- Email: [email protected]
New World monkey phylogeny based on X-linked G6PD DNA sequences
MOLECULAR PHYLOGENETICS AND EVOLUTION Molecular Phylogenetics and Evolution 27 (2003) 121–130 www.elsevier.com/locate/ympev
New World monkey phylogeny based on X-linked G6PD DNA sequences Michael E. Steiper* and Maryellen Ruvolo Department of Anthropology, Harvard University, 11 Divinity Avenue, Cambridge, MA 02138, USA Received 24 May 2002; revised 16 August 2002
Abstract The Platyrrhini, or New World monkeys, are an infraorder of Primates comprised of 16 genera. Molecular phylogenetic analyses have consistently sorted these genera into three groups: the Pitheciidae (e.g., saki and titi monkeys), Atelidae (e.g., spider and howler monkeys), and Cebidae (e.g., night monkeys, squirrel monkeys, and tamarins). No consensus has emerged on the relationships among the three groups or within the Cebidae. Here, 0.8 kb of newly generated intronic DNA sequence data from the X-linked glucose-6-phosphate dehydrogenase (G6PD) locus have been collected from nine New World monkey taxa to examine these relationships. These data are added to 1.3 kb of previously generated G6PD intronic DNA sequence data [Mol. Phylogenet. Evol. 11 (1999) 459]. Using distance and parsimony-based techniques, G6PD sequences provide support for an initial bifurcation between the Pitheciidae and the remaining platyrrhines, linking Atelidae and Cebidae as sister taxa. Bayesian methods provided a conflicting phylogeny with Atelidae as outgroup. Within the Cebidae, a sister relation between Aotus and the Cebus/Saimiri clade is favored by parsimony analysis, but not by other analyses. Potential reasons for the difficulty in resolving family level New World monkey phylogenetics are discussed. Ó 2002 Elsevier Science (USA). All rights reserved. Keywords: New World monkeys; Platyrrhines; Platyrrhini; Glucose-6-phosphate dehydrogenase; G6PD; Phylogenetics; Cebidae; Atelidae; Pitheciidae
1. Introduction The Platyrrhini, or New World monkeys, are an infraorder of Primates comprised of approximately 16 genera. The group is fairly ancient, having originated in the Oligocene, and New World monkeys currently range from southern Mexico to northern Argentina (Fleagle, 1999). The phylogenetics of this group have been addressed by multiple researchers using both morphological (e.g., Ford, 1986; Kay, 1990; Rosenberger, 1984) and molecular data sets, including nuclear sequences from e-globin (Harada et al., 1995; Porter et al., 1997a, 1999, 1997b, 1995; Schneider et al., 1993), interphotoreceptor retinoid-binding protein (IRBP) (Barroso et al., 1997; Harada et al., 1995; Schneider et al., 1996), G6PD (von Dornum and Ruvolo, 1999), b2 -microglobulin
* Corresponding author. Fax: 1-617-496-8041. E-mail address: [email protected] (M.E. Steiper).
(Canavez et al., 1999), von Willebrand factor (vWF) (Chaves et al., 1999), and mitochondrial sequences from 12S and 16S rRNA (Horovitz and Meyer, 1995; Horovitz et al., 1998) and COII (von Dornum, 1997). One main result from these analyses is that New World monkeys can be roughly sorted into six groups, which are congruent with the Ônatural groupsÕ of Groves (1989): (1) callitrichids, including Callimico; (2) capuchins (Cebus) and squirrel monkeys (Saimiri); (3) owl monkeys (Aotus); (4) the Pitheciinae (e.g., saki monkeys); (5) titi monkeys (Callicebus); and (6) the Atelinae (e.g., spider monkeys) with Alouatta (howler monkeys). Based on molecular analyses, these six groups consistently fall into three monophyletic groups, the Cebidae (1, 2, and 3), the Pitheciidae (4 and 5), and the Atelidae (6), which Goodman et al. (1998) rank as families. Two major questions remain regarding platyrrhine phylogeny: First, what are the relationships among these three monophyletic families? Second, what are the species relationships within the Cebidae?
1055-7903/02/$ - see front matter Ó 2002 Elsevier Science (USA). All rights reserved. doi:10.1016/S1055-7903(02)00375-5
122
M.E. Steiper, M. Ruvolo / Molecular Phylogenetics and Evolution 27 (2003) 121–130
Here, 0.8 kb of newly generated intronic DNA sequence data have been collected from the X-linked glucose-6-phosphate dehydrogenase (G6PD) locus bearing on these questions. These data are added to 1.3 kb of intronic DNA sequence data from G6PD previously generated by von Dornum and Ruvolo (1999) (and abbreviated here as G6PD DE). The new data are from two downstream introns, termed G and H. At least one taxon from each of GrovesÕ (1989) six well-supported Ônatural groupsÕ is sampled in this data set (including two callitrichids [Leontopithecus and Saguinus], both Cebus and Saimiri, and both Ateles and Alouatta). In this study, Aotus, Cebus, Saimiri, Leontopithecus, and Saguinus represent Cebidae; Callicebus and Pithecia represent Pitheciidae; and Ateles and Alouatta represent Atelidae. Together, this enhanced G6PD data set (abbreviated as G6PD DEGH) comprises 2.1 kb of DNA sequence data from nine New World monkey species.
2. Materials and methods 2.1. Data collection DNA sequence data were collected from nine New World monkeys. The following DNA sequences were prepared from the same specimens as reported in von Dornum and Ruvolo (1999): Ateles fusciceps (Atf1 in von Dornum and Ruvolo, 1999), Saimiri boliviensis (Smb1), Alouatta caraya (Alc1), Pithecia pithecia (Pip1), Saguinus fuscicollis (Sgf1), and Leontopithecus rosalia (Lnr1). Genomic DNA was prepared as in von Dornum and Ruvolo (1999). The remaining sequences were generated partly from samples in von Dornum and Ruvolo (1999) and partly from samples from other individuals because the original DNA samples were depleted, as follows. The Callicebus sequence is from Callicebus moloch (Ccm1) for introns D, E, and G, while intron H sequence is from Callicebus moloch (Davis RPRC ID# CMO-00005). The Cebus DNA sequence is from Cebus apella (Cbp8) for introns D, E, and G, and intron H is from Cebus capuchinus (Bronx Zoo ID# 881198/200493). The Aotus sequence is from Aotus trivirgatus (Aot2) for introns D and E, while introns G and H are from A. trivirgatus (Bronx Zoo ID# 646-96). Both intron G and intron H were amplified using nested or semi-nested PCR, in which two rounds of PCR are used to obtain the desired DNA fragment. After the first round, a diluted product is used as a template for the second reaction, and the product from the second reaction is used for DNA sequencing. PCR amplification was performed using either the AmpliTaq Gold Kit (Perkin–Elmer, Foster City, CA) or Taq PCR Master Mix Kit (Qiagen, Valencia, CA). Reactions were 25 ll in size and contained 2.5 ll GeneAmp 10 PCR
buffer (final concentration 1, provided in kit), 10 mM of each dNTP, 1.125 U AmpliTaq Gold, and 1–5 ll unquantified template DNA. For both PCR rounds, an initial denaturation step (10 min with AmpliTaq Gold or 4 min with Qiagen Taq; 95 °C) was followed by 40 cycles of denaturation (45 s; 95 °C), annealing (45 s; temperatures given in Table 1), and extension (60 s; 72 °C) using Perkin–Elmer 9600 and 9700 thermocyclers. After cycling, reactions were held for 10 min at 72 °C and stored at 4 °C. After the first round of PCR, an aliquot of this reaction was diluted from [1:5] to [1:100] and used as template DNA for the second round of PCR. DNA fragments obtained at the second round were sequenced. MgCl2 concentrations, annealing temperatures, and primer sets used for each specific PCR are given in Table 1, and primer sequences in Table 2. Fragments were electrophoretically separated on 1–2% agarose gels. Ethidium bromide stained gels were visualized using UV light and a /v 174 RF/HaeIII digested DNA (GibcoBRL, Gaithersburg, MD) size standard. If multiple DNA bands were present, the desired fragment was excised from the gel and purified using a QIAquick Gel Extraction Kit (Qiagen, Valencia, CA). When only the desired band was present, PCR samples were run through a Microcon membrane column (Millipore, Bedford, MA) to remove unincorporated oligonucleotides. After purification, the DNA was again electrophoretically separated to roughly estimate DNA concentration for sequencing.
Table 1 PCR parameters and primer combinations used for G6PD sequencing. ‘‘F’’ indicates forward primer; ‘‘R,’’ reverse Round one Intron G Callicebus and Aotus AotGE-P004 (F); AotGH-M005 (R) Anneal 53 °C; 3.0 mM MgCl2 All other species G7-P92 (F); NWG8-M002 (R) Anneal 56 °C; 2.5 mM MgCl2 Intron H Callicebus AotGE-P004 (F); G9-M39 (R) Anneal 53 °C; 2.5 mM MgCl2 All other species GG-P168 (F); G9-M39 (R) Anneal 52 °C; 2.5 mM MgCl2
Round two
NWG7-P001 (F); NWG8-M002 (R) Anneal 53 °C; 3.0 mM MgCl2 NWG7-P001 (F); NWG8-M002 (R) Anneal 56 °C; 2.5 mM MgCl2
NWG7-001 (F); CMOG9-M005 (R) Anneal 48 °C; 2.5 mM MgCl2 NWG8-P003 (F); G9-M39 (R) Anneal 52 °C; 2.5 mM MgCl2
M.E. Steiper, M. Ruvolo / Molecular Phylogenetics and Evolution 27 (2003) 121–130 Table 2 Primers used in this study to amplify and sequence G6PD introns G and H in New World monkeys AotGE-P005 AotGH-M004 CMOG9-M005 G7-P92 G9-M39 GG-P168 NWG7-P001 NWG8-M002 NWG8-P003
50 -GGTGTTGAGCCAGGGGGTCG-30 50 -ACAGATGAGCCCACGACGGAG-30 50 -CTGCACCTCTGAGATAC-30 50 -AAGGAGCCCTTTGGCACTGAGGGTC-30 50 -CAATGTGGTCCTGGGCCAGTACGTG-30 50 -AGGGCTGGACCCCTACACAG-30 50 -GGGTCGTGGGGGCTATTTCG-30 50 -CACGGACGTCATCAGAGTTG-30 50 -CCCGCCTCCACCAACTCTG-30
2.2. Sequencing DNA was fluorescently labeled for sequencing with the dye-terminator method using the ABI PRISM dye terminator cycle sequencing ready reaction kit (Perkin– Elmer, Foster City, CA). Each 20 ll sequencing reaction contained 8 ll terminator ready reaction mix, 3.2 pmol primer (PCR primers were used for sequencing), and 3–8 ll template. Reactions were brought to 20 ll volume with deionized water. Using a Perkin– Elmer 9600 thermocycler, each reaction was cycled 25 times at 96 °C for 10 s, 50 °C for 5 s, and 60 °C for 4 min. After cycling, reactions were purified using Centri-Sep columns (Princeton Separations) or Edge Gel filtration/Centri-Flex columns (Edge Biosystems/ AGTC, Gaithersburg, MD), then dried, and stored at )20 °C. Dried sequencing reactions were resuspended with 2.5–5 ml 5:1 formamide:25 mM EDTA with blue dextran (50 mg/ml). Samples were denatured and electrophoresed for 14 h through 6% Long Ranger acrylamide gels on an ABI 373 automated DNA sequencer. Data in the form of electropherograms were collected using Data Collection v.1.2.1 (ABI). To obtain each speciesÕ sequence at least two electropherograms, usually derived from both DNA strands, were collected (i.e., DNA was sequenced in both directions). Electropherograms were processed with AutoAssembler v.1.4 (ABI) to generate one consensus sequence for each taxon. 2.3. Alignment DNA sequences from G and H were first aligned by ClustalW (Thompson et al., 1994) and then manually confirmed by eye. G6PD G and H DNA sequences for the nine taxa were added to the alignment of von Dornum and Ruvolo (1999) for G6PD D and E introns without modification. Homo sapiens was used as outgroup (GenBank Accession No. X55448). The final alignment was 2110 bp in length and the data set was then converted to Nexus file format for phylogenetic analyses.
123
2.4. Distance analysis Genetic distances between sequences were calculated for tree building and calculating branch lengths. Distance trees were built using the neighbor joining distance algorithm (Saitou and Nei, 1987) and the correction method of Hasegawa, Kishino, and Yano (HKY) (1985) in PAUP* v.4b4-10 (Swofford, 2002). 2.5. Parsimony analysis The most parsimonious tree reconstructions for the G6PD DEGH data set were found using an exhaustive search in PAUP* v.4b4-10 (Swofford, 2002). A weighting matrix was used in which tranversions count as two steps, and transitions, one. This matrix is based on the empirically observed excess of transitions in observed molecular data (Wakeley, 1996). When coded as informative data, alignment gaps do not yield different tree topologies, even under different weighting schemes. For all trees shown here, gaps are ignored. Bootstrap values were calculated using the same parameters, with 1000 resampling replicates of 10 heuristic search additions each. Unambiguous sites were culled from the PAUP* apomorphy list. 2.6. Maximum likelihood (ML) analysis The most likely tree reconstruction for the G6PD DEGH data set was estimated with a heuristic search consisting of 50 random taxon additions in PAUP* v.4b4-10 (Swofford, 2002). Initially, five ML search replicates were used to estimate the transition to transversion ratio and the c correction (based on six classes). Homo was the chosen outgroup. These estimates were then used in the larger search. Bootstrap values were calculated using the same parameters, with 500 resampling replicates of 10 heuristic search additions each. 2.7. Bayesian analysis A Bayesian analysis of the data set was implemented using MrBayes (Huelsenbeck and Ronquist, 2001). A general time-reversible model of molecular evolution was chosen. A 6-category gamma shape parameter for rate variation was estimated. Base frequencies were empirically derived. Homo was the chosen outgroup. No prior trees were defined; a random initial tree was generated. The analysis was run for 1,000,000 generations, saving a tree every 100 generations. For the Markov Chain Monte Carlo (MCMC) analysis, four chains were used. The initial 10,000 generations or 1000 trees were removed from the analysis as Ôburn-in,Õ the phase of the MCMC analysis where likelihoods are rapidly changing, before finally leveling off. The posterior probabilities for individual clades and alternate taxon arrangements were
124
M.E. Steiper, M. Ruvolo / Molecular Phylogenetics and Evolution 27 (2003) 121–130
generated along with a consensus tree using the sumt command. This analysis was carried out four times, with concordant results.
3. Results The New World monkey G6PD DNA sequence data from introns G and H are shown as an alignment in Fig. 1. Intron G is approximately 393 bp in length. Intron H is approximately 450 bp long, but only 429 bases were used in the analyses because several bases could not be verified in the 30 -most region. The total alignment of all four introns was 2110 bp in length. The pairwise HKY distance matrix for the data set is given in Table 3. One most parsimonious (MP) tree was reconstructed for this alignment (Fig. 2). The tree was 985 steps in length and had a consistency index of 0.88. Bootstrap values are
indicated above tree branches; the numbers of unambiguous sites supporting each branch are shown below. In this tree Pitheciidae is the most basal platyrrhine lineage and Aotus is allied to the Cebus/Saimiri clade. The distance tree (Fig. 3) is congruent with the MP tree, except for the placement of Aotus, which is sister to the callitrichids (Saguinus/Leontopithecus). Branch lengths are given on the tree. The ML tree is shown in Fig. 4 ()Ln likelihood score 6631.57895). The a estimate for the c correction was 0.87871. The estimated transition/transversion ratio was 2.52237. The ML tree yields a trichotomy at two nodes in the tree, at the most basal node (between Atelidae, Pitheciidae, and Cebidae), and within the Cebidae between the callitrichids, Cebus/Saimiri, and Aotus. In the ML bootstrap analysis, which shows values of taxon groupings that are not recovered in the ML tree, a Pitheciidae/Cebidae clade (with Atelinae as most basal
Fig. 1. Intronic data from G6PD for nine New World monkeys. Sequences are available in GenBank (Accession Nos. AF514756–AF514772). Intron G spans base pairs 1287–1680 and intron H spans base pairs 1681–2110. Data are numbered to follow von Dornum and Ruvolo (1999). Based on the Homo sapiens GenBank sequence (Accession No. X55448), intron G corresponds to the region between exons 7 and 8 (from base pair 15,469 to 15,833) and intron H corresponds to the region between exons 8 and 9 (from base pair 15,928 to 16,337).
M.E. Steiper, M. Ruvolo / Molecular Phylogenetics and Evolution 27 (2003) 121–130
125
Fig. 1. (continued)
platyrrhine) is recovered with a 39.5% value. The two alternative arrangements for the most basal taxon have values of <10%. Within the Cebidae trichotomy, betterresolved alternative relationships had bootstrap values of less than 10%. The consensus tree built from the Bayesian analysis is shown in Fig. 5. The average likelihood score for the trees is )6646.9378 and the a value for the c correction was 0.8815. The analysis yielded Atelidae as the most basal platyrrhine clade with a Pitheciidae/Cebidae clade having a 39% posterior probability, mirroring the ML bootstrap analysis. (The alternative partitions, Atelidae/ Pitheciidae and Cebidae/Atelidae, have 29 and 32% posterior probabilities, respectively.) Within the Cebidae, there is a 40% posterior probability for a callitrichid and Cebus/Saimiri clade. Posterior probabilities for alternative Cebidae arrangements are 36% for a Cebus/ Saimiri and Aotus partition, and 24% for an Aotus and callitrichid partition. Comparing this analysis to that of G6PD DE shows that the G6PD DEGH MP tree is identical to one of the
three most parsimonious trees found in G6PD DE (von Dornum and Ruvolo, 1999), when pruned to the nine taxa analyzed here. The smaller G6PD DE data set (trimmed to the nine taxa sampled before analysis) recovers a tree similar to Fig. 2, except that the most basal relationships are represented by a trichotomy among Pithecia/Callicebus, Ateles/Alouatta, and the Cebidae. Thus, based exclusively MP analysis phylogenetic placements are increasingly resolved with a longer data set, which has been shown to be an important determinant for recovering a tree closest to the true gene tree for that locus (Rosenberg and Kumar, 2001). However, topological discordance exists based on the type of optimality criteria used in the analysis of the G6PD DEGH data set.
4. Discussion Each of the three platyrrhine families forms wellsupported monophyletic clades in every analysis. Based on MP analysis, Cebidae is supported by 13 unambig-
126
M.E. Steiper, M. Ruvolo / Molecular Phylogenetics and Evolution 27 (2003) 121–130
Fig. 1. (continued)
uous synapomorphies and a bootstrap value of 84%, Pitheciidae is supported by 12 unambiguous synapomorphies and a 94% bootstrap value, and Atelidae has a bootstrap value of 95% and is defined by nine unambiguous synapomorphies. Each group is also recovered in the distance analysis, with high bootstrap values in the ML analysis, and with 99–100% posterior probabilities in the Bayesian analysis. The MP analysis reconstructs the most basal divergence among New World monkeys between the Pitheciidae and a Cebidae/Atelidae clade. This initial bifurcation is found in the smaller G6PD data set (von Dornum and Ruvolo, 1999) and the unpublished COII mtDNA data set (von Dornum, 1997). A study of the vWF locus (Chaves et al., 1999), which has a diverse sampling of New World monkeys sufficient to address the questions posed here, also supports a Cebidae/ Atelidae clade. The bootstrap support for Cebidae/ Atelidae is 51% with the G6PD data set. Although only a modest value, bootstrap values over 50% are generally thought to be supportive of a grouping and an under-
estimate of the probability of the clade (Hillis and Bull, 1993). In opposition to the MP results, the Bayesian analysis and to a lesser degree the ML analysis support Atelidae as the most basal clade. This arrangement is in agreement with the 16S (Horovitz and Meyer, 1995) and IRBP (Barroso et al., 1997; Harada et al., 1995; Schneider et al., 1996) data sets. In b2 -microglobulin (Canavez et al., 1999) and e-globin (Harada et al., 1995; Porter et al., 1997a, 1999, 1997b, 1995; Schneider et al., 1993) Cebidae is reconstructed as the most basal taxon. A Cebidae clade, here containing Aotus, Cebus, Saimiri, Leontopithecus, and Saguinus has thus far been proposed with good support by the e-globin (Harada et al., 1995; Porter et al., 1997a, 1999, 1997b, 1995; Schneider et al., 1993), IRBP (Barroso et al., 1997; Harada et al., 1995; Schneider et al., 1996), b2 -microglobulin (Canavez et al., 1999), G6PD DE (von Dornum and Ruvolo, 1999), and vWF (Chaves et al., 1999). [Cebidae is not found to be monophyletic for 16S or 12S rRNA data sets (Horovitz and Meyer, 1995; Horovitz et al., 1998), however.] Within the Cebidae clade, the callitri-
M.E. Steiper, M. Ruvolo / Molecular Phylogenetics and Evolution 27 (2003) 121–130
127
Fig. 1. (continued)
Table 3 Distance matrix of G6PD DEGH intron sequences generated using the HKY85 correction algorithm, implemented with PAUP* v.4b4-10 (Swofford, 2002)
1. 2. 3. 4. 5. 6. 7. 8. 9. 10.
Homo Ateles Aotus Cebus Saimiri Alouatta Pithecia Saguinus Leontopithecus Callicebus
1
2
3
4
5
6
7
8
9
– 0.15452 0.16831 0.17159 0.18146 0.16589 0.15932 0.18074 0.16905 0.16008
– 0.06234 0.06723 0.07632 0.04564 0.06225 0.07428 0.06090 0.05488
– 0.06732 0.07680 0.07010 0.07613 0.07395 0.06058 0.06951
– 0.07212 0.07387 0.08624 0.07251 0.06979 0.07774
– 0.08186 0.08907 0.09175 0.07705 0.08237
– 0.07256 0.08390 0.06638 0.06482
– 0.08577 0.07958 0.05398
– 0.04991 0.08262
– 0.07195
Values are numbers of substitutions per site.
chids (Leontopithecus and Saguinus) form a strongly supported clade with the G6PD DEGH data, as observed in all other analyses of molecular data sets. Cebus and Saimiri also form a strongly supported clade, con-
gruent with many morphological (Ford, 1986; Rosenberger, 1984) and molecular data sets [e-globin (Harada et al., 1995; Porter et al., 1997a, 1999, 1997b, 1995; Schneider et al., 1993), 12S (Horovitz et al., 1998), COII
128
M.E. Steiper, M. Ruvolo / Molecular Phylogenetics and Evolution 27 (2003) 121–130
Fig. 2. Most parsimonious cladogram for the New World monkey G6PD sequences. Bootstrap values are given above branches and numbers of unambiguous sites supporting each grouping are located below branches. Tree was generated with PAUP* v.4b4-10 (Swofford, 2002).
Fig. 4. Maximum likelihood tree for the G6PD DEGH data set. Branch lengths are given on the tree. Bootstrap values are given in the legend, along with shorter branch lengths. Tree was generated with PAUP* v.4b4-10 (Swofford, 2002).
Fig. 5. Bayesian inference of G6PD DEGH phylogeny generated using MrBayes (Huelsenbeck and Ronquist, 2001). Posterior probilities appear above branches (where 40 ¼ 40% or 0.40). Fig. 3. Distance tree of G6PD sequences reconstructed using neighborjoining and the HKY correction algorithm (generated in PAUP* v.4b4-10, Swofford, 2002).
(von Dornum, 1997), IRBP (Barroso et al., 1997; Harada et al., 1995; Schneider et al., 1996), b2 -microglobulin (Canavez et al., 1999), and G6PD DE (von Dornum and Ruvolo, 1999)].
With the additional G6PD data, Aotus forms a moderately supported clade with Cebus and Saimiri in the MP tree, but not in the other analyses. In the distance tree analysis, Aotus is sister to the callitrichids. The Bayesian analysis allies the callitrichids and Cebus/Saimiri. In the ML analysis, a trichotomy is recovered. When the von Dornum and Ruvolo (1999) data set is
M.E. Steiper, M. Ruvolo / Molecular Phylogenetics and Evolution 27 (2003) 121–130
pruned to include only the nine taxa included here, an Aotus, Cebus, and Saimiri clade is found with a bootstrap value of 58% in a MP analysis. An Aotus and Cebus/Saimiri clade has been recovered for b2 -microglobulin (Canavez et al., 1999) (parsimony analysis only), and although the sampling of vWF did not include Saimiri, an Aotus/Cebus clade is well supported (Chaves et al., 1999). With e-globin (Harada et al., 1995; Porter et al., 1997a, 1999, 1997b, 1995; Schneider et al., 1993), Aotus and the callitrichines form a clade to the exclusion of Cebus/Saimiri. Callitrichines and Cebus/ Saimiri form a clade in IRBP (Barroso et al., 1997; Harada et al., 1995; Schneider et al., 1996) in agreement with the Bayesian analysis.
129
moplasious changes that can swamp phylogenetic signal. A second way to overcome this problem is to sequence longer stretches of DNA from less rapidly evolving loci and to generate sequences from additional independent loci (Saitou and Nei, 1986), a strategy which has worked successfully for resolving hominoid phylogenetic relationships (Chen and Li, 2001; Ruvolo, 1997). In the final analysis, it is likely that molecular data sets from many additional independent loci, as well as additional morphological research on fossil and extant taxa, will be necessary to resolve the relationships among the families of New World monkeys.
Acknowledgments 5. Conclusions Considerable topological discord exists at two regions within the platyrrhine phylogenetic tree: the most basal bifurcation and the relationships within the Cebidae. These differences occur both within the current data set and among the different molecular analyses. Given these conflicting results, one interesting question is: Why are family level phylogenetics of the New World monkeys so difficult to resolve? A primary contributing factor may be the nature of the platyrrhine evolutionary radiation itself. An examination of the G6PD distance tree of the New World monkeys is consistent with a scenario of relatively rapid diversification. The average branch length from the most basal platyrrhine node to the basal node of the Cebidae, Atelidae, and Pitheciidae is 0.008. The average branch length from the same basal node to the tips of the tree is 0.041. Therefore, the internodal length is 0.008/0.041 ¼ 20% of the total tree length. In other words, short internodal branches exist deep within the tree. Because the G6PD sequences for all taxa except Ateles are behaving in a clock-like manner judging by TajimaÕs (1993) pairwise relative rate test, it is appropriate to broadly estimate divergence times from these data. Using a 26 million year date to calibrate the most basal platyrrhine node [approximated from Fleagle (1999, p. 431)], each New World monkey family originates roughly five million years later. Such an ancient and rapid diversification hampers phylogenetic reconstruction because when internodes are short relative to the terminal branches, little time passes for synapomorphies to accrue (DeFilippis and Moore, 2000; Lanyon, 1988; Saitou and Nei, 1986). One solution to the problem of resolving phylogenetic bifurcations that occur in rapid succession is to choose a locus that accumulates substitutions rapidly (Saitou and Nei, 1986). However, this is not a viable option for New World monkeys because their radiation is ancient; over long time periods, a rapidly evolving gene can undergo multiple substitutions at the same site, leading to ho-
We thank Charlene Dickinson, George Amato (Bronx Zoo), and the Davis RPRC for samples. Two anonymous reviewers provided comments that greatly improved the manuscript. This work was supported by a NSF grant to M.R. (SBR-9319021) and by a NSF Graduate Research Fellowship and a Mellon Grant for Research to M.S.
References Barroso, C.M.L., Schneider, H., Schneider, M.P.C., Sampaio, I., Harada, M.L., Czelusniak, J., Goodman, M., 1997. Update on the phylogenetic systematics of New World monkeys: further evidence for placing the pygmy marmoset (Cebuella) within the genus Callithrix. Int. J. Primatol. 18, 645–668. Canavez, F.C., Moreira, M.A., Ladasky, J.J., Pissinatti, A., Parham, P., Seuanez, H.N., 1999. Molecular phylogeny of New World Primates (Platyrrhini) based on b2 -microglobulin DNA sequences. Mol. Phylogenet. Evol. 12, 74–82. Chaves, R., Sampaio, I., Schneider, M.P., Schneider, H., Page, S.L., Goodman, M., 1999. The place of Callimico goeldii in the Callitrichine phylogenetic tree: evidence from von Willebrand factor gene intron II sequences. Mol. Phylogenet. Evol. 13, 392– 404. Chen, F.C., Li, W.H., 2001. Genomic divergences between humans and other hominoids and the effective population size of the common ancestor of humans and chimpanzees. Am. J. Hum. Genet. 68, 444–456. DeFilippis, V.R., Moore, W.S., 2000. Resolution of phylogenetic relationships among recently evolved species as a function of amount of DNA sequence: an empirical study based on woodpeckers (Aves: Picidae). Mol. Phylogenet. Evol. 16, 143–160. Fleagle, J.G., 1999. Primate Adaptation and Evolution. Academic Press, San Diego. Ford, S.M., 1986. Systematics of the New World monkeys. In: Swindler, D.R., Erwin, J. (Eds.), Comparative Primate Biology, vol. I: Systematics, Evolution and Anatomy. Alan R. Liss, New York, pp. 73–135. Goodman, M., Porter, C.A., Czelusniak, J., Page, S.L., Schneider, H., Shoshani, J., Gunnell, G., Groves, C.P., 1998. Toward a phylogenetic classification of Primates based on DNA evidence complemented by fossil evidence. Mol. Phylogenet. Evol. 9, 585–598. Groves, C.P., 1989. A Theory of Human and Primate Evolution. Oxford University Press, New York.
130
M.E. Steiper, M. Ruvolo / Molecular Phylogenetics and Evolution 27 (2003) 121–130
Harada, M.L., Schneider, H., Schneider, M.P., Sampaio, I., Czelusniak, J., Goodman, M., 1995. DNA evidence on the phylogenetic systematics of New World monkeys: support for the sistergrouping of Cebus and Saimiri from two unlinked nuclear genes. Mol. Phylogenet. Evol. 4, 331–349. Hasegawa, M., Kishino, H., Yano, T., 1985. Dating of the human–ape splitting by a molecular clock of mitochondrial DNA. J. Mol. Evol. 22, 160–174. Hillis, D.M., Bull, J.J., 1993. An empirical test of bootstrapping as a method for assessing confidence in phylogenetic analysis. Syst. Biol. 42, 182–192. Horovitz, I., Meyer, A., 1995. Systematics of New World monkeys (Platyrrhini, Primates) based on 16S mitochondrial DNA sequences: a comparative analysis of different weighting methods in cladistic analysis. Mol. Phylogenet. Evol. 4, 448–456. Horovitz, I., Zardoya, R., Meyer, A., 1998. Platyrrhine systematics: a simultaneous analysis of molecular and morphological data. Am. J. Phys. Anthropol. 106, 261–281. Huelsenbeck, J.P., Ronquist, F., 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17 (8), 754–755. Kay, R.F., 1990. The phyletic relationships of extant and fossil Pitheciinae (Platyrrhini, Anthropoidea). J. Hum. Evol. 19, 175– 208. Lanyon, S.M., 1988. The stochastic mode of molecular evolution: what consequences for systematic investigations? The Auk 105, 565–573. Porter, C.A., Czelusniak, J., Schneider, H., Schneider, M.P., Sampaio, I., Goodman, M., 1997a. Sequences of the primate e-globin gene: implications for systematics of the marmosets and other New World primates. Gene 205, 59–71. Porter, C.A., Czelusniak, J., Schneider, H., Schneider, M.P., Sampaio, I., Goodman, M., 1999. Sequences from the 50 flanking region of the e-globin gene support the relationship of Callicebus with the pitheciins. Am. J. Primatol. 48, 69–75. Porter, C.A., Page, S.L., Czelusniak, J., Schneider, H., Schneider, M.P.C., Sampaio, I., Goodman, M., 1997b. Phylogeny and evolution of selected primates as determined by sequences of the e-globin locus and 50 flanking regions. Int. J. Primatol. 18, 261–295. Porter, C.A., Sampaio, I., Schneider, H., Schneider, M.P., Czelusniak, J., Goodman, M., 1995. Evidence on primate phylogeny from eglobin gene sequences and flanking regions. J. Mol. Evol. 40, 30– 55.
Rosenberg, M.S., Kumar, S., 2001. Incomplete taxon sampling is not a problem for phylogenetic inference. Proc. Natl. Acad. Sci. USA 98, 10751–10756. Rosenberger, A.L., 1984. Fossil New World monkeys dispute the molecular clock. J. Hum. Evol. 13, 737–742. Ruvolo, M., 1997. Molecular phylogeny of the hominoids: inferences from multiple independent DNA sequence data sets. Mol. Biol. Evol. 14, 248–265. Saitou, N., Nei, M., 1986. The number of nucleotides required to determine the branching order of three species, with special reference to the human–chimpanzee–gorilla divergence. J. Mol. Evol. 24, 189–204. Saitou, N., Nei, M., 1987. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4, 406–425. Schneider, H., Sampaio, I., Harada, M.L., Barroso, C.M., Schneider, M.P., Czelusniak, J., Goodman, M., 1996. Molecular phylogeny of the New World monkeys (Platyrrhini, Primates) based on two unlinked nuclear genes: IRBP intron 1 and e-globin sequences. Am. J. Phys. Anthropol. 100, 153–179. Schneider, H., Schneider, M.P., Sampaio, I., Harada, M.L., Stanhope, M., Czelusniak, J., Goodman, M., 1993. Molecular phylogeny of the New World monkeys (Platyrrhini, Primates). Mol. Phylogenet. Evol. 2, 225–242. Swofford, D.L., 2002. PAUP* beta version. Phylogenetic analysis using parsimony (*and other methods). Sinauer Associated, Sunderland, MA. Tajima, F., 1993. Simple methods for testing the molecular evolutionary clock hypothesis. Genetics 135, 599–607. Thompson, J.D., Higgins, D.G., Gibson, T.J., 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22, 4673–4680. von Dornum, M.J., 1997. DNA sequence data from mitochondrial COII and nuclear G6PD loci and a molecular phylogeny of the New World monkeys (Primates, Platyrrhini). PhD dissertation, Harvard University, Cambridge, MA. von Dornum, M., Ruvolo, M., 1999. Phylogenetic relationships of the New World monkeys (Primates, Platyrrhini) based on nuclear G6PD DNA sequences. Mol. Phylogenet. Evol. 11, 459–476. Wakeley, J., 1996. The excess of transitions among nucleotide substitutions: new methods of estimating transition bias underscore its significance. Trends Ecol. Evol. 11, 158–163.
New World monkey phylogeny based on X-linked G6PD DNA sequences Michael E. Steiper* and Maryellen Ruvolo Department of Anthropology, Harvard University, 11 Divinity Avenue, Cambridge, MA 02138, USA Received 24 May 2002; revised 16 August 2002
Abstract The Platyrrhini, or New World monkeys, are an infraorder of Primates comprised of 16 genera. Molecular phylogenetic analyses have consistently sorted these genera into three groups: the Pitheciidae (e.g., saki and titi monkeys), Atelidae (e.g., spider and howler monkeys), and Cebidae (e.g., night monkeys, squirrel monkeys, and tamarins). No consensus has emerged on the relationships among the three groups or within the Cebidae. Here, 0.8 kb of newly generated intronic DNA sequence data from the X-linked glucose-6-phosphate dehydrogenase (G6PD) locus have been collected from nine New World monkey taxa to examine these relationships. These data are added to 1.3 kb of previously generated G6PD intronic DNA sequence data [Mol. Phylogenet. Evol. 11 (1999) 459]. Using distance and parsimony-based techniques, G6PD sequences provide support for an initial bifurcation between the Pitheciidae and the remaining platyrrhines, linking Atelidae and Cebidae as sister taxa. Bayesian methods provided a conflicting phylogeny with Atelidae as outgroup. Within the Cebidae, a sister relation between Aotus and the Cebus/Saimiri clade is favored by parsimony analysis, but not by other analyses. Potential reasons for the difficulty in resolving family level New World monkey phylogenetics are discussed. Ó 2002 Elsevier Science (USA). All rights reserved. Keywords: New World monkeys; Platyrrhines; Platyrrhini; Glucose-6-phosphate dehydrogenase; G6PD; Phylogenetics; Cebidae; Atelidae; Pitheciidae
1. Introduction The Platyrrhini, or New World monkeys, are an infraorder of Primates comprised of approximately 16 genera. The group is fairly ancient, having originated in the Oligocene, and New World monkeys currently range from southern Mexico to northern Argentina (Fleagle, 1999). The phylogenetics of this group have been addressed by multiple researchers using both morphological (e.g., Ford, 1986; Kay, 1990; Rosenberger, 1984) and molecular data sets, including nuclear sequences from e-globin (Harada et al., 1995; Porter et al., 1997a, 1999, 1997b, 1995; Schneider et al., 1993), interphotoreceptor retinoid-binding protein (IRBP) (Barroso et al., 1997; Harada et al., 1995; Schneider et al., 1996), G6PD (von Dornum and Ruvolo, 1999), b2 -microglobulin
* Corresponding author. Fax: 1-617-496-8041. E-mail address: [email protected] (M.E. Steiper).
(Canavez et al., 1999), von Willebrand factor (vWF) (Chaves et al., 1999), and mitochondrial sequences from 12S and 16S rRNA (Horovitz and Meyer, 1995; Horovitz et al., 1998) and COII (von Dornum, 1997). One main result from these analyses is that New World monkeys can be roughly sorted into six groups, which are congruent with the Ônatural groupsÕ of Groves (1989): (1) callitrichids, including Callimico; (2) capuchins (Cebus) and squirrel monkeys (Saimiri); (3) owl monkeys (Aotus); (4) the Pitheciinae (e.g., saki monkeys); (5) titi monkeys (Callicebus); and (6) the Atelinae (e.g., spider monkeys) with Alouatta (howler monkeys). Based on molecular analyses, these six groups consistently fall into three monophyletic groups, the Cebidae (1, 2, and 3), the Pitheciidae (4 and 5), and the Atelidae (6), which Goodman et al. (1998) rank as families. Two major questions remain regarding platyrrhine phylogeny: First, what are the relationships among these three monophyletic families? Second, what are the species relationships within the Cebidae?
1055-7903/02/$ - see front matter Ó 2002 Elsevier Science (USA). All rights reserved. doi:10.1016/S1055-7903(02)00375-5
122
M.E. Steiper, M. Ruvolo / Molecular Phylogenetics and Evolution 27 (2003) 121–130
Here, 0.8 kb of newly generated intronic DNA sequence data have been collected from the X-linked glucose-6-phosphate dehydrogenase (G6PD) locus bearing on these questions. These data are added to 1.3 kb of intronic DNA sequence data from G6PD previously generated by von Dornum and Ruvolo (1999) (and abbreviated here as G6PD DE). The new data are from two downstream introns, termed G and H. At least one taxon from each of GrovesÕ (1989) six well-supported Ônatural groupsÕ is sampled in this data set (including two callitrichids [Leontopithecus and Saguinus], both Cebus and Saimiri, and both Ateles and Alouatta). In this study, Aotus, Cebus, Saimiri, Leontopithecus, and Saguinus represent Cebidae; Callicebus and Pithecia represent Pitheciidae; and Ateles and Alouatta represent Atelidae. Together, this enhanced G6PD data set (abbreviated as G6PD DEGH) comprises 2.1 kb of DNA sequence data from nine New World monkey species.
2. Materials and methods 2.1. Data collection DNA sequence data were collected from nine New World monkeys. The following DNA sequences were prepared from the same specimens as reported in von Dornum and Ruvolo (1999): Ateles fusciceps (Atf1 in von Dornum and Ruvolo, 1999), Saimiri boliviensis (Smb1), Alouatta caraya (Alc1), Pithecia pithecia (Pip1), Saguinus fuscicollis (Sgf1), and Leontopithecus rosalia (Lnr1). Genomic DNA was prepared as in von Dornum and Ruvolo (1999). The remaining sequences were generated partly from samples in von Dornum and Ruvolo (1999) and partly from samples from other individuals because the original DNA samples were depleted, as follows. The Callicebus sequence is from Callicebus moloch (Ccm1) for introns D, E, and G, while intron H sequence is from Callicebus moloch (Davis RPRC ID# CMO-00005). The Cebus DNA sequence is from Cebus apella (Cbp8) for introns D, E, and G, and intron H is from Cebus capuchinus (Bronx Zoo ID# 881198/200493). The Aotus sequence is from Aotus trivirgatus (Aot2) for introns D and E, while introns G and H are from A. trivirgatus (Bronx Zoo ID# 646-96). Both intron G and intron H were amplified using nested or semi-nested PCR, in which two rounds of PCR are used to obtain the desired DNA fragment. After the first round, a diluted product is used as a template for the second reaction, and the product from the second reaction is used for DNA sequencing. PCR amplification was performed using either the AmpliTaq Gold Kit (Perkin–Elmer, Foster City, CA) or Taq PCR Master Mix Kit (Qiagen, Valencia, CA). Reactions were 25 ll in size and contained 2.5 ll GeneAmp 10 PCR
buffer (final concentration 1, provided in kit), 10 mM of each dNTP, 1.125 U AmpliTaq Gold, and 1–5 ll unquantified template DNA. For both PCR rounds, an initial denaturation step (10 min with AmpliTaq Gold or 4 min with Qiagen Taq; 95 °C) was followed by 40 cycles of denaturation (45 s; 95 °C), annealing (45 s; temperatures given in Table 1), and extension (60 s; 72 °C) using Perkin–Elmer 9600 and 9700 thermocyclers. After cycling, reactions were held for 10 min at 72 °C and stored at 4 °C. After the first round of PCR, an aliquot of this reaction was diluted from [1:5] to [1:100] and used as template DNA for the second round of PCR. DNA fragments obtained at the second round were sequenced. MgCl2 concentrations, annealing temperatures, and primer sets used for each specific PCR are given in Table 1, and primer sequences in Table 2. Fragments were electrophoretically separated on 1–2% agarose gels. Ethidium bromide stained gels were visualized using UV light and a /v 174 RF/HaeIII digested DNA (GibcoBRL, Gaithersburg, MD) size standard. If multiple DNA bands were present, the desired fragment was excised from the gel and purified using a QIAquick Gel Extraction Kit (Qiagen, Valencia, CA). When only the desired band was present, PCR samples were run through a Microcon membrane column (Millipore, Bedford, MA) to remove unincorporated oligonucleotides. After purification, the DNA was again electrophoretically separated to roughly estimate DNA concentration for sequencing.
Table 1 PCR parameters and primer combinations used for G6PD sequencing. ‘‘F’’ indicates forward primer; ‘‘R,’’ reverse Round one Intron G Callicebus and Aotus AotGE-P004 (F); AotGH-M005 (R) Anneal 53 °C; 3.0 mM MgCl2 All other species G7-P92 (F); NWG8-M002 (R) Anneal 56 °C; 2.5 mM MgCl2 Intron H Callicebus AotGE-P004 (F); G9-M39 (R) Anneal 53 °C; 2.5 mM MgCl2 All other species GG-P168 (F); G9-M39 (R) Anneal 52 °C; 2.5 mM MgCl2
Round two
NWG7-P001 (F); NWG8-M002 (R) Anneal 53 °C; 3.0 mM MgCl2 NWG7-P001 (F); NWG8-M002 (R) Anneal 56 °C; 2.5 mM MgCl2
NWG7-001 (F); CMOG9-M005 (R) Anneal 48 °C; 2.5 mM MgCl2 NWG8-P003 (F); G9-M39 (R) Anneal 52 °C; 2.5 mM MgCl2
M.E. Steiper, M. Ruvolo / Molecular Phylogenetics and Evolution 27 (2003) 121–130 Table 2 Primers used in this study to amplify and sequence G6PD introns G and H in New World monkeys AotGE-P005 AotGH-M004 CMOG9-M005 G7-P92 G9-M39 GG-P168 NWG7-P001 NWG8-M002 NWG8-P003
50 -GGTGTTGAGCCAGGGGGTCG-30 50 -ACAGATGAGCCCACGACGGAG-30 50 -CTGCACCTCTGAGATAC-30 50 -AAGGAGCCCTTTGGCACTGAGGGTC-30 50 -CAATGTGGTCCTGGGCCAGTACGTG-30 50 -AGGGCTGGACCCCTACACAG-30 50 -GGGTCGTGGGGGCTATTTCG-30 50 -CACGGACGTCATCAGAGTTG-30 50 -CCCGCCTCCACCAACTCTG-30
2.2. Sequencing DNA was fluorescently labeled for sequencing with the dye-terminator method using the ABI PRISM dye terminator cycle sequencing ready reaction kit (Perkin– Elmer, Foster City, CA). Each 20 ll sequencing reaction contained 8 ll terminator ready reaction mix, 3.2 pmol primer (PCR primers were used for sequencing), and 3–8 ll template. Reactions were brought to 20 ll volume with deionized water. Using a Perkin– Elmer 9600 thermocycler, each reaction was cycled 25 times at 96 °C for 10 s, 50 °C for 5 s, and 60 °C for 4 min. After cycling, reactions were purified using Centri-Sep columns (Princeton Separations) or Edge Gel filtration/Centri-Flex columns (Edge Biosystems/ AGTC, Gaithersburg, MD), then dried, and stored at )20 °C. Dried sequencing reactions were resuspended with 2.5–5 ml 5:1 formamide:25 mM EDTA with blue dextran (50 mg/ml). Samples were denatured and electrophoresed for 14 h through 6% Long Ranger acrylamide gels on an ABI 373 automated DNA sequencer. Data in the form of electropherograms were collected using Data Collection v.1.2.1 (ABI). To obtain each speciesÕ sequence at least two electropherograms, usually derived from both DNA strands, were collected (i.e., DNA was sequenced in both directions). Electropherograms were processed with AutoAssembler v.1.4 (ABI) to generate one consensus sequence for each taxon. 2.3. Alignment DNA sequences from G and H were first aligned by ClustalW (Thompson et al., 1994) and then manually confirmed by eye. G6PD G and H DNA sequences for the nine taxa were added to the alignment of von Dornum and Ruvolo (1999) for G6PD D and E introns without modification. Homo sapiens was used as outgroup (GenBank Accession No. X55448). The final alignment was 2110 bp in length and the data set was then converted to Nexus file format for phylogenetic analyses.
123
2.4. Distance analysis Genetic distances between sequences were calculated for tree building and calculating branch lengths. Distance trees were built using the neighbor joining distance algorithm (Saitou and Nei, 1987) and the correction method of Hasegawa, Kishino, and Yano (HKY) (1985) in PAUP* v.4b4-10 (Swofford, 2002). 2.5. Parsimony analysis The most parsimonious tree reconstructions for the G6PD DEGH data set were found using an exhaustive search in PAUP* v.4b4-10 (Swofford, 2002). A weighting matrix was used in which tranversions count as two steps, and transitions, one. This matrix is based on the empirically observed excess of transitions in observed molecular data (Wakeley, 1996). When coded as informative data, alignment gaps do not yield different tree topologies, even under different weighting schemes. For all trees shown here, gaps are ignored. Bootstrap values were calculated using the same parameters, with 1000 resampling replicates of 10 heuristic search additions each. Unambiguous sites were culled from the PAUP* apomorphy list. 2.6. Maximum likelihood (ML) analysis The most likely tree reconstruction for the G6PD DEGH data set was estimated with a heuristic search consisting of 50 random taxon additions in PAUP* v.4b4-10 (Swofford, 2002). Initially, five ML search replicates were used to estimate the transition to transversion ratio and the c correction (based on six classes). Homo was the chosen outgroup. These estimates were then used in the larger search. Bootstrap values were calculated using the same parameters, with 500 resampling replicates of 10 heuristic search additions each. 2.7. Bayesian analysis A Bayesian analysis of the data set was implemented using MrBayes (Huelsenbeck and Ronquist, 2001). A general time-reversible model of molecular evolution was chosen. A 6-category gamma shape parameter for rate variation was estimated. Base frequencies were empirically derived. Homo was the chosen outgroup. No prior trees were defined; a random initial tree was generated. The analysis was run for 1,000,000 generations, saving a tree every 100 generations. For the Markov Chain Monte Carlo (MCMC) analysis, four chains were used. The initial 10,000 generations or 1000 trees were removed from the analysis as Ôburn-in,Õ the phase of the MCMC analysis where likelihoods are rapidly changing, before finally leveling off. The posterior probabilities for individual clades and alternate taxon arrangements were
124
M.E. Steiper, M. Ruvolo / Molecular Phylogenetics and Evolution 27 (2003) 121–130
generated along with a consensus tree using the sumt command. This analysis was carried out four times, with concordant results.
3. Results The New World monkey G6PD DNA sequence data from introns G and H are shown as an alignment in Fig. 1. Intron G is approximately 393 bp in length. Intron H is approximately 450 bp long, but only 429 bases were used in the analyses because several bases could not be verified in the 30 -most region. The total alignment of all four introns was 2110 bp in length. The pairwise HKY distance matrix for the data set is given in Table 3. One most parsimonious (MP) tree was reconstructed for this alignment (Fig. 2). The tree was 985 steps in length and had a consistency index of 0.88. Bootstrap values are
indicated above tree branches; the numbers of unambiguous sites supporting each branch are shown below. In this tree Pitheciidae is the most basal platyrrhine lineage and Aotus is allied to the Cebus/Saimiri clade. The distance tree (Fig. 3) is congruent with the MP tree, except for the placement of Aotus, which is sister to the callitrichids (Saguinus/Leontopithecus). Branch lengths are given on the tree. The ML tree is shown in Fig. 4 ()Ln likelihood score 6631.57895). The a estimate for the c correction was 0.87871. The estimated transition/transversion ratio was 2.52237. The ML tree yields a trichotomy at two nodes in the tree, at the most basal node (between Atelidae, Pitheciidae, and Cebidae), and within the Cebidae between the callitrichids, Cebus/Saimiri, and Aotus. In the ML bootstrap analysis, which shows values of taxon groupings that are not recovered in the ML tree, a Pitheciidae/Cebidae clade (with Atelinae as most basal
Fig. 1. Intronic data from G6PD for nine New World monkeys. Sequences are available in GenBank (Accession Nos. AF514756–AF514772). Intron G spans base pairs 1287–1680 and intron H spans base pairs 1681–2110. Data are numbered to follow von Dornum and Ruvolo (1999). Based on the Homo sapiens GenBank sequence (Accession No. X55448), intron G corresponds to the region between exons 7 and 8 (from base pair 15,469 to 15,833) and intron H corresponds to the region between exons 8 and 9 (from base pair 15,928 to 16,337).
M.E. Steiper, M. Ruvolo / Molecular Phylogenetics and Evolution 27 (2003) 121–130
125
Fig. 1. (continued)
platyrrhine) is recovered with a 39.5% value. The two alternative arrangements for the most basal taxon have values of <10%. Within the Cebidae trichotomy, betterresolved alternative relationships had bootstrap values of less than 10%. The consensus tree built from the Bayesian analysis is shown in Fig. 5. The average likelihood score for the trees is )6646.9378 and the a value for the c correction was 0.8815. The analysis yielded Atelidae as the most basal platyrrhine clade with a Pitheciidae/Cebidae clade having a 39% posterior probability, mirroring the ML bootstrap analysis. (The alternative partitions, Atelidae/ Pitheciidae and Cebidae/Atelidae, have 29 and 32% posterior probabilities, respectively.) Within the Cebidae, there is a 40% posterior probability for a callitrichid and Cebus/Saimiri clade. Posterior probabilities for alternative Cebidae arrangements are 36% for a Cebus/ Saimiri and Aotus partition, and 24% for an Aotus and callitrichid partition. Comparing this analysis to that of G6PD DE shows that the G6PD DEGH MP tree is identical to one of the
three most parsimonious trees found in G6PD DE (von Dornum and Ruvolo, 1999), when pruned to the nine taxa analyzed here. The smaller G6PD DE data set (trimmed to the nine taxa sampled before analysis) recovers a tree similar to Fig. 2, except that the most basal relationships are represented by a trichotomy among Pithecia/Callicebus, Ateles/Alouatta, and the Cebidae. Thus, based exclusively MP analysis phylogenetic placements are increasingly resolved with a longer data set, which has been shown to be an important determinant for recovering a tree closest to the true gene tree for that locus (Rosenberg and Kumar, 2001). However, topological discordance exists based on the type of optimality criteria used in the analysis of the G6PD DEGH data set.
4. Discussion Each of the three platyrrhine families forms wellsupported monophyletic clades in every analysis. Based on MP analysis, Cebidae is supported by 13 unambig-
126
M.E. Steiper, M. Ruvolo / Molecular Phylogenetics and Evolution 27 (2003) 121–130
Fig. 1. (continued)
uous synapomorphies and a bootstrap value of 84%, Pitheciidae is supported by 12 unambiguous synapomorphies and a 94% bootstrap value, and Atelidae has a bootstrap value of 95% and is defined by nine unambiguous synapomorphies. Each group is also recovered in the distance analysis, with high bootstrap values in the ML analysis, and with 99–100% posterior probabilities in the Bayesian analysis. The MP analysis reconstructs the most basal divergence among New World monkeys between the Pitheciidae and a Cebidae/Atelidae clade. This initial bifurcation is found in the smaller G6PD data set (von Dornum and Ruvolo, 1999) and the unpublished COII mtDNA data set (von Dornum, 1997). A study of the vWF locus (Chaves et al., 1999), which has a diverse sampling of New World monkeys sufficient to address the questions posed here, also supports a Cebidae/ Atelidae clade. The bootstrap support for Cebidae/ Atelidae is 51% with the G6PD data set. Although only a modest value, bootstrap values over 50% are generally thought to be supportive of a grouping and an under-
estimate of the probability of the clade (Hillis and Bull, 1993). In opposition to the MP results, the Bayesian analysis and to a lesser degree the ML analysis support Atelidae as the most basal clade. This arrangement is in agreement with the 16S (Horovitz and Meyer, 1995) and IRBP (Barroso et al., 1997; Harada et al., 1995; Schneider et al., 1996) data sets. In b2 -microglobulin (Canavez et al., 1999) and e-globin (Harada et al., 1995; Porter et al., 1997a, 1999, 1997b, 1995; Schneider et al., 1993) Cebidae is reconstructed as the most basal taxon. A Cebidae clade, here containing Aotus, Cebus, Saimiri, Leontopithecus, and Saguinus has thus far been proposed with good support by the e-globin (Harada et al., 1995; Porter et al., 1997a, 1999, 1997b, 1995; Schneider et al., 1993), IRBP (Barroso et al., 1997; Harada et al., 1995; Schneider et al., 1996), b2 -microglobulin (Canavez et al., 1999), G6PD DE (von Dornum and Ruvolo, 1999), and vWF (Chaves et al., 1999). [Cebidae is not found to be monophyletic for 16S or 12S rRNA data sets (Horovitz and Meyer, 1995; Horovitz et al., 1998), however.] Within the Cebidae clade, the callitri-
M.E. Steiper, M. Ruvolo / Molecular Phylogenetics and Evolution 27 (2003) 121–130
127
Fig. 1. (continued)
Table 3 Distance matrix of G6PD DEGH intron sequences generated using the HKY85 correction algorithm, implemented with PAUP* v.4b4-10 (Swofford, 2002)
1. 2. 3. 4. 5. 6. 7. 8. 9. 10.
Homo Ateles Aotus Cebus Saimiri Alouatta Pithecia Saguinus Leontopithecus Callicebus
1
2
3
4
5
6
7
8
9
– 0.15452 0.16831 0.17159 0.18146 0.16589 0.15932 0.18074 0.16905 0.16008
– 0.06234 0.06723 0.07632 0.04564 0.06225 0.07428 0.06090 0.05488
– 0.06732 0.07680 0.07010 0.07613 0.07395 0.06058 0.06951
– 0.07212 0.07387 0.08624 0.07251 0.06979 0.07774
– 0.08186 0.08907 0.09175 0.07705 0.08237
– 0.07256 0.08390 0.06638 0.06482
– 0.08577 0.07958 0.05398
– 0.04991 0.08262
– 0.07195
Values are numbers of substitutions per site.
chids (Leontopithecus and Saguinus) form a strongly supported clade with the G6PD DEGH data, as observed in all other analyses of molecular data sets. Cebus and Saimiri also form a strongly supported clade, con-
gruent with many morphological (Ford, 1986; Rosenberger, 1984) and molecular data sets [e-globin (Harada et al., 1995; Porter et al., 1997a, 1999, 1997b, 1995; Schneider et al., 1993), 12S (Horovitz et al., 1998), COII
128
M.E. Steiper, M. Ruvolo / Molecular Phylogenetics and Evolution 27 (2003) 121–130
Fig. 2. Most parsimonious cladogram for the New World monkey G6PD sequences. Bootstrap values are given above branches and numbers of unambiguous sites supporting each grouping are located below branches. Tree was generated with PAUP* v.4b4-10 (Swofford, 2002).
Fig. 4. Maximum likelihood tree for the G6PD DEGH data set. Branch lengths are given on the tree. Bootstrap values are given in the legend, along with shorter branch lengths. Tree was generated with PAUP* v.4b4-10 (Swofford, 2002).
Fig. 5. Bayesian inference of G6PD DEGH phylogeny generated using MrBayes (Huelsenbeck and Ronquist, 2001). Posterior probilities appear above branches (where 40 ¼ 40% or 0.40). Fig. 3. Distance tree of G6PD sequences reconstructed using neighborjoining and the HKY correction algorithm (generated in PAUP* v.4b4-10, Swofford, 2002).
(von Dornum, 1997), IRBP (Barroso et al., 1997; Harada et al., 1995; Schneider et al., 1996), b2 -microglobulin (Canavez et al., 1999), and G6PD DE (von Dornum and Ruvolo, 1999)].
With the additional G6PD data, Aotus forms a moderately supported clade with Cebus and Saimiri in the MP tree, but not in the other analyses. In the distance tree analysis, Aotus is sister to the callitrichids. The Bayesian analysis allies the callitrichids and Cebus/Saimiri. In the ML analysis, a trichotomy is recovered. When the von Dornum and Ruvolo (1999) data set is
M.E. Steiper, M. Ruvolo / Molecular Phylogenetics and Evolution 27 (2003) 121–130
pruned to include only the nine taxa included here, an Aotus, Cebus, and Saimiri clade is found with a bootstrap value of 58% in a MP analysis. An Aotus and Cebus/Saimiri clade has been recovered for b2 -microglobulin (Canavez et al., 1999) (parsimony analysis only), and although the sampling of vWF did not include Saimiri, an Aotus/Cebus clade is well supported (Chaves et al., 1999). With e-globin (Harada et al., 1995; Porter et al., 1997a, 1999, 1997b, 1995; Schneider et al., 1993), Aotus and the callitrichines form a clade to the exclusion of Cebus/Saimiri. Callitrichines and Cebus/ Saimiri form a clade in IRBP (Barroso et al., 1997; Harada et al., 1995; Schneider et al., 1996) in agreement with the Bayesian analysis.
129
moplasious changes that can swamp phylogenetic signal. A second way to overcome this problem is to sequence longer stretches of DNA from less rapidly evolving loci and to generate sequences from additional independent loci (Saitou and Nei, 1986), a strategy which has worked successfully for resolving hominoid phylogenetic relationships (Chen and Li, 2001; Ruvolo, 1997). In the final analysis, it is likely that molecular data sets from many additional independent loci, as well as additional morphological research on fossil and extant taxa, will be necessary to resolve the relationships among the families of New World monkeys.
Acknowledgments 5. Conclusions Considerable topological discord exists at two regions within the platyrrhine phylogenetic tree: the most basal bifurcation and the relationships within the Cebidae. These differences occur both within the current data set and among the different molecular analyses. Given these conflicting results, one interesting question is: Why are family level phylogenetics of the New World monkeys so difficult to resolve? A primary contributing factor may be the nature of the platyrrhine evolutionary radiation itself. An examination of the G6PD distance tree of the New World monkeys is consistent with a scenario of relatively rapid diversification. The average branch length from the most basal platyrrhine node to the basal node of the Cebidae, Atelidae, and Pitheciidae is 0.008. The average branch length from the same basal node to the tips of the tree is 0.041. Therefore, the internodal length is 0.008/0.041 ¼ 20% of the total tree length. In other words, short internodal branches exist deep within the tree. Because the G6PD sequences for all taxa except Ateles are behaving in a clock-like manner judging by TajimaÕs (1993) pairwise relative rate test, it is appropriate to broadly estimate divergence times from these data. Using a 26 million year date to calibrate the most basal platyrrhine node [approximated from Fleagle (1999, p. 431)], each New World monkey family originates roughly five million years later. Such an ancient and rapid diversification hampers phylogenetic reconstruction because when internodes are short relative to the terminal branches, little time passes for synapomorphies to accrue (DeFilippis and Moore, 2000; Lanyon, 1988; Saitou and Nei, 1986). One solution to the problem of resolving phylogenetic bifurcations that occur in rapid succession is to choose a locus that accumulates substitutions rapidly (Saitou and Nei, 1986). However, this is not a viable option for New World monkeys because their radiation is ancient; over long time periods, a rapidly evolving gene can undergo multiple substitutions at the same site, leading to ho-
We thank Charlene Dickinson, George Amato (Bronx Zoo), and the Davis RPRC for samples. Two anonymous reviewers provided comments that greatly improved the manuscript. This work was supported by a NSF grant to M.R. (SBR-9319021) and by a NSF Graduate Research Fellowship and a Mellon Grant for Research to M.S.
References Barroso, C.M.L., Schneider, H., Schneider, M.P.C., Sampaio, I., Harada, M.L., Czelusniak, J., Goodman, M., 1997. Update on the phylogenetic systematics of New World monkeys: further evidence for placing the pygmy marmoset (Cebuella) within the genus Callithrix. Int. J. Primatol. 18, 645–668. Canavez, F.C., Moreira, M.A., Ladasky, J.J., Pissinatti, A., Parham, P., Seuanez, H.N., 1999. Molecular phylogeny of New World Primates (Platyrrhini) based on b2 -microglobulin DNA sequences. Mol. Phylogenet. Evol. 12, 74–82. Chaves, R., Sampaio, I., Schneider, M.P., Schneider, H., Page, S.L., Goodman, M., 1999. The place of Callimico goeldii in the Callitrichine phylogenetic tree: evidence from von Willebrand factor gene intron II sequences. Mol. Phylogenet. Evol. 13, 392– 404. Chen, F.C., Li, W.H., 2001. Genomic divergences between humans and other hominoids and the effective population size of the common ancestor of humans and chimpanzees. Am. J. Hum. Genet. 68, 444–456. DeFilippis, V.R., Moore, W.S., 2000. Resolution of phylogenetic relationships among recently evolved species as a function of amount of DNA sequence: an empirical study based on woodpeckers (Aves: Picidae). Mol. Phylogenet. Evol. 16, 143–160. Fleagle, J.G., 1999. Primate Adaptation and Evolution. Academic Press, San Diego. Ford, S.M., 1986. Systematics of the New World monkeys. In: Swindler, D.R., Erwin, J. (Eds.), Comparative Primate Biology, vol. I: Systematics, Evolution and Anatomy. Alan R. Liss, New York, pp. 73–135. Goodman, M., Porter, C.A., Czelusniak, J., Page, S.L., Schneider, H., Shoshani, J., Gunnell, G., Groves, C.P., 1998. Toward a phylogenetic classification of Primates based on DNA evidence complemented by fossil evidence. Mol. Phylogenet. Evol. 9, 585–598. Groves, C.P., 1989. A Theory of Human and Primate Evolution. Oxford University Press, New York.
130
M.E. Steiper, M. Ruvolo / Molecular Phylogenetics and Evolution 27 (2003) 121–130
Harada, M.L., Schneider, H., Schneider, M.P., Sampaio, I., Czelusniak, J., Goodman, M., 1995. DNA evidence on the phylogenetic systematics of New World monkeys: support for the sistergrouping of Cebus and Saimiri from two unlinked nuclear genes. Mol. Phylogenet. Evol. 4, 331–349. Hasegawa, M., Kishino, H., Yano, T., 1985. Dating of the human–ape splitting by a molecular clock of mitochondrial DNA. J. Mol. Evol. 22, 160–174. Hillis, D.M., Bull, J.J., 1993. An empirical test of bootstrapping as a method for assessing confidence in phylogenetic analysis. Syst. Biol. 42, 182–192. Horovitz, I., Meyer, A., 1995. Systematics of New World monkeys (Platyrrhini, Primates) based on 16S mitochondrial DNA sequences: a comparative analysis of different weighting methods in cladistic analysis. Mol. Phylogenet. Evol. 4, 448–456. Horovitz, I., Zardoya, R., Meyer, A., 1998. Platyrrhine systematics: a simultaneous analysis of molecular and morphological data. Am. J. Phys. Anthropol. 106, 261–281. Huelsenbeck, J.P., Ronquist, F., 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17 (8), 754–755. Kay, R.F., 1990. The phyletic relationships of extant and fossil Pitheciinae (Platyrrhini, Anthropoidea). J. Hum. Evol. 19, 175– 208. Lanyon, S.M., 1988. The stochastic mode of molecular evolution: what consequences for systematic investigations? The Auk 105, 565–573. Porter, C.A., Czelusniak, J., Schneider, H., Schneider, M.P., Sampaio, I., Goodman, M., 1997a. Sequences of the primate e-globin gene: implications for systematics of the marmosets and other New World primates. Gene 205, 59–71. Porter, C.A., Czelusniak, J., Schneider, H., Schneider, M.P., Sampaio, I., Goodman, M., 1999. Sequences from the 50 flanking region of the e-globin gene support the relationship of Callicebus with the pitheciins. Am. J. Primatol. 48, 69–75. Porter, C.A., Page, S.L., Czelusniak, J., Schneider, H., Schneider, M.P.C., Sampaio, I., Goodman, M., 1997b. Phylogeny and evolution of selected primates as determined by sequences of the e-globin locus and 50 flanking regions. Int. J. Primatol. 18, 261–295. Porter, C.A., Sampaio, I., Schneider, H., Schneider, M.P., Czelusniak, J., Goodman, M., 1995. Evidence on primate phylogeny from eglobin gene sequences and flanking regions. J. Mol. Evol. 40, 30– 55.
Rosenberg, M.S., Kumar, S., 2001. Incomplete taxon sampling is not a problem for phylogenetic inference. Proc. Natl. Acad. Sci. USA 98, 10751–10756. Rosenberger, A.L., 1984. Fossil New World monkeys dispute the molecular clock. J. Hum. Evol. 13, 737–742. Ruvolo, M., 1997. Molecular phylogeny of the hominoids: inferences from multiple independent DNA sequence data sets. Mol. Biol. Evol. 14, 248–265. Saitou, N., Nei, M., 1986. The number of nucleotides required to determine the branching order of three species, with special reference to the human–chimpanzee–gorilla divergence. J. Mol. Evol. 24, 189–204. Saitou, N., Nei, M., 1987. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4, 406–425. Schneider, H., Sampaio, I., Harada, M.L., Barroso, C.M., Schneider, M.P., Czelusniak, J., Goodman, M., 1996. Molecular phylogeny of the New World monkeys (Platyrrhini, Primates) based on two unlinked nuclear genes: IRBP intron 1 and e-globin sequences. Am. J. Phys. Anthropol. 100, 153–179. Schneider, H., Schneider, M.P., Sampaio, I., Harada, M.L., Stanhope, M., Czelusniak, J., Goodman, M., 1993. Molecular phylogeny of the New World monkeys (Platyrrhini, Primates). Mol. Phylogenet. Evol. 2, 225–242. Swofford, D.L., 2002. PAUP* beta version. Phylogenetic analysis using parsimony (*and other methods). Sinauer Associated, Sunderland, MA. Tajima, F., 1993. Simple methods for testing the molecular evolutionary clock hypothesis. Genetics 135, 599–607. Thompson, J.D., Higgins, D.G., Gibson, T.J., 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22, 4673–4680. von Dornum, M.J., 1997. DNA sequence data from mitochondrial COII and nuclear G6PD loci and a molecular phylogeny of the New World monkeys (Primates, Platyrrhini). PhD dissertation, Harvard University, Cambridge, MA. von Dornum, M., Ruvolo, M., 1999. Phylogenetic relationships of the New World monkeys (Primates, Platyrrhini) based on nuclear G6PD DNA sequences. Mol. Phylogenet. Evol. 11, 459–476. Wakeley, J., 1996. The excess of transitions among nucleotide substitutions: new methods of estimating transition bias underscore its significance. Trends Ecol. Evol. 11, 158–163.