Polymorphism in the Freshwater Isopod Asellus Aquaticus (L)

Heredity 58 (1987) 69—73 The Genetical Society of Great Britain Received 19 March 1986

Components of fitness and the PGI polymorphism in the freshwater isopod Asellus aquaticus (L). 1. Fecundity selection

A. F. Shihab* and Department of Biology, University of Essex, D. J. Heath Colchester, England. Differences in reproductive output were studied between genotypes in a population of the isopod Asellus aquaticus polymorphic at the PGI locus. Reproductive output was measured both in terms of the numbers of eggs and the number of offspring per female. Significant differences were found between the three most common genotypes (3/3, 2/3, 3/4) for both measures. 3/3 had the highest reproductive output. Relative to this the output of 2/3 was 65 per cent and of 3/4 only 40 per cent. This pattern was repeated in three separate years.

INTRODUCTION alleles at this locus, the frequencies of which seem to be correlated (both spatially and temporally) Isopodsare useful organisms for studying the with water temperature and oxygen concentration. dynamics of genetic variation in natural popula- The work presented here formed part of a detailed tions. This is because they allow the investigator 3 year study into the dynamics of this variation in to follow the sampling schedule devised by Chris- one population and describes differences in repro- tiansen and Frydenberg (1973) designed to detect ductive output between genotypes. Further papers fitness differences between allozyme genotypes in will consider aspects of sexual and zygotic natural populations during the reproductive phase selection. of the life cycle. Because isopods are viviparous it enables the reproductive output of females of known genotype to be compared; where the species METHODS has an annual life cycle and breeds only once this is equivalent to an estimate of lifetime reproductive Fulldetails of the methods are given by Shihab output. Using samples of fertilised females, sterile (1985) so only an outline is given here. The popula- females, males and random offspring allows a tion selected for sampling occurred in a freshwater detailed analysis of sexual and gametic selection lake on the University of Essex campus (map ref. (Christiansen and Frydenberg, 1973). In species TM168 031241). Samples of isopods were taken that display precopula (Ridley, 1983) an indepen- with a fine mesh net every two weeks from the dent analysis of sexual selection can be achieved, same area of the lake, from 15 March 1982 to 30 a refinement not available in an organism such as August 1984. Four alleles at the PGI locus were Zoarces viviparus (Christiansen, Frydenberg and detected by horizontal starch gel electrophoresis. Simonsen, 1973). These alleles were numbered 1, 2, 3, 4 in order of The freshwater isopod Asellus aquaticus has an increasing mobility. Of the 16 possible genotypes annual life cycle in most of Britain (Shibhab, 1985) three predominated, 3/3 (frequency between 0'60- and has a precopula making it an almost ideal 075), 2/3 (0.10-0.30), and 3/4 (005-020). organism for these studies. Verspoor (1983) Another three were less common, 1/3 (0.07—0.01), demonstrated that this species is polymorphic for 2/4 (0.01-0.04), and 2/2 (0.02) and the remainder the enzyme phosphoglucose isomerase (PGI, were very rarely encountered. There were two dis- EC5.3 1.9). There are four relatively common tinct breeding seasons each year. In the first * Departmentof Biology, College of Science, University of ovigerous females, mating animals and females Basrah, Basrah, Iraq. carrying young were found from January-May. 70 A. F. SHIHAB AND D. J. HEATH

The offspring were released in June, the parents gave rise to the late summer cohort were of smaller then dying. These offspring comprise the early average size and had lower egg production than summer cohort (e.s.c.), a few of which grow those that bred earlier, but 3/3 still produced more sufficiently to breed from late July—October, pro- eggs. When individual pairs of mean values are ducing the late summer cohort (l.s.c.) (Shihab, compared 3/3 consistently produced more eggs 1985). than 2/3 and 3/4 but there were no detectable To estimate the mean number of eggs carried, differences between the latter two genotypes. In ovigerous females were measured (body length) 1982 and 1983 3/3 produced significantly more the eggs dissected out and counted and the female eggs than 1/3 and also 2/4 (in 1983) in the early electrophoresed. To estimate fecundity (mean summer breeding but not in the late summer breed- number of offspring released), individual fertilised ing. However sample sizes of these rarer genotypes females captured in the field were isolated in petri were small. dishes of lake water with decaying leaves for food. The number of young released was counted, the (b)Mean offspring number (fecundity) female was then measured and genotyped. To investigate possible paternal effects on female fer- Table1(b) shows a similar pattern to the previous tility mating pairs (in precopula) were collected one with significant differences in the mean num- from the field and placed in petri dishes as before. bers of offspring of different genotypes every year After offspring release both parents were elec- but no differences in mean female size. Once again trophoresed and numbers of offspring counted. the 3/3 genotype produced the most offspring Mean numbers of eggs and young for females of (about 80, in the early summer) compared to the different genotypes were compared using single other genotypes which produced between 30 and classification analysis of variance (Sokal and 60. Comparison of pairs of means shows that the Rohlf, 1981). When the tests yielded significant differences between 3/3 and 2/3, 3/4 and 1/3 were differences, a posteriori, unplanned comparisons significant although there were no detectable among means were made by the GT2 method differences between the latter three genotypes. (Sokal and Rohlf, 1981). Similar tests were There is a suggestion that both 2/4 and 2/3 had employed on mean female body length since both higher fertilities than 3/4 and 1/3 but this was not fertility and fecundity in Asellus are positively significant. Late summer breeding females were correlated with body size (Ridley and Thompson, once again smaller and produced fewer offspring 1979). than early summer breeders but again 3/3 did significantly better than 2/3 and 3/4. Mean offspring numbers for a given genotype were RESULTS always lower than mean egg numbers. The figures imply that up to 50 per cent of the eggs carried by Theseare presented in summary form in table 1(a) a female fail to develop into offspring. egg number, 1(b) offspring number and 1(c) male effects; full data are given by Shihab (1985). (c)Male effects Onlyone female genotype (3/3) was sufficiently (a)Mean egg number common to provide enough data for this analysis. Datawere collected from the females which pro- Each column in table 1(c) gives the average num- duced the early summer cohort in all three years, ber of offspring (o) produced by 3/3 females found but the later, smaller breeding season was only in precopula with males of each genotype. Only analysed in 1983. In all three years and for both the data for 1983 showed significant heterogeneity, breeding seasons (in 1983) the results were con- with females fertilised by 3/3 males producing the sistent. The mean numbers of eggs produced by most offspring, significantly more than those fer- different genotypes were significantly different but tilised by 2/3 males, The direction of this difference this could not be attributed to differences in female was the same in 1982, but not significant, possibly due to the smaller sample size. body size between genotypes. The general pattern was also the same with the 3/3 homozygote consistently producing the highest numbers of eggs DISCUSSiON (about 120, on average) compared to the other common genotypes (2/3, 3/4, 1/3, 2/4) which Thereare marked, consistent, differences between produced around 70-90 eggs. The females which genotypes in fecundity. The fitness of 2/3 and 3/4 PGI POLYMORPHISM IN ASELLUS 71

Table IEggnumberand offspring number (fecundity) of PGI genotypes. e =averageno of eggs, o =averageno of offspring, =averagebody lengthoffemale,n=sample size, bpm=broadpouchmortality(%), l.s.c. =latesummer cohort,e.s.c.=early ** summer cohort, Fvariance ratio,d.f.= degrees of freedom, p<001, p<0001.Meanvalues connected by horizontal barsare significantlydifferent. For tables 1(a)and1(b)thevertical columns labelled genotype refertofemales. For table 1(c) the columns refer to males (see text for explanation)

Genotype 2/4 2/3 3/3 3/4 1/3 2/2 F d.f.

(a) Egg Number

1982 e 10442 92161 79120 — 8.47*5* 4, 124 12050 76100 e.s.c 1 921 888 917 886 877 — 097 n 7 18 78 15 11

1983 e 7518 7945 11898 7856 7666 — 1507*** 4,113 e.s.c 1 827 872 908 912 905 234 n 11 20 62 16 9

1983 e — 3390 5350 3283 3266 3550 643*5* 4,42 l.s.c 1 515 527 516 550 537 031 n — 10 24 6 3 4

1984 e 8277 85178 124120 6858 6987 1822*** 4, 101 e.s.c 1 866 886 903 888 875 047 n 9 19 53 17 8 (b)Offspring Number

1982 o 5800 48126 7995 32162 2750 11.705*5 4,104 e.s.c I 925 855 852 853 856 063 n 4 14 62 16 8 bpm 4445 4788 3365 5881 6381

1983 o 5933 52100 8184 39116 3900 17.94*** 4,98 e.s.c I 883 838 863 850 861 039 n 6 17 53 18 9 bpm 2108 3455 3112 5015 4912

1983* o — 29100 472 26140 2600 7.04*5 3,29 l.s.c I — 535 530 5.30 516 016 n 7 18 5 3 bpm — 1445 1454 1958 2678

1984 o 5360 55105 831 33175 3257 18.44*5* 4,91 e.s.c 1 880 857 859 843 850 007 n 5 19 49 16 7 bpm 3524 3582 3252 5078 5338 (c)MaleEffect 1982 o 400 585 647 536 515 238 4,46 e.s.c n 2 10 32 6 1

1983 43.7 5211 7113 524 606 510 3•56 5,156 e.s.c n 4 26 106 18 6 2

relative to 3/3 are about 065 and O4 respectively, scarce. Marinkovic and Ayala (1975) showed that Such large fitness differences arising from fecun- one of the homozygotes at the esterase-5 locus in dity selection between allozyme genotypes have Drosophila pseudoobscura had a fecundity of about never been reported although information is 90 per cent compared to the two other genotypes. 72 A. F. SHIHAB AND D. J. HEATH

A similar difference was observed between octanol criminate between these hypotheses on the avail- dehydrogenase genotypes but not between MDH able data but they could all affect the proportion genotypes. Edwards and Heath (1983) reported of eggs which successfully develop in different that genotypes at the phosphoglucose isomerase genotypes and thus contribute to brood pouch locus in the estuarine isopod Sphaerona rugicauda mortality. differed in fecundity; fast homozygotes and In the absence of some disadvantage this strong heterozygotes had a fitness of about 08 compared directional selection in favour of 3/3 should lead to slow homozygotes. Sassaman (1978), on the rapidly to a population monomorphic for the 3 other hand detected no fecundity differences allele yet all the indications are that genotype between LDH genotypes in the woodlouse frequencies have been stable over the past 3 years Porcellio scaber, and the same was true of the (Shihab, 1985). Fitness differences between some investigation by Christiansen et a!. (1973) of of the other genotypes in survival and mating esterase genotypes in the fish Zoarces viviparus. ability have been clearly demonstrated (Shihab, Part of the fecundity difference observed here 1985) hut no disadvantage of the 3/3 genotype is clearly due to differences in the numbers of has been found. Its frequency in the population unfertilised eggs carried by the different genotypes. increases to a maximum of 75 per cent during the The differences in both cases are in the same direc- production of the early summer cohort (as the tion and of the same order of magnitude. In all fecundity differences would predict). The fre- cases females carry more eggs than they produce quency then falls to 60 per cent in December offspring implying that not all eggs develop suc- and January suggesting that it has a lower winter cessfully. This brood pouch mortality has been survival but there is no experimental evidence described previously in Asellus (Steel, 1961) and for this. is a common feature of isopod reproductive bio- An additional interesting observation from logy (Heath and Khazaeli, 1985) although it is not table I concerns the 2/2 genotype (frequency 002— clear whether it arises pre- or post-fertilisation or 003). This does not appear amongst the females whether it results from expulsion or resorption of which contribute towards the early summer cohort reproductive products. Clearly if there are differen- in 1983 despite the relatively large sample sizes ces in the rate of brood pouch mortality between but is proportionately over-represented in females genotypes they will modify fecundity differences producing the late summer cohort. Its rarity in due to differential egg production. There is some early breeding animals is supported by an indepen- evidence of this from table 1. 3/3 homozygotes dent analysis of sexual selection (Shihab, 1985). always have the lowest brood pouch mortality, 2/2 is less common than expected among animals followed by 2/4 and 2/3, then 3/4 and 1/3, (Brood mating to produce the early summer cohort. The pouch mortality is simply calculated as the average data presented here would suggest that its disad- number of offspring produced by a genotype vantage in not reproducing early in the year may expressed as a percentage of the average number be counterbalanced by an above average ability of of eggs carried by that genotype and subtracted 2/2 juveniles to grow and reproduce later in the from 100.) year in which they were born. Additional evidence for this comes from the Obviously the data presented here do not iden- limited data on male effects which showed that tify the target of selection as variation at the PGI 3/3 females mated with 3/3 males produced the locus rather than variation at a linked locus. This most offspring. A similar effect was described by can only be resolved by the direct approach as Verspoor (1983) who suggested that the reproduc- employed by Hilbish and Koehn (1985), Watt tive output of pairs where the male was 2/2 could (1983) and Burton and Feldman (1983) where in be 30 per cent higher than those where the male vitro differences in allozyme biochemistry are was 2/3 or 3/3. Since effects of maternal genotype linked to in vivo differences in biochemistry, phy- on egg numbers are controlled for, the subsequent siology and fitness. There is in the literature a differences in fecundity must have other causes. growing body of evidence showing both that PGI Such effects could arise because the paternal allozymes vary in their in vitro properties e.g. genotype (diploid) affects fertilisation success by Hoffman (1981), Hall (1985) and that these better "outfitting" of sperm prior to ejaculation differences can translate into in vivo differences in (Gilbert and Richmond 1982) or because the fitness e.g., Watt (1983). This is not surprising given haploid sperm genotype affects fertilisation the role of PGI in glycolysis and makes it not efficiency. Finally the diploid zygote genotype may unreasonable to posit a direct causal connection affect prenatal viability, It is not possible to dis- in Asellus. PGI POLYMORPHISM IN ASELLUS 73

REFERENCES

BURTON, R. S. AND FELDMAN, M. w.1983.Physiologicaleffects HOFFMAN, R. .i. 1981.Evolutionary genetics of Metridium senile of an allozyme polymorphism: glutamate—pyruvate trans- I. Kinetic differences in phosphoglucose isomerase allozy- aminase and response to hyperosmotic stress in the mes. Biochem. Genet., 19, 145-144. copepod Tigriopus californicus. Biochem. Genet., 21, 239— MARINKOVIC,D. AND AYALA, F. J. 1975.Fitness of allozyrne 251. variants in Drosphila pseudoobscura. II. Selection at the CHRISTLANSEN,F. B. AND FRYDENBERG, o.1973. Selection Est-5, 0db and Mdh-2 loci. Genet. Res., 24, 137-149. component analysis of natural polymorphisms using popu- RIDLEY,M. 1983.The explanation of organic diversity. Claren- lation samples including mother-offspring combinations. don Press, Oxford. Theor.Pop.Biol., 4, 425-445. RIDLEY,M. AND THOMPSON, D. i. 1979.Size and mating in CFIRISTIANSEN, F. B., FRYDENBERG, 0. AND SIMONSEN, V. Asellus aquaticus (Crustacea: Isopoda). Z.Tierpsychol.,51, 1973.Genetics of Zoarcespopulations.IV. Selection com- 3 80-397. ponent analysis of an esterase polymorphism using popula- SASSAMAN,C. 1978.Dynamics of a lactate dehydrogenase tion samples including mother-offspring combinations. polymorphism in the woodlouse Porcellio scaber Latr: Hereditas, 73, 291-304. Evidence for partial assortative mating and heterosis in EDWARDS,J. P. AND HEATH, D. .i. 1983.Dynamics ofan enzyme natural populations. Genetics, 88, 591—609. polymorphism in the isopod. Sphaeroma rugicauda SHI1-IAB,A. 1985.Dynamics of an enzyme polymorphism in (Leach). 11. Sexual, gametic and fecundity selection. Hered- the isopod Asellus aquaticus. Ph.D. Thesis, University of ity51, 477-486. Essex. GILBERT,D. 0. AND RICHMOND, R. L. 1982.Esterase 6 in SOKAL,R. R. AND ROHLF, F. .j. 1981.Biometry. W. H. Freeman Drosophila melanogaster: Reproductive function of active and Company, San Fransisco. and null males at low temperature. Proc. NatI, Acad. Sci. STEEL,E. A. 1961.Some observations on the life history of USA, 79, 2962-2966. Asellus aquaticus (L) and Asellus meridianus (Racovitza). HALL,.i. o.,1985. Temperature related kinetic differentiation Proc. Zool. Soc. Lond., 137, 71-87. of glucose phosphate isomerase alleloenzymes isolated vERSPOOR,E. 1983.Allozyme frequencies in Western European from the blue mussel, Mytilus edulis. Biochem. Genet. 23, populations of Asellus aquaticus (L) Isopoda and their 705-729. association with water pollution. Bin!. J. Lion. Soc., 19, HEATH,D. J. AND KHAZAELI, A. A. 1985.Population dynamics 275—293. of the estuarine isopod, Sphaeroma rugicauda. Estuarine WATT,w.B.1983.Adaptation at specific loci II. Demographic Coastal and Shelf Sci., 20,105-116. and biochemical elements in the maintenance of the Colias HILBISH,T. J. AND KOEHN, R. K. 1985.The physiological basis PGI polymorphism. Genetics, 103, 691-724. of natural selection at the Lap locus. Evolution, 39, 1309- 13 17.