MOLECULAR PHYLOGENY OF CONGENERIC MONOGENEAN PARASITES (DACTYLOGYRUS): A CASE OF INTRAHOST SPECIATION

Evolution, 58(5), 2004, pp. 1001–1018

MOLECULAR PHYLOGENY OF CONGENERIC MONOGENEAN PARASITES (DACTYLOGYRUS): A CASE OF INTRAHOST SPECIATION ANDREA SˇIMKOVA´,1,2,3 SERGE MORAND,4,5 EDOUARD JOBET,6,7 MILAN GELNAR,1,8

AND

OLIVIER VERNEAU2,9

1 Department

of Zoology and Ecology, Faculty of Science, Masaryk University, Kotla´rˇska´ 2, 61137 Brno, Czech Republic Fonctionnelle et Evolutive, UMR 5555 CNRS-UP, Universite´, 52 Avenue Paul Alduy, 66860 Perpignan Cedex, France 3 E-mail: [email protected] 4 Centre de Biologie et de Gestion des Populations, Campus international de Baillarguet, CS 30016, Montferrier sur Lez cedex, 34988 Montpellier, France 5 E-mail: [email protected] 6 Laboratoire Ge ´ nome et De´veloppement des Plantes, UMR 5096 CNRS-UP, Universite´, 52 Avenue Paul Alduy, 66860 Perpignan Cedex, France 7 E-mail: [email protected] 8 E-mail: [email protected] 9 E-mail: [email protected] 2 Parasitologie

Abstract. Dactylogyrus species (Dactylogyridae: Monogenea) are a group of monogenean gill parasites that are highly specific to freshwater fish of the family Cyprinidae. Dactylogyrus species were sampled from 19 cyprinids and one percid collected in Europe. Using partial 18S rDNA and ITS1 sequences, a phylogeny of 51 Dactylogyrus species was reconstructed to investigate the patterns of parasite speciation and diversification. Three main Dactylogyrus lineages were recognized from all phylogenetic trees, that is, analysis of 18S rDNA alone and combined 18SrDNA and ITS1. The first lineage associates the Dactylogyrus species of Cyprinus carpio and Carassius auratus of the Cyprininae; the second associates Dactylogyrus species of the Gobioninae, Pseudorasbora parva of the Rasborinae, and Ctenopharyngodon idella of the Cyprininae; and the third associates Dactylogyrus species of the Leuciscinae and Alburninae and Barbus barbus of the Cyprininae. Our results suggest that the genus Dactylogyrus is of quite recent origin and that these three lineages separated from each other in a very short period of time. Host subfamily mapping onto the parasite tree inferred from analysis of the combined dataset showed that the Cyprininae could be plesiomorphic hosts for Dactylogyrus. Dactylogyrus parasites would have secondarily colonized the Percidae and representatives of the Leuciscinae, Alburninae, Gobioninae, and Rasborinae. Comparison of host and parasite phylogenetic relationships indicated that a very high number of parasite duplications occurred within two of the three Dactylogyrus lineages. Dactylogyrus diversification can be mainly explained by sympatric intrahost speciation events that seem to be correlated to strict host specificity. Moreover, the present study shows that the congeneric parasites speciating within one host tend to occupy niches within hosts differing at least in one niche parameter. Key words.

Cyprinidae, Dactylogyrus, intrahost speciation, molecular phylogeny, Monogenea, niche preference. Received October 20, 2003.

The search for patterns and processes of evolution to understand speciation and diversification of parasites has been one of the main subjects in evolutionary biology (Brooks and McLennan 1993). As suggested by Brooks and MacLennan (1993), one important point is the investigation of host specificity that can be the basis of parasite speciation. Parasite speciation may occur in allopatry, that is, on two distinct geographically isolated host species, or sympatry, that is, either on distinct host species by a host shift or in the same host species by duplication. Cospeciation in host-parasite assemblages was first reported for pocket gophers and their chewing lice (Hafner and Nadler 1988; Hafner et al. 1994). In that case, allopatric parasite speciation was involved, following geographical isolation of their hosts. Thus, when host lineages speciate, parasites on the descendant host species may also speciate, which is illustrated by congruent host and parasite phylogenies. The presence of sister species in small areas (e.g., isolated islands or lakes) might imply sympatric speciation (Coyne and Price 2000), which can present an outcome of competition for resources (Dieckmann and Doebeli 1999). When considering only parasites, sympatric speciation might be the result of association of the host with particular ecological conditions (Klassen and Beverley-Burton 1987; Gue´gan and Agne`se

Accepted January 8, 2004.

1991), thus facilitating host switches. In that case, phylogenetically closely related parasite species infesting different host species are exemplified by incongruent host and parasite phylogenies. However, sympatric speciation may also be considered when a single host species is infested by a monophyletic parasite lineage. In that case, it represents intrahost speciation or parasite duplication, that is, the parasite speciates without a corresponding host cospeciation event and this leads to two or more lineages of parasites being present on single host species (Paterson and Gray 1997). Vickery and Poulin (1998) suggested that this could explain the occurrence of congeneric parasite species in the same host species. Finally, it is sometimes difficult to distinguish between sympatric and allopatric speciation when parasite extinction has occurred during host evolution (Page et al. 1998; Paterson and Banks 2001). In that case, the phylogenetic patterns of parasites cannot be easily explained, even in the light of host relationships. The Monogenea, as a group of ectoparasites with a direct life cycle that predominantly live on the gills and skin of fish, seem to be an appropriate model for studying the process of parasite diversification, mainly because of their high species richness and morphological and ecological diversity (Poulin 2002). Moreover, monogeneans are highly host (re-

1001 q 2004 The Society for the Study of Evolution. All rights reserved.

1002

´ ET AL. ANDREA SˇIMKOVA

stricted to one or a few host species) and niche specific (restricted to several microhabitats within the same host species), and it has been shown that congeneric species could coexist on the same host because of a low level of interspecific competition (Rohde 1977, 1979, 1991; Sˇimkova´ et al. 2000). To date, several models of congeneric monogenean species have been investigated to study the process of parasite diversification using phylogenies (Littlewood et al. 1997; Desdevises et al. 2002; Huyse and Volckaert 2002). Congeneric gill parasites belonging to Lamellodiscus (Monopisthocotylea), a group of parasites specific to marine fish of the Sparidae, were used as a model by Desdevises et al. (2002). These parasites are either generalists, infesting different host species, or specialists that may infest one host species. Desdevises et al. (2002) suggested that the genus was old and that intraspecific morphological variability could explain their ability to colonize different host species. They proposed that the rapid speciation of Lamellodiscus that infest hosts living in sympatry happened without any cospeciation and intrahost speciation. Similarly, no intrahost speciation was reported for Gyrodactylus (Monopisthocotylea), a genus occurring in multiple parasite lineages on single host species of freshwater and marine fish (Huyse and Volckaert 2002). Zietara and Lumme (2002) proposed the host switch as the model of Gyrodactylus speciation facilitated by their particular reproductive system. Polystomes (Polyopisthocotylea) are geographically widespread endoparasitic monogeneans infecting mainly amphibians and freshwater turtles. These parasites are morphologically very similar, are highly host specific, and congeneric species can be found on the same host species within turtles, but infecting different niches (i.e., oral cavities, urinary bladders, and conjunctival sacs). Littlewood et al. (1997), by investigating phylogenetic relationships of six polystome species of turtles, showed no intrahost speciation and concluded that the occurrence of congeneric species within the same host species could be the result of either host switching or cospeciation. Therefore, one may ask whether parasite speciation may occur within a single host. Moreover, within one host species, different congeneric species can occupy different niches, that is, microhabitats within the host (Rohde 1989). Therefore, one could also ask whether phylogenetically closely related congeneric species parasitize similar or different microhabitats within hosts. In the case of site-specific polystome monogeneans, congeneric species infecting the same niche within the different host species (even if geographically isolated host species) are more closely related than congeners infecting different sites of the same host species (Littlewood et al. 1997). Following the hypotheses of Rohde (1991), niche segregation or reproductive isolation between congeneric species parasitizing the same host species has been predicted, to prevent competition and to increase intraspecies mating contacts (Sˇimkova´ et al. 2000; Morand et al. 2002). The model of congeneric monogeneans of the genus Dactylogyrus (Dactylogyrinae: Dactylogyridae: Monopisthocotylea) was investigated in this study for the following reasons: Dactylogyrus is a highly diversified group within the Monogenea (Gusev 1985) with more than 900 nominal species mainly restricted to freshwater fish of the Cyprinidae (Gibson

et al. 1996), although, occasionally reported on the Percidae (Gusev 1985; Valtonen et al. 1990; Cone et al. 1994; Hayward 1997). In this genus, a high number of congeneric species coexisting on the same host species has been reported (Rohde 1989; Kennedy and Bush 1992; Sˇimkova´ et al. 2000, 2002), and several studies have been conducted on host specificity within the Cyprinidae (Gue´gan and Agne`se 1991; Gue´gan and Lambert 1991; El Gharbi et al. 1994; Lambert and El Gharbi 1995; Sˇimkova´ et al. 2001). The coexistence of Dactylogyrus species on the same host was suggested to be facilitated by niche distances and differing morphology of reproductive apparatus (Sˇimkova´ et al. 2002; Jarkovsky´ et al. 2004). Similarly, the Cyprinidae is the most diverse family among freshwater fish with about 2000 species (Helfman et al. 1997). Seven subfamilies are recognized, among them six with representatives occurring in Europe (Acheilognathinae, Alburninae, Cyprininae, Gobioninae, Leuciscinae, and Rasborinae), the most diverse being the Leuciscinae (Winfield and Nelson 1991). It has been believed for a long time that cyprinids originated in Asia and later dispersed to Europe, Africa, and North America (Winfield and Nelson 1991). Nevertheless, it is still questioned whether eastern Asia was an important interchange for cyprinids or a center of speciation (Durand et al. 2002). The present study carried out phylogenetic analyses, based on partial 18S, or small subunit, ribosomal RNA gene (rDNA) and ribosomal RNA gene internal transcribed spacer 1 (ITS1) sequences obtained from 51 Dactylogyrus species sampled from cyprinids and one percid species of central Europe. The process of parasite speciation and diversification, especially whether congeneric parasites coexisting on the same host species have speciated via intrahost speciation and whether phylogenetically related parasites occupied similar or different microhabitats, was investigated. MATERIALS

AND

METHODS

Parasite Sampling A total of 51 Dactylogyrus species, including specialist and generalist species, were collected from freshwater fish of the Morava River basin in the Czech Republic (central Europe). Among them, 49 species parasitizing fish species belonging to the family Cyprinidae (19 cyprinid species) and two species parasitizing Gymnocephalus cernuus belonging to the family Percidae, were recognized. A total of 44 Dactylogyrus species were collected only on one host species, and seven Dactylogyrus species were collected on more than one host species. For generalist species, several individuals of the different host species were collected. Dactylogyrus species with their hosts are given in Table 1. This sample is representative of the fauna living in central Europe (71 species of Dactylogyrus and 34 host species are known). Fish sampling was performed as described in Ergens and Lom (1970). Parasites were removed from the gills, placed on slides, covered by a coverslip, and identified using a light microscope equipped with phase contrast, differential interference contrast, and digital image analysis (MicroImage 4.0 for Windows, Olympus Optical Co., Hamburg, Germany). The sclerotized parts of the parasite attachment organ (the

1003

MOLECULAR PHYLOGENY OF DACTYLOGYRUS

TABLE 1. Dactylogyrus sequences obtained in this study with their host species and GenBank accession numbers for partial 18SrDNA and ITS1. Numbers 1, 2, and 3 on the right column refer to the Dactylogyrus lineages deduced from the parasite neighbor-joining tree. Dactylogyrus species and their hosts investigated for niche preference are indicated by asterisks. Fish species

Abramis ballerus Abramis brama* Abramis sapa Alburnus alburnus*

Aspius aspius Barbus barbus* Abramis bjoerkna*

Carassius auratus*

Chondrostoma nasus Ctenopharyngodon idella Cyprinus carpio Gobio albipinatus Gobio gobio Gymnocephalus cernuus Leuciscus cephalus*

Leuciscus idus

Phoxinus phoxinus Pseudorasbora parva Rutilus rutilus*

Scardinius erythrophthalmus*

Dactylogyrus species

auriculatus chranilowi falcatus wunderi* zandti* propinquus alatus* fraternus* minor* parvus* ramulosus tuba carpathicus* dyki malleus* cornoides* cornu* distinguendus* sphyrna* anchoratus* dulkeiti* formosus* inexpectatus* intermedius* vastator* chondrostomi ergensi vistulae lamellatus achmerovi anchoratus extensus finitimus cryptomeres amphibothrium hemiamphibothrium fallax folkmanovae* nanoides* prostae* vistulae* vranoviensis* crucifer ramulosus tuba vistulae borealis squameus caballeroi* crucifer* fallax* nanus* rarissimus* rutili* similis* sphyrna crucifer difformis* difformoides* izjumovae*

opisthaptor; i.e., the central hooks called anchors, seven pairs of marginal hooks, one connective bar [dorsal in this case] or two connective bars [dorsal and ventral]) and reproductive organs (vaginal armaments and copulatory organs) were used

Accession number

Dactylogyrus lineage

AJ564112 AJ564117 AJ564130 AJ564164 AJ564165 AJ564147 AJ564109 AJ564136 AJ564143 AJ564146 AJ564149 AJ564158 AJ564115 AJ564127 AJ564142 AJ564118 AJ564119 AJ564125 AJ564155 AJ564111 AJ564126 AJ564135 AJ564138 AJ564139 AJ564159 AJ564116 AJ564128 AJ564160 AJ564141 AJ564108 AJ490161 AJ564129 AJ564133 AJ564123 AJ564110 AJ564137 AJ564132 AJ564134 AJ564144 AJ564148 AJ564161 AJ564163 AJ564122 AJ564150 AJ564157 AJ564162 AJ564113 AJ564156 AJ564114 AJ564120 AJ564131 AJ564145 AJ564151 AJ564152 AJ564153 AJ564154 AJ564121 AJ490160 AJ564124 AJ564140

3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 1 1 1 1 1 1 3 3 3 2 1 1 1 2 2 3 3 3 3 3 3 3 3 3 3 3 3 3 2 3 3 3 3 3 3 3 3 3 3 3 3

for parasite determination according to Gusev (1985). Some parasite specimens were fixed in a mixture of glycerine and ammonium picrate and deposited in the collection of the Department of Zoology and Ecology, Masaryk University (Brno,

´ ET AL. ANDREA SˇIMKOVA

1004

Czech Republic). Other parasites were preserved in 95% ethanol before DNA extraction. DNA Extraction and Polymerase Chain Reaction Individual parasites were removed from ethanol, dried, and DNA was extracted using the standard phenol-chloroform method (Sambrook et al. 1989). For generalist species, individuals from different host species were investigated. Total DNA obtained from individual parasites was resuspended in 25 ml distilled water. Partial 18S rDNA and entire ITS1 region were amplified in one round using primer S1 (59ATTCCGATAACGAACGAGACT-39) that anneals in the terminal region of the 18S gene (Sinnappah et al. 2001) and H7 (59-GCTGCGTTCTTCATCGATACTCG-39; Sinnappah et al. 2001) or IR8 primers (59-GCTAGCTGCGTTCTTCATCGA-39; Sˇimkova´ et al. 2003) that anneal in the 5.8S rDNA. Each amplification reaction was performed in a final volume of 25 ml containing 1.5 units of Taq polymerase, 1X buffer, 1.5 mM MgCl2, 200 mM of each dNTP, 0.8 mM of each polymerase chain reaction (PCR) primer and 3 ml DNA. PCR was carried out with the following steps: 4 min at 958C followed by 35 cycles of 1 min at 958C, 1 min at 558C, and 1 min 30 sec at 728C, and 10 min of final elongation at 728C. Electrophoresis was performed on a 1% agarose gel stained with ethidium bromide for DNA visualization. PCR products were cut from the gel and purified using siliconized glasswool. DNA was precipitated and resuspended in sterile water. Some of the PCR products were sequenced directly. Others were ligated into the pGEM-T vector and cloned in Escherichia coli JM109 competent cells when the amount of purified DNA after PCR was too low for direct sequencing. Recombinant colonies were checked following Sekar’s procedure (1987) and plasmids were purified with the Wizard Plus miniprep kit (Promega, Madison, WI). At least three clones were sequenced per species. DNA Sequencing The direct sequencing of purified PCR products as well as recombinant plasmids was performed using the same primers as for PCR or universal primers (T7 and SP6) supplied by Promega. Moreover, for sequencing species included in the D. minor group (group 3, see Results), two forward primers IF8 (59-AACTGTTCAATCATCGTCGTG-39; Sˇimkova´ et al. 2003) and newly designed IF9 (59-ATCCGCCGACTCTGACTGGA-39) and two reverse primers IR5 (59-TACGGAAACCTTGTTACGAC-39; Sinnappah et al. 2001) and newly designed IR9 (59-RRGACTCACCCGAAGGGAG-39) were used. For sequencing of species included in the D. anchoratus group (groups 1 and 2, see Results), the same internal primers were used, with the exception of IF9, which was replaced by IF10 (59-YMTTCTCCCTTCGGGTGAGT39) also designed for this study. Sequences obtained, including partial 18S rDNA, complete ITS1, and partial 5.8S rDNA ranged from 962 to 1130 bp. Primers IF8 and IR5 anneal in the 18S rDNA at positions 234–254 and 471–490, respectively, of D. difformis sequence accession no. AJ490160. The primers IF9 and IR9 anneal in the ITS1 region at positions 223–243 and 348–367, respectively, of D. difformis, and IF10 anneals in the ITS1 region

at positions 274–294 of D. anchoratus accession no. AJ490161. Sequencing was carried out using ABI Prism Big Dye Terminator Cycle Sequencing kit (Applied Biosystems, Foster City, CA) and electrophoresis was performed on an automated sequencer (ABI373 DNA Sequencer). Sequences were read and corrected using Sequencher software (Gene Codes Corp., Ann Arbor, MI). New sequences were deposited in EMBL (see Table 1 for their accession numbers). Phylogenetic Analyses of Dactylogyrus DNA sequences were aligned using the ED program in the MUST package (Philippe 1993). Gaps and ambiguously aligned regions were removed. First, phylogenetic analysis using partial 18S rDNA alone was performed to polarize Dactylogyrus evolution as the variability of this marker allows for unambiguous alignment with the outgroup. For this analysis seven other monogenean species belonging to different subfamilies of the Dactylogyridae were included, following the phylogenetic relationships previously inferred for Dactylogyridae (Sˇimkova´ et al. 2003). Among them two species, Thaparocleidus vistulensis (accession no. AJ490165) of the Ancylodiscoidinae and Cleidodiscus pricei (AJ490168) of the Ancyrocephalinae, were used for rooting the tree. Five other species previously found to form a group including Dactylogyrus species were also used (i.e., Pseudodactylogyroides apogonis, Pseudodactylogyrus anguillae, and P. haze of the Pseudodactylogyrinae and Thylacicleidus sp. and Pseudohaliotrema sphincteropus of the Ancyrocephalinae; AB065115, AJ490162, AB0651141, AJ490169, AJ287568, respectively). For generalist species, only the sequence of the individual parasitizing the most commonly infested host was retained, so 58 species were analyzed. Distance trees were generated with a neighbor-joining (NJ) algorithm based on Kimura two-parameter distances (Kimura 1980) and performed with PAUP* 4b10 (Swofford 2002). Support values for internal nodes were estimated using a bootstrap resampling procedure with 1000 replicates (Felsenstein 1985). Maximum likelihood (ML) and NJ analyses based on ML distances were also conducted using PAUP from the best appropriate model (TrNf 1 G 1 I in this case) selected by the ModelTest program (Posada and Crandall 1998). The search of the best ML tree was done using a branch-swapping algorithm (TBR, tree bisection reconnection). One hundred replicates for ML and 1000 replicates for NJ were calculated using the same model as above, using the TBR branch-swapping algorithm for ML. Maximum parsimony (MP) analysis was also performed using heuristic search with stepwise random addition sequence running on unweighted informative characters. Support values for internal nodes were estimated after 1000 replicates using the TBR branch-swapping algorithm. Further analyses were performed using combined sequences from partial 18S rDNA and ITS1 without including other species belonging to the Dactylogyridae as ITS1 sequences of Dactylogyrus species could not be aligned with ITS1 sequences of the seven species of Dactylogyridae used in the analysis of 18S rDNA. The analyses of combined 18S rDNA and ITS1 data were done to obtain a better phylogenetic resolution within the Dactylogyrus genus. NJ analysis was

1005

MOLECULAR PHYLOGENY OF DACTYLOGYRUS

TABLE 2. Fish species used for phylogenetic analysis, with the exception of Gobio albipinatus, for which the cytochrome b gene has not yet been sequenced. Asterisks indicate species for which Dactylogyrus species were recorded. The taxonomic position and GenBank accession numbers for cytochrome b sequences are indicated for each species. Order

Family

Cypriniformes

Balitoridae Characidae Cyprinidae

Subfamily

Cyprininae

Gobioninae Rasborinae Leuciscinae

Alburninae Perciformes

Percidae

Salmoniformes

Salmonidae

performed as described above from Tamura Nei distances including a proportion of invariable characters and gamma distribution (TrN 1 I 1 G). One thousand bootstrap replicates were conducted following the same evolutionary model. ML analyses were also performed on the TrN 1 I 1 G model selected by ModelTest. One hundred replicates were calculated following the same model and using the same procedure described for the analyses of 18S rDNA alone. MP analysis was performed as described in the case of 18S rDNA. Host Phylogeny The phylogeny of cyprinids has been previously investigated using mitochondrial markers (Briolay et al. 1998; Gilles et al. 1998, 2001; Zardoya and Doadrio 1999; Zardoya et al. 1999; Durand et al. 2002), but not all species used in our study were included. Thus, we reconstructed the host phylogeny from cytochrome b sequences retrieved from GenBank, including 18 representatives of the Cyprinidae, one of the Balitoridae, one of the Characidae, and six of the Percidae. The tree was rooted with Oncorhynchus mykiss from the Salmonidae. All species with their accession numbers are given in Table 2. The ModelTest program was used for selecting the best evolutionary model for ML analysis (TVM 1 I 1 G). Support values for internal nodes were estimated using a bootstrap resampling procedure with 100 replicates following a branchswapping algorithm (TBR) on the same model. NJ analysis was performed using evolutionary parameters selected by ModelTest. One thousand replicates were calculated follow-

Species

Crossostoma lacustre Astyanax mexicanus Barbus barbus* Carassius auratus* Cyprinus carpio* Ctenopharyngodon idella* Gobio albipinatus* Gobio gobio* Pseudorasbora parva* Abramis ballerus* Abramis bjoerkna* Abramis brama* Abramis sapa* Aspius aspius* Chondrostoma nasus* Leuciscus cephalus* Leuciscus idus* Phoxinus phoxinus* Rutilus rutilus* Scardinius erythrophthalmus* Alburnus alburnus* Ammocrypta clara Etheostoma kennicotti Gymnocephalus cernuus* Perca flavencens Perca fluviatilis Zingel streber Oncorhynchus mykiss

Accession number

M91245 AF045997 Y10450 AF051858 X61010 AF051860 no sequence available Y10452 Y10453 AY026409 Y10442 Y10441 AY026408 AY026398 AY026402 Y10446 AY026397 Y10448 Y10440 Y10444 Y10443 AF045350 AF045341 AF045356 AJ001521 AF045358 AF045360 L29771

ing the same model. MP analysis was performed as described for previous analyses. Mapping of Host Subfamilies onto the Parasite Tree The host subfamily for each Dactylogyrus species was mapped onto the NJ tree inferred from analysis of combined data on 51 Dactylogyrus species, using MacClade version 4.0.1 with Farris optimization (Maddison and Maddison 1992). Host species were divided into subfamilies according to Nelson (1994) and are given in Table 2. For parasite species that have been recorded on different host species (generalists) of different host subfamilies, mapping was performed with the host subfamily from which parasites were examined and analyzed. It should be noted that this host subfamily corresponds to the most infested host species. Comparison of Host and Parasite Phylogenies TreeMap 1.0 (Page 1994) was used to represent host-parasite associations using the Dactylogyrus tree inferred from analysis of combined data and the cyprinid tree after excluding fish species not infested by Dactylogyrus. Gobio albipinatus was added to the fish topology as a sister species of G. gobio considered as a monophyletic group (Barusˇ and Oliva 1995). Mapping of Niche Preference The mapping of niche preference was performed using the same methodology as for mapping of host subfamilies. For

´ ET AL. ANDREA SˇIMKOVA

1006

niche preference, data concerning niche position (i.e., the position on the gills within the hosts) were recorded for 33 Dactylogyrus species obtained from eight different cyprinid species (asterisks in Table 1). We could not obtain niche data for all Dactylogyrus species from all host species investigated because of either low host sample or low parasite sample. Some fish species were found only rarely, and even when using a relatively high number of host individuals we obtained only a few individual specimens of several Dactylogyrus species. The parasite niche was investigated as previously described in Sˇimkova´ et al. (2000). For a description of the gill arch, see Figure 6A. The niche position of each parasite individual was recorded (i.e., its position on the gill arch, gill segment, gill area), and the position where the maximum number of parasite individuals was found was considered as the preferred niche. The preferred niche (gill arch, segment, and area) was mapped onto the phylogenetic tree obtained from the NJ analysis using combined 18S rDNA and ITS1 sequences. RESULTS Molecular Divergence within Generalist Species For each of the seven generalist species investigated, we sequenced at least two individuals recovered from each different host species. Whereas no differences were detected between individuals collected from the same host species, differences were observed between individuals collected from different host species. Pairwise comparisons showed 0.4% molecular divergence between D. anchoratus specimens from Cyprinus carpio and Carassius auratus; no difference between D. crucifer specimens from Rutilus rutilus, Scardinius erythrophthalmus, and Leuciscus idus; 0.3% molecular divergence between D. fallax specimens from R. rutilus and Leuciscus cephalus; 1% molecular divergence between D. ramulosus specimens from Aspius aspius and L. idus; 0.6% molecular divergence between D. sphyrna specimens of Abramis bjoerkna and R. rutilus; and no difference between D. tuba specimens of Aspius aspius and L. idus. Finally, 0.2% divergence was calculated between D. vistulae specimens from Chondrostoma nasus and L. idus; the same order of divergence (1.2%) was found between D. vistulae specimens from C. nasus and L. idus with specimens from L. cephalus. The genetic divergence between the two most closely related species was 1.4% (D. caballeroi and D. crucifer), and the genetic divergence between other pairs was more than 2.6%. Phylogenetic Analysis Using Partial 18S rDNA For the phylogenetic analyses based on partial 18S rDNA alone we used all Dactylogyrus species sequenced for this study and seven other dactylogyrids. The sequence alignment comprised 440 unambiguously alignable positions after removing gaps and ambiguous aligned regions, of which 150 were variable with 106 parsimony informative. The first analysis was performed using NJ with Kimura two-parameter because of equal nucleotide base frequencies (about 25%). The tree obtained from the Kimura two-parameter distances

with bootstrap proportions (BP) supported the monophyly of Dactylogyrus (BP 5 84; see Fig. 1). ML analysis was performed on the TrNf 1 G 1 I model, with the following parameters: substitution rate matrix: A-C 5 1.000, A-G 5 4.1211, A-T 5 1.000, C-G 5 1.000, C-T 5 5.4533, G-T 5 1.000; proportion of invariable sites 5 0.4342; and rate heterogeneity approximated by a gamma distribution (four rate categories), a 5 0.5912. Surprisingly, the best ML tree (not shown) shows nonmonophyly of Dactylogyrus. Indeed, the group including D. cryptomeres, D. finitimus, D. squameus, and D. lamellatus was positioned as the sister group of the Pseudodactylogyrinae. Following NJ analysis using TrNf 1 G 1 I, species of Pseudodactylogyrinae are not nested within Dactylogyrus species. Nevertheless, the monophyly of Dactylogyrus was again not supported. MP analysis was also performed. In MP analyses, 100 equally parsimonious trees with 393 steps were retained (consistency index, CI 5 0.550, retention index, RI 5 0.639). However, BP values for MP show a very weak phylogenetic resolution. No statistically significant differences among NJ on Kimura two-parameter, ML tree, NJ based on ML distance (TrNf 1 G 1 I), and the equally most parsimonious 100 trees were found (Shimodaira-Hasegawa test implemented in PAUP* 4b10, P . 0.05). This incongruence (monophyly vs. nonmonophyly of Dactylogyrus) may be due to different evolutionary rates between Dactylogyrus species and the other species used as outgroup. Li (1997) stated that ML methods may become inconsistent if the rate of evolution is assumed to be uniform when in fact it is not. We observed highly variable substitution rates among the branch lengths in the tree inferred from the ML approach, suggesting that this assumption may be violated. Furthermore, we noted that rate heterogeneity was different between Dactylogyrus species and other dactylogyrids. Finally, the high number of parameters for the TrNf 1 G 1 I model involves a higher variance of branch-length estimation and therefore may lead to incorrect reconstruction of internal short branches (Nei and Kumar 2000). For those reasons, we retained the NJ tree based on the Kimura two-parameter distances that shows the monophyly of Dactylogyrus. Within Dactylogyrus we observe three monophyletic lineages, which are moderately supported (BP values). The first (further referenced as clade 1) includes the Dactylogyrus species from two species of the Cyprininae; the second (clade 2) associates one Dactylogyrus species from the Cyprininae, one from the Rasborinae, and two from the Gobioninae; and the third (clade 3) includes mainly Dactylogyrus species parasitizing species of the subfamilies Leuciscinae and Alburninae, three Dactylogyrus species of Barbus barbus (subfamily Cyprininae), and two of the family Percidae. While no clear phylogenetic pattern is observed among the three clades, relationships between species within clades 1 and 2 are better resolved than relationships between species of clade 3 (see Fig. 1). Phylogenetic Analysis Using Combined Data (18S rDNA and ITS1) The partition homogeneity test as implemented in PAUP* 4b10 (Swofford 2002) was used to test the congruence of the

MOLECULAR PHYLOGENY OF DACTYLOGYRUS

1007

FIG. 1. Neighbor-joining tree based on Kimura two-parameter distances inferred from analysis of partial 18S rDNA sequences of 51 Dactylogyrus species including seven representatives of the Dactylogyridae (Thaparocleidus vistulensis, Cleidodiscus pricei, Pseudohaliotrema shincteropus, Thylacicleidus sp., Pseudodactylogyroides apogonis, Pseudodactylogyrus anguillae, P. haze), with T. vistulensis and C. pricei taken as outgroup. Numbers along branches indicate bootstrap percentages resulting from the different analyses in the order: neighbor joining/maximum parsimony/maximum likelihood. Values lower than 50 are indicated with dashes.

1008

´ ET AL. ANDREA SˇIMKOVA

two datasets: 18S and ITS1. As no significant difference was found (P 5 0.100), further analyses were performed with the 51 Dactylogyrus species examined in this study using combined sequences from partial 18S rDNA and ITS1. The alignment of the partial 18S rDNA and ITS1 sequences comprised 712 unambiguously alignable positions after gaps were removed, of which 246 were variable with 201 parsimony informative. The ML analysis was performed on the model TrN 1 I 1 G with the following parameters: substitution rate matrix: A-C 5 1.0000, A-G 5 3.0453, A-T 5 1.0000, C-G 5 1.0000, C-T 5 3.7850, G-T 5 1.0000; the proportion of invariable sites 5 0.5341; and rate heterogeneity approximated by a gamma distribution (four rate categories), a 5 0.7423. The topology of the best ML tree was similar to the topology inferred from NJ analysis on the Tamura Nei distances including one proportion of invariable characters and gamma distribution (TrN 1 I 1 G). MP analysis was also performed and 100 equally parsimonious trees with 951 steps were retained (CI 5 0.428, RI 5 0.621). The strict consensus displays similar topology to the NJ and ML trees, that is, it supports the clustering of species found from the NJ and ML analyses with BP of more than 50%. No statistically significant differences among NJ, ML, and trees obtained by MP were found (Shimodaira-Hasegawa test implemented in PAUP* 4b10, P . 0.05), therefore we only present the NJ tree giving BP values obtained from all three analyses (Fig. 2). The phylogenetic relationships inferred from 18S rDNA analysis (Fig. 1) do not allow us to clearly specify which of the three reported lineages diverged first. Therefore, the tree obtained from combined 18S rDNA and ITS1 data was drawn to reflect the phylogenetic relationships described in the analysis of 18S rDNA sequences, that is, three Dactylogyrus groups. Each of the three Dactylogyrus groups is well supported (BP values more than 90). Moreover, the midpoint rooting technique was applied as implemented in PAUP* 4b10 and highlights the first Dactylogyrus group as the most basal lineage. Several general trends are inferred from all analyses. The first lineage (group 1, Fig. 2) is divided into two sister subgroups. The first associates two specialists found on C. carpio that are related to the two specialists found on C. auratus. The second one associates four species found on C. auratus. However, one of them, D. anchoratus, is also found on C. carpio. The second lineage (group 2, Fig. 2) is composed of four species. Two Dactylogyrus species found on Gobio species are related to D. squameus from P. parva. The fourth species, D. lamellatus from Ctenopharyngodon idella, is the most basal. All associations within the first and second lineages are well supported by BP in all analyses. The third lineage (group 3, Fig. 2) can be divided in two groups: the first one associates two Dactylogyrus species of the Percidae (Gymnocephalus cernuus), and the second associates all remaining Dactylogyrus species, that is, parasites of the Leuciscinae, A. alburnus of the Alburninae, and B. barbus of the Cyprininae. Phylogenetic interrelationships among species of this group are in general moderately or poorly supported or unresolved with the exception of one subgroup of generalist parasites that associates species of the

Leuciscinae and Alburninae (((D. similis, D. alatus) D. sphyrna) D. vistulae). In all cases, parasites of the three host subfamilies are monophyletic. Conversely, the Dactylogyrus species of the two species that are the most common hosts for Dactylogyrus, L. cephalus and R. rutilus, do not form monophyletic groups. Phylogeny of Cyprinid Species Using Cytochrome b The alignment of cytochrome b sequences comprised 1140 alignable positions, of which 545 were variable with 487 parsimony informative. The selected model for ML analysis was TVM 1 I 1 G with the following parameters: A-C 5 0.6912, A-G 5 8.1493, A-T 5 0.4777, C-G 5 0.3696, C-T 5 8.1493, G-T 5 1.0000; the proportion of invariable sites 5 0.4887; and rate heterogeneity approximated by a gamma distribution (four rate categories), a 5 0.8008. The best ML tree is reported in Figure 3. Whereas the monophyly of the Cyprinidae is well supported, the monophyly of the different host subfamilies is moderately supported. The only representative of Alburninae is nested within Leuciscinae. The Gobioninae and Rasborinae are sister subfamilies, but only one species from these two subfamilies was examined. Finally, monophyly of the Cyprininae was not found, but the divergence of this subfamily appears to be the most basal event within the Cyprinidae. NJ using the parameters selected by ModelTest was also performed. Three equally parsimonious trees with 2909 steps were retained (CI 5 0.331, R 5 0.429), but no statistically significant differences were found among the topologies of the trees obtained from ML, NJ, and MP analyses (Shimodaira-Hasegawa test implemented in PAUP* 4b10, P . 0.05), and BP values obtained from ML, NJ, and MP analyses are reported in Figure 3. Mapping Host Subfamilies onto the Parasite Phylogenetic Tree The fish subfamilies were mapped onto the parasite phylogeny inferred from the analysis of the combined dataset (18S rDNA and ITS1). This mapping was done with the assumption that the first Dactylogyrus group has the most basal divergence, following the results obtained from the analysis of combined 18S 1 ITS1 using a midpoint rooting technique. However, following the results using only partial 18S rDNA, where lineages 1 and 2 are not well supported, three hypotheses should be mentioned when considering which lineage is basal. When groups 1 (Fig. 4) or 2 (not shown) are considered as the most basal taxa, the Cyprininae is hypothesized to be a plesiomorphic host subfamily for Dactylogyrus. In that case, the Percidae would have been secondarily colonized by Dactylogyrus species. However, when group 3 (not shown) is considered as the most basal taxon, the plesiomorphic host subfamily is equivocal. It could be representatives of the Cyprininae, Leuciscinae, or Percidae. In all cases, parasites of B. barbus (Cyprininae) and of A. alburnus (Alburninae) are derived from host fish of the Leuciscinae. Comparison of Host and Parasite Phylogenies Although phylogenetic relationships are not fully resolved for the parasites, several cases of parasite duplication within

MOLECULAR PHYLOGENY OF DACTYLOGYRUS

1009

FIG. 2. Neighbor-joining tree based on Tamura-Nei distances, including one proportion of invariable characters and gamma distribution, inferred from analysis of combined partial 18S rDNA and ITS1 sequences of 51 Dactylogyrus species. Numbers along branches indicate bootstrap percentages resulting from different analyses in the order: neighbor joining/maximum parsimony/maximum likelihood. Values lower than 50 are indicated with dashes.

1010

´ ET AL. ANDREA SˇIMKOVA

FIG. 3. Maximum-likelihood tree inferred from analysis of cytochrome b sequences of 18 representatives of the Cyprinidae, one of the Balitoridae, one of the Characidae, six of the Percidae, and one of the Salmonidae. Selected model was TVM 1 I 1 G. Numbers along branches indicate bootstrap percentages resulting from different analyses in the order: maximum likelihood/maximum parsimony/neighbor joining. Values lower than 50 are indicated with dashes.

host species (i.e., intrahost speciation) can be recognized (Fig. 5). Dactylogyrus duplications were found within the host species A. alburnus, which is infested by three specialist parasites; within S. erythrophthalmus, which is also infested by three specialist parasites; within C. nasus, which is infested by two specialists; within A. bjoerkna for the three specialist parasites D. cornoides, D. cornu, and D. distinguendus; within L. cephalus for the two specialists D. nanoides and D. folkmanovae; within A. brama infested by two specialists D. zandti and D. wunderi; within B. barbus infested by two specialists; and within G. cernuus, which is infested

by two specialist parasites. Those duplications were supported by BP of more than 50 except those found within S. erythrophthalmus, A. brama, and B. barbus (see Figs. 2, 5). Two duplication events can be recognized within R. rutilus. The first one is characterized by the two specialist parasites D. rutili and D. nanus (BP . 70 from all analyses) and the generalist parasite D. fallax (BP , 50), and the second one by the specialist D. caballeroi and the generalist D. crucifer (BP . 90 from all analyses), the two generalist species infesting predominantly R. rutilus. All those duplication events occurred within lineage 3. Similarly, some duplication events

MOLECULAR PHYLOGENY OF DACTYLOGYRUS

1011

FIG. 4. Mapping of host subfamilies onto the parasite neighbor-joining tree inferred from analysis of combined data. The first Dactylogyrus group is considered as the most basal lineage within Dactylogyrus following the most likely scenario of Dactylogyrus evolution.

were also reported within lineage 1. Dactylogyrus duplications were found within the host species C. carpio (BP . 60), which is infested by two specialist parasites, and within C. auratus for which two duplication events were reported. The first gave rise to two specialist species (D. vastator and D. intermedius, BP 5 100) and the second (BP . 95) to three specialist parasites and one generalist, that is, D. anchoratus also infests C. carpio. No duplication event was reported within lineage 2. Besides intrahost duplication events, several cospeciation events (numbers in Fig. 5) can also be recognized from the comparison of host and parasite phylogenies. In lineage 2, the two sister species D. finitimus and D. cryptomeres parasitize two congeneric host species, G. albipinatus and G. gobio respectively (a in Fig. 5). Moreover, Dactylogyrus of Gobio species are a sister group to D. squameus of P. parva (b in Fig. 5). Because Gobio species and P. parva form a monophyletic group, cospeciation events can be suggested to explain similarities in host and parasite relationships. Similarly, the parasite species infesting A. alburnus from the Alburninae are closely related to Dactylogyrus species in-

festing S. erythrophthalmus from the Leuciscinae (lineage 3, c in Fig. 5), but this cospeciation is not supported by bootstrap (Fig. 2). Even if the Alburninae are recognized to be polyphyletic and nested within Leuciscinae, A. alburnus is a sister species of S. erythrophthalmus (Figs. 3, 5). Thus, it is likely that these two groups of parasites cospeciated with their host species and later diversified within hosts. Dactylogyrus prostae would have switched to L. cephalus before intrahost speciation occurred on S. erythrophthalmus. Finally, several host switches must be invoked to explain discordant host and parasite relationships. That is the case, for instance, for Dactylogyrus species parasitizing B. barbus that are nested within the species of the Leuciscinae. Mapping of Niche Preference onto the Phylogenetic Tree The preferred niche of each Dactylogyrus species was mapped onto their phylogeny. For parasite species found on more than one host species (D. sphyrna and D. crucifer), the position on the fish species with the highest parasite abundance was chosen. This position, however, was similar to

1012

´ ET AL. ANDREA SˇIMKOVA

FIG. 5. Tanglegram of Dactylogyrus and cyprinid species deduced from comparison of the parasite neighbor-joining tree inferred from analysis of combined data (18SrDNA and ITS1 sequences) with the topology of a fish phylogeny that includes only species infested by Dactylogyrus species. Gobio albipinatus was added to the fish phylogeny as a sister species of Gobio gobio considering that they are monophyletic. Intrahost duplications are depicted on the trees. Dactylogyrus speciation by intrahost duplication supported by bootstrap percentage greater than 50 are given in bold. Potential coevolution events are indicated by a, b, and c.

MOLECULAR PHYLOGENY OF DACTYLOGYRUS

that observed when considering individuals of a given Dactylogyrus species obtained from more than one host species. The first mapping shows the preferred position on the gill arch (Fig. 6B). The position on the second gill arch seems to be ancestral and several changes toward the first, third, and fourth arches were noted. The evidence for strong changes in the arch position was found within the groups of parasites on C. auratus. The clade including the specialist Dactylogyrus species parasitizing B. barbus, L. cephalus, S. erythrophthalmus, and A. alburnus shows a shift from the position on the second to the third gill arch with changes toward first and second gill arches for parasites of S. erythrophthalmus. The preferred segment position on the gill arches was mapped onto the phylogeny (Fig. 6C). The position on the dorsal segment is found to be the ancestral state with several changes toward the medial and ventral segments for the species in terminal positions. Changes in preferred segment position were found between the branches within monophyletic groups of the lineages of Dactylogyrus species parasitizing the same host species. The preferred area position on the gill arches was mapped onto the phylogeny (Fig. 6D). The position on the distal area is found to be the ancestral state with several changes toward the central and proximal areas. The same observation as in the case of preferred segment position was noted, that is, the changes in preferred areas are found between the branches within the group of Dactylogyrus species parasitizing the same host species. However, when Dactylogyrus species of one clade and parasitizing the same host (i.e., recognized as a case of intrahost duplication) were separated by their arch position, then the same segment or area positions are often found in the phylogenetic tree. Looking at changes for all positions, we can note that no sister groups have identical niches and at least one of the niche parameters (arch, segment, or area) has changed. DISCUSSION The Dactylogyrus-Cyprinidae Model In general, parasites with direct life cycle and free-living infectious stages attaching to the external surface of their hosts are believed to have a narrow host range (i.e., they are specialists) and to have coevolved with their hosts, in contrast to parasites with indirect life cycles (Littlewood et al. 1997; Paterson and Poulin 1999). Therefore, highly specific monogeneans could be considered an appropriate model for studying the evolution of host-parasite associations to investigate patterns and processes of parasite speciation. Nevertheless, different levels of correspondence between patterns of parasite speciation and host specificity have been shown across different models of congeneric monogenean parasites (Littlewood et al. 1997; Bentz et al. 2001; Desdevises et al. 2002; Huyse and Volckaert 2002), suggesting importance of the host species in the process of parasite diversification. When considering that evolution of freshwater fish is more constrained than that of marine fish (i.e., it is correlated to the history of freshwater movements and dispersals; Tsigenopoulos and Berrebi 2000), then freshwater fish may represent more suitable models for studying host-parasite coevo-

1013

lutionary patterns (e.g., case study of marine monogenean parasites by Desdevises et al. 2002). Therefore, we studied highly specific monogenean parasites of the genus Dactylogyrus that infest freshwater fish of the Cyprinidae. This study represents a large parasite sampling from central European fish. Dactylogyrus Origin On the basis of all phylogenetic analyses using partial 18S rDNA separately or combined 18S rDNA and ITS1 data, three main lineages of Dactylogyrus species were recognized: (1) the clade of parasites of C. carpio and C. auratus, both representatives of the Cyprininae; (2) the lineage that includes parasite species of C. idella (Cyprininae), P. parva of the Rasborinae (this species is considered as Gobioninae by Chen et al. 1984) and Gobio of the Gobioninae; and (3) the lineage that includes parasite species of B. barbus (Cyprininae), Leuciscinae, Alburninae, and Percidae, all of European origin. Phylogenetic relationships between species within each of the three main lineages were better resolved when analyzing combined data. The 18S rDNA represents a well-conserved gene that evolves relatively slowly (Hillis and Dixon 1991). This explains why it has been widely used for the study of plathyhelminth relationships (Littlewood and Olson 2001; Olson and Littlewood 2002). However, fast-evolving sequences, for instance the ITS1 region, could provide useful phylogenetic information for resolving relationships within groups of recent origin (Booton et al. 1999). Desdevises et al. (2002) showed in the case of monogenean species belonging to Lamellodiscus that ITS1 sequences could not be aligned among species, which suggested an older age of this group of parasites than hypothesized from a morphological point of view. It may also be the result of an increased substitution rate in this genus. Our results inferred from analysis of ITS1 sequences suggest that Dactylogyrus is of quite recent origin and that the three lineages separated from each other in a very short period of time. This is confirmed by our alignment of ITS1 sequences derived from all Dactylogyrus species and by the lack of basal resolution in the genus. Plesiomorphic Host Subfamily Results inferred from mapping host subfamily onto the parasite phylogenetic tree could suggest three scenarios concerning plesiomorphic hosts for Dactylogyrus. Evolutionary scenarios considering either lineage 1 or 2 as displaying the most basal divergence, both support the hypothesis that Asiatic Cyprininae could be the ancestral hosts. However, the scenario involving basal divergence of lineage 3 proposes an equivocal solution where either representatives of the Cyprininae, Leuciscinae, or Percidae would be plesiomorphic hosts. In the latter case, if we consider that Dactylogyrus first evolved within percids, then it would imply that Dactylogyrus species became extinct within the Percidae, except in one species, G. cernuus, and at least in the Characidae and Balitoridae, which are the most basal taxa within Cypriniformes. Indeed, the majority of Dactylogyrus species infest representatives of the Cyprinidae. For these reasons, we consider this possibility very unlikely. The second solution is to consider the Leuciscinae as the primitive host subfamily for Dactyl-

1014

´ ET AL. ANDREA SˇIMKOVA

FIG. 6. Mapping of niche within hosts preferred by Dactylogyrus species. (A) Gill arch division: gill segments: D, dorsal; M, medial; V, ventral; gill area: d, distal; c, central; p, proximal; gill surface: in, inner; out, outer; gill hemibranch: A, anterior; P, posterior. (B) Gill arch, (C) gill segment, and (D) gill area were mapped onto the parasite phylogeny.

ogyrus. In that case, it would imply that Dactylogyrus originated in the ancestor of the Leuciscinae and secondarily colonized the Percidae and Cyprininae. This scenario also seems unlikely if we consider that east-central Asia is the center of speciation of the Cyprininae. How then can we explain host switches from European to Asian host fish, whereas the divergence of Asian cyprinins appears basal within the Cyprinidae (Fig. 3)? Indeed, C. auratus and C. carpio, the hosts for eight parasite species in our study, are considered to have originated in Asia and been recently imported in Europe. Furthermore, there is no paleontological evidence for their presence in Europe during the Pleistocene (Barusˇ and Oliva 1995). However, it should be noted that the origin of different subfamilies of the Cyprinidae and dispersal migration routes are still debated (Durand et al. 2002). When considering the above arguments and host subfamily mapping, we propose that Dactylogyrus originated within the Cyprininae and secondarily colonized ancestral hosts of the Leuciscinae and several species of the Percidae. Concerning

the distribution of the latter family, that is, its Africa and eastern Asia, limited diversification and extensive diversity in North America, it is that host transfer occurred in Europe from fish prininae.

absence in in Europe, very likely of the Cy-

Pattern and Process of Parasite Diversification The nested position of Dactylogyrus species parasitizing B. barbus within the species of the Leuciscinae suggests host switching from host fish of the Leuciscinae. It has been proposed that Western Palearctic Barbus would have originated in eastern Asia and migrated to Europe and Asia during the Miocene–Upper Oligocene (Tsigenopoulos and Berrebi 2000). However, many species belonging to the genus Barbus are of European origin (Cunha et al. 2002). It has also been shown that Barbus divergence is of the same order as the radiation of the Leuciscinae genera and that both are considered to have their center of speciation in Siberia (Durand

MOLECULAR PHYLOGENY OF DACTYLOGYRUS

FIG. 6.

et al. 2002). The biogeography of this species could explain when this capture of parasites has occurred. Whether numerous host switches can be inferred from the comparison of host and parasite relationships (Fig. 5), their interpretations would at the moment be too speculative. However, we can assume that parasites of B. barbus of the Cyprininae and of A. alburnus of the Alburninae are derived from parasites of Leuciscinae host fish. Although several nodes are not well supported, the molecular phylogeny shows a consistent pattern of relationships of Dactylogyrus species, as well as a very high number of intrahost speciation events (i.e., parasite duplications), compared to very few cospeciation events. Eight duplication events were recognized within lineage 3 across seven host species and three within lineage 1 across two host species. These speciation events gave rise mainly to specialist species, that is, parasite species that infest a single host species. However, there are several exceptions where two duplication events gave rise to both specialist and generalist parasite species. Within the first parasite clade of R. rutilus (D. ca-

1015

Continued.

balleroi, D. crucifer, and D. rarissimus), D. crucifer mainly infests R. rutilus and occasionally L. idus and S. erythrophthalmus. Within the second parasite clade of R. rutilus (D. rutili, D. nanus, and D. fallax), D. fallax has been more often reported from R. rutilus than from L. cephalus. Within the parasite clade of C. auratus (D. formosus and D. anchoratus), D. anchoratus has also been reported from C. carpio. However, because of low sampling across the two latter host species, we are currently unable to know which of these two host species is preferred for D. anchoratus. Nevertheless, sympatric speciation seems to be closely correlated to strict host specificity as suggested by Vickery and Poulin (1998). Our observations suggest that the majority of Dactylogyrus parasites first speciated by intrahost duplication and secondarily colonized new host species that are closely related. This confirms the suggestion of Gusev (1985) who defined socalled basic host species for many generalist species (host species when Dactylogyrus species occurs in high abundance) and rare host species (other phylogenetically closely related species). Our results contradict the investigations of Gue´ gan

1016

´ ET AL. ANDREA SˇIMKOVA

and Agne`se (1991) on congeneric African cyprinids and their Dactylogyrus species. The comparison of Dactylogyrus phylogeny based on morphological grounds with a host phylogeny (with fish species belonging only to a single genus) based on genetic markers revealed no intrahost speciation, but did suggest cospeciation and host switching events (Gue´ gan and Agne`se 1991). Parasite duplication is a coevolutionary scenario that has been recently considered to explain host-parasite relationships (Paterson and Gray 1997; Johnson et al. 2003). Paterson and Poulin (1999) showed that intrahost speciation could be more likely than cospeciation events within chondracanthid copepods and their teleost hosts. They suggested that the large geographical distances among hosts could explain this pattern of speciation. However, in the case of Dactylogyrus species, intrahost speciation is observed within hosts that are not geographically isolated but living in sympatry. In the tree reported in Figure 3, four of the several generalist species, D. alatus, D. similis, D. sphyrna, and D. vistulae, form a clade of large generalists with similar morphological features (mainly strongly developed central hooks and similar shape of the third pair of marginal hooks, which differentiate this group from other Dactylogyrus species). Although two species (D. alatus and D. similis) were sequenced from only one host species, their presence on other host species was recorded during this study (D. alatus on L. idus and D. similis on L. cephalus were not sequenced due to the low number of parasite individuals collected from those hosts). According to Gusev (1985) and Moravec (2001), those species may parasitize a wide host range. Like the generalists included among a majority of specialists from the molecular phylogeny (see above), those four generalist species also occur in high abundance in one host species and only rarely on other host species. Because those species form a clade, it could be hypothesized that the ancestral parasite was able to colonize a wide range of hosts with consequent speciation of new species within one host species. New parasite species preferentially choose this host species and occasionally infest closely related host species, maybe because of their large body size. The presence of several large species could preclude species coexistence. Molecular divergence among generalist Dactylogyrus species collected from the different host species was about 1%. This could suggest that there is low gene flow between different populations of the same Dactylogyrus species parasitizing different host species, which may potentially reduce interbreeding among the different populations. However, such suggestion needs further investigation in the future. When comparing the mode of Dactylogyrus species speciation with that of Gyrodactylus (also a monogenean genus with high number of parasite species), host switch was accepted for Gyrodactylus, and adaptive radiation was suggested to be the consequence of host switch to a new family (Zietara and Lumme 2002). Those differences may be explained by different reproductive strategy (viviparous Gyrodactylus vs. oviparous Dactylogyrus). In the case of Gyrodactylus species, only one specimen may be required achieve a successful switch to a new host if the host defense system is tolerated. As Dactylogyrus species form groups of phylogenetically

closely related species within a host, it could be hypothesized that their mode of speciation (i.e., intrahost speciation) should be closely related to the evolution of attachment organs (i.e., opisthaptor) morphology and reproductive system, as suggested by Rohde (1989). However, the morphology of the attachment apparatus of the Dactylogyrus genus is complex, including central and marginal hooks and connective bars. The number of the morphological characters studied so far would not allow for phylogenetic reconstruction based on morphology when using a parasite sample as large as the 51 species in the present study. The evolution of complex characters and the comparison of these with molecular evolution require further studies in the future. Gusev (1985) suggested that there are some morphological features, mainly in the sclerotized parts of opisthaptor and reproductive organs, that differ between Dactylogyrus species of Cyprininae and Leuciscinae subfamilies, and some of these may be common for several Dactylogyrus species occurring on the same or closely related species. However, there is no evidence that morphological characters help determine the groups within Dactylogyrus, in contrast to the genus Gyrodactylus, where the protonephridial system and opisthaptor have been used (Malmberg 1970). However, the morphological characters of the attachment apparatus and reproductive organs for Dactylogyrus species are sufficient for species discrimination, and there is no need to use the molecular markers for facilitating species discrimination, as suggested Zietara and Lumme (2002) for Gyrodactylus species. Our findings on intrahost speciation are closely related to hypotheses explaining species coexistence. Coexistence of Dactylogyrus species on the same host is facilitated by differing morphology of copulatory organs, or niche center distances, which prevent competition and increase intraspecific mating contacts (Sˇimkova´ et al. 2002). The present study shows that parasites speciating within one host (intrahost speciation) tend to occupy niches differing at least in one parameter (gill arch, segment, or area). The same result was obtained in a study of Dactylogyrus assemblages coexisting on one fish species (i.e., roach) by Sˇimkova´ et al. (2000). However, molecular phylogeny indicates that Dactylogyrus species parasitizing roach do not form a monophyletic group. In that case, isolation by niche segregation is important for congeneric species. Moreover, when different Dactylogyrus species occupy similar or closely situated niches, reproductive barriers could also reinforce reproductive isolation between congeners (Rohde and Hobbs 1986; Rohde 1991; Morand et al. 2002; Jarkovsky´ et al. 2004). Finally, the mapping of niche position within the fish gills indicates that there is an ancestral position (possibly a site more protected against water current), from which other positions seem to be derived. Conclusions Phylogenetic relationships of Dactylogyrus species combined with host subfamily mapping suggest that the genus originated recently within fish of the Cyprininae. They would have secondarily evolved by colonizing different host species of the subfamilies Rasborinae, Gobioninae, Leuciscinae, and Alburninae and the family Percidae by host switching and

MOLECULAR PHYLOGENY OF DACTYLOGYRUS

cospeciation. The main process of Dactylogyrus diversification corresponds to sympatric speciation (i.e., intrahost speciation), which seems correlated to strict host specificity. This kind of speciation was previously very rarely recognized in host-parasite associations and the model Dactylogyrus-Cyprinidae presents the first case study where intrahost speciation has been recorded for monogenean parasites. Moreover, the present study shows that congeneric parasites speciating within one host tend to occupy niches differing at least in one niche parameter. ACKNOWLEDGMENTS This study was supported by Research Project of the Masaryk University, Brno, project number J07/98:143100010. ASˇ was founded by the Grant Agency of the Czech Republic, project number 524/03/P108. The stay of ASˇ in France was supported by a postdoctoral fellowship from the Ministe` re de la Recherche in France. We would like to thank M. Ondra´cˇkova´, R. Sonek, M. Nova´kova´, and K. Korˇ´ınkova´ of Masaryk University, Brno, the Czech Republic, for their help in collecting material; P. Jurajda from the Institute of Vertebrate Biology, Brno, Czech Republic, in electrofishing. We are very grateful to C. O. Cunningham from FRS Marine Laboratory, Aberdeen, Scotland, for correction of English; J.-F. Martin from Centre de Biologie et de Gestion des Populations, Montferrier, France, and two reviewers for helpful comments. LITERATURE CITED Barusˇ, V., and O. Oliva. 1995. Petromyzontes and Osteichthyes. Academy of Science of Czech Republic, Prague. Bentz, S., S. Leroy, L. du Preez, J. Mariaux, C. Vaucher, and O. Verneau. 2001. Origin and evolution of African Polystoma (Monogenea: Polystomatidae) assessed by molecular methods. Int. J. Parasitol. 31:697–705. Booton, G. C., L. Kaufman, M. Chandler, R. Oguto-Ohwayo, W. Duand, and P. A. Fuerst. 1999. Evolution of the ribosomal RNA Internal transcribed spacer one (ITS-1) in cichlid fishes of the Lake Victoria region. Mol. Phylogenet. Evol. 11:273–282. Briolay, J., N. Galtier, R. M. Brito, and Y. Bouvet. 1998. Molecular phylogeny of Cyprinidae inferred from cytochrome b DNA sequences. Mol. Phylogenet. Evol. 9:100–108. Brooks, D. R., and D. A. McLennan. 1993. Parascript: parasites and the language of evolution. Smithsonian Institution Press, Washington, DC. Chen, X. L., P. Q. Yue, and R. D. Lin. 1984. Major groups within the family Cyprinidae and their phylogenetic relationships. Acta Zootaxonomica Sin. 9:424–498. Cone, D., T. Eurell, and V. Beasley. 1994. A report of Dactylogyrus amphibothrium (Monogenea) on the gills of European ruffe in western Lake Superior. J. Parasitol. 80(3):476–478. Coyne, J. A., and T. D. Price. 2000. Little evidence for sympatric speciation in island birds. Evolution 54:2166–2171. Cunha, C., N. Mesquita, T. E. Dowling, A. Gilles, and M. M. Coelho. 2002. Phylogenetic relationships of Eurasian and American cyprinids using cytochrome b sequences. J. Fish Biol. 61: 929–944. Desdevises, Y., S. Morand, O. Jousson, and P. Legendre. 2002. Coevolution between Lamellodiscus (Monogenea: Diplectanidae) and Sparidae (Teleostei): the study of a complex host-parasite system. Evolution 56:2459–2471. Dieckmann, U., and M. Doebeli. 1999. On the origin of species by sympatric speciation. Nature 400:354–357. ¨ nlu¨, and P. Berrebi. 2002. Durand, J.-D., C. S. Tsigenopoulos, E. U Phylogeny and biogeography of the family Cyprinidae in the

1017

Middle East inferred from cytochrome b DNA: evolutionary significance of this region. Mol. Phylogenet. Evol. 22:91–100. El Gharbi, S., E. Birgi, and A. Lambert. 1994. Monoge`nes Dactylogyridae parasites de Cyprinidae du genre Barbus d’Afrique Nord. Syst. Parasitol. 27:45–70. Ergens, R., and J. Lom. 1970. Causative agents of fish diseases. (In Czech.) Academia Praha, Prague. Felsenstein, J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39:783–791. Gibson, D. I., T. A. Timofeeva, and P. I. Gerasev. 1996. A catalogue of the nominal species of the monogenean genus Dactylogyrus Diesing, 1850 and their host genera. Syst. Parasitol. 35:3–48. Gilles, A., G. Lecointre, E. Faure, R. Chappaz, and G. Brun. 1998. Mitochondrial phylogeny of the European cyprinids: implications for their systematics, reticulate evolution, and colonization time. Mol. Phylogenet. Evol. 10:132–143. Gilles, A., G. Lecointre, A. Miquelis, M. Loerstcher, R. Chappaz, and G. Brun. 2001. Partial combination applied to phylogeny of European cyprinids using the mitochondrial control region. Mol. Phylogenet. Evol. 19:22–33. Gue´gan, J.-F., and J.-F. Agne`se. 1991. Parasite evolutionary events inferred from host phylogeny: the case of Labeo species (Teleostei, Cyprinidae) and their dactylogyrid parasites (Monogenea, Dactylogyridae). Can. J. Zool. 69:595–603. Gue´gan, J.-F., and A. Lambert. 1991. Dactylogyrids (Platyhelminthes: Monogenea) of Labeo (Teleostei: Cyprinidae) from West African coastal rivers. J. Helminthol. Soc. W. 58:85–99. Gusev, A. V. 1985. Metazoan parasites. I. Identification key to parasites of freshwater Fish. Vol. 2. O. N. Bauer, ed. Nauka, Leningrad. (In Russian) Hafner, M. S., and S. A. Nadler. 1988. Phylogenetic trees support the coevolution of parasites and their hosts. Nature 332:258–259. Hafner, M. S., P. D. Sudman, F. X. Villablanca, T. A. Spradling, J. W. Demestes, and S. A. Nadler. 1994. Disparate rates of molecular evolution in cospeciating hosts and parasites. Science 265:1087–1090. Hayward, C. J. 1997. Helminth ectoparasites of sillaginid fishes (Perciformes: Percoidei) have low species richness. Folia Parasitol. 44(3):173–187. Helfman, G. S., B. B. Collette, and D. E. Facey. 1997. The diversity of fishes. Blackwell Science, Madlen, MA. Hillis, D. M., and M. T. Dixon. 1991. Ribosomal DNA: molecular evolution and phylogenetic inference. Q. Rev. Biol. 66:411–453. Huyse, T., and A. M. Volckaert. 2002. Tangled trees: comparing the phylogenies of Gyrodactylus and its goby host. Pp. 157–175 in T. Huyse, ed. Evolutionary associations between Gyrodactylus and its goby host: Bound forever? Ph.D. diss., Katholieke Universiteit Leuven, Belgium. Jarkovsky´, J., S. Morand, A. Sˇimkova´, and M. Gelnar. 2004. Reproductive barriers between congeneric monogenean parasites (Dactylogyrus: Monogenea): attachment apparatus morphology or copulatory organ incompatibility? Parasitol. Res. 92:95–105. Johnson, K. P., R. J. Adams, R. D. M. Page, and D. H. Clayton. 2003. When do parasite fail to speciate in response to host speciation? Syst. Biol. 52:37–47. Kennedy, C. R., and A. O. Bush. 1992. Species richness in helminth communities: the importance of multiple congeners. Parasitology 104:189–197. Kimura, M. 1980. A simple method for estimating evolutionary rate of base substitutions through comparative studies of nucleotide sequences. J. Mol. Evol. 16:111–120. Klassen, G. J., and M. Beverley-Burton. 1987. Phylogenetic relationship of Ligictaluridus spp. (Monogenea: Ancyrocephalidae) and their ictalurid (Siluriformes) hosts: an hypothesis. Proc. Helminthol. Soc. Wash. 54:84–90. Lambert, A., and S. El Gharbi. 1995. Monogenean host specificity as a biological and taxonomic indicator for fish. Biol. Conserv. 72:227–235. Li, W.-H. 1997. Molecular evolution. Sinauer Associates, Sunderland, MA. Littlewood, D. T. J., and P. D. Olson. 2001. Small subunit rDNA and the Platyhelminthes: signal, noise, conflict and compromise. Pp. 262–278 in D. T. J. Littlewood and R. A. Bray, eds. Inter-

1018

´ ET AL. ANDREA SˇIMKOVA

relationships of the Platyhelminthes. Taylor and Francis, London. Littlewood, D. T. J., K. Rohde, and K. A. Clough. 1997. Parasite speciation within or between host species? Phylogenetic evidence from site-specific polystome Monogeneans. Int. J. Parasitol. 27:1289–1297. Maddison, W. P., and D. R. Maddison. 1992. MacClade: analysis of phylogeny and character evolution. Sinauer Associates, Sunderland, MA. Malmberg, G. 1970. The excretory systems and the marginal hooks as a basis for the systematics of Gyrodactylus (Trematoda, Monogenea). Ark. Zool. 23:1–235. Morand, S., A. Sˇimkova´, I. Matejusova´, L. Plaisance, O. Verneau, and Y. Desdevises. 2002. Investigating patterns may reveal processes: evolutionary ecology of monogeneans. Int. J. Parasitol. 32:111–119. Moravec, F. 2001. Checklist of the Metazoan parasites of fishes of the Czech Republic and the Slovak Republic. Academia Praha, Prague. Nei, M., and S. Kumar. 2000. Molecular evolution and phylogenetics. Oxford Univ. Press, New York. Nelson, J. S. 1994. Fishes of the world. John Wiley & Sons, New York. Olson, P. D., and D. T. J. Littlewood. 2002. Phylogenetics of the Monogenea: evidence from a medley of molecules. Int. J. Parasitol. 32:233–244. Page, R. D. M. 1994. Parallel phylogenies: reconstructing the history of host-parasite assemblages. Cladistics 10:155–173. Page, R. D. M., P. L. M. Lee, S. A. Becher, R. Griffiths, and D. H. Clayton. 1998. A different tempo of mitochondrial DNA evolution in birds and their parasitic lice. Mol. Phylogenet. Evol. 9:276–293. Paterson, A. M., and J. Banks. 2001. Analytical approaches to measuring cospeciation of host and parasites: through a glass, darkly. Int. J. Parasitol. 31:1012–1022. Paterson, A. M., and R. D. Gray. 1997. Host-parasite co-speciation, host switching and missing the boat. Pp. 236–250 in D. H. Clayton and J. Moore, eds. Host-parasite evolution: general principles and avian models. Oxford Univ. Press, Oxford, U.K. Paterson, A. M., and R. Poulin. 1999. Have chondracanthid copepods co-speciated with their hosts? Syst. Parasitol. 44:79–85. Philippe, H. 1993. MUST, a computer package for management utilitarians for sequences and trees. Nucleic Acids Res. 21: 5264–5272. Posada, D., and K. A. Crandall. 1998. Modeltest: testing the model of DNA substitution. Bioinformatics 14:817–818. Poulin, R. 2002. The evolution of monogenean diversity. Int. J. Parasitol. 32:245–254. Rohde, K. 1977. A non-competitive mechanism responsible for restricting niches. Zool. Anz. 199:164–172. ———. 1979. A critical evaluation of intrinsic and extrinsic factors responsible for niche restriction in parasites. Am. Nat. 114: 648–671. ———. 1989. Simple ecological systems, simple solutions to complex problems? Evol. Theor. 8:305–350. ———. 1991. Intra- and interspecific interactions in low density populations in resource-rich habitats. Oikos 60:91–104. Rohde, K., and R. Hobbs. 1986. Species segregation: competition

or reinforcement of reproductive barriers? Pp. 189–199 in M. Cremin, C. Dobson, and E. Noorhouse, eds. Parasites lives, papers on parasites, their hosts and their association to Honour. JFA Sprent, University of Queensland Press, St. Lucia. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Sekar, V. 1987. A rapid screening procedure for the identification of recombinant bacterial clones. Biotechniques 5:11–13. Sˇimkova´, A., Y. Desdevises, M. Gelnar, and S. Morand. 2000. Coexistence of nine gill ectoparasites (Dactylogyrus: Monogenea) parasitizing the roach (Rutilus rutilus L.): history and present ecology. Int. J. Parasitol. 30:1077–1088. ———. 2001. Morphometric correlates of host specificity in Dactylogyrus species (Monogenea) parasites of European cyprinid fish. Parasitology 123:169–177. Sˇimkova´, A., M. Ondracˇkova´, M. Gelnar, and S. Morand. 2002. Morphology and coexistence of congeneric ectoparasite species: reinforcement of reproductive isolation? Biol. J. Linn. Soc. 76: 125–135. Sˇimkova´, A., L. Plaisance, I. Matejusova´, S. Morand, and O. Verneau. 2003. Phylogenetic relationships of the Dactylogyridae Bychowsky, 1933 (Monogenea: Dactylogyridea): the need for the systematic revision of the Ancyrocephalinae Bychowsky, 1937. Syst. Parasitol. 54:1–11. Sinnappah, N. D., L. H. S. Lim, K. Rohde, R. Tinsley, C. Combes, and O. Verneau. 2001. A paedomorphic parasite associated with a neotenic amphibian host: phylogenetic evidence suggests a revised systematic position for Sphyranuridae within anuran and turtle polystomatoineans. Mol. Phylogenet. Evol. 18:189–201. Swofford, D. L. 2002. PAUP*: phylogenetic analysis using parsimony and other methods. Version 4.0b10. Sinauer Associates, Sunderland, MA. Tsigenopoulos, C. S., and P. Berrebi. 2000. Molecular phylogeny of north Mediterranean freshwater barbs (genus Barbus: Cyprinidae) inferred from cytochrome b sequences: biogeographic and systematic implications. Mol. Phylogenet. Evol. 14:165–179. Valtonen, E. T., M. Prot, and R. Rahkonen. 1990. Seasonality of two monogeneans from two freshwater fish from an oligotrophic lake in northeast Finland. Int. J. Parasitol. 20:101–107. Vickery, W. L., and R. Poulin. 1998. Parasite extinction and colonisation and the evolution of parasite communities: a simulation study. Int. J. Parasitol. 28:727–737. Winfield, I. J., and J. S. Nelson. 1991. Cyprinid fishes: Systematics, biology and exploitation. Fish and Fisheries Series no. 3. Chapman and Hall, London. Zardoya, R., and I. Doadrio. 1999. Molecular evidence on the evolutionary and biogeographical patterns of European cyprinids. J. Mol. Evol. 49:227–237. Zardoya, R., P. S. Economidis, and I. Doadrio. 1999. Phylogenetic relationships of Greek Cyprinidae: molecular evidence for at least two origins of the Greek cyprinid fauna. Mol. Phylogenet. Evol. 13:122–131. Zietara, M. S., and J. Lumme. 2002. Speciation by host switch and adaptive radiation in a fish parasite genus Gyrodactylus (Monogenea, Gyrodactylidae). Evolution 56:2445–2458. Corresponding Editor: R. Poulin