When houseguests become parasites: Sympatric in ants

Stewart H. Berlocher* Department of Entomology, University of Illinois at Urbana–Champaign, 320 Morrill Hall, 505 South Goodwin Avenue, Urbana, IL 61801

pecies are part of the common related , and we are finally, if still argue that was not coinage of . Taxonomists dimly, seeing some of the specific genes limited to a handful of special cases, but name them, developmental biolo- that underlie (3). was quite common. Bush was driven to gists deconstruct them, physiolo- But fundamental questions remain his stance by the observation that one S unanswered. A critical question is how gists compare them, ecologists count could find many groups of closely re- them, conservation biologists conserve reproductive isolation could evolve in lated, sympatric species that use differ- them, and evolutionary biologists study the face of ; random mating ent ecological niches. Conspicuous their multiplication and . It within a population and gene flow be- among these groups are parasites of ani- may be that only the individual is a tween neighboring populations are enor- mals and plants, which are often highly more important biological unit than the mously powerful homogenizing forces. specialized ecologically and frequently species. It is thus no surprise that Dar- How could selection tear a single popu- mate on the , a factor that poten- win named his great work On the Origin lation into two reproductively isolated tially links any to a new host of Species, nor is it a surprise that many with a reduction in gene flow between evolutionary biologists today concen- the new and ancestral populations. trate their efforts on understanding spe- We might expect Another group of organisms that has cies formation, or speciation. Riitta attracted undying speculation about Savolainen and Kari Vepsa¨la¨inen (1), in to see speciation in sympatric speciation are the social para- a study in this issue of PNAS, have con- sites in the Hymenoptera (7). Social tributed significantly to a long-standing response to humanity’s parasites of the wasps, ants, and bees debate in , over have fascinating natural histories. Of the whether geographic isolation is a neces- modification of usual social insect divisions into female sary first step in speciation, in a study of reproductives (queens), males, and non- social parasites in Myrmica ants (2). The the globe. reproductive females (workers), only resolution of this argument has impor- two occur in social parasites: female re- tant consequences beyond evolutionary productive and males (7). Workers are biology. If species do not have to wait species? For Darwin, who had an over- often totally absent. Social parasite spe- for the next geological event to speciate, riding faith in the power of natural se- cies live by tricking workers of the host then speciation could conceivably be lection, this was not a difficulty. But species into feeding themselves and their very rapid. This, in turn, means that the with the mathematical sophistication progeny. They are the ultimate bad process might be conducive to study in that the new science of population ge- houseguests. Social parasites occur in real time. We might expect to see spe- netics brought to the ‘‘modern synthe- most groups of social Hymenoptera, in- ciation in response to humanity’s inces- sis’’ of the 1930s (4) came an under- cluding bumble bees, several wasp sant modification of the globe. Anthro- standing of just how strong selection groups, and many ant groups. pogenic change might drive not only had to be to overcome gene flow. The How does social parasitism involve such well known as the evo- solution that , Theodosius sympatric speciation? The entomologist lution of insecticide, herbicide, and anti- Dobzhansky, and others (5) proposed is Emery in 1909 noted that social para- biotic resistance, but also, for example, that geographic barriers could break the sites almost always parasitized close rel- speciation of herbivorous insects adapt- chain of gene flow between populations. atives, an observation now known as ing to plants moved to new parts of the If two populations exchange no genes ‘‘Emery’s rule’’ (7). Others in the de- world. because of geographic isolation, then cades since (7) have been entranced by To fully understand Savolainen and these populations can evolve completely the idea that the social parasites could Vepsa¨la¨inen’s results, some history is in independently, adapting to new chal- evolve from their hosts. The most often order. In the 144 years since publication lenges until genetic reproductive isola- proposed scenario is from a of , enormous tion evolves as an evolutionary byprod- state of polygyny (multiple queens) progress has been made in illuminating uct. The two new species that have within a species that all share equally in the things we need to know to under- evolved in geographic isolation (or allo- worker production, to polygyny with stand speciation (2). We now know a lot patry) can later coexist in the same area some queens ‘‘cheating’’ to increase about geographic isolation and the bio- (in ). Geographic (or allopat- their fitness by producing mostly repro- geography of closely related species. We ric) speciation is such an obvious solu- ductives, to among know the kinds of features that prevent tion to the problem of evolving repro- cheaters, to finally a new parasitic spe- mating between closely related species ductive isolation, and the evidence for it cies reproductively isolated from its an- or, in other words, that produce ‘‘repro- so great (5), that it became the only ac- cestor (1, 7). Such a sympatric process ductive isolation.’’ We understand that cepted mode of speciation (excepting would produce species pairs that are many close species differ in ecology, and speciation via and a few sister species (nearest relatives), which we appreciate the consequences of lack other similar mechanisms, which must of ecological differences. We know a occur in sympatry). great deal about how to count differ- But a few evolutionary biologists, such See companion article on page 7169. ences between the genomes of closely as Guy Bush (6), started in the 1960s to *E-mail: [email protected].

6896–6898 ͉ PNAS ͉ June 10, 2003 ͉ vol. 100 ͉ no. 12 www.pnas.org͞cgi͞doi͞10.1073͞pnas.1332533100 Downloaded by guest on September 30, 2021 COMMENTARY

should pair together in a . But in evolutionary biology it can be a long, rocky road from scenario to proof. ‘‘Proving’’ sympatric speciation has been a particularly tough proposi- tion. The only potentially simple cases involve divergence in historical time (8, 9). For example, the host of the fruit fly Rhagoletis pomonella arose from the ancestral, native, hawthorn- infesting population in Ϸ140 years ago, and has evolved sub- stantial reproductive isolation from the ancestral race. The catch here, however, Fig. 1. (A) Species tree (blue area) and tree (lines). Sampling just one allele per species reveals only is that reproductive isolation between part of what could be known about a speciation (red allele branches, or ), whereas sampling the races is only partial; they are not extensively (red ϩ light brown clades) reveals much more about timing, geographical patterns of events, generally accepted as “complete” etc. Dark brown extinct clades are unknowable. (B) Allele tree for a host (red) and parasite (green) that species. would support sympatric speciation, because the parasite is not only sister species to the host but also For the vast majority of cases, the hy- originated later than the host (parasite originate at a later time). (C) A pattern that has been pothesis of sympatric speciation must observed for mitochondrial genes implying both interspecific gene flow and rapid evolution, possibly stand or fall based on inferences about because of ‘‘selective sweeps’’ of selected through the species (tree labeled as in B). the evolutionary past. This is certainly true for hymenopteran social parasites; history of the Myrmica social parasites ever since the pioneering work of Avise it is unlikely that we will see the evolu- would facilitate preferential mating be- and colleagues (10) in the late 1970s, tion of new ones in historical time. What kinds of observations can be used tween newly originated parasites (1). very intricate and informative genealogi- to test the competing hypotheses of Using the powerful tools of modern cal relationships of haplotypes (or al- sympatric and in phylogeny estimation based on molecu- leles) exist within species trees. Such a the past? Fundamentally, they are (i) lar data, it is now possible to infer phy- ‘‘gene tree’’ within a ‘‘species tree’’ is the types of intermediate stages seen logenies with some confidence. As shown in Fig. 1A. When only one allele and (ii) the phylogenetic relationships Savolainen and Vepsa¨la¨inen point out, a is sampled per species or population, as observed in the relevant populations relatively simple and straightforward is the case for most of Savolainen and and species. inference is that if the parasites and Vepsa¨la¨inen’s work, then one can see Over many decades, Ernst Mayr has hosts each form a monophyletic group, only allele clades like those shown in pointed out that a parasite could adapt then sympatric speciation can be cleanly red in Fig. 1. Such sampling is suitable to a new host in a geographically iso- rejected. That is, if one finds that the for determining relationships at the spe- lated population, evolve reproductive social parasites are all grouped together cies and higher level. However, other isolation there, and then come back into on one branch or of a phyloge- extant lineages, like those shown in light sympatry with the original form (5). netic tree, and their hosts are all brown, could also tell us much about the Mayr’s allopatric hypothesis for the exis- grouped on another branch, then the history of speciation events. Fig. 1B pro- tence of sympatric, specialized species parasites cannot have evolved from the vides a hypothetical example of how the thus requires allopatric intermediate hosts. This is, in fact, the exact finding study of gene trees, which would be pos- populations in varying degrees of adap- of several studies of social parasites and sible with greater sampling than done in tation to a new host; the competing their hosts in wasps and bees (see ref. 1 the present Myrmica study, could pro- sympatric hypothesis requires sympatric and references therein), in which the vide a powerful test of sympatric specia- ‘‘host races’’ in varying degrees of adap- ‘‘strict form’’ of Emery’s rule is violated. tion. Here, all of the parasite alleles be- tation to new hosts (8). For social para- In Myrmica, however, Savolainen and long to one clade (are monophyletic). sites, the corresponding allopatric sce- Vepsa¨la¨inen did find evidence for host– But the host alleles are not monophy- nario has had the parasite attaining full parasite sister species pairs. Using trees letic. Although they all have a common species status in allopatry, then becom- built with sequences for three mitochon- ancestor, the parasite alleles are also ing sympatric with the host, and then drial genes, they have shown that two of derived from one of the clades (so that evolving into a social parasite, leaving three Myrmica social parasites are in the host is said to be paraphyletic). its old free-living lifestyle behind (1, 7). fact sister species of their host species, What would such a finding tell us? It The allopatric explanation thus requires and thus follow the strict version of Em- would tell us that the host was older that one should, if enough groups are ery’s rule, whereas the third apparently than the parasite, which is critically im- studied, see all possible intermediate parasitizes several close relatives, and portant because this is what the sympat- forms, such as sister species that have thus supports at least the ‘‘loose version’’ ric hypothesis demands. just come into sympatry and are par- of Emery’s rule. Altogether, Savolainen But there is another complication tially adapted to parasitism. Although and Vepsa¨la¨inen make a good case for here. As Savolainen and Vepsa¨la¨inen more extensive searching needs to be sympatric speciation in Myrmica from the point out, interspecific gene flow (or done, such intermediates have not been standpoint of both intermediate stages, more precisely passage of entire mito- reported for Myrmica (1) or for ants in and from phylogeny. chondria across species lines) is a signif- general (7). Savolainen and Vepsa¨la¨inen In the near future we can look for- icant problem for testing the sympatric also point out that social parasitism and ward to even more thorough tests of the vs. allopatric hypotheses. Based on stud- its hypothesized precursor state of po- allopatric and sympatric hypotheses for ies in Drosophila and other organisms lygyny are both common in the tribe to the evolution of social parasites in Hy- (ref. 11 and references therein), mito- which Myrmica belongs, and that the menoptera. As has been appreciated chondrial gene trees from closely related

Berlocher PNAS ͉ June 10, 2003 ͉ vol. 100 ͉ no. 12 ͉ 6897 Downloaded by guest on September 30, 2021 species may look like those in Fig. 1C. Sequencing independent nuclear genes Lamarckian inheritance as to be a In this case, neither of the species is would address this possibility. staunch consolidator on Darwinian se- monophyletic, or even paraphyletic, and One of the more conspicuous divi- lection. But there is little doubt that the moreover the sequences are quite simi- sions among evolutionary biologists is most effective challengers, like Darwin lar. This pattern seems to be caused by between what could be called the con- himself, are driven by a core of facts a combination of interspecific gene flow solidators and the challengers. Consoli- that keep the challenge from being just and ‘‘selective sweeps,’’ in which a single dators wish to bind those things about an exercise in rhetoric, that keep the strongly selected base position can drag evolution which are undeniably true, challenge vital. Sympatric speciation is a an entire, nonrecombining mitochon- such as , , challenge to conventional thinking that will not die, because the challengers drial sequence to high frequency. In and allopatric speciation, into larger ex- now have the support of a growing base some cases, species trees based on mito- planatory frameworks that organize as of empirical support (ref. 8 and refer- chondrial DNA may be positively mis- much as possible of the spectacular ences therein). When the idea of sympa- leading (11). As Savolainen and Vep- complexity of nature. Challengers (oth- tric speciation is combined with the sa¨la¨inen (1) discuss, the sister pairing of erwise known as troublemakers) are ir- great font of evolutionary ideas that the English sample of the parasite Myr- revocably drawn to ideas that, at least in arise from the study of social insects, as mica microrubra and its host M. rubra, the short run, threaten consolidation. Of Savolainen and Vepsa¨la¨inen’s contribu- and the corresponding pairing of host course, this is simplistic; a challenger tion (1) makes clear, we should expect and parasite in Finland, may indicate with respect to sympatric speciation both scientific challengers and challeng- gene flow between host and parasite. might be so threatened by, say, ing research for many years to come.

1. Savolainen, R. & Vepsa¨la¨inen, K. (2003) Proc. S. H. (Oxford Univ. Press, Oxford), pp. 3–15. 9. Feder, J. L. (1998) in Endless Forms: Natl. Acad. Sci. USA 100, 7169–7174. 5. Mayr, E. (1963) Animal Species and Evolution Species and Speciation, eds. Howard, D. J. & 2. Howard, D. J. & Berlocher, S. H., eds. (1998) (Harvard Univ. Press, Cambridge, MA). Berlocher, S. H. (Oxford Univ. Press, Oxford), Endless Forms: Species and Speciation (Oxford 6. Bush, G. L. (1969) Evolution (Lawrence, Kans.) 23, pp. 130–144. Univ. Press, Oxford). 237–251. 10. Avise, J. C. (1994) Molecular Markers, Natural 3. Ting, C.-T., Tsaur, S.-C. & Wu, C.-I. (2000) Proc. 7. Ho¨lldobler, B. & Wilson, E. O. (1990) The Ants History, and Evolution (Chapman & Hall, New Natl. Acad. Sci. USA 97, 5313–5316. (Harvard Univ. Press, Cambridge, MA). York). 4. Berlocher, S. H. (1998) in Endless Forms: Species 8. Berlocher, S. H. & Feder, J. (2002) Annu. Rev. 11. Machado, C. A. & Hey, J. (2003) Proc. R. Soc. and Speciation, eds. Howard, D. J. & Berlocher, Entomol. 47, 773–815. London B, 10.1098րrspb.2003.2333.

6898 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.1332533100 Berlocher Downloaded by guest on September 30, 2021