A TEST FOR REPRODUCTIVE SEPARATION OF ALTERNATE GENERATIONS IN A BIENNIAL SPIDER, ARANEUS DIADEMATUS (ARANEAE, ARANEIDAE)

5 Pages·2002·5.2 MB·English

by  Jes Johannesen

2002. The Journal of Arachnology 30:65-69
A TEST FOR REPRODUCTIVE SEPARATION OF ALTERNATE
GENERATIONS IN A BIENNIAL SPIDER,
ARANEUS DIADEMATUS (ARANEAE, ARANEIDAE)
Jes Johannesen: Institut fiir Zoologie, Abt. Okologie, Universitat Mainz,
SaarstraBe 21, D-55099 Mainz, Germany. E-mail: [email protected].
uni-mainz.de
S0ren Toft: Department of Zoology, University of Aarhus, Bldg. 135, DK-8000
o
Arhus C, Denmark
ABSTRACT. In Denmark, two seemingly distinct size-classes, 3"^ and 4'*’ instar juveniles and repro-
ductive adults, of Araneus diadematus are found during every breeding season in autumn, indicating a
We
non-overlapping biennial life-cycle. tested the hypothesis that alternate generations might experience
a degree of reproductive isolation, using the distribution of nuclear (allozyme) and maternal (mtDNA)
genetic markers. Individuals of a locality behaved as belonging to a random mating population, irrespective
of size. No differences were found between any size-class pairs, within and between 2 yr, or among
geographically distant samples. Processes that may lead to this result are discussed: the biennial devel-
opment may be incomplete; or there may be migrational influx of genes from southern annual populations.
There is no evidence for sexual differences in life-cycle length.
Keywords: Araneae, life history, biennial reproduction, mitochondrial DNA, allozymes
Several species of spiders at northern lati- reach sexual maturity (Toft 1976). In Den-
tudes need an extended time to reach sexual mark, A. diadematus reaches maturity in July-
maturity compared to their more southern August and mates through August and early
conspecifics (Toft 1976 and references there- September. The females lay a single egg sac
in). The ecological background promoting in October. The egg sac overwinters, while the
such a life-cycle shift may result from lower parent generation disappears before winter.
temperature and shorter seasonal length (Roff Reproduction in spring has never been ob-
1992). Life-cycle shifts may also be viewed served. In the breeding season two seemingly
as a strategy to maximize fecundity by avoid- distinct size-classes can be observed: ca. 3*^^
ing intraspecific competition (Polis 1984; Roff to 4“^ instar juveniles (i.e., young of previous
1992). Most recorded life-cycle shifts in spi- year’s reproduction) and sexually mature ani-
ders involve species that extend their annual mals (two years old). This pattern has consis-
life-cycle in south and central Europe to a bi- tently been observed during the past 30
ennial life-cycle in northern Europe (Almquist years(Toft 1976; Toft pers. obs.). Analysis of
1969; Toft 1976). If the biennial development size structure in the field produced no evi-
is perfect, cohorts that mature in even and un- dence for different developmental patterns in
even years will never mix. Dividing conspe- males and females (Toft 1976). It seems as if
cifics into separately reproducing cohorts can populations of A. diadematus have split into
foster sympatric race formation and, ultimate- two non-overlapping cohorts.
ly, speciation. In this manner, the 13-yr and The combination of a narrow breeding sea-
17-yr periodical cicadas have evolved into son and the repeated annual occurrence of two
distinct evolutionary units (Archie et al. 1985; distinct size classes led us to speculate wheth-
Marshall 2000). er A. diadematus experiences some degree of
The life-cycle of Araneus diadematus reproductive isolation, i.e., whether size clas-
We
Clerck 1757 is generally described as annual ses constitute two discontinuous cohorts.
(Bonnet 1930). However, in northern regions tested reproductive separation assuming two
of Europe A. diadematus needs two years to hypotheses. First, reproductive separation can
65
THE JOURNAL OF ARACHNOLOGY
66
A
be estimated as a departure from random mat- ton 1993). total of 121 spiders were inves-
ing between cohorts. This leads to the well- tigated (see Table 1). Three buffer systems
known Wahlund-effect (Wahlund 1928). Fur- were used: Tris-Maleic acid pH = 7.0 (Rich-
thermore, given a 2-yr life-cycle, juveniles of ardson et al. 1986, adjusted with maleic acid
a certain year should correlate genetically with to pH = 7.0 from pH = 7.8) for Aat-2, Mpi,
adults of the following year, relative to adults Pgd. Tris-Citrate pH = 8.2 (Richardson et al.
of the same year. The same reasoning holds 1986) for Gpd, Pgi, Idh. Tris-Glycine pH =
for adults of a certain year and juveniles of 8.5 (Hebert & Beaton 1993) for Adh, Fum,
the next year. Second, a sex-specific devel- Pgm. All enzymes were run at 250V for 30
opmental cycle is possible (cf. Levy 1970) min.
though it seems to be rare, at least in temper- Departure from random mating among ju-
ate species (Schaefer 1987). If, for example, veniles and adults was tested as a departure
females mature in 2 years and males in year,
1 from Hardy-Weinberg proportions for all spi-
no differentiation of allozyme frequencies
ders collected at Mols in the years 1998 and
would be possible. However, if femamlets’DNbiA- 1999, respectively, with the program GENE-
ennial development is complWetee, POP (Raymond & Rousset 1995a). Differen-
might possibly have diverged. tested this
tiation among samples was quantified using
hypothesis using maternally inherited mito-
&
DNA. the ^ST estimator of Weir Cockerham
chondrial
(1984). Fgy is defined between 0 and 1, where
METHODS Fsj = 0 signifies no variance among samples
Juvenile and 4^*’ instars) and adult spi- and Est = 1 signifies that all variance is dis-
ders were collected at Mols, Denmark in the tributed between samples. Fsj-analyses are
autumn of 1998 and 1999 (M98-J(uvenile), based on allele frequencies summed over all
M98-A(dult), M99-J, M99-A) and from the loci. Divergence initially may proceed at a
300 km distant island of Bornholm, near single locus, e.g., under selection, before ge-
Hammern, in 1998 (H98-J, H98-A). Each netic drift leads to divergent allele frequencies
size-class, at each location, was treated as a at other loci (e.g., McKechnie et al. 1975; Jo-
sample. All collections were done in the hannesson et al. 1995). Therefore, we tested
month of September, i.e., at a time when all for allele frequency homogeneity at each lo-
the breeders of the year had matured. Born- cus between all population pairs using exact
holm is an isolated island in the Baltic Sea, tests (Raymond & Rousset 1995b) with GE-
approximately 30 km southeast of Sweden NEPOP. The relationship among juveniles and
and 80 km north of Poland. The Bornholm adults, and geographic locations was investi-
samples were used to analyze the influence of gated applying maximum likelihood analysis
geographic distance relative to genetic diver- with the program package PHYLIP (Felsen-
gence between size groups, under the rationale stein 1993). This algorithim assumes diver-
that if two distinct reproductive groups exist, gence by genetic drift only. Branch lengths are
then correlations within size-groups should be
relative to divergence, and each is tested for
independent of geographic distance.
significance.
All cohort related hypotheses were tested
Einally, we sequenced the mitochondrial
by applying biparental inherited markers (al- DNA gene NDl for 580 base pairs. NDl has
lozymes) and maternally inherited markers
proven to be highly variable in other species
(mitochondrial DNA). Nine polymorphic loci,
of spider (Hedin 1997; Croucher 1998; Johan-
could be constantly scored in all samples: Aat- &
2 (E.C. 2.6.L1), Fum (E.C. 4.2.1.2), Gpd nesen Veith in press, Johannesen pers.
(E.C. 1.LL8), /J/i-7, -2 (E.C. 1.1.1.42), PgJ obs.). Initially, two individuals from H98-J,
(E.C. 1.1.1.44), Mpi (E.C. 5.3. 1.8), Pgi (E.C. and three individuals from all other samples
=
5.3. 1.9), Pgm (E.C. 5.4.2.2). The locus Adh were analysed, 17. Amplification pro-
(E.C. 1.1. 1.1) was also polymorphic but tocol and NDl primers can be found in Hedin
DNA
stained too weakly for juveniles to be ana- (1997). fragments were sequenced in
lyzed, and was omitted from further analysis. both directions. Sequences were aligned using
Electrophoretic analysis was done by cel- the program Sequence Navigator and succes-
&
lulose acetate electrophoresis (Hebert Bea- sively checked manually.
JOHANNESEN & TOFT—BIENNIAL GENERATIONS OF ARANEUS DIADEMATUS 67
—
Table 1. Allele frequencies at nine polymorphic loci for A. diadematus. For locality abbreviations
please refer to the text.
M98-J M98-A M99-J M99-A H98-J H98-A
Fum 1 0.85 0.90 0.89 0.85 0.80 0.70
2 0.15 0.10 0.11 0.15 0.20 0.30
Aat-2 1 0.96 0.96 0.92 1.00 1.00 1.00
2 0.04 0.04 0.08 0 0 0
Gpd 0.21 0.27 0.13 0.25 0.15 0.07
1
2 0.02 0 0 0 0 0
3 0.77 0.73 0.87 0.75 0.85 0.93
Idh-1 0.81 0.81 0.92 0.90 0.80 0.91
1
2 0.19 0.19 0.08 0.10 0.20 0.09
Idh-2 0.04 0.04 0 0 0.17 0
1
2 0.96 0.96 1.00 1.00 0.83 1.00
Pgd 1 0 0 0.13 0.03 0 0.04
2 0 0 0.05 0 0 0
3 0.98 1.00 0.79 0.93 0.90 0.89
4 0.02 0 0.03 0.05 0.10 0.07
Pgi 1 0.02 0.02 0.05 0.03 0 0.04
2 0.62 0.40 0.53 0.53 0.40 0.50
3 0.25 0.44 0.34 0.26 0.45 0.35
4 0 0.04 0 0 0.10 0
5 0.12 0.10 0.08 0.18 0.05 0.11
Pgm 0.02 0 0 0 0 0
1
2 0.87 0.90 0.87 0.77 0.85 0.87
3 0.12 0.10 0.13 0.22 0.10 0.13
4 0 0 0 0 0.05 0
Mpi 1 0 0 0.04 0.06 0.09 0.02
2 0.65 0.26 0.54 0.63 0.50 0.50
3 0.31 0.52 0.32 0.31 0.41 0.35
4 0.04 0.22 0.11 0 0 0.13
N
26 24 19 20 10 22
RESULTS were found for a single deviating sample, rel-
Only a single mtDNA haplotype, thus no ative to all other samples.
mtDNA differences among samples, was ob- Thus, neither hypothesis could be con-
served (Genbank number AY036084). We did firmed. There were no genetic differences be-
tween juveniles and adults of the same year,
not sequence further individuals after this re-
no genetic correlations between alternate gen-
sult.
The allozyme allele frequencies are shown erations, and no differences among females
from different years. Furthermore, there were
in Table 1. There were no allozyme-genetic
no genetic differences between geographically
correlations between alternate generations, nor
were there deviations from Hardy-Weinberg distant samples.
proportions among generations within the
DISCUSSION
same year at the same locality, P > 0.50. No
allozyme differentiation was found among all This study was inspired by the notion that
samples, = 0.018 ± 0.013, and no internal the recurrent appearance in autumn of two
branches of the maximum likelihood phono- size classes of A. diadematus, coupled to its
gram were significantly greater than zero (tree narrow breeding season, could be related to
not shown). No among-sample allele frequen- reproductively separated cohorts. However, a
cy differences were found in the loci Pgi, lack of differentiation among samples was ob-
Pgm, Idh-1, —2, Aat and Fum. For the loci served for nuclear genetic markers and no var-
Mpi, Pgd, and Gpd allele number differences iation was observed in maternally inherited
THE JOURNAL OF ARACHNOLOGY
68
mtDNA markers. No correlations between the recurrent occurrence of two size classes
size classes or sexes were found. Thus, the suggests strong selection for maintaining bi-
genetic markers were not able to separate the ennial life cycles. The phenomenon of diver-
size groups into two reproductive cohorts. gence of biennial northern and southern an-
A
lack of genetic correlation does not prove nual populations remains an intriguing
that A. diadematus size-classes are, to some question. Immigration from central popula-
extent, reproductively isolated. Lack of genet- tions is thought to impede genetic adaptations
ic correlations can be a consequence of large of peripheral populations (Hoffmann & Blows
population sizes that cause genetic drift to 1994). Given a high dispersal potential, it is
have little power in separating cohorts. How- possible that immigrants from southern pop-
ever, genetic homogeneity at nine allozyme
ulations with an annual life-cycle can hinder
loci suggests that cross-generation matings
separation of distinct cohorts, in that mobility
may not be uncommon. It is unlikely that too- among populations will break down any sign
small sample sizes resulted in a lack of sig-
of intra-population cohort divergence. There-
nificant correlations because alleles of the
fore, the question arises about the extent of
most polymorphic loci, where the sample er-
dispersal mobility of A. diadematus.
ror should be greatest, were all homogeneous-
Reports on ballooning activity of A. dia-
ly distributed. It is more likely that the bien-
dematus are few and inconclusive. During
nial life-cycle of A. diadematus is not m
analysis of seven year’s catches by a 12.2
complete, i.e., that a small fraction of the pop-
high Rothamsted suction trap near Copenha-
ulation (and probably both sexes) cross the
generations by having annual or triannual life- gen, Denmark, Toft (pers. obs.) found four in-
cycles. Such mixed life-cycles are well known dividuals, all 3rd or 4th instar juveniles. Three
in several species of spiders (Toft 1976, were captured in August-October of their first
1983), but they are recognized easily only by year, and one in June of the second year. An-
standard analysis of size distributions if the imals that are caught in this trap are assumed
developmental lines are numerically well rep- to be long-distance dispersers, because the
resented. More extensive sampling or labora- catch is independent of the immediate sur-
tory breeding studies may elucidate this pos- roundings of the trap due to its height. Though
sibility. the number trapped is rather low, the data sug-
Uniform distributions of allozyme alleles gest that dispersal is predominantly in medi-
from two locations 300 km apart indicated um-sized instars, in which both autumn and
that gene flow may occur over large distances. spring dispersal is possible. Relatively, bal-
High mobility by ballooning (Foelix 1996) is looning activity may be higher than the figures
a possible reason for lack of differentiation seem to indicate if related to the populations
both between the geographically distant sam- on the ground. Araneus diadematus is a wide-
&
ples (Ramirez Haakonsen 1999) and be-
spread species, but it occurs only locally in
tween alternate cohorts. Furthermore, espe- high densities. It may thus be sufficiently ac-
cially for mtDNA, lack of variation may be
tive as a ballooner to explain the lack of geo-
influenced by post-glacial colonization due to
graphical differentiation. The zone of transi-
founder events. Colonization is not expected
tion between annual and biennial life-cycle of
to influence the variation of nuclear genes as
A. diadematus is unknown but may be only a
much because the effective population size of
few hundred kilometres south of Denmark.
nuclear genes is four fold that of mtDNA. The
level of allozyme polymorphism in A. diade- Even if it is unlikely that the distance may be
bridged regularly by single individuals, accu-
matus is similar to that found in other spiders
mulated dispersal over some generations may
examined in northern Europe. These studies
were able to separate local populations (Ped- create the gene flow necessary to prevent ge-
ersen & Loeschcke 2001; Johannesen & Veith netic divergence of alternate biennial genera-
2001) and even delimit micro-structures with- tions, even if there are no local “cross-overs”
in populations (Johannesen et al. 1998; Schaf- between alternate generations. In conclusion,
er et al. 2001). we suggest that both intra- and inter-popula-
In the face genetic homogeneity, caused ei- tion gene flow may play a role in preventing
ther by gene flow or a recent life history split, intra-population differentiation.
JOHANNESEN & TOFT—BIENNIAL GENERATIONS OF ARANEUS DIADEMATUS 69
LITERATURE CITED McKechnie, S.W., PR. Ehrlich, & R.R. White.
1975. Population genetics of Euphydryas butter-
Almquist, S. 1969. Seasonal growth of some dune-
living spiders. Oikos 20:392-408. flies. I. Genetic variation and the neutrality hy-
& pothesis. Genetics 81:571-594.
Archie, J., C. Simon, D. Wartenberg. 1985. Geo- &
Pedersen, A. aa. V. Loeschcke. 2001. Conser-
graphical patterns and population structure in pe-
vation genetics of peripheral populations of the
riodical cicadas based on spatial analysis of al-
mygalomorph spider Atypiis affinis (Atypidae) in
lozyme frequencies. Evolution 39:1261-1274.
northern Europe. Molecular Ecology 10:1133-
Bonnet, P. 1930. La mue, Fautonomic et regenera-
142.
tion chez les araignees. Bulletin de la Societe 1
Polls, G.A. 1984. Age structure component of niche
d’Histoire Naturelle de Toulouse 59:237-700.
width and intraspecific resource partitioning: can
Croucher, P.J.P. 1998. Evolutionary interactions of
age groups function as ecological species. Amer-
two colonizing species of large house spider (Ar-
— ican Naturalist 123:541-564.
aneae: Tegenaria Spp.) testing the reinforce- &
Ramirez, M.G. K.E. Haakonsen. 1999. Gene flow
ment hypothesis. PhD thesis, University of York.
among habitat patches on a fragmented land-
Foelix, R.R. 1996. Biology of Spiders. Second Edi-
scape in the spider Argiope trifasciata (Araneae:
tion. Oxford University Press, Oxford.
Araneidae). Heredity 83:580-585.
Felsenstein, J. 1993. PHYLIP (Phylogeny Inference Raymond, M. & F. Rousset. 1995a. GENEPOP
Package) version 3.5c. Department of Genetics,
(V.1.2): population genetics software for exact
University of Washington, Seattle.
& tests and ecumenicism. Journal of Heredity 86:
Hebert, P.D.N. M.J. Beaton. 1993. Methodolo-
248-249.
gies for allozyme analysis using cellulose acetate Raymond M. & F. Rousset. 1995b. An exact test
electrophoresis. Helena Laboratories, Beaumont,
for population differentiation. Evolution 49:
Texas.
1280-1283.
Hedin, M.C. 1997. Speciational history in a diverse Richardson, B.J., P.R. Baverstock, & M. Adams.
clade of habitat-specialized spiders (Araneae: 1986. Allozyme Electrophoresis. A Handbook
Nesticidae: Nesticus): inferences from geograph- for Animal Systematics and Population Studies.
ic-based sampling. Evolution 51:1929-1945. Academic Press, Inc. San Diego.
Hoffmann, A.A. & M.W. Blows. 1994. Species bor- Roff, D.A. 1992. The Evolution of Life Histories.
ders: Ecological and evolutionary perspectives. Chapman & Hall, New York.
TREE 9:223-227.
Schaefer, M. 1987. Life cycles and diapause. Pp.
&
Johannesen, J., T. Baumann, A. Seitz M. Veith. 331-347, In Ecophysiology of Spiders (W. Nen-
1998. The significance of relatedness and gene
twig, ed.). Springer-Verlag, Berlin.
flow on population genetic structure in the sub- Schafer, M.A., A. Hille & G.B. Uhl. 2001. Geo-
social spider Eresus cinnaberinus (Araneae: Er- graphic patterns of genetic subdivision in the cel-
esidae). Biological Journal of the Linnean Soci- lar spider Pholciis phalangioides (Araneae). He-
ety 63:81-98. redity 86:94-102.
Johannesen, J. & M. Veith. 2001. Population His- Toft, S. 1976. Life-histories of spiders in a Danish
tory of Eresus cinnaberinus (Araneae: Eresidae) beech wood. Natura Jutlandica 19:5-39.
Colour Variants at a Putative Species Transition. Toft, S. 1983. Life cycles of Meta segmentata
Heredity 87:114-124. (Clerck, 1757) and Meta mengei (Blackwall,
&
Johannesson, K., B. Johannesson U. Lundgren. 1869) in Western Europe (Arachnida: Araneae:
1995. Strong natural selection causes microscale Tetragnathidae). Verhandlungen des Naturwis-
allozyme variation in a marine snail. Proceedings senschaftlichen Vereins Hamburg 26:265-276.
of the National Academy of Science 92:2602- Wahlund, S. 1928. Zusammensetzung von Popula-
2606. tion und Korrelationserscheinung vom Stand-
Levy, G. 1970. The life cycle of Thomisus onustiis punkt der Vererbungslehre aus betrachtet. Here-
(Thomisidae: Araneae) and outlines for the clas- ditas 11:65-106.
sification of the life histories of spiders. Journal Weir, B.S. & C.C. Cockerham. 1984. Estimating E-
of Zoology London 160:523-536. statistics for the analysis of population structure.
Marshall, D.C. 2000. Reproductive character dis- Evolution 38:1358-1370.
placement and speciation in periodical cicadas,
with description of a new species, 13-year Mag- Manuscript received 1 November 2000, revised 15
icicada neotredecim. Evolution 54:1313—1325. August 2001.