Phylogeny of the Bovidae ( Artiodactyla, Mammalia ) , Based on

Phylogeny of the Bovidae ( Artiodactyla, Mammalia ) , Based on Mitochondrial Ribosomal DNA Sequences’

John Gatesy, * Deborah Yelon,j- Rob DeS’alle,~~g and Elisabeth S. Vrba* p-f *Department of Geology and Geophysics, TDepartment of Biology, Yale University; and SDepartment of Entomology, American Museum of Natural History

Portions of the 12s and 16s mitochondrial ribosomal genes for 16 species repre- senting nine tribes in the mammal family Bovidae were compared with six pre- viously published orthologous sequences. Phylogenetic analysis of variable nucleotide positions under different constraints and weighting schemes revealed no robust groupings among tribes. Consensus trees support previous hypotheses of monophyly for four clades, including the traditional subfamily Bovinae. However, the basal diversification of bovid tribes, which was largely unresolved by morphological, immunodiffusion, allozyme, and protein sequence data, remains unresolved with the addition of DNA sequence data. The intractability of this systematic problem is consistent with a rapid radiation of the major bovid groups. Several analyses of our data show that monophyly of the Bovidae, which was weakly supported by previous morphological and molecular work, is questionable.

Introduction

The mammal family Bovidae comprises all artiodactyl ungulates possessing nondeciduous horn cores and sheaths and includes antelopes, cattle, sheep, and goats. Since its first appearance in the Miocene -20 Mya (Savage and Russell 1983; Vrba 1985), the Bovidae underwent a rapid radiation that is unparalleled in other mammals of large body size. This diversification has resulted in > 120 extant species (Honacki et al. 1982, pp. 326-343) and -300 extinct species (Savage and Russell 1983, pp. 2 1 l-40 1) distributed .over Africa, Eurasia, and North America. Because of the excep- tionally good fossil representation, some extant bovid species may be traced back in the geological record for millions of years (Gentry 1978). The rare combination of large past and present diversities marks the bovids as prime candidates for macro- evolutionary research in which patterns of speciation and extinction of taxa in time are of interest (Vrba 1984, 1987).

Systematic work on the bovids has been difficult, leading one worker to comment that the Bovidae is one of the most troublesome groups of mammals to classify (Simp- son 1945 ). Previous work employing morphology (Simpson 1945 ), immunodiffusion (Lowenstein 1986b), protein sequences (Beintema et al. 1986 ), and allozymes (Geor- giadis et al. 199 1) has offered limited insight into the basal subdivisions of the Bovidae.

In the present study, we have sequenced segments of mitochondrial ribosomal DNA (mt-rDNA) to examine bovid relationships at the intertribal level of the traditional systematic framework (Simpson 1945). Previous work using similar stretches

1. Key words: Bovidae, evolutionary radiation, mitochondrial ribosomal DNA, secondary structure.

Address for correspondence and reprints: John Gatesy, % Rob DeSalle, Department of Entomology, American Museum of Natural History, 79th Street and Central Park West, New York, New York, 10024.

Mol. Biol. Evol. 9(3):433-446. 1992. 0 1992 by The University of Chicago. All rights reserved. 0737-4038/92/0903-0005$02.00

433

Dow

nloaded from https://academ

ic.oup.com/m

be/article/9/3/433/1037260 by guest on 23 Decem

ber 2021

434 Gatesy et al.

of mt-rDNA has elucidated artiodactyl phylogeny at both higher and lower levels (Miyamoto et al. 1989; Kraus and Miyamoto 199 1). Therefore mt-rDNA genes should be evolving at an appropriate rate for intermediate levels of divergence. A comparison between nine bovid tribes and other pecoran artiodactyls allows bovid monophyly to be assessed on the basis of mt-rDNA sequences.

Material and Methods Tissues and Sequences

Sequences and tissue were collected from various sources (table 1). Represen- tatives of (a) nine tribes within the Bovidae, (b) three closely related pecoran artiodactyl

Table 1 Artiodactyl Taxa Examined in Present Study

SOURCE OF TISSUE OR FAMILY BOVID TRIBE SPECIES COMMON NAME SEQUENCE

Bovidae Caprini Capra hircus Domestic goat Kraus and Miyamoto 199 1 Hemitragus Himalayan tahr New York Zoological

jemlahicus Society Cephalophini Cephalophus Maxwell’s duiker New York Zoological

maxwelli Society Reduncini Kobus kob Kob San Diego Zoo

K. ellipsiprymnus Waterbuck San Diego Zoo Antilopini Gazella thomsoni Thomson’s gazelle New York Zoological

Society Alcelaphini Damaliscus dorcas Blesbok New York Zoological

Society D. hunteri Hunter’s hartebeest San Diego Zoo

Hippotragini Oryx leucoryx Arabian oryx San Diego Zoo 0. dammah Scimitar-homed San Diego Zoo

oryx 0. gazella Fringe-eared oryx San Diego Zoo Addax Addax San Diego Zoo

nasomaculatus Hippotragus Roan San Diego Zoo

equinus H. niger Sable San Diego Zoo

Bovini Bos taurus Domestic cow Anderson et al. 1982 Tragelaphini Tragelaphus Nyala New York Zoological

angasi Society Boselaphini Boselaphus Nilgai San Diego Zoo

tragocamelus Antilocapridae

Cervidae

Giraffidae

Tragulidae

Antilocapra Pronghom Mathew Cronin, Yale americana University (tissue)

Hydropotes Chinese water deer Miyamoto et al. 1990 inermis

Muntiacus reevesi Muntjac Miyamoto et al. 1990 Giraffa Giraffe S. Tanhauser (unpublished

camelopardalis sequence) Tragulus napu Mouse deer Kraus and Miyamoto 199 I

NOTE.-Classification is according to Simpson (1945).

Dow


ic.oup.com/m


ber 2021

Bovid mt-rDNA Phylogeny 435

families-Cervidae (antlered deer), Giraffidae (giraffes), and Antilocapridae (pronghorn antelope)-and (c) a more distantly related tragulid (mouse deer) were included.

DNA Treatment

Total genomic DNA was extracted from brain, kidney, and liver tissues according to the protocol of Caccone et al. ( 1987). Conserved primer pairs L109 1 + H 1478 (Kocher et al. 1989) and 16sar + 16sbr (S. R. Palumbi, personal communication), in 1: 50 proportions, were used to amplify single-stranded segments of the 12s and 16s mtrDNA genes by the polymerase chain reaction (PCR). These fragments were concentrated and desalted in Centricon 30 columns. Purified single strands were then sequenced with the limiting primer of the unbalanced PCR amplification by the dideoxy method (Gyllensten and Erlich 1988 ) . Radiolabeled sequencing products were resolved by vertical acrylamide gel electrophoresis and were visualized by autoradiography. Approximately 250 bases of the 12s gene and 350 bases of the 16s gene were sequenced for each sample. Results were checked by sequencing each fragment at least two times. Both strands were sequenced in the majority of samples. The remainder were sequenced multiple times in the same direction from different PCR amplifications. The orthologous DNA sequences, having a limited amount of length variability, were aligned by eye (fig, 1).

Phylogenetic Analysis

The variable nucleotide positions and indels were analyzed using Swofford’s ( 1990) PAUP 3.0. All nucleotide-substitution characters were specified as unordered. Indels were coded as single characters, regardless of length. Where indels of different length overlap in the same region (i.e., at positions B and C in fig. 1)) each indel size was assigned a character state. Transformations between any two indel character states were accorded a cost of 1. Because the number of taxa in our sample precluded the use of algorithms that are guaranteed to find all minimum-length cladograms, we heuristically searched for most-parsimonious trees. Unless otherwise stated, the fol- lowing analyses employ nearest-neighbor-interchange branch swapping of 100 ran- domly constructed initial trees (Swofford 1990).

Bootstrap analyses consisted of 100 heuristic replications. Each component heuristic search employed both initial simple addition of taxa and nearest-neighbor-interchange branch swapping (Swofford 1990). Up to 100 trees were held for each bootstrap replication.

The Tragulidae are universally designated as the extant sister group of the pecoran artiodactyls, according to both morphological (Gentry and Hooker 1988) and molecular data (Irwin et al. 199 1) . Therefore Trugulus nupu was used to root our trees.

Our data were subdivided and constrained in various combinations. In addition, a weighting scheme that corrects for stem-position covariance was employed. These procedures are described in the Results section.

Results Initial Search

When applied to nucleotide sequence data, the original parsimony method of Fitch ( 197 1) assumes all character transformations to be equally probable. In fact, huge differences in the occurrence of some classes of DNA transformation have been recognized (Brown et al. 1982 ) . Nevertheless, the result of an equally weighted search for hierarchical structure in a data set may be used as the basis for comparison with other analyses that assume certain evolutionary regularities.

Dow


ic.oup.com/m


ber 2021

Dow


ic.oup.com/m


ber 2021


We first examined all characters, weighting them equally with no phylogenetic constraints. The strict-consensus tree of the 12 shortest trees found is shown in figure 2A. The Bovidae do not fall out as a unified group in this analysis, as the cervids and Antilocupru cluster within the bovids. One hundred bootstrap replications show that few bovid intertribal groupings are robustly supported. However, genera are confidently placed within proper tribes.

Partitioning the Data

Previous theoretical (Pamilio and Nei 1988) and empirical work (Novacek et al. 1988; Smith 1989; Kraus and Miyamoto 199 1) has shown that different genes, structural areas of genes, and classes of nucleotide transformations produce conflicting resolutions of phylogenetic problems. Therefore our data were partitioned into classes that a priori may be expected to behave differently because of unique evolutionary properties.

C.h.(Cap) H.j.(Cap)

C.m(Ceph) K.k.(Red) K.e.(Red) G.t.(Ant) D.d.(Alc) D.h.(Alc) O.l.(Hip) O.d.(Hip) O.g.(Hip) A.n.(Hip) H.e.(Hip) H.n.(Hip)

I I G.c." I T.n.*

A

3 99

I 99

96

B Ftc .2.-Strict-consensus trees of shortest trees found by using all of the equally weighted data. A, No

constraints. B, Bovid monophyfy enforced. Nodes supported by 250% of bootstrap replicates are indicated. Branch lengths are not proportional to the number of character changes. C.h. = Capra hircus; H.j. = Hem- itragus jemlahicus; C.m. = Cephalophus maxwelli; Kk. = Kobus kob; Le. = Kobus ellipsiprymnus; G.t. = Gazella thomsoni; D.d. = Damaliscus dorcas; D.h. = Damaliscus hunteri; 0.1. = Oryx leucoryx; 0.d. = 0. dammah; 0.g. = 0. gazella; A.n. = Addax nasomaculatus; H.e. = Hippotragus equinus; H.n. = Hip- potragus niger; B.t. = Bos taurus; T.a. = Tragelaphus angasi; Bo.t. = Boselaphus tragocamelus; A.a. = An- tilocapra americana; H.i. = Hydropotes inermis; Mr. = Muntiacus reevesi; G.c. = Giraffa camelopardalis; and T.n. = Tragulus napu. Bovid species’ tribal affinities, according to the classification of Simpson ( 1945), are shown in parentheses: Cap = Caprini; Ceph = Cephalophini; Red = Reduncini; Ant = Antilopini; Ale = Alcelaphini; Hip = Hippotragini; Bov = Bovini; Trag = Tragelaphini; and Bos = Boselaphini. Nonbovids are marked by asterisks.

Dow


ic.oup.com/m


ber 2021

438 Gatesy et al.

Different mitochondrial DNA genes evolve at different rates (Anderson et al. 1982). Our comparisons indicate that 34% of the sites in the 16s gene are varied, while only 26% of the sites in the 12s gene are varied in the taxa sampled here. Separate phylogenetic analyses of these two genes show a higher level of character support for the 16s gene, according to the retention index of Farris ( 1989; also see table 2).

Stem regions and loop regions of mt-rDNA are subject to different selective and structural constraints. It is not surprising that independent analyses of stem characters and loop characters from the same ribosomal DNA (rDNA) molecule have produced broadly contrasting phylogenetic results (Wheeler and Honeycutt 1988). Stem regions and loop regions of our artiodactyl sequences were assigned according to the secondary- structure models of cow 12s mt-rRNA (Gutel et al. 1985) and cow 16s mt-rRNA (Gutel and Fox 1988 ). Loop regions were found to be much more varied than the stem regions, with only 20% of sites within stems being varied and with 40% being

Table 2 Summary of Parsimony Searches

CHARACTERS

AND WEIGHTING CONSTRAINT

STRICT

No. OF No. OF CONGRUENCES

STEPS/NO. EXTRA RETENTION

OF TREES STEAL INDEX~ S L B D P

All:

Equal _. Equal _. Equal _. Equal _. Equal Equal _. Equal

Stems l/2 Stems l/2 Stems 213 Stems 2/3

Transversions: Equal Equal _.

12s:

None 499112 Bovids monophyletic 50519 Beintema 38612 Georgiadis distance 507112 Georgiadis parsimony 49819 Simpson 51013 Lowenstein 51416 None 45619 Bovids monophyletic 46116 None 470.3313 Bovids monophyletic 475.6716

0 0.507 6 0.495 2 0.417

29 0.450 20 0.469

5 0.485 9 0.478 0 . . . 5 . . . 0 . . . 5.33 . . .

None 97184 0 0.610 Bovids monophyletic 97112 0 0.610

Equal Equal _.

16s:


Equal _. Equal

Stems:


Equal Equal _.

Loops:


Equal _. None 36916 0 0.484 Equal Bovids monophyletic 373115 4 0.473

0 0 1 0 0 1 2 2 0 0 1 1 3 0 0 00061 0 0 0 1 2 2 2 2 1 0 1 4 1 0 0 00110 1 2 2 0 0 00110 1 2 2 0 0

1 2 1 0 0 1 2 1 0 0

00100 0 0 0 0 0

0 0 0 0 0 00100

00110 1 1 2 0 0

00100 1 2 2 0 0

’ Source: Fanis (1989). b Data are number of monophyletic intertribal groups shared between strict consensus of minimum-length trees and

previous phylogenetic hypotheses. S = Simpson (I 945); L = Lowenstein (19866); B = Beintema et al. (1986); D = Georgiadis et al. ( 199 I) distance analysis; and P = Georgiadis et al. (I 99 I) parsimony analysis.

Dow


ic.oup.com/m


ber 2021


varied within loop regions. When the two groups of data are analyzed separately, the retention index is higher for the stem data (table 2).

Finally, transversions-purine-to-pyrimidine change or pyrimidine-to-pm-me change-were analyzed. In vertebrate mt-rDNA (Hixson and Brown 1986) as well as in nuclear rDNA (Larson and Wilson 1989; Hedges et al. 1990), transversions tend to occur much less frequently than transitions. Only 37% of the varied nucleotide positions in our data set were characterized by transversions. Transversion parsimony analysis (Lake 1987) of these positions produced 84 equally parsimonious trees re- sulting in the strict-consensus tree of figure 3A. The retention index is higher for the transversion data set than for the overall data set (table 2).

Constraints

The total data set and all subgroups of data were also analyzed under the constraint of bovid monophyly. The complete data set required six additional steps to account for bovid monophyly (fig. 2B). This is approximately a 1% increase in tree length. Similarly, extra evolutionary steps were necessary to retain a monophyletic Bovidae in the partitioned data (table 2). The sole exception was the analysis based on transversions, in which bovid monophyly was found 12 times among the 84 shortest trees that resulted when no constraints were applied. The strict consensus of these 12 shortest trees is presented in figure 3B.

To assess the compatibility of our data with previous phylogenetic hypotheses for the Bovidae, we enforced topological constraints that preserved monophyletic groupings of these past studies. Under these phylogenetic constraints, parsimony anal-

C.h.(Cap)

‘I H.j.(Cap) C.m.(Ceph)

I K.k.(Red) K.e.(Red) G.t.(Ant)

I 1 D.d.(Alc) D.h.(Alc) O.l.(Hip)

I O.d.(Hip) O.g.(Hip) A.n.(Hip) a-J - H.e.(Hip) H.n.(Hip) B.t.(Bov)’

T.a.(Trag) t Bo.t.(Bos)

I

A.a.*

A

I H.i.* M.r.* G.c.” T.n.*

FIG. 3.-Strict-consensus trees of shortest trees found by using only transversions. A, No constraints. B, Bovid monophyly enforced. Species’ abbreviations, tribal affinities, and nonbovid denotation are as in fig. 2. Branch lengths are not proportional to the number of character changes.

Dow


ic.oup.com/m


ber 2021

440 Gatesy et al.

ysis of our data for relevant taxa produced trees that were consistently longer than our most-parsimonious results (table 2). The allozyme-distance tree of Georgiadis et al. ( 199 1) was least accountable by our data, costing 29 additional steps, or 6% of total tree length. The allozyme-parsimony tree of Georgiadis et al. ( 199 1) was also at odds with our data, requiring an additional 20 steps (table 2).

Weighting of Characters

Partitioning data into subgroups and then excluding certain subgroups is the most extreme form of character weighting. More-subtle philosophies have been suggested in the past (Neff 1986; Lake 1987; Williams and Fitch 1989; Wheeler 1990). We attempted to account for structural and selective constraints of nucleotide transformations in stem regions by weighting our data.

Wheeler and Honeycutt ( 1988) noted that rDNA stem nucleotide positions covary because natural selection tends to maintain secondary Watson-Crick base-pairing structure in the functional rRNA. Once an initial change in a stem position occurs, the pairing nucleotide will be selected to change as well. In a reanalysis of arthropod nuclear rDNA sequences, Wheeler and Honeycutt ( 1988) found a statistically signif- icant number of instances in which changes in one stem position were compensated by changes in the complementary stem position. The authors suggested that, because of this character covariance, nucleotide substitutions in stems could not be interpreted as independent evolutionary events and should be either down-weighted by one half or even disregarded altogether.

Compensatory changes in stem regions can be inferred from our mt-rDNA sequences (e.g., sites E and F in fig. 1) . Sites where compensation differences are observed should be given less weight, but the amount of weight reduction that is appropriate is not obvious. It is not clear either how often compensation takes place or how quickly it occurs. Figure 4 illustrates the evolution of two complementary stem nucleotide positions mapped onto one of the shortest trees for the total data set with bovid monophyly enforced. In this instance, a “doubling up” of convergence is evident on the lines leading to Gazella thomsoni and to Oryx gazella+Addax. Compensation is instantaneous at the level of resolution provided by the data, and homoplasy is com- pounded. Alternatively, the change in position D leading to Damaliscus hunteri is not compensated. This lag in compensation is ~2.5 Myr, the minimum estimate of the divergence time for D. hunteri and D. dorcas (Vrba 1979, 1985). A longer lag can be seen on the line leading to Bos taurus. Here, compensation is delayed 3 11.5 Myr, the minimum estimate for the common node of B. taurus and Tragelaphus angasi (Hill et al. 1985; E. S. Vrba, unpublished data). Compensation is delayed in T. angasi- but not to the same extent that it is in B. taurus.

If compensatory change is instantaneous, then complementary pairs of stem nu- cleotides should be down-weighted by one-half. As suggested above, this is not necessarily the case in our data. Most taxa do not show perfect Watson-Crick base pairing according to the proposed secondary rRNA structure of the domestic cow. This may be a result of either an inaccurate secondary-structure model for divergent mt-rRNA molecules, lags in compensation, or limited functional importance of some stem nucleotide pairings.

We followed a variation of the weighting procedure of Steele et al. ( 1988, 199 1). Complementary base-pairing nucleotide positions were identified. The thirty-two pairs of sites that were characterized by more than one type of Watson-Crick base pairing were then down-weighted relative to uncompensated sites. The strict-consensus trees derived from these data were more resolved but never inconsistent with the analysis

Dow


ic.oup.com/m


ber 2021

Bovid mt-rDNA Phylogeny 44 1

A D (G>A)

6.t. c

A (C>Tl T.a. T

I Bo.t. C

1 ’ Cm. C

I C.h. C

A (C>T) 0.g. T

h

-I D (G>A) 1 A.n. T

I He. c

1 I- D.d. C

1 D(G,A) Dh c .

A.a. C

M.r. C

D A

A

G

G

A

G

G

A

A

G

G

A

G

G

FIG. 4.-Unambiguous state transformations for complementary stem characters A and D on one of our shortest trees. A case of covarying convergence in these nucleotide positions can be seen both in the line leading to G.t. and in the line leading to O.g.+A.n. In the transformation of character A on the line leading to B.t., there is a lag 3 11.5 Myr, while the line leading to T.a. is complemented in this same time period. In the D.h. lineage, compensatory change in character A has been delayed G2.5 Myr. Unambiguous lag in compensation is represented by the boldface line. Species abbreviations are as in fig. 2.

of the equally weighted data, whether the complementary nucleotide pairs were down- weighted by the maximum, by one-half, or less severely. For instance, the analyses constrained by bovid monophyly additionally supported the sister group Redun- cini+Antilopini while the unconstrained analyses supported Alcelaphini+Hippotragini.

Discussion

Factors that have contributed to the general confusion in bovid systematics include morphological convergence, rapid radiation, and questionable monophyly of the group. First, it has been demonstrated that convergent evolution is rampant in the Bovidae as well as in other ruminants (Vrba 1979; Janis and Scott 1987). For instance, hyp- sodonty, the lack of horns in females, the relative decrease in length of the premolar

Dow


ic.oup.com/m


ber 2021

442 Gatesy et al.

tooth row, as well as numerous other functional characters are suspected to have arisen multiple times within the family, obscuring true evolutionary relationships.

Second, the pattern of bovid evolution inferred roughly from the fossil record is one of episodic speciation (Vrba 1985). First fossil appearances of many bovid subfamilies and tribes are clustered in time, suggesting a geologically near-instantaneous origination of the major bovid groups in the Miocene. It is possible that such fossil patterns may result from taphonomic biases, as some geological time frames are much better sampled than others. However, the same pattern of rapid radiation among bovid tribes was discovered by Lowenstein ( 1986b) and Georgiadis et al. ( 199 1). Using independent measures of molecular distance, they found that many of the bovid tribes were genetically equidistant from each other. If rates of molecular change are assumed to be, on the average, constant through long time, then this is indicative of a rapid splitting of lineages. Such radiations are notoriously difficult to resolve with either morphological or molecular data (e.g., eutherian mammal orders; Wyss et al. 1987; Novacek et al. 1988). When character change is approximately clocklike, rapid cladogenesis allows little time for synapomorphic changes to accumulate, thereby re- ducing robustness of groupings.

Finally, it is not firmly established that the bovids form a natural clade. Specifically, there is only one morphological character that unambiguously defines the bovids: their nondeciduous horn cores and horn sheaths (Janis and Scott 1987). Molecular evidence for a monophyletic Bovidae is equally sparse. In fact, a paraphyletic status has been presented three times in the literature (Miyamoto and Goodman 1986; Irwin et al. 199 1; Kraus and Miyamoto 199 1) . Similarly, Beintema et al. ( 1986) offered no support for bovid monophyly.

Kraus and Miyamoto ( 199 1) hypothesized that the basal radiation of bovids was near or even within the initial radiation of the pecoran families. Our sequence data hint at the possibility that the bovids are paraphyletic. For instance, Antilocapra clusters with the Boselaphini in our total unconstrained data analysis (fig. 2A). In most analyses, extra evolutionary steps are invoked to satisfy the condition of bovid monophyly, although the additional number of steps is not great in comparison with overall tree length (table 2). Because several independent molecular studies have also hypothesized bovid paraphyly, and because no detailed cladistic morphological hypothesis has been proposed, the evolutionary uniqueness of the group remains questionable.

The paucity of support for bovid monophyly is emphasized by the lack of resolution of its nearest evolutionary relatives. The Bovidae along with the families An- tilocapridae (pronghorn antelope), Giraffidae (giraffes), Cervidae (antlered deer), and Moschidae (musk deer) form the extant components of the infraorder Pecora. Both molecular (Beintema et al. 1986; Lowenstein 1986a; Miyamoto and Goodman 1986; Irwin et al. 199 1; Kraus and Miyamoto 199 1) and morphological (Romer 1966; Hamilton 1978; Leinders and Heintz 1980; Webb and Taylor 1980; Janis and Scott 1987; Gentry and Hooker 1988) systematists have investigated the phylogenetic relationships within this group, yet there is no consensus on its evolutionary history. There are only 15 possible rooted cladograms relating the families Bovidae, Cervidae, Antilocapridae, and Giraffidae; 11 have been proposed in the literature. The phylogenetic confusion at this higher level makes rigorous outgroup analysis (Maddison et al. 1984) problematic for the Bovidae.

It is not surprising that hierarchical organization of the basal radiation of bovids has proved to be an intractable problem. Simpson’s ( 1945) classification based on morphology (fig. 5A) allocated the family into 13 tribes, most of which are thought

Dow


ic.oup.com/m


ber 2021

Caprini

Cephalophini

Antilopini

Reduncini

Hippotragini

Alcelaphini

A B

C

Hippotragini

Reduncini

Caprini

Cephalophini

D

FOG. 5.-Previous hypotheses of botid interrelationships for tribes sampled in present study. A, Simpson ( 1945) morphology. B, Strict consensus of hypotheses A, C, and D. C, Lowenstein ( 1986b) immunodiffusion. D, Beintema et al. ( 1986) pancreatic-type ribonuclease protein sequence. E, Georgiadis et al. ( 1991) ahozymes- neighbor-joining tree of Nei’s genetic distances. F, Georgiadis et al. ( 199 1) allozymes-parsimony consensus tree.

Dow


ic.oup.com/m


ber 2021

444 Gatesy et al.

to be monophyletic. However, subfamilial relationships of tribes were not necessarily assumed to be natural. Lowenstein ( 1986b) presented an immunodifhtsion phenogram with some resolution, but many tribes were sunk into a large polytomy (fig. 5C). A parsimony analysis of amino acid replacements in pancreatic-type ribonuclease sequences (fig. 5D; Beintema et al. 1986) retained Simpson’s subfamily Bovinae (Tra- gelaphini+Bovini+Boselaphini), in agreement with Lowenstein (fig. 5B). Georgiadis et al.‘s ( 199 1) distance and parsimony analyses of allozymes (fig. 5E and F) allow contrasting interpretations of bovid history. However, outgroups were too distant to root these networks. All of the above contributions are marked by both a lack of resolution at the tribal level and a general incongruence with each other. The strict consensus of these trees is bereft of structure.

Although some of our strict-consensus trees are well resolved, a bootstrap analysis of the equally weighted total data set revealed no strong evolutionary ties among bovid tribes. Because no robust intertribal relationships are obvious from our analyses, we interpreted our results by searching for monophyletic groups shared by our study and previous studies (table 2). Strict-consensus trees from our work are characterized by four groupings strictly congruent with past analyses: Tragelaphini+Bovini (Bein- tema et al. 1986), Hippotragini+Alcelaphini (Georgiadis et al. 199 1 ), Bo- vini+Tragelaphini+Boselaphini (Simpson 1945; Beintema 1986; Lowenstein 1986b), and Reduncini+Hippotragini+Antilopini+Cephalophini+Caprini+Alcelaphini (Lowenstein 1986b). Weakly supported clades unique to our shortest trees may be tested by future systematic work. The radiation of the bovids will not be resolved until additional morphological and molecular characters are analyzed.

Acknowledgments

We thank Oliver Ryder (San Diego Zoo) and Dan Wharton (New York Zoological Society) for providing tissue samples. Susan Tanhauser and Michael Miyamoto allowed access to unpublished sequences. George Amato, Diane Bridge, Matt Dick, Cheryl Hayashi, and Annie Williams read early drafts of the manuscript. Leo Buss provided use of his computer. Walter Fitch provided many useful comments. This research was supported by NSF grant BSR-8907673 awarded to E.V. and R.D.

LITERATURE CITED

ANDERSON, S., M. H. DEBRUIJN, A. R. COULSON, I. C. EPERON, F. SANGER, and I. G. YOUNG. 1982. Complete sequence of bovine mitochondrial DNA: conserved features of the mammalian mitochondrial genome. J. Mol. Biol. 156:683-717.

BEINTEMA, J. J., W. M. FITCH, and A. CARSANA. 1986. Molecular evolution of pancreatic-type ribonucleases. Mol. Biol. Evol. 3:262-275.

BROWN, W. M., E. M. PRAGER, A. WANG, and A. C. WILSON. 1982. Mitochondrial DNA sequences of primates: tempo and mode of evolution. J. Mol. Evol. 18:225-239.

CACCONE, A., G. D. AMATO, and J. R. POWELL. 1987. Intraspecific DNA divergence in Dro- sophila: a study on parthenogenetic D. mercatorurn. Mol. Biol. Evol. 4:343-350.

FARRIS, J. S. 1989. The retention index and the resealed consistency index. Cladistics 5:417- 419.

FITCH, W. M. 1971. Toward defining the course of evolution: minimum change for a specific tree topology. Syst. Zool. 20:406-4 16.

GENTRY, A. W. 1978. Bovidae. Pp. 540-572 in V. J. MAGLIO and H. B. S. COOKE, eds. Evolution of African mammals. Harvard University Press, Cambridge, Mass.

GENTRY, A. W., and J. J. HOOKER. 1988. The phylogeny of the Artiodactyla. Pp. 235-272 in M. J. BENTON, ed. The phylogeny and classification of the tetrapods, vol. 2: Mammals. Clarendon, Oxford.

Dow


ic.oup.com/m


ber 2021


GEORGIADIS, N. J., P. W. KAT, H. OKETCH, and J. PATTON. 199 1. Allozyme divergence within the Bovidae. Evolution 44:2 135-2 149.

GUTELL, R. R., and G. E. FOX. 1988. A compilation of large subunit RNA sequences presented in a structural format. Nucleic Acids Res. 16[ Suppl.] : R 17%R3 13.

GUTELL, R. R., B. WEISER, C. WOESE, and H. F. NOLLER. 1985. Comparative anatomy of 16- S-like ribosomal RNA. Prog. Nucleic Acids Res. 32: 155-2 15.

GYLLENSTEN, U. B., and H. A. ERLICH. 1988. Generation of single-stranded DNA by the polymerase chain reaction and its application to direct sequencing of the HLA-DQA locus. Proc. Natl. Acad. Sci. USA 85:7652-7656.

HAMILTON, W. R. 1978. Fossil giraffes from the Miocene of Africa, and a revision of the phylogeny of the Giraffoidea. Philos. Trans. R. Sot. Lond. [B] 282:165-229.

HEDGES, S. B., K. D. MOBERG, and L. R. MAXSON. 1990. Tetrapod phylogeny inferred from 18s and 28s ribosomal RNA sequences and a review of the evidence for amniote relationships. Mol. Biol. Evol. 7:607-633.

HILL, A., R. DRAKE, L. TAUXE, M. MONAGHAN, J. C. BARRY, A. K. BEHRENSMEYER, G. CURTIS, B. FINE JACOBS, L. JACOBS, N. JOHNSON, and D. PILBEAM. 1985. Neogene paleontology and geochronology of the Baring0 Basin, Kenya. J. Hum. Evol. 14:759-773.

HIXSON, J. E., and W. M. BROWN. 1986. A comparison of the small ribosomal RNA genes from the mitochondrial DNA of the great apes and humans: sequence, structure, evolution, and phylogenetic implications. Mol. Biol. Evol. 3: 1- 18.

HONACKI, J. H., K. E. KINMAN, and J. S. KOEPPL. 1982. Mammal species of the world. Allen, Lawrence, Kans.

IRWIN, D. M., T. D. KOCHER, and A. C. WILSON. 199 1. Evolution of the cytochrome b gene of mammals. J. Mol. Evol. 32: 128-144.

JANIS, C. M., and K. M. SCOTT. 1987. The interrelationships of higher ruminant families with special emphasis on the members of the Cervoidea. Novitates 2893:1-85.

KOCHER, T. D., W. K. THOMAS, A. MEYER, S. V. EDWARDS, S. PAABO, F. X. VILLABLANCA, and A. C. WIL,SON . 1989. Dynamics of mitochondrial DNA evolution in animals: amplification and sequencing with conserved primers. Proc. Natl. Acad. Sci. USA 86:6 196-6200.

KRAUS, F., and M. M. MIYAMOTO. 199 1. Rapid cladogenesis among the Pecoran ruminants: evidence from mitochondrial DNA sequences. Syst. Zool. 40: 117- 130.

LAKE, J. A. 1987. A rate-independent technique for analysis of nucleic acid sequences: evolutionary parsimony. Mol. Biol. Evol. 4: 167-19 1.

LARSON, A., and A. C. WILSON. 1989. Patterns of ribosomal RNA evolution in salamanders. Mol. Biol. Evol. 6: 13 l- 154.

LEINDERS, J. J. M., and E. HEINTZ. 1980. The configuration of the lacrimal orifice in pecorans and tragulids ( Artiodactyla; Mammalia) and its significance for the distinction between Bo- vidae and Cervidae. Beaufortia 30: 155- 160.

LOWENSTEIN, J. M. 1986~. Bovid relations based on serum immunology. S. Afr. J. Sci. 82:77- 78.

-. 19866. Molecular phylogenetics. Annu. Rev. Earth Planet Sci. 14:7 l-83. MADDISON, W. P., M. J. DONOGHUE, and D. R. MADDISON. 1984. Outgroup analysis and

parsimony. Syst. Zool. 33:83-103. MIYAMOTO, M. M., and M. GOODMAN. 1986. Biomolecular systematics of eutherian mammals:

phylogenetic patterns and classification. Syst. Zool. 35:230-240. MIYAMOTO, M. M., F. KRAUS, and 0. A. RYDER. 1990. Phylogeny and evolution of antlered

deer determined from mitochondrial DNA sequences. Proc. Natl. Acad. Sci. USA 87:6 127- 6131.

MIYAMOTO, M. M., S. M. TANHAUSER, and P. J. LAIPIS. 1989. Systematic relationships in the a&dactyl tribe Bovini (family Bovidae), as determined from mitochondrial DNA sequences. Syst. Zool. 38:342-349.

NEFF, N. A. 1986. A rational basis for a priori character weighting. Syst. Zool. 35:110-123. NOVACEK, M. J., A. R. WYSS, and M. C. MCKENNA. 1988. The major groups of eutherian

Dow


ic.oup.com/m


ber 2021

446 Gatesy et al.

mammals. Pp. 3 l-7 1 in M. J. BENTON, ed. The phylogeny and classification of the tetrapods, vol. 2: Mammals. Clarendon, Oxford.

PAMILO, P., and M. NEI. 1988. Relationships between gene trees and species trees. Mol. Biol. Evol. 5:568-583.

ROMER, A. S. 1966. Vertebrate paleontology. University of Chicago Press, Chicago. SAVAGE, D. E., and D. E. RUSSELL. 1983. Mammalian paleofaunas of the world. Addison-

Wesley, Reading, Mass. SIMPSON, G. G. 1945. The principles of classification and a classification of mammals. Bull.

Am. Museum Nat. Hist. 85:1-350. SMITH, A. B. 1989. RNA sequence data in phylogenetic reconstruction: testing the limits of its

resolution. Cladistics 5:321-344. STEELE, K. P., K. E. HOLSINGER, R. K. JANSEN, and D. W. TAYLOR. 1988. Phylogenetic

relationships in green plants-a comment on the use of 5S ribosomal RNA sequences by Bremer et al. Taxon 37:135-138.

-. 199 1. Assessing the reliability of 5s rRNA sequence data for phylogenetic analysis in green plants. Mol. Biol. Evol. 8:240-248.

SWOFFORD, D. 1990. PAUP: phylogenetic analysis using parsimony, version 3.0. Illinois Natural History Survey, Champaign.

VRBA, E. S. 1979. Phylogenetic analysis and classification of fossil and recent Alcelaphini (Family Bovidae, Mammalia). J. Linnean Sot. (Zool.) 11:207-228.

-. 1984. Evolutionary pattern and process in the sister group Alcelaphini-Aepycerotini (Mammalia: Bovidae), Pp. 62-79 in N. ELDREDGE and S. M. STANLEY, eds. Living fossils. Springer, New York.

-. 1985. African Bovidae: evolutionary events since the Miocene. S. Afr. J. Sci. 81:263- 266.

-. 1987. Ecology in relation to speciation rates: some case histories of Miocene-recent mammal clades. Evol. Ecol. 1:283-300.

WEBB, S. D., and B. E. TAYLOR. 1980. The phylogeny of hornless ruminants and a description of the cranium of Archaeomeryx. Bull. Am. Museum Nat. Hist. 167: 117- 158.

WHEELER, W. C. 1990. Combinatorial weights in phylogenetic analysis: a statistical parsimony procedure. Cladistics 6:269-275.

WHEELER, W. C., and R. L. HONEYCUTT. 1988. Paired sequence difference in ribosomal RNAs: evolutionary and phylogenetic implications. Mol. Biol. Evol. 5:90-96.

WILLIAMS, P. L., and W. M. FITCH. 1989. Finding the minimal change in a given tree. Pp. 453-470 in B. FERNHOLM, K. BREMER, L. BRUNDIN, H. J~~RNVALL, L. RUTBERG, and H. E. WANNTROP, eds. The hierarchy of life. Elsevier, Amsterdam.

WYSS, A. R., M. J. NOVACEK, and M. C. MCKENNA. 1987. Amino acid sequence versus morphological data and the interordinal relationships of mammals. Mol. Biol. Evol. 4:99- 116.

WALTER M. FITCH, reviewing editor

Received June 13, 1991; revision received October 14, 1991

Accepted October 14, 199 1

Dow


ic.oup.com/m


ber 2021

Documents

Phylogeny of the Bovidae ( Artiodactyla, Mammalia ) , Based on