Copeia, 2001(3), pp. 656±667 Phylogenetic Relationships of Cyprinid Fishes in Subgenus Notropis Inferred from Nucleotide Sequences of the Mitochondrially Encoded Cytochrome b Gene JOSEPH P. B IELAWSKI AND JOHN R. GOLD Cytochrome b sequences (1140 bp) from 16 species considered to be members of subgenus Notropis and several outgroup taxa were analyzed phylogenetically. Phy- logenetic analysis of cytochrome b was sensitive to sampling of outgroup taxa, prob- ably resulting from substitutional rate heterogeneity among outgroups. Maximum- likelihood analysis was more robust to these effects as compared to maximum par- simony. Both maximum parsimony and maximum-likelihood analysis supported re- moval of Notropis candidus and Notropis shumardi from subgenus Notropis. Monophyly of the 14 remaining species was supported by the maximum likelihood analysis. Regardless of the method of analysis or use of outgroup taxa, phylogenetic analysis of cytochrome b indicated strong support for the following hypothesis of relation- ships: ((Notropis rubellus, Notropis suttkusi), (Notropis amoenus, Notropis stilbius, ((No- tropis atherinoides, Notropis oxyrhynchus), (Notropis amabilis, Notropis jemezanus)))). This phylogenetic hypothesis implies that morphological characters associated with increased olfaction were derived independently among members of subgenus No- tropis, perhaps in response to life in more turbid habitats. HE subgenus Notropis is an ecologically di- group (Notropis shumardi, Notropis candidus, N. T verse yet morphologically conservative oxyrhynchus, Notropis. jemezanus); (3) the photo- group of North American cyprinids. Distribu- genis group (Notropis photogenis, N. amoenus, No- tions of the 17 nominal species range from tropis stilbius); and (4) the ariommus group (N. highly restricted (Notropis oxyrhynchus, found ariommus, Notropis scepticus, Notropis semperasper, only in the Brazos River, TX) to nearly conti- Notropis telescopus). Coburn (1982) proposed a nental (Notropis atherinoides, found throughout hypothesis of relationships for these species the central United States from the Gulf of Mex- groups, placing the shumardi group sister to the ico to Canada). Habitats are primarily riverine atherinoides group and the photogenis group sister and range from broad channels of turbid rivers to the ariommus group. (Notropis girardi) to clear creeks and streams Aside from placement of N. rubellus (formerly (Notropis ariommus). Subgenus Notropis are of subgenus Hydrophlox) in subgenus Notropis broadcast spawners, scattering eggs and sperm (Dowling and Brown, 1989; Mayden, 1991), without preparation of substrate, although at modi®cation of subgenus Notropis since it was least two species, Notropis amoenus and Notropis de®ned by Coburn (1982) has centered primar- rubellus, are known to spawn as nest associates ily on members of the N. shumardi species with other species ( Johnson and Page, 1992). group. Based upon a cladistic analysis of varia- Members of subgenus Notropis are relatively tion in chromosomal NORs, Amemiya and Gold small and round bodied and lack striking col- (1990) hypothesized that N. oxyrhynchus (for- oration. However, two species, N. rubellus and merly of N. shumardi species group) was more Notropis suttkusi, display red and orange colora- closely related to N. atherinoides (N. atherinoides tion on the head and base of ®ns during the species group). Based upon reinterpretation of breeding season. characters associated with the olfactory capsule, Coburn (1982) conducted a detailed mor- Coburn and Cavender (1992) placed N. girardi phological study of N. atherinoides, the type spe- (formerly of subgenus Alburnops) in the N. shu- cies for genus Notropis, and employed a cladistic mardi species group and moved N. oxyrhynchus analysis to de®ne a monophyletic subgenus No- (formerly of N. shumardi species group) into the tropis based upon synapomorphies shared be- N. atherinoides species group. Coburn and Cav- tween N. atherinoides and other species of genus ender (1992) also raised the question of af®ni- Notropis. Coburn (1982) described four distinct ties of species placed in the N. shumardi species species-groups within subgenus Notropis: (1) the group by noting that characters of the olfactory atherinoides group (N. atherinoides, N. amabilis, N. capsule used to diagnose the N. shumardi species rubellus, Notropis perpallidus); (2) the shumardi group were distributed widely outside of sub- q 2001 by the American Society of Ichthyologists and Herpetologists BIELAWSKI AND GOLDÐPHYLOGENETICS OF SUBGENUS NOTROPIS 657 genus Notropis. Prior to the analysis of Coburn nitrogen or 95% ethanol, and stored at 280 C (1982), N. shumardi was placed outside of sub- until processed. Collection localities, drainage, genus Notropis in subgenus Alburnops (Snelson, and catalog number of voucher specimens are 1968). Two recently described species, N. can- presented in the Materials Examined section. A didus (Suttkus, 1980) and N. suttkusi (Hum- single specimen of Notropis semperasper was col- phries and Cashner, 1994), represent geograph- lected from the James River, Virginia, but reli- ically distinct populations formerly considered able sequences of its cyt b gene could not be to be part of N. shumardi and N. rubellus, re- obtained. The cyt b gene (1140 bp) was se- spectively. quenced directly for representatives of 24 spe- The purpose of the present investigation was cies of Notropis; cyt b sequences of N. rubellus to use complete sequences of the mitochondri- were kindly provided by T. E. Dowling of Ari- ally encoded cytochrome b (cyt b) gene to infer zona State University. a hypothesis of phylogenetic relationships among species of subgenus Notropis and, in par- DNA extraction, ampli®cation, and sequencing.Ð ticular, to investigate phylogenetic af®nities of Whole ®sh were ground in liquid nitrogen, and species in the N. shumardi species group. Mito- DNA was extracted by phenol-chloroform treat- chondrial genes have been useful in studies of ment and ethanol precipitation (Sambrook et the phylogenetic relationships of several groups al., 1989). DNA preparations were used as tem- of North American cyprinids. Schmidt and Gold plate for primer-directed ampli®cation of target (1995) used partial cyt b sequences to infer phy- DNA sequences via polymerase chain reaction logenetic af®nities of Notropis topeka, and com- (PCR). PCR ampli®cation used the primers LA, plete cyt b sequences provided the basis for a LC, HA, HB, and HD described in Schmidt et hypothesis of relationships among species of the al. (1998). PCR thermal pro®le consisted of 35± cyprinid genera Luxilus (Dowling and Naylor, 45 cycles of 95 C denaturation for 1 min, 48±50 1997) and Lythrurus (Schmidt et al., 1998). Cyt C annealing for 1 min, and 72 C extension for b even has been used to investigate cyprinid re- 45 sec. Excess primers, nucleotides, and poly- lationships at the subfamilial level (Briolay et merase were removed from DNA ampli®cation al., 1998). Other mitochondrial genes also have products by using Prep-A-Genet DNA puri®ca- been used effectively in studies of North Amer- tion system (BioRad). Double-stranded DNA ican cyprinids (e.g., Simons and Mayden, 1997; ampli®cation products were sequenced directly Broughton and Gold 2000). with ABI PRISMy (Perkin Elmer) dye-termina- Five species of subgenus Notropis (N. candidus, tor cycle-sequencing kits and an Applied Biosys- N. girardi, N. jemezanus, N. oxyrhynchus, and N. tems (Perkin Elmer) 377 automated DNA se- shumardi) exhibit pronounced enlargement of quencer. Sequencing used primers LA, LC, HA, brain structures (olfactory bulb and facial lobe) HB, and HD described in Schmidt et al. (1998). associated with olfaction (Coburn, 1982, pers. The cyt b sequence of an individual was deter- comm.). Previously, four of these were consid- mined from a minimum of two independent se- ered a monophyletic group (shumardi species quencing-reactions for each primer. group; Coburn 1982). However, with the excep- tion of N. shumardi and N. candidus, our analysis Patterns of sequence variation.ÐMeasures of nu- of cyt b indicated that species with pronounced cleotide composition were obtained by using enlargement of brain structures associated with PAUP* version 4.0d63 (D. L. Swofford, Sinauer, olfaction were not closely related. Because these Sunderland, MA, 1998, unpubl.). Homogeneity species also tend to prefer more turbid waters of nucleotide composition was tested among (Huber and Rylander, 1992), our ®ndings sup- taxa by using chi-square tests of contingency ta- port the hypothesis that enlarged olfactory ca- bles of nucleotide counts, as implemented in pabilities might have evolved within subgenus PAUP*. Maximum likelihood provided the sta- Notropis to compensate for reduced visibility in tistical basis for testing expectations of the mo- turbid waters of large rivers. lecular clock (Yang et al., 1995). Likelihood ra- tio tests of the molecular clock were conducted by using the optimal topology obtained from MATERIALS AND METHODS maximum-likelihood analysis of all ingroup and Taxa examined.ÐCyt b sequences were acquired outgroup taxa. from 16 of 17 nominal species of subgenus No- tropis and from eight species belonging to ®ve Maximum-likelihood analysis of DNA substitution other assemblages within genus Notropis (see models.ÐPrior to estimating a topology, an op- Materials Examined). Whole ®sh were captured timal DNA substitution model for analysis of the with minnow seines, transported in either liquid cyt b was selected