Sea Urchin Hox Genes: Insights into the Ancestral Hox Cluster Ellen Popodi, * Jessica C. Kissinger, 1_ Mary E. Andrews,” and Rudolf A. RafJ” *Department of Biology, Indiana University; and tNational Institutes of Health, Bethesda, Maryland We describe the Hox cluster in the radially symmetric sea urchin and compare our findings to what is known from clusters in bilaterally symmetric animals. Several Hox genes from the direct-developing sea urchin Heliocidaris erythrogramma are described. CHEF gel analysis shows that the Hox genes are clustered on a 5300 kilobase (kb) fragment of DNA, and only a single cluster is present, as in lower chordates and other nonvertebrate metazoans. Phylogenetic analyses of sea urchin, amphioxus, Drosophila, and selected vertebrate Hox genes confirm that the H. erythrogramma genes, and others previously cloned from other sea urchins, belong to anterior, central, and posterior groups. Despite their radial body plan and lack of cephalization, echinoderms retain at least one of the
anterior group Hox genes, an orthologue of Hox3. The structure of the echinoderm Hox cluster suggests that the ancestral deuterostome had a Hox cluster more similar to the current chordate cluster than was expected. Sea urchins have at least three Abd-B type genes, suggesting that Abd-B expansion began before the radiation of deuterostomes. Introduction A major ancestral role of evolutionarily conserved patterning genes, such as the Hox genes, appears to have been in specifying anterior-posterior identity in the animal nervous system. It has been hypothesized that changes in the number and organization of Hox genes have played a role in the evolution and diversification of chordate body plans (Pendelton et al. 1993). At least some of these genes are present in sea urchins (Ruddle et al. 19946), animals that do not have an obvious anterior-posterior polarity. Is the genomic structure of these genes conserved in this group? If so, are the genes expressed in an overlapping pattern in tbe nervous system? If the cluster structure is conserved, examination of the expression pattern may provide a new tool in the quest to explain the evolutionary origins of the echinoderm radial body plan. The Hox genes are found in all animal phyla examined to date, and their chromosomal organization has been highly conserved as well. The genes are clustered and there is a relationship between chromosomal position, activation time, and anterior expression limit. The conservation of these features suggests that clustering of these genes may be important for precise spatiotemporal gene regulation and embryonic patterning. To date, the complete organization of the Hox gene clusters has been reported from only a few widely divergent, bilaterally symmetric taxa (insects, nematodes, vertebrates, and cephalochordates). The role of these genes in axis specification during early development has probably been conserved since the beginning of the metazoan radiation. Assessment of this hypothesis requires a better knowledge of the genes present in other animal phyla. Echinoderms and chordates shared a bilaterally symmetric ancestor (Turbeville, Schulz, and Raff 1993). Abbreviations: bp, base pair(s); CHER contour clamped homogenous electric field; kb, kilobase or kilobase pairs; PCR, polymerase chain reaction; UTR, untranslated region(s). Key words: Heliocidaris erythrogramma, Hox homeobox, echinoderms, sea urchin, ancestral Hox cluster, radial symmetry. Address for correspondence and reprints: Ellen Popodi, Department of Biology, Indiana University, Bloomington, Indiana 47405. E-mail:
[email protected]. Mol. Bid. Evol. 13(8):1078-1086. 1996 0 19% by the Society for Molecular Biology and Evolution. ISSN: 0737-4038
1078
Although chordates retained that bilateral symmetry, echinoderms have evolved a pentamerally symmetrical adult body plan. These body plans have manifested an extraordinary stability for over 530 million years (Gould 1989). As part of a study on the origin of the echinoderm body plan, we have examined the Hox genes of the direct-developing sea urchin Heliociduris erythrogrumma. We have cloned the 3’ exons of five Hox genes from H. erythrogramma. These exons encode the C-terminal portion of the protein including the entire homeodomain (encoded by the homeobox). Although Hoxtype homeobox genes have been cloned from other sea urchins (Tripnuestes gratillu [Dole&i et al. 1986; Dole&i, Wang, and Humphreys 1988; Wang et al. 19901, Strongylocentrotus purpuratus [Zhao 19921, Paracentrotus lividus [Di Bernard0 et al. 1994]), this is the largest collection from a single species so far presented. When all sea urchin species are considered, a total of eight different Hox-type homeobox genes have been identified. We demonstrate that these eight genes are physically linked, and thus are clustered, like those of bilaterally symmetric animals. Phylogenetic analyses of the relationships of sea urchin, insect, and chordate Hox genes indicate that most of the urchin genes cloned so far represent members of the 5’ and central portions of the cluster. One sea urchin gene is a probable orthologue of a 3’ Hox gene. We discuss the implications of our current knowledge .of the sea urchin Hox gene cluster for the evolution of Hox gene cluster organization. Materials and Methods Genomic Library Construction, Screening, and Sequencing A library of H. erythrogramma genomic DNA was constructed in lambda DASH (Stratagene). Six genomic equivalents (4.4 X lo5 plaque-forming units) of an unamplified portion of the library were screened with a homeobox-containing fragment of the T. grutillu Hboxl cDNA. Hybridizing fragments were subcloned into Bluescript (Stratagene) and sequenced using the chaintermination method (Sanger, Nicklen, and Coulson 1977) with the Sequenase Sequencing Kit (United States Biochemical) and radiolabeled nucleotides, or with dyelabeled primers (Li-Cor, Inc.) employing the Sequitberm
Sea Urchin Hox Genes
Sequencing Kit (Epicenter Technologies). Radiolabeled sequencing products were run on standard sequencing gels and analyzed by autoradiography. The dye-labeled sequencing products were analyzed on a Li-Cor automated sequencer (Li-Car, Inc.). Polymerase Chain Reaction (PCR) Amplification of Homeoboxes A range of 10 pg to 10 ng of purified plasmid or phage DNA was used as the substrate for PCR amplification of homeoboxes. The primers used were: Hoxlf: 5’-TGGARCTGCAGAARGART-3’ a n d Hox4r: 5’CATCCGGCCGTTYTGRAACC-3’. These primers were directed against highly conserved regions of the homeobox and amplified a 117-bp fragment. Sequence Comparisons The Genetics Computer Group (GCG) sequence analysis package (Genetics Computer Group, Inc. 1994) was used to analyze the sequences and generate amino acid and nucleotide alignments. Multiple sequence alignments were generated using Pileup (GCG) and then refined by eye. Alignments were then imported into PAUP 3.1.1 (Swofford 1993) or PHYLIP 3.5 (Felsenstein 1993). All parsimony analyses utilized 100 random addition replicates and 100 bootstrap replicates. When indicated, certain analyses were constrained in the number of trees that could be saved in each replicate because of the high computational burden of swapping on equally parsimonious trees. In these instances at least 500 random addition replicates were performed. Neighborjoining analyses were performed and bootstrapped using the PHYLIP programs SEQBOOT, DNADIST, NEIGHBOR, and CONSENSE. Assumptions included the Kimura two-parameter model and a transitionhransversion ratio of 2.0. Third positions were ignored in all nucleotide analyses. The data sets were bootstrapped 100 times. Maximum-likelihood analyses were performed with the PHYLIP program DNAML. Empirical frequencies were used and global swapping was permitted. Parsimony trees were further analyzed with MacClade Version 3 (Maddison and Maddison 1992). The following genes were used in the alignment presented in figure 6. Drosophila: lb, X13103; p b , X63728; Dfd, X05136; Scr, X14475; antp, M20704; Ubx, X76210; abd-A, X54453; Abd-B, X51663; en, M29285. Mouse: Hoxa-I, X06023; Hoxa-2, M93148; Hoxa--3, K02591; Hoxa-4, X13538; Hoxa-5, M36604; Hoxa-6, M11988; Hoxa-7, M17192; Hoxa-9, M28449; Hoxa-IO, L08757; Hoxa-II, U20370 Hoxb-1, X59474; Hoxb-2, M34004; Hoxb-3, X66 177; Hoxb-4, M36654; Hoxb-5, M26283; Hoxb-6, X56461; Hoxb-7, X06762; Hoxb-8, X13721; Hoxb-9, M34857; Hoxc-4, S62287; Hoxc-5, U28071; Hoxc-6, X16511; Hoxc-8, X07646; Hoxc-9, X55318; Hoxc-IO, D11330; Hoxc-12, U04839; Hoxc-Z3, U04838; Hoxd-Z, X60034; Hoxd-3, X73573; Hoxd-4, JO3770; Hoxd-8, X56561; Hoxd-9, X62669; Hoxd-IO, X62669; Hoxd-11, X 6 0 7 6 2 ; H o x d - 1 2 , X58849; Hoxd-13, S80533; EN-I, L12703; EN-2, L12705. Amphioxus: amphihoxl, 235142; amphihox2, 235 143: amDhihox3. X68045: amuhihox4. 235 144: am-
1079
phihox5, 235145; amphihox6, 235146; amphihox7, 235 147; amphihox8, 235 148; amphihox9, 235 149; amphihoxl0, 235 150. CHEF (Contour Clamped Homogenous Electric Field) Gel Electrophoresis To obtain intact DNA, we lysed sperm suspended in agarose. The DNA was digested with restriction enzymes that cut at rare sites, Not I or S’ I, and large fragments were separated using CHEF gel electrophoresis. Southern blots of gels resolving fragments from 47 to 2,200 kb were probed with fragments specific for individual sea urchin Hox genes. Each blot was then stripped and reprobed with one or more different Hox genes to confirm that the genes were hybridizing to the same bands. Removal of the previous probe was confirmed before the next probe was applied. To rule out the possibility of artifactual hybridization, we reprobed one blot for the engrailed gene, which is not linked to Hox genes. Hybridization was viewed both by standard autoradiography and with a Molecular Dynamics phospho-imager. Results Isolation of Hox genes from H. erythrogramma Sixteen positive phage clones were obtained from a screen of six genomic equivalents of an unamplified genomic library. We used a portion of the TgHboxl cDNA (Angerer et al. 1989) that contained the entire homeobox as a probe (HBl probe). PCR amplification using primers specific for conserved regions near either end of the Hox-type homeobox confirmed the presence of homeoboxes in 11 clones. Subsequent sequencing sorted the clones into five different Hox-type gene classes. Three of the gene types had been cloned from other sea urchin species: Hboxl and Hbox6 from T. gratilla (TgHboxl and TgHbox6-Dolecki et al. 1986; Wang et al. 1990), and Hbox7 from S. purpuratus (SpHbox7Zhao 1992). The remaining two were unique until a recent PCR survey of Hox genes in sea urchins described limited portions of these genes, and named them Hbox9 and HboxZO (Ruddle et al. 1994b). Following the established nomenclature for sea urchin Hox genes, all H . erythrogramma genes are designated by the species abbreviation and the Hbox gene number and not by the designation “Hox” until the absolute identities are determined. We have sequenced the entire coding sequences of the 3’ exons for each of the five H. erythrogramma Hox genes; these sequences have been deposited with GenBank under the following accession numbers: HeHboxl, U30935; HeHbox6, U31447; HeHbox9, U31563; HeHbox7, U31564; and HeHboxZO, U31600. One of five phage clones containing the HeHboxZ homeobox also contained the coding sequence of the 5’ end of the gene (U46002). The predicted HeHboxl intron is 12.5 kb long. Linkage We used CHEF gel electrophoresis to examine the linkage of the eight sea urchin Hox genes. All genes gave exactlv the same hvbridization oattem (fig. 1A).
Sea Urchin Hox Genes 1 Antennapedia
:’ Qrou9 zur&ihox-1
10
20
30
40
50
60
~QQ&Y~&lYLtiIE~i~IK~
61
70
KTKGBIT
...A..“...A..N.N.T ‘TJ.. ........*
..px
.........Q
..p.E
1 1
m a-l m b-l
1
m d-l dIab
ps,Ar.a.np,orx.Lr.. . . . . . . . x.. . .*. . . . . . nc.p.t.Jm.“. . . . . . . . . . Q ..r4E ~.s..~.~.n T.. . . . . . . . . . . . ‘A.. . . . .NT.*.n.r.“. . . . . . . . . . o ..w
KENGFSTFGSXSPAG KEOLLPISPATPFGSD Fm2GFx?AGFmxPKE F.EGLLATAAS"ASII(L KEGLIPADILTQHSTS
Anphihox-2 m a-2 mb-2 dpb
sa.L.a.7Jr.rQ.L.. . . . . . . Y.x.XJcxp..x...sY.n.n “‘V.. . . . . . . . a*.m sa.L.r*..n.r ‘L.. . . . . . . . . x..c.p ..” . ..*r& .n.. . . .V.“. . . . . . . . x.pQr sa.n.**..)rr.n. . . . . . . . . . x..c.p..“...AL .n.. . . .“.“. . . . . . . . n.RQT pa.n.r*..)rr ‘L.. . . . . . . . . x..c .p . . . . . . M.D.. . . . “‘“. . . . . . . . n.RQT
TKGFSEICTDWTAAE CCKEWJSEDKFKNLE E... LSKTDDEDHKDSLKGD
PlIibCdl
p..N.rp~.~.L TJ.. ..........C.P..“.M.KS .N ................y .R G..*q.*..S*.L”. ...........C.p._“.M.m.N.. .............. y . ..Q $. ...TA...P.L”. ...........M.p..“.M.)& ‘N.. ..............y ..* S..*.T*..S*.L”. ...........c.p ..“.M.~.N ‘S.. ............y ..&J S..V.T...S*.L”. ...........c.p ..“.M.NL ‘N.. ..............y ..m PR.L.TA..NT ‘L.. ........K..C .p.. ....AS.D.. ...“.“. ....... H.pQT
KvKmGa;sw WT... KAF.GLASssooPsPAG KAKGILJiSPAWSPER LSKTIxmmmSIxGD
s..s.rp* ............................. . ..G ....................
NvKSISQLIS*EmaN
1 1 1 1 1
Amphihox-3 m a-3 m b-3 m d-3 dpb
1 1
PMM;._.TK.LT.. .....yq(. ppJAT”r.-.~.LT..
.......I(. ...*..“...AS.Q.N.* .“.
poDL._.TR.LT..
.......K..,g.A
1081
..” . ..*T.E.N.* .“. .........Q ..p.E
cantra1 S.P..i ydlscr OIOUPI HemmX9 1 1 1
Arrphihox-4 m a-4 mb-4 m c-4 m d-4 d Dfd
1
&r&ihax-5 m a-5 mb-5 m c-5 dScr
1 1
1 HtiC lTgHbax3 ~mphihox- 6
1
1 1 1 1 1 1
................. ..D.
p_.s.T*...Q ‘V.. ..................... T...S p..S.T*...Q .“. .......y..
“‘V.. ........... m
..... ..”........S “““““‘“““M
s.
P..S.T*...Q .“. .......y.. ............
..S ............... ..~
p..s.T*...*.“.......................=...p.................. p..*.q.*...H ‘I.. ...... y.. ............T.“.S ............... ..D. N..T.T* ..................................................... (J..J&.l.A ............................. .. ..~.................~. G..*.T* ......................... ...... ..~.................~. (J..S.q.S .......................... ..~ .... .................m T..*.TS ........................................ .............
WSsAsooT KLFwrmF"sSI?rm lummF.IRsGcTAGRA RLpNlxvRsApPAGAA KLP?cmGFssssSsCs KLm
*ral9 I .G . . . . . . . . . . . . . . . . y “‘S . ..~.............................~S.~ . . . . . . . . ..*..........y......K......*.xJ ..S . . . . . . . . . . . . . . . . . . .
G.........Q...........S..".....F...*S.G.S................K
..
~.....................~....~........~.G ..................... G~.................................~ . . . . . . . . . . . . . . . . . . . . . . . . G~...................y . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .R....I.S..........................~ . . . . . . . . . . . . . . . . . . . . . ..S
dUbx d AM-A
KIPSLNATTINQLTNM KLINSINRSGEDSEAK KLLSASQLSAEEEW NLTSTISGaoGoATAD .K...................T .H ....... ..M.....................L . ..L QRIKwNEQMgAQAQ . . . . . . . . ..F . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . FJ+vKEB
Aq&ihox-7 m a-7 m b-7
. . . . . . . . . . . . . . . . . . . . . ..K . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........................................................... . ............................................................
Artphihox-8 m b-8 c-8
.R......S..............~..............G................~ . ..A
m a-6 m b-6 m c-6
1
T..S.T*...*‘V.. ..................... ,g.G
m m
d-8
.~~.....~...........L”P”“K”“VS”‘O’....~ ............... .~~.....~...........L”P”“K”“VS”‘G’....~ ............... .K......S.F.........L ..p ....K.. ..XJS’T’A.. . ..“. ..............
Kl5SLKcQPAESETSS mEsQAFrAAPEnwP KTSGPGITGQDKAEGE
AMxPPiammmsS malIcFPSSF.CEQEELa NmmnaFmBBKvBE NKDKFPASF.PxlcGD
z-9.lik. gzou9‘r , Ii-7 I,2 HboXlO 1.2 lwkx4 1.2
1.2 1,2 1.2 1.2
Amphihox.9 m a-9 mb-9 m c-9 m d-9 d AH-B
2
Anphihox-10 m a-10 m c-10 m d-10
2
m a-11 m d-11
2
m c-12 m d-12
2
m c-13 m d-13
.R~.Rp.SKL,.Iy.. ....*A.M.. ..)J..$.m,Q “D.. . ..“. ......... M..L. m.Kp..KF “F.. ....LY’M.. ..D..SH.SR “S.. . ..“. .........L ..m GRKK.r.-...K F.. ....... L “M.. ..D..L...RL ‘S.. . ..“. ........ .&f..* .. s-.cp “‘F.. ....... LY.M.. .. E..Y..SQ -. .....J.. ........M ..s m.Cp..K H.. ....... L “M.. ..D..Y.“.p& .N ... ..“. ......... M..I. S-.Cp ..I(. .........L ..M .. ..~..~.~.~~.~.~...“..........~..~. m.cp “K.. ........L ....... D..Y.“.R” ‘N.. . ..“. ......... M..M. m.(Tp “K.. ........L “M.. ..D..Y.“.RI ‘N.. ...v.. ........ M ..s _.KP.SW.. ....... L..A.“SKQK.W.L.pJq ‘Q.. ...v.. ........ N ..m _.CP..K ..I.. .....~..~.“,g.~..Q..~~.~~ ..“. .........M.M. -.cp ..m.. ....... L “M.. .. E ..L ..SRM..D ..“. ......... I,..M. G-.cp ..m.. ....... L ..M ....E..L..SKTIN..D ..“. ......... L,..$$. G-.cp ..m.. ....... L “M.. ..E ..l_ ..S_..D ..“. .........L.-m TRKK'CP"K"IR"'R"F'SV'I~ ~.~..D.."..........E..~.
~~.~p..~..*~...~..~..“.~~.~Q~~~.~..~ ..lJ.. ........ E..L. KL’LA”‘G”L”‘EFI”*--R’LSD.~.~ ~.~p..~*.~~...~..~“.~~~~.Q~.~.~~.~.~~ .“. ......... K.pJjTJ ... ..y~~~.~.~.~~.~ “‘VT.. .... ~~.“p..~~.~...~.yA~.~~NKTa(.~.~A.~.~ . ..“. ....... xJ.D ..JyJ
KQRQBWXPPQ' KDFAKDE' NGAGm' KEK?zMEQs* KEKCPKa-J* QNW%Q_sN KAREEQIFXQEKGKQ RENRIFXlTANXNFS' RINF.IFELTSNEWFT* RFNRIRELTANLWS* FDRLQYYSANPLL' mm' QREPRLALY' sm*
FIG. 2.-Alignment of mouse, amphioxus, sea urchin, and Drosophila Hox gene homeodomains, and downstream amino acid sequences. Amino acid identity to Anrp is indicated by dots. Termination codons are indicated by *. Genes included in the general Hox analyses summarized in table 1 are indicated with a superscript “1.” Genes included in the Abd-B-type gene analyses are indicated by a superscript “2.”
1082 Popodi et al.
Table 1 Analyses that Place Sea Urchin Genes in Specific Groups Gene
Cluster Associated with Gene
Analyses that Support the Clustep
HeHbox7..
. . . Abd-B-likeb HeHboxlO . . . Abd-B-likeb TgHbox4 . . . Abd-B-likeb TgHbox3 . . AnzplUbx-likec (Hox-6-8)
All analyses All analyses All analyses ML, D,, PN,, PA,, PA*d, D,, PN,, D,, P N , AntpllJbx-like’ + Abd-B-likeb ML, LM,, PN4 HeHboxl . . AntpllJbx-likec ( H o x - 6 - 8 ) ML Dd, PN,, D,, PN,, D,, P N , AntpllJbx-likec + Abd-B-likeb a,, ML,, PN., HeHbox6 . . . AntpKJpx-like’ (Hox-6-S) ML,, D,, PN,, PA,, PA*,, D,, PN,, D,, PN, AntplUpx-likec + Abd-B-likeb ML, W,,, PN HeHbox9 . . . . . . DfdlScr (Hex-4-5) Q,, Da, PA, PA*, Dfdlsr + AntplUbx-likec (Hox-4-8) D, Scr + AntplUbx-likec (Hox-5-8) PN,,, PlHoxll . . . . pb, Hox2 ML, D,, PN,, PA,, PA*, Hox3 only PA,, PA*,, D,, PN,, PA,, PA*,, P&, ML,, PN, lablpb, Hox-l-3 ML. D,, PN,, ML,, PN,, D, a Types of analyses: ML = maximum likelihood; D = distance by neighbor-joining using nucleotides; PN = parsimony using nucleotides; PA = parsimony using amino acids. Subscript indicates the species whose genes were included in the dataset; d = Drosophila; a = amphixous; m = mouse, 4 = all four species. * = engrailed genes were not inlcuded in these analyses. bAbd-B-like genes = Abd-E, Hex-9-13 (only 9 and 10 were used in these analyses), and TgHbox4, HeHbox7, and HeHboxlO. ’ AnrpKJbx-like genes = Anrp, Vbx, abd-a, and Hex-64 groups and the sea urchin genes HeHboxl, HeHbox6, and TgHbox3.
tion when identifying orthologous relationships among Hox genes. For brevity we used only one of the five mouse Abd-B-like homeoboxes in the above analyses. To try to determine the specific identity of the sea urchin Abd-Blike Hox genes, we assembled data sets containing the three Abd-B-like sea urchin homeoboxes, representatives of each of the five Abd-B-like groups from mice, the two Abd-B-like homeoboxes of amphioxus, and the single Abd-B homeobox of flies. The genes used in these analyses are indicated by a superscript “2” in figure 2. We applied three inference methods (maximum likelihood, ML; distance, D; and parsimony, P) to the nucleotide data set using engrailed genes as an outgroup. Three data sets were analyzed using only the parsimony method: (1) the nucleotide data set using Hox-7/S genes as the outgroup and (2 and 3) amino acid data sets with either engruiled or Hox-7/8 outgroups. Ruddle et al. (1994b), basing their analyses on an 81-nucleotide region amplified by PCR, observed that the three sea urchin Abd-B-like genes look most like the Hox-9 group. Our analyses do not support this inference. The difference in results could be attributed to the additional phylogenetic signal provided when using the entire homeobox. We conclude that it is premature to assume that all three sea urchin genes are most like the Hox-9 group. The three sea urchin sequences Hboxl, Hbox3, and Hbox6 cluster with AntplUbx-like genes. These genes do not always form a monophyletic clade (e.g., fig. 4B and C): sometimes it is paraphyletic, with the Abd-B clade contained within it (fig. 4C). However, in every analysis the sea urchin genes behave like the AntplUbxlike genes of the other species. As with the Abd-B-like genes, how the genes behave varies with the inference method used and the set of genes included.
HeHbox9 appears to be an orthologue of Scr or Dfd. There are two hypotheses about the evolution of these groups. Some authors propose that the Dfd-like and &r-like genes evolved from a common ancestral gene (e.g., Holland 1992; Btirglin and Ruvkun 1993); others (e.g., Akam et al. 1994) propose that they evolved from different genes, with Scr evolving from the ancestor of the middle group genes. In our analyses these genes rarely form a monophyletic group and are usually positioned outside of the AntplUbx-like group. Current hypotheses propose that the DfdLScr-like genes diverged before the AntplUbx-like group genes. If Hbox9 is a Dfdl &r-like gene then it should form a sister group to the AntplUbx-like group. This is often the case in our analyses (fig. 4A and C). Also, when only sea urchin genes are compared, the Hbox9 sequence forms the sister group to a clade consisting of the Hboxl, Hbox3, and Hbox6 genes (data not shown). The predicted amino acid sequence of Hbox9 immediately downstream of the homeobox resembles Scrlike (Hox-5) sequences more than it does Dfd-like (Hox-4) sequences (fig. 2). Three residues C-terminal to the homeodomain of the Hbox9 gene can be aligned with the Hox-5 sequences. A small gap is required to align Hbox9 to the same region of Hox-4 proteins. This may indicate that Hbox9 is an orthologue of Hox-5 genes, but we need to know the sequence of the other central group gene before we can make this assignment. Discussion
We have shown that the eight known sea urchin Hox genes are clustered within 300 kb in the H. crythrogrumma genome and that only one cluster is present. We compared all eight genes to Hox genes of Drosophila, amphioxus, and mouse. The homeobox is the
Sea Urchin Hox Genes 1083
r musenl
F mushoxal
L mushoxb8 .-.h,.“7
L l-
B)
I Ihv
L
L Ubx Scr
10
mushoxb4 ‘t Dfd
10
FIG. 3.-Trees generated by distance analysis with neighbor-joining using nucleotides (A) or parsimony analysis with amino acids (B, C). A, Amphioxus and sea urchin homeoboxes using first and second codon positions only: the outgroup is Heen. B, Drosophila and sea urchin homeodomains; the outgroup is labial. This tree is one of three equally parsimonious trees of length 145. C, Mouse, amphioxus, Drosophila, sad sea urchin homeodomains, the outgroup is engrailed. This is one of 244 equally parsimonious trees of length 32 1. Due to the large number of equally parsimonious trees, tree C was generated in an analysis constrained to save no more than the four shortest trees in each of 500 random addition replicates. Bootstrap values >30% are indicated above or below the branch. The scale bars shown underneath B and C represent 10 changes.
only region in the 3’ exons sufficiently conserved to be informative. The information contained within the homeobox is insufficient to allow precise identification of orthologous relationships between the sea urchin genes and the Hox genes of other organisms. However, it is possible to place the sea urchin genes into identifiable classes. They represent anterior, central, and posterior portions of the cluster. HboxZI is an anterior group gene, and may represent an orthologue of Hox-3 (Di Bernard0 et al. 1995; table 1). Hbox9 appears to be an orthologue of Hex4 or Hox-5. Hboxl, Hbox3, and Hbox6 are Antpl Ubx-like (Hox-6-8). Hbox4, Hbox7, and HboxlO are Abd-B-like (posterior group, Hox-9-13) genes. Additional information is required before the sea urchin genes can be assigned to individual paralogy groups. Position in the Hox cluster, relative order of expression, expression patterns, and additional protein sequence should be taken into consideration when assigning orthologous relationships. In some instances, the amino terminal regions of orthologous genes show conservation of a few short
regions (Krumlauf 1992; Btirglin 1994). These regions are usually coded by exons upstream of the homeobox. With two exceptions, TgHboxl (Angerer et al. 1989) and SpHbox7 (Zhao 1992), upstream regions have not been cloned from echinoderms. Using the cDNA sequences from TgHboxl we identified the coding sequences of the 5’ exon of HeHboxl. The predicted amino acid sequence was compared to sequences in GenBank using the BLAST search algorithm. The closest matches were to vertebrate Hox-7 genes. Such information, combined with the gene’s position within the cluster, has been used to assign the amphioxus gene amphihox6 to the Hox-6 group. We do not yet know the position of the HeHboxI gene in the sea urchin cluster. Figure 4 diagrams the sea urchin Hox cluster relative to the prototypical vertebrate cluster. Some genes have not yet been found in sea urchins. Although Ruddle et al. (1994b) have done PCR surveys, closely targeted, exhaustive searches remain to be done. The two 3’ genes that are missing from sea urchins are present in other organisms. All arthropods and chordates have three 3’
1084 Popodi et al.
Sea Urchin
HBox7 HBoxlO HBox4
HBoxl HBox6 HBox3
Hbox9
Hboxl 1
Prototypical Vertebrate
Amphioxus:
Arthropod Abd-8 F IG. 4.Summary of the We do not know if there are comparison of the four clusters the box representing each gene. Akam et al. (1994).
I
I
I
I
-
I
abd-A Ubx
ASP
Scr Dfd
zen
pb lab
sea urchin Hox cluster and comparison to other Hox clusters. The top of the drawing depicts a sea urchin cluster. additional Hox genes in sea urchins (indicated by ?). The prototypical vertebrate Hox cluster is based on a present in mammals (Kappen, Schughart, and Ruddle 1989). The number of the paralogy group is underneath The amphioxus cluster is taken from Garcia-Femandez and Holland (1994). The arthropod cluster is taken from
group genes, and it seems likely that echinoderms do also. Analysis of zen from beetles and grasshoppers demonstrates that zen and Hox-3 genes are orthologous (Falciani et al. 1996). This strongly suggests that three 3’ group genes were present in the last common ancestor of arthropods and chordates. This ancestor predates the divergence of echinoderms and chordates. Recent PCR studies suggest that even Platyhelminthes (primitive, triploblastic bilaterians) have three genes resembling Hox-l-3 (Balavoine and Telford 1995), implying that the ancestor of all triploblastic animals had three 3’ genes. Another central group gene, probably the orthologue of Dfd, has yet to be isolated from sea urchins. We know of three Abd-B-like (posterior) genes in sea urchins; there are five Abd-B-like paralogue groups in vertebrates. The total number of Abd-B-like genes in deuterostomes other than vertebrates is unknown. The Hox cluster is thought to have evolved from three to five precursor genes via a series of gene duplications (Kaufman, Seeger, and Olsen 1990; Schubert, Nieselt-Struwe, and Gruss 1993; Ruddle et al. 1994~). Comparisons of the structures of the Hox clusters in a few animals have led to a number of hypotheses on the evolution of the cluster in relation to the divergence of animal phyla. Below we outline current ideas on the evolution of the Hox cluster, and point out where the sea urchin information confirms or conflicts with the specific hypotheses. Vertebrates have four copies of the Hox cluster, but most animals have a single copy. The ancestor of arthropods and chordates is hypothesized to have had a
single Hox cluster (Holland 1992; Krumlauf 1992; Ruddle et al. 1994a, 1994b). Sea urchins provide another example of a single Hox cluster in deuterostomes, supporting the concept that a single Hox cluster was present in the common ancestor of protostomes and deuterostomes. The Hox cluster present in the arthropod-chordate ancestor was thought to consist of five or six genes (e.g., Holland 1992; Akam et al. 1994; Ruddle et al. 1994a). According to these hypotheses a single AntplUbx-like gene was independently duplicated to produce three genes in each lineage (arthropods and chordates). The sea urchin cluster contains three AnrplUbx-like genes. It is striking that all these phyla have three such genes. Although our analyses cannot refute the proposal that the common ancestor of deuterostomes and arthropods had a single AnrplUbx-like gene, they do not support it either (fig. 2). A more parsimonious interpretation would be that three AntpllJbx-like genes were present in the ancestor. The five Abd-B-like paralogue groups (Hox-9-13) found in vertebrates arose by duplications of a single Abd-B-like gene before the divergence of vertebrates (Ruddle et al. 1994~). Pendleton et al. (1993) have proposed that those duplications could have been related to the evolution of features unique to vertebrates. The amphioxus cluster (see fig. 4) contains at least two Abd-Blike genes, implying that the expansion of these genes began before the divergence of cephalochordates and chordates (Garcia-Fernandez and Holland 1994). The presence of three Abd-B-like genes in sea urchins suggests that this expansion may have begun much earlier
Sea Urchin Hox Genes 1085
than previously thought. The 5’ end of the cephalochordate cluster was not mapped further than amphihox9 and amphihoxl0, therefore, we do not know if orthologues of Hox-II-13 are present in cephalochordates. The homeoboxes of the genes from the extreme 5’ end (Hex-Z Z-13) are quite different than the HOX-9 and Hox10 homeoboxes and may not be detectable with probes or primers based on sequences of these latter two genes (or Abd-B). The identification of proposed homologues of Hox-IO, Hox-12, and Hox-13 in ascidians (Ruddle et al. 1994b; Di Gregorio et al. 1995) suggests that more than two Abd-B-like genes were present in the ancestor of ascidians and chordates. It seems likely that the ancestral deuterostome cluster contained several posterior group genes, perhaps all five. The ancestor of arthropods and deuterostomes may well have had a Hox cluster much like that of modern arthropods. Nematodes are the only other phylum in which the structure of the Hox cluster is currently known. Cuenorhabditis &guns has four Hox genes in its cluster. We did not include nematodes in our analyses because the exact relationship of nematodes to arthropods and chordates is unclear. Nematodes have generally been considered to have diverged before the arthropod and chordate lineages split. Recent 18s ribosomal DNA sequence analyses imply that nematodes are a sister group of the arthropods (unpublished data). If nematodes diverged before arthropods and chordates split, then another step in the evolution of the Hox cluster can be outlined (Btirglin and Ruvkun 1993; Wang et al, 1993). However, it is not clear if the small number of genes present in Caenorhabditis elegans represents the ancestral state, or gene loss within this lineage. The Hox cluster may have contained many of its present complement of genes early in bilaterian evolution. The evolution of radial symmetry in echinoderms apparently did not involve a radical change in the structure of the Hox cluster. How did radial symmetry arise? Perhaps the five radii represent the remnants of five axes. This and other scenarios are outlined in Raff and Popodi (1996). Examination of the expression patterns of the Hox genes may help to differentiate among alternative explanations for the evolution of the radial symmetry of echinoderms. Acknowledgments
We thank Heather Sowden for help in collecting Heliocidaris, and the University of Sydney and Sydney
Aquarium for their hospitality and help with facilities. We thank Dr. Mimi Zolan for assistance with CHEF gel electrophoresis, Ruth Walters for technical assistance, and Dr. J. A. Bolker for reading the manuscript. This work was supported by NIH grant ROl NS3 1661 and an NIH Genetics Training grant to J.C.K. LITERATURE CITED AKAM, M., M. AVEROF, J. CASTELLI-GAIR, R. DAWES, E FALCIANI, and D. FERRIER. 1994. The evolving role of Hox genes in arthropods. Development 1994 Supplement, pp. 209-215.
ANGERER , L. M., G. J. DOLECKI , M. L. GAGNON, R. LUM, G. W ANG, Q. YANG, T. HUMPHREYS, and R. C. A NGERER. 1989. Progressively restricted expression of a homeo box gene within the aboral ectoderm of developing sea urchin embryos. Genes Dev. 3:370-383. BALAVOINE , G., and M. J. T ELFORD . 1995. Identification of planarian homeobox sequences indicates the antiquity of most hox/homeotic gene subclasses. Proc. Natl. Acad. Sci. USA 92:7227-723 1. BORGLIN, T. C. 1994. A comprehensive classification of homeobox genes. Pp. 27-71 in D. DUBOULE , ed. Guidebook to the homeobox genes. Oxford University Press, New York. BORGLIN, T. C., and G. R UVKUN. 1993. The Caenorhabditis elegans homeobox gene cluster. Curr. Opin. Genet. Dev. 3: 615-620. DI BERNARDO , M., R. Russo, I? OLIVERI, R. MELFI, and G. SPINELLI . 1994. Expression of homeobox-containing genes in the sea urchin (Paracentrotus lividus) embryo. Genetica 94: 141-150. DI GREGORIO, A., A. S PAGNUOLO , E RISTORATORE, M. PISCHETOLA, E ANIELLO, M. BRANNO, L. CARIELLO, and R. DI LAURO. 1995. Cloning of ascidian homeobox genes provides evidence for a primordial chordate cluster. Gene 156: 253-257. DOLECKI, G. J., G. W ANG, and T. HUMPHREYS . 1988. Stageand tissue-specific expression of two homeo box genes in sea urchin embryos and adults. Nucleic Acids Res. 16: 11543-l 1558. DOLECKI, G. J., S. WANNAKRAIROJ, R. LUM, G. WANG, H. D. RILEY, R. CARLOS, A. WANG, and T. HUMPHREYS. 1986. Stage-specific expression of a homeo box-containing gene in the non-segmented sea urchin embryo. EMBO J. 5:925930. FALCIANI, E, R. SCHROEDER , M. AKAM, D. TAUTZ, R. DENELL, and S. BROWN. 1996. Class 3 Hox genes in insects and the origin of ten. Proc. Natl. Acad. Sci. USA (in press). FELSENSTEIN , J. 1993. PHYLIP (phylogeny inference package) 3.5~. Distributed by the author. Department of Genetics, University of Washington, Seattle. GARCIA -FERNANDEZ , J., and P. W. H OLLAND. 1994. Archetypal organization of the Amphioxus Hox gene cluster. Nature 370:563-566. GENETICS COMPUTER GROUP, INC. 1994. Program manual for the Wisconsin package. Version 8, September 1994. Sequence comparisons programs of the GCG sequence analysis package. 575 Science Drive, Madison, Wisconsin 53711. GOULD, S. J. 1989. Wonderful life. Norton, New York. HOLLAND, P. W. H. 1992. Homeobox genes in vertebrate evolution. BioEssays 14:267-273 KAPPEN , C., K. S CHUGHART , and E H. RUDDLE. 1989. Two steps in the evolution of antennapedia-class vertebrate homeobox genes, Proc. Natl. Acad. Sci. USA 86:5459-5463. KAUFMAN , T C., M. A. SEEGER, and G. OLSEN . 1990. Molecular and genetic organization of the Antennapedia gene complex of Drosophila melanogaster. Adv. Genet. 27:309362. KRUMLAUF , R. 1992. Evolution of the vertebrate Hox homeobox genes. BioEssays 14:245-252. MADDISON , W. I!, and D. R. M ADDISON . 1992. MacClade Version 3. Sinauer Associates, Inc., Sunderland, Mass. PENDLETON , J. W., B. K. NAGAI, M. T. MURTHA , and E H. RUDDLE. 1993. Expansion of the Hox gene family and the evolution of chordates. Proc. Natl. Acad. Sci. USA 90: 6300-6304. RAFF, R. A., and E. M. POPODI. 1996. Evolutionary approaches to analyzing development. In J. D. F ERRARIS and S. R. PAL-
1086 Popodi et al.
eds. Molecular zoology: advances, strategies and protocols. John Wiley and Sons, New York (in press). RUDDLE, E H., J. L. BARTELS, K. L. BENTLEY , C. KAPPEN, M. T. MURTHA , and J. W. PENDLETON . 1994~. Evolution of Hox genes. Annu. Rev. Genet. 28423-442. RUDDLE, E H., K. L. B ENTLEY, M. T. MURTHA, and N. RISCH. 1994b. Gene loss and gain in the evolution of the vertebrates. Development (Cambridge, U.K.) 1994 Supplement, pp. 155-161. SANGER, E, S. NICKLEN, and A. R. COULSON. 1977. DNA sequencing with chain-terminating inhibtors. Proc. Natl. Acad. Sci. USA 745463-5467. SCHUBERT , E R., K. N IESELT-STRUWE , and I? GRUSS. 1993. The antennapedia-type homeobox genes have evolved from three precursors separated early in Metazoan evolution. Proc. Natl. Acad. Sci. USA 90:143-147. SWOFFORD , D. L. 1993. PAUP: phylogenetic analysis using parsimony. Version 3.1.1. Computer program distributed by the Illinois Natural History Survey, Champaign, Ill. UMBI,
TURBEVILLE, J. M., J. R. SCHULZ , and R. A. R AFF. 1993. Deuterostome phylogeny and the sister group of the chordates: evidence from molecules and morphology. Mol. Biol. Evol. 11648-655. WANG, B. B., M. M. M ULLER-IMMERGLUCK, J. AUSTIN, N. T ROBINSON , A. CHISHOLM , and C. KENYON . 1993. A homeotic gene cluster patterns the anteroposterior body axis of C. elegans. Cell 74~29-42. W ANG, G. V. L., G. J. D OLECKI , R. CARLOS , and T. HUMPHREYS . 1990. Characterization and expression of two sea urchin homeobox gene sequences. Dev. Genet. 11:77-87. ZHAO, A. Z. 1992. Temporal regulation of the Ll late H2B histone gene during sea urchin embryonic development by an antennapedia class homeoprotein. Ph.D. dissertation, University of Southern California, Department of Biology, Los Angeles, Calif. G ARY O LSEN, reviewing editor Accepted June 12, 1996
