Lactate Dehydrogenase (LDH) Gene Duplication During Chordate ...

Lactate Dehydrogenase (LDH) Gene Duplication During Chordate Evolution: The cDNA Sequence of the LDH of the Tunicate StyeZaplicata David W. Stock, *iclJoseph M. Quattro, *2 Gregory S. Whitt,t_ and Dennis A. Powers* *Department of Biological Sciences, Hopkins Marine Station, Stanford University; Evolution, University of Illinois at Urbana-Champaign

and TDepartment

of Ecology,

Ethology

and

L-Lactate dehydrogenase (L-LDH, E.C. 1.1.1.27) is encoded by two or three loci in all vertebrates examined, with the exception of lampreys, which have a single LDH locus. Biochemical characterizations of LDH proteins have suggested that a gene duplication early in vertebrate evolution gave rise to Ldh-A and Ldh-B and that an additional locus, Ldh-C arose in a number of lineages more recently. Although some phylogenetic studies of LDH protein sequences have supported this pattern of gene duplication, others have contradicted it. In particular, a number of studies have suggested that Ldh-C represents the earliest divergence among vertebrate LDHs and that it may have diverged from the other loci well before the origin of vertebrates. Such hypotheses make explicit statements about the relationship of vertebrate and invertebrate LDHs, but to date, no closely related invertebrate LDH sequences have been available for comparison. We have attempted to provide further data on the timing of gene duplications leading to multiple vertebrate LDHs by determining the cDNA sequence of the LDH of the tunicate Styelu plicatu. Phylogenetic analyses of this and other LDH sequences provide strong support for the duplications giving rise to multiple vertebrate LDHs having occurred after vertebrates diverged from tunicates. The timing of these LDH duplications is consistent with data from a number of other gene families suggesting widespread gene duplication near the origin of vertebrates. With respect to the relationships among vertebrate LDHs, our data are not consistent with previous claims that Ldh-C represented the earliest divergence. However, the precise relationships among some of the main lineages of vertebrate LDHs were not resolved in our analyses.

Introduction Comparisons of nuclear genome sizes and numbers of loci encoding isozymes and other proteins in a variety of chordates led Ohno (1970) to propose that episodes of large-scale gene duplication had occurred during the evolution of the group. The more recent discovery that many genes involved in the control of development are members of multigene families with more representatives in vertebrates than in invertebrates has led to increased interest in documenting the patterns of gene duplication in chordate evolution (Holland 1992; Holland et al. 1994). The sizes of some of these families, coupled with the difficulty in determining gene number in a large sample of taxa, suggest the value of continued examination of smaller, more extensively studied gene families in detecting episodes of gene duplication in chordate evolution. One such gene family encodes L-lactate dehydrogenase (L-LDH, E.C. 1.1.1.27), which catalyzes the interconversion of pyruvate (the end product of glycolysis) and L-lactate with the concomitant interconversion of NADH and NAD+ (Markert 1984). Extensive surveys of LDH locus number, expression pattern, kiand immunochemical similarities in netic properties, representatives of the major taxa of jawed vertebrates (Marker& Shaklee, and Whitt 1975; Fisher et al 1980; Whitt 1984) have suggested that they all possess orthologs of Ldh-A and Ldh-B. LDH-A is expressed pre’Present address: Department State University.

of Anthropology,

* Present address: Department of Biology, ence, University of South Carolina. Key words: tunicate,

molecular

phylogeny,

The Pennsylvania

Program

in Marine Sci-

gene duplication.

Address for correspondence and reprints: David W. Stock, Department of Anthropology, The Pennsylvania State University, University Park, Pennsylvania 16802. E-mail: [email protected]. Mol. Biol.

Evol. 14(12): 1273-1284. 1997 0 1997 by the Society for Molecular Biology

and Evolution.

ISSN: 0737-4038

dominantly in white skeletal muscle and is kinetically suited to the reduction of pyruvate to lactate under conditions of low oxygen availability, while LDH-B is predominantly expressed in more aerobic tissues and is tailored to the oxidation of lactate to pyruvate (Markert 1984). An additional locus encoding an isozyme with more variable kinetic properties (LDH-C) is expressed in the sperm of mammals and columbid birds and in a variety of tissues or the liver or eye of ray-finned fishes (Markert, Shaklee, and Whitt 1975). Jawless fishes possess either a single locus (lampreys) or, in the case of hagfishes, two loci with tissue expression patterns similar to those of jawed vertebrate Ldh-A and Ldh-B (Markert, Shaklee, and Whitt 1975). Markert, Shaklee, and Whitt (1975) proposed that an initial gene duplication gave rise to Ldh-A and Ldh-B early in vertebrate evolution and that Ldh-C is the result of a more recent duplication. Subsequent phylogenetic analyses of vertebrate LDH sequence data have supported some aspects of this “traditional” scenario for LDH evolution and challenged others (Li et al. 1983; Hendriks et al. 1988; Crawford, Constantino, and Powers 1989; Stock and Whitt 1992; Quattro, Woods, and Powers 1993; Tsuji et al. 1994; Quattro et al. 1995). Some points that remain controversial or poorly resolved are the phylogenetic position of the LDH-Cs, the relationship of cyclostome LDHs to gnathostome LDHs, and the monophyly of gnathostome LDH-As. The former two issues involve the earliest duplication events leading to the multigene family in vertebrates. Such duplications would be further elucidated by an examination of the relationship of invertebrate and protochordate LDHs to those of vertebrates. To date, the only complete LDH sequence available from an invertebrate animal is that of the nematode Caenorhabditis elegans (Tsoi and Li 1994), a represen1273

1274

Stock et al.

tative of a group that is quite distant from vertebrates. Among protochordates, tunicates and amphioxus have been reported to possess single LDH loci (Fisher et al. 1980; Baldwin, Mortimer, and Patak 1988). Immunochemical analyses and comparisons of amino acid composition were interpreted as evidence that tunicate LDH was more closely related to the LDH-C of a teleost fish than to either LDH-A or LDH-B (Baldwin, Mortimer, and Patak 1988), while Fisher et al. (1980) suggested that the single-locus condition of amphioxus and tunicates predated the divergence of these groups from vertebrates. We have attempted to provide additional resolution of the early duplication history of chordate LDHs by cDNA sequence determination and phylogenetic analysis of an LDH locus from the tunicate Sty& plicata.

Adult specimens of the tunicate Styelu plicata were obtained from Gulf Specimen Company (Panacea, Fla.). A section of the body wall was cut from a single specimen, and the inner and outer surfaces were removed with a scalpel. The remaining portion was ground to a fine powder in liquid N2 using a mortar and pestle. Total RNA was extracted from 2 g of powdered tissue by homogenization in the presence of sodium tri-isopropylnaphthalene sulfonate followed by phenol/chloroform extraction, ethanol precipitation, and precipitation from sodium acetate, as previously described (Stock et al. 1991). Transcription

Total RNA was reverse transcribed with MMLV reverse transcriptase and either random hexamers or a dT adaptor primer (5’-GATGCTGCAGAGCTCAAGC(T)is3’) as described by Stock and Powers (1995). RT-PCR Amplification Styela Ldh

RACE Amplification

of Internal

Fragment

of

A number of degenerate PCR primers (designed from an alignment of bacterial and vertebrate LDH protein sequences-Stock and Whitt 1992; Stock and Powers 1995) were tested for their ability to amplify a product of the expected size from Styela first-strand cDNA. Of these degenerate primers, only Ldhll3FhF (5’GCGAATTCGTNGGNATGGCNTGYGCNRT-3’, where the underlined sequences indicate restriction sites, and ambiguous bases are indicated by the appropriate IUPAC codes) and LdhB736FhR (S’-GCCGCTGCAGCCCANBWNGTRTANCCYTT-3’) amplified a fragment of Styela L+dh. The sequence obtained from this fragment was used to design gene-specific primers for additional RT-PCR (S’-GCCGGAGCTCAAGGTGAAGTAATGGACT-3 ‘, used in combination with primer LdhB736FhR) and RACE amplifications (see below). Conditions for RT-PCR were as described by Stock and Powers (1995). Weak bands were reamplified in a similar PCR reaction using as a template either a dilution of a band cut from a low-melting-point agarose gel (NuSieve-FMC) or the DNA adhering to a syringe

of the 3’ End of Styela Ldh

The 3’ end of the Styelu LDH cDNA was amplified by 3’ RACE, essentially as described by Frohman (1990). A dT-adaptor reverse transcription was used as the template for an initial PCR with an adaptor primer (5’-GATGCTGCAGAGCTCAAGCTT-3’) and the gene-specific primer 5’-GCCGGAGCTCGRTGGMCCWGAAGGYTGGGA-3’. PCR conditions were as described by Stock and Powers (1995). The initial PCR was then used as a template for a second PCR with the adaptor primer 5’GATGCTGCAGAGCTCAAG-3’ and the gene-specific primer 5’-GCCGGAGCTCACAYAARCARGTYRTYGATG-3’. RACE Amplification

Materials and Methods RNA Extraction

Reverse

needle stabbed into a band of the expected size (band stab technique of Bjourson and Cooper 1992).

of the 5’ End of Styela Ldh

5’ RACE amplification followed Frohman (1990). RNA was reverse transcribed with the degenerate LDH primer 5’-GCCGCTGCAGTCNACNACYTGYTTRTG3’. The first-strand cDNA was tailed with dATP using terminal deoxynucleotidyl transferase. PCR amplification of the dA-tailed template was with the gene-specific primer 5’-GCCGCTGCAGTGGCCGATACGGAGTAATCT-3’, the dT adaptor primer, and one of the adaptor primers used above. The PCR reaction was subjected to electrophoresis on an agarose gel, and the band stab technique (Bjourson and Cooper 1992) was used to reampufy a faint band of the expected size with the nested LDH primer 5’-GCCGCTGCAGCCRTGYTGNARRTCCAT-3’ and the same adaptor primer as in the first reaction. Cloning PCR products were cloned into Ml3 mp18 and mp 19 or pBluescript KS+ and KS - (Stratagene) using standard procedures (Sambrook, Fritsch, and Maniatis 1989). PCR products were either digested with restriction enzymes (recognizing sites at the 5’ ends of the primers or in the region between the primers), made blunt-ended with the Klenow fragment of DNA polymerase I, or cloned directly using the t-vector technique (Marchuk et al. 1991). Sequencing Single-stranded DNA for sequencing was either extracted from Ml 3 phage particles or obtained from pBluescript-containing cultures infected with VCSM13 helper phage (Stratagene) according to Sambrook, Fritsch, and Maniatis (1989). DNA sequencing was by dideoxy chain termination using a modified T7 DNA polymerase (Sequenase, version 2.0, United States Biochemicals). In order to minimize errors associated with the sequencing of cloned PCR products, each nucleotide was determined from at least six clones, with at least three sequences representing each strand. Sequence

Assembly

and Alignment

The cDNA sequence of the LDH of Styelu plicata was assembled from the consensus of overlapping clones using the SEQMAN contig assembly program in

Tunicate

the DNASTAR package (DNASTAR, Inc.). The sequence described in this paper has been deposited in GenBank under accession number AF023168 (fig. 1). Additional lactate dehydrogenase sequences were obtained from the GenBank (release 92) or PIR database. D-LDHs (E.C. 1.1.1.28) were not included in the analyses, as they represent an, at best, distantly related protein family or families (Kochhar et al. 1992). The LLDH sequences used and their accession numbers (GenBank, unless otherwise noted) are: lamprey (Petromyzon marinus) LDH (M74064); dogfish shark (Squalus acanthias) LDH-A (U38893); killifish (Fundulus heteroclitus) LDH-A (I-43525), LDH-B (M33969), and LDH-C (L07336); frog (Xenopus Zaevis) LDH-A (UO7179), LDH-B (UO7176), and LDH-C (U07175); duck LDH-B (503869); chicken LDH-A (X53828) and LDH-B (PIR accession A00346); cattle LDH-A (D90143); pig LDH-A (U07178) and LDH-B (UO7180); rabbit LDH-A (M22585); human LDH-A (X02152), LDH-B (YOO711), and LDH-C (502938); mouse LDHA (YOO309), LDH-B (X51905), and LDH-C (X04752); rat LDH-A (X01964) and LDH-C (U07177); fox (Vulpes vulpes) LDH-C (U19868); nematode (C. elegans) LDH (U15420); barley LDH-A (M55685) and LDH-B (M55684); rice LDH (D13817); maize LDH (211754); the LDHs of the apicomplexan protists ToxopZusma gondii (U35 118) and Plasmodium falciparum (M93720); and the L-LDHs of the eubacteria Streptococcus mutuns (M72545), S. thermophilus (D 13405), Luctococcus lactis (M88490), Luctobacillus casei (M76708), Luctobacillus sake (U26688), Lactobacillus plantarum (D90340), Pediococcus acidilactici (X70927), Bacillus stearothermophilus (M19396), B. caldolyticus (M19394), B. caldotenax (M19395), B. psychrosacchar118 and LDH-B-X55 119), B. olyticus (LDH-A-X55 subtilis (PIR accession A25805), B. megaterium (M22305), Mycoplasma hyopneumoniae (X67286), M. genitalium (U39733), Thermus aquaticus (D00585), T. caldophilus (X045 19), Deinococcus radiodurans (D63899), Thermotoga maritima (X74302), BiJidobacterium longum (M33585), and Synechocystis sp. (D64003). Nucleotide sequences were translated into protein sequences according to the authors’ feature specifications. An alignment of amino acid sequences was constructed using CLUSTAL W (Thompson, Higgins, and Gibson 1994) and the default settings. Length variability among chordate LDH sequences in this alignment was limited to the amino terminal arm (positions l-43 in fig. 2, as defined by Li et al. 1983) and the extreme carboxy terminus. No attempt was made to improve the alignment of the former region because of extensive sequence variability, but the alignment of the later region was adjusted manually. Invertebrate and plant LDH sequences require the addition of only a few additional gaps to the alignment of chordate LDHs, but the alignment of bacterial and protist sequences appeared less reliable. Rather than attempting additional alignment methods or parameters, we conducted phylogenetic analyses with and without the latter two groups of LDHs.

Phylogenetic

LDH

1275

Analyses

Positions l-44 in the alignment of LDH sequences (fig. 2) were omitted from all analyses because of the difficulty in obtaining a reliable alignment of this region. Phylogenetic analyses were conducted using both neighbor-joining and parsimony methods. Amino acids were used as characters in both types of analysis because of the large amount of sequence divergence among some of the taxa and the fact that we were interested in early branching events in LDH evolution (Kumar, Tamura, and Nei 1993). Neighbor-joining trees were constructed from pairwise amino acid distances (gamma corrected with a = 2.0) using the MEGA package of programs (Kumar, Tamura, and Nei 1993). Parsimony analyses were conducted with PAUP, version 3.0s (Swofford 1991), and in most analyses, amino acids were treated as unordered characters with all possible transformations weighted equally. Preliminary analyses with amino acid replacements weighted according to the minimum number of nonsynonymous nucleotide substitutions required (PROTPARS weighting-Felsenstein 1993) gave similar results but were not pursued extensively because of the computational time required. Parsimony analyses with equal weighting were performed with 100 random addition sequences of taxa and TBR branch swapping (Swofford 1991), while those with PROTPARS weighting used 20 random addition sequences. The CONSTRAINTS feature of PAUP was used to determine the lengths of the shortest trees matching a number of alternative phylogenetic hypotheses, In the case of constrained analyses using PROTPARS weighting, the PROTPARS program of the PHYLIP package (Felsenstein 1993) was used to test whether the length differences between constrained trees and the best unconstrained tree were significant according to the method of Templeton (1983). Bootstrapping was used to assess the degree of support for particular nodes. Neighborjoining analyses were conducted with 500-1,000 bootand bootstrap parsimony analyses strap replications, were conducted with 10 replications of a random addition sequence for each of 100 bootstrap replications.

Results Features

of Styelu Ldh

The 1,168-nt cDNA sequence of S. plicatu LDH is shown in figure 1. Three of 10 clones of the 3’ end had an additional 10 nucleotides before the poly-A tail (not shown). No consensus polyadenylation signal (Sheets, Ogg, and Wickens 1990) was found in either class of clone, suggesting that the reverse transcription primer may have bound upstream of the true poly-A tail. Of the sites at which nucleotide differences from the sequence shown in figure 1 were found in individual clones of PCR products, only seven exhibited the same difference in more than one clone. Two of these differences occurred in the 3’ untranslated region (positions 116 1 and 1164), three were silent (positions 923, 97 1, and 1037), and only two encoded different amino acids (N and P at nucleotide positions 454 and 583, respectively). These latter two differences were only found in

1276

Stock et al.

GATTGGAGTATTTGGTTGTATTAAGAATTTTGATAATTAAACACAA TCATGGCTGATCTGAAGGGATTCGATATAATGTCCGAGTTATTCA MADLKGFDIMSELF CGGAAGTCTGTCCAGACATTCCAAAATCTGGACCATCCAAAGTCA TEVCPDIPKSGPSKV CTGTAGTTGGAGTTGGGATGGTGGGGATGGCGTGTGGGATGAGTG TVVGVGMVGMACGMS TGGTGTTGAAGGGACTATGTACAGATCTTGTACTAGTCGATGTTG VVLKGLCTDLVLVDV TTCAGGACAAACTTCAAGGTGAAGTAATGGACTTGCAACATGGTA VQDKLQGEVMDLQHG GTTTGTTTCTAGAGAATATTCTATGGAGACAAAGATTACT SLFLENIKVYGDKDY CCGTATCGGCCAACAGTCGGATTGTCATAGTAACTGCTGGTGCTC SVSANSRIVIVTAGA GTCAACZUXCTGGGGAATCTAGATTGAGTTCAACGG RQQPGESRLSLVQRN TCAACATTTTCTTCCACAAATTGCGAAATATAGTC VNIFKHIIPQIAKYS CAAGn=CTATTCTTGTTATn=TTTCAAATCCAGTTGACTT PSAILVIVSNPVDLM CATATGTTGCCTGGAAATTGTCAAACTTTCCTCGTAATCGTGTGA TYVAWKLSNFPRNRV TTGGTTCTGGZUuZGAATCTTGATTCTGCJUiGATTTCGACATCTTA IGSGTNLDSARFRHL TCGCGGAAAAATIY3AATTTGTCACCTGTCAGTGTTCATG IAEKLNLSPVSVHGW TCATCGGCGAACATGGCGATTCAAGTGTTCCAATGTGGAG IIGEHGDSSVPMWSG TTAACGTGAGTGGGZUATGTCGATTCATCCAAGAATAG VNVSGKCLNSIHPRI GATATCCGGATGGACCAGAAGGTTGGGATAAAATACATACATAAACAAG GYPDGPEGWDKIHKQ TTGTCGATGGGGCATACGATGTAATAAGATTAAAAGGATACACAA VVDGAYDVIRLKGYT ACTGGGCTATTGGTCTGAGCTGTGCAGAATTGCTTGCA?iCTATTT NWAIGLSCAELLATI TACATCACAGACATAGAATCCATCCAGTTACTTGTTTTGTGAAGG LHHRHRIHPVTCFVK GCCGATATGGAATCACTGATGATGTATGCCTATCTCTACCTTGTG GRYGITDDVCLSLPC TTTTGZUXX'GTAATGGAGTCAACTCTATTGT~TGTTGATTTGA VLNCNGVNSIVNVDL CTGCTGAAGAAGAAGCAATTGCAATGACTATTG TAEEEAMIKKSAMTI CTGATGTG CAAAAAGGCCTGAAATGGTAAAATTTACCAGAATACT ADVQKGLKW* ACATAAAAAGCTTTCATTGTGCTCTTTCTCAAAGCGCTTTCAAGT CCCTTTAATGTCATGGCCGATATGCTAAAAGACCCAGACTTGT FIG. l.+DNA cata L_dh.

and deduced

amino acid sequence

45 90 14 135 29 180 44 225 59 270 74 315 89 360 104 405 119 450 134 495 149 540 164 585 179 630 194 675 209 720 224 765 239 810 254 855 269 900 284 945 299 990 314 1035 329 1080 338 1125 1168

of Styela pli-

2 of 11 and 3 of 15 clones and may have been PCR errors. Two of the silent differences (positions 923 and 971) may be the result of heterozygosity, as two bands of relatively equal intensity were seen at these positions on autoradiographs of directly sequenced (uncloned) independent PCR products (data not shown). The Styelu LDH protein that we inferred from the nucleotide sequence (assigning the start codon according to the properties of vertebrate translation start sites listed by Kozak 1996) is 338 amino acids long including the initial M; vertebrate LDHs, in comparison, range from 332 to 334 amino acids in length. The additional length of the tunicate sequence is in the N-terminal arm (AbadZapatero et al. 1987 and included references), a region that has previously been shown to also exhibit the greatest degree of sequence variability among vertebrate LDHs and is reduced or absent in bacterial sequences (Li et al. 1983). The only other region of length difference among chordate LDHs reported to date is near the carboxy terminus (fig. 2); at the penultimate position, all vertebrate LDH-Bs and the LDH-B-derived teleost fish

LDH-C (Quattro, Woods and Powers 1993) exhibit an extra amino acid that is missing in mammal LDH-Cs, lamprey LDH, and all vertebrate LDH-As. Styelu LDH, like that of the nematode and the less reliably alignable plant LDHs is also missing an amino acid at this position. The tunicate sequence, with its greater similarity to vertebrate sequences in this region, provides additional evidence that the absence of the amino acid is the primitive condition and that the amino acid was added in the common ancestor of LDH-Bs and teleost fish LDH-C. Interestingly, there are no reversals of this character evident in the vertebrate sequences examined to date. Summaries of the pairwise sequence identities of LDHs are shown in table 1. Styelu LDH is approximately equally similar to all vertebrate LDHs, with the highest similarities to certain LDH-As and the lowest to mammalian LDH-Cs. The lowest within-vertebrate similarity is lower than some of the similarities of vertebrate and Styela LDH. Compared to other groups outside the Vertebrata, Styela-to-vertebrate similarities are approximately the same as those between nematode and vertebrate LDH, and greater than those between plants and vertebrates, apicomplexans and vertebrates, and bacteria and vertebrates. Nonorthologous Cyanobacteria

LDHs of Apicomplexan

Protists

and

A neighbor-joining analysis of 54 LDH sequences is shown in figure 3. The tree has been rooted at the midpoint; it cannot be rooted in a manner that makes eukaryotic LDHs a monophyletic group because of the clustering of the cyanobacterial Synechocystis LDH with the Apicomplexan protist LDHs of Plasmodium and Toxoplasma and the location of this group on a branch LDH and all other bacterial uniting Bifidobacterium LDHs. The node uniting Synechocystis with Plasmodiurn and Toxoplasma is very strongly supported (100% bootstrap support). This result, along with the location of the midpoint root, suggested to us that these three LDHs might not be orthologous to the other LDHs in the tree. A search of GenBank using BLASTP revealed that the closest matches to these sequences were to the malate dehydrogenases (MDH, E.C. 1.1.1.37) of three eubacteria (Bacillus subtilis [Jin, de Jesus-Berries, and Sonenshein 19961, Bacillus sp. [Wynne et al. 19961, and Chlorojlexus aurantiacus [Synstad, Emmerhoff, and Sirevag 19961). The cytoplasmic and mitochondrial MDHs of eukaryotes and eubacterial MDHs have been shown to be related to L-LDHs in both sequence and, especially, three-dimensional structure (Birktoft et al. 1982). Sequence comparisons suggested that LDHs and MDHs represent two distinct groups whose members are more closely related among themselves than to members of the other group (Iwabe et al. 1989). Recently, an MDH from the archaebacterium HuZoarcuZa marismortui was cloned that exhibits greater sequence similarity to LDHs than to other MDHs (Cendrin et al. 1993). The MDHs of the three eubacteria mentioned above appear to exhibit a similar relationship to LDHs (Jin, de Jesus-Berrios and Sonenshein 1996; Synstad, Emmerhoff, and

Tunicate 10

Hsa Hsa Fhe Sac Pma Hsa Fhe Fhe Spl Cel Hvu Blo Taq Mhy Bst Pfa ssp

Hsa Hsa Fhe Sac Pma Hsa Fhe Fhe SPl Cel Hvu Blo Taq MhY Bst Pfa Ssp

A C A A B B C A

A C A A B B C

A

Hsa Hsa Fhe Sac Pma Hsa Fhe Fhe SPl Cel Hvu Blo Taq MhY Bst Pfa Ssp

A C A A

Hsa Hsa Fhe Sac Pma Hsa Fhe Fhe SPl Cel Hvu Blo Taq MhY Bst Pfa SSP

A C A A

B B C A

B B C

A

20

30

40

I

I

I

I

110

120

130

I

I

I

I

140

150

160

I

I

I

250

260

270

280

290

I

I

I

I

EEQTP----DDENS..... ..PVG.C... SQEPRS.... SKERDPP... ..A.VP.... SSAE.P.... SSPE.P.... DIPKSG.... PVENS..... ATP.SSAVPH TTVK...... -----..... IKLMK.... . MKNNG..... __m.****. IACQS.....

VTANSKLVII .S...RI..V .... ..V .VV .S.G....V. . ..G.R ..W ..I .VV ........ .R ..VV ..S..RI .VV .S...RI ..V I..G...C W ..K..D . ..V ICRDADM.V. DLEGARV ..V DLKDADFIV. DCRDAD..V. DL.GADV ..V A..G.DV.V.

210

220

230

240

I

I

I

I

I

VHPLSCHGWV . ..T.....I L..S.....I . .SS...... .NSA.....I I..S.....I I.AS.FN... I.SS.FN..I LS.V.V...I IA.S.....I .NAQDVQAYM .NVKNV.AYI .A.Q.V.AY. .S.N.VQAY. .A.QNV.AYI .C.RDVNAHI AC.SDINTL.

310

320

330

340

350

IKGLYGIKDV .E. :iIi::::::i V~:iiI:i.. .C........ V.DF....N. . . . . . . . . .H C..QH.VH.. .S.I...... V..M...EN. .S........ V.DMF..GE. MN.I...... A..M...G.. RH.1.. .TCF V..R...T.. S.N.FAL..N V..FH..N.. .VL AS.FH..S.G Q Tiii::. S. L.DFH..S.. EKG.FT..LF TPNE.VEE. TKNIMIAGAN LR.E..FHG. ENAILT..AY LD....ER.. .KK.LIC..L LE.Q..HS.. QS.IL.AA.Y LD.A..L...

-DVFLSVPCI .EL.. .I..V .E.. ..I..V . . . ...L..V . . . ...L..V .E....L... .E....L..V .E.Y..L..V . . .C..L..V . ..Y..L.W HE....L.AR .-ICM.. .TL .-.A..L.R. .-.NIG..W .-.YIG..AV .-1.GGT.W .-I..G...R

LGQNGISDLV ..R..V..V. ..NS.LT.VI .DNH...NI. ..SE..T.II .NAR.LTSVI .NGS.VGSV. .NSG.VGSV. .NC..VNSI. ..SA.LTHV. ..RG..LGVA .NRQ.VNNTI ..AR.VEATL ..A.. .EKII INR...RNI ..A..VEQVI ..CR.VE.IL

KVTLTSEEEA .IN.N..... HM..KP...K .MK.KPD..Q NQ..KK.... NQK.KDD.V. NM...AA.V. NM...D..V. N.D..A.... .QQ..EA.VQ EMD..EA.A. NTPVSDK.L. YPR.NE..RQ EIS.NDK.KE EIE.NDD.KN ELQ.N...-E.Q..P..K.

FIG. 2.-Alignment killifish, Sac-dogfish

I

of representatives shark, Pma-lamprey,

I

VWSGMNVAGV L...V..... . . ..V.I... ..... ..... . . ..V..... . . ..V..... . . ..A..... T il:::V::&i . . ..V..... I..SIS.G.M L.ESATIG.. . ..SAQ.G.. AY.NIKI..E . ..QAYIGVM LKRYIT.G.1 LPRYCT.S..

I

1277

I

I

I

180

190

200

--QNKITWG VGAVGMACAI SILMKDLADE LALVDVI--E DKLKGEMMDL QHGSLFLRTP ...C ...I .. T .......... ..L .......... ..A ..L .............. ..FS .S ..R..V .... ..M....S .. .V.L...C ...... ..M ....... ..A .... ..A...K .H V .M Y .H.A ::Hh:;:: ::b::::i:: :+:iR::::: .::::.v ...... :i:::::::: L:::::.K.A ..N ......... .G.S ........ ..L .................. ..Q .. ..R..V ...... E::::::; :: .LR..C ...... ..M ... .R ......... ..L...K .S ..R..V .... ..Q .... ..V T..LRE ........ ..V ... ..V ............ ..K .. ..PS.V .... ..M.....G M .VVL.G.CTD .V....V ..Q . ..Q..V ........ ..JZN I ..HG.V .... ..Q .... ..Y . ..QQN..N. .C....V ..A ........... ..LA.T.H C RRLT..S.I. A.N....I .Q T..TQN .... I....AL ..P . ..R..AL .. ..AAA..PR V -T.LA.I. A . . ..STL .F AAAQRGI.R. IV.E.IA ..K ERVEA.VL.M .. ..S.YP .V ::-V.VGI . . S.F..S.T .Y ALVLQGV.R. VV .. .LD ..R KLAQAHAE.I L.ATP.AH-. ..PI..ALI. A.N..NSFL Y AAMNQG..S. YGII.IN ..P .FAD.NAF.F EDA.AS.PF. ..GARVV.I. A.F..ASYV F ALMNQGI ... IV.I.AN ... S.AI.DA ..F N..KV.APK. . .KA..VL . . S.MI.GVM.T L.VQ.N.G-D VV.F.IV ..K NMPH.KAL.T S.TNVMAYSN ..-WQV .... A.N..RTL .Q RLVQQNV.-N VV.L.IV ..P GLPQ.IAL .. MAAQSVEEYD

FRYLMGERLG . . . . I..K.. . .H . . . .KFH . .. . . . . . . . . .. . .S . . . . . . . . .A.K.. .M.A.... ::F..ADK.. ..H.IA.K.N ..F.LS.K.N ..F.IA.H.D L.F.IAQQT. ..A.LAQH.L LQFAIAK.AK ..F.L..YFS LK.YISQK.N LKAFIAMK..

I

I

TAGARQQEG- ----ESRLNL VQRNVNIFKF IIPNVVKYSP . . . . .. . . . . . . . ..T..A. ..... A.M.1 . ..AI.H ... .................... .............. I ..... ............. DI..H .. .................... .................... .............. I ..... ..V ......V ...... .. ..V ................................ ......v ..c . ..g:::: . ..V ......... ..R .... D ........ H ...LI.RH .. ..... ..P ........ ..S. ......... H . ..QIA .... .... ..R ...... ..T..S. ..... E ... G . ..QL ..... ..ALYR K .V.P.AEH .. .... ..IP ..... ..T .... L . ..P..KP .... ..Q...E. .iT .. .L.A .M..L..VA. A..VA.RP ..... ..T..Q. LD..AQV.A D VV.KIL.AA. . ..RP.KP ..... ..T..E. .AD.IR.IRE .ALK.KESGF C...N.KP ..... ..T..D. .DK.IA..R S .VES.MASGF . ..FTKAP .K SDKEWN.DD. LPL.NK.MIE .GGHIK.NC. . ..LPRRP .... ..M..DD. LGK.A..VA Q GAREALR ...

KIVS-GKDYN .VT......S ... A.D ...S ......... S ...A.D . ..S .A.D ..S :;I.A.D.. ..A .A.D ..S :~NG.D .. . .S TVKA.DT ..S R.S-..T .AA S.DG.SD.PE VW.R.SGW.E IS .. .RYE.K VDIW.HG ..D CK..GSNT .D SKIIGTNE.E

LGEHGDSSVP I......... V........A I......... I......... . .. . . . . ..A . . . . ..T... . . . . ..T... I......... I........A V....... .A A......E.. V......E.L M...... .FV I.....TEL. V.A..NKM.L ..G...LML.

80

I

_-____---_ -MATLK---- -DQLIYNLLK .......... ..S.V ..... .E.. .EK.IE .......... .-MSTQ .... .EK..SHVM. .................... ..K..GH.AT ST .GK..HEMVP :::::::::: 1:: ............ .EK..APVAE .......... ..SSVL .... .QK..TP.AS ........... ..SVL .... .HK..TP.AC ........... ..D..GFD I MSE.FTEVCP .......... MAS.I ..... .-EVFAEIAA MHKASSLSEL GFDAGGASSG FFRPVADGCP .......... .----- .... .------MAE .......... .----- .... .--------.......... .----- .... .MKKFNKKGE .......... .----- .... .------MNA .......... .----- .... .__--____.......... .----- .... .-MN.LEYAP

LRRVHPVSTM

70

LDH

SLKTLHPDLG A....D.K.. ..Q..N.NM. . ..E...E.. G.QS.N..I. ..QE.N.EM. ..QK.N.EI. N.Q..N.NI. C.NSI..RI. T.HEIK..I. PAFKSLR.SH PMCDWT.LP. D.EAFAQARCFCAYSKLTPIRK.VESKP.QEFINN-PITE.I.P--

I

170

I

I

I

NCKLLIVSNP VDILTYVAWK ISGFPKNRVI GSGCNLDSAR D..I.V ........ ..IV ... ..L.VT ............. ..I..V .............. L....RH .... ..T ...... D.1I.V .... ..V ....... L..L.MH.I. .......... ..I..V .............. L..L..H ............. D.IIIV ......... ..T .. L..L..H ............. ..TI.V .... ..V....T .. L..L..H .... ..T ...... D.IIIV .... ..V....T .. L..L.MH .... ..T ...... SA1.V ..... ..LM ...... L.N..R ..... ..T ...... DTCI.V .... ..V....T .. L..L.RE ..F ...T ...... DAL..V ..V ....... L....AS ..T....S. .AIYMLIT .........A.H..Q. LT.L.E.QIF ..... ..T ...... EAV...AT .. ..VM.Q..Y R L..L.PE ..V . ..TI..T .. SGISI..A ... ..I.RAYR D A...SDQK ... ..TV..T .. QGLF.VAT ...... ..AT .. F..L.HE .... ..TI..T .. .AFIIV.T .. ..VMVQLLH Q H..V...KI. .L.GV..TS. .AI.IV.T .. L.VM..L ... VT.L.SQ ..M .MAGV .....

TDKDKEQWK..S...H... A.G.S.N... . . . . ..N... .P..G.D... ..N.S.N... ..G....... ..F.E.N... YPDGP.G.D. EKT.N.H.EA RSF.EAALE. H.PLDADKRE GRALTPDDRL GIDSS-NYEK GEEAQ-KDLE KL-ISDAELE ------.TIE

EVHKQVVESA NI....IQ.. .L..K..DG. KL..D..D.. S......D.. . ...M..... AT..A..D.. .T..M..D.. KI.....DG. .I..K..D.. GIRRA..GG. .I.QE.KNA. RIDEG.RRA. .LEYP.SRR. RIFVN.RDA. AIFDRT.NT. .LVERTRNGG

360

370

I

I

YNIKLK--G ..I....... . . . .. . . . . . .. . .. . . . . . .. . . . . .. . . . . . . . . . .. . . . . .. . . . . . . . . .. . . . . . .D..R..... ..I....... . ...G..... .KI.NG.... .RI.EG.... ..I.NR...R .QI.EK.... L.IVN.H..A A.IAA.LQT.

YTSWAIGLSV . . .. .. . . . . . . . . . ..M.. . . . . . . . . .. . . . . . .. . . . ..N....... ..N....F.. ..N....... ..N......C .. ........ . . . . . ..Y.. A.NY...M.G A.YYG..AGL A.FYG..AAI A.YYG.AMGL SPYV.PAAAI TAYY.PAS.A

300

I

ADLAESIMKN M..VG..L.. . ..V...V.. .. . ..T.... . .. ..T.L.. . ..I..ML.. . ..T...V.. . ..T..L.R. .E.LAT.LHH .KI.QG.FS. .S..A.LLRD V.II.AVLHD .R.TRA.LTD .KIVSN.I.D .RVTRA.LH. IEM...YL.D .VMV...LR.

I _--

RLKKSADTLW GIQKELQ-FLF....E... N...D.I... Q.V...E... .V....T.L. Q.Q...T... D...D.K... QVQ...E... NV....T... Q......... D...D.KDL. Q......... . . ..D.KDL. Q.QG..S... D...D.RDI. MI....M.IA DV..G.K.W. K.HN..KA.L N.NG1V.M. Q.RR..K... ENCQL.D.L. A..R..E..K ETAAQFG... A.RR..EI.K .AASA.G... KFA..VAIID K.YQDAIKNI .FHH..A..K SVLARAFTR. --.AKF.EiAI AET.RMKALA A.HL..FiAVR LNIDVALAMV

... ... ... ...

of major groups Spl-tunicate,

of LDH sequences. Abbreviations of taxa are as follows: Hsa-human, FheCel-nematode, Hvu-barley, Blo--B$dobacterium Zongum, Taq-Z’hermus aquaticus, Mhy-Mycoplasma hyopneumoniae, Bst-Bacillus stearothermophilus, Pfa-Plasmodium falciparum, Ssp-Synechocystis sp. A, B, and C refer to the different isozymes within a species. Dashes indicate alignment gaps and periods indicate matches to the sequence of human LDH-A.

Sirevgg 1996; Wynne et al. 1996). In addition, all of these LDH-like MDHs are tetrameric like LDH, rather than dimeric like other MDHs. Although the MDH activity of the cloned sequences from Haloarculu, BacilZus, and Chlorojkxus has been relatively well charac-

terized (Jin, de Jesus-Berries and Sonenshein 1996; Synstad, Emmerhoff, and Sirevag 1996; Wynne et al. 1996), the LDH activity of Plasmodium LDH has not been as conclusively demonstrated (Bzik, Fox, and Gonyer 1993), and the Toxoplasma and Synechocystis sequences

1278

Stock et al.

Table 1 Summary

of Pairwise Differences Pairwise

Among LDH Sequences

Comparisons

Range of Distance@

Bacteria to vertebrate .................. Apicomplexa to vertebrate. ............. Plant to vertebrate. .................... Nematode to vertebrate ................ Tunicate to vertebrate. ................. Within vertebrates. .................... Tunicate to vertebrate LDH-A. .......... Tunicate to lamprey ................... Tunicate to vertebrate LDH-B ........... Tunicate to mammal LDH-C ............ Tunicate to killifish LDH-C .............

0.59-0.73 0.68-0.73 0.46-0.54 0.35-0.46 0.36-0.44 SO.40 0.35-0.39 0.36 0.37-0.40 0.39-0.44 0.38

a Distances were calculated using the alignment in figure 2, with gaps in the pairwise comparisons deleted and all other positions considered. Distances are proportions of nonidentical residues.

have been identified as LDHs only by sequence similarity to Plasmodium LDH (Kaneko et al. 1995; Yang and Parmley 1995). Regardless of whether the “LDHs” of Plasmodium, Toxoplasma, and Synechocystis actually function as LDHs or MDHs, we suggest that they are not orthologous to the other LDHs in figure 3. Strong support for this hypothesis is provided by the presence LDH (Fig. 3) in addition in B. subtilis of a conventional to the LDH-like MDH described above. Phylogeny

of Eukaryotic

LDHs

The midpoint rooting of figure 3, which supports a monophyletic group of nonprotistan eukaryotic LDHs distinct from bacterial LDHs, is consistent with the results of simultaneous analysis of LDH and MDH (Iwabe et al. 1989). This monophyletic group of plant and animal LDHs was found in 100% of the bootstrap NJ trees. Plant LDHs also form a monophyletic group with 100% bootstrap support, although this is not surprising given that all of the species are members of the family Poaceae. The monophyly of animal LDHs was found in 99% of the bootstrap NJ tree. All of the above clusters were also found in parsimony analyses (not shown). Phylogenetic

Position

of Styela LDH

The phylogeny of the LDHs of the three animal phyla matches that of the organisms. In the tree in figure 3, nematode LDH clusters outside of urochordate (Styela) and chordate LDHs, but with only moderate bootstrap support (60%). Figure 3 shows significant support (97% of the bootstrap trees) for Styelu LDH being outside of a monophyletic group of all of the vertebrate LDHs. This result indicates that all of the gene duplications within the vertebrates examined occurred after the divergence of tunicates and vertebrates. In an attempt to provide increased resolution of the relationships of tunicate and vertebrate LDHs, we conducted a neighbor-joining analysis without the more distantly related and less reliably alignable bacterial and protistan LDHs and have rooted the resulting tree between plant and animal LDHs (fig. 4). Such an analysis shows higher support for the theory that the tunicate LDH is more closely related to vertebrate LDH than is nematode LDH (92% bootstrap support), but lower support for the

monophyly of vertebrate LDHs (87% of the bootstrap trees). We used maximum-parsimony analyses to evaluate the relative support for alternative hypotheses on the phylogenetic relationships of tunicate LDH (fig. 4 and table 2). It has been suggested based on immunochemical comparisons that tunicate LDH is most closely related to LDH-C and lamprey LDH (Baldwin, Mortimer, and Patak 1988). These authors assumed that mammal and fish LDH-Cs were orthologous, although this relationship was strongly contradicted by a subsequent phylogenetic analysis of a teleost LDH-C sequence (Quattro, Woods and Powers 1993). Not surprisingly, given this latter result, the forced grouping of Styela LDH, lamprey LDH, and mammal and fish LDH-Cs required a large number of extra steps (31 for equal weighting, 46 for PROTPARS weighting) relative to the most parsimonious trees and in the case of PROTPARS weighting was significantly worse as judged by the method of Templeton (1983). Relaxing the requirement that mammal LDH-C be a part of this cluster (Baldwin, Mortimer, and Patak’s [1988] immunochemical data were based on fish LDH-C) still results in trees significantly worse than unconstrained trees. Clustering of tunicate LDH with mammalian LDH-Cs could also be rejected, but a specific relationship with fish LDH-C could not. Additional phylogenetic positions of the tunicate LDH inside the vertebrate LDH clade (not necessarily suggested by any authors) were also evaluated with respect to the number of extra steps. The clustering of the tunicate LDH with gnathostome LDH-B and teleost LDH-C required the Styela fewest additional steps. Hypotheses clustering LDH with LDH-C remained among the least likely. Phylogeny

of the Vertebrate

LDHs

The phylogenetic relationships among vertebrate LDHs were investigated with plant and invertebrate outgroups as shown in figure 4. Parsimony and neighborjoining analyses gave similar topologies, although the bootstrap values were generally higher with neighborjoining. The values discussed below are from this latter analysis. A branch uniting all of the jawed vertebrate (gnathostome) LDH-Bs as well as teleost fish (killifish) LDH-C, Xenopus LDH-A, and Xenopus LDH-C was found in 90% of the bootstrap trees. The tree topology supports a number of gene duplications within this clade. Ninety-three percent of the bootstrap trees support a derivation of teleost LDH-C from teleost LDH-B after this group diverged from tetrapods, as initially proposed on immunochemical grounds (Whitt 1969; Marker-t, Shaklee, and Whitt 1975) and subsequently supported at the protein sequence level (Quattro, Woods, and Powers 1993). Two duplications are also indicated (with 100% bootstrap support) in the Xenopus lineage after it diverged from amniotes. One is likely to be the result of tetraploidy (Kobel and Du Pasquier 1986), while the other may be a gene duplication postulated by Wolff and Kobel (1982). Stock and Powers ( 1995) concluded from similar results that LDH-A and LDH-C have been incorrectly designated in Xenopus, with LDH-A probably remaining to be cloned. Aside from these likely dupli-

Tunicate

LDH

1279

_c Cattle LDH-A

45 51

Pig LDH-A i Rabbit LDH-A loo_ Mouse LDH-A 77_ ' Rat LDH-A - Human LDH-A 81 Chicken LDH-A Fox LDH-C Human LDH-C ‘M-“i;wL;FC -

Dogfish shark LDH-A Lamprey LDH Kiilifish LDH-A Killifish LDH-B Killifish LDH-C

,

100 58

Tunicate LDH Nematode LDH

100

100

(

8. longum LDH T. aquaticus LDH T. caldophilus LDH D. radiodurans LDH T. maritima LDH M. hyopneumoniae LDH M. genitalium LDH

87

P. acidilactici LDH 95

S. mutans LDH

FIG. 3.-Neighbor-joining tree of 54 LDH sequences. The tree is rooted at the midpoint of the longest path length. Amino acid distances were calculated with a gamma correction (a = 2). Alignment positions l-44 were omitted from consideration, and the pairwise deletion option of MEGA was used to treat alignment gaps. Numbers at nodes indicate the results of a bootstrap analysis with 1,000 replications. Bootstrap support within mammal LDH isozymes has been omitted for clarity. Numbers in a larger font indicate the support for the interrelationships among the major groups of eukaryotic LDHs.

cations, the phylogeny of the LDH-B genes matches the well-accepted organismal phylogeny of (teleost fish, (frogs, (birds, (mammals)))). Although LDH-Bs and their derivatives form a well-supported clade in figure 4, the sister group of this clade (lamprey LDH + gnathostome LDH-A + mammal LDH-C) is found in only 47% of the bootstrap trees. Within this latter group, there is strong support, however, for the theory that mammalian LDH-Cs are derived from within LDH-A. The strongest contradiction of the proposals that LDH-C is the earliest divergence among vertebrate LDHs (e.g., Li et al. 1983; Baldwin, Mortimer, and Patak 1988; Crawford, Constantino, and Powers 1989; Baldwin, Patak, and Mortimer 1992; Tsoi and Li 1994; Tsuji et al. 1994; Mannen et al. 1996) is pro-

vided by the node uniting dogfish shark LDH-A with mammal LDH-C and tetrapod LDH-A (96% bootstrap support). Further support for a derivation of mammalian LDH-Cs from LDH-A is provided by the node uniting them with chicken and mammal LDH-As (80% of the bootstrap trees). Mammalian LDH-Cs are specifically placed as the sister group of bird and mammal LDH-As (77% bootstrap support), a position that requires gene losses in a number of amniote groups, as discussed by Matson (1989). This grouping may be due to a rapid rate of evolution of LDH-C, which is obvious from inspection of the within-mammal divergences relative to those of LDH-A and LDH-B (fig. 4; Whitt 1984). A clustering of mammal LDH-C and mammal LDH-A, which does not require additional gene losses, is found

1280

Stock et al. Cattle LDH-A

and 48% bootstrap support) and one strongly supported one (96% bootstrap support). The additional gene duplications and losses required by the tree in figure 4 over a tree matching the organismic phylogeny have been discussed by Tsuji et al. (1994), Quattro et al. (1993, and Stock and Powers (1995).

Pig LDH-A

IL

rt 99

Rabbit LDH-A

Mouse LDH-A

60

Rat LDH-A

77 27

Human LDH-A

1

8C) I

-

Chicken LDH-A

I

,

I-

Fox LDH-C

-

Human LDH-C

97

Mouse LDH-C I

Rat LDH-C

Dogfish shark LDH-A I-

Killifish LDH-A

Xenopus LDH-A

88 60

~ 96

Duck LDH-B Chicken LDH-B

91

L Mouse LDH-B Tunicate LDH Nematode LDH

FIG. 4.-Neighbor-joining and parsimony analyses of 24 vertebrate LDH sequences with plant and invertebrate outgroups. The topology indicated was obtained from a neighbor-joining analysis conducted as described in figure 3. Numbers above the nodes indicate the results of a bootstrap neighbor-joining analysis with 1,000 replications. Numbers below the branches indicate the support obtained in a bootstrap parsimony analysis. Some sequences with close relatives were deleted from the parsimony analyses to reduce the amount of time required; these were cattle LDH-A, pig LDH-A, rabbit LDH-A, rat LDH-A, Xenopus LDH-B, duck LDH-B, pig LDH-B, and barley LDHB. Three most-parsimonious trees were found with 100 replications of a random addition sequence and TBR branch swapping (Swofford 1991). These differed from the topology shown by placing mammalian LDH-Cs within or as the sister group of mammalian LDH-As (the only aspect of the three trees in which they differed among each other) and by grouping killifish LDH-A with lamprey LDH. The bootstrap parsimony results indicated were obtained with 100 bootstrap replications, each with 10 replications of a random addition sequence and TBR branch swapping.

in the most parsimonious trees and in 48% of the bootstrap trees from a maximum-parsimony analysis (not shown). The phylogeny of gnathostome LDH-A and lamprey LDH (suggested to be orthologous to LDH-A by Stock and Whitt [1992]) in figure 4 does not match the generally accepted phylogeny of the organisms; moving the LDH-A of the killifish to a sister group relationship with tetrapod LDH-A would bring the two phylogenies into agreement. Such a position of the killifish LDH-A is contradicted by two weakly supported nodes (47%

Discussion Gene Duplications Leading to Multiple Vertebrate LDHs Occurred After the Divergence of Tunicates Vertebrates

and

Our analyses strongly support a sister group relationship between the LDH of Styela plicata and a clade consisting of all of the vertebrate LDHs that have been sequenced to date (those of lamprey and jawed vertebrates). If it is assumed that the locus we sequenced is the only one encoding LDH in the genome of Styela (see below), then this phylogenetic result suggests that all of the LDH duplications represented by the present collection of vertebrate sequences occurred after the divergence of vertebrates and tunicates. This pattern of duplication is consistent with the hypotheses of Markert, Shaklee, and Whitt (1975) and Fisher et al. (1980), which were formulated largely in the absence of sequence data. Our phylogenetic results strongly contradict the hypothesis of Baldwin, Mortimer, and Patak (1988) that tunicate LDH is more closely related to the LDH of lampreys and even more so to the LDH-C of mammals and teleost fish than it is to the LDH-A and LDH-B of jawed vertebrates. This hypothesis was based on immunochemical comparisons of the LDH of the tunicate Pyura stolonifera (distantly related to S. plicata) with the LDH-A, LDH-B, and LDH-C of a teleost fish, as well as on comparisons of amino acid compositions of tunicate and other vertebrate LDH isozymes. The lack of agreement of our phylogenetic conclusions based on sequence comparisons with earlier conclusions based on cruder immunochemical and amino acid composition analyses suggests that hypotheses on the relationships of echinoderm LDH (Baldwin, Patak, and Mortimer 1992) and hagfish LDHs (Baldwin, Mortimer, and Patak 1988 and included references) to the LDHs of lamprey and jawed vertebrates that are based on similar evidence are no longer persuasive. LDH Locus Number Invertebrate Groups

in Tunicates

and Other

Our PCR amplifications of Styela L.dh did not provide any evidence for the existence of more than one locus in this species. A number of the primers we used are coniplementary to regions with amino acid sequences conserved among multiple vertebrate paralogs. This suggests that if the Styela locus we amplified were orthologous to a specific type of vertebrate LDH, then the primers would be capable of amplifying Styela genes orthologous to other vertebrate genes. However, amplification of additional loci may have failed for a number of reasons, including restricted expression or unique substitutions along the Styela lineage. A more systematic determination of the number of loci present is pos-

Tunicate

Table 2 Lengths of Constrained

Trees in Parsimony

128 1

Analyses PROTPARSC

EQUAL WEIGHTS~ CONSTRAINTS Nonee..................................................

(tunicate, (tunicate, (tunicate, (tunicate, (tunicate, (tunicate, (tunicate, (tunicate,

LDH

killifish C, mammal C, lamprey) . . . . . . . . . . . . . . . . . . . killifish C, lamprey). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . killifish C) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . mammal C). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . all A except Xenopus, lamprey). . . . . . . . . . . . . . . . . . . . all A except Xenopus, mammal C, lamprey) . . . . . . . . . all B, all Xenopus). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . all B, all Xenopus, killifish C) . . . . . . . . . . . . . . . . . . . . .

Length

Extra Steps

Length

928

0

1,103

959 948 936 935 936 933 939 931

31 20 8 7 8 5 11 3

1,149 1,128 1,119 1,120 1,118 1,113 1,122 1,111

Extra Stepsd

0 46” 25” 16 17” 15 10 19 8

a Constraints consisted of requiring the indicated group of sequences to be monophyletic with no other relationships specified inside or outside this group. See figure 4 for the sequences included in the analyses. b Lengths of trees calculated with all amino acid transformations weighted equally. c Lengths of trees calculated with amino acid transformations weighted according to the number of nonsynonymous nucleotide substitutions required. d Asterisks indicate topologies judged to be significantly worse than the most parsimonious tree by the test of Templeton (1983) as implemented in PROTPARS (Felsenstein 1993). ’ eUnconstrained trees placed the tunicate outside of all vertebrate LDHs.

sible with the characterization of LDH enzymatic activity; such investigations should detect additional loci that have diverged extensively in sequence but not in function. The most thorough study of this type in tunicates is that of Baldwin, Mortimer, and Patak (1988) on P. stoloniferu. Nondenaturing electrophoresis and histochemical staining of crude extracts of whole animals and a number of tissues revealed evidence for a single locus, as did analyses of purified LDH activity. Similar electrophoretic investigations of LDH activity suggest that echinoderms (Baldwin, Patak, and Mortimer 1992) and amphioxus (Fisher et al. 1980) also possess a single LDH locus. Assuming the generally accepted phylogeny (Peterson 1995) of (echinoderms, (tunicates, (amphioxus, (vertebrates)))), the simplest explanation of these locus numbers and our phylogenetic results is that the ancestral state of deuterostomes (a single LDH locus) has been retained by echinoderms, tunicates, and amphioxus, while LDH duplications have occurred within vertebrates after their divergence from amphioxus. This hypothesis was proposed previously by Fisher et al. (1980). ’ If, in spite of evidence to the contrary, an additional tunicate LDH locus did exist, the implications for LDH evolution in vertebrates would depend on the phylogenetic position of the new locus in the LDH tree. Clustering with the existing tunicate locus would imply a duplication in tunicates after their divergence from vertebrates and would not alter hypotheses of the timing of vertebrate gene duplications. Because of the phylogenetic position of the known tunicate locus, any other position of the new locus would require that one of the two types of loci was lost in all vertebrates. Placement of the new locus within the clade consisting of vertebrate LDHs would require the loss of at least one type of locus in the tunicate as well. In other words, even if an additional tunicate locus is discovered, it will not be possible to postulate that particular vertebrate duplications occurred earlier than the divergence of tunicates and vertebrates without also postulating a more complex pattern of gains and losses.

The Location

of the Root in the Tree of LDHs

We described above the evidence that the LDHs of Synechocystis, Plasmodium, and Toxoplasma are not orthologous to the other LDHs in the tree and may in fact be more closely related to tetrameric MDHs. With the exception of a weakly supported node placing Bijidobacterium LDH closer to plant and animal LDHs than are other bacterial LDHs, the tree in figure 3 (rooted at the midpoint) depicts a phylogeny of the remaining LDHs corresponding to the organismal phylogeny of (bacteria, (plants, (nematodes, (tunicates, (vertebrates))))). This provides support for a root separating monophyletic groups of bacterial and eukaryotic LDHs (joining an internal branch with 100% bootstrap support in fig. 3), as any other root position would require a greater number of gene duplication and loss events. The clustering of bacterial LDH outside of vertebrate LDHs has also been obtained in trees of LDHs rooted with MDH sequences (Iwabe et al. 1989; Cendrin et al. 1993; Wynne et al. 1996). LDH Evolution

from Bacteria

to Vertebrates

Determining whether large numbers of LDH gene duplications and losses have occurred outside of chordates requires more extensive collection of LDH sequences. Our phylogenetic results obtained with LDHs from a relatively small number of nonchordate taxa imply an evolutionary pattern of a single locus persisting in most lineages with occasional (and usually recent) gene duplications. Examples of gene duplications in nonvertebrate taxa include the appearance of LDH-A and LDH-B of Bacillus psychrosaccharolyticus after its divergence from the other Bacillus species in the tree and the LDH-A and LDH-B of barley, which resulted from a duplication after the divergence of barley from the other two members of the family Poaceae (fig. 3). Although the number of LDH loci in invertebrates has not been well characterized, Drosophila melanogaster is believed to have a single locus for L-LDH (Alahiotis et al. 1983), and the evidence for a single locus in an echinoderm (Baldwin, Patak, and Mortimer 1992) has been

1282

Stock et al.

discussed above. A number of insects and molluscs are believed to lack LDH activity altogether, which would be loss events according to our phylogenetic trees (e.g., fig. 3). However, the biggest remaining question in the evolution of invertebrate LDHs relates to the orthology of the loci encoding the proteins or enzyme activities that have been described (Long 1976; Siebenaller et al. 1983). All of the LDH sequences that we have analyzed and all known chordate LDHs are specific for the Lstereoisomer of lactate (Whitt 1984). A number of groups of invertebrates have D-LDHs, however, and the presence of D- or L-LDH appears to be mutually exclusive for any given taxon (Long 1976). Sequence analyses have shown that bacterial D-LDHs form a separate family from L-LDHs that is much more distantly related to the latter than are MDHs (Kochhar et al. 1992). No invertebrate D-LDH sequences are available to allow a determination of their relationship to L-LDH, although Siebenaller et al. (1983) suggested that the two types of LDH were not orthologous based on the sequence of a few short peptide fragments from a D-LDH of a horseshoe crab. If this hypothesis is correct, then L-LDH would have to have been lost and replaced by D-LDH at least four times in invertebrates (Long 1976). Relationships

Within

Vertebrate

LDHs

The relationships of vertebrate LDHs that we obtained in our analyses are generally consistent with those of Stock and Whitt (1992), Quattro, Woods, and Powers (1993), Quattro et al. (1995), and Stock and Powers (1995). There was strong support for the derivation of teleost LDH-C from LDH-B after the divergence of teleosts and tetrapods and a derivation of mammalian LDH-C from LDH-A. This contradicts other sequence analyses that suggested that LDH-C represents the earliest divergence among vertebrate LDHs (Li et al. 1983; Crawford, Constantino, and Powers 1989; Tsoi and Li 1994; Tsuji et al. 1994; Mannen et al. 1996). The results of these analyses differed from ours largely in the location of the root of the vertebrate LDH tree. We conducted a series of phylogenetic analyses with varying subsets of outgroup sequences to the vertebrate LDHs (not shown) and obtained a root in or between mammalian LDH-Cs and other LDHs only with highly divergent sequences as outgroups. The analyses of Tsoi and Li (1994) and Mannen et al. (1996), in which the position of LDH-C differed from ours with respect to topological relationships as well, were conducted with the unweighted pair-group with arithmetic mean (UPGMA) method, one that has been shown to be highly sensitive to unequal rates of evolution (Huelsenbeck and Hillis 1993). Our analyses also agreed with previous studies in providing strong support for recent duplications of LDH-B in Xenopus and the monophyly of a group consisting of LDH-B and its derivatives in teleosts and Xenopus. Although the lamprey LDH clusters with LDH-As as in the analyses of Stock and Whitt (1992), this result was not strongly supported, and at least one node strongly contradicted the orthology of killifish LDH-A with tetrapod and shark LDH-As. Such a result was also ob-

tained in other studies that included teleost LDH-A sequences (Tsoi and Li 1994; Tsuji et al. 1994; Quattro et al 1995). This phylogenetic result cannot be brought in line with the orthology of LDH-As by rerooting the tree. Whether or not it is artifactual may be addressed by analyses of the LDH-As of other groups of primitive bony fishes. The interrelationships of four major lineages of LDHs (shark LDH-A + tetrapod LDH-A + mammal LDH-C; teleost LDH-A, lamprey LDH, and jawed vertebrate LDH-Bs and derivatives) were not well resolved in our analyses, as indicated by low bootstrap support. The persistence of at least some of these low bootstrap values in the absence of an outgroup (analysis not shown) suggests that the sequence of more closely related outgroups (amphioxus is the only candidate invertebrate) will not by itself resolve all aspects of the phylogeny of vertebrate LDHs. The collection of additional vertebrate sequences may decrease or eliminate support for potentially artifactual groupings but is unlikely to result in their replacement with strongly supported alternatives. The resolution of the phylogeny of vertebrate LDHs may require analyses more sophisticated than the rather simple methods we have used, such as the assignment of different expected rate categories to different sites and the consideration of structural or functional information in weighting changes. Gene Duplications

in Early Vertebrate

Evolution

Regardless of the exact pattern of gene duplication leading to multiple LDH loci in vertebrates, our analyses suggest that all duplications occurred after the divergence of tunicates and vertebrates and that at least some of them occurred early in the diversification of vertebrates. This evolutionary pattern is the same as that suggested by Markert, Shaklee, and Whitt (1975) and Fisher et al. (1980). In the latter study, it was specifically proposed that the earliest LDH duplications occurred in jawless vertebrates after the divergence of vertebrates from cephalochordates (amphioxus). Ohno (1970) proposed that extensive gene duplications had occurred at two or more points in chordate phylogeny, based largely on evidence from genome size estimates. The first of these episodes of large-scale gene duplication was proposed to have occurred in the lineage leading to amphioxus and vertebrates after the divergence of tunicates, and the second in the fish or amphibian ancestors of amniotes. Ohno suggested that the latter duplication was due to polyploidy but that the first may have been due either to polyploidization or extensive tandem duplication. Fisher et al. (1980) interpreted the evolution of four isozyme systems (including LDH) as indicating one or more rounds of polyplodization in the early vertebrate or prevertebrate lineage. Holland et al. (1994) and Sharman and Holland (1996) surveyed the phylogenetic relationships and number of loci encoding genes involved in the regulation of development and also concluded that widespread gene duplications occurred at two periods in chordate evolution. These authors proposed a slightly different timing for the duplication episodes than Ohno (1970), however. The first phase was

Tunicate LDH

proposed to have occurred close to the origin of vertebrates after the divergence of this lineage from amphioxus, and the second near the origin of gnathostomes. Our results indicating that the duplication of LDH loci in vertebrates occurred after the divergence of tunicates is consistent with both the hypotheses of Ohno (1970) and of Holland et al. (1994) and Sharman and Holland (1996). A more precise understanding of the timing of the duplications leading to multiple LDH loci and their relationship to the duplications of other genes awaits the sequencing of the LDHs of amphioxus and hagfish. That the duplications leading to LDH-A and LDHB involved large genomic regions is suggested by the linkage relationship of these two genes in mammals; mouse and human Ldh-Bs are located on chromsomes 6 and 12p, respectively, and mouse and human Ldh-As are located on chromosomes 7 and llp (Debry and Seldin 1996). In both species, the chromosome segments containing Ldh-A and Ldh-B contain additional loci with paralogous relationships between the two regions, and it has been proposed that these chromosomal segments are ancestrally related, having arisen by a process such as polyploidization (Lundin 1993). A relatively recent origin of mammalian Ldh-C from Ldh-A, as we have suggested based on phylogenetic analyses of LDH genes, is consistent with the close linkage of these two loci (perhaps within 100 kb-Stubbs et al. 1994). Ldh-C and Ldh-A have also been shown to be linked in some fish species (Morizot 1990) despite evidence from sequence analyses that the former gene is closely related to Ldh-B. This latter result suggests that linkage arrangements in modern species may bear a complex relationship to gene duplication histories. Nevertheless, it would be interesting to compare phylogenies of chordate gene families containing paralogs linked to mammalian or fish LDH loci with the duplication history that we have proposed for the LDH gene family. Phylogenetic congruence between LDH and other linked gene families would provide additional evidence for our hypothesis.

BALDWIN, J., A. PATAK,

(Holothuria

1283

and K. MORTIMER.1992. Echinoderm

atra) lactate dehydrogenase: affinities with the

putative ancestral chordate enzyme. Biochem. Syst. Ecol. 20535-539. BIRKTOFT,J. J., R. T FERNLEY,R. A. BRADSHAW,and L. J. BANASZAK.1982. Amino acid sequence homology among the 2-hydroxy acid dehydrogenases: mitochondrial and cytoplasmic malate dehydrogenases form a homologous system with lactate dehydrogenase. Proc. Natl. Acad. Sci. USA 79:6166-6170. BJOURSON,A. J., and J. E. COOPER. 1992. Band-stab PCR: a simple technique for the purification of individual PCR products. Nucleic Acids Res. 20:4675. BZIK, D. J., B. A. Fox, and K. GONYER.1993. Expression of Plasmodium falciparum lactate dehydrogenase in Escherichia coli. Mol. Biochem. Parasitol. 59: 155-166. CENDRIN,E, J. CHROBOCZEK,G. ZACCAI, H. EISENBERG,and M. MEVARECH.1993. Cloning, sequencing, and expression in Escherichia coli of the gene coding for malate dehydrogenase of the extremely halophilic archaebacterium HaZoarcula marismortui. Biochemistry 32:4308-43 13. CRAWFORD,D. L., H. R. CONSTANTINO,and D. A. POWERS. 1989. Lactate dehydrogenase-B cDNA from the teleost Fundulus heteroclitus: evolutionary implications. Mol. Biol. Evol. 6:369-383. DEBRY,R. W., and M. E SELDIN.1996. Human/mouse homology relationships. Genomics 33:337-35 1. FELSENSTEIN,J. 1993. PHYLIP phylogeny inference package. Version 3.5~. University of Washington, Seattle. FISHER, S. E., J. B. SHAKLEE,S. D. FERRIS,and G. S. WHIT-T. 1980. Evolution of five multilocus isozyme systems in the chordates. Genetica 52/53:73-85. FROHMAN,M. A. 1990. RACE: rapid amplification of complementary DNA ends. Pp. 28-38 in M. A. INNIS,D. H. GELFAND,J. J. SNINSKY,and T J. WHITE, eds. PCR protocols: a guide to methods and applications. Academic Press, New York. HENDRIKS,W., J. W. M. MULDERS,M. A. BIBBY,C. SLINGSBY, H. BLOEMENDAL, and W. W. DE JONG. 1988. Duck lens Ecrystallin and lactate dehydrogenase B4 are identical: a single-copy gene product with two distinct functions. Proc. Natl. Acad. Sci. USA 85:7114-7118. HOLLAND,I? W. H. 1992. Homeobox genes in vertebrate evolution. BioEssays 14:267-273. Acknowledgments HOLLAND,I? W. H., J. GARCIA-FERNANDEZ, N. A. WILLIAMS, and A. SIDOW. 1994. Gene duplications and the origins of We wish to thank Patricia Schulte for the gift of vertebrate development. Development (Suppl.): 125-l 33. competent bacterial cells. This study was supported by HUELSENBECK, J. P, and D. M. HILLIS. 1993. Success of phyNSF grants BSR-87-17417 (to G.S.W.), BSR-87-18425 logenetic methods in the four-taxon case. Syst. Biol. 42: (to D.A.P.), BSR-88-15362 (to G.S.W. and D.W.S.), 247-264 BSR-90-22648 (to D.A.P.), and DEB 96-29572 (to IWABE, N., K.-I. KUMA, M. HASEGAWA,S. OSAWA, and T. D.A.P.), an NSF Predoctoral Fellowship in Genetics (to MIYATA. 1989. Evolutionary relationship of archaebacteria, D.W.S.), and an Alfred F? Sloan Postdoctoral Fellowship eubacteria, and eukaryotes inferred from phylogenetic trees of duplicated genes. Proc. Natl. Acad. Sci. USA 86:9355in Studies of Molecular Evolution (to J.M.Q.). 9359. JIN, S., M. DE JESCJS-BERR~OS, and A. L. SONENSHEIN. 1996. LITERATURE CITED A Bacillus subtilis malate dehydrogenase gene. J. Bacterial. 178:560-563. ABAD-ZAPATERO, C., J. I? GRIFFITH,J. L. SUSSMANN,and M. KANEKO,T., A. TANAKA,S. SATO, H. KOTANI,T. SAZUKA,N. G. ROSSMANN.1987. Refined crystal structure of dogfish MIYAJIMA,M. SUGIURA,and S. TABATA. 1995. Sequence M4 apo-lactate dehydrogenase. J. Mol. Biol. 198:445-467. analysis of the genome of the unicellular cyanobacterium ALAHIOTIS,S. N., A. ONOUFRIOU, M. FOTAKI, and M. PELESynechocystis sp. strain PCC6803. I. Sequence features in CANOS. 1983. Drosophila lactate dehydrogenase: developthe 1 Mb region from map positions 64% to 92% of the mental aspects. Biochem. Genet. 21:199-211. genome. DNA Res. 2:153-166. BALDWIN,J., K. MORTIMER,and A. PATAK. 1988. Do ascidians KOBEL, H. R. and L. Du PASQUIER.1986. Genetics of polypossess the ancestral subunit type of vertebrate lactate dehydrogenase? J. Exp. Zool. 246: 109-l 14. ploid Xenopus. Trends Genet. 2: 3 10-3 15.

1284

Stock et al.

KOCHHAR, S., I? E. HUNZIKER, I? LEONG-MORGENTHALER,and H. HJOTTINGER. 1992. Evolutionary relationship of NAD+dependent D-lactate dehydrogenase: comparison of primary structure of 2-hydroxy acid dehydrogenases. Biochem. Biophys. Res. Commun. 184:60-66. KOZAK, M. 1996. Interpreting cDNA sequences: some insights from studies on translation. Mamm. Genome 7563-574. KUMAR, S., K. TAMURA, and M. NEI. 1993. MEGA: molecular evolutionary genetics analysis. Version 1.Ol . Pennsylvania State University, University Park. LI, S. S.-L., W. M. FITCH, Y.-C. E. PAN, and E S. SHARIEF. 1983. Evolutionary relationships of vertebrate lactate dehydrogenase isozymes A4 (muscle), B, (heart), and C4 (testis). J. Biol. Chem. 258:7029-7032. LONG, G. L. 1976. The stereospecific distribution and evolutionary significance of invertebrate lactate dehydrogenases. Comp. Biochem. Physiol. 55B:77-83. LUNDIN, L. G. 1993. Evolution of the vertebrate genome as reflected in paralogous chromosomal regions in man and the house mouse. Genomics 16:1-19. MANNEN, H., S. C.-M. TSOI, D. B. PICKFORD, J. A. DONALD, L. J. GUILLETTE, and S. S.-L. LI. 1996. Sequences of the lizard cDNAs encoding lactate dehydrogenase (LDH) isozymes A (muscle) and B (heart). Gene 171:303-304. MARCHUK, D., M. DRUMM, A. SAULINO, and E S. COLLINS. 199 1. Construction of t-vectors: a rapid and general system for direct cloning of unmodified PCR products. Nucleic Acids Res. 19: 1154. MARKERT, C. L. 1984. Biochemistry and function of lactate dehydrogenase. Cell Biochem. Funct. 2:131-134. MARKERT, C. L., J. B. SHAKLEE, and G. S. WHITT. 1975. Evolution of a gene. Science 189: 102-l 14. MATSON, R. H. 1989. Distribution of the testis-specific LDHX among avian taxa with comments on the evolution of the LDH gene family. Syst. Zool. 38:106-l 15. MORIZOT, D. C. 1990. Use of fish gene maps to predict ancestral vertebrate genome organization. Pp. 207-234 in Z.-I. OGITA and C. L. MARKERT, eds. Isozymes: structure, function, and use in biology and medicine. Wiley-Liss, New York. OHNO, S. 1970. Evolution by gene duplication. Springer, New York. PETERSON, K. J. 1995. A phylogenetic test of the calcichordate scenario. Lethaia 28:25-38. QUATTRO, J. M., D. D. POLLOCK, M. POWELL, H. A. WOODS, and D. A. POWERS. 1995. Evolutionary relations among vertebrate muscle-type lactate dehydrogenases. Mol. Mar. Biol. Biotechnol. 4:224-23 1. QUATTRO, J. M., H. A. WOODS, and D. A. POWERS. 1993. Sequence analysis of teleost retina-specific lactate dehydrogenase C: evolutionary implications for the vertebrate lactate dehydrogenase gene family. Proc. Natl. Acad. Sci. USA 90: 242-246. SAMBROOK, J., E. E FRITSCH, and T. MANIATIS. 1989. Molecular cloning: a laboratory manual. 2nd edition. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. SHARMAN, A. C., and I? W. H. HOLLAND. 1996. Conservation, duplication, and divergence of developmental genes during chordate evolution. Neth. J. Zool. 46:47-67. SHEETS, M. D., S. C. OGG, and M. I? WICKENS. 1990. Point mutations in AAUAAA and the poly (A) addition site: effects on the accuracy and efficiency of cleavage and polyadenylation in vitro. Nucleic Acids Res. 18:5799-5805. SIEBENALLER, J. E, T. L. ORR, B. B. OLWIN, and S. S. TAYLOR. 1983. Comparison of the D-lactate stereospecific dehydrogenase of kmulus polyphemus with active-site regions of L-lactate dehydrogenases. Biochim. Biophys. Acta 749: 153-162.

STOCK, D. W., K. D. MOBERG, L. R. MAXSON, and G. S. analysis of the 18s ribosomal WHI’IT. 199 1. A phylogenetic

RNA sequence of the coelacanth Latimeria chalumnae. Environ. Biol. Fish. 32:99-l 17. STOCK, D. W., and D. A. POWERS. 1995. The cDNA sequence of the lactate dehydrogenase-A of the spiny dogfish (SquaZus acanthias): corrections to the amino acid sequence and an anlysis of the phylogeny of vertebrate lactate dehydrogenases. Mol. Mar. Biol. Biotechnol. 4:284-294. STOCK, D. W., and G. S. WHITT. 1992. Evolutionary implications of the cDNA sequence of the single lactate dehydrogenase of a lamprey. Proc. Natl. Acad. Sci. USA 89:17991803. STUBBS, L., E. M. RINCHIK, E. GOLDBERG, B. RUDY, M. A. HANDEL, and D. JOHNSON. 1994. Clustering of six human 11~15 gene homologs within a 500-kb interval of proximal mouse chromosome 7. Genomics 24:324-332. SWOFFORD,D. L. 1991. PAUP phylogenetic analysis using parsimony. Version 3.0s. Illinois Natural History Survey, Champaign. SYNSTAD, B., 0. EMMERHOFF, and R. SIREVAG. 1996. Malate dehydrogenase from the green gliding bacterium Chlorojlexus uuruntiucus is phylogenetically related to lactic dehydrogenases. Arch. Microbial. 165:346-353. TEMPLETON, A. R. 1983. Phylogenetic inference from restriction endonuclease cleavage site maps with particular reference to the evolution of humans and the apes. Evolution 371221-244. THOMPSON, J. D., D. HIGGINS, and T. J. GIBSON. 1994. CLUSTAL W improving the sensitivity of progressive multiple sequence alignment through sequence weighting, positionspecific gap penalties and weight matrix choice. Nucleic Acids Res. 22:4673-4680. TSOI, S. C.-M., and S. S.-L. LI. 1994. The nucleotide and deduced amino-acid sequences of a cDNA encoding lactate dehydrogenase from Cuenorhubditis eleguns: the evolutionary relationships of lactate dehydrogenases from mammals, birds, amphibian, fish, nematode, plants, bacteria, mycoplasma, and plasmodium. Biochem. Biophys. Res. Commun. 205:558-564. TSUJI, S., M. A. QURESHI, E. W. Hou, W. M. FITCH, and S. S.L. LI. 1994. Evolutionary relationships of lactate dehydrogenases (LDHs) from mammals, birds, an amphibian, fish, barley, and bacteria: LDH cDNA sequences from Xenopus, pig and rat. Proc. Natl. Acad. Sci. USA 91:9392-9396. WHITT, G. S. 1969. Homology of lactate dehydrogenase genes: E gene function in the teleost nervous system. Science 166: 1156-l 158. 1984. Genetic, developmental and evolutionary aspects of the lactate dehydrogenase isozyme system. Cell Biochem. Funct. 2: 134-139. WOLFF, J., and H. R. KOBEL. 1982. Lactate dehydrogenase of Xenopus luevis luevis and Xenopus borealis depends on a multiple gene system. J. Exp. Zool. 223:203-210. WYNNE, S. A., D. J. NICHOLLS, M. D. SCAWEN, and T. K. SUNDARAM. 1996. Tetrameric malate dehydrogenase from a thermophilic Bacillus: cloning, sequence and overexpression of the gene encoding the enzyme and isolation and characterization of the recombinant enzyme. Biochem. J. 3171235-245. YANG, S., and S. E PARMLEY. 1995. A bradyzoite stage-specifically expressed gene of Toxoplasma gondii encodes a polypeptide homologous to lactate dehydrogenase. Mol. Biochem. Parasitol. 73:291-294. SHOZO YOKOYAMA,

Accepted

August

reviewing

29, 1997

editor