Archaea and bacteria demonstrate evidence of extensive genetic exchange via horizontal gene transfers (Table 1). For example, DeLong [24] reported rDNA sequences characteristic of archaebacteria (i.e., Archaea) in a previously unknown environment for these organisms, that of oxygenated coastal surface waters. This observation generated the following hypothesis: Eubacteria, Archaea and Eukarya “…reside and compete in the ocean’s photic zone under the pervasive influence of light” [25]. This hypothesis leads to several predictions, one of which is that if Archaea and Eubacteria are to compete for light-limited resources, they must both possess genes involved in the utilization of photic energy. It is therefore significant that such genes have been isolated from Archaebacteria and Eubacteria (e.g., [25–27]). Specifically, photoproteins termed proteorhodopsins have been detected [26]. Furthermore, the detection of the shared genetic architecture for utilizing light energy has thus been ascribed to the lateral transfer of genes into the Archaea.

Frigaard et al. [25] also detected evidence for lateral transfer between planktonic bacteria and Archaea as well. This transfer was suggested to be adaptive in nature given that the proteorhodopsin genes isolated from euryarchaeotes were present in those isolates taken from photic, but not subphotic, regions of the water column. This finding was consistent with an adaptive scenario in which the organisms in the light-limited zones gained no benefit from the transfer of the photo-response genes, while those organisms living in the photic zone did [25]. From their findings, Frigaard et al. [25] made the general conclusion that “…lateral gene dispersal mechanisms, coupled with strong selection for proteorhodopsin in the light, have contributed to the distribution of these photoproteins among various [marine] members of all three of life’s domains.”

In addition to the above examples, cyanobacteria have also been shown to possess genetic components most likely resulting from horizontal transfer events. Shi and Falkowski [28]–in an examination of 682 loci–discovered widespread disagreement between phylogenies constructed for 13 cyanobacteria genomes. This discordance among phylogenetic hypotheses was apparently due to large-scale horizontal gene transfer. Indeed, of the 682 orthologs analyzed only 323 were placed into a “core set” whose evolutionary histories seemed congruent [28]. The majority of the loci thus appeared to have been potentially affected by genetic exchange. For example, it was hypothesized that the unique presence of the genetic architecture for nitrogen fixation originated from exchange with a heterotrophic prokaryotic lineage [28]. Likewise, Swingley et al. [29] also discovered evidence for the transfer of the genes (i.e., to produce chlorophyll d) that provided the cyanobacterial species, Acaryochloris marina, with the ability to utilize far-red light for photosynthesis. The close physical association of the cyanobacterium with other oxygenic phototrophs (e.g., Prochloron) and the selective benefit (to the recipients) of being outside the competitive milieu of organisms possessing chlorophyll a and/or b likely facilitated the acquiring of this function [29].

Worden et al. [30] suggested that members of the picoeukaryotic genus, Micromonas, could play a role as “sentinel organisms” for monitoring climate-change driven perturbations in oceanic systems. This potential utility as biogeochemical-indicator species is due to their distribution, and thus adaptation, across tropical to polar marine ecosystems.

Significantly, the differentiation and adaptive evolution of various isolates belonging to the Micromonas clade is likely the result of the combined action of horizontally-acquired genes and selection. In particular, a large fraction of the genes identified as “unique” to these eukaryotes lineages analyzed shared significant similarity with clades of prokaryotes. Furthermore, the two isolates examined by Worden et al. [30] were highly divergent at these loci reflecting this combination of reticulation and differential selection. Indeed, this pattern of divergence and acquisition (via horizontal transfer) reflects the evolutionarily dynamic nature of these protist lineages [30], and also indicates the potential for genetic exchange to underlie adaptive evolution [5–7].

The exchange of transposable elements (i.e., transposons) among distantly-related terrestrial organisms has been recognized for more than a decade (reviewed in [31]). For example, the relatively recent introduction of transposons known as “Type II class elements” was detected for the cosmopolitan invertebrate, Drosophila melanogaster. Likewise, a recent horizontal exchange of this class of element was inferred between D. melanogaster and D. willistoni, two species that last shared a common ancestor >50 million years ago [31].

Unlike the terrestrial lineages mentioned above, genomic information for marine invertebrates is relatively limited [32]. Tests for the horizontal exchange of transposable elements have been limited by this lack of genomic data. However, recent work has not only identified various Type II transposons, but has also identified instances of apparent horizontal exchange between phylogenetically-unrelated organisms that occur in close spatial proximity. Specifically, Casse et al. [32] identified mariner-like elements in the genomes of four different marine invertebrates (Table 1). Two of the species, Cancer pagurus and Maia brachydactila, are coastal crustaceans, while the remaining two species, Ventiella sulfuris and Bythograea thermydron are hydrothermal vent-associated organisms (an amphipod and crab, respectively [32]).

Though phylogenetically highly divergent lineages, the mariner-like elements isolated from the genomes of the coastal and the hydrothermal vent species demonstrated high levels of sequence similarity to the organism with which it was spatially associated [32]. In particular, the two hydrothermal organisms possessed elements that shared 99.5% similarity. Transposable elements isolated from the two coastal species likewise exhibited >99% sequence similarity. These findings led to the conclusion that, as for terrestrial invertebrates, horizontal transfer had resulted in the exchange of the Type II class elements between the unrelated, but sympatrically-distributed, coastal and hydrothermal vent species [32].

Table 1 lists three examples of marine plant clades that demonstrate the effect of introgressive hybridization and hybrid speciation (both homoploid–i.e., diploid hybrid derivative lineages–and polyploid). Of these, we will consider only the eelgrass, Zostera. This aquatic angiosperm, like other seagrasses, is an important constituent of the marine coastal communities in which it occurs. Though eelgrass genotypes have the capacity to reproduce asexually via rhizomatous growth, they also reproduce sexually, with their seeds having the ability to be transported over long distances by rafting on detached plants (e.g., up to 50km along the Northern European coast [33]).

Pollen flow among eelgrass populations appears limited (reviewed in [33]); however, gene flow via male gametes does occur at low frequencies. For example, in a sample of 28 Zostera “meadows” located both in the California Channel Islands and the adjacent mainland, Coyer et al. [34] detected genetic variation indicative of significant levels of gene flow, particularly for the coastal populations. Furthermore, their samples from the eelgrass sites included two species identified as Zostera marina and Z. pacifica. In addition to the genetic connectivity caused by clonal growth and sexual reproduction, these samples reflected admixtures of the two species’ genomes. Specifically, the assignment of genotypes to classes of Z. marina, Z. pacifica, or “introgressed”, utilizing microsatellite marker loci, detected hybrid/parental assemblages at the Santa Catalina, San Clemente and San Diego sites [34]. Each of these sites was suggested as a possible example of anthropogenically-mediated introgressive hybridization thus reflecting the extensive impact of humans on coastal environments. These results also led to the conclusion that introgressive hybridization may occur throughout the global distribution of Zostera populations, thereby contributing to the genetic variability in numerous eelgrass species [34].

The literature concerning the role of reticulate evolution in the origin and diversification of some coral clades is extensive. For example, descriptions of widespread introgressive hybridization leading to an enrichment of genetic and morphological variation are replete for reef corals belonging to the genus Acropora [19]. In this regard, Hatta et al. [35] found high rates of experimental, interspecific fertilization between naturally hybridizing species of acroporids. In contrast, Knowlton et al. [36] and Levitan et al. [37] defined barriers to reproduction between different species and morphotypes of Montastraea corals, but with regional differences in the strength of isolation. In addition, Fukami et al. [38] detected a north to south hybridization gradient in these same Montastraea lineages using molecular and morphological analyses. This latter study provided evidence that introgression between these species occurred mostly in the northern portion of their distribution. Likewise, an analysis of the family Faviidae documented extensive paraphyly across numerous clades [39]. It was suggested that introgressive hybridization may have contributed to the observation of paraphyly. Specifically, Huang et al. [39] argued “Introgression…may have resulted in such disparity, where the gene tree does not resemble the species tree, and neither is well-correlated with morphological evolution…” Additional evidence for introgressive hybridization within this clade of corals has also come from the fossil record. Specifically, morphological variation across the fossil record of this coral genus led to the inference of introgressive hybridization during the Pleistocene period [21].

Examples of hybrid speciation and introgression affecting coral evolution have been found within the genera Alcyonium and Pocillopora as well. McFadden & Hutchinson [40] tested the hypothesis of hybrid speciation giving rise to members of two genera of European soft corals (i.e., Alcyonium and Bellonella). Specifically, they analyzed sequence variation in the rRNA internal transcribed spacer (“ITS-1”) region of the putative hybrid lineages, Alcyonium hibernicum and Bellonella bocagei. The following patterns of genetic variability did indeed support the hypothesis of hybrid derivation for each of these lineages: (1) A. hibernicum possessed a combination of two different sequence variants that were characteristic of divergent clades of soft corals, and (2) B. bocagei possessed divergent ITS-1 types inferred to be recombinants between the same two divergent sequence families [40].

Species belonging to the tropical eastern Pacific scleractinian coral genus, Pocillopora, are the dominant reef-building organisms in this region [41]. Unlike their congeners found throughout most of the genus’ geographical distribution, these species have shifted from internal brooding of larvae to free-spawning [42]. Interestingly, the transition in reproductive mode was correlated with introgressive hybridization among five species of these tropical corals. Like the study of Alcyonium and Bellonella, rRNA ITS sequence data were collected from Pocillopora individuals distributed in the tropical eastern Pacific [41]. These data revealed a sharing of sequence variants among Pocillopora damicornis, P. eydouxi, P. elegans, P. inflata and P. effusus. Not only did Combosch et al. [41] infer introgressive hybridization as the source of this genetic variation, but they also concluded that the gene flow was largely unidirectional, with P. damicornis acting as the recipient of the allelic variability. Thus, numerous coral assemblages, including Indo-Pacific acroporids, the European soft corals and the eastern Pacific Pocillopora clades, reflect signatures of reticulate evolution.

Diatoms of the genus Pseudo-nitzschia are probably best known as the organisms responsible for ‘harmful algal blooms’ (HABs) during which large quantities of the neurotoxin, domoic acid is produced [43]. Domoic acid is the causative agent of ‘amnesic shellfish poisoning’ (ASP), a syndrome often caused (as the name suggests) by ingestion of shellfish laden with this neurotoxin [43]. Such poisoning events prior to the 1980’s had been ascribed mainly to toxin-producing dinoflagellate or cyanobacteria, taken up by shellfish that were subsequently eaten by humans [43]. However, a particular episode of poisoning in 1987 led to the discovery of Pseudo-nitzschia as the source of a low molecular weight amino acid (i.e., domoic acid) leading to ASP [43].

At least 12 species of Pseudo-nitzschia are now known to have the capacity to produce domoic acid (reviewed in [43,44]). Most of these species are now believed to be cosmopolitan [43,44], with genetic and morphological data indicative of multiple strains and/or varieties within some of the species [45,46]. For the present discussion it is significant that the evolution of multiple strains/varieties and their occurrence in sympatric associations have led to introgressive hybridization. This introgression has been documented using a variety of molecular and morphological markers. For example, D’Alelio et al. [45] detected admixtures of the Pseudo-nitzschia multistrata strains “A” and “B” (using the internal transcribed spacer region (ITS) of the ribosomal RNA genes) in samples from the Gulf of Naples. They thus found individual genomes defined as having A + B haplotypes. Because the individuals belonging to the A and B categories could not be defined on the basis of any other molecular or morphological parameter examined, it was concluded that this pattern of ITS admixture was due to the overlap of conspecific, but somewhat divergent, populations that had arisen either in situ or allopatrically [45].

Like the results from P. multistrata, analyses of P. pungens populations also detected patterns indicative of both divergence and introgression [46]. However, unlike P. multistrata, the P. pungens lineages were characterized by divergence in multiple genomic (i.e., ITS and chloroplast loci) and morphological characters [46]. Furthermore, the degree of divergence in the DNA and morphological characters allowed a relatively detailed definition of introgressive hybridization in a hybrid zone in the northeast Pacific. Casteleyn et al. [46] detected individuals exhibiting mixtures of morphological traits and chloroplast/nuclear haplotypes indicative of both first-generation and advanced-generation hybrids. As with all of the other examples given in this review, the detection of introgression within P. multistrata and P. pungens, suggests a broader base of genetic exchange among marine organisms than has been previously appreciated.

Like marine groups such as corals and cyanobacteria, mussels have a rich literature indicating extensive genetic exchange between divergent lineages (Table 1). For example, Arnold [5,7] has reviewed in detail work documenting the effect of introgression on the genetic structure of widely dispersed species of the genus, Mytilus. However, reticulate evolution is not limited to these near-shore taxa.

O’Mullan et al. [51] and Won et al. [52] reported genetic analyses of deep-sea hydrothermal vent mussels across an area of sympatry between the species, Bathymodiolus azoricus and B. puteoserpentis. Both studies detected introgressed individuals along a ridge in the area of overlap between the northern B. azoricus and southern B. puteoserpentis. In the first of the analyses, morphometric and genetic data (from both nuclear and mtDNA loci) “revealed a mixed population with gene frequencies and morphology that were broadly intermediate to those of the northern and southern species…” [51]. The spatially restricted nature of the hybrid individuals suggested the presence of selection against at least some of the hybrid genotypes. This latter hypothesis was supported by cytonuclear disequilibrium estimates [52]. In particular, Won et al. [52] discovered a pattern indicative of parental migration into the zone and restriction of the hybrids to the zone of overlap due to selectively disadvantageous interactions between genes inherited from the two species [52]. Notwithstanding the evidence for selection acting against some hybrid genotypes, recombination between the two hydrothermal vent mussel genomes–at least within the hybrid zone–was apparent.

Table 1 reflects the growing literature indicating the role of reticulate evolution within several different species complexes commonly known as coral reef fish. Two of these clades–Plectropomus and Acanthurus–exemplify the outcomes of introgressive hybridization and introgressive hybridization/hybrid speciation, respectively [60,61]. In particular, member lineages of these unrelated genera possess mosaic genomes and/or phenotypes reflecting contributions from multiple species. Furthermore, some of these hybrid lineages have been recognized as species.

van Herwerden et al. [60] defined the nuclear and mtDNA variation in two species of grouper, Plectropomus maculatus and P. leopardus, found along both the eastern and western Australian coastlines. The patterns of phylogenetic differentiation led these workers to infer both introgressive hybridization and incomplete lineage sorting as causal for discordances among the genetic markers. In particular, (1) the lack of reciprocal monophyly in mtDNA phylogenies for the eastern populations of the two species, but (2) the resolution of species-specific clades for the western samples, suggested an impact of introgression on the east coast lineages [60]. In contrast, incomplete lineage sorting was inferred as the cause of the discordances found among the nuclear-based phylogenies. Thus, one of three loci generated clades containing only one of the species. Two of the nuclear loci produced admixed clades containing samples from both P. maculatus and P. leopardus [60]. It is, however, possible that the discordance among the nuclear phylogenies, like those from the mtDNA, could also reflect the role of introgression.

A study of genetic variation in an area of sympatry in the eastern Indian Ocean between the coral reef surgeonfish, Acanthurus leucosternon and A. nigricans, also resolved patterns indicative of reticulate evolution (Figure 2). Marie et al. [61] collected DNA sequence data for three distinctive morphotypes, two reflecting A. leucosternon and A. nigricans and the third being a hypothesized hybrid species (i.e., “A. cf. leucosternon”). Sequence information was obtained from both mtDNA and nuclear loci. These data allowed simultaneous tests for introgression between A. leucosternon and A. nigricans, and the hybrid origin of A. cf. leucosternon. Both the introgression and hybrid speciation hypotheses were supported by the mtDNA data [61]. First, admixed clades of all three species were defined by the mtDNA sequence information (Figure 2). Indeed, the extent and directionality of introgression suggested concern that the A. leucosternon lineage might be lost from the region of overlap [61]. Second, mtDNA haplotypes characteristic of allopatric populations of both A. leucosternon and A. nigricans were detected in the sample of A. cf. leucosternon individuals (Figure 2); this is consistent with a hybrid origin for the “intermediate color patterns” possessed by A. cf. leucosternon [61].

Reticulate evolution in the form of introgressive hybridization has been well defined for numerous marine turtle taxa. For example, Karl et al. [62] used an analysis of both mtDNA and nuclear loci to test for the infrequent formation of hybrids among loggerhead, Kemp’s ridley, hawksbill and green sea turtles (Caretta caretta, Lepidochelys kempii, Eretmochelys imbricata and Chelonia mydas, respectively). Likewise, Bass et al. [63] detected divergent mtDNA haplotypes in Brazilian samples of hawksbill turtles that were identical, or nearly identical, to those found in loggerhead samples. Each of these analyses thus suggested the likelihood of low-frequency introgression in a number of marine turtle clades. Furthermore, the findings of Bass et al. [63] indicated the possibility of a large effect from introgressive hybridization on some lineages; 10 of 14 Brazilian “hawksbill” animals possessed mtDNA haplotypes most similar to loggerhead turtles.

Recently, Lara-Ruiz et al. [64] analyzed hawksbill populations from the state of Bahia in Brazil. This region contains > 90% of the E. imbricata nesting sites in Brazil. The high frequency of introgression in the Brazilian hawksbill samples suggested a decade earlier by Bass et al. [63] was confirmed by the much larger sample of 119 individuals. Over half of the turtles sampled (i.e., 67) possessed the expected mtDNA sequences characteristic of E. imbricata. Yet, of the remaining 52 individuals, 50 reflected introgression of mtDNA from C. caretta (loggerheads), while two possessed mtDNA from L. olivacea (i.e., the olive ridley lineage; [64]). Thus, introgressive hybridization among marine turtle lineages is taxonomically diverse and, in some cases, extensive in terms of the proportion of the population impacted.

Human-mediated environmental changes have been demonstrated to be catalysts for bouts of genetic exchange among a diverse array of organisms (see [6,7] for reviews). In this regard, introgressive hybridization among species of the fur seal genus, Arctocephalus reflects the role of anthropogenic processes [65,66]. Analyses of various fur seal populations thus suggested that the observed introgression was at least partly caused by the extinction (or near-extinction) of populations of Arctocephalus gazella, A. tropicalis and A. forsteri (Antarctic, subantarctic and New Zealand fur seals, respectively) due to human harvesting. The introgression among the fur seals was hypothesized to have occurred due to the recolonization of islands by multiple species occupied formerly by a single taxon [65,66].

In their analysis of the genetic structure of Macquarie Island, Lancaster et al. [65] collected mtDNA and nuclear sequence variation for 1007 pups sampled over an eight year period (Figure 3). Though A. gazella genotypes predominated, hybrids among all three species were also detected, with the percentage of hybrid pups averaging ca. 23% and varying from 17–30% across years (Figure 3). Four hybrid categories were identifiable from the Macquarie Island samples. These included: (1) A. gazella x A. tropicalis; (2) A. gazella x A. forsteri, (3) A. tropicalis x A. forsteri and (4) A. gazella x A. tropicalis x A. forsteri (Figure 3) [65]. Though lower reproductive success was detected for some of the hybrid classes, the presence of “multiple mating strategies” have led to the establishment of this genetically diverse hybrid zone among the three fur seal taxa [67,68]. Likewise, Kingston and Gwilliam [66] also detected introgression on Iles Crozet, but only between subantarctic and Antarctic fur seals. Furthermore, their estimated frequencies of hybridization were lower than that of Lancaster et al. [65]; these authors estimated the frequency of F₁ hybrids at 1% of the total population and 1.6% of the pups. They also concluded that at least 2.4% (and possibly as much as 4%) of the population consisted of backcross progeny [66].