Development 111, 821-828 (1991)Printed in Great Britain © The Company of Biologists Limited 1991

821

Molecular cloning and expression of a novel homeobox gene AHoxi of the

ascidian, Halocynthia roretzi

HIDETOSHI SAIGA1*, ATSUSHI MIZOKAMI1, KAZUHIRO W. MAKABE2, NORIYUKI SATOH2 and

TAKASHI MITA1

^Department of Molecular Biology, University of Occupational and Environmental Health, Yahatanishi-ku, 1-1 Iseigaoka, Kitakyushu 807,Japan^Department of Zoology, Kyoto University, Kyoto 606, Japan

* Author for correspondence

Summary

We have isolated a novel ascidian homeobox gene,designated AHoxl, by screening the genomic DNA ofHalocynthia roretzi with the Bombyx mori Antennapediatype homeobox as a probe. The AHoxl gene encodes aprotein that consists of 741 amino acids. The homeoboxof AHoxl is interrupted by 2 introns each of which isabout 300 bp in length and it shows about 70 %similarity at a deduced amino acid level to that ofDrosophila H2.0. This suggests that AHoxl is one of themost diverged homeobox genes so far characterized.Northern blot hybridization with an AHoxl probeshowed the presence of single transcripts approximately2.8 kb in length in larvae, juveniles and some adulttissues. The expression of AHoxl is scarcely detectedduring the course of early development but it increasesto a moderate level at the larval stage. After metamor-

phosis, the level of AHoxl expression increases asdevelopment proceeds. In situ hybridization to thejuvenile 7 days after metamorphosis showed that the siteof AHoxl expression is the epithelium of digestive tract.Among the adult tissues examined, digestive tract,digestive gland and coelomic cells were the major sites ofthe expression of AHoxl. In gonad, body wall muscleand pharyngeal epithelium, the expression of AHoxl isrelatively weak. These results suggest that AHoxl isprimarily expressed in the tissues of endodermal originand that the gene expression may be associated withdifferentiation of the endodermal tissues.

Key words: ascidian, homeobox, molecular cloning, geneexpression, developmental expression, in situ hybridization.

Introduction

The homeobox was originally discovered as a smallsegment of conserved nucleotide sequence among theDrosophila genes that are involved in the formation ofaxes and body segments, and in the specification of thebody segments (for review see Gehring, 1987; Akam,1987). The homeodomain, the translated form of thehomeobox, has a helix-turn-helix motif, which confers aDNA binding capacity to the homeobox gene productand therefore it has been suggested that homeoboxgenes encode transcription factors. There are severallines of evidence supporting this hypothesis (Quian etal. 1989; Ingham, 1988; Scott et al. 1989, referencestherein). Homeobox genes have been found to bewidely distributed among various animal species, fromnematodes to insects and vertebrates on the phylogen-etic tree (McGinnis et al. 1984; Holland and Hogan,1986) where they play key roles in development.

Since the last century, it has been noted that the earlydevelopment of ascidians, which belong to the class

Protochordata, is exceptionally autonomous. Thisfeature enabled Conklin, Ortolani and Nishida andSatoh to trace the fate of each blastomere (Conklin,1905; Ortolani, 1955; Nishida and Satoh, 1983, 1985).The cell lineage in early development has beenestablished completely up to the 110-cell stage inHalocynthia roretzi (Nishida, 1987). It is thought thatthe autonomous development of the ascidians is due tofactors, named determinants, which appear to beinvolved in the determination of the fate of eachblastomere and are localized heterogeneously in thefertilized egg cytoplasm (for recent reviews see Satoh,1987 and Satoh et al. 19906; Uzman and Jeffery, 1986;Whittaker, 1987; Venuti and Jeffery, 1989). Themolecular identity of a determinant, however, is still anenigma (see reviews by Jeffery, 1985; Satoh et al.1990a).

We are interested in studying the homeobox genes ofascidians for two reasons. (1) Since ascidians belong tothe class Protochordata located phylogenetically be-tween invertebrates and vertebrates, the structure of

822 H. Saiga and others

the homeobox genes of ascidians should shed some lighton the evolutionary aspect of homeobox genes. (2)Since it is highly Likely that expression of homeoboxgenes is involved in determining cell differentiation, theexpression patterns of homeobox genes in ascidians ornematodes, which exhibit typical mosaic development,are of particular interest. This might lead to betterunderstanding of the molecular identity of the so-calleddeterminants. Therefore, we have started isolatinghomeobox genes from the ascidian, Halocynthia ror-etzi- In this report, we describe the isolation andmolecular characterization of a homeobox gene fromthe Halocynthia genome. Although we used theAntennapedia (Antp) type homeobox as a probe, theisolated homeobox gene, designated AHoxl, was not ofthis type but, unexpectedly, it has a similar homeoboxto that of H2.0. This is a Drosophila homeobox geneshowing tissue-specific expression rather than region-specific expression (Barad et al. 1988). Likewise theexpression of the isolated ascidian homeobox gene,AHoxl, is primarily detected in the endoderm duringthe course of development.

Materials and methods

AscidianAdult ascidians, Halocynthia roretzi, were purchased fromfishermen near Asamushi Marine Biological Station, TohokuUniversity, Mutsu Bay, Aomori, Japan. Adult specimenswere dissected and tissues were frozen quickly with liquidnitrogen. The frozen samples were kept at —80°C until use.

Naturally spawned eggs were fertilized with a suspension ofnon-self sperm, and fertilized eggs were raised in filtered seawater at 13—15 °C. Embryogenesis proceeded synchronouslyamongst batches of eggs. Tadpole larvae hatched about 33 hafter fertilization. They were allowed to accomplish metamor-phosis naturally. Juveniles that adhered to plastic dishes werecultured for 13 days in aquaria with circulating naturalseawater. Samples at appropriate stages were collected by lowspeed centrifugation and were frozen with liquid nitrogen.

Construction of genomic and subgenomic librariesHalocynthia roretzi genomic DNA was prepared from thegonad of one adult individual according to Blin and Stafford(1976).

Genomic DNA was digested partially with Sau3A andligated to the BamFfl-digested AEMBL3 vector. To constructa subgenomic library, genomic DNA was digested completelywith EcoRI and separated on a 1 % preparative agarose gel.The DNA fragments of about 2.7 kb in length were recoveredby electroelution and ligated to the £coRI-digested Agtllvector. Ligated materials were packaged and propagated inNM539 and Y1088 for genomic and subgenomic libraries,respectively.

The probes used for screening of the libraries were asfollows; a Hindi-Avail fragment of pBm3.6 about 250 bp inlength, which is the genomic clone of a Bombyx mori Antp-type homeobox gene (Hara and Suzuki, unpublished), and aPsd-TaqI fragment of pHBl about 200 bp in length, whichcontains a sea urchin homeobox as described by Dolecki et al.(1986).

For screening of the iibrary, hybridization was carried outunder reduced stringency conditions essentially as described

by McGinnis et al. (1984). In brief, hybridization was carriedout at 37 °C for 30-40 h in the mixture consisting of 38%formamide, 6xSSC, 0.2% polyvinylpyrrolidone, 0.2%Ficoll, 0.1% BSA, 0.5% SDS, 50mM sodium phosphatepH7.0, 20^gml"1 E. coli DNA and 1-5xK^ctsminimi"1

probe labelled by random priming using [a--32?] dCTP as thelabelled nucleotide. After hybridization, the filters werewashed in 2xSSC, 0.2% SDS at 37°C for 30min twice.

Preparation of poly (A) RNATotal RNA from embryos or adult tissues was preparedessentially according to Ullrich et al. (1977). H. roretziembryos were developed at 15 °C, collected by low-speedcentrifugation and kept at -80°C. Frozen packed embryos oradult tissues were suspended in 4 M guanidine thiocyanate,0.1M sodium acetate, 5mM EDTA, pH5.0 and werehomogenized using a Polytron homogenizer (Kinematics).After low-speed centrifugation to remove cell debris, thehomogenate was layered into a tube for a Beckman SW4OTirotor one-third filled with 5 M CsCl, 0.1M sodium acetate,5mM EDTA, pH5.0 and subjected to centrifugation at35000revsmin~^ for 15 h at 20 °C. The pellet was dissolved ina small volume of 10 mM Tris-HCl pH7.6, 10 mM EDTA,0.5% SDS, extracted with phenol and ether, and the RNAwas recovered by ethanol precipitation.

Poly(A) RNA was selected using Oligotex-dT30 Latexbeads (Roche Japan) according to manufacturer's procedure.Normally 1-3 % of the total RNA was recovered as poly(A)RNA. Quality of the poly (A) RNA was comparable to thatprepared using the ordinary oligo(dT)-cellulose column.

cDNA library constructioncDNA was synthesized from gonad poly(A) RNA using acDNA synthesis kit (Pharmacia). Synthesized cDNA wasligated to the Agtll vector. The ligated materials werepackaged in vitro as described in 'Construction of genomicand subgenomic libraries'.

Southern and northern blotting2 ng of genomic DNA digested with Eco RI, Hindlll or BgUlwas separated on a 1 % agarose gel. After electrophoresis, thegel was soaked in 0.5 M NaOH, 1.5 M NaCl for 15min andsubjected to blotting onto a nylon membrane (Hy-bond N,Amersham) using a low-pressure vacuum blotting apparatus(Vac Gene, LKB-Pharmacia). For northern blotting, poly(A)RNAs (10 ^g per lane) were separated on a 1 % agarose gelcontaining 6 % formaldehyde and were transferred to a nylonmembrane filter using a Vac Gene apparatus and IOXSSC astransferring buffer. Southern or northern blot hybridizationwas carried out according to the standard procedure (Maniatisetal. 1982).

In situ hybridizationIn situ hybridization using RNA probes and autoradiographywas carried out according to the method described byTomlinson et al. (1987). [35S]ATP-labelled probes wereprepared from a 1108 bp fragment of AHoxl cDNA(corresponding to nucleotides 209-1317 in Fig. 2 in Results)subcloned in the pGEM-Blue vector (Promega). The senseand antisense RNAs were transcribed with T7 and SP6 RNApolymerases, after linearization of the plasmid DNA withHindlll and EcoRI, respectively. Juveniles 7 days aftermetamorphosis were fixed in ice-cold 95 % ethanol:acetic acid(3:1, v/v), and embedded in Tissue-Prep (Fisher ScientificCo.). The specimens were sectioned at 8/an and attached tosubbed glass slides. After treatment with 1 /igmF1 proteinaseK, the specimens were postfixed in 4% paraformaldehyde,

A homeobox gene of an ascidian 823

treated with freshly prepared 0.25% acetic anhydride, 0.1Mtriethanolamine, and air dried. ApproximatelylxH^ctsmin"1 of 35S-labelled RNA probe was applied toeach slide in 50 /1 hybridization buffer (50% formamide,lxDenhardt's solution, 10% dextran sulfate, 2xSET, 0.01MDTT, 500/igmT1 TRNA, 500/igmT1 poly(A)). Hybridiz-ation reaction was carried out at 45 °C approximately for 12 h,and the slides were washed at high stringency (in 0.2xSSC at45 °C for 45min). The washed slides were dehydrated with95% ethanol, air dried and subjected to autoradiography.The autoradiographs were stained through the emulsion withhaematoxylin-eosin.

Results

Isolation of a homeobox genePrior to library screening, we carried out Southernblotting analysis of H. roretzi genomic DNA todetermine the conditions for hybridization so as to givesignals with reasonable strength. This was done using agenomic DNA probe of the Antp-type homeobox of silkworm or sea urchin. As shown in Fig. 1, under reducedstringency hybridization conditions, both probesresulted in several bands on a membrane filter but mostcoincided with those detected by ethidium bromidestaining of the gel. A 2.7 kb band detected upon EcoKldigestion, however, was distinct from those detected byethidium bromide staining. To facilitate efficientcloning, the ZscoRI-digested genomic DNA fragments

BE H Bg E H Bg

1kb

21.2 "

5.02 -

3.53 -i

2.03 _

1.38 —

0.85 —

Fig. 1. Detection of homeobox homology in the genomicDNA of the ascidian Halocynthia roretzi. 10 /ig of genomicDNA digested with EcoKl (E), Hindlll (H), or Bglil (Bg)was separated on a 1% agarose gel. Genomic Southernblots were hybridized under reduced stringency conditionswith the Bombyx homeobox probe (left) and the sea urchinhomeobox probe (right). An arrowhead indicates the2.7kb band, which was distinct from the ethidium bromide-stained bands. The size marker used was end-labelledADNA fragments digested with EcoRI and HindlU.

of about 2.7 kb in length were recovered, and fromthese a subgenomic library was constructed. Wescreened the subgenomic library under the reducedstringency hybridization conditions using the Bombyxhomeobox probe, since it gave lower background thanthe sea urchin probe. We were able to isolate a positiveclone, designated 62, which had a 2.7 kb insert. Bydetermination of the nucleotide sequence, it wasidentified as a clone containing a part of the homeobox.Using DNA fragments from the clone 62, we havefurther isolated three cDNA clones corresponding tothe genomic DNA fragment included in the clone 62from the adult gonad cDNA library as well as genomicclones covering the cDNAs from the H. roretzi genomiclibrary.

Structure of the ascidian homeobox gene, AHoxlFig. 2 shows the nucleotide sequence of cDNA of theascidian homeobox gene designated AHoxl. It has asingle long open reading frame consisting of 741 aminoacid residues and including a highly diverged homeo-domain. One structural feature of the AHoxl homeo-box is the presence of two introns of 319 and 314 bp inlength (data not shown) located at position 9 and afterposition 44, respectively. This is in contrast to theengrailed (en), invested (inv) and labial (lab) which haveone intron in the homeobox locating at position 17 (enand inv) or after position 44 (lab). Comparison of theamino acid sequence of the AHoxl homeodomain withthose of several representative Drosophila homeoboxgenes is shown in Fig. 3. The amino acid sequence ofthe AHoxl homeodomain is highly diverged from all ofthe known homeodomains with less than 50 % identityover 60 residues. The only one exception is thehomeodomain of H2.0 of Drosophila to which theAHoxl homeodomain shows 70% identity, thoughoutside the homeodomain, there is no significantsimilarity. AHoxl also lacks the sequences correspond-ing to the M repeat and the YPWM peptide, which areconserved upstream of the homeobox in many of thehomeobox genes of Drosophila and vertebrates.

To see whether the Halocynthia genome containsother genes possessing a similar homeobox to that ofAHoxl, Southern genomic hybridization underreduced stringency conditions was carried out using theDNA fragment from the clone 62 as a probe, whichcontained the last one-third of the AHoxl homeobox.As shown in Fig. 4A, a 2.7 kb band was detected in theZscoRI-digested genomic DNA as expected. This andintense bands seen in the Hindlll- and BglH-digestedDNA correspond to the AHoxl gene. Other thanthese, no significant band was detected. Upon Hindllldigestion, two bands were detected as shown moreclearly in Fig. 4B. These results suggest that two copiesof AHoxl are present in the genome, possiblyrepresenting two allelic forms, and it is unlikely that agene containing a similar homeobox to that of AHoxl ispresent in the Halocynthia genome.

Expression of AHoxlWe have examined expression of the AHoxl gene in

824 H. Saiga and others

1

176

20151

45226

70301

95376

120451

14 5526

170601

195676

220751

245826

270901

295976

3201051

3451126

3701201

3951276

4201351

4451426

4701501

4951576

1651

5451726

5701801

5951876

6201951

6452026

6702101

6952176

7202251

2326

2401

2476

TTT

ATA

TACA

kAAA

kAAA

CTGC

VGTA

AAT

FTTT

RACA

DGAC

LTTG

TACC

PCCA

QCAA

FTTC

QCAA

DGAT

NAAT

FTTT

kAAG

EGAG

ACA

WTGG

DCAT

FTTT

ATG

RCGA

VGTC

NAAT

DGAT

CGA

ATA

GTA

TGA

CTC

STCC

HCAT

kAAC

RAGA

PCCA

GTT

STCA

VGTT

DGAT

kAAA

EGAA

RAGA

PCCC

YTAC

YTAC

VGTA

AGCC

STCC

NAAT

DGAC

GCG

PCCA

ACCT

EGAG

SAGC

kAAG

AGCA

RAGA

HCAC

TTC

GCA

TTA

AAT

GCA

LTTA

IATC

kAAG

SACC

IAAA

AAC

STCC

IATT

HCAT

NAAC

DGAT

SACT

YTAT

RAGA

STCT

EGAC

LTTC

DGAC

STCA

NAAT

AAA

FTTT

PCCA

NAAC

SAGT

WTGG

IAAA

RCGG

EGAA

CTC

ATC

AAT

CAT

TTA

GGGA

EGAA

QCAA

kAAG

GGCA

CCA

SAGC

DGAC

IATT

kAAA

QCAG

FTTT

LTTG

EGAA

PCCA

EGAG

STCT

IAAA

GGGA

NAAT

TTA

AGCA

HCAC

PCCA

STCC

NAAT

PCCG

HATC

IATT

ATT

AAC

TAC

TTA

AAA

GCCA

EGAA

NAAC

kAAA

IATA

TTG

STCA

IAAG

NAAT

RAGA

QCAG

STCT

RAGG

AGCA

DGAT

GGGC

YTAT

GGGG

MATG

PCCT

TTA

AGCA

YTAT

YTAT

QCAA

RCGA

EGAA

kAAG

EGAA

CCA

AAC

GTA

TTG

AGC

LCTC

ECAA

RAGA

DGAT

STCT

C

AGT

EGAA

RCGG

QCAA

LCTT

IATA

STCT

AGCT

NAAC

AGCA

DGAC

GGGT

IATA

SAGC

QCAA

GAA

STCA

GGGA

FTTT

NAAC

AGCA

RAGA

WTGG

IAAA

AAA

AAT

TTT

ACT

HATG

RCCA

RAGA

FTTT

SAGC

NAAC

AAA

AGCA

kAAG

DGAT

RCGA

SAGT

CGGA

YTAT

VGTT

SAGT

kAAG

RCGA

QCAG

FTTT

TACG

NAAC

VGTT

AGCG

FTTC

PCCA

VCTT

RAGG

RAGG

EGAA

TCA

AAA

GTA

TCT

EGAA

kAAA

TACC

EGAA

TACC

ACCC

TAT

SAGT

SAGC

NAAC

LCTT

IACC

LTTG

YTAC

EGAA

TACT

EGAA

IAAG

AGCT

DGAT

IATT

EGAA

RAGA

AGCG

NAAT

HCAT

FTTC

IAAC

QCAA

IAAA

TGC

CTG

CAC

ATT

iAAA

STCA

SAAT

RACG

VGTT

GGGA

TTC

DGAC

AGCG

CTCT

DGAC

RAGA

QCAA

NAAT

NAAT

FTTC

AGCA

DGAC

C,GGT

FTTC

GGGG

GCT

IAAA

QCAA

EGAG

STCT

SAGC

CTC

EGAG

TACG

CGC

CAA

C

ATA

,1ATG

STCC

NAAC

YTAT

LTTG

RAGG

AAG

SAGT

DGAT

DGAT

SACC

VGTG

DGAT

IATT

DGAT

YTAC

NAAC

YTAC

SAGT

PCCC

SAGC

AAA

SAGT

VGTT

LCTT

ECAA

LCTT

AGCG

IATA

QCAA

CAC

TCC

TCT

HCAC

SACT

IATC

STCT

kAAG

1ATT

CCA

IAAA

STCA

LTTA

SAGT

EGAA

PCCA

QCAA

NAAT

PCCT

IAAA

KAAC

YTAT

PCCT

EGAA

TTG

YTAC

NAAT

NAAC

TACA

MATG

DGAT

kAAA

TACC

CGA

ATC

AAT

STCA

1ATT

TACA

SAGT

FTTT

QCAG

TCG

STCA

DGAC

AGCC

EGAA

MATC

FTTT

kAAA

YTAC

IATA

LTTC

GGGT

YTAC

IAAA

YTAC

CCT

DCAT

PCCG

TACA

QCAA

QCAA

ACCC

HATG

PCCT

ACC

AAC

TAT

kAAA

PCCC

TACA

STCA

GGGA

DGAT

sTCTC

TGC

VTGC

SAGC

AGCT

MATC

AGCA

YTAT

IAAG

RAGA

FTTC

LCTG

QCAA

FTTC

QCAA

TAT

PCCA

TACA

LTTG

NAAC

RCGT

LCTA

kAAA

STCT

CAG

AAG

CTA

STCT

TACC

MATG

AAC

IATT

EGAA

AAT

CTGT

SAGT

TACA

LCTT

NAAT

IAAA

IATT

TACT

TACT

IAAA

STCA

TACT

GGGA

IATT

GGGG

AGCA

AGCA

DGAC

RAGA

RCGT

SAGT

NAAT

DGAT

ACA

AGA

TTC

VCTT

LTTC

ACCA

HCAT

DGAC

RAGA

GAC

TACA

EGAA

LCTT

NAAC

HCAC

NAAT

YTAC

EGAG

EGAG

DGAT

TACC

GGGA

SAGT

NAAT

NAAC

LTTG

GGGC

TACT

SAGT

GGGA

LTTG

RCGC

ECAA

GTA

TTA

ATA

STCA

AGCA

MATG

ACCC

SAGC

FTTC

QCAG

KAAT

DGAC

EGAG

DGAC

QCAA

LCTA

HCAC

EGAA

DGAT

LCTT

STCT

EGAA

NAAC

TACA

ATC

RAGA

NAAC

TACC

EGAA

LCTT

TACC

CGGT

GGGG

CCA

TTA

CAT

PCCT

VGTA

IAAC

QCAA

IATA

TACA

TTA

NAAC

AGCA

QCAA

LCTT

AGCC

PCCT

IAAA

STCA

SAGT

CTCT

QCAA

RAGA

NAAC

QCAA

GTA

sTCC

NAAT

GGGT

EGAA

EGAA

DGAT

LCTG

EGAA

ATA

TAT

ACA

VGTC

pCCA

RAGA

EGAA

LCTT

ECAG

CGGT

NAAT

TACT

SACT

TACA

FTTC

NAAC

IAAC

LTTG

KAAT

VGTC

TACC

NAAT

STCA

ICGA

TACA

YTAT

FTTC

FTTC

VGTA

IAAA

AGCA

VGTG

VGTG

AAG

TTT

GCT

PCCA

EGAA

RCGC

QCAA

kAAA

AGCA

ACCA

EGAA

CGGA

RCGT

PCCA

AGCC

AGCA

LTTG

RAGG

GGGG

SAGC

STCA

DGAT

STCA

SAGC

QCAA

FTTC

AAG

LCTT

RCGC

SAGT

QCAG

CCA

IATC

TTC

TAT

AGT

FTTC

CTGC

LTTC

STCT

NAAT

CTCT

RCCA

EGAA

NAAC

IATT

YTAT

RAGA

FTTC

LTTA

STCT

SAGC

VGTT

GGGG

STCT

TACC

MATG

TAT

SAGT

STCT

SAGT

GCGT

FTTT

VCTT

ICTA

NAAC

ACT

ATC

ACA

NAAC

EGAA

LTTA

STCA

IAAA

TACT

RAGA

AGCC

ECAA

VGTT

DGAT

ECAA

LCTG

HAAT

PCCG

RAGA

GGGC

NAAT

LCTT

DGAC

IAAA

CCA

RCGA

HATG

RCGT

TACT

QCAA

IAAA

HCAT

VGTC

AAT

ATA

TTT

sAAC

STCA

DGAT

ECAC

NAAC

NAAC

TACT

BCCT

ACCA

AGCT

QCAA

NAAC

PCCC

RCGA

SAGC

NAAT

DGAC

RCGT

AGCG

kAAG

DGAT

RAGA

FTTC

LTTA

QCAA

VGTC

SAGT

1ATC

IATT

DGAT

TTC

CAT

TTT

SACC

MATG

PCCC

ECAA

ACCA

TACA

AGCT

YTAC

DCAC

LTTG

LTTG

NAAC

FTTT

NAAT

EGAA

LTTG

RAGA

FTTT

GGGA

PCCA

NAAT

CCT

NAAT

PCCG

YTAC

kAAA

QCAA

WTCG

LCTT

TAA

ACG

ATT

ACT

NAAC

CCGT

QCAA

NAAT

ECAA

SACC

TACA

kAAC

DGAT

EGAA

STCA

ECAA

YTAC

DGAC

TACA

kAAA

LTTA

LCTA

PCCA

EGAG

LTTA

TTT

NAAT

GGGT

CCGC

kAAA

kAAA

FTTC

GCGT

TGA

GTT

ATT

TTT

NAAC

NAAC

NAAT

FTTT

kAAA

STCC

sACTY

TAT

PCCT

IATA

RCCT

UTCG

HATG

RAGG

kAAA

VGTC

STCA

NAAT

LCTC

QCAA

LCTC

TCA

SAGT

NAAT

HCAT

RACA

YTAC

QCAG

QCAA

AAA

TCC

CTA

75

19150

44225

69300

94375

1 19450

144525

1 69600

194675

219750

244825

269900

294975

3191050

3441125

3691200

3941275

4191350

4441425

4691500

4941575

5191650

1725

5691800

5941875

6191950

6442025

6692100

6942175

7192250

7422325

2400

2475

2500

Fig. 2. The structure of the homeobox gene AHoxl of H. roretzi- The nucleotide sequence and deduced amino acidsequence of AHoxl are shown, numbered left to right. The amino acids are numbered starting with the putative initiatingmethionine residue. The homeobox region indicated by underlining is interrupted by two introns (arrowheads).

A homeobox gene of an ascidian 825

AHoxi RKWNRAVFSLMQRRGLEKSFQSQKYVAKPERRKLADALSLTDAQVKIWFQNRRMKWRQEI 100 %

H2.0 -S-S NL—K IQ—Q IT—D AR-N V HTR 7 0Antp —RG-QTYTRY-TLE E-HFNR-LTRRR-IEI-H--C — ER-I KK-N 46 .6Scr T-RQ-TSYTRY-TLE E-HFNR-LTRRR-IEI-H—C--ER-I LKK-H 43.3Abd-B VRKK-KPY-KF-TLE E-L—A—S-QK-WE — RN-Q--ER NKKNS 50Dfd P-RQ-TAYTRH-ILE E-HYNR-LTRRR-IEI-HT-V-SER-I KKDN 38.3Lab NNSG-TN-TNK-LTE E-HFNR-LTRAR-IEI-NT-Q-NFN QKKRV 4 0Eve VRRY-TA-TRE-LGR E-YKEN — SR-R-CE—AQ-N-PESTI-V DKRQR 4 0En E-RP-TA—SE-LAR-KRE-NENR-LTERR-QQ-SSE-G-NE--I K-A-IKKST 36.7Prd QRRC-TT—AS-LDE—RA-ERTQ-PDIYT-EE--QRTN—E-RIQV--S ARL-KQH 33.3

Fig. 3. Comparison of the homeodomain of AHoxl with those of the Drosophila homeobox genes. The identical residuesare shown by a dash (—). Similarity between the homeodomains of AHoxl and other homeobox genes is indicated by thepercentage on the right margin.

M E H Bg

kb

21.2 -

6 . 0 2 ^

3.53 -

2 . 0 3 "

1.38 —

Fig. 4. Southern blot hybridization of Halocynthia genomicDNA with the AHoxl probe. (A) 2^g of DNA isolatedfrom one adult individual was digested with EcoBI (E),Hindlll (H), and BgUl (Bg), and transferred to a nylonmembrane and hybridized with a 250 bp genomic DNAfragment encompassing the second intron-exon boundaryand the end of the homeobox. The size marker (M) wasend-labelled ADNA fragments digested with EcoBI andHindlll. B is an autoradiogram from the same filter as inA but with a shorter exposure to show more clearly thetwo bands in lane H in A.

early embryos, juveniles and adult tissues. Poly(A)RNAs were prepared from embryos at various stagesand subjected to northern blot hybridization. As shownin Fig. 5, a single band of 2.8 kb in length was detected.During early development, the 2.8 kb transcript wasscarcely detectable (Fig. 5A). The multiple bands seenin the higher molecular weight region in Fig. 5A may benon-specific hybridization to abundant classes ofmRNAs which became apparent due to a long exposuretime. When embryos reached the larval stage, the levelof expression of AHoxl obviously increased.

In Fig. 5B, AHoxl expression was followed injuveniles up to approximately 2 weeks after metamor-phosis. The level of AHoxl expression increased asdevelopment progressed up to day 7 and remained highthereafter.

The expression of AHoxl in various adult tissues wasalso examined. As shown in Fig. 5C, intense expressionwas detected in digestive tract and digestive gland, andto a lesser extent in endostyle. All of these organs are ofendodermal origin. In addition, a considerable level ofAHoxl transcription was detected in coelomic cells,which are thought to be derived from embryonicmesenchyme cells. Very weak expression was alsodetected in pharyngeal epithelium (ectodermal tissue)and in gonad and body wall muscle (mesodermaltissues).

Following this, we carried out in situ hybridization tolocalize the site of expression on sections from juveniles7 days after metamorphosis. As shown in Fig. 6A,B,the AHoxl expression was detected in endodermal cellsthat were differentiating to form the digestive system.Hybridization signals were observed only on theepithelium of the digestive system, but on both theoutside surface, which faces to the future pharyngealepithelium, and the inside surface, which faces thecoelom. No accumulated grains above backgroundwere observed in epidermal cells, the nervous system orpharyngeal epithelium (Fig. 6C,D). At this stage,hybridization signals did not accumulate on coelomiccells (Fig. 6C,D). In situ hybridization with the senseRNA probe showed no significant accumulation ofgrains in those particular regions (Fig. 6E,F).

Discussion

Homology researches employing zoo-blot hybridizationhave demonstrated that the genomes of a variety ofanimals including ascidians contain regions cross-reacting with the Drosophila Antp-type homeobox(McGinnis et al. 1984; Holland and Hogan, 1986).However, little was known about the homeobox genesof ascidians. One exception is an Antp-typ& homeoboxgene isolated from the genomic DNA of Phallusiamammilata (Gehring, personal communication),though it is uncertain whether this Phallusia homeoboxgene is expressed in embryos, larvae, juveniles andadult organisms. In the present study, we have isolatedan ascidian homeobox gene AHoxl from H. roretzi-

The nucleotide sequence of AHoxl shows only alimited degree of similarity to the Antp-type homeoboxsequence, though we used the Antp-type homeobox for

826 H, Saiga and others

A BGo F 16 64 Ga N eTB mTb La La 1 3 5 7 9 11 13

V

- 3.53 -* * '• * flu

-2 .03 _

_0.95

cGo Ph Dt Dg En Mu Be

kb

- 3.53

_ 2.03

_ 0.95

Fig. 5. Expression of AHoxl examined by northern blot hybridization.(A) Northern blot of 10 fig of poly(A) RNA from gonad which includesoocytes, sperm and somatic cells (Go), fertilized eggs (F), 8- and 16-cellembryos (16), 64-cell embryos (64), gastrulae (Ga), neurulae (N), early-tailbud embryos (eTb), middle-tailbud embryos (mTb) and larvae (La).The blot was hybridized with a AHoxl probe (nucleotides 1504-1760 inFig. 2) and washed in O.lxSSPE at 60°C. (B) 10//g of poly(A) RNAisolated from larvae (La) and juveniles at 1 day (1), 3 days (3), 5 days(5), 7 days (7), 9 days (9), 11 days (11) and 13 days (13) aftermetamorphosis, was loaded on the lanes indicated. The blot washybridized and washed as in A. (C) 10 fig of poly(A) RNA from gonad(Go), pharyngeal epithlium (Ph), digestive tract (Dt), digestive gland(Dg), endostyle (En), body wall muscle (Mu) and coelomic (blood) cells(Be) of adult animals was hybridized with the AHoxl probe. Washingwas carried out as in A. The size marker used was end-labelled ADNAfragments digested with EcoRl and HindUl. Note that exposure wascarried out for 10 days with A and for 20 h with B and C. Therefore theintensities of the bands for larvae or gonad poly(A) RNA are different inA, B and C, so can be used as controls for comparison of the relativeintensity of AHoxl expression between the blots.

the screening. Two uninterrupted matches to the probe,one of 14 base pairs and the other of 10 base pairs,present in the last one-third of the homeobox regionenabled us to isolate AHoxl. On the contrary, theisolated homeobox gene has an unexpectedly similarhomeodomain to that of H2.0 of Drosophila. Since ithas recently been pointed out that the AntR-typehomeobox genes clustering on the chromosome(s) ofDrosophila and vertebrates have a common ancestor(Graham et al. 1989), and both Antp- and en-typehomeobox genes have been isolated in the sea urchingenome (Dolecki et al. 1986; Dolecki and Humphreys,1988), one would expect an ascidian to also possessgenes containing the Antp-type homeobox. However,we have not been successful in isolating an Antp-typehomeobox gene from the H. roretzi genome using theAntp homeobox of a silkworm, a sea urchin orDrosophila and the Scr homeobox of Drosophila asprobes. At the present stage, a possible explanation forthis may be the difference in codon usage between theprobes and the H. roretzi genome so that isolation byvirtue of cross hybridization was not possible.

Expression of AHoxl was very restricted in earlyembryos, but became obvious during the later stages ofdevelopment. Transcripts were evident in larvae andjuveniles. The results of in situ hybridization with the

juveniles suggest that the primary site of expression is inendoderm. There are considerable variations amongascidian species in the patterns of temporal differen-tiation in juvenile tissues. H. roretzi is one of the speciesthat exhibit retarded differentiation. Although theendoderm is evident in embryos at later stages, it stillappears to contain undifferentiated cells. Differen-tiation of the endoderm into the digestive tract beginsduring metamorphosis (e.g. Mita-Miyazawa et al.1987). Such temporal development of the endodermseems to be compatible with the developmentalexpression pattern of AHoxl. Furthermore, relativelylarge amounts of the AHoxl mRNA were found inadult endodermal tissues such as the digestive tract anddigestive gland. These results suggest that the AHoxlexpression is associated with differentiation of theendodermal tissues. Together with the similarity of theAHoxl homeobox to that of Drosophila H2.0, theseobservations are reminiscent of those concerning theexpression of the Drosophila H2.0 homeobox gene,which is restricted to visceral mesoderm (Barad et al.1988). Although the function of H2.0 has not beenidentified either, it is possible that the function ofAHoxl in endoderm of a deuterostome is similar to thatof H2.0 in visceral mesoderm of a protostome.

On the other hand, intense expression of AHoxl was

A homeobox gene of an ascidian 827

observed in coelomic or blood cells. In ascidians, six tonine different types of blood cells are found within theopen circular system (reviewed by Goodbody, 1974;Wright, 1981). Each type of blood cell has a character-istic morphology. Various functions such as coagu-lation, nutrition, immune responses, heavy metalaccumulation, gonad and germ cell formation, and

tunic formation have been attributed to either one orseveral types of blood cells. In spite of extensive studieson the structure and function of ascidian blood cells,little is known about their embryological origin, thoughit is suggested that coelomic cells originate fromembryonic mesenchyme cells (Cowden, 1968). Recentstudies have shown that the A7.6 cells of the 64-cell

bs ns

as

ph

en

ep

ds. /

CO

tu

r Ph

ep

cods

tu

CO

ep

cite"

Fig. 6. In situ hybridization to sectioned specimens of 7-day-old juveniles with antisense (A-D) and sense RNA probes (E,F). as, atrial siphon; bs, buccal siphon; co, coelom; ds, digestive system; en, endostyle; ep, epidermis; ns, nervous system;oc, ocellus; ot, otolith; ph, pharynx; tu, tunic. Scale bars, 100 pm. (A) Light photomicrograph of a section through themidline of the juvenile and (B) dark-field image of A. Hybridization signals above background are evident only on theepithelium of the digestive system (ds). (C) Light photomicrograph of paracross section of the juvenile and (D) dark-fieldimage of C, showing the grains are restricted to the epithelium of the digestive system (ds). (E) Light photomicrograph ofparasagittal section of the juvenile and (F) dark-field image of E. By contrast to the antisense probe, hybridization with thesense probe did not show any significant accumulation of grains. Bright spots down left in the photograph are nonspecific,irrespective of grains.

828 H. Saiga and others

embryo of H. roretzi give rise to the trunk-lateral cells(TLCS) of a tadpole larva (Nishida and Satoh, 1985;Nishida, 1987) and that a monoclonal antibody specificto TLCs stains coelomic cells of juvenile and adultbasophilic blood cells, suggesting the developmentalrelationship between TLCs and a type of blood cell(Mishide et al. 1989). At present, however, we are notcertain which type(s) of blood cells expresses AHoxl orof the embryonic origin of the blood cells with AHoxlexpression. In mammalian hematopoietic cell lineages,it has been reported that a variety of homeobox genesare expressed and their expression patterns are differ-ent depending on the cell lineage (Kongsuwan et al.1988; Shen etal. 1989). The cell lineage in which AHoxlis expressed is also a subject for further research.

The authors thank Drs W. Hara and Y. Suzuki, G. Doleckiand T. Humphreys and A. Kuroiwa for kindly providing uswith pBM3.6, the genomic clone of Bombyx mori Antp-Xyyzhomeobox, pHBl, the sea urchin homeobox clone andDrosophila homeobox probes, respectively. Thanks are alsodue to Drs W. R. Jeffery and B. Swalla for technical advicewith in situ hybridization, and we are grateful to Drs T.Matsui, M. Nomoto and A. Ison for comments and criticalreading of the manuscript.

This work was supported by Grants-in-Aid for PriorityArea No. 62124047 and No. 02221101 from Ministry ofEducation, Science and Culture of Japan to H.S.

References

AKAM, M. (1987). The molecular basis for metameric pattern inthe Drosophila embryo. Development 101, 1-22.

BARAD, M., JACK, T., CHADWICK, R. AND MCGINNIS, W. (1988).

A novel, tissue-specific, Drosophila homeobox gene. EMBO. J.7, 2151-2161.

BLIN, N. AND STAFFORD, D. W. (1976). A general method forisolation of high molecular weight DNA from eukaryotes.Nucleic Acids Res. 3, 2303-2308.

CONKLIN, E. G. (1905). The organization and cell lineages of theascidian egg. / . Acad. natn. Sci., Philad. 13, 1-119.

COWDEN, R. R. (1968). The embryonic origin of blood cells in thetunicate Clavelina Ricta. Trans. Am. microsc. Soc. 87, 521-524.

DOLECKI, G. J. AND HUMPHREYS, T. (1988). An engrailed classhomeo box gene in sea urchins. Gene 64, 21-31.

DOLECKI, G. J., WANNAKRAIROJ, S., LUM, R., WANG, G., RJLEY,

H. D., CARLOS, R., WANG, A. AND HUMPHREYS, T. (1986).Stage-specific expression of a homeo box-containing gene in thenon-segmented sea urchin embryo. EMBO J. 5, 925-930.

GEHRING, W. J. (1987). Homeo boxes in the study ofdevelopment. Science 236, 1245-1252.

GOODBODY, I. (1974). The physiology of ascidians. Adv. MarineBiol. 12, 1-149.

GRAHAM, A., PAPALOPULU, N. AND KRUMLAUF, R. (1989). The

murine and Drosophila homeobox gene complexes havecommon features of organization and expression. Cell 57,367-378.

HOLLAND, P. W. H. AND HOGAN, B. L. M. (1986). Phylogeneticdistribution of Antennapedia-like homeo boxes. Nature 321,251-253.

INGHAM, P. W. (1988). The molecular genetics of embryonicpattern formation of Drosophila. Nature 335, 25-34.

JEFFERY, W. R. (1985). Specification of cell fate by cytoplasmicdeterminants in ascidian embryos. Cell 41, 11-12.

KONGSUWAN, K., WEBB, E., HOUSIAUX, P. AND ADAMS, J. M.

(1988). Expression of multiple homeobox genes within diversemammalian haematopoietic lineages. EMBO J. 7, 2131-2138.

MANIATIS, T., FRITSCH, E. F. AND SAMBROOK, J. (1982). Molecular

Cloning. Cold Spring Harbor Laboratory.MCGINNIS, W., GARBER, R. L., WIRZ, J., KUROIWA, A. AND

GEHRING, W. J. (1984). A homologous protein-coding sequencein Drosophila homeotic genes and its conservation in othermetazoans. Cell 37, 403-408.

MITA-MIYAZAWA, I., NISHIKATA, T. AND SATOH, N. (1987). Cell

and tissue-specific monoclonal antibodies in eggs and embryos ofthe ascidian Halocynthia roretzi. Development 99, 155-162.

NISHIDA, H. (1987). Cell lineage analysis in ascidian embryos byintracellular injection of a tracer enzyme. III. Up to the tissue-restricted stage. Devi Biol. Ill, 526-541.

NISHIDA, H. AND SATOH, N. (1983). Cell lineage analysis inascidian embryos by intracellular injection of a tracer enzyme. I.Up to the eight-cell stage. Devi Biol. 99, 382-394.

NISHIDA, H. AND SATOH, N. (1985). Cell lineage analysis inascidian embryos by intracellular injection of a tracer enzyme.II. The 16- and 32-cell stages. Devi Biol. 110, 440-454.

NISHIDE, K., NISHIKATA, T. AND SATOH, N. (1989). A monoclonal

antibody specific to embryonic trunk-lateral cells of juvenile andadult basophilic blood cells. Dev. Growth Differ. 31, 595-600.

ORTOLANI, G. (1955). The presumptive territory of the mesodermin the ascidian germ. Experimentia 11, 445—446.

QUIAN, Y. Q., BILLETER, M., OTTTNG, G., MULLER, M., GEHRING,W. J. AND WUTHRICH, K. (1989). The structure of theAntennapedia homeodomain determined by NMR spectroscopyin solution: Comparison with prokaryotic repressors. Cell 59,573-580.

SATOH, N. (1987). Towards a molecular understanding ofdifferentiation mechanisms in ascidian embryos. BioEssays 7,51-56.

SATOH, N., DENO, T., NISMDA, H., NISHIKATA, T. AND MAKABE,K. W. (1990ft). Cellular and molecular mechanisms of musclecell differentiation in ascidian embryos. Int. Rev. Cytol, 122,221-258.

SATOH, N., NISHJKATA, T. AND MAKABE, K. W. (1990a). Eggcytoplasmic components responsible for larval muscle celldifferentiation of ascidian embryos. In Developmental Biology(ed. E. Davidson, J. Ruderman and J. W. Posakony) NewYork: Wiley-Liss, Inc.

SCOTT, M. P., TAMKUN, J. W. AND HARTZELL m , G. W. (1989).

The structure and function of the homeodomain. Biochim.biophys. Acta 989, 25-48.

SHEN, W.-F., LARGMAN, C , LOWNEY, P., CORRAL, J. C , DETMER,K., HAUSER, C. A., SIMOMTCH, T. A., HACK, F. M. AND

LAWRENCE, H. J. (1989). Lineage-restricted expression ofhomeobox containing genes in human hematopoietic cell lines.Proc. natn Acad. Sci. U.S.A. 86, 8536-8540.

TOMLINSON, C. R., BATES, R. L. AND JEFFERY, W. R. (1987).

Differential expression of a muscle actin gene in muscle celllineages of ascidian embryos. Development 101, 751-765.

ULLRICH, A., SHINE, J., CHIRGWIN, J., PICTET, R., TISCHER, E.,

RUTTER, W. J. AND GOODMAN, H. M. (1977). Rat insulin genes:Construction of plasmids containing the coding sequences.Science 196, 1313-1319.

UZMAN, J. A. AND JEFFERY, W. R. (1986). Cytoplasmicdeterminants for cell lineage specification in ascidian embryos.Cell Differ. 18, 215-224.

VENUTI, J. M. AND JEFFERY, W. R. (1989). Cell lineage anddetermination of cell fate in ascidian embryos. Int. J. Devi. Biol.33, 197-212.

WHITTAKER, J. R. (1987). Cell lineages and determinants of cellfate in development. Am. Zool. 27, 607-622.

WRIGHT, R. K. (1981). Urochordates. In Invertebrate Blood Cells,vol 2, pp. 565-626, (ed. N. A. Ratcliffe and A. F. Rowley)Orland, Miami: Academic Press.

(Accepted 12 December 1990)