Nucleotide sequences of the Erwinia chrysanthemi ogl and pelE genes negatively regulated by the kdgR gene product

Nucleotide sequences of the Erwinia chrysanthemi ogl and pelE genes negatively regulated by the kdgR gene product

Gene, 85 (1989) 125-134 Elsevier 125 GENE 03295 Nucleotide sequences of the Enviniu clrvysantlremi ogl and pelE genes negatively regulated by the k...

1MB Sizes 0 Downloads 16 Views

Gene, 85 (1989) 125-134 Elsevier

125

GENE 03295

Nucleotide sequences of the Enviniu clrvysantlremi ogl and pelE genes negatively regulated by the kdgR gene product (Recombinant DNA; pectin degradation; soft-rot; phytopathogen; site)

promoter; terminator; repressor-binding

Sylvie Reverchon, Yue Huang, Claude Bourson and Janine Robert-Baudouy Laboratoire de G&&ique Molkculaire des Mcrootgankmes, Instihrt National des Sciences Appliqukes, 69621 Villeurbanne (France) Received by J. Davison: 19 April 1989 Revised: 2 July 1989 Accepted: 3 July 1989

SUMMARY

The nucleotide sequences of the coding and regulatory regions of the genes encoding oligogalacturonate lyase (OGL) and pectate lyase e isoenzyme (PLe) from Erwinia chrysanthemi 3937 were determined. The ogl sequence contains an open reading frame (ORF) of 1164 bp coding for a 388~amino acid (aa) polypeptide with a predicted A4, of 44 124. A possible transcriptional start signal showing homology with the Escherichia coli promoter consensus sequence was detected. In addition, a sequence 3’ to the coding region was found to be able to form a secondary structure which may function as an Rho-independent transcriptional termination signal. For the pelE sequence, a long ORF of 1212 bp coding for a 404-aa polypeptide was detected. PLe is secreted into the external medium by E. chrysanthemi, and a potential signal peptide sequence was identified in the pelE gene. In the 5’ upstream pelE coding region, a putative promoter resembling E. coli promoter consensus sequences was detected. Furthermore, the region immediately 3’ to the peZE translational stop codon may function as an Rho-independent translational termination signal. In strain 3937, the synthesis of OGL and PLe, as well as the other enzymes involved in the pectin-degradative pathway (particularly the kdgT product), are known to be regulated by the KdgR repressor, which mediates galacturonate and polygalacturonate induction. Synthesis of these enzymes is also regulated by the CRP-CAMP complex which mediates catabolite repression. Analysis of the regulatory regions of ogl and pelE allowed us

Correspondence to: Dr. S. Reverchon, Laboratoire de Genttique Moltculaire des Microorganismes, Batiment 406, Institut National des Sciences Appliquees, 20 Avenue Albert Einstein, 69621 Villeurbanne (France) Tel. 78-94-80-88; Fax 72-44-08-00.

Abbreviations: aa, amino acid(s); Ap, ampicillin; bp, base pair(s); /?Gal, B-galactosidase; CAMP, cyclic adenosine monophosphate; CRP, catabolite activator protein; IPTG, isopropylp-D-thiogalactopyranoside; kb, kilobase or 1000 bp; KDG, 0378-I 119/89/$03.50

0 1989 Elsevier

Science Publishers

B.V. (Biomedical

2-keto-3-deoxygluconate; KdgR, general repressor of genes involved in pectinolysis; kdgT, gene encoding KDG permease; Km, kanamycin; LB, Luria-Bertani medium; nt, nucleotide(s); OGL, oligogalacturonate lyase; ogl, gene encoding OGL; ORF, open reading frame; PGA, polygalacturonate; PL, pectate lyase; PLe, pectate lyase e isoenxyme; peE, gene encoding PLe; R, resistant; RBS, ribosome-binding site(s); Tc, tetracycline; wt, wild type; XGal, 5-bromo-4-chloro-3-idolyl-/I-D-galactopyranoside; ::, novel joint/fusion; [ 1, denotes plasmid-carrier state. Division)

126

to identify possible CRP-binding sites for these two genes. Furthermore, comparative study of the regulatory regions of the ~1, kdgT and peiE genes revealed the existence of a highly conserved sequence which could correspond to a whole or partial KdgR-b~d~g site.

the pme and pelA.B,C,D,E Erwinia chvsanthemi is a pathogenic enterobacterium causing soft-rot disease in plants. It has been demonstrated that pectolytic enzymes produced by this organism are directly involved in pathogenesis (Collmer and Keen, 1986). Pectin is the major component of plant cell walls and its degradation leads to maceration of plant tissue. During the last live years, the pectin catabolic pathway has been exten(Hugou~eux-Cotte-Pattat and sively studied Robert-Baudouy, 1987). In E. chrysanthemi strain 3937, there are twelve structural genes whose products are likely to be required for pectin degradation. These genes include pme (pectin me~ylesterase), peL4, pelB, pelC, pelD, peIE (five isoenzymes of pectate lyases which are differentiated by isoelectric focusing), ogl (oligogalacturonate lyase), kdul (4keto-S-deoxyuronate isomerase), kduD (2-keto-3deoxy~uconate o~doreductase), MgK (2-keto-3deoxygluconate kinase), kdgA (2-keto-3-deoxy-6phosphogluconate aldolase), and kdgT (2-keto-3deoxygluconate permease). All of these genes constitute independent transcriptional units and are distributed over five difIerent regions of the E. chrysanthemi chromosome: pelA, D, E and pme are near the pro marker; pe1B.C are between the ile and leu markers; ogl, kdgT and kduD are near tip; kdgA is between trp and his; kdgK is near xyI; and kduI has not yet been localized (H~ou~eux-Cotte-Pattat et al., 1989). All of these genes, except for kdu1, have been cloned: pme, and pelA,B,C,D,E (Kotoujansky et al., 1985), ogl, kduD (Reverchon and RobertBaudouy, 1987a), kdgT (Condemine and RobertBaudouy, 1987b), kdgA and kdgK (HugouvieuxCotte-Pattat, unpublished results). Furthermore, it has been shown that expression of these genes is under the control of the negative regulatory kdgR gene, which mediates induction by PGA and galacturonate (Condemine and Robert-Baudouy, 1987a). Whereas kdgR seems to be the unique regulatory gene governing kdgA, kdgK and kdgT expression,

genes are only partially regulated by the kdgR gene product, and it seems that several regulatory elements exist for this second group of genes (Hugouvieux-Cotte-Pattat et al., 1986; Reverchon and Robed-Baudouy, 1987b). In this study, we decided to sequence the ogi and pelE genes for two reasons: (1) the ogl and pelE gene products have the same catalytic activity, i.e., both cleave by transelimination of the CIl-4 glycosidic bond between two galacturonate residues. We wished to determine if there were some conserved regions in the polypeptide sequences of OGL and PLe which could be involved in the catalytic sites of these enzymes. (2) Our second purpose was to determine whether the regulatory regions of ogl and pefE present common sequences. Indeed, since the kdgR gene product appears to act on the expression of several genes of the pectin-degradative pathway, we assume that these genes may possess a common KdgR-bind~g site.

MATERIALS

AND METHODS

(a) Bacterial strains, plasmids, media and growth conditions The bacterial strains and plasmids used in this study are shown in Table I. E. ch~sant~mi and E. coli cells were usually grown at 30” C and 37°C in LB medium or in M63 minimal medium (Miller, 1972), respectively, supplemented with a carbon source (0.2%) and when required, with aa (40 pg/ml) and antibiotics at the following concentrations: 50 pg Ap/ml; 20 pg Km/ml; 10 ,ug Tc/ml. Recombinant pBluescript plasmid derivatives were identilied using the fiGal complementation assay on Ap-LB medium containing XGal indicator and IPTG inducer. The presence of ogl on the various constructed plasmids was tested by their ability to restore growth, on PGA as sole carbon source, to an

127 TABLE I Bacterial strains and plasmids Strain

Genotype or description”

Source or reference

reo4l,lac~,endAl,gyrA96,thi,hsdRl7,supE44,relAl,[F’proAB+,

Stratagene

E. coli

XLl-blue

IacIQ, lacZAM15, TnZO] E. chrysanthemi 3937

wt strain isolated from Saint Paulia

A595

Inn-T’, IacZ, arg-10, ogl

Reverchon and Robert-Baudouy (1987a)

ApR, IacZ’ pBR322 derivative with 1.5-kb HindIII-PstI fragment harboring a part of ogl gene pBluescript with 2.3-kb NsiI-PstI fragment harboring ogl gene of strain 3937 As pOGLl0, but with the 2.3-kb fragment inserted in the opposite orientation relative to the Zac promoter of pBluescript pBR322 derivative with 1.8-kb NsiI-NaeI fragment harboring a part of ogl gene ApR, IacZY KmR cassette pBluescript derivative with ogl::ZucZY KmR protein fusion pBluescript with 1.7-kb DraI-DraI fragment harboring pelE gene of strain 3931 As pPLeD1, but with the 1.7-kb fragment inserted in the opposite orientation relative to the lac promoter of pBluescript

Stratagene Reverchon and Robert-Baudouy (1987a)

Plasmids

pBluescript pOGL9 pOGLl0 pOGLl1 pOGLl2 pLKC48 1 pOGLl3 pPLeD 1 pPL.eD2

This work This work This work Tiedeman and Smith (1988) This work This work This work

a Genotype symbols are according to Bachmann (1983). ImrT” indicates that the transport system encoded by the gene 1mrT which mediates entry of lactose, melibiose and rafSnose into the cells, is constitutively expressed. lacZ’ indicates that 3’ end of this gene is truncated.

ogl mutant. For pelE subcloning experiments, plasmid constructs were tested for PL activity using the PGA agar plate detection assay, as previously described (Reverchon et al., 1985). (b) Plasmid constructions

Standard molecular cloning techniques employed in this study (small- and large-scale plasmid DNA extraction, restriction enzyme digestion, agarose gel electrophoretic analysis, DNA ligation and transformation of E. coli cells) were performed as described by Maniatis et al. (1982). DNA-modifying enzymes Boehringer-Mannheim were obtained from Biochemicals or from New England Biolabs. Introduction of plasmid DNA into E. chrysanthemi was performed by electroporation.

RESULTS AND DISCUSSION

(a) Sequence of the ogl gene

Subcloning experiments from plasmid pOGL1 (Reverchon and Robert-Baudouy, 1987a) have allowed us to construct the plasmids, pOGLl0 and pOGL11, carrying a 2.3-kb %I-PstI fragment inserted in both orientations with respect to the pBluescript lac promoter. Both plasmids complemented the ogl mutation, leading us to conclude that we had cloned the entire gene and its promoter region. The complete nt sequence of this fragment was determined from both strands (Fig. 1A). The observed sequence contains 450 bp of the C-terminal end of Mu introduced by in vivo cloning of ogl using RP4 : : mini-Mu (Reverchon and Robert-Baudouy,

a ‘3

2::

%



---P-A 100 DP

u Fig. 1. Sequencing strategy for the ogl and peIE genes. For nt sequence analysis, a nested series of deletion clones was created using the exonuclease III/m~g bean nuclease method of Henikoff (1984) with the pBluescript plasmids. In addition, some DNA fragments were subcloned using restriction endonucleases. Sequencing was done by the chain termination method on double-stranded DNA templates. Two complementary universal 17-nt oligodeoxyribonucleotides (Ml3 primer and Ml3 reverse primer from Boehringer) were used as primers and [ ?S]dATP (Amersham) was used to label the products. Extension of primer was done with Polfk or, in the case of compression regions, with reverse transcriptase. The resulting data were analyzed using the Mac Molly (SoRGene, Berlin) program. The orientation and length of thin arrows indicate the direction and extent of sequence derived from each deletion. (A) The NsiI-PstI fragment carrying ogl has been subcloned from plasmid pOGL1 (Reverchon and Robert-Baudouy, 1987a) and has been sequenced on both strands. This fragment contains 450 nt of the C-terminal end of Mu introduced by in vivo cloning of cgt using RP4::~i-Mu. The length of the ogl gene and the direction of transcription are indicated by a hatched arrow. (8) The majority of the sequence of the 1.7-kb DruI fragment carrying peLE was determined on both strands, except for one small region. Where information from only one strand was obtained, it was from areas of the sequencing gels in which the reading was unambiguous. The boxes at the top represent the sequence determined on both strands. The length of the peL!Xgene and the direction of transcription are indicated by a hatched arrow.

1987a). In this sequence, a single long ORF of 1164 bp from nt 271 to 1435 was detected (Fig. 2A), with a direction of transcription from the Hind111 site to the PstI. This ORF potentially codes for a 3%aa polypeptide with a deduced IV, of 44 124, which

probably corresponds to OGL based on further ogl subcloning experiments. Indeed, plasmids, pOGL9 and pOGL12 (harboring the NindIII-PstI fragment and NsiI-NaeI fragment, respectively), were shown to be unable to complement the ogl mutation, indicating that the ogf gene overlaps the Hind111 (285 bp) and NueI (1392 bp) sites. A purine-rich Shine-Dalgamo sequence (Shine and Dalgarno, 1974) was located at nt 259, appropriately positioned 5’ to the presumed translational start codon. Possible transcriptional initiation signals were found at nt 111 (GTGAAA, -35) and nt 139 (TAAAAT, -10). Comp~son of this putative ogi promoter with the consensus promoter of E. coli (Hawley and McClure, 1983) showed identity in three out of six nt in the -35 sequence and in tive out of six nt in the -10 sequence. The spacing between the -35 and -10 sequences was 17 bp, which is typical for strong promoters (Hawley and McClure, 1983). The region imme~ately 3’ to the tr~slation~ stop of the ogi gene (nt 145 1-1474) presents a stretch of T preceded by a G + C-rich dyad symmetry which is typical of Rho-independent transcription termination sites (von Hippel et al., 1984). The ogl transcriptional direction and reading frame were confirmed by constructing, an ogl: : l&Z fusion. The 6.3-kb HindIIINruI lucZY KmR cassette of plasmid pLKC481 was inserted between the Hind111 (286 bp) and H&c11 (1106 bp) sites of the ogl gene, leading to the plasmid pOGL13. In E. coli XLIblue[pOGL13], jIGal synthesis was not induced by galacturonate and was reduced by twofold in the presence of glucose (Table II). The absence of induction of the fusion by g~act~onate in E. co& may reflect a lack of interchangeability between the E. coii and E. chryxmthemi kdgR gene products or alternatively, the high copy number of pOGL13 may hide the regulatory effects. In contrast, the E. coli CRP-CAMP system appears to recognize the ogl CRP-binding site. The characterization of the ogi gene product by the bacteriophage T7 RNA pol~~ase/T7-promoter system (Tabor and Richardson, 1985) revealed two polypeptides of 44 and 41 kDa, respectively (data not shown). This result suggests that the 41-kDa polypeptide could correspond to the mature OGL obtained after cleavage of the 44-kDa precursor, indicating that OGL could be periplasmic.

129 TABLE II Expression of ogl::lac fitsion carried on pBluescript derivative in Escherichiacoli Strain[plasmid] a

Carbon source

Inducer b

XLIblue[pLKC%l]

Glycerol Glycerol Glucose

None Galacturonate None

0.1 0.3 0.2

XLIblue[pOGLlO]

Glycerol Glycerol Glucose

None Galacturonate None

0.2 0.4 0.2

XLIblue[pOGL13] 0gl::lacZ

Glycerol Glycerol Glucose

None Galacturonate None

f3Ga.lactivityc

64 51 28

a The plasmid pLKC481 bore the la&Y KmR cassette without a promoter, pOGLl0 bore a functional ogl gene and pOGL13 bore au 0gl::lacZ fusion. b Cells were grown at 3O’C for 14-16 h in the presence or absence of the inducer, galacturonate, at 0.2% (w/v). Cells were then toluenized and the resulting extracts used for @Gal assays (Miller, 1972). Each experiment was repeated three times and the data presented in this table are those of a representative experiment. c Specific activity is expressed as nmol of o-nitrophenol liberated from o-nitrophenyl-#i-D-galactopyrauoside/min/mg of bacterial dry weight.

(b) Sequence of the pelE

gene

Using the cloning data from the B374 E. chtysanthemi peiD, pelE and peiA genes (Reverchon et al., 1986), we subcloned a 3.6-kb ~~~dIII-C~ff~ fragment bearing the pelE and pek4 genes of E. chrysanthemi strain 3937 and then a 1.7-kb DraI-DruI fragment containing only the pelE gene. The sequencing strategy for peiE is shown in Fig. 1B and the nt sequence of the &@I fragment is presented in Fig. 2B. A single ORF from nt 251 to 1462 was detected (Fig. 2B) with a direction of transcription from the EcoRI to the HpaI sites as previously predicted by mini-Mu insertion (Reverchon et al., 1986). This ORF codes for a 404-aa polypeptide with a deduced M, of 43095. Upstream from the presumed start codon, we have identified a purine-rich sequence at nt 239 that would correspond to an RBS. We have also found a putative promoter showing strong agreement with the E. coli promoter consensus sequence. Indeed, the -35 region at nt 120 (TTCACA) is highly conserved since five nt out of six are homologous to that of the consensus sequence, and the -10 region at nt 142 (CATAAA) differs by two nt from the consensus. In addition, the spacing between the -35 and -10 regions was 17 bp (Fig. 2B) and as described by Drew et al. (1985)

from Gr~-negative bacteria, an A + T-rich region was observed around the presumed pelE promoter. Indeed, the A + T content of the promoter-flanking region from nt 1 to 119 is 79%, whereas the overall A + T content for the pelE gene is 5 1% . The 3’ end of the peIE gene is indicated by a TAA stop codon at nt 1463, immediately followed by a classical Rhoindependent transcriptional termination signal (nt 1485 to 1511) (Fig. ZB). The PLe being secreted in the external medium by E. chlysanthemi, a potential signal peptide sequence has been identified with a typical cleavage site (/): Ala-Ser-Ala/Ala, that would give a leader peptide of 41 aa. This classical cleavage site presumably accounts for the fact that PLe is efficiently exported to the periplasm of E. coli cells (Reverchon et al., 1985). The length of this presumed PLe signal sequence is rather long, since, in prokaryotes, signal sequence length is generally between 21 and 23 aa (von Heijne, 1985). (c) Comparison of OGL and PL sequences The E. chrysanthemi 3937 PLe aa sequence has been compared with the E. chrysanthemi EC16 PLe, PLa, and truncated PLd aa sequences (Tamaki et al., 1988; Keen and Tamaki, 1986). Our sequence was found to possess considerable similarity with all

130

131

three EC16 isoenzymes. However, the 3937 PLe showed stronger similarity to the EC16 truncated PLd than with to PLa or PLe. From these results, we believe that the truncated PL of strain EC16 more likely corresponds to the PLe of strain 3937. If we supposed that the EC16 truncated pel gene corresponds in fact to pelE’, we could see that the genetic organization of the peL4DE gene family is the same in the two strains (Fig. 3). In addition, Tamaki et al. (1988) previously observed that two short regions of conserved aa were present in the four PL isoenzymes of E. chtysanthemi EC16, as well as in the two PL isoenzymes of Erwinia carotovora subsp. atroseptica. These conserved regions also occur in the E. chrysanthemi 3937 PLe as well as in the 3937 PLa (C.B., unpublished results). I =

a I

II

I

I

I

I

--WIA

I

II

MIE

III

I

Id

I

3937

PIID

I

EC16

--_)aPIlA

pa7

PIIE

p.,A

prlE

PdD

TamaLl’s

nomenclatura

according 5937 nomc”clat”re

Fig. 3. Comparative gene organization of the pefA, pelE and pelD genes in E. chrysmthemi strains EC16 and 3931. In strain EC16, a truncated pelD gene has been detected based on sequence data and peL4DE gene organization was previously established by Tamaki et al. (1988). In strains 3937 and B374, peL4DE organization was determined by Reverchon et al. (1986) based on the polarity of expression of lacZ insertions. In strain 3937, the orientation of pelA has been redefined using nt sequence data (C.B., unpublished results) and is contrary to that suggested previously using the 1acZ fusion technique (Reverchon et al., 1986). In this strain, all three genes, peL4, pelE and pelD, were usually transcribed in the same direction, as shown in this figure. By sequence comparison (RESULTS AND DISCUSSION, sectionc) it seemed that the EC16 truncated pel gene corresponds, in fact, to pelE’ according to nomenclature adopted for strain 3937 which would suggest that genetic organization of the peL4DE gene family is similar in strains EC16 and 3937.

Comparison of the OGL and PL aa sequences did not reveal any homologous regions. This result suggests that the catalytic sites of these enzymes are probably different, although they both cleave by transelimination of the al-4 glycosidic bond between two galacturonate residues. (d) Structure regions

of three kdgR-regulated

promoter

Since expression of the ogl and pelE genes is subject to catabolite repression, we looked for the existence of a CRP-binding site. For ogl, a sequence showing similarity to the consensus E. coli CRPbinding site (de Crombrugghe et al., 1984) was found at nt 28 to 48 (Fig. 4). This sequence contains a TCTGA block instead of the highly conserved consensus pentamer, TGTGA, and a second block which is imperfectly symmetric: TCTGG. For pelE a potential CRP-binding site has been identified, immediately 5’ to the -35 region (Fig. 4). We have also noticed interesting features in the regulatory regions of ogl, pelE and kdgT. In ogl, we have found two palindromic sequences: TTTTATAAAA and TACCGGTA and two dyad symmetries: GTTTTAT(AAA)ATAAAAC and GAAAC(GTT)GTTTC (Fig. 4). Such structures are very often involved in regulatory mechanisms in prokaryotes. Indeed, operator sites generally show symmetries which relate to the fact that regulatory proteins act in dimer form (Gicquel-Sanzey and Cossart, 1982). A rather long direct repeat of the sequence, GAAAAATGAAA, was also found in the ogl regulatory region. Concerning the pelE regulatory region, we have detected one palindromic sequence upstream from the presumed promoter, TITITATTAATAAAAA, and one instance of dyad symmetry between the promoter and the initiation codon: CAAATGG(ACA)CCATITG. Neither of these particular sequences are common to the two genes: ogl and pelE, however the ogl palindrome, TITTATAAAA, and the pelE palindrome, TTTTTAT-

Fig. 2. Nucleotide sequences determined for ogl (A) and pelE (B). The sequences are those of the noncodmg strands. The deduced aa sequences of OGL and PLe are shown and selected restriction sites are specified. Translational start and stop codons are boxed. The putative RBS (S.D.) and potential -10 and -35 regions of promoters are underlined. The potential Rho-independent termination sites are indicated by facing arrows representing inverted repeat of the G + C-rich dyad symmetry preceding the poly(T) sequence. (A) The first 14 nt correspond to the C-terminal end of Mu introduced by in vivo cloning of ogl by RP4 : : mini-Mu. (B) The upward arrow indicates the potential signal sequence cleavage site of PLe.

132 w

CNP G-‘XG-G&G-G..T-ATT-TAT-TCT-GAC-GTT-TGG-TCT-GGG-~-TTA-TTG-CCG-

Potmnthl.

Rreeot.rN

-3s Consanm~

-10

Ogl

ogl

G+A-Tc~TTT-~L*-AAA-TAA-AAC-CAC-GAT-CACkdgR box?

TTT-TTC-ACG-&TA-i,GC-GCT-MG-GAT-TTA-C

kdgR box?

TTGACA (N)l?

TATMT

GTGiMA

TAAAAT

,N,17

p.d*

TTCACA (N)17

kdgT

TTGACG (NI 17 TATTTT

C?+TAAi,

CCC-AM-GGT-T,P.A-MG-CTT-TCT-TCT-TTT-TCG-

ND

Potential CR+bindha

TTT---ACT-TCA-TTT-TTA-ATT-TAR-TCA-TM-A CRP

TAC-AGT-~-TA~TTP-ATT-TTT-TAT-TM-TM-~~~-T~-T~-T~-TCG-T~~-~~-10

COnSenS”

+:+GT+={ T-GM

CM-AATLAGA-CAC-TCA-ACC-GCR-TM-ACA-TTC-G~-~-G~-MA-

pelE

nit08

-33

cql

AAN=

TAT

m

NNTANNI

CGTTTU:

NTT

m

CTG

kdgRbox? ND -W\A-CTC-ACG-TAT-GK-TTC-CGT-TTC-AAC-ACA-

Potential

GTT-MC-CCC-GGC-GCT-GGh-GCC-GI\T-GCA-GCC--

K&R-binding

nitam

-33 CGDTM-WVt-CtC-GTA-CTA-CGG~~T~A~-AGC-~-TCG-~C~TC-~~T-~-C~ffG-10

kdg7-

Consensus

ATGAAA fN!S

TTTCAT

oqu

ATGAAA 1N,5

TTTTAT

og12

ATGAAA IN15

TTTCTA

pelE

ATGAAA ,N, 5 TTTCGT

kdqT

MGAAA

GCDAAC-GTA-A’PC-CTA-~T-TCA-~T=G=-~G-ATA-~G-TTT-~T=TA-~T-CAT-~T-CTT-~-

TTC-GAT-ATA-MT-GAA-ATA-ATG-ATC-TAC-TAA-

+~ZZ+*TT~J~~TT-GTT-TGC-AAG-

kdgR box?

ORF

C~-TCI\-CTT-TTC-TCT-TCC-GGC-TM-ACT-G-TG

43AT

GGC-TGT-

,815

TTTCAT

#D

Fig. 4. Comparison of regulatory regions of the three genes, pelE, kdgT and ogl, controlled by the KdgR repressor. The presented sequences are those of the noncoding strand. Translation start codons are indicated by ORF. The putative RRS and the potential -10 and -35 regions of promoters are underlined and marked SD. CRP indicates the posstbie CRP-binding sites. The spiral lines refer to the palindromic sequences and dyad symmetries mentioned in RESULTS AND DISCUSSION, sectiond. The potential KdgRrepressor-binding sites are boxed.

TAATAAAAA, were both exclusively constituted by A and T. Such extremely A + T-rich sequences were also observed in the regulatory regions of the E. chrysanthemi B374 pelE and pelD genes (F. Van Gijsegem, personal communication) and could be important for regulation of these genes. However, since such sequences did not occur in the regulatory region of another MgR-regulated gene, kdgT (Allen et al., 1989), these A + T-rich sequences probably do not correspond to a KdgR-bong site, Further comparison of the regulatory regions of ogl, peZE and kdgT revealed the existence of a highly conserved sequence, which could correspond to a KdgR-binding site (Fig. 4). From this observation, we propose the potential KdgR-binding site con-

sensus sequence, ATGAAA(N),‘MTCAT. This sequence presents the typical symmetry of operator sites (Gicquel-Sanzey and Cossart, 1982). In ogl, we found two potential KdgR-binding sites, one between the -35 and -10 regions and the second between the presumed promoter and the initiation codon. Of interest is that a part of the proposed KdgR consensus sequence, ATGAAA, is included in the direct repeat, GAAAAATGAAA, found in the ogl regulatory region. Concernin the peiE and kdgT regulatory regions, we found only one potential KdgR-binding site in each region which was in both cases located between the promoter and the start codons. Binding of a repressor to such a site would probably prevent transcription by steric hindrance.

133

Moreover, the existence of one or two potential KdgR boxes (such as in the ogl regulatory region) could correspond to a gradation of affinity of the KdgR repressor for the various kdg operators, similar to the situation described for the arginine biosynthesis genes in E. coli (Piette et al., 1982). Given all these considerations, i.e., the presence of motifs related to the defined consensus sequence, ATGAAA(N),TTTCAT, in the promoter regions of three E. chqsanthemi 3937 genes regulated by kdgR, in addition to some appropriate locations, we believe that this sequence is a good candidate for the KdgRrepressor-binding site. (e) Conclusions

In this work, we have shown that: (1) all of the sequences required for synthesis of OGL are located within a 1.8-kb DNA segment, in which we have identified an ORF of 1164 bp coding for a protein with an M, of 44 124. This ORF is preceded by classical transcriptional and translational initiation signals and is followed by an E. coli type Rho-independent transcription termination site. The identity of this ORF has been confirmed by constructing an ogl: : IacZ fusion which is correctly expressed under appropriate conditions in E. coli. Analysis of the ogl gene product by the T7-promoter system suggests that OGL could be a pex-iplasmic enzyme. (2) Concerning PLe synthesis, we have demonstrated that the regulatory and coding regions of the pelE gene are located in a 1.7-kb DraI fragment. Within this fragment, we have detected an ORF of 1212 bp encoding a 404-aa polypeptide of M, 43 095. As in ogl, this ORF is preceded by classical transcriptional and translational initiation signals and is followed by an E. coli type Rho-independent transcription terminator. Analysis of the PLe aa sequence revealed a rather long potential signal peptide of 41 aa. Comparison of the E. chrysanthemi 3937 PLe aa sequence with other E. chrysanthemi PL sequences, confirmed the existence of two short regions of conserved aa as previously described by Tamaki et al. (1988). (3) To determine whether there were some conserved regions in OGL and PL enzymes which could be involved in their catalytic sites, we have compared the OGL and PLe aa sequences. This comparison did not reveal any significant homology, suggesting

that the catalytic sites of these enzymes are probably different, or alternatively that the catalytic sites may be more closely related in their three-dimensional conformations and simple match sequence comparison did not permit us to detect these sorts of features. (4) We also wished to determine if there were some conserved sequences in the ogl, pelE and kdgT regulatory regions likely to mediate KdgR regulation of these genes. Indeed, comparing regulatory sequences 5’ to the E. chrysanthemi 3937 ogl, pelE and kdgT genes (which are all regulated by the kdgR gene product), we have identified a highly-conserved sequence which could be a partial or entire KdgR repressor-binding site. From these observations, we have proposed the following sequence, ATGAAA(N),TTTCAT, as a consensus sequence for these potential KdgR-binding sites. This sequence possesses the classical symmetry of operator sites. However, to determine the KdgR binding site without ambiguity, comparison with regulatory regions of other genes governed by kdgR will be necessary. With this in mind, we are at the present attempting to clone two KdgR-regulated genes, kdgK and kduD, to determine the nt sequences of their 5’ regions. In addition, a mutational approach will be taken to define the essential and nonessential sequences for the KdgR-mediated regulation. (5) Furthermore, we have also noticed the presence of palindromic sequences and dyad symmetries in ogl and pelE regulatory regions. Such sequences did not occur in the kdgT regulatory region. These observations could be related to the fact that kdgR seems to be the only regulatory gene controlling kdgT expression, whereas, at least for pelE and probably also for ogl, expression is also affected by other regulatory elements. Both the ogl and pelE sequences contain regulatory regions, which are palindromic sequences exclusively composed of A and T. Such sequences could correspond to a binding site for another common regulatory protein, whereas other sequences, particularly the dyad symmetries, may mediate some specific regulation. The overall structural features of the nt sequences of the ogl and peZE regulatory regions suggest that the regulation of transcription involves the interaction of several regulatory proteins over a significant portion of the DNA. A detailed definition of the binding proteins, their binding sites and their interaction with the RNA polymerase will be re-

134

quired to establish the overall mechanisms involved in this regulatory process.

ACKNOWLEDGEMENTS

We wish to thank N. Hugouvieux-Cotte-Pattat for valuable discussions and advice. We are grateful to Barbara Angus and Jay Hinton for correcting our English and to S. Colombani for technical assistance. We thank all the members of the Erwiniu BAP ELWW for exchange of unpublished information, strains, phages and plasmids during our work. This work was supported by grants from the Commission of the European Communities (Biotechnology Action Program), from an Action d’Intervention sur Programme of the Centre National de la Recherche Scientilique, from Etablissement Publique Regional Rhone-Alpes and from the Minis&e de la Recherche et de 1’Enseignement Superieur (Aide sur Fonds de la Recherche).

REFERENCES Allen, C., Reverchon, S. and Robert-Baudouy, J.: Nucleotide sequence of the Erwinia chrysanthemi 2-keto-3-deoxygluconate permease. Gene 83 (1989) 233-241. Bachmann, B.S.: Linkage map ofEscherichia coli K-12. Edition 7. Microbial. Rev. 47 (1983) 180-230. Collmer, A. and Keen, N.T.: The role of pectic enzymes in plant pathogenesis. Annu. Rev. Phytopathol. 24 (1986) 383-409. Condemine, G. and Robert-Baudouy, J.: Tn5 insertion in kdgR, a regulatory gene of the polygalacturonate pathway in E. chrysanthemi. FEMS Microbial. Lett. 42 (1987a) 39-46. Condemine, G. and Robert-Baudouy J.: 2-Keto-3-deoxygluconate transport system in Erwinia chrysanthemi. J. Bacteriol. 169 (1987b) 1972-1978. De Crombrugghe, B., Busby, S. and But, H.: Cyclic AMP receptor protein: role in transcription activation. Science 224 (1984) 831-837. Drew, H.R., Weeks, J.R. and Travers, A.A.: Negative supercoiling induces spontaneous unwinding of a bacterial promoter. EMBO J. 4 (1985) 1025-1032. Gicquel-Sanzey, B. and Cossart, P.: Homologies between different procaryotic DNA-binding regulatory proteins and between their sites of action. EMBO J. 1 (1982) 591-595. Hawley, D.K. and McClure, W.R.: Compilation and analysis of Escherichia coli promoter DNA sequences. Nucleic Acids Res. 11 (1983) 2237-2255. Henikoff, S.: Unidirectional digestion with exonuclease HI creates targeted breakpoints for DNA sequencing. Gene 28 (1984) 351-359.

Hugouvieux-Cotte-Pattat, N. and Robert-Baudouy, J.: Hexuronate catabolism in Envinia chrysanthemi. J. Bacterial. 169 (1987) 1223-1231. Hugouvieux-Cotte-Pattat, N., Reverchon, S., Condemine, G. and Robert-Baudouy, J.: Regulatory mutants affecting the synthesis of pectate lyases in Erwinia chrysanthemi. J. Gen. Microbial. 132 (1986) 2099-2106. Hugouvieux-Cotte-Pattat, N., Reverchon, S. and RobertBaudouy, J.: Expanded linkage map of Erwinia chrysanthemi strain 3937. Mol. Microbial. 3 (1989) 573-581. Keen, N.T. and Tamaki, S.: Structure of two pectate lyase genes from Erwinia chrysanthemi EC16 and their high-level expression in Escherichia coli. J. Bacterial. 168 (1986) 595-606. Kotoujansky, A., Diolez, A., Boccara, M., Bertheau, Y., Andro, T. and Coleno, A.: Molecular cloning of Erwinia chrysanthemi pectinase and cellulase structural genes. EMBO J. 4 (1985) 781-785. Maniatis, T., Fritsch, E.F. and Sambrook, J.: Molecular Cloning. A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1982. Miller, J.H.: Experiments in Molecular Genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1972. Piette, J., Cunin, R., Boyen, A., Charlier, D., Crabeel, M., Van Vliet, F., Glansdorff, N., Squires, C. and Squires, CL.: The regulatory region of the divergent argECBH operon in Escherichia coliKl2. Nucleic Acids Res. 10 (1982) 8031-8048. Reverchon, S. and Robert-Baudouy, J.: Molecular cloning of an Erwinia chrysanthemi oligogalacturonate lyase gene involved in pectin degradation. Gene 55 (1987a) 125-133. Reverchon, S. and Robert-Baudouy, J.: Regulation of expression of pectate lyase genes pelA, pelD, pelE in Envinia chrysanthemi. J. Bacterial. 169 (1987b) 2417-2423. Reverchon, S., Hugouvieux-Cotte-Pattat, N. and RobertBaudouy, J.: Cloning of genes encoding pectolytic enzymes from a genomic library of the phytopathogenic bacteria Erwinia chtysanthemi. Gene 35 (1985) 121-130. Reverchon, S., Van Gijsegem, F., Rouve, M., Kotoujansky, A. and Robert-Baudouy, J.: Organization of a pectate lyase gene family in E. chrysanthemi. Gene 49 (1986) 215-224. Shine, J. and Dalgamo, L.: The 3’-terminal sequence of Escherichia coli 16s ribosomal RNA: complementarity to nonsense triplets and ribosome binding sites. Proc. Natl. Acad. Sci. USA 71 (1974) 1342-1346. Tabor, S. and Richardson, C.: A bacteriophage T7 RNA polymerase/promoter system for controlled exclusive expression of specific genes. Proc. Natl. Acad. Sci. USA 82 (1985) 1074-1078. Tamaki, S., Gold, S., Robeson, M., Manulis, S. and Keen, N.T.: Structure and organisation of the pel genes from Erwinia chrysanthemi EC16. J. Bacterial. 170 (1988) 3468-3478. Tiedeman, A. and Smith, J.: IacZY gene fusion cassettes with Kan’ resistance. Nucleic Acids Res. 16 (1988) 3587. von Heijne, G.: Signal sequences. The limits of variation. J. Mol. Biol. 184 (1985) 99-105. von Hippel, P.H., Bear, D.G., Morgan, W.D. and McSwiggen, J.A.: Protein-nucleic acid interactions in transcription: a molecular analysis. Annu. Rev. Biochem. 53 (1984) 389-446.