Sequence analysis of the cellulase-encoding celY gene of Erwinia chrysanthemi: a possible case of interspecies gene transfer

Gerzr, 106 (1991) 109-114 ce) 1991 Elscvier Science Publishers

GENE

B.V. All rights reserved.

109

0378-l 119/91/$03.50

06062

Sequence analysis of the cellulase-encoding interspecies gene transfer DNA;

(Recombinant

phytopathogenic

bacteria;

Annick Guiseppi, Jean Luc Aymeric,

ceZY gene of Erwinia chvysanthemi:

endoglucanase;

nucleotide

sequence;

NtrA promoter;

a possible case of

horizontal

transfer)

Brigitte Cami, FrCdCric Barras and Nicole Creuzet

Lahoratoire de Chitnie BactPrienne, CNRS,

13277 Marseille Cedex 9 (France)

Received by J.-P. Lecocq: 13 March 1991 Revised/Accepted: 13 May/24 May 1991 Received at publishers: 23 July 1991

SUMMARY

The Erwiniu chr~~sunthemi (strain 3937) celY gene encoding the minor endoglucanase (EGY) was sequenced. The analysis of the upstream region allowed us to identify an in vivo active promoter recognized by the NtrA (d4) holoenzyme. No similarity was found between the predicted amino acid (aa) sequences of EGY and either the Er. chrpunthenzi major endoglucanase, EGZ, or the Er. curotovoru CelS endoglucanase. In contrast, a very high level of identity, both at the nucleotide and the predicted aa levels, was found between celY and an EG-encoding gene from Cellulomonas udu, a Gram + bacterium taxonomically distant from Er. chrpunthemi. By comparing the molar G + (2% of the cellulase-encoding genes and that of Er. chr~sunthemi and C. udu chromosomal DNAs, we speculate that celY was transferred from Er. chr~sunthemi to C. udu.

INTRODUCTION

Er. chrysunthemi, a member of the Enterobacteriaceae, is capable of causing soft-rot disease of plants (Chatterjee and Vidaver, 1986; Collmer and Keen, 1986; Kotoujansky, 1987). At least part of this capacity is mediated by the

Correspondence CNRS,

to: Dr. F. Barras,

31 Chemin

Joseph

Laboratoire

Aiguier,

de Chimie

Bacterienne,

BP 71. 13277 Marseille

Cedex

9

(France) Tel. (33)91224148; Abbreviations:

Fax (33)917189

aa, amino

acid(s);

14. Ap, ampicillin;

CeNulomonas;CAT, Cm acetyltransferase; cellulase(s);

celY, gene encoding

chloramphenicol;

bp, base pair(s);

cat, gene encoding

EGY; celZ, gene encoding

CMC, carboxymethylceilulose;

C.,

CAT; Gel, EGZ;

Cm,

EG, endoglucanase(s);

EGY, Er. chrysanthemi minor EG; EGZ, Er. chrysanthemimajor EG; Er., Erwinia; kb, kilobase or 1000 bp; LB, Luria-Bertani (medium); nt, nucleotide(s); PAGE,

ntrA, gene encoding

polyacrylamide-gel

protease( Shine-Dalgarno transcriptional

R, resistance/resistant; sequence; sigma factors;

54-kDa d4; ORF, open reading frame;

electrophoresis;

Pel, pectate

RBS,

SDS, sodium

ribosome-binding dodecyl

wt, wild type.

sulfate;

lyase(s);

Prt,

site; SD, d4 and d’,

action of cell-wall depolymerising enzymes, such as pectinases, proteases, and cellulases. For all three activities there are multiple enzymatic forms. Questions that arise out of this apparent redundancy relate to its physiological advantage for the bacterium and to the evolutionary mechanism that established it. Er. chrysunthemi produces five Pel, three Prt and two EG. Genetic, biochemical and immunological criteria led to classify the five (or four in EC16 strain) Pel in two groups, each containing closely related enzymes (Collmer and Keen, 1986; Tamaki et al., 1988; Van Gijsegem, 1989). In contrast, no similarity is found between Pel of the two different groups. Since the structuralpel genes of one group are tandemly repeated on the chromosome, it is thought that successive duplications led to the present-day composition of each Pel group. The same duplication situation was observed for two (or three in EC16 strain) proteases which are encoded byprt genes tandemly repeated on the chromosome and sharing a high level of nt sequence similarity, suggesting the occurrence of a duplication event in the course of these genes’ evolution (Delepelaire and

Wandersman, 1989; Dahler et al., 1990). Concerning the cellulases, work in this laboratory has shown that, in Er. chrysanthemi strain 3937 grown in minimal medium, two enzymes are produced, referred to as EGZ and EGY, which account for 95 “/, and 5 y0 of the total activity, respectively (Boyer et al., 1987a,b). The analysis of the EGZ aa sequence revealed the existence of region sharing a good level of similarity (approx. 36:;) with cellulases from organisms taxonomically unrelated to Et-. chqwanthemi such as Bacillus and Clostridium (Guiseppi et al., 1988). This common region was later found to correspond to the catalytic domain (Py et al., 1991). We wcrc interested in knowing whether EGY shares some sequence relatedness with EGZ and whether, as with the pectinolytic and proteolytic enzymes, duplication could account for the evolution of the cellulolytic capacity of the bacterium.

EXPERIMENTAL

AND DISCUSSION

(a) Cloning and sequencing of the celY gene The celY gene was isolated from a genomic bank of Er. chrysanthemi 3937 constructed in the J.L47-1 vector (Boyer et al., 1987b). Various plasmid derivatives were constructed, introduced into E. coli and the resulting transformants scored for the Gel’ phenotype on plates containing M9 medium supplemented with glycerol and CMC. The celY gene was located on a 2. I-kb BanlHI-SphI DNA restriction fragment which, when transferred on a pUCl8 vector, gave rise to the pMH18 plasmid. DNA fragments from pMH 18 were randomly cloned into a M 13mp8 phage and nt sequences were obtained by the dideoxynucleotide chain-termination method of Sanger et al. (1977). Fragments that had escaped the shotgun cloning step were subsequently subcloned and sequenced using the same procedure.

Fig. 1 The nt sequcncc of the EGY protein. sequence

identical

EGY protein marked

of the celY gene and the deduced

EGY aa sequence

to the N-terminal

purified

is indicated

aa sequence

from E. co/i is underlined.

by a downward

arrow.

aa sequence

as ORFI.

The aa

directly determined The processing

The nt and aa are numbered

for

site is

on the left

margin. The inverted repeat sequences in both the 5’-noncoding and the 3’-noncoding regions are indicated by convergent arrows. The sequence matching that of NtrA-dependent promoters is underlined with the consensus -24/-12 positions indicated. The position of the MnlI and 7’aqI restriction

sites which were used for fusing the NtrA-dependent

to the CAT cartridge

are indicated

on the same DNA strand

(see section c for details).

as EGY; ORF3

promoter ORF2 lies

lies on the complementary

DNA strand and the putative RBS and aa sequences are given above the nt sequence (see section b for details). The aa is aligned with second nt of each codon. The celY sequence is deposited M74044 with GenBank, Los Alamos. NM.

under

accession

No.

111 (b) Analysis of the nt sequence

A large ORF that extends

over 1000 bp was identified

(Fig. 1). The ORF ends at nt 997 with UAA stop codon while four in-frame start codons were found, i.e., ATG, at nt -144, ATG, at nt -138, ATG, at nt 1 and ATG, at nt 13 (Fig. 1). The search for the start codon used for celY translation was determined using the following criteria: (i) EGY was purified from the periplasmic space of E. coli; (ii) the N-proximal aa sequence of the purified EGY protein is Ala-Asn-Gly-Trp-Glu-Be-Tyr-Lys-Ser-Arg (Boyer et al., 1987a); (iii) the EGY protein migrates on an SDS-PAGE at 35 kDa, and (iv) upstream from the chosen ATG, an SD sequence should lie within 5 to 9 bp (Storm0 et al., 1982). The N-proximal aa sequence listed in (ii) was found in the region between nt 70 and 99. Taking (i) into account the idea was supported that EGY is first synthesized as a precursor with a cleavable signal sequence. Considering the usual size of signal peptides in bacteria, the first aa of that precursor should be located 20-40 aa upstream from the Ala-Asn-. . sequence (Oliver, 1985). This rules out the use of both ATG, at nt -144 and ATG, at nt -138 which, if used as initiators, would lead to the synthesis of a signal peptide of approx. 70 aa in size. Furthermore, neither ATG, nor ATG, are preceded by an SD sequence. In contrast, both ATG, and ATG, are qualified as start codons for the synthesis of a protein exhibiting signal peptides of 23 and 19 aa, respectively. Furthermore, both are preceded by SD sequences 5 and 6 nt upstream. respectively. We are unable to predict which one of ATG, and ATG, is used for celY translation initiation. Considering the fact that signal sequences are usually made of short charged segments at the N terminus, followed by a large hydrophobic core, does not help to differentiate the more likely possibility, since in both cases, one single positively charged residue is found immediately after the ATG. It is worth noting that the predicted size of the mature protein will be 33 990 M, (309 aa) in very close agreement with that estimated from SDS-PAGE (Boyer et al., 1987a). Two small ORFs were also identified. One, referred to as ORF2, is located upstream from EGY from nt -217 to nt -97 (Fig. 1). Its actual translation seems unlikely since the only sequence resembling an SD lies 13 nt upstream from the ATG (Fig. 1). The second ORF, referred to as 0RF3, was found on the complementary DNA strand, from nt 1115-607 (Fig. 1) and is preceded by an SD sequence. A peculiar feature of this ORF is that it is amazingly rich in positively charged residues: out of 172 aa, 32 Arg residues and 10 His residues are present. Arg-rich motifs have been found in bacteriophage antiterminator proteins recognizing RNA hairpins and a consensus sequence has been proposed (Lazinski et al., 1989). This latter was not found in ORF3 and the role, if any, of this hypothetical peptide requires further characterization.

(c) Identification

of a ntrA-dependent

promoter

Regions flanking the celY coding region were analysed for the presence of transcription signals. The analysis of the upstream region failed to identify a sequence closely matching the consensus sequences of E. co/i 070-dependent promoter. In contrast, a sequence matching the sequences recognized by the E. coli d4-containing RNA polymerase, e.g., GG + 10 bp 4 GC, was to be found at nt -1361-135 and nt -124/-123, respectively (Kustu et al., 1989). TO know whether this sequence was functioning in vivo, the 96-bp M&I-7’uqI fragment (Fig. 1) containing this sequence was fused upstream from the promoter-less car gene, referred to as CAT cartridge (Close and Rodriguez, 1982). This was done in two steps: first, the M&I-TuqI restriction fragment

was introduced

into the SmaI-AccI

sites of a

pUC19 plasmid giving rise to pMH 191, and second, a HindIII-Hind111 restriction fragment carrying the CAT cartridge was cloned downstream from the Er. chrysanthemi MnlI-TuqI restriction fragment in pMH 191. The ligation mixture was used to transform E. coli MC4100 and ApR colonies were selected. Next, the ApR colonies were screened on LB plates containing Cm (30 pg/ml) and a subset of the ApR CmR colonies was analysed for their plasmid content. All were found to contain a pMH191 derivative in which the CAT cartridge had been inserted downstream from the Er. chrysanthemi M&I-TuqI restriction fragment. One plasmid, referred to as pAJ24, was introduced into an E. coli MC4100 derivative strain lacking a functional ntrA gene. Transformants were scored on LB plates containing concentrations of Cm ranging from 15 to 30 pg/ml; in no case was growth observed. This implied that the expression of the cat gene was dependent upon the existence of a functional ntrA gene. By inference we hypothesized that celY gene expression could be controlled by the E. coli d4 holoenzyme. To test this assumption pMH18 was introduced into both NtrA + and NtrA strains. When scored on CMC plates, the NtrA + transformants were surrounded by a typical halo, indicating cellulase activity, while the NtrA _ transformants were not. This result prolonged the hypothesis that the celY gene expression is under the control of an ntrA-dependent promoter. Two potential palindromic structures were identified. One lies downstream from the stop codon, is followed by a run of four T residues and is predicted to exhibit a free energy of formation of -60 kJ/mol if folded in RNA; these features make this sequence a good candidate for acting as a Rho-independent transcriptional terminator (Platt, 1986; Fig. 1). The second structure lies upstream from the coding region (Fig. 1). It is not followed by a run of T residues and as such is unlikely to act as a terminator. Nonetheless, its stability, if formed (free energy of formation of (- 125 kJ/mol) and its location immediately upstream from the ntrA-dependent promoter could endow this structure with a regulatory role yet to be characterized.

112 1 1

1

19

Gl”

---

---

Gly

---

LYS

---

-__

__-

--_

___

---

---

___

---

---

__-

-_-

55

l.y&

---

---

-G-

---

MG

---

--c

*w

__c

__c

-__

__T

__-

--*

___

___

___

55

ATT

---

---

-c-

---

cx.x

---

_-G

TCG

--T

__G

---

--c

__-

-_I2

--_

---

---

19

Ile

---

---

*la

--_

Gly

---

-__

__-

___

___

___

___

___

___

___

___

___

37

___

net

___

___

His

Tyr

*sp

___

___

Ile

___

___

___

*sn

___

___

*sn

___

_-G

A__

---

_--

c**

T-C

G-T

--c

___

*TC

___

___

__T

__c

___

___

A-C

___

109

-4

c__

---

---

GCG

A-T

A-C

-_T

-_-

CCG

___

_-_

__c

__G

--_

___

C-G

-__

37

---

Lc?U

---

---

*la

As”

*srJ

---

-_-

pro

--_

---

--_

LYS

___

___

Gl”

___

55

___

Gin

___

His

___

___

*sn

Thr

Thr

Se=

___

___

___

___

___

___

___

*sp

163

_-_

C-A

---

CA_

---

--G

*--

-c-

*_c

_G_

__c

T-_

__c

___

-__

--T

_-c

G-T

163

--_

G-C

---

AC_

--_

__c

G--

-*-

T-T

_*_

__G

c__

__T

___

___

__c

_-T

A-C

55

---

Asp

---

Thr

---

--_

Asp

LYS

se=

*sn

___

--_

___

___

--_

_-_

_--

*sn

___

---

LYS

__-

Gly

*sn

LYS

___

___

---

I\=”

_--

___

271

___

--T

_--

_--

--_

T__

T-A

-

__-

GG_

__T

*-_

___

___

___

--c

_-T

___

271

-__

--rJ

_--

_--

--_

c__

C-G

CGC

__-

CA_

__G

c-_

___

___

--_

--G

__c!

___

91

--_

---

_--

_--

---

_-_

---

*rg

-_-

Gl”

Lys

Gl”

-_-

--_

---

LYS

_--

---

109

Le”

Gin

---

---

---

Ser

---

Gin

Lys

Ala

Ile

Ile

Ala

Ser

Asn

Ile

Ile

Gln

325

CTG

CA-

--G

--G

--C

AG-

---

CAG

AAS, G-G

A-C

A-C

GCC

AGC

-AT

A-C

A-T

CAG

325

GCC

AC-

-4

--C

--T

GC-

---

*CC

GCC

T-C

C-G

C-G

M

TAT

-CG

G-G

G-G

ACT

109

Ala

Thr

---

---

---

Ala

---

Thr

Ala

Ser

Leu

Leu

Lys

Tyr

Thr

Val

Val

Thr

73 217 217 73

91

127

--_

---

---

---

Thr

___

___

___

___

___

Ala

Tyr

___

___

___

Lys

___

Ser

373

_-T

_-G

--C

___

*CC

___

___

T-G

__C

___

4-C

T-T

--T

__C

___

AJ,G ___

*G_

379

_-C

__C

__T

___

CAG

___

___

C-C

-_G

___

-TG

A-G

--G

--T

___

CGC

___

GA-

_--

Gl"

___

___

___

___

___

"al

Lys

___

___

___

*q

___

Asp

127

Fig. 2.

145

TyT

"al

Il.= ---

-__

___

---

___

_-_

&.u

___

-_-

___

__-

_--

Asp

---

---

433

T-T

G__

-T-

___

___

--G

--G

___

___

C-G

___

___

___

___

__C

_*-

___

__T

433

C_C

C--

-A-

___

___

__C

-_C

___

___

A-C

___

___

___

___

--G

-C-

___

--G

145

His

Leu

*sn

___

___

___

___

___

___

Ile

___

___

___

___

___

*Ia

___

___

163

As"

---

Ser

---

---

Gln

"al

---

---

Gl"

---

Ile

Asp

---

Ser

Le"

Ser

---

487

A-C

-4

-GC

--T

--T

CAG

-TG

---

---

CA-

C--

ATT

GAC

---

A-C

-T-

T-A

T--

487

G-G

--G

-CG

-4

--G

*CC

-CC

---

---

AC-

T--

CAG

AGT

---

G-G

-A-

G-G

C--

163

Gl"

---

Thr

---

---

Thr

Ala

---

---

Thr

---

Gl"

Ser

---

Gly

Gl"

Ala

---

181

"al

---

Glu

---

Arg

Phe

---

Gin

Val

Gly

---

---

Thr

---

---

Ala

---

---

541

G-C

--*

G--

___

C-T

-TC

--T

C-G

GTC

GGG

___

__G

-CG

___

___

_C_

_--

--_

541

C-G

--G

C--

___

G-C

_GG

--G

*-*

TCG

CAT

___

__C

-GC

___

___

_T_

___

___

181

Len

---

Gin

---

Gly

Trp

---

Lys

Ser

His

---

---

Ser

---

---

Val

---

---

199

*s,, ___

___

___

Se=

___

*Ia

___

___

Thr

Ala

___

___

Se=

___

Phe

_-_

Tyr

595

AAC

___

___

___

TC-

___

GC-

___

_-G

-CG

-CC

___

___

T__

__T

T-C

_--

-A_

595

CGG

___

___

___

AA-

___

CT-

___

__C

_a

_AG

___

___

C-_

__G

A-G

__-

-T_

199

Aq

___

___

_--

~ys

___

~,eu ___

__-

~ys

Glu

__-

___

~10

---

,,et ---

Phe

217

_-_

---

11e

---

---

---

---

---

Le"

Tyr

---

Tyr

---

Ala

Lys

Thr

Thr

Ala

649

--C

__C

__T

_-_

___

___

___

___

T-G

-*C

___

TAT

__C

G-C

A-A

_C_

A-G

GC-

649

--T

__G

__C

___

___

___

___

___

A-C

-CG

___

GTC

_-T

C-G

C-C

_G_

G-C

TT-

217

--_

--_

~eu

---

---

---

---

---

Ile

Ser

---

Val

---

Pro

His

Ser

Ala

Leu

235

-_-

"al

---

Phe

Gl,, ~eu

Tyr

Trp

Arg

As,, ---

---

---

---

Thr

---

---

---

703

--G

_TG

___

_TC

C-G

CTG

-AC

TG-

_GT

-AC

__T

-4

---

---

ACG

--G

---

--C

703

__C

-CA

___

-GG

A-A

EC

_GG

AT-

_*G

-GT

--C

__G

___

___

C&A

--T

---

--G

235

--_

Ala

---

Trp

Lys

Ala

Tq,

"et

Gl"

Ser

---

---

---

---

Gin

---

---

---

253

---

"al

Asp

---

Le"

---

ser

---

Thr

___

Thr

Tyr

___

___

Gl,, _--

---

757

___

G-T

G-T

m-G

CTG

---

-GT

---

AC-

--G

A-T

_*C

___

___

C*G

_--

--T

757

___

A-C

A-C

e-T

___

-CC

___

GAG

GT-

__C

C-G

_GG

___

---

GCC

---

--C

253

---

11e

AS"

---

---

Thr

---

Gl"

"al

---

Pro

Trp

---

---

Ala

---

---

270

___

-__

___

_-_

---

-__

---

--_

net

---

Glu

808

T__

___

___

___

__C

_-,J C-G

___

A-G

___

A-C

808

C__

___

___

___

__T

_-T

T-*

___

C-T

___

G-A

270

___

---

---

_--

---

-__

---

---

Le"

---

As"

113 (d) The celY

gene as an example

of interspecies

gene

transfer The EGY deduced aa sequence was compared to those of other cellulases. No similarity was found with EG produced by Er. chrysanthemi and Er. carotovoru, i.e., EGZ and CelS, respectively (Guiseppi et al., 1988; Saarilahti et al., 1990). In contrast, an unexpectedly high level of similarity was found between EGY and an EG produced by Cellulomonas udu, a Gram + cellulolytic bacterium (Nakamura et al., 1986). The level of identity amounts to approx. 58% throughout the first 279 aa in the mature proteins (Fig. 2). A closer analysis revealed the existence of five regions, namely, from aa 4-115, 127-173, 192-226, 245-253 and 266-279, which exhibit a level of identity ranging from 67-81%. These blocks are interrupted by short segments (less than 20 aa) in which the level of similarity drops to less than 20%. This high relatedness between both aa sequences prompted us to perform a similar comparison at the nt sequence level. In the region corresponding to the first 279 aa both nt sequences were found to share an overall identity of 62% (Fig. 2). When considering the reading frames, 57% of the codons (100/279) were found to be identical, while 23% (64/279) were conservatives. Among the latter, a vast majority (55/64) differed only by the nature of the nt at the third position (Fig. 2). Given the phylogenetic distance between Er. chrysanthemi and C. uda, such a similarity suggested that the gene had been exchanged between both bacteria. The question arose as to which organism was the donor. An answer might be obtained by analysing the codon usage of the two genes studied and comparing it with the codon usage prevailing in both organisms (Van Vliet et al., 1988). Unfortunately, too few genes in either of the two bacteria have been characterized and even though the codon usage of Er. chrysanthemi might be close to that of E. cofi, the use of such a criterion might be misleading. Another possibility is to use the GC% as a criterion. The C. uda chromosome is thought to be G + C-rich, approx. 72% while that of Er. chrysanthemi is approx. 57 % (Lelliot and Dickey, 1986; Stackebrandt and Keddie, 1986). The GC% value of the celY gene is 58%, in close agreement with that of the Er. chrysanthemi chromosome. In contrast, the GC% value of the C. uda gene is 6 1y0 , i.e., much lower than that of the host chromosome and quite close to that of Er. chrysanthemi. Hence it is tempting to speculate that the celY gene

Fig. 2. Comparison glucanases. sponding

of the

to the ET. chrysanthemi cellulase,

the aa and the nt sequences between details).

Er. chrysanthemi and

the

C. uda endo-

The upper two lanes are the aa and the nt sequences of the C. uda cellulase.

both nt or aa sequences

corre-

while the lower two lanes are

are indicated

Only the differences

(see section d for further

was recently

acquired

by C. uda from Er. chrysanthemi.

In

this context it is interesting that both organisms are soil bacteria and that strains of Cellulomonas have been isolated from rotting

sugar-cane

stalks.

(e) Conclusions (1) Comparison of the aa sequence, as deduced from the celY gene sequence and the N-proximal sequence of the purified EGY protein, shows that EGY is first synthesized as a precursor with a cleavable signal sequence. We wish to note that analysis of Er. chrysanthemi cultures by Western blot, using anti-EGY antibody, and by enzymatic test did not allow us to assess whether EGY was extracellularly secreted or was periplasmically located; the reason for this failure is linked to the poor level of expression of the celY gene, at least in usual laboratory growth media. (2) The celY ORF is preceded by a sequence matching -24 and -12 regions of promoters the consensus recognized by the d4 containing RNA polymerase and cefY gene expression in E. coli was indeed shown to be dependent upon the presence of a functional ntrA gene. It was proposed that ntrA-dependent promoter function requires the action of an activator, itself activated under particular growth conditions, e.g., nitrogen limitation and anaerobiosis (Kustu et al., 1989). Varying several growth parameters such as O,, iron, osmolarity, carbon sources, did not permit a better expression with the exception of phosphate starvation where a threefold increase was observed (Aymeric et al., 1988; unpublished data). Similarly, no cultural conditions could be defined that would strongly influence EGZ synthesis (unpublished). The next step will be to study celY gene expression within a bacterium growing in the presence of plant compounds. (3) The EGY sequence is totally unrelated to that of the other Er. chrysanthemi cellulase EGZ. Hence, in contrast to the situation prevailing with the pectinolytic and proteolytic activity, duplication is unlikely to have been instrumental in strengthening the cellulolytic activity of Er. chrysanthemi. In this context, it might be interesting to note that homologous pel and homologous prt genes are tandemly repeated while the ceIZ and celY genes are located on different areas in the chromosome (Aymeric et al., 1988). (4) The EGY sequence displays a strikingly high level of similarity with an EG of C. uda. Cellulases have been classified in seven families by use of hydrophobic cluster analysis and the similarity with the C. udu EG undoubtedly places EGY in family D (Henrissat et al., 1989). The nt sequence of both structural genes is remarkably similar. Taken into account the taxonomic distance between the two organisms and the G + C% values of both the cellulase genes and the host chromosomes, we propose that celY constitutes a typical example of horizontal transfer from Er. chrysanthemi to C. uda.

114 ACKNOWLEDGEMENTS

Cellulases families revealed (1989) 83-95.

We wish to thank Marc Chippaux for his help and interest in this study. Thanks are due to Jean Pierre Belaich for pointing out the similarity between C. udu and Er. chrysanthemi enzymes.

This work was supported by grants from the Research Contract BAP 021 l/F of the European Communities, from the CNRS and from the Fondation pour la

Recherche

Medicale.

Kotoujansky, Annu.

A.: Molecular

chanism. Lazinski,

Microbial.

D. and Weiss. D.: Expression

genes is probably

united by a common

D., Grzadzielska,

E. and Das, A.: Sequence-specific

nition of RNA hairpins

by bacteriophage

conserved

motif. Cell 59 (1989) 207-218.

arginine-rich

R.S.: Erwinia.

antiterminators

of Systematic

and Wilkins,

MD, 1986, pp. 469-476.

Baltimore,

K., Norihiko,

Bacteriology,

M. and Kitamura,

J.L., Guiseppi,

A., Pascal,

M.C. and Chippaux,

and regulation of the ccl genes in Erwinia Genet. 211 (1988) 95-101. Boyer, M.H., Cami, B., Chambost, Characterization

of a new endoglucanase 162 (1987a)

Mol. Gen.

M. and Cattaneo,

J.:

Cellulase

Boyer, M.H., Cami, B., Kotoujansky, A., Chambost, J.P., Frixon, C. and Cattaneo, J.: Isolation of the gene encoding the major endoglucanase from Erwinia 351-356. Chatterjee,

Microbial.

Lett.

41 (1987b)

A.K.: Genetics

to phytopathogenic

bacteria.

of pathogenicity

and characterization

gene cartridge:

mapping

of the

a new approach

of extrachromosomal

elements.

to the Gene

20

Dahler,

A. and Keen, N.T.: The role of pectic enzymes Annu. Rev. Phytopathol.

G.S.,

Barras,

extracellular Bacterial. Delepelaire,

metalloproteases

in plant patho-

24 (1986) 383-409.

F. and Keen,

N.T.: Cloning

from Erwinia

of genes

santhemi. Protease

chrysanthemi

importance

of gene expres-

I., Haiech,

J., Chippaux.

M. and Barras,

of His98 and Glu133 residues

organization

for catalysis.

F.: and

Prot. Eng. 4

(1991) 325-333. Saarilahti,

H.T..

Henrissat,

identified

Stackebrandt,

B. and Palva,

from Erwinia

E. and Keddie,

Vol. 2. Williams Sanger,

F., Nicklen,

E.T.: CelS: a novel endosubsp. coromvora. Gene

caromvoru

R.M.: Cellulomonas.

terminating

and Wilkins,

inhibitors.

Proc.

In: Sneath.

of Systematic

Baltimore,

S. and Co&son,

C.: Protease

endoglucanase

secretion

by Erwinia

and secreted

Z from Erwinia

chry-

as zymogens

chrysanthemi

from Bacillus subtilis and alkalophilic

N.: and

Bacillus.

Mol.

P., Lemesle, L. and Mornon.

J.-P.:

M., Tomme,

G.D.,

Schneider,

translational

initiation

Tamaki,

Natl.

P.H.A.

Bacteriology,

MD, 1986, pp. 1325-1329.

A.R.: DNA sequencing Acad.

Sci.

T.D. and Gold,

wjith chain-

IJSA

74

(1977)

L.M.: Characterization

sites in E. co/i. Nucleic

S.J.. Gold, S., Robeson,

EC16. J. Bacterial.

of

Acids Res. 10 (1982)

Van Gijsegem,

F.: Relationship

cluster in Erwinia 1415-1424.

M., Manulis,

of the pel genes

S. and Keen. N.T.: Strucfrom Erwinirr

chrysanthemi

170 (1988) 3468-3478. chrysumhemi

between

the pel genes of the pe/ADE

strain B374. Mol. Microbial.

Van Vliet, F., Boyen, .4. and Glansdorff.

N.: On interspecies

fer: the case of the NrgF gene of Escherirhia 139 (1988) 493-496.

2 (1988) 159-164.

B., Claeyssens,

Stormo,

ture and organisation

B and C are synthesized

between

endoglucanases Microbial.

J.

EC16.

without a signal peptide. J. Biol. Chem. 264 (1989) 9083-9089. Guiseppi, A., Cami, B., Aymeric, J.L., Ball, G. and Creuzet,

Henrissat,

and the regulation

55 (1986) 339-372.

2971-2995.

encoding

172 (1990) 5803-5815. P. and Wandersman,

Homology

termination

Rev. Biochem.

5463-5467.

(1982) 305-316. genesis.

39

90 (1990) 9-14.

factors:

Adv. Plant Pathol. 4 (1986)

R.L.: Construction

chloramphenicol-resistance transcriptional Collmer,

in Escherichio cd. Annu. Rev. Microbial.

and Holt, J.G. (Eds.), Bergey’s Manual

Close, T.J. and Rodriguez,

of a cellulase

(19X6)247-254.

EGZ of Erwinin chr_vsanrhemi: structural

glucanase

A.K. and Vidaver,

application l-213.

FEMS

chrysanfhemi.

sion. Annu.

Py. B., Bortolli-German.

from Erwinia chrysanthemi.

311-316.

1

(1985) 615-648. Platt, T.: Transcription

J.P., Magnan,

Eur. J. Biochem.

Oliver, D.: Protein secretion

M.: Mapping

chrysanthemi.

a

Vol. 1. Williams

K.: Sequence

gene of Cel/u/omorzcrs u&r CB4. J. Biotechnol. Aymeric,

recogrequires

In: Krieg, N.R. and Holt, J.G.

(Eds.), Bergey’s Manual Nakamura.

me-

Rev. 53 (1989) 367-376.

Lelliot, R.A. and Dickey,

REFERENCES

by soft-rot erwinias.

25 (1987) 405-430.

E.. Keener, J., Popham,

of d4 (ntrA)-dependent

Gene 8 I

cluster analysis.

genetics ofpathogenesis

Rev. Phytopathol.

Kustu, S., Santero,

by hydrophobic

3 (1989)

gene trans-

cdi. Ann. Inst. Pasteur