Type IV collagen in sponges, the missing link in basement membrane ubiquity

Bid Cell (1996) 88. 374 0 Elsevier, Paris

37

Original article

Type IV collagen in sponges, the missing link in basement membrane ubiquity*

Nicolas Boute a, Jean-Yves Exposito a, Nicole Boury-Esnault b, Jean Vacelet b, Nobuhiro Noro c, Koyomi Miyazaki c,

Katsutoshi Yoshizato c, Robert Garrone a**

aInstitut de Biologie et Chimie des Prote’ines, CNRS-UPR 412, Universite’ Claude Bernard, 7, passage du Vercors, 69367 Lyon cedex 07; b Centre Oceano lo i g q ue de Marseille, Universite’ de la Mediterranee,

URA-CNRS 41, Station marine d’Endoume, 13007 Marseille, France; c Yoshizato Morphomatrix Project ERATO, JST, 13-26, Kagamiyama 3-chome, Higashi, Hiroshima 739, Japan

(Received 2 December 1996; accepted 27 January 1997)

Summary - Basement membrane structures, or their main component, type IV collagen, have been detected in all multicellular animal species, except sponges. We cancel this exception by the demonstration of type IV collagenous sequences in a new marine sponge species by cDNA and genomic DNA studies. One of these sequences is long enough to demonstrate the specific characteristics of type IV collagen chains. The 12 cysteines are at conserved positions in the carboxyl-terminal non-helical NC1 domain, as are the interruptions in the carboxyl-terminal end of the triple helical domain. The gene organization of the region coding for the NC1 domain is similar to that of the human genes COLAAZ, COUA4 and COL4A6. An additional, shorter sequence suggests the presence of a second chain. The expected tissue localization of this collagen has been confirmed using polyclonal antibodies raised against a sponge recombinant protein. These results demonstrate that type IV collagen is representated in all animal phyla. It is actually the only known ubiquitous collagen and it has at least two different alpha chains in all the species where it has been characterized.

collagen / basement membranes / cysteine / evolution / porifere

Introduction

Basement membranes are sheet-like complexes of extracellular matrix proteins, highly structured and containing spe- cifically type IV collagen as scaffold [2, 4, 23, 35, 36, 371. They underlie epithelial, endothelial tissues, most glial cells and they surround several mesenchymal cells and especially all the muscle cell types [21, 261. Basement membranes exert not only a mechanical function as supporting structures, but they also play a major biological role as molecular sieves and in the control of cell differentiation and stability [28, 341. The latter function has been related to the modular organization of most of its components [lo]. The type IV collagen molecule is a long triple helix containing several interruptions, which is characterized by a specific carboxyl- terminal, non-collagenous NC1 domain [2, 4, 23, 35-371. Among the six different (XI chains (numbered 1 to 6) of type IV collagen it is generally assumed (although not definitely demonstrated) that the more frequent molecular stoichiome-

* The nucleotide sequence data reported in this paper have been submitted to the GenBank/EMBL Data bank with accession numbers X95815 (mRNA) and X95816 (gene) ** Correspondence and reprints Abbreviations: NCI, carboxyl-terminal non-collagenous domain; RT-PCR, reverse transcriptase polymerase chain reaction; UTR, untranslated region; IPTG, isopropylthio-ED galactoside.

try is of two al, a3 or a5 chains for one a2, a4 or ar6 chain, respectively [ 181. The corresponding human genes are found in pairs (COIAAlCOL4A2 on chromosome 13; COLAASCOIAA4 on chromosome 2; COLAAS-COLAA on chromosome X) and have head-to-head orientations with overlaping promoter regions [ 183, This special organization of the genes and their homologies has led to an evolutionary hypothesis from an ancestral gene involving first a duplica- tion/inversion, differentiation of two gene families (the even and uneven genes) and successive duplications [25]. Early ultrastructural studies clearly showed the presence of basement membrane structures in all multicellular organisms, except in the most ‘primitive’, the sponges [ 151. More recent studies have confirmed the wide distribution of type IV collagen in invertebrates, for example in Drosophila [9, 141, sea urchins [ 131 and nematodes [22] and the frequent occur- rence of other basement membrane components such as laminin [4, 19, 241. In this context, the absence of typical basement membranes in sponges, which possess well-defined epithelia [29], is unexpected. This apparent lack of collagen IV is especially puzzling as sponges contain an abundant extracellular matrix composed of several collagen types [1, 11, 12, 16, 301. There is, however, a group of marine sponges, Homoscleromorpha, with reduced skeleton content, which possess a structure partially ressembling the morphology of basement membranes [15, 161. We have looked for collagen IV in Pseudocorticium jarrei, a newly described homoscleromorph sponge devoid of inorganic skeleton [ 61.

Materials and methods

Materials and preliminary procedures

The sponges were collected by scuba diving in littoral caves in the western Mediterranean Sea (France) and immediately frozen in liquid nitrogen. Total RNA, poly(Aj+ RNA and DNA extracts were classically prepared and characterized as described previously [ 11, 121. A cDNA library was prepared using the. kit ‘cDNA synthesis system plus RPN.lZ%YPZ’ (Amersham). A total of 106 phages per ml library were obtained. The cloning procedure used the kit ‘cDNA cloning system ,%gt IO’ (Amersham).

RT-PCR umpl$ication

Two separate experiments have been performed. In the first, the sequences of the primers used to obtain the screening probe by RT-PCR amplification were: SGCIGT(T/C)CACTCCCAAGA 3’ (in 5’ position) and 5’ TCCA(A/G)GGCAGGAICCAGG 3’ (in 3’position). The partial sequence of the genomic DNA was also obtained by PCR using the oligonuleotides 5’ GGACCTGGA- GACGACGGATC 3’. located within the triple helical-coding sequence, and 5’ GTGCAGTGCTCTACCCTACC 3’, located within the UTR sequence, as 5’ and 3’ primers, respectively. The second series of experiments was undertaken with the aim to clone sponge matrix metalloproteinases. In this case, the sequences of the primers were: S’TG(TIC)GGTGTIC- CIGA(T/C)GT3’ (in 5’ position) and S’ICCIGGICC (A/G)TC(A/G)AA3’ (in 3’ position).

Computer unalysis

The sequences were analyzed by the computer program ‘Anthe- prot’ [ 171. A phylogenetic tree was realized using the ‘Clustal W’ method 1331.

Recombinant peptide production und purificcttion

A derivative of pT7-7 plasmid [32] modified by the addition oi nucieotides encoding six histidine residues [8] was used. A 677-bp cDNA fragment containing the sequence coding for the sponge NC1 domain was synthesized by PCR amplification. The two nucleotides used as primers were S’TATGAATTCACAGCCA GACAACGACATTC S’carrying an EcoRI site and S’TATCTG- CAGGGCTCTCCGCCCCACCCATTC 3’ containing a PstI site. The DNA fragment enclosed within these two restriction sites was isolated and ligated into the modified plasmid. The integrity of the construction was verified by sequencing. The expression of the plasmid was analyzed after transformation in the host E coli strain BL21 (DE3) inducible by IPTG. 3 h post-induction, the bacteria were collected and disrupted by sonication. As the recombingfit protein was insoluble, the pellet obtained by centrifugation was treated by 5 M guanidinium hydrochloride. The centrifugation cleared solution was purified by several runs in a Ni-NTA resin column (Qiagen, Chatsworth, CA). SDS-PAGE of the purified recombinant protein was performed, the corresponding band was transferred on Selex 20 Polypropylene. membrane (Schleicher and Schuel, Dassel, Germany) and sequenced in an Applied Biosystem 473A protein sequencer,

immunologicui clnd miwphologicul proc-e&w.\

Rabbit immunization was performed at the Valbex center in rhc lnstitut Universitaire de Technologie (Lyon, France) by Six subcu- taneous injections of the renatured recombinant protein still ligated to the Ni-agarose beads. The first four injections.contined I50 pg of recombinant protein in complete Freund’s adjuvant for the first and in uncompkte adjuvant for the following; 250 pg in uncomplete adjuvant for the last two injections._ELISA tested immunized rabbit serum was used first in immunoblotting experi - merits to demonstrate the staining of the protein band appearing irl the extract from stimulated transformed bacteria; For immunofluo~~ rescence experiments, performed on sponge frozen sections, the anti-collagen antibodies were revealed with an anti-rabbit I& conjugated to fluorescein isothiocyanate. For electron microscopy. specimens were fixed in 2.5% glutaraldehyde in a mixture of 0.4 M sodium cacodylate buffer and sea water (45, v/v) and post- fixed in 2% osmium tetroxide in sea water. Embedding was realized in Araldite and many1 acetate and lead citrate stained thin sections were observed with an Hitachi HU. 600 transmisSion elec- fron microsdope.

Results and discussion

Obtention of the screening probe

Considering that the sequence flanked by cysteines 7 and X is a highly conserved region of the NC1 domain in type 1V collagen chains, a specific probe corresponding to this sequence has been constructed by RT-PCR arttplification using degenerated~primers. The probe encoded for the poly- peptide V~TS~~V~D~~G~TL~T~F~FLQQTAA- QAEGTmGLEfPm, where the underlined residukq appear conserved in human, mouse, bovine, Drosophila, sea urchin, and nematodes. This probe hasbeen used i& screen a P jurrei cDNA library and yielded several positive clones The sequencing of three of them (PCCl, PCC2, PCC81 revealed a partial type IV collagenous sequence (fig 1). In Northern blotting experiments, the largest insert, used as a probe, hybridized to a sponge mRNA-of 6.9 kb (not shownh This size corresponds to the known sizes of colIagen IV mRNAs (5.3 to 10 kb).

Deduced umino mid sequence of the PCC8 cDNA clone

The PCCS cDNA clone encodes for a part of a type IV coi- lagen chain (PjCol4cl) with 628 residues of the triple helical domain and a 226 residues long carboxyl-terminal NCI domain.The triple helical domain is composed of 206 class& cal collagenous~ Gly-X-Y motifs and five short intermptitins. As shown in figure 2 the last imperfection of the collagenous domain is located 70 residues from the end. of this domain, within an area of conserved interruptions iri all vertebrate and invertebrate collagens IV [23]. Despite the presence of imperfections in the succession of the Gly-X-Y

Fig 1. A. Schematic representation of the Pseudocorticium jarrei type IV collagen (black box, NC1 domain; incomplete white box, sequenced part of the triple helical domain: the imperf~ons are marked by vertical lines) and cDNA clones (PCCl, PcC2 and PCS&f alignment, With their size (in kbase pairs) indicated-on the right. 3’ UTR, unstranslated sequence in 3’; TGA, stop codon; the restriction sites Suu3A (S), BumHI (B) and &ndIII (H) are indicated on the largest clone. B. Deduced amino acid sequence. interruptions in-the triple helical domain are marked below the amino acids by asterisks a&putative glycosylation sites are underlined. The MC1 domain is in&&ted by the boxed area. The black dots in the NC1 domain, located between the residues E and F and Q and Q. respe&veIy, indicate ihe.oosi- tions of the introns in the corresponding nucleotide sequence.

Type IV collagen in sponges 39

triple helical domain I I I

NC1 3’ UTR domain TGA

I

PCCI

TFA

PCC2

0,94kb

1,67kb

s s /’ s BsH,F STGA =- 3,5kb

PCC8

B GITGDSGPRGEKGEEGERGSFGDPGEAGSQGFKGEPGPKGMVGGKGQQGEQ

GRLGGSGPRGTQGPQGDRGSQGTRGESGDQGNAGSAGERGSNGRAGDVGP

PGSNGPAGLKGDDGEEGPDGNPGPKGSPGEEGLPGFMGFKGMKGDVGGQGP

DGDVGENGFKGEPGPA~~GPPGPNGDVGPKGER~~GTLGLKGMKGDPGPVG

ANGSVGESGERGRKGGVGRKGVIGDEGGDGEIGV NGTAGDGGQKGGLGQVGI

LGDQGEKGAEGARGASGMKGVKGEVGPQYFGDQGPAGSPGVKGLSGVAGEE

GPEGPQGPKGEAGPAGPPGPPGGPGPEGLPGINGLQGDEGNEGRPGENGDVG

GKGDVGPQGEQGDSGQRGEVGEKGRKGDFGQEGDRGEPGPF~GNDGRDGF

DGNDGSQGEPGPKGRKGDPGDSGDQGPKGSVGRLGLPGDIGDEGMPGERGR

KGLKGQTGAKGDTGMDGDRGNPGASGVNGINGMTGDQGRQGRDGDEGNEGD

VGEKGNSGPQGSRGPGGDEGDQGITGGQGPKGRRGDVGPQGIRGMG~~~GM

PGLPGDGGNEGGTGDQGPPGPPGPPGDIGEVGVVGAIGEKGGNGSKGDPGIKG

GPGDDGSVGVGGGD);LLLVVHSQTTNIPQCPNDYTRLWVGYSLLQLTGNGLGV

GQDLGDPGSCMPSFHPMPVVRCNPMQRCiFARRKDESYWLSTNATRPPlPVSG

SDlEEHlSRCSVCESNSlSlAVHSQDSNVPDCFPGWVTLWTGFSF~TAAQAEG

TGQGLESPGSCLQHFRSTPFIGCGGRGQCSYDSVSGSYWMIVLDALNPFQDTEP

GTYPVSDIEKRLSRCRVCEWVGRRA

I P~CoMa

_- I I I I I I C. elegans ai (IV)

-- I I I I I I C. elegans a2( IV1

_- I I I I I I I A. suum c&?(tV)

-_ I I I I I m B. makyia2(tVj

I I m I I n I 1 s. purpuratus 3a

-_ n 1 4 II II 4 Human a4(lVj

I I I 4 II m 4 -_ I t I m I Human a6(IVj

-- I m I I I 4 Human al (h/j

-- I n I 4 II 4 I Human a3(tVj

I I -_ I 4 I 4 m I 4 Human a5(IV)

Fig 2. Schematic comparisons of the imperfection locations in the triple helical domain of available type IV collagen chain sequences. Imperfections are marked by rectangles, the length of which is proportional to the size of the imperfection. The starting point is the first Gly residue of the NC1 domain which is strictly conserved in aB the a (IV) chains. Considering that in most vertebrate and invertebrate a (IV) collagen chains there are several residues between this conserved Gly residue and the beginning of the NC1 domain, the slight shift of the sponge chain (PjCol4a) imperfections might be explained by a deletion in this junctional area. The nematodes quoted are Caenorhab- ditis elegans (C elegans); Brugia malayi (B-malayi); Ascuris suum (A suum). The sea urchin species is Strongylocentrotus pu@uratus (S purpuratus).

motifs, type IV collagen triple helical domains are stable. This stability is probably ensured by the unequal distribution of charged or hydroxylabie residues between~position X and Y of the Gly-X-Y motifs [7]. In the PjCol4a triple helical domain, positively charged residues (arginine and lysinej are preferentially present in Y position when negatively charged ones (aspartic acid and glutamic acid) are often in X position. Indeed, PjCol4a has a high aspartic acid content (5 1 residues) and 31 of which are in the second position of the repeated triplet Gly-X-Y. Like the other type IV collagen chains, the carboxyl-terminal NC1 domain contains 12 cysteines at conserved positions and it is also composed of two subdomains sharing 39% homology.

Identification qf a second type IV collagen chain

The second series of experiments, primarily performed to look for the presence of matrixins in sponges allowed the isolation of a clone containing a sequence (not shown) com- parable to a part of the previous NC1 sequence. The deduced protein sequence suggested a 129 amino acid

domain including four cysteines (presumably cystemes L eo 4). As this sequence was too short, it was not used for sequence comparisons.

Comparative sequence analysis

Sequence comparisons showed that 69 of the 79 amino a&s conserved between human, sea urchin, DYosopkila atid r elegans type IV NC1 domains are present in the sponge chain (fig 4). This sponge NC1 domain has about 45--4&% identity with the NC1 domain of the uneven a (IV) group and 46-50% identity with the even group. Unfortunately the differences between the identity rates were not significant enough to determine if PjCo14aris an even or an uneven a (IV) chain. Moreover, the Clustal W mu&i-alignment intro- duced some gaps in the different chains and some oft them were more specific of even or-uneven a (IV) chains (fig4). However, the aligned PjCol4a chain- contains the gaps spe: cific of the two groups of chains. It can be assumed then that PjCol4a is a mixed chain and that the differentiation between the two groups of chains probably occurred afrex


0.051

Human cc6(lV)

C. elegans 02(W)

Drosophila al (IV)

S. purpuratus 4a

C. elegans al (IV)

Mouse a5(IV)

Human cr5(lV) -

S. purpuratus 3a -(PjCol4a) P. jarrei

Fig 3. Phylogenetic tree elaborated with the Clustal W method using the available amino acid sequences of the NC1 domains of several 01 (IV) collagen chains. Two groups (even and uneven) of a: chains are clearly established from which the sponge Pseudo- corticiumjurrei (P jurrei) chain (PjCol4cz) is excluded. The nema- tode and sea urchin species quoted are the same as in figure 2. The length of each branch is proportional to the percentage of diver- gence between the sequences. The scale is indicated on the upper left of the figure.

the separation of the phylum Porifera. This point of view is strenghthened by the fact that a reconstituted phylogenetic tree sets apart the sponge NC1 domain (fig 3).

Gene organization

Partial genomic analysis revealed that the gene organization of the region coding for the NC1 domain is similar to that of the human genes coding for the collagen IV chains CR, a4 and a6 [27, 311. The sponge gene PjCOL4A has two introns in this region, as do the human genes COL4A2, COLAA and COL4A6. The most 3’ intron is exactly at the same location as in other human COL.4 genes and is in addition simi- larly flanked by unsplit codons (fig 5). Those two introns are 59 and 58 pb long and their ends are conserved according to the gt-ag rule.

Protein localization

A recombinant sponge NC 1 domain, tagged with six histines at the carboxyl-terminal end was produced in Escherichia coli and purified by Ni-agarose chromatography. After con-

firming the expected sequence, the renaturated protein was injected into rabbits to obtain polyclonal antibodies. The specificity of these antibodies was checked in immunoblots: they stained several bands corresponding to bacterial proteins and above all the band corresponding to the recombinant sponge protein. The antibodies were then used in immunofluorescence experiments and they stained a thin layer at the internal side of epithelia corresponding to the location of the basement membrane-like structure as seen in transmission electron microscopy (fig 6).

Type IV collagen is the only ubiquitous collagen

Considering the size of the mRNA, the sequence of the NC1 domain and especially the position of the twelve cysteines, the location of the interruptions in the collagenous domain and the gene organization, it can be assumed that a type IV collagen has been partially characterized for the first time in sponges. Moreover, the detection of a second sequence, even partial, suggests that this sponge type IV collagen is made of more than one a chain, as it is a case for all known type IV collagens, since the discovery of a second chain in Drosophila (Yasothomsrikul et al, submitted). In the fresh- water sponge species Ephydatia miilleri, which does not contain basement membrane-like structure, short-chain collagens have been characterized previously [ 121. Interest- ingly, these collagens possess a C-terminal domain sharing homologies with the NC1 domain of type IV collagen [ 121. It cannot be stated that other sponges do not contain basement membrane collagen, however, the presence of a collagen IV in P jarrei suggests that the homoscleromorph sponges, in which a basement membrane-like structure has been observed, do possess a true basement membrane pro- ceeding from this collagen type. It remains to be determined whether the observed structure is only composed of type IV collagen or if it is a complex of collagen and other proteins, especially of laminin. The EDTA extraction of the closely related sponge of the same group, Oscarella tuberculata, yielded abundant tenascin-like structures, but only a few cross-shaped laminin structures [20]. However, unexpected truncated laminin isoforms might have been misrecognized. Biochemical analyses were unsuccessful, this collagen being unextractable as are the other sponge collagens [ 161. The presence of this collagen type and other relatively evolved features as the expression of tenascin-like protein [20], the fact that homosclerophorids are the only sponges possessing a spermatozoon with acrosome [3], and highly structured epithelia [5] indicate a functional link between appearance of basement membranes and epithelial differentiation during evolution [16]. One or another variety of collagen chains is certainly present in all multicellular animal species, however, type IV collagen is the only collagen type constantly represented within each phylum.

Acknowledgments

The authors are greatly indebted to JC Cortay, D Nbgre and A Cozzone for helpful advice in overexpression methods. They gratefully acknowledge the contributions of MM Boutillon during the protein sequencing, C Geourjon in his help in sequence comparisons, Y Descollonges and R Willems in the immunisation experiments and immunofluorescence studies and A Bosch for the photographic artwork. They are grateful to D Hulmes, G Misevic and 0 Popescu for helpful discussions and to R Fleischmajer for reading a draft of this manuscript.

Human cXl(IV) ---SVDHGFLVTRHSQTIDDPQC?SGTK~LYHGYSLLYVQGN~RAHGQDLGTAGS~~~~~ Human a3 (IV) ---ATWTTRGFVFTRHSQTTAIPSCPEGTVPLYSGFSFLFVQGNQ~~GQDLGTLGSC~~Q Human CZ5(IV) ------SVAHGFLITRHSQTTDAPQCPQGTLQVYEGFSLLYHGQDLGTAGS~I,~ s.purpuratus 3a -----.--.S-----.GFFITRHSQTTSIPQCPQGTARMn7HGYSLLFVQGNERGHGQDLGKPGSCLM C.elegans al(W) ------APSRGFTFAKHSQTTAVPQCPPGASQLWEGYSLLYVQGNG~SGQDLGQPGS~L~ Drosophila oll(IV)APAALDYLTGrLrTRHSQSETVPACSAGHTELWTGYSLL~DGNDYAHNQDL-.---GS~~~~ C. elegans a2 (IV! - -----YRDGFVLVKHSQTTEVPRCPEGQTKLWDGYSLI~YIEGNEKSHNQDLGHAGS~~~~ S.purpuratus 4a -----.-.-PNRGHFITRHSQSRNVFSCPAGTVELWRGFSVLFSMGNGH~HQDLGDAGS~~~ Human a2 (IV) ----SVSIGYLLVKHSQTDQEPMCPVGMNKLWSGYSLLYFEGQE~~QDLGLAGS.~L~~ Human a4(IV) - ---EGYLGGFLLVLHSQTDQEPTCPLGMPRLWTGYSLLYLEGQEKAHNQDLGLAG~~~L~ Human a61 IV) -- ~-MRV',YTI~VKHSQSEQVPPCE'IGMSQLWGYSLI,FVEGQE~HNQDLGFAG~~~,~ Pjc014a --- --------GLLLWHSQTTNIPQ~PNDYTRLWVGYSLLQLTGNGLGVGQDLGDPGS~~~

T ,i f * * * A i*i* -C** tx*

4 5 p

Human al(IV) KFSTMPFLFCNINNVCNFASRNDYSYWLSTPEPMPMSMAP~~A~~A Human a3 (IV) RFTTMPFLFCNVNDVCNFASRNDYSYWLSTPALNPMNMAPLEPY~SRCTVCEGPA Human a5(IV) RFSTMPFMFCMINNVCNFASRNDYSYWLSTPEPMPMPMSMQPLKGQSIQPFISRCAVC~A~A S.purpuratus 3a RFSTMPFLFCN~NNVCHVASRNDYSYWLSTTEPMPMNMRPIRGGQLQPFISRCWCEA~A C.elegans al(IV) KFNTMPFMFCNMNSVCHVSSRNDYSFWLSTDEPMTPMMNP~EV~~ Drosophila al(IV)RFSTLPVLSCGQNNVCNYASRNDKTFWLTTNAAIPVENIE~RQYISR~W~~EAPA C.elegans a2(IV) RFSTMPFLF~DFNNVCNYASRNDKSYWLSTSEAIP--MMPERE~EPYISR~~~~EAP~ S.purpuratus 4a RFSTMPFLFCNFNNtrCNYASRNDRSYWLTTNEPLP--MMPLMNQQ~DPYISRCTV~EAPT Human a2 (IV) RFSTMPFLYCNPGDVCYYASRNDKSYWLSTTAPLP--MMPVAEDE~KPYISRCS~~~~?~P~ Human CX4(IV) VFSTLPFAYCNIHQVCHYAQRNDRSYWLASAAPLP--MMPCEA~A Human a6( IV) RFSTMPFIYCNINEVCHYARRNDKSYWLSTTAPIP--MMPSQTQIPQYISR~SV~EAPS PjCol4a SFHPMPWRCNPMQRCEFARRKDESYWLSTNATRP--PIPVSGSDIEEH~SR~SVCES~~S

*A* i * *A* ** A' *** **t ii 7 Y

Human al(IV) MVMAVHSQTIQ-IPPCPSGWSSLWIGYSFVMHTSAGAEGSGQALASPGSCLEEFRSA~-F-F'IE: Human a3 (IV) lAIAVHSQTTDIPPCPHGWISLWKGFSFIMFTSAGSEGTGQALASPGSCLEEF~SPFL~ Human a5( IV) WIAVHSQTIQIPHCPQGWDSLWIGYSFMMHTSAGAEGSGQA~SPGSCLEEFRSAPF~~ S.purpuratus 3a QVLTVHSQTVNIPDCPDRWGVLWlGYSFMMHTGAGGEGSGQMLSSPGS~L~DFRSSFF~~ C.elegans al(Iv) QIIAVHSQDTSVPQCPQGWSGMWTGYSFVMWTAAGAEGTGQSLQSPGSCLEEF~VPF~~ Drosophila al(IV)NVIAVHSQTIEVPDCPNGWEGLWIGYSFLMHTAVGNGGGGQALQSPGSCLEDFRA'TPFS" C.elegans a2(Iv) NTIAVHSQTIQIPNCPAGWSSLWIGYSFAMHTGAGAEGC~QSLSSPGS~LEDFR~rP~~~~~ S.purpuratus 4a QSLAIHSQSQEIPQCPGGWRSLWTGYSFTMYTAA-SEGGGQGLESVGSCL~NFRATPF~~~ Human a2(IV) IAIAVHSQDVSIPHCPAGWRSLWIGYSFLMHTAAGDEGGGQSLVSPGSCLEDFRATF?.~~ Human a4(IV) QAVAVHSQDQSIPPCPQTWRSLWIGYSFLMHTGAGDQGGGQALMSPGSCLED~R~PFL~ Human a6(IV) QAIAVHSQDIrIPQCPLGWRSLWIGYSFLMHT~GAEGGGQSLVSPGSCLEDF~TPF:~~ Pjcol4a rSIAVHSQDSNVPDCFPGWVTLWTGFSFI~QQT~QAEGTGQGLESPGSCLQHFRST~~~~

*** **>A * i x ** j1 * * ** * c ***,A ++ r+ ;>

‘; ic ;1 12

Human al(IV) CHG-.RGTCNYYANAYSFWLATIERSEMFKKPT-PSTLKAG-ELRTHVSReQVC~R~T Human a3 (IV) CHG-RGTCNYYSNSYSFWLASLNPERMFRKPI-PSTVKAG-ELEK~ISKCQVCMKKRH Human a5 (IV) CHG-RGTCNYYANSYSFWLATVDVSDMFSKPQ-SETLKAG-DLRTR~SR~QV~MKRT S.purpuratus 3a CHG-DGKCNYYATTYSFWLSSITGNAQFTMPE-SETLKAG C.elegans al(IV! CHG-RGTCNYYATNHGFWPSIVDQDKQFRKPM-SQTLKAG-GLKDRVSRCQVCLKNR Drosophila al(lV)CNGAKGTCHFYETMTSFWMYNLESSQPFERPQ-QQTI~G-ERQSHVSRCQVCM~SS C.elegans aZ(IV) CNGARGSCHYFANKFSFWLTTIDNDSEFKVPE-SQTLKSG-NLRTRVSRCQV~VKSTDGR~ S.purpuratus 4a CNG--RGNCHFFSNEYSFWLTVIDEEDQFA~PR-KRTIKSG-QLQSWSRCRVCQ~~ Human a2 (IV) CNGGRGTCHYYANKYSFWLTTIF-EQSFQGSPSADTLKAG~ Human a4(IV) CQGRQGTCHFFANKYSFWLTTVKADFEFSSAPAPDTLKESQAQRQK~SRCQV~V~~S Human a6( IV) CSGARGTCHYFANKYSFWLTTVEERQQFGELPVSETZKAGCMKSI~ rjcol4a CGG-IGQCSYDSVSGSYWMIVLDALNPFQDTE-PGTYPVS-DIEK~L~RCRV~~~~RR~

lr * * * A* * A *** **


human Col4A4 S S

PjCol4A

s S S

Fig 5. Schematic comparison of the exon/intron organization of different human type IV collagen genes and the sponge PjCol4A gene in the region coding for the NC1 domain. The exons are indicated by boxes and the introns by interconnecting bars. Hatched, black and empty boxes represent sequences coding for triple helical domain, NC1 domain and UTR region respectively. The letters S point to exon borders with split codons. The introns are not drawn to the actual scale. Modified from [31].

References

Aho S, Turakainen H, Onnela ML, Boedtker H (1993) Char- acterization of an intronless collagen gene family in the marine sponge Microciona prolifera. Proc Nat1 Acad Sci USA 90,7288-7292 Aumailley M (1993) Structure and function of basement membranes components: laminin, nidogen, collagen IV and BM-40. Adv Mel Cell Biol6, 183-206 Bacetti B, Gaino E, Sara M (1986) A sponge with acrosome: Oscarella lobularis. J Ultrastruct Mel Struct Res 94, 195-198 Beck K, Gruber T (1995) Structure and assembly of basement membrane and related extracellular matrix proteins. In: Prin- ciples of Cell Adhesion (Richardson PD, Steiner M, eds) CRC Press, Boca Raton, 2 19-252 Boury-Esnault N, De Vos L, Donadey C, Vacelet J (1984) Comparative study of the choanosome of Porifera: I. The Homoscleromorpha. J Morph01 180, 3-l 7 Boury-Esnault N, Muricy G, Gallissian MF, Vacelet J (1995) Sponges without skeleton: a new mediterranean genus of Homoscleromorpha (Porifera, Demospongiae). Ophelia 43, 25-43 Braze1 D, Oberblumer I, Dieringer H, Babel W, Glanville W, Deutzmann R, Kuhn K. (1987) Completion of the amino acid sequence of the alpha 1 chain of human basement membrane collagen (type IV) reveals 21 non-triplets interruptions located within the the collagenous domain. Eur J Biochem 168,529-536 Cortay JC, Negre D, Scarabel M, Ramseier TM, Vartaki NB, Reizer NB, Saier MH, Cozzone AJ (1994) In vifro asymmet- ric binding of the pleiotropic regulatory protein, FruR, to the ace operator controlling glyoxylate shunt enzyme synthesis. J Biol Chem 269,14885-14891 Cecchini JP, Knibiehler B, Mirre C, Le Parco Y (1987) Evi- dence for a type-IV-related collagen in Drosophila melano-

Fig 6. Localization of the sponge type IV collagen. A. Immunofluorescence staining of a lamina underlying the surface epithelium of Pseu- docorticium jarrei on a frozen section. x 100. B. Electron microscope demonstration of the basement membrane-like layer (arrowheads) lining an epithelial cell in Pseudocorticium jarrei. x 20 000.

4 Fig 4. Sequence comparisons of the NC1 domains of several vertebrate and invertebrate collagen (r (IV) chains. The multiple alignments were done using the Clustal W method [33]. The quoted species are the same as in figure 2. The 12 cysteine residues are indicated by numbers; the asterisks point to residues strictly conserved in all the chains; the triangles indicate residues conserved in all the chains except in the sponge chain PjCol4a:

44

10

11

I2

13

14

1s

16

17

18

19

20

21

22

gaster. Evolutionary constancy of the carboxyl-termmai non- collagenous domain. Eur J Biochem 165,587-593 Engel J, Efimov VP, Maurer P (1994) Domain organizations of extracellular matrix proteins and their evolution. Develop- ment (suppl) 1994, 35-42 Exposito JY, Garrone R (1990) Characterization of a fibrillar collagen gene in sponges reveals the early evolutionary appearance of two collagen gene families. Proc Natl Acad S0 USA 87,6669-6673 Exposito JY, Le Guellec D, Lu Q, Garrone K (1991)Shori chain collagens in sponges are encoded by a family of closely related genes. J Biol Chem ‘266.2 1923-2 1928 Exposito JY, Suzuki H, Geourjon C, Garrone R, Solursh M. Ramirez F (1994) Identification of a cell lineage-specific gene coding for a sea urchin c~‘2 (IV)-like collagen chain. JBiol Chem 269, 13167-13171 Fessler JH, Fessler LI (1989) Drosophilu extracellular matrtx. Annu Rev Cell Biol5,309-339 Garrone R (1984) Formation and involvement of extracellular matrix in the development of sponges: a primitive multicellular system. In: The Role of Extracellular Mutrix in Develop- ment (Trelstad RL, ed), Alan R, Liss, New York, 461-477 Garrone R (1985) The collagen of the Porifera. In: Biology qf Invertebrate and Lower Vertebrate Colkugens (Bairati A. Garrone R, eds) Plenum Press, London, 157-175 Geourjon C, Deleage G (1993) Interactive and graphic coupling between multiple alignements, secondary structure predictions and motif/pattern scanning into proteins. Comput Appl Biosci 9, 87-91 Hudson BG, Reeders SY, Tryggvason K (1993)Type 1V coi- lagen: structure, gene organization and role in human dis- eases. J Biol Chem 268,26033-26036 Humbert-David N, Chandrasekaran S, Tanzer ML, Garrone R (1995) Laminin extraction and disorganization of collagen fibrils in snail muscle by EDTA treatment. Biol Cell 83. 39-47 Humbert-David N, Garrone R (1993) A six-armed tenascin- like protein extracted from the Porifera Oscarella tuberculnttl (Homosclerophorida). Eur J Biochem 216,255-260 Inoue S (1989) Ultrastructure of basement membranes. ltzt Rev Cytol 117, 57-98 Kramer JM (1994) Structures and functions of collagens in Caenorhahditis elegans. FASEB J 8. 329-336

Kiihn K (1994) Basement membrane (type IV) coiiagen. Matrix Biol 14,439-445 Mackrell AJ, Kusche-Gullberg M, Garrison K, Fessier, JFi (1993) Novel Drosophila laminin A chain reveals structural relationships between lamiilin subunits. FASEB J 7, 375-38~1 MariyamaM, Kalluri R, Hudson B, Reeders ST (1992) The a4(IV) chain of basement membrane collagen. J Viol CIlenl 267, 1253-1258 Merker HJ (1994) Morphology of the basemenr membrane. Micr Res Tech 28,95-124 Oohashi H, Ueki V, Sugimoto M, Ninomiya Y 11995) IX& tion and structure of the COLA6 gene encoding the human rxh(IV) collagen chain and comparison with other type fV collagen genes. J Biol Chem 270,26863-26867 Paulsson M (1992) Basement membrane proteins: ~IFUCWT, assembly, and cellular interactions. Crit Rev Riochem Mrii Rio1 27,93-127 Pavans de Ceccatty M (1986) Cytoskeletal organizatron bni: tissue pattern of epithelia in the sponge Ephydutia m&llerC. J Morph &9,45-65 Shimizu K, Yoshizato K, (1993) Involvement of coltasen synthesis in&sue reconstitution by dissociated sponge cells. Dev Growth Differ 35,293-300 Sugimoto M, Oohashi T, Yoshioka I-I, Matsuo N. Nmom:y,; Y (1993) cDNA isolation and partial gene structure of the human &(IV) collagen chain. FEBS Lett 330, 122-128 Tabor S, Richardsun CC (1985) A bacteriophage T7 RNA polymeraselpromoter system for controlled exclusive expres sion of specific genes. Proc Nat/ Acrid Sci USA 82 1074-1078 Thompson JD, Higgins DG, Gibson TJ (1994) CLUSTAl., W : improving the sensitivity of progressive multiple sequepce alignement through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res 23~ 4673-4680 Timpl R (1989) Structure and biological activity of hasemem membrane proteins. Eur 3 Biochem 180,48?-502 Timpl R (1996) Macromolecular organization of basement membranes. Curr Opin Cell Biol8.618-624 Yurchenco PD. O’Rear 53 (1994) Basal lamina assembly. Curr Opin Cell Biol6.674-681 Yurchenco PD, Schnitty JC ( 1990) Molecular archltecttire cti basement membranes. FASEB J 4, 1577- 1590

Type IV collagen in sponges, the missing link in basement membrane ubiquity

Documents

Transcript of Type IV collagen in sponges, the missing link in basement membrane ubiquity