 hr.
Separate samples were collected sequentially on Carolina Biological Supply
instant Drosophila medium (CBSIDM), CBSIDM diluted 50~ by weight with
Fuller's earth (which Throckmorton, personal communication, believes is
ingested but not metabolized), and CBSIDM fortified with glucose--5% by
weight. The flies were allowed to develop and later were analyzed for Zw
genotype. Each set of data was investigated for deviations from Hardy-
826 Giesel

Fig. 1. Electropherogram showing some variants of G6PD

gene-enzymes.Columns 1, 2, and 9-12 are Zw a homozygotes;
the flies in columns 3-8 are Zw ~ homozygotes. The loci are as
follows: slow, G6PDs; fast = G6PDf.

Weinberg expection by comparing observed with expected genotype fre-

quencies by 7~ dr.
The third set of experiments was intended to provide data on the develop-
ment times, average ages at reproduction, and mean numbers of eggs per
day per female of Z w heterozygotes and homozygotes when raised under the
environmental conditions presented in Table I. For this purpose, lines of flies
homozygous for the Z w allele giving only G6PD~ activity (Zw BB) and for the
Z w allele giving only G6PDf activity (Zw AA) were isolated. These were then
outcrossed to give 15 nonidentical F~ lines of each Z w genotype. These mated
pairs of flies were allowed to oviposit for nine successive 8-hr periods. Each
Biology of Duplicate Gene System 827

Table I. Experimental Regimes

Food 28 C 25 C 19 C
Low" X X X
Medium" X X X
High b X X X

" CBSIDM.
CBSIDM plus 5% (wt) sucrose.

sample was developed in one of the environments shown in Table I. In this

way, the responses of full sibs to a variety of environments could be obtained
for each genotype and genotypes could be compared with respect to develop-
ment time (measured by counting flies at 4-hr intervals as they emerged).
After eclosion, virgins from the same culture were pair-mated and placed on
fresh medium of the same type where the females were allowed to oviposit.
Females were transferred to new medium every 24 hr for 12 days and egg
production per day was measured. Thereby data on genotype- and environ-
ment-specific fecundity distributions were produced.
Statistical methods used in analyzing these data are presented in the
appropriate places in the Results section. The structural loci were segregating
for all of their alleles in all experiments.

RESULTS

Genetic Inference

An electropherogram of some typical flies is presented in Fig. 1. The flies in

columns 1, 2, and 9-12 are Z w A homozygotes but are clearly segregating for
different allelozymes. We see here that G6PDe segregates for three alleles.
The flies in columns 3-8 are from a true-breeding Z w B line. Again (two)
distinct phenotypes can be seen. Note, for future reference, that Z w ~ flies and
the Z w A flies, particularly those shown in columns 11 and 12, demonstrate
secondary activity for G6PDf and G6PDs, respectively.
The results of the G6PDs mapping experiments are shown in Table II.
These demonstrate that this gene is linked to the far left end of chromosome
III.'
G6PDf has not yet been mapped. However, a cross between a 1/3 male
(heterozygous) and a 1/1 female (homozygous) yielded 56 progeny (males and
females) which demonstrated the presence of both alleles and 44 which had
only band 1. A Z~ af value of 1.44 allows us to accept the conclusion that this
set of phenotypes is due to the segregation of alleles of an autosomal locus.
828 Giesel

Table II. Results of Mapping Experiments of G6PD~"

ri G6PD l'l 106 IDH s's G6PD 1'1 69

ri G6PD 1'2 49 IDH ~'s G6PD 1'2 35
+ G6PD 1'I 55 IDH s'e G6PD 1'I 30
+ G6PD 1'2 114 IDH~'f G6PD 1'2 83
Percent crossover = 0.321 +0.053 Percent crossover = 0.29+0.062
Z~ as = 41.24 Z~ df = 36.92

"AO-G6PD percent crossover indistinguishable from 50Yo; eS-G6PD

percent crossover indistinguishable from 50%.

In the test for genetic independence, 104 flies taken from a cage popula-
tion were scored for genotypes of the loci. The data were organized into a
3 x 3 matrix of genotypes. This revealed that all nine possible genotypes were
present. Moreover, when gamete frequencies were taken it was found that
coupling gametes were significantly in excess of repulsion gametes (X] af =
17.58), which suggests that the loci may be linked and were in this case in a
state of linkage disequilibrium. Specific data on the exact linkage relationships
of the two loci will be published elsewhere.
Thus it becomes necessary to redefine the function of the Z w locus. This
locus acts to determine the in vitro, and probably in vivo, relative activities of i

the G6PD duplicate gene-enzymes. Z w heterozygous females show roughly

equal activity for these enzymes, while Z w homozygotes and hemizygous
males display, upon electrophoresis, primarily the appropriate gene-enzyme
(Zw bb is primarily active for G6PDs, Z w a" for G6PDf). However, both Z w
homozygotes can, within basic limits set by the Z w genotype, exhibit activity
for both structural gene-enzymes. Work designed to. define the reasons for
this variability of expression of Z w genotypes is in progress.

Components-of-Fitness Analysis
Data from the first experiment revealed, upon inspection, that the hetero-
zygote genotype of the Z w locus was always in excess of Hardy-Weinberg
expectation in the cage populations--even though allele frequencies were
relatively constant at a b o u t p Z w ~ = 0.45-0.52. The data were then examined
graphically to detect evidence of association between extent of heterozy.gote
excess and population rate of increase for the corresponding period. A positive
correlation was revealed between extent of excess and population density for
the eight cases in which sample sizes were large and the nonrandom fre-
quencies of genotypes could be confirmed by X2 (Fig. 2). This suggests that
(some component of) heterozygote relative fitness must increase with popula-
tion density, or, since in these experiments food level was constant while
Biology of Duplicate Gene System 829

40 I I I i I

,m 3o 0 0
¢.~
X D
Lid
LU
o~
2o

t.O
l--
Od
I
I0-

oo ; ~; io ~o .'o
POPULATION 51ZE (THOUSANDS}
Fig. 2, Associations between population density and
Z w heterozygote excess (observed frequency minus
expected frequency).

population density varied, with degree of crowding or competition for

available food resources (see Discussion for another possible reason for the
heterozygote excesses).
The second set of data were analyzed by comparing observed and ex-
pected frequencies of the three Z w genotypes of flies taken as egg samples at
the same time from the base population and raised at 25 C with food main-
tained at high, normal, and low levels. The results of these comparisons are
presented in Table III. The only significant deviation from expected frequen-
cies is found in the sample raised under the most trophically deprived condi-
tions. The 2 6 ~ excess of heterozygotes is highly significant. Since the flies
used were derived from egg samples, this heterozygote excess must have been
due to the larval viabilities of the heterozygotes being higher than those of

Table III. Frequencies Zw Phenotype of Flies Taken from the Same

Base Population but Developed on Low, Normal, or Fortified Food

Phenotype

Condition ss sf ff Heterozygote excess

HF 0.13 0.46 0.4 Equilibrium

MF NO 30 87 63 0.08 (N.S.)
N E 38 73 70
LF NO 29 148 16
N E 56 96 41.3 0.26;p < 0.05
830 Giesel

either homozygote class. Again, there seems to be a direct relationship between

(a component of) heterozygote relative fitness and extent of trophic depriva-
tion (see Discussion).
The third set of experiments was designed specifically to investigate
genotypic performances in terms of development time, average age at repro-
duction, and average number of eggs laid per female per day in environments
which differed in food quality and (constant) temperature.

Development Time
The data were first analyzed by three-way analysis of variance to determine
the relative importance of the independent variables--food level, temperature,
and genotype--and their interactions in explaining the variance of the depen-
dent variable--egg to adult development time. Results of this analysis are
presented in Table IV. Here we see that all three independent variables are
important in the explanation of sample variance in development time and that
significant interactions exist between the two environmental variables--
temperature and genotype--and the environmental variables and genotype
of the Z w locus. As expected for a poikilotherm, food level is relatively
unimportant by itself but interacts strongly with temperature. Since it is of
major interest to this study to define the reactions of the heterozygotes vs.
those of the homozygotes to the environment variable levels, mean develop-
ment times of the two broad genotypic classes---heterozygotes and homozy-
gores--were calculated for the three different temperatures and for the food
levels. In Table V we see that at low temperature, which, for a poikilotherm,

Table IV. Results of the Three-Way Analysis of Variance of Food Level, Zw

Genotype, and Temperature and Their Interaction with Respect to Egg-Adult
Development Timea

Component of variance ss ~ss Fvalue

Food level 17,945 0.49 10.27d
Temperature 649,711 17.8 248a
Genotype 7,284 0.2 2.78b
Food levelx temperature 19,838 0.54 3.78c
Food level× genotype 4,211 0.1 N.S.
Temperature × genotype 20,610 0.56 2.26b
Food level× temperature × genotype 43,194 1.2 2.91b
Total 3,667,947

a Sample size 1245.

bp<0.05.
Cp<0.01.
alp<0.005.
Biology of Duplicate Gene System 831

Table V. Mean Heterozygote and Homozygote Development

Time at 25, 28, and 19 C and High, Medium, and Low
Food Levels

Genotype
Homozygotes significance Heterozygotes

At 19 C
LF 301 P < 0.01 286
MF 289 N.S. 283
HF 264 N.S. 269
At 25 C
HF 253 N.S. 247
MF 270 N.S. 230
LF 269 N.S. 271
At 28 C
HF 210 P<0.01 185
MF 216 P < 0.05 205
LF 225 P<0.01 193.5

means low f o o d d e m a n d , the two g e n o t y p i c classes have the same develop-

m e n t time except at low food, where the h e t e r o z y g o t e is superior.
However, at 28 C h e t e r o z y g o t e d e v e l o p m e n t time relative to t h a t o f the
h o m o z y g o t e class is lower for all t r o p h i c treatments. The p a t t e r n o f differences
suggests t h a t heterozygotes are m o r e efficient utilizers o f available nutrients
t h a n h o m o z y g o t e s . A n o b s e r v a t i o n which is n o t shown in T a b l e V is t h a t
Z w BB flies have significantly lower d e v e l o p m e n t times t h a n the Z w aa flies.

Fecundity
Because o f generally low viabilities, there were insufficient d a t a for a three-
w a y analysis o f variance. Therefore, m e a n p e r d a y p e r female fecundities a n d

Table VI. Comparison of Mean and Variance of Fecundity per Day and Age at
Peak Fecundity of Z w Homozygotes and Heterozygotesa

Fecundity/day Age at peak fecundity

Genotypes Mean Variance of means Mean Variance of means

Homozygotes 12.89 38.08 7.2 115

Heterozygotes 12.65 3.8 a 8.48 9.73 b

"Means and variances are computed over all environmental treatments for which
there were complete data.
bp (variance ratio F) < 0.01.

K
832 Giesel

average ages at reproduction were calculated and then collected over the
treatments in which all gentotypes were represented, to form genotypic means
and variances over all experimental treatments. Results of this treatment of
the data are presented in Table VI. They are remarkably clear. There are no
genotypic differences in either average egg production per day or average
age of reproduction. However, genotypic differences in variance of these mean
values engendered by experimental conditions are striking and highly signi-
ficant (F ratio, P<0.01). These data suggest, more clearly that any other set
presented here, that the heterozygotes are buffered against variation in
trophic conditions.

DISCUSSION
This set of experiments, designed to evaluate fitness differentials among
genotypes of Zw, demonstrated that Zw heterozygotes are more "fit" under
conditions of trophic deprivation than either Z w homozygote. This fact is
particularly well demonstrated by the third set of experiments, in which the
development times and fecundity distributions of genotypes could be unam-
biguously compared. The first two sets of data, which showed large excesses
of the Z w heterozygote phenotype, share a common ambiguity; Z w homo-
zygotes can, depending on genetic background and trophic conditions,
exhibit phenotypes approaching and sometimes indistinguishable from those
of Z w heterozygotes (Giesel, in preparation). Thus the heterozygote excess
which was noted to increase with degree of trophic deprivation may be due
to two factors. Higher viability of heterozygotes is one, and a set of regulatory
phenomena producing heterozygotelike homozygote phenotypes is the other.
Nevertheless, the fitness polymorphism shown here is inescapable.
The fitness differentials are most likely to be due to the heterozygote's
greater ability to utilize the products of both structural forms of G6PD,
which have quite different substrate and cofactor kinetics, "on demand."
G6PD S is the more efficient utilizer of low G6P levels, and G6PDf is less
adversely affected by low levels of cofactor (see Steel et aL, 1968; Komma,
1968). Thus the Zw heterozygotes, which run both structural forms of G6PD
simultaneously (at least more so than do Z w homozygotes), should more
readily and more efficiently metabolize low levels of dietary lipid and low
levels of dietary glucose than either Z w homozygote.

REFERENCES
Brewer, G. J. (1970). An Introduction to Isozyme Technique, AcademicPress, New York,
186 pp.
Komma,D. J. (1968). Glucose6-phosphate dehydrogenasein Drosophila: A sex influenced
electrophoretic variant. Biochem. Genet. 1:229.
Biology of Duplicate Gene System 833

Somero, G. H., and Hochachka, P. W. (1973). Strategies of Biochemical Adaptation,

Saunders, Philaddphia.
Steel, M. W., Young, W. J., and Childs, B. (1968). Genetic regulation of glucose 6-phosphate
dehydrogenase in Drosophila melanogaster: Starch gel electrophoretic variation due to
molecular instability. Biochem. Genet. 2:159.
Young, W. J., Porter, J. E., and Childs, B. (1964). Glucose 6-phosphate dehydrogenase in
Drosophila: X-linked electrophoretic variants. Science 143:140.