Volume 305, number 2, 91-96

FEBS 11195

June 1992

© 1992 Federation of European Biochemical Societies 00145793/92/$5.00

Fibronectin type III-like sequences and a new domain type in prokaryotic depolymerases with insoluble substrates Christian K. Hansen Institute de Mlcrobiologia Dioquimiea, Facuhad de Biolag[a, CSIC/Universldad de Salamanca, 37071 Salamanca, Spain Received 13 April 1992; revised version received 11 May 1992 Fibronectin type Ill-like sequences are present in many Froteins from higher eukaryotes and are involved in protein-protein interactions, heparin binding and cell adhesion. A nine-member family of bacterial sequences is shown to be significantly homologous to the type Ill-like sequences. All the sequences are contain~ in secreted depolynacrases acting on complex, energy-rich insoluble substrates, in which they apparently do not participate in catalysis or substrata-binding, their ezact function remaining unclear. Furthermore, a new family of sequences, present in some cellulases, is present~. Amylase; Cellulas¢; Cellulose binding-domain; Chitinase; Fibronectin; Poly.3.hydroxybutyrate depolymerase

1, INTRODUCTION Depolymerization of complex, insoluble, and sometimes crystalline natural polymers, such as cellulose, xylan, chitin and starch, often involves protein domains with catalytic and substrata-binding capabilities [1]. Trichoderma reesei [2], Celhdomonas fhni [3]: Bacillus lautus [4,5], and some Clostridla [6-8] produce efficient and well-characterized cellulose- and xylan-degrading enzyme systems. While fungai, aerobic- and some anaerobic bacterial systems are believed to be composed of various non-associated polypeptides with catalytic and substrata-binding domains~ Cl. thermocellum and CI. cellulovorans produce a high molecular weight (2-4 MDa) cellulosome which consists of catalytic polypeptides joined to a large cellulose-binding protein [9,10]. By conventional primary amino acid sequence analysis and hydrophobic cluster analysis [11], nine families of catalytic domains (A-I) [1,12] and five families of cellulose-binding domains (I-V) have been identified fN.R. Gilkes, personal communication; Coutinho et el., paper submitted) which, in the non-associated enzyme systems, are joined by linker sequences [1]. In contrast, catalytic subunits of the cellulosome contain a 24 amino acid repeated sequence which binds to the central protein (CbpA) [13]. CbpA must in turn have cellulose-, subunit- and perhaps cell.binding functions. Shoseyov et el. [14] showed that CbpA of CI. cellulovorans conAbbreviations: Fn3(s), fibronectia type lll-like sequence(s). Correspondence address: C.K. Hansen, ¢o Per Lin/t Jergen~en, Nero Nordisk A/S, Nero Alle 1, DK-2880 Bagsvterd, Denmark. Fax: (45) 4449 0070.

Published by ElsevierScience Publishers B. 14

tains a family III cellulose-binding domain, a hydrophobic, probably subunit-binding domain type, and a hydrophilic domain, homologous to regions in endo-p1,4-glucanase EG-Z of CI. stercorarium [15]. It was recently reported that CenB of Cell. fimi, putative protein ORFX of Cell. flavigena [16], and chitinas¢ AI orB. circulans WL-12 [17] contain sequences homologous to type III repeats (FnYs) in fibronectin, a eukaryotic extracellular matrix adhesion molecule [18]. Fn3's in fibroneetin are probably involved in protein-protein interactions, heparin-binding and cell adhesion [18]. The presence of Fn3-1ike sequences and hydrophilic (71. cellulot,orans CbpA-like sequences in several hydrolytic enzymes acting on insoluble natural polymers is reported. 2. MATERIALS AND METHODS The PCGENE software package (Genofit, Switzerland) was on PC-AT compatible microcomputers for alignment of protein ~e. quences (PALIGN and PCLUSTAL) and construction of phylogenetic trees (PCLUSTAL), while the FASTA software paekal~e was mad for ~.anning data banks. PCOMPARE was used to calculate the alignment scores, using the MDM~ matrix [19]. HCAPLOT was uze.d on Apple computers to present sequences for hydrophobia cluster analysis [11].

3. RESULTS 3.1. Fn3's Comparison of the reported Fn3's of chitinase AI of B. circulans WL-12 [IT], CenB of Cell. fimi [16], and putative protein ORFX of Cell. flavigena [20] with the SWISS-PROT Protein 3equence Database release 20, using the FASTA program, revealed that an g-amylase-

91

Volume 305, number 2

CthAAP CthAAP AfaP3HB BciChAl BciChAl CfiCenB CfiC~nB CfiCenB CflORFX

FEBS L E T T E R S

894 1126 340 458 553 644 742 840 106

Q--NLTAPQ-PI~DLKAVSGNGKVDLSWSVVDKAV ..... SYNI%RSTVKGGL PSNDVBTPTA~-~LQQPGZE~SR~A'LDWSPSADDV--AI~GZ'KI%'KSSSETGP T G - - - Q A G S A P T G L A V T A T T S T S V S L S W N A V A N A . . . . . S S Y G V ~ R . . . . NGS PPVDTTAPSVPGNARSTGUTANSVTLAWN~STDNV--GVTG~?~/N .... GAN 9G~DTQA~APTNLASTAQTTSSITLSWTASTDNV--~"~TGMD%"/N .... GTA TTTDTTPPTTPGTPVATGVTTVGASLSWAASTDAG-SGVAG~ELYRVQGTTQT TTGET~PPTTPGTPVASA%"~STGATLAWAPST--GDPAVS~MDVLRVQGTTTT PPVDTVAPTVPGTPVASNVATTGATLTWTASTDSGGSGLA~"~EVLRVSGTTQT DPTDTQAPSVPSGLTAGTVTETSVALSWTASTDNV--GVTGMD%'/~ .... NGS

DT A~T g

B a c t e r i a l (9) ~ibroneotin (16) T i t i n (~200) T w i t c h i n (31) C y t o t a o t i n (8-11) Phosphatase (16) C y t o k i n ~ r~o. (40) Total

941 I17~ 380 504 599 695 792 892 153

VT~S~ ASTD V G V GZ V Y ~ S V W TGYRV S V I'- W P DGGS X GYIVEKRD

L

PG P LD P TVP P

PEV DV D T ~- W PP D D G G A P I L TD W P V L N L LW GD L VW

V

P

T

LW

~ K Y Y ~

W W

D

Y

YV

Y E K I A S N V T Q T - T Y T D T ~ . V T N G L K Y V X A V T A V D N D G ~ E S A L SN . . . . . . . . E V F IK I A T V S D S V Y N Y V D T D V V N G ~ A V D T S Y N R T ~ S N T V K A ~ P D I IP I KVGSATATA .... ~TD SGL~AGTTYSMTVTAVDP TAGESQP SAA-VS .... AT L A T S V T G T T . . . . AT X S G L T A G T S ~ T F T ~ K ~ E D A A G N L S R ~ S N A - % " E V S T T A Q L A T T V T G T T . . . . AT I S G L A A D T S ~ T F ~ A A G N V S R ~ g N A - ~ S V K T A A E L V G T T T A A A . . . . Y IR L D L T P G T A Y S ~ V A G N V S ~ A A - ~ T ~ T T D -WAQTTVPT .... VTRSGLTP STAYTYAVKAKNVAGDVS~L SAP-~TFTTAAL V A S P T T A T . . . . V A R A G L T P A T A Y S ~ V % ~ K K D G A G N V S A V S SP - ~ T F T T L - KVGSSSGTT .... YSD~LTAATA~QYSVAA~DAAG~VSQRS 8A- L S V T T K S -

B a c t e r i a l (9) Fibronectin (16) T i t i n (~200)

V

T w i t c h l n (31) C y t o t a o t i n (8-11) Phosphatase (16) C y t o k i n e rec. (40)

V

Total

T

VT T %"~

P PGPP

consensus

CthAAP CthAAP AfaPSHB BciChAl BciChAl CfiCenB C£iCenB CfiCenB CflOB~X

June 1992

T

T

T

YT T

G L T GT X Y V A K D A G N V S A A $ GL PO EY V V A VT L E G E ~ F R V A N A ~ P

V G L E G ~ Z E F R V /%VN A G S D P S T. P G E Y V L A K S P F L PGR ~ V T SG L T. X V S

con3ensus

•

G EY

A ~T V AK '2 GR S

F

V A

Fig. 1. Alignment el" Fn3's. Homolo[~ous regions in CI. thernmhydrosu/phuricum 0~-amylase-pullulanase (CthAAP) [21], AI. faecalts poly-3h~/droxybutyrate depolymerase (AI'aP3HB) [22], B. circu/ans ehitinase AI (BciChAl) [17], Celt.fitnl CcnB (CfiCgnB) [16], and Ct,ll.flavigena ORFX (CflORFX) [20] are shown in the upper part of the alignment. A consen.~ussequence for the bacterial sequences (Bacterial) was construct~ with amino acids conserved in at least 5 ~quenge,s, with residue~ eon~rved i , 8-9 sequences shown in bold, and completely conserved residues underlined. Amino acich in the individual sequences, which are identical to the consensus sequence, are also shown in bold. Partial comensus sequences of Fn3's in fibroneetin [26], titin [23], twitehin [24], eytotaetin [25], a protein-tyrosin© phosphata~¢ [27] and s~vcral cytokine receptors [281are shown below. A total con~nsus sequence (bottom) was constructed by showing residues conserved in at least 4 of the individualconsensus sequences, with completely conserved re,~idugsin bold. Residues in the individual consensus sequences identical to the total consensus are shown in bold. Numbers in parentheses refer to the number of sequences present in the protein or group of proteins.

pullulanase from (71. therrnohydrosutphuricura [21] and a poly-3-hydroxybutyrate depolymgrase from Alcaligenes faecali~' [22] also contain such motifs, repeated twice in the former. Alignment of the bacterial Fn3's shows conservation of 6 amino acid residues, of which 3 a~ aromatic (Fig. 92

1). The latter and a proline are also conserved in FnYs of the eukaryotie proteins: titin [23], twitchin [24], cytotactin [25], fibron~tin [26], a protein-tyrosin¢ phosphatase [27] and several eytokiae receptors [28]. Pairwise comparizon ot" the bacterial s~quen~s with PALIGN showed that they have 19.2-71.6% identical,

Volume 305, n u m b e r 2

FEBS L E T T E R S

June 1992

and 33.7-83.2% identical plus similar amino acids (Table I). The homology of these sequences and Fn3's in fibronectin was assessed with PCOMPARE, which assigns significance to homologous sequence-pairs with SD values larger than 3. Each of the bacterial sequences was compared to the other bacterial Fn3's and to the 16 relents in fibroneetin. SD values of 3.98-22.58 were obtained when the bacterial sequences were compared, clearly demonstrating the significance of their similarity (Table 1). In addition, each bacterial sequence scored SD values above 3 when compared to at least 6 of the 16 FnYs in fibronectin. The phylogenetic tree of Fn3's clearly shows distinct prokaryotic and eukaryotic sequence clusters (Fig. 2).

.....

~

m

m

~

_

_

_

_

C~&API C~AAP2 Afa~I-IB [---- B¢iChA11 lkiChAl2 CflORFX CfiCenBl CfiCenB2 CfiCenB3 FIIII FIII2 Fill3

F!.t_t5

3.2. Hydrophilie CI. cellulovorans CbpA-like se~,'ences By comparing the hydrophilic repeats of Cl. eellulovorans CbpA [14], and domains B and B' of Ct. stercorarium EG-Z [15] with the protein sequence databank, endo-fl-l,4-glucanase EG-B of B. lautus [4] was shown to contain a homologous domain, surrounded by putative linker sequences (SSTASS and SSTTGTTSS) (Fig. 3) [1]. Alignment of the sequences shows that 8 amino acids are generally conserved. Pairwise comparison of the sequences showed 22.1-63.2% identical, and 38.3-83.3% identical plus similar amino acids (Table II). SD values larger than 3 (3.53-17.78) were obtained in all pairwise evaluations. Program HCAPLOT was used to present region B of EG-Z, hydrophilic repeat 1 of CbpA and the homologous region of EG-B for hydrophobic cluster analysis (Fig. 4). Similar cluster shapes (hydrophobic regions) ($1, $2 and $3) are observed around the conserved asparagine-glycine, aspartic acid-tyrosine and phenylalanine residues, respectively, emphasizing their probable importance for the conserved residue micro-environment.

Fin4 FIIIll l~IIIl3 FIII9 FIIII5 FIIII2 FIIII0 FIII8 FIII14 FIII7 FIII16 FIII6 Fig. 2. Phylogenetic tree Ibr the Fn3's of b~terial enzymes and fibron~tin. Repeats 1-16 from fibron~tin ar~ d.~.3gnat~ Fill l, FIII2, etc. Bacterial sequences are designated as in Fig. 1, r~peats from the same protein carrying a suffix to indicate their position in the protein (e,B. CfiCenBi, CfiCenB2, CfiCenB3, etc.).

4. DISCUSSION FnYs were until recently only reported in proteins from higher eukaryotes such as the (i) muscle proteins, i.e. titin, twitehin, smooth muscle myosin light chain kinase, C-protein and 86K protein [23], (ii) cell adhesion proteins, i.e. neural adhesion protein L1 [29], fi-

Table 1 Pairwise alignment of bacterial fibron~ctin type III-like sequences Domain CthAAPl CthAAP2 AfaP3HB BoiChA11 BeiChA12 CflORFX CfiCenBl CfiCenB2 CflCenB3 Fib re n~tin rope, 1-16

CthAAPI 7.17 9,19 3.98 4.90 5,98 8.06 4,93 5,12 11/411

CthAAP2

AfaP3HB

31.5/46.1

34.1/43.5 24.7/40.0

8.01 7.26 9.70 9.58 6.48 6.56 4.54 12/3/1

7.11 10.46 14,20 8,21 6.09 5,99 1511/0

'BciChAII 25,8/41.5 22,1/40,0 36.5/54.1 22.27 20.06 13.26 12.04 15.5"t 10/4/2

BciChAl2

CflORFX

CfiCenBl

CfiCenB2

CfiCenB3

23.6/39,3 20.0/42,2 43.5/60,0 71.6/83,2

21.3/38,2 24.5M1.5 50.0/76,6 52,1/68,1 53.2/73.4

24.7/33.7 21.4/41.8 38.8/50.6 46.3!66.3 45.3/64.2 45.7/60.6

23.6/34.8 23.5/40,8 29.4/48.2 41.1/57.9 44.2/61.0 38.3/57.4 60.2/72.4

23.6/38.0 19,2/37.4 28.2/48.2 45.3/66.4 42.1/63,2 48.9/62.7 65.3/78.6 61.2/76.5

16.03 10.31 13,27 14.03 71712

12,19 13.51 13.78 6/911

22.58 21.19 7/4/5

18.17 1i 14/I

11/4/I

The upper triangle shows pereentase identitiestidentities + similarities. The lower triangle shows sisniflcanee scorns for comparisons of the sequences. The scores, in SD units, were cakalah.'d from 50 random runs with the PCOMPARE program using the MDM~a matrix [19]. A bias of 60 and a gap penalty of 60 were used to d e i s t the relationships. The bottom line indicates the number of repeats in fibron~tin scoring abovedtmtw©en 2 and 3/tmlow 2 SD units when c,ompar~ to the respective bacteria[ scquen~'s.

93

Volume 305, number 2

FEBS LETTERS

CoeCbpA-I C¢eCbpA-2 CoeCbpA-3 CceCbpA-4 Cat~GZ-B Cag~GZ-B'

BIa~GB

21359415351628667-

760385-

CooCbpA-I CoeCbpA-2 CceCbpk-3 Cc@CbpA-4 Cst~GZ-B

ADI/KTTMTLNGNTFKTXTD~GTA~ASTDYSVSGND'~T X SEAT ADVATTMTLNGYTYNGXTGL ...... TTSDMS ISGNVVKISQAY A D A A I T~TIaNGNTIPSA~KN- G T A T ~ V K G T D ~ T V S E ~ ISEAM ADVV~S ~ G N T ¥ S A~ K N - G T T T L V K G T D Y T X S G S T%"EI SEA'~ ~,NI Q I / ~ I~NGNTLNOI KY-GNTYLP.IgGTDYTVSGDTVT ~ LKS ~ RDVKVKLVPNG~IT L L A V K K -D G ~ . ~ V L G R D Y 8 IDGD~VTTFPJgY KDTSVQLNLNGNTLTSLS -~rNGT¢LKSGTD~TLNS SRLTFKASQ ADV VTM ~

Consensus

25763215781671710-

Cs~GZ-B"

803-

BIa~GB

428-

June 1992

I

T L

GTDYT

S

VT~SKAY

LAKQSV ..... GTTTLt~NFS AGN---PQKLV-- ITVVDTPVEA LAKQPV ..... GDLTLTYNFSNGNKTATAKLV--VS IKDAP-KT L A L - Q T . . . . . G T V T ~ ¥ V F D K G N S R / ~ V V V A V K E IQ IVN ST I TP L A T L A D . . . . . GSATIdEFVFNQGAS A K L R L T . . . . . II/PAVVDP LNSFDT ..... STVQLIFDFSA@RDPVLT~ X IDT . . . . . . . ~T L A D Q P V . . . . . G R V T L T F D F D R G T D P V L T ~N ITD SI%QV~TGV ~Q L T K L T S L G K L G V N & T I V T K F N R G A D W K ~ N V % / L Y N T L a K L S STTGT

Consensus ~ G TL ¥ E • V FiB, 3. Alisnment of reBions homoloBous to Ct. celhdavorara' CbpA hydrophilic repeats (CceCbpAI-Cc¢CbpA4) [14], Homologou-~ r¢~ions in C/. • Iercorarlum EG-Z (C~tEGZB and CstEGZB') [15] and B, lautus EG-B (BlaEGB) [4] are also shown. The consensus sequcn~ was constructed with residu~ conserved in at least 4 sequcn~s. Amino acids conserv~ in 6--7 saquences are shown in bold, while completely conserved residues are underlined. In the individual scquencgs, amino acids id,ntieal to the gonsgnsus ,~eqtlen~ are also shown in bold,

bronectin [18] and cytotactin [25], and (iii) a family of cytokine receptors [28]. In many of these proteins, Fn3's are known to interact with other proteins, heparin and the cell, although the molecular basis of the interactions in most cases is unclear. TI:e function of Fn3's in the bacterial proteins is unknown, although they would be expected to be nonessential for the catalytic and substrate-binding activities since they are found in enzymes with activity against different substrates, namely cellulose, starch, chitin and poly-3-hydroxybutyrate. For example, CcnB-activity on soluble substrate and qualitative cellulose-binding activity has been shown to be independent of the presence of the FnYs [16]. Likewise, in chitinase A1, elimination of the Fn3 and a region of unknown fanctior~ (l~rhaps a chitin-binding domain), did not abolish activity against insoluble chitin, although some difference

was observed and attributed to loss of the chitin-binding ability [17]. Similarly, the poly-3-hydroxybutyrate depolymcrase was shown to lose substrate-binding capability and activity against insoluble substrates upon treatment with trypsin, an event accompanied by deletion of approximately 60 amino acids, probably from the C-terminus and not including the Fn3 [30] (Fig. 5). The role of FnYs in bacterial proteins, therefore, remains unclear, although functions such as binding to other components of the hydrolytic system, the substrate or the cell can be imagined. The role of hydrophilic regions in CbpA, EG-Z and EG-B is also unclear, since none of the regions have been characterized individually, However, an EG-Z peptide composed of domains C', B, B' and C has been shown to bind to cellulose [15], although it remains unclear if this binding is mediated by C-sequences alone

Table 11 Pairwis¢ alignments of CbpA hydtophilic-like sequences Domain CceCbpA-l Cc~CbpA-2 CocCbpA-3 Cc~CbpA-4 CstEGZ-B CstEGZ-B' BIaEGB

CceCbpA-I 10,80 I 1.70 ! 5,24 15.02 11.98 6.28

Cc~CbpA-2

C.ceCbpA-3

CceCbpA-4

63,2/69,1

50.0/59.7 55.9/64.7

51,2/67.2 51.5/60.3 77..2/83.3

13.40 10,42 9,91 l 1.84 3.53

17.42 1147 i 4.20 7.34

17.78 12.38 6.22

CstEGZ.B 47.9154.B 36.8/47,1 48.6/55,5 48.0/57.3 12,21 6.94

CstEGZ-B'

BIaEGB

43,8/52,0 41,2/45.6 43,1/55,6 41.3/50.6 42.7/52.0

28.8/49,3 22.1/38.3 31.9/51,3 36.0/50.7 29.3/42,6 29.3/44.0

6.65

The upper righttriangleshows per~ntagc identities/identities ~-similarities.Tho lower lef~triansleshow.~significancesco~s {'orthe comparisons of th~ ~quences,calculatedas in Table I.A biasof 60 and a gap penaltyof 60 were used to detectthe relationships.

94

Volume 305, number 2

FEBS L E T T E R S

g

N

D

~

June 1992

~

D

~

D

s,I CccCbpA

z

rs D K

".9'

[]

m

u

Is, I o oA a

•

I

•

*

~ g~

:MNN

o ' "

Fig. 4. Alignment of CI. celltd~:orans CbpA hydrophilic domain-like sequences by hydrophobic cluster analysis [i 1]. Sequenc,es were presenlcd with the HCAPLOT program and aligned by eye. Conserv~l regions Sl, $2 and S3 arc shown. Abbreviations are as in Fig. 3. Open square, thrconine; filled sqtmrc, serin¢; star, praline; diamond, gly¢ine.

(cellulose-binding domains) or is enhanced by the Bsequences. CbpA, as a whole, has been proposed to contain the substrate-,subunit-, and perhaps cell-binding functions of the CI. cellulovorans 'c¢llulosome' [14]. It shall be interestingindeed to determine the roles played by these regions in bacterial enzymes and perhaps contribute in the investigationof FnYs present in so many eukaryotic proteins. Acknmv[edgements: Drs. Nell R. Gilkes and Roi H. Doi are thanked for making data available prior to publication. Dr. Ramon $antamaria, in whose laboratory this work was carried out, Miss MajBrit: Kaltoft, and Drs. Pierre B6guin, Peter Carisson, Bernard Hanrissat and Per L. Jergensen are thanked for helpful discussions. The financialsupport from General DirectorateXII of the Commission of European Communities is greatlyappreciated.

REFERENCES [1] Gilkes, N.R., H¢nrissat, B., Kilbarn, D.G., Miller Jr., R.C. and Warren, R.A.J. (1991) Microbiol. Roy. 5~, 303-315. [2] Mandels, M., Sternbcrg, D. ~tnd Andrcoti, A.E. (1975) in: Proceedin~ of Symposium on Enzymatic Hydrolysis of Cellulose

(Bailey, M., Enari, T.M. and Linko, M. cds.) pp. 81-109, Aulanko, Finland. [3] Gilkes, N.R., Langsford, M.L., Kilbum, D.G., Miller Jr., R.C. and Warren, R.A.J. (1984) J. Biol. Chem. 259, 10455-10459. [4] Jergens~n, P.L. and Hanson, C.K. (1990) Gone 93, S5--60. [5] Hanson, C.K. (1992) Ph.D. Thesis, The Tgchnical University of Denmark, Copenhagen. [6] Weimer, P.J. And Zcikus, J.G. (1977) Appl. Environ. Mi~robiol. 33, 289-297. [7] Bronnenmeier, K. and Staudenbauer, W.L. (1988) Appl. Microbiol. Biotcchnol. 27, 432-436. [8] Bleat, R., Mah, R.A. and Robinson, R. (1984) Appl. Environ. Mierobiol. 48, 88-93. [91 Lone.d, R. and Bayer, E.A. (1988) Adv. Appl. Microbiol. 33, 1-46.

[10] Shoseyov, O. and Doi, R.H. (1990) Proc. Natl. Acad. S¢i. USA 87, 2192-2195. [111 Gaboriaud, C , Bissery, V., l~neheritrit, T. and Mornon, J.-P. (1987) FEB$ L~tt. 224, 149--155. [12] Henrissat, B., Claeysscns, M., Tommc, P., Lamesle, L. and Motnon, J.-P. 0989) Gone Sl, 83-95. [13] Tokatlidis,K., Salamiton,$.,l~guin,P.,Djurjati,P.and Aul)crt, J.-P.(1991)FEBS L-.tt.291,185-158. [14] Shoseyov,O., Takagi, M., Goldstein,M.A. and D,oi,R.H. 0992) Prec. Natl. Acad. Sci. USA (in press).

95

Volume305,number2

June 1992

FEBS L E T T E R S

C lostridium ceUulovorans CbpA

Bacilluslautu~strainPL236EG-B ~

ClostridiwnstercorariumCelZ

~

CellulomonasfimiCenB

~ B a c i l l u s

circulansChitinas¢A1 Cloatridiwnthermohydrosulfuricum tz-amylase-pullulanase

L______.~Alcaligenesfaecalispoly-3-hydroxybutyrat¢depolymeras¢ Scale I ~

Humanfibroneetin I

200 aa

FibronecfintTl~ M repeat ~

Signal ~ peptid¢

Hydrophili¢CbpA-like ~ domain

Possiblesubstratebindingregion

Fig. 5. Domain organization in the bacterial sequences aligned in Figs. 1 and 3, and fibronectin. Letters indicate membership of a fl-glycanase catalytic domain family [1], and roman numerals indicate membership e r a cellulose.binding domain family (N.R. Gilkes, ~rsonal communication; Coutinho, et al., paper submitted). Other codes are explained in the figure.

[!$] Jauris, S,, R~¢knagel, K.P., $¢hwam, W.H., Kmtzseh, P., Bronnenm¢icr, K. and Staudenbauer, W.L. (1990) Moi, Gem G~net. 223, 258-267. [16] Moinke, A., Gilkes, N.R., Kilburn, D.G., Miller Jr., R.C. and Warren, R.A,,I. 0991) J, Bacteriol, 173, 7126-7135, [17] Watanabe, T., Suzuki, K., Oyanagi, W., Ohnishi, K. and Tanaka, H. (1990) J. Biol. Chem. 265, 15659-15665. [181 Ruoslahti, E. (1988) Anna. Rev. Bioehem. 57, 375-413. [19] Dayhoff, M.O. 0978) in: Mlas of Protein Sequgnc~ and Struttare, vol. 5, suppl. 3 (Dayhoff, M.O. ed.) pp. 1-8, National Biomedical Foundation, Washi~ll~ton, De. [20] Al-Tawh~d, A.R, 0988) M,Se. Thesis, Trinity Collese, Dublin. [21] Mzlasniemi, H., Paloheimo, M, and Hemi0, L. 0990) L Gem Microbiol. 136, 447--.454. [22] Saito, S., Suzuki, K., Yamamoto, J., Fukui, T., Miwa, K., To. mita, K., Nakanbhi, S., Odani, S., Suzuki, J.-I. and lshikawa, K, (1989) J. Bacteriol. 171,184--189.

96

[23] Labeit, S., Barlow, D.P., Gautel, M., Gibson, T., Holt, J., Hsieh, C.-L., Franek¢, U., Leonard, K., Wardale, J., Whitins, A. and Triniek, J. (1990) Nature 345, 273-276. [24] Benian, fi.M., Kiff, J.E., Neekelmann, N., Mocrman, D.G. and Waterston, R.H. (1989) Nature 342, 45-50. [25] Jon~s, F.S., Hoffmann, S., Cunninsham, B.A. and Ed~Iman, G.M. 0989) Proc. Natl. Acad. Sci. USA 86. 1905-1909. [26] Kornblihtt, A.R., Umezawa, K., ¥ibe-F~lersen, K. and Baralle, F.E. 0985) EMBO J. 4, 1755-1759. [27] Krucgcr, N,X., Streuli, M. and Suite, H. (1990) EMBO J. 9, 3241-3252. [28] Bazan, J.F. (1990) Prec. Natl. Aead. $ei. USA 87, 693~-6938. [29] Moos, M., Tacke, R., Soberer, H., Teplow, D., Frah, K. and Schner, M. 0988) Nature 334, 701-703. [30] Fukui, T., Narikawa, T., Miwa, K., Shirakura, Y., Saito, T. and Tomita, K. 0988) Biochim. Biophys. Acta 952, 164-171.