Members of two different species possess a similar-looking structure that they use in a similar

Introduction

Convergent evolution refers to the evolution in different lineages of structures that are similar or ‘analogous’, but that cannot be attributed to the existence of a common ancestor; in other words, the fact that the structures are analogous does not reflect homology. A similarity may reside at the phenotypic level, in which case the lineages share the overt trait, but the underlying DNA sequences are different. Convergent evolution occurs when species occupy similar ecological niches and adapt in similar ways in response to similar selective pressures. Traits that arise through convergent evolution are referred to as ‘analogous structures’. They are contrasted with ‘homologous structures’, which have a common origin. The opposite of convergent evolution is ‘divergent evolution’, whereby related species evolve different traits.

Well-documented cases of convergent evolution of similar DNA sequences are not plentiful; such cases are usually restricted to a few amino acids. Convergent evolution can mislead phylogenetic inference because it mimics shared ancestry. Standard phylogenetic methods are not equipped to differentiate between the two. When convergent evolution is mistaken for homology, this produces a phylogenetic tree that is falsely reticulate or bushy in appearance, that is, species appear to originate from a common ancestor when in fact that is not the case.

Epigenetic Explanation of Evolutionary Convergences

As pointed out earlier, Darwin himself believed in the possibility of convergent evolution based on directed change “without the aid of any form of selection” rather than on random variability (gene mutations in modern terminology):

If the varying individual did not actually transmit to its offspring its newly acquired character, it would undoubtedly transmit to them, as long as the existing conditions remained the same, a still stronger tendency to vary in the same manner.

Darwin (1872)

Evolutionary phenotypic changes in metazoan morphology may imply modification of existing developmental pathways, reactivation of ancestral pathways, or evolution of new pathways. It is important to bear in mind that neither modification of existing pathways nor evolution of new developmental pathways implies obliteration or irreversibility of previous or ancestral developmental pathways; gene products involved in these pathways generally are still present and functionally unchanged. Only the spatiotemporal pattern of their expression is epigenetically changed and regulated.

The epigenetic paradigm would relate the occurrence of evolutionary convergences with the similarity of solutions to problems arising from the adverse effects of environmental conditions, with the limited number of developmental algorithms (signal cascades) and constraints on modification of these algorithms. Since these signal cascades start with neural signals, one basic prediction from the view of the epigenetic paradigm would be that the frequency of evolutionary convergences would dramatically increase with the evolution of the nervous system, coinciding with the Cambrian explosion.

This prediction seems to have been validated in a recent study by Vermeij showing that before the Cambrian explosion, evolutionary convergences have been rare events, but later evolutionary innovations appear repeatedly. Vermeij estimated that only 23.3% of first 56 convergences occurred during the first 2.5 billion years of life on Earth, and 76.7% during the last 0.5 billion years since the Cambrian explosion. Only 4% of the singular nonconvergent innovations have occurred during the last <250 million years, indicating a clear tendency of increased frequency of evolutionary convergences in the course of metazoan evolution (Vermeij, 2006):

If these inferences are correct, they would imply that history during its early phases was substantially more contingent, that is, more dependent on singular circumstances, than are more recent historical episodes. In other words, unique “frozen accidents” were more common in the very distant past than in more recent times.

Vermeij (2006)

No relation seems to exist between frequencies of gene mutations and the increased frequency of evolutionary convergences in metazoans.

What might have determined the abrupt change from the singularity to repeatability of evolutionary innovations after the Cambrian explosion? The increasing trend toward evolutionary convergence in metazoans may be related to the great informational revolution that characterized and enabled their evolution, the advent of the epigenetic information, and the evolution of the nervous system. In contrast to the randomness of the production of the new genetic information via gene mutations, the epigenetic information is not random but is generated in neural circuits. Hence, it is replicable and reemergent. The nonrandomness of epigenetic information may be the causal basis of the ubiquity of the phenotypic convergence observed in the metazoan world.

Glossary

Convergent evolution

Distant or unrelated organisms evolve the same anatomical, morphological body plan characteristics, or ecological function.

The combined plant and animal communities plus their physical environment.

The energy flow, productivity, element cycling, and resilience of ecosystem structure; synonymous with “behavior of ecological systems” and “ecological processes.”

The organisms, their communities, biodiversity, and habitats that comprise an ecosystem.

Polyphyletic suites of species that share ecological characteristics and play equivalent roles in natural communities and ecosystems. Commonly, organisms with convergent anatomical, morphological, physiological, behavioral, biochemical, or trophic characteristics are grouped together.

Organisms that use similar resources in similar ways. Depending on the application, guilds can be synonymous with functional groups.

1.2.9 Application to drug discovery

Convergent evolution from unrelated species has led to extremely varied compositions of venom across a range of species lineages. For drug discovery, this means there is a very diverse compound library of potentially millions of compounds from an estimated 220,000 venomous species currently known [1]. The Australian funnel web spiders have the most complex venoms venom discovered to date, the venom of the Darling Downs funnel web spider (Hadronyche infensa) contains over 3000 different peptides [6]. Cone snails have over 1000 venom peptides [78]. Platypus venom has 83 peptides from 13 different families [48]. In contrast the Northern short-tailed shrew has much less complex venom with only seven proteins isolated so far [5].

So, if venoms cause high mortality and morbidity in so many people why would venom components be considered as potential drugs? Looking at the problem from a pharmacological standpoint the opportunities are obvious, and it is not a surprise that receipt of a random dose of a complex mixture of pharmacologically active compounds is dangerous. The clear action of these compounds led to early experimentation and application as traditional medicines with variable results [79] but now, in the hands of drug discovery scientists, their true potential is being realised.

IV.A. Groupings among Mobile Organisms

Convergent evolution is well-known and documented in the terrestrial realm. Marsupial and placental mammals have converged to similar morphologies and ecological function (Fig. 5). The limited variations on the mammalian body plan are evident in the wolf and catlike carnivores, the arboreal gliders, fossorial herbivores, anteaters, and subterranean insectivores that evolved independently in Australia for the marsupials and on the other continents for the placentals.

Convergent functions can be found among dissimilar-looking organisms. Distantly related marine molluskan herbivores, chitons and limpets, provide an excellent example. Chitons (class Polyplacophora) and snails (class Gastropoda) evolved in the Cambrian at the beginning of the molluskan diversification. The true limpets, Patellogastropoda, evolved from the Archaeogastropoda, much later in the Triassic. However, the radula (the teeth) of these two groups are functionally similar (Fig. 6). Both groups have relatively few teeth that contact the substrate, and those that do are hardened by mineralization of iron or silica compounds (note the black teeth in Fig. 6). Both groups have strong buccal musculature for applying downward forces. Within the molluskan body plan, only chitons and limpets have such a large foot area-to-mass relationship and an excavating-type radula. These morphological and anatomical characteristics, along with their size and mobility, allow species of these groups to specialize on large and expansive macrophytes such as sea grasses, kelp, and encrusting coralline algae. Although species diversity is much higher in other groups of mollusks such as nonlimpet gastropods, it is the functional characteristics of these two groups that make them capable of consuming and even trophically specializing on the tough or limestone-imbedded cells of kelp or coralline algae, respectively.

Convergent functions are numerous among terrestrial organisms. For example, a diversity of flowering plants and their pollinators possess similar morphological and anatomical characteristics despite significant phyletic separation among the plants and the pollinators. The geometry of the flowers, such as the length and width of the floral tube, as well as the placement of nectaries are common features among the plants. The flying characteristics and mouthparts are convergent among the pollinating insects, butterflies, moths, bats, and birds. It is surprisingly easy to find functional similarities among distantly related organisms.

3.1 Conflict Theory

The evolutionary convergence of genomic imprinting in mammals and plants implies that imprinting affects inclusive fitness related to reproductive success. Under the conflict theory of genomic imprinting, the endosperm/placenta, which acquires nutrient from the maternal sporophyte for nourishment of the embryo, is a participant in a conflict of interest among the mother, father, and offspring (Haig & Westoby, 1989). Against a background in which siblings from a single mother are destined to compete against each other, offspring of different fathers would strive to obtain a greater share of maternal resources with no regard to the interests of the mother. In contrast, the mother has an equal interest in all the offspring and strives to secure adequate nutrient for all sibs and for herself. Because of these conflicting interests, the paternal genome attempts to make its possessors larger, while the maternal genome attempts to make them smaller. Genomic imprinting has a predominant role in this process. It has been demonstrated in animals that imprinted genes have a role in brain function, and their effects are often manifested via actions on social behavior throughout an organism’s lifetime (Isles, Davies, & Wilkinson, 2006). Recently, it has been reported that the imprinted gene Meg1 of maize plays a significant role in the maternal nourishment of the embryo and promotes seed growth in a gene dosage-dependent manner (Costa et al., 2012). Thus, in contrast to the expectations of the conflict theory, the maternally expressed imprinted Meg1 gene enhances the growth of endosperm. The applicability of the conflict theory to plants still requires substantiation; certainly, it is known that most plant-imprinted genes have minimal roles in resource allocation. Current thinking holds that the action to optimize resource flow between generations is a part of the driving force to birth and/or to retention of uniparental gene expression patterns.

2 The aaRS conundrum

The convergent evolution of the two unrelated classes of aminoacyl-tRNA synthetases present a baffling evolutionary history. Protein families are defined by shared structural and functional features. Extremely ancient and related biological reactions may be catalyzed by homologous enzymes that evolved from a single ancestor, or by unrelated proteins that converged upon a similar biochemical activity. Convergent evolution generates protein families with common functions whose members need not share the same three-dimensional structure. On the other hand, proteins that share the same fold and carry out the same functions most likely evolved through the duplication of ancestral genes that diverged through mutation and selection [8].

Consider the aminoacylation of tRNAs with their cognate amino acids by the aaRSs. For the vast majority of coded amino acid there is a single, cognate, universal aaRS. Thus, roughly speaking, there are 20 different aaRS. The most parsimonious possibility would be that these enzymes all evolved from a single ancestral protein. Alternatively, if convergent evolution played a role in their evolution, a vast number of molecular ancestors for aaRS would be possible, but this second scenario should not result in any structural or functional correlations between enzymes of different origins. Again, neither of these two possibilities remotely approximates the observed structural and functional relationships among the aaRS [9] (Fig. 1).

All known aaRS evolved from two different ancestral active site domains and are thus classified as class I or class II depending on the fold architecture of these domains [4,5]. Structural and functional considerations show that these catalytic domains evolved from two distinct genetic sequences [10,11]. They have different folds, they recognize different ATP conformations, and they use different mechanisms to attach amino acids to the 3′ end of tRNAs (reviewed in this volume by Hendrickson and Alexander). Thus, aaRSs converged onto the same function from two completely different genetic origins. However, ample evidence indicates that the two classes of aaRSs did not evolve independently, but directed by selection forces acting on both classes simultaneously.

If one only considers the 20 canonical amino acids in the universal Genetic Code, there are 20 aaRS (one per amino acid). Theoretically, the number of extant enzymes originating from each primordial domain could range from 1 to 19. For example, there may be a single class I aaRS and 19 class II aaRSs. In reality, both classes contain a similar number of universally distributed enzymes (11 in class I and 10 in class II), an indication that some coordination occurred between the evolution of both classes. Of course, minor variations exist. For example, the difference in numbers between the two classes is due to the fact that one enzyme (lysyl-tRNA synthetase, or LysRS) can be found with a class I or class II architecture in different species [12] (Fig. 1). Two additional class II aaRS, pyrrolysyl-tRNA synthetase (PylRS) and phosphor-seryl-tRNA synthetase (SepRS) have limited phylogenetic distributions [13–15]. Thus, both aaRS classes are, unexpectedly, of similar size, an evolutionary example of convergence that is likely not coincidental.

Although the internal classification of aaRSs within each class is subject to controversy, to a large extent the organization of the classes is similar, and each can be subdivided into a similar number of subclasses that also hold similar numbers of enzyme [4,5], but see [16]. Again, these similarities could be serendipitous, or they may reflect the existence of common evolutionary constraints acting during the evolution of each aaRS class (Fig. 1A). When the internal organization of the two aaRS classes are compared, a striking level of symmetry between them is apparent (Fig. 1). It is important to notice that this symmetry is not perfect, and can vary slightly when different criteria are used to evaluate the relationships between enzymes [1,4,5,16,17] (Fig. 1). In addition, lateral gene transfer (LGT) events are common among aaRS genes, which increases the complexity of the phyletic relationships within each class [1,16–18]. Thus, in a few cases, the subclass to which an aaRS belongs is controversial, and two cases (LysRS, and glycyl-tRNA synthetase or GlyRS) exist where two independent origins for these enzymes are clear, giving rise to two distinct forms of each of these proteins [12,19]. However, many of the unexpected coincidences between class I and class II aaRS remain in all published analyses, and the nature of the forces that shaped the apparently linked features of the two aaRS classes remains an open question.

The similarities between the two classes extend to the substrate specificities of the enzymes. The largest subgroups within each class are of similar size and recognize polar or apolar branched amino acids. Charged amino acids and their amidated derivatives (Glu, Asp, Lys, Gln, and Asn, with the exception of Arg) are recognized by class Ib or class IIb enzymes (Lys can be used by class Ib or class IIb structures). Aromatic residues are also split between the two classes and belong to smaller subclasses (Fig. 1A).

Class similarities are mirrored in the structure of the Genetic Code, as reported by several authors [20–25] (Fig. 1). Most class Ia and IIa aaRS charge tRNAs with NAN (class I) and NGN (class II) anticodons with polar or branched amino acids. Class I and II aaRS cognate aminoacylate tRNAs with NUN anticodons with charged amino acids. While class I and II aaRS recognize tRNAs with NNA anticodons to charge them with aromatic residues (Fig. 1B). These relationships have been interpreted to indicate that aaRS evolution was shaped by the development of amino acid biosynthetic pathways and the evolution of the Genetic Code [26,27].

The mode of interactions between the two aaRS classes and their tRNA substrates generates an additional surprising observation: While nearly all class I enzymes bind the tRNA on the major groove side of the acceptor stem, class II enzymes typically use the opposite, minor groove side to bind this substrate [6,28,29] (Fig. 3). Remarkably, the aaRSs for aromatic amino acids from class Ic and IIc behave exactly to the contrary: TrpRS and TyrRS recognize their tRNA substrates from the minor groove side, while PheRS does the opposite [30,31].

Thus, class I and class II aaRS are two functionally convergent families of enzymes, whose supposedly independent evolution has generated two groups of proteins that display striking structural and functional symmetries. These parallels strongly support the idea that the evolution of both classes was somehow coordinated, either genetically or structurally, despite the fact that they have different ancestors [32]. The nature of such coordination might explain the surprising features of these enzyme families.

Community Structure

The convergent evolution of ecologically “equivalent” species, such as many Australian marsupials and their placental counterparts in other regions, suggests that independently evolving communities might converge toward similar structure. Certainly, properties such as vegetative morphology and the architecture of vegetation converge; for example, “Mediterranean” vegetation (chaparral, matorral, maquis) is dominated by sclerophyllous, small-leaved shrubs in several parts of the world. Such similarities, however, arise simply from each species' independent adaptation to physical environmental factors. Whether or not community properties such as stability, food-web structure, or species diversity converge as a result of coevolution is a different, much more difficult question. Simple models of interspecific competition and of food webs suggest that the coevolutionarily stable equilibria might be rather few, so that coevolution might yield some predictable structure (MacArthur and Levins, 1964). Whatever the ideal applicability of these models might be, however, the opportunity for such pervasive effects of coevolution may generally be rather slight. Gene flow among conspecific populations that interact with different ensembles of species may prevent finely tuned coevolution. Moreover, paleoecological studies have shown that throughout the Pleistocene, species have had highly individual histories of change in geographic distribution, so that many of today's assemblages of species are very recent. Except for specialized associations, as of host-specific parasites that have moved about with their hosts, there has been little time for coevolutionary adjustments in many of today's species assemblages. Nevertheless, paleontologists have documented rather steady levels of diversity at both global and local levels over vast periods of time (107–108 years), despite turnover of taxa (Jablonski and Sepkoski, 1996). These observations suggest that whatever convergence or constancy of community structure exists may be attributable more to purely ecological rules of community assembly rather than to coevolution.

An unusually clear example of convergent multispecies assemblages is provided by the anoline lizards mentioned earlier, which form monophyletic groups of morphologically and ecologically equivalent species on each of the Greater Antilles islands. However, similar habitats in different parts of the world generally are not very similar in species richness. For example, lizard diversity in Australia exceeds that in corresponding habitats in southern Africa. Nevertheless, in some instances variation in species richness among habitats shows similar patterns; lizards are more diverse in deserts than in wetlands in both Australia and Africa. Thus, habitats seem to have consistent effects on species coexistence, and perhaps on the evolution of resource partitioning (Ricklefs and Schluter, 1993).

Coevolution may also affect community properties such as stability, that is, the property of returning to equilibrium after a disturbance. The study of artificially constructed food webs has identified some patterns of interaction among component species that are conducive to stability. One such aspect is connectance, that is, the ratio of actual to potential interspecific links. High levels of connectance are associated with reduced stability, as a disturbance to one species will affect many other species in the community. Thus processes that reduce connectance, or generate compartments within a community, such as a subset of species that interact predominantly with one another, contribute to stability. To the extent that coevolution between parasites and hosts and between plants and herbivores leads to increased levels of specialization (and thus compartmentalization), it can thus indirectly enhance community stability.

The extensive convergent evolution of subterranean mammals across the planet represents a global, natural experiment of life without light under darkness stress. Adaptive convergence comprises structural and functional reductions and expansions (hypertrophies) through molecular and organismal evolutionary tinkering involving regression, progression, and convergence of eye evolution advanced differentially in subterranean mammals, climaxing in the blind subterranean mole rat, Spalax. The mosaic evolution of the Spalax eye is described in terms of morphology, physiology, and molecular biology. Eye development is initiated normally but at an early stage displays degenerative features – though neurogenesis develops normally. The main regressive feature is the drastic, relative reduction of retinal input to the superior culliculuis. By contrast, the structure of the ‘non-image-forming’ visual pathway involved in photoperiodic perception is well developed in Spalax. The retinal photopigments, coneopsin, rhodopsin, and melanopsin, function in photoentrainment, and the somatosensory brain cortex is remarkably developed. The differential regression and progression of visual structures, photopigments, and brain underwent adaptive evolutionary tinkering through natural selection to life underground at both the molecular and organismal structures.

D Convergent Evolution or Common Ancestry?

In evolutionary biology, convergent evolution is defined as the process whereby distantly related organisms independently evolve similar traits to adapt to similar necessities. VLRs and TCRs/BCRs both serve as antigen receptors, but are evolutionarily unrelated. Thus, the use of distinct receptors in jawed and jawless vertebrates can be regarded as a prime example of convergent evolution.13 However, the overall design of the AIS in jawed and jawless vertebrates seems too similar to be accounted for solely by convergent evolution. Particularly striking is the observation that both jawed and jawless vertebrates have two major populations of lymphoid cells presumed to have similar specialized immune functions145 (Fig. 6). To account for this, it seems more reasonable to assume that VLRA+ cells and T cells evolved from a common ancestor and that, likewise, VLRB+ cells and B cells shared common ancestry; most likely, a common ancestor of all vertebrates was equipped with two lineages of lymphoid cells.7 Recent evidence indicates that, contrary to a commonly held belief, T cells and B cells do not share an immediate common ancestor, but differentiate from myeloid-T and myeloid-B progenitors, respectively.149,150 If T cells and B cells are distantly related as suggested by these studies, it is not surprising if the two lineages of lymphoid cells diverged at an earlier stage in evolution than previously thought.151

In summary, authentic T cells and B cells, as defined by surface expression of TCRs and BCRs, are unique to jawed vertebrates (Fig. 1). However, jawless vertebrates have at least two populations of lymphoid cells that likely share common ancestry with T cells and B cells of jawed vertebrates.