The transcriptional activator CorR is involved in ... - Springer Link

Download PDF
20 downloads 2367 Views 483KB Size Report
Comment
Organic acids were extracted from cell-free bacterial supernatants. (1.5 ml) of cultures ... molecular mass, whereas the MBP-CorP fusion was approximately 6 ...
Ó Springer-Verlag 1999

Mol Gen Genet (1999) 262: 250±260

ORIGINAL PAPER

L. Wang á C. L. Bender á M. S. Ullrich

The transcriptional activator CorR is involved in biosynthesis of the phytotoxin coronatine and binds to the cmaABT promoter region in a temperature-dependent manner Received: 25 February 1999 / Accepted: 14 June 1999

Abstract A modi®ed two-component regulatory system consisting of the histidine protein kinase CorS and two highly homologous response regulators, CorR and CorP, controls biosynthesis of the polyketide phytotoxin coronatine (COR) by Pseudomonas syringae pv. glycinea PG4180 in a temperature-dependent manner. COR synthesis is maximal at 18° C but does not occur at 28° C. Fusions of CorR and CorP to the maltose-binding protein (MBP) were overproduced in Escherichia coli and P. syringae PG4180, and tested for functionality by complementation of corR and corP mutants of PG4180, respectively. The cmaABT promoter region was de®ned by deletion mapping, and the DNA-binding capability of CorR and CorP was examined by gel retardation assays. When overproduced in P. syringae at 18° C and puri®ed, MBP-CorR was shown to bind speci®cally to a 218-bp DNA fragment corresponding to positions )841 to )623 bp upstream of the transcriptional start site of the cmaABT operon. In contrast, MBP-CorP and MBP itself, when overproduced in P. syringae and E. coli at 18° C and 28° C, respectively, did not bind to the 218-bp fragment or to any other DNA fragment analyzed. The CorP protein lacks typical DNA-binding motifs, suggesting that it might modulate the function of CorR. However, addition of puri®ed MBP-CorP did not alter the DNA-binding activity of MBP-CorR. On the other hand, this activity was completely abolished when MBPCorR was overproduced at 28° C or in a corS mutant, indicating that the binding of CorR depended on the Communicated by A. Kondorosi L. Wang á M. S. Ullrich (&) Max-Planck-Institut fuÈr terrestrische Mikrobiologie, AG OÈkophysiologie, Karl-von-Frisch-Strasse, D-35043 Marburg, Germany E-mail: [email protected], Tel.: +49-6421-178600; Fax: +49-6421-178609 C. L. Bender Department of Entomology and Plant Pathology, Oklahoma State University, 110 Noble Research Center, Stillwater, OK 74078-3032, USA

growth temperature at which it was produced and was controlled by CorS. In addition, overproduction of MBP-CorR in a corP mutant of PG4180 also yielded inactive protein, underlining the importance of CorP for CorR activation. We propose that CorR is activated by CorS at low temperature and that CorP is required for this activation before CorR can bind to DNA. Key words Coronatine synthesis á Response regulator á DNA binding á Two-component regulatory system á Temperature

Introduction Virulence factors of pathogenic bacteria are often subject to tightly coordinated regulation by various environmental factors (Mekalanos 1992). In many cases, this control is achieved via regulatory proteins that function in a multistep manner, thereby requiring a system for monitoring the environmental conditions that modulate the expression of virulence genes. Previous research on these regulatory cascades has indicated that temperature, osmolarity, pH, and host-borne factors can impinge on virulence gene expression (Venkatesan et al. 1988; Bernardini et al. 1990; Mekalanos 1992; Nakayama and Watanabe 1995, Hurme and Rhen 1998). In contrast to virulence factors in human and animal bacterial pathogens, relatively little is known about how temperature acts to regulate virulence in phytopathogenic bacteria. In this context, the temperature-dependent production of the phytotoxin coronatine (COR) by Pseudomonas syringae has been investigated in detail (Palmer and Bender 1993; Ullrich et al. 1995; Rangaswamy et al. 1997; Budde et al. 1998; Rohde et al. 1998). COR is a non-host-speci®c chlorosis-inducing phytotoxin produced by several pathovars of P. syringae that cause disease on soybean, tomato, Prunus spp., ryegrass and crucifers (Mitchell 1982; Mitchell et al. 1983; Wiebe and Campbell 1993). Structurally, COR consists of two distinct moieties, coronafacic acid (CFA)

251

and coronamic acid (CMA), which function as intermediates in the pathway to COR and are joined together by an amide bond (Mitchell et al. 1994; Parry et al. 1994). COR is thought to function as a structural mimic of intermediates in the octadecanoic signaling pathway of higher plants (Feys et al. 1994; Weiler et al. 1994; Krumm et al. 1995). Distinct regions required for the synthesis of CFA and CMA have been identi®ed within the 32.8-kb plasmid-borne COR biosynthetic gene cluster of P. syringae pv. glycinea PG4180 (Ullrich and Bender 1994; Ullrich et al. 1994; Bender et al. 1996; Rangaswamy et al. 1998; Fig. 1). COR biosynthesis in P. syringae pv. glycinea PG4180 is optimal at 18° C and undetectable at 28± 30° C. (Palmer and Bender 1993; Rangaswamy et al. 1997; Budde et al. 1998). Expression of transcriptional fusions of a promoterless û-glucuronidase (GUS) gene Fig. 1 Physical and functional maps of the COR biosynthetic gene cluster of P. syringae pv. glycinea PG4180. The gene clusters for coronafacic acid and coronamic acid synthesis are shown at the top. Expanded maps of the genes involved in CMA biosynthesis and regulation of COR production are shown below the physical map. The location of the cmaABT upstream region is shown by the horizontal arrow. The origins of the insert DNAs in the plasmids used in this study and their orientations with respect to the uidA and malE reporter genes are marked by horizontal bars (bottom). Glucuronidase activities of the cmaABT:uidA fusions at 18° C and the distance between the distal end of each and the transcriptional start site (TSS) of cmaABT (Ullrich and Bender 1994) are shown in the columns adjacent to each construct. Restriction enzyme sites: B, BamHI; E, EcoRI; Eh, EheI; H, HindIII; P, PstI; Sm, SmaI; S, SstI

(uidA) to promoter regions upstream of the CMA and CFA biosynthetic gene clusters (cmaABT and c¯/CFA) was found to be controlled by temperature (Ullrich and Bender 1994; Liyanage et al. 1995; Palmer et al. 1997). GUS expression was optimal when PG4180 cells containing the cmaABT::uidA or c¯::uidA transcriptional fusion were incubated at 18° C, and four- to ten-fold lower at 28° C. The temperature dependence of COR biosynthesis has been demonstrated in numerous CORproducing P. syringae strains (Rohde et al. 1998). For cmaABT:uidA, an upstream region roughly 400±1500 bp long was previously shown to be required for transcriptional activation (Ullrich and Bender 1994). Nucleotide sequencing and mutational analyses have indicated that three regulatory genes, corS, corP, and corR, are required for the temperature-dependent transcription of the cmaABT and c¯/CFA operons (Ullrich et al. 1995; Fig. 1). Insertion mutations in each of these regulatory genes abolish transcriptional activation of the COR biosynthetic promoters. The predicted amino acid sequence of CorS is related to those of histidine protein kinases that function as environmental sensors in twocomponent regulatory systems. Furthermore, CorR and CorP contain the highly conserved phosphoacceptor domains that are typical for response regulators in the two-component regulatory paradigm. Two-component regulatory systems are widespread among bacteria and mediate signal transduction from an external stimulus to gene expression, via reactions involving phosphorylation

252

and dephosphorylation of the participating proteins. CorR and CorP are nearly identical in their N-terminal regions; however, only CorR contains a helix-turn-helix (H-T-H) motif typical of response regulators in the twocomponent family (Ullrich et al. 1995). The absence of such a DNA-binding motif in CorP might indicate a function as a modulator or phosphorelay linking the signal transduction system with other regulatory circuits in the bacterial cell. It is reasonable to speculate that CorR binds to the cmaABT and c¯/CFA promoter regions, whereas CorP may modulate signal transduction. Recently, PenÄaloza-VaÂzquez and Bender (1998) showed that CorR binds to a 278-bp DNA fragment in the c¯/ CFA promoter region. However, the precise mechanism by which CorS, CorP, and CorR interact remains unclear. In this study, deletion analysis was used to de®ne the cmaABT promoter region. CorR and CorP were over-

produced as fusions to the maltose-binding protein (MBP) in Escherichia coli and P. syringae, and assayed for the ability to bind to the cmaABT promoter region. Binding of MBP-CorR to the cmaABT promoter region was demonstrated, and the e€ect of temperature and the involvement of CorS in CorR-mediated DNA binding were investigated. Furthermore, we studied the function of CorP and tested its in¯uence on the DNA-binding ability of CorR.

Materials and methods Bacterial strains, plasmids and growth conditions The bacterial strains and plasmids used in this study are listed in Table 1. P. syringae derivatives were routinely cultured on King's B (KB) medium (King et al. 1954) at 28° C. Hoitink-Sinden medium optimized for COR production (HSC) (Palmer and Bender

Table 1 Bacterial strains and plasmids used in this study Strain/plasmid Relevant characteristics E. coli DH5a JM109 P. syringae PG4180 PG4180.D4 PG4180.P2 PG4180.F7 Plasmids pBluescript II SK pMAL-c2 pLW045 pBBR1MCS pMUH34 pAP06.415 pLWB045 pMBPMU pRGMU1 pBSL14 pRGMU1P pRGMU7 pRGMU8 pRGMU9 pRGMUD1 pRGMUD2 pRGMUD3 pRGMUD4 pLW048 pLW044 pLW042 pLW038 pLW033 pLW026 pLW022 pLW015 pLW011

Reference/source Sambrook et al. (1989) Stratagene

Wild-type, COR+ Kmr, COR), corR corS Gmr, COR), corR Kmr, COR), corP

R. E. Mitchell Bender et al. (1993) PenÄaloza-VaÂzquez et al. (1996) Bender et al. (1993)

Apr, high copy number cloning vector

Stratagene

Apr, ColE1 origin, tac promoter, encodes malE Apr, contains corP on a 0.45-kb BamHI fragment; derived from pMUH34 by PCR cloning in pMAL-c2 Cmr, 4.7-kb broad-host-range cloning vector Tcr, contains corR, corP, and corS on a 3.4-kb HindIII-EcoRI fragment in pRK415 Tcr, contains malE::corR translational fusion in pRK415

New England Biolabs This study

Cmr, contains malE::corP translational fusion on 2.1-kb EcoRV-XbaI fragment derived from pLW045 in pBBR1MCS Tcr, chimeric plasmid constructed from pMAL-c2 and pRK415 r

r

Sm Sp , contains a 3.1-kb PstI fragment spanning the cmaABT upstream region in pRG960sd Kmr, 3.9-kb plasmid carrying multiple cloning sites and a Kmr cassette Smr Spr Kmr, contains polylinker and Kmr cassette of pBSL14 between SalI sites of pRGMU1 Smr Spr, contains a 1.5-kb PstI-AatI fragment in the SmaI site of pRG960sd Smr Spr, contains a 1.3-kb EcoRV-AatI fragment in pRG960sd Smr Spr, contains a 1.1-kb SmaI-AatI insert in pRG960sd Smr Spr, deletion derivative of pRGMU1P Smr Spr, deletion derivative of pRGMU1P Smr Spr, deletion derivative of pRGMU1P Smr Spr, deletion derivative of pRGMU1P Apr, contains 483-bp BamHI fragment derived from pRGMU1 by PCR cloning in pBluescript II SK Apr, contains 445-bp SmaI fragment derived from pRGMU1 in pBluescript II SK Apr, contains 420-bp SmaI-EheI fragment derived from pRGMU1 in pBluescript II SK Apr, contains 380-bp HindIII-SmaI fragment derived from pLW044 in pBluescript II SK Apr, contains 328-bp SmaI-BamHI fragment derived from pLW048 in pBluescript II SK Apr, contains 265-bp HindIII-BamHI fragment derived from pLW048 in pBluescript II SK Apr, contains 218-bp BamHI-HindIII fragment derived from pLW048 in pBluescript II SK Apr, contains 155-bp SmaI-BamHI fragment derived from pLW048 in pBluescript II SK Apr, contains 110-bp SmaI-HindIII fragment derived from pLW044

Kovach et al. (1994) Ullrich et al. (1995) PenÄaloza-VaÂzquez et al. (1996) This study PenÄaloza-VaÂzquez et al. (1996) Ullrich and Bender (1994) Alexeyev (1995) This study Ullrich and Bender (1994) Ullrich and Bender (1994) Ullrich and Bender (1994) This study This study This study This study This study This This This This This This This This

study study study study study study study study

253 1993) was used to grow P. syringae strains for protein puri®cation and for growing cells prior to GUS assays. E. coli JM109 and DH5a were used as hosts in cloning experiments and cultured in Luria-Bertani medium (LB) at 37° C. Protein concentrations were determined by the Bradford assay (Sambrook et al. 1989) and con®rmed by gel electrophoresis with known standards. Antibiotics were added to media in the following concentrations (lg/ml): ampicillin, 40; tetracycline, 12.5; kanamycin, 12.5; chloramphenicol, 12.5. Construction of translational fusions Plasmids pMBPMU and pAP06.415 were constructed previously (PenÄaloza-VaÂzquez et al. 1996; Table 1). pMal-c2, which contains the malE gene encoding the maltose-binding protein (MBP) under control of the tac promoter (Ptac), was used for overproduction of CorP. The corP gene was ampli®ed by PCR using plasmid pMUH34 as template DNA and the primers f± corP (5¢- GACGGATCCATGCCGAGCTCTTCG-3¢) and r-corP (5¢- GACGGATCCGCGCATTTCAACCAAAGCCTG-3¢) (the underlined sequences represent recognition sites for BamHI). A malE::corP fusion was constructed by cloning the 450-bp PCR fragment into BamH1-digested pMal-c2, resulting in pLW045 (Fig. 1). To facilitate overproduction of MBP-CorP in P. syringae, pLW045 was digested with EcoRV and XbaI to obtain a 2.1-kb fragment harboring the malE::corP fusion under Ptac control. This fragment was subsequently cloned in pBBR1MCS, a broad-host-range plasmid that replicates in P. syringae, resulting in pLWB045. Plasmids pAP06.415, pLWB045, and pMBPMU were introduced into PG4180, PG4180.P2 (corR), PG4180.F7 (corP), and PG4180.D4 (corR corS) by triparental matings (Gerhardt et al. 1994), and MBP-CorR, MBP-CorP, and MBP were overproduced in the respective P. syringae recipients. Protein overproduction and puri®cation For protein overproduction in E. coli, 500 ml of LB medium was inoculated with 5 ml of an overnight culture and incubated at 250 rpm at 18, 28, or 37° C. At an OD600 of 0.45±0.5, cultures were induced with 0.3 mM IPTG (isopropyl-b-D-thiogalactopyranoside) and then incubated for an additional 4±24 h at 18, 28, or 37° C. Cells were then pelleted, resuspended in protein extraction bu€er (50 mM TRIS-HCl, pH 7.4, 200 mM NaCl, 1 mM EDTA, 5 mM DTT, 1 mM magnesium acetate, 50 lg/ml lysozyme and 2 lg/ml DNase), and placed on ice for 30 min. Bacterial cells were lysed by four passages through a French pressure cell. MBP and MBP fusions were puri®ed from cell lysates by anity chromatography on amylose columns as recommended by the matrix manufacturer (New England Biolabs, Schwalbach, Germany). For overproduction of recombinant proteins in P. syringae, 100 ml of HSC medium was inoculated with 3 ml of an overnight culture and incubated at 18 or 28° C and 280 rpm. At an OD600 of 0.45±0.5, cultures were induced with 5 mM IPTG and incubated for additional 24±48 h. Proteins were isolated from P. syringae using the methods described for E. coli. DNA manipulations Agarose gel electrophoresis, restriction digestion, puri®cation of DNA fragments from agarose gels, and PCR were performed using standard techniques (Sambrook et al. 1989). Nested deletions were constructed using the Erase-a-Base system (Promega, Mannheim, Germany). To obtain suitable endonuclease recognition sites for subsequent exonuclease III digestion, a polycloning site derived from plasmid pBSL14 was inserted between the two SalI sites of pRGMU1 to yield pRGMU1P. Recognition sites used for deletion analysis were XbaI (susceptible site) and ApaI (resistant site). The primers 450F (5¢-TCGGATCCTCCTCGATCCTGCTG-3¢) and

450R (5¢-TAGGATCCAAGAACTGCTCCAT-3¢) (BamHI sites are underlined), and plasmid pRGMU1, were used to amplify by PCR a 483-bp fragment which was cloned into pBluescript SK+ to yield pLW048 (see Fig. 4). A 420-bp fragment was obtained by SmaI + EheI digestion of pRGMU1 and cloned into pBluescript SK+ to generate pLW042, and a 445-bp SmaI fragment from pRGMU1 was inserted into pBluescript SK+ to yield pLW044 (see Fig. 4). Fragments of 380, 328, 265, 218, 155, and 110 bp were derived from pLW048 and pLW044 and cloned into pBluescript SK+ to yield pLW038, pLW033, pLW026, pLW022, pLW015, and pLW011, respectively (Table 1; Fig. 4). The nucleotide sequence of all cloned fragments was determined and shown to be identical to the previously published DNA sequence of the cmaABT upstream region (GenBank Accession Nos. U14657 and U33327). Large-scale preparations of plasmid DNA were isolated from E. coli by alkaline lysis and puri®ed with the Qiagen Tip 100 kit (Qiagen, Hilden, Germany). Plasmid DNA was isolated from P. syringae as described by Kado and Liu (1981). Nucleotide sequencing was performed by the dideoxynucleotide method (Sambrook et al. 1989) using the Thermo Sequenase Fluorescent Labeled Primer Cycle Sequencing kit (Amersham-Buchler, Braunschweig, Germany). Sequence data were aligned and processed with the Lasergene software package, Version 4.1 (DNASTAR, Madison, Wis.). Detection and quantitation of COR synthesis Organic acids were extracted from cell-free bacterial supernatants (1.5 ml) of cultures grown to an OD600 of approximately 3.0 and analyzed for the presence of COR by the high-pressure liquid chromatography (HPLC) method described by Palmer and Bender (1993). Glucuronidase assays Plasmid pRGMU1 and deletion derivatives containing cmaABT::uidA transcriptional fusions (Fig. 1) were used to measure GUS activity in P. syringae. Bacteria were incubated in HSC medium as described above until cultures reached an OD600 of 3.0 at 18 and 28° C. Cells were subsequently harvested by centrifugation and processed as described previously (Budde et al. 1998). All assays included six replicates, and GUS activity was quanti®ed by ¯uorometric analysis (Xiao et al. 1992) using a Fluorolite-1000 microplate reader (Dynatech, Denkendorf, Germany). Electrophoretic mobility shift assays The 3¢ end of each DNA fragment (400 ng) was labeled with 1 nmol of digoxigenin (DIG)-11-ddUTP (Boehringer-Mannheim, Mannheim, Germany) using 1 U of terminal deoxynucleotidyl transferase at 37° C for 15 min in 20 ll of assay bu€er (200 mM potassium carcodylate, 25 mM TRIS-HCl pH 6.6, 0.25 mg bovine serum albumin, and 5 mM CoCl2). The labeled DNA fragments were then precipitated with 95% ethanol, washed with 70% ethanol, suspended in 20 ll of TEN bu€er (10 mM TRIS-HCl pH 8.0, 1 mM EDTA, 0.1 M NaCl), and stored at )20° C. For electrophoretic mobility shift assays, DIG-labeled DNA fragments (10 ng) were incubated with 1±500 ng of each fusion protein and 1 lg of poly(dI±dC) at 0, 18, or 28° C for 30 min in 5´ gel retardation bu€er [50 mM TRIS-HCl (pH 7.5), 50 mM KCl, 5 mM EDTA (pH 8.0), 5 mM DTT, and 50% glycerol]. Samples were then separated on 5% polyacrylamide gels in 0.25´ TBE bu€er at 160 V for 40±70 min, electroblotted onto positively-charged nylon membranes (Boehringer-Mannheim), and ®xed to the membranes by UV cross-linking. DNA fragments were visualized using anti-DIG/alkaline phosphatase conjugate and chemoluminescent detection, as directed by the manufacturer (Boehringer-Mannheim).

254

Results Overproduction and puri®cation of MBP-CorR and MBP-CorP Plasmids pAP06.415 and pLWB045 were used to overproduce MBP-CorR and MBP-CorP, respectively, in E. coli and P. syringae. After induction with IPTG, MBP-CorR and MBP-CorP were produced in both bacterial species as soluble proteins that had molecular masses of approximately 64 and 52 kDa, respectively (Fig. 2, lanes 3 and 6). The molecular mass of MBP-CorR corresponded to the previously predicted molecular mass, whereas the MBP-CorP fusion was approximately 6 kDa smaller than predicted (Ullrich et al. 1995). MBP-CorR and MBP-CorP were puri®ed by anity chromatography on amylose columns, and the puri®ed fusion proteins were shown to be free of major contaminants (Fig. 2, lanes 4 and 7). The malE::corR fusion restores transcriptional activity in a corR mutant Plasmid pAP06.415 was previously shown to restore COR biosynthesis to the corR mutant PG4180.P2, which indicated that the MBP-CorR fusion was functional in vivo (PenÄaloza-VaÂzquez et al. 1996). In the present study, we investigated whether the MBP-CorR fusion could complement PG4180.P2 for transcriptional

Fig. 2 Overproduction and puri®cation of MBP fusions of CorR and CorP in P. syringae pv. glycinea PG4180 grown at 18° C. Proteins were fractionated by SDS PAGE (10% polyacrylamide). Lanes: 1, molecular weight markers; 2, PG4180 (pAP06.415) uninduced; 3, PG4180 (pAP06.415) induced with IPTG; 4, MBP-CorR puri®ed by anitiy chromatography; 5, PG4180 (pLWB045) uninduced; 6, PG4180 (pLWB045) induced with IPTG; 7, MBP-CorP puri®ed by anity chromatography

activation of the cmaABT::uidA fusion in pRGMU1. No GUS activity was detected in PG4180.P2 (pRGMU1), regardless of the incubation temperature. However, the wild-type PG4180 containing pRGMU1 synthesized 550 and 80±100 U of GUS activity per mg of protein at 18 and 28° C, respectively. When pAP06.415 was introduced into PG4180.P2 (pRGMU1) and protein expression was induced by addition of IPTG, GUS activities of 389 ‹ 24 and 127 ‹ 19 U were measured in bacteria grown at 18 and 28° C, respectively. These results demonstrated that recombinant MBP-CorR was able to restore transcriptional activation of the cmaABT operon in a corR mutant of PG4180. The malE::corP fusion restores COR biosynthesis and transcriptional activation of cmaABT in a corP mutant Plasmid pLWB045 carrying a malE::corP fusion was introduced into the corP mutant PG4180.F7 and PG4180.F7 (pRGMU1). Mutant PG4180.F7 harbors a transposon insertion in the corP gene, lacks COR synthesis, and showed no GUS expression when carrying plasmid pRGMU1 (Ullrich et al. 1995; Fig. 3). Strains

Fig. 3A, B Complementation of the corP mutant PG4180.F7 by overproduction of MBP-CorP. A Coronatine production by PG4180, PG4180.F7 (corP), and PG4180.F7 (pLWB045) grown at 18° C. B GUS activities of transconjugants harboring pRGMU1 and pLWB045 grown at 18 and 28° C. All values represent the averages of three experiments with two replicates

255

PG4180.F7 (pLWB045) and PG4180.F7 (pRGMU1, pLWB045) were incubated in HSC medium at 18 and 28° C, induced with IPTG, and analyzed for COR synthesis and glucuronidase expression, respectively. Results were compared with those for PG4180 and PG4180.F7, with or without plasmid pRGMU1, and are summarized in Fig. 3; they demonstrated that plasmid pLWB045 restored COR synthesis (Fig. 3A) and temperature-dependent cmaABT transcription (Fig. 3B) in the corP mutant PG4180.F7. These results suggested that recombinant CorP was functional and was required for activation of cmaABT expression. Deletion analysis of the cmaABT promoter region Transcriptional fusions were previously used to demonstrate that the DNA region located upstream of cmaABT contains a thermoregulatable promoter (Ullrich and Bender 1994; Budde et al. 1998). To de®ne more precisely the DNA region involved in transcriptional activation of the cmaABT transcript, a series of deletions were constructed from the 5¢ end of pRGMU1P, resulting in pRGMUD1, pRGMUD2, pRGMUD3 and pRGMUD4 (Fig. 1). GUS activity was evaluated in PG4180 transconjugants grown at 18° C and containing these deletion derivatives or plasmids pRGMU1, pRGMU7, pRGMU8, or pRGMU9 (Ullrich and Bender 1994) (Fig. 1). The results indicated that a region located 460± 1500 bp upstream of the cmaABT transcriptional start site was required for activation of the uidA reporter gene (Fig. 1). Since deletion derivatives with inserts larger than that in pRGMUD1 did not mediate a further increase in GUS activity, subsequent experiments focused on the DNA region located between nucleotides )460 and )1500. Nine DNA fragments (Fig. 4) spanning this region were cloned in E. coli DH5a, puri®ed, and used in subsequent DNA binding studies. Electrophoretic mobility shift assays The results of the GUS assays reported above indicated that the DNA region cloned in plasmids pLW048, pLW044 and pLW042 (Fig. 4) might interact with a transcriptional activator. This hypothesis was examined by electrophoretic mobility shift assays using insert DNA derived from these plasmids together with puri®ed MBP, MBP-CorR, or MBP-CorP. When these proteins were puri®ed from E. coli, no DNA binding to target DNA fragments was observed, regardless of the assay conditions (data not shown), indicating that the fusion proteins might be improperly folded or nonfunctional when produced in E. coli. Therefore, in all subsequent experiments fusion proteins were derived from PG4180. Initially, PG4180 cultures producing translational fusions were grown at 18° C since this temperature is optimal for cmaABT transcription. When the fragments cloned in pLW048 and pLW044 were used as target DNA, a protein-DNA complex was

Fig. 4 Restriction map of the cmaABT upstream region in P. syringae pv. glycinea PG4180. The locations of subclones used for electrophoretic mobility shift studies are indicated and the distance from the proximal end of each to the transcriptional start site (TSS) of cmaABT are shown in the column on the right. Results of DNA binding assays with MBP-CorR are given in the left column adjacent to each construct. Abbreviations: (+), DNA binding; ()), no DNA binding; primer 450R and 450F, primer-binding sites. The region required for binding of MBP-CorR binding is indicated by the stippled column

observed with 25 ng of MBP-CorR puri®ed from PG4180 (pAP06.415) (Fig. 5A, lanes 1+ and 2+). This complex was not observed with 200±1000 ng of MBP or MBP-CorP puri®ed from PG4180 (pMBPMU) and PG4180 (pLWB045), respectively (Fig. 5A, lanes 4) and 4+ and data not shown). No band retardation was observed when the DNA cloned in pLW042 was used as a target fragment and incubated with 25±500 ng of MBP-CorR (Fig. 5A, lanes 3) and 3+), indicating that the CorR-binding region had been subcloned on plasmids pLW048 and pLW044. To de®ne the DNA region responsible for binding more precisely, we used the DNA fragments subcloned in pLW038, pLW033 (Fig. 5B), pLW026, pLW022 (Fig. 5C), pLW015, and pLW011 for gel shift assays. The results are summarized in Fig. 4 and show that the 218-bp insert contained in pLW022 was the smallest DNA fragment that allowed MBP-CorR binding. All DNA fragments were also examined for the capacity to bind to MBP and MBP-CorP using 200±1000 ng of puri®ed protein. However, no band retardation was observed with either MBP (Fig. 5C, right side) or MBP-CorP (Fig. 5B, right side), indicating that MBP-CorP does not bind to the cmaABT promoter region. These results furthermore suggested that the MBP portion of MBP-CorR was not involved in the DNA-binding activity of the fusion protein. Speci®city of the MBP-CorR binding reaction To demonstrate the speci®city of the DNA-protein complex formed by the 328-bp fragment of pLW033 and

256

Fig. 6 Competition assay using MBP-CorR and the 328-bp target fragment in pLW033. DNA-protein complex formation is indicated by band shifts. All lanes contained 25 ng of MBP-CorR and 10 ng of labeled 328-bp fragment from pLW033. In addition the lanes contained the following amounts of unlabeled 328-bp fragment as competitor DNA. Lanes: 1, 1 ng; 2, 10 ng; 3, 20 ng; 4, 50 ng; 5, 100 ng; 6, 200 ng; 7, 400 ng

containing MBP-CorR and the labeled fragment from pLW033. Addition of the unlabeled 420-bp insert from pLW042, which does not bind MBP-CorR, did not reduce the degree of complex formation (data not shown), indicating that the interaction of MBP-CorR with the labeled target DNA was highly speci®c. Interaction of CorR and CorP

Fig. 5A±C Electrophoretic mobility shift analysis of DNA fragments incubated with the puri®ed proteins MBP-CorR, MBP-CorP, and MBP. DNA-protein complex formation is indicated by band shifts following electrophoresis in 5% polyacrylamide gels. A Binding of MBP-CorR to 445-bp and 483-bp DNA fragments, but not to a 420bp fragment [(+) protein added; ()), no protein added]. Lanes: 1, MBP-CorR (25 ng) with 445-bp fragment (pLW044); 2, MBP-CorR (25 ng) with 483-bp fragment (pLW048); 3, MBP-CorR (250 ng) with 420-bp fragment (pLW042); 4, MBP (500 ng) with 445-bp fragment (pLW044). B Comparison of the ability of 25, 100, and 250 ng of MBP-CorR, and 25, 100, and 250 ng of MBP-CorP to bind the 328bp fragment cloned in pLW033 [()), no protein added]. C Comparison of the ability of 25 and 50 ng of MBP-CorR, and 100 and 500 ng of MBP to bind the 218-bp fragment of pLW022 [()), no protein added]

MBP-CorR, increasing amounts (1±400 ng) of the unlabeled target fragment were added to the reaction mixture (Fig. 6). The results clearly showed a progressive reduction in binding of labeled DNA as increasing amounts of unlabeled fragment were added. In another experiment, increasing amounts of unlabeled insert DNA from pLW042 were added to reaction mixtures

Possible interactions between CorR and CorP were investigated by adding various amounts of puri®ed MBPCorP (25±1000 ng) to DNA binding reaction mixtures containing MBP-CorR and the DIG-labeled 328-bp fragment of pLW033. In control experiments, increasing amounts of MBP were evaluated for their e€ect on the binding of MBP-CorR to the labeled 328-bp fragment (data not shown). Interestingly, MBP-CorP had no effect on the binding of MBP-CorR to the target fragment. Large amounts of MBP-CorP (1000 ng) interfered with the binding of MBP-CorR but this interference was nonspeci®c since the same amount of MBP also interfered with complex formation (data not shown). These results demonstrate that the recombinant CorP protein does not interact with MBP-CorR during binding of the latter to the cmaABT promoter region. Since MBP-CorP was shown to be required for restoration of COR synthesis and cmaABT transcription in the corP mutant PG4180.F7 (see above), we next tested whether lack of native CorP had any in¯uence on the DNA-binding activity of MBP-CorR. To do so, plasmid pAP06.415 carrying the malE::corR translational fusion was introduced into mutant PG4180.F7 (corP). Subsequently, MBP-CorR was overexpressed at 18° C, puri®ed from the transconjugant PG4180.F7 (pAP06.145), and analyzed in electrophoretic mobility shift assays

257

with the DNA target derived from pLW033 (Fig. 7A). The results indicated that MBP-CorR binding to the DNA fragment was completely abolished when the recombinant protein was expressed in a corP mutant (Fig. 7A, lanes 6±10), suggesting that native CorP is indispensable for the DNA-binding activity of MBPCorR. E€ect of temperature and CorS on DNA binding The temperature-dependent transcriptional activation of the cmaABT operon (Ullrich et al. 1995; Budde et al. 1998) prompted us to investigate whether MBP-

CorR could bind to the target DNA region in a temperature-dependent manner. Binding assays with puri®ed MBP-CorR and the 328-bp fragment derived from pLW033 were repeated at 0, 18, and 28° C; however, no signi®cant di€erence in binding eciency was observed when reactions were conducted at these three temperatures (data not shown). Consequently, temperature may not have any obvious e€ects on the DNA binding of MBP-CorR during the short incubation period prior to electrophoresis of the DNA-protein complex. Therefore, we investigated the e€ect of temperature on MBP-CorR binding by incubating PG4180 (pAP06.415) at 18 and 28° C prior to protein puri®cation. Equal amounts of MBP-CorR could be obtained from cultures grown at both temperatures, suggesting that incubation at 28° C had no negative e€ect on MBP-CorR production (data not shown). Bandshifts were observed when MBP-CorR was puri®ed from PG4180 (pAP06.415) grown at 18° C as described in other experiments (Fig. 7B, lanes 2±4). However, the DNA binding ability of MBP-CorR was completely abolished when the fusion protein was overproduced at 28° C (Fig. 7B, lanes 5±7). We previously demonstrated that corS, which encodes the putative cognate histidine protein kinase for CorR, was not expressed at the transcriptional level when PG4180 cultures were grown at 28° C (Ullrich et al. 1995). Since corS is required for optimal expression of the cmaABT promoter, we investigated whether MBP-CorR was produced in a functional form in the corR corS mutant PG4180.D4. MBP-CorR was puri®ed from PG4180.D4 (pAP06.415) grown at 18° C and shown to be produced at levels comparable to those expressed in PG4180 (pAP06.415) (data not shown). When MBP-CorR was puri®ed from PG4180.D4 (pAP06.415) and used in gel shift experiments, the fusion protein did not bind to the 328-bp fragment of pLW033 (Fig. 7B, lanes 8±10). Similar experiments were repeated with DNA fragments derived from pLW044 and pLW022, and gave the same result (data not shown). These observations indicated that a functional copy of corS was required for the DNA-binding activity of MBP-CorR.

Discussion Fig. 7 A DNA-binding ability of MBP-CorR puri®ed from PG4180.F7 (corP) and from PG4180 wild type grown at 18° C. Lanes: 1, no protein; 2±5, MBP-CorR (10, 25, 50, and 100 ng) puri®ed from PG4180 (pAP06.415) grown at 18° C; 6±10, MBP-CorR (10, 25, 50, 100, and 500 ng) puri®ed from the corP mutant PG4180.F7 (pAP06.415) grown at 18° C. B DNA-binding ability of MBP-CorR puri®ed from PG4180 cultures grown at 18 and 28° C, or overproduced in the corS mutant PG4180.D4 at 18° C. DNA-protein complex formation is indicated by band shifts. Lanes: 1, no protein added; 2±4, MBP-CorR (10, 25, and 50 ng) puri®ed from PG4180 (pAP06.415) grown at 18° C; 5±7, MBP-CorR (25, 100, and 250 ng) puri®ed from PG4180 (pAP06.415) grown at 28° C; 8±10, MBP-CorR (25, 100, and 250 ng)puri®ed from PG4180.D4 (pAP06.415) grown at 18° C

The thermoregulated synthesis of COR, an important virulence factor of P. syringae pv. glycinea PG4180, is an excellent model system in which to study the e€ects of temperature on plant-microbe interactions. The present study was aimed at de®ning the function of CorR as a transcriptional activator of the cmaABT operon, expression of which is required for COR biosynthesis. Furthermore, the results of this study allow a better understanding of the interactions between CorR, its cognate sensor kinase, CorS, and CorP, the second response regulator protein involved in regulation of COR biosynthesis.

258

It was previously shown that MBP-CorR binds to DNA located upstream of the c¯/CFA operon of P. syringae PG4180 (PenÄaloza-VaÂzquez and Bender 1998). Here, we have demonstrated that MBP-CorR also binds to the upstream sequence of cmaABT, the second biosynthetic operon of the COR gene cluster. This ®nding is compatible with the sequence similarities between the two DNA regions reported by PenÄalozaVaÂzquez and Bender (1998). The present study has provided direct insight into the possible interactions between CorR, CorS, and CorP, based on genetic and biochemical evidence. It was shown that CorS is indispensable for the DNA-binding activity of CorR. The interaction of the three proteins yielded active MBPCorR only at 18° C, a growth temperature at which corS expression is induced (Ullrich et al. 1995), but not when MBP-CorR was derived from a corS mutant grown at 18° C. Using mutant complementation, recombinant CorP protein was shown to function in the regulation of COR biosynthesis and in transcriptional activation of the cmaABT operon. Interestingly, all attempts to demonstrate a direct interaction between CorR and CorP during DNA binding of CorR failed. Finally, we provide direct evidence that CorP is essential for the DNA-binding activity of CorR. In summary, these results strengthen the postulate that CorP might have an indispensable modulatory or phospho-relay function when the CorSRP protein triad is present in its native stoichiometry. Other researchers have successfully used MBP translational fusions to investigate the activity of response regulators (Lee et al. 1993; Boucher et al. 1994; Grob and Guiney 1996). In the current study, the MBPCorR fusion protein exhibited DNA-binding activity when overproduced and puri®ed from its native organism, P. syringae, but not when produced in E. coli. Therefore, MBP-CorR might be improperly folded or lack essential factors required for DNA-binding activity (such as CorS and CorP) when overproduced in E. coli. The in vivo activity of CorR as a transcriptional activator of the cmaABT operon demonstrated in this study agrees with earlier ®ndings (PenÄaloza-VaÂzquez and Bender 1998). In the earlier report, recombinant CorR was shown to restore COR biosynthesis to the corR mutant PG4180.P2; we have now demonstrated that MBP-CorR can also restore transcriptional activity to this mutant. Deletion mapping and electrophoretic mobility shift assays indicated that the region required for MBP-CorR binding is located between nucleotides )841 and )623 with respect to the cmaABT transcriptional start site. This is in agreement with previous results (Ullrich and Bender 1994) indicating that a distal upstream region is required for transcriptional activation of this operon. Far upstream binding sites have also been identi®ed for other regulatory proteins (Reitzer et al. 1989; Kato and Chakrabarty 1991; Mohr et al. 1991; Huang et al. 1994). Recently, speci®c binding of MBP-CorR to the c¯/CFA promoter was demonstrated using as a probe a region

located 650 bp upstream of the c¯/CFA transcriptional start site (PenÄaloza-VaÂzquez and Bender 1998). The MBP-CorR binding site in the c¯/CFA promoter was previously de®ned by gel retardation and DNaseI footprinting analyses and delimited to a 54-bp sequence between )650 and )704 with respect to the transcriptional start site (PenÄaloza±VaÂzquez and Bender 1998). In the present study, the CorR-binding region in the cmaABT upstream sequence showed 40% identity to the 54-bp sequence required for MBP-CorR binding to the c¯/CFA promoter region (PenÄaloza-VaÂzquez and Bender 1998). The homologous region in the cmaABT transcript was located in a 60-bp sequence located between nucleotides )757 and )698 with respect to the transcriptional start site. Both the c¯/CFA and cmaABT transcripts are controlled by CorR and optimally expressed at 18° C (Ullrich and Bender 1994; Liyanage et al. 1995; Ullrich et al. 1995; PenÄaloza-VaÂzquez and Bender 1998). Although recombinant MBP-CorP was capable of restoring COR synthesis and transcriptional activation of cmaABT to a corP mutant, it did not bind to the cmaABT promoter region. This result is consistent with the fact that CorP lacks a typical C-terminal H-T-H DNA-binding domain. However, the N-terminal receiver domains of CorR and CorP are very similar to each other, suggesting a shared speci®city for the same phosphodonor (Ullrich et al. 1995). Possible functions for CorP have been previously proposed and include participation in a phosphorelay, formation of a CorRCorP heterodimer, or function as a modulator of signal transduction (Ullrich et al., 1995). In the present study, we show that the addition of MBP-CorP to electrophoretic mobility shift assays did not a€ect the DNAbinding function of MBP-CorR, which lessens the probability that a CorP-CorR heterodimer is involved in binding. However, we cannot completely discount possible interference by MBP in the formation of a CorPCorR heterodimer. Furthermore, the expression of MBP-CorR in a corP mutant abolished the DNAbinding capability of MBP-CorR, suggesting that CorP is essential for the activation of recombinant MBPCorR. These results are consistent with the hypothesis that CorP might have a modulatory function, might transduce signals from regulatory circuits other than the COR regulatory system, or might initially be activated by CorS to subsequently activate CorR. Obviously, the function of CorP must play an important role in regulation of COR biosynthesis since the corP mutant PG4180.F7 completely lacks COR synthesis and is not able to transcribe cmaABT (Ullrich et al. 1995 and this report). Experiments to further de®ne the regulatory functions of CorP and CorS are currently underway in our laboratories. The incubation temperature used for the gel retardation assays did not a€ect the ability of MBP-CorR to bind target DNA fragments. Whether DNA supercoiling (TseDinh et al. 1997) in¯uences this DNA-protein interaction remains to be tested in vivo or with closed circular DNA

259

targets. However, when MBP-CorR was overproduced in P. syringae grown at 28° C, the DNA-binding ability of the fusion protein was abolished. Furthermore, no binding was obtained when MBP-CorR was puri®ed from a corS mutant of PG4180, possibly because CorR needs to be phosphorylated by CorS (or CorP) prior to binding. Other researchers have shown that response regulators bind their target DNA sequences more eciently when phosphorylated (Walker and DeMoss 1994; Ste€en et al. 1996; Arthur et al. 1997). Interestingly, addition of lowmolecular-weight phosphodonors, such as acetyl phosphate, did not result in conversion of puri®ed but inactive MBP-CorR to its functional form (L. Wang and M. Ullrich, unpublished data), suggesting that phosphorylation of CorR might speci®cally require CorS or CorP. Cullen et al. (1996) have reported a similar phenomenon for the NtrC/NtrB regulatory system which controls nitrogen metabolism in Rhodobacter capsulatus. It is also important to note that transcription of the corS promoter is repressed at 28° C, but strongly induced at 18° C (Ullrich et al. 1995). This may explain the inactivity of MBP-CorR when puri®ed from P. syringae cultures incubated at 28° C. In summary, we provide the ®rst direct evidence that CorS is required for the DNA-binding activity of CorR. The CorP protein might in turn meditate or modulate this interaction to guarantee a balanced signal transduction in vivo. Our data also indicate that MBP-CorR remained functional during the puri®cation process, a feature that will be useful for the phosphorylation assays and protein-protein interaction studies to be conducted in our laboratories. Acknowledgements This work was supported by grants from the Max-Planck-Gesellschaft and Deutsche Forschungsgemeinschaft (M. Ullrich) and by the National Science Foundation MCB9603618 (C. Bender). We thank Ulrich GraÈdler and Nicole RoÈssel for technical help during initial steps of this project and Erhard Bremer and Uwe VoÈlker for stimulating discussions.

References Alexeyev M (1995) Three kanamycin resistance gene cassettes with di€erent polylinkers. BioTechniques 18:53±55 Arthur M, Depardieu F, Gerbaud G, Galmand M, Leclercq R, Courvalin P (1997) The VanS sensor negatively controls VanRmediated transcriptional activation of glycopeptide resistance genes of Tn1546 and related elements in the absence of induction. J Bacteriol 179:97±106 Bender CL, Liyanage H, Palmer D, Ullrich M, Young S, Mitchell R (1993) Characterization of the genes controlling biosynthesis of the polyketide phytotoxin coronatine including conjugation between coronafacic and coronamic acid. Gene 133:31±38 Bender CL, Palmer P, PenÄaloza-VaÂzquez A, Rangaswamy V, Ullrich M (1996) Biosynthesis of coronatine, a thermoregulated phytotoxin produced by the phytopathogen Pseudomonas syringae. Arch Microbiol 166:71±75 Bernardini ML, Fontaine A, Sansonetti PJ (1990) The two-component regulatory system OmpR-EnvZ controls the virulence of Shigella ¯exneri. J Bacteriol 172:6274±6281 Boucher PE, Menozzi FD, Locht C (1994) The molecular architecture of bacterial response regulators. Insights into the activation mechanism of the BvgA transactivator of Bordetella pertussis. J Mol Biol 241:363±377

Budde IP, Rohde BH, Bender CL, Ullrich MS (1998) Growth phase and temperature in¯uence promoter activity, transcript abundance, and protein stability during biosynthesis of the Pseudomonas syringae phytotoxin coronatine. J Bacteriol 180:1360±1367 Cullen PJ, Bowman WC, Kranz RG (1996) In vitro reconstitution and characterization of the Rhodobacter capsulatus NtrB and NtrC two-component system. J Biol Chem 271:6530±6536 Feys BJ, Benedetti CE, Penfold CN, Turner JG (1994) Arabidopsis mutants selected for resistance to the phytotoxin coronatine are male sterile, insensitive to methyl jasmonate, and resistant to a bacterial pathogen. Plant Cell 6:751±759 Gerhardt P, Murray RG, Wood WA, Krieg NR (1994) Methods of general and molecular bacteriology. American Society for Microbiology, Washington, DC Grob P, Guiney DG (1996) In vitro binding of the Salmonella dublin virulence plasmid regulatory protein SpvR to the promoter regions of spvA and spvR. J Bacteriol 178:1813± 1820 Huang KJ, Schieberl JL, Igo MM (1994) A distant upstream site involved in the negative regulation of the Escherichia coli ompF gene. J Bacteriol 176:1309±1315 Hurme R, Rhen M (1998) Temperature sensing in bacterial gene regulation ± what it all boils down to. Mol Microbiol 30:1±6 Kado CI, Liu ST (1981) Rapid procedure for detection and isolation of large and small plasmids. J Bacteriol 145:1365±1373 Kato J, Chakrabarty AM (1991) Puri®cation of the regulatory protein AlgR1 and its binding in the far upstream region of the algD promoter in Pseudomonas aeruginosa. Proc Natl Acad Sci USA 88:1760±1764 King EO, Ward MK, Raney DE (1954) Two simple media for the demonstration of pyocyanin and ¯uorescein. J Lab Clin Med 44:301±307 Kovach ME, Philipps RW, Elzer PH, Roop III RM, Peterson KM (1994) pBBR1MCS: a broad-host-range cloning vector. BioTechniques 16:800±802 Krumm T, Bandemer K, Boland W (1995) Induction of volatile biosynthesis in the Lima bean (Phaseolus lunatus) leucine- and isoleucine conjugates of 1-oxo-and 1-hydroxyindan-4-carboxylic acid: evidence for amino acid conjugates of jasmonic acid as intermediates in the octadecanoid signaling pathway. FEBS Lett 377:523±529 Lee HS, Berger DK, Kustu S (1993) Activity of puri®ed NIFA, a transcriptional activator of nitrogen ®xation genes. Proc Natl Acad Sci USA 90:2266±2270 Liyanage H, Palmer DA, Ullrich M, Bender CL (1995) Characterization and transcriptional analysis of the gene cluster for coronafacic acid, the polyketide component of the phytotoxin coronatine. Appl Environ Microbiol 61:3843±3848 Mekalanos, JJ (1992) Environmental signals controlling expression of virulence determinants in bacteria. J Bacteriol 174:1±7 Mitchell RE (1982) Coronatine production by some phytopathogenic pseudomonads. Physiol Plant Pathol 220:83±89 Mitchell RE, Hale CN, Shanks JC (1983) Production of di€erent pathogenic symptoms and di€erent toxins by strains of Pseudomonas syringae pv. tomato not distinguishable by gelimmunodi€usion assay. Physiol Plant Pathol 23:315±322 Mitchell RE, Young SA, Bender CL (1994) Coronamic acid, an intermediate in coronatine biosynthesis by Pseudomonas syringae. Phytochemistry 35:343±348 Mohr CD, Hibler NS, Deretic V (1991) AlgR, a response regulator controlling mucoidy in Pseudomonas aeruginosa, binds to the FUS sites of the algD promoter located unusually far upstream from the mRNA start site. J Bacteriol 173:5136±5143 Nakayama S, Watanabe H (1995) Involvement of cpxA, a sensor of a two-component regulatory system, in the pH-dependent regulation of expression of Shigella sonnei virF gene. J Bacteriol. 177:5062±5069 Palmer DA, Bender CL (1993) E€ects of environmental and nutritional factors on production of the polyketide phytotoxin coronatine by Pseudomonas syringae pv. glycinea. Appl Environ Microbiol 59:1619±1623

260 Palmer DA, Bender CL, Sharma, SB (1997) Use of Tn5-gusA5 to investigate environmental and nutritional e€ects on gene expression in the coronatine biosynthetic gene cluster of Pseudomonas syringae pv. glycinea. Can J Microbiol 43:517±525 Parry RJ, Mhaskar SV, Lin MT, Mafoti R (1994) Investigations of the biosynthesis of the phytotoxin coronatine. Can J Chem 72: 86±99 PenÄaloza-VaÂzquez A, Bender CL (1998) Characterization of CorR, a transcriptional activator which is required for biosynthesis of the phytotoxin coronatine. J. Bacteriol 180:6252±6259 PenÄaloza-VaÂzquez A, Rangaswamy V, Ullrich M, Bailey AM, Bender CL (1996) Use of translational fusions to the maltosebinding protein to produce and purify proteins in Pseudomonas syringae and assess their activity in vivo. Mol Plant-Microbe Interact 9:637±641 Rangaswamy V, Ullrich M, Jones W, Mitchell R, Parry R, Reynolds P, Bender CL (1997) Expression and analysis of coronafacate ligase, a thermoregulated gene required for production of the phytotoxin coronatine in P. syringae. FEMS Microbiol Lett 154:65±72 Rangaswamy V, Jiralerspong S, Parry R, Bender CL (1998) Biosynthesis of the Pseudomonas polyketide coronafacic acid requires monofunctional and multifunctional polyketide synthase proteins. Proc Natl Acad Sci USA 95:15469±15474 Reitzer LJ, Movsas B, Magasanik B (1989) Activation of glnA transcription by nitrogen regulator I (NRI)-phosphate in Escherichia coli: evidence for a long-range physical interaction between NRI-phosphate and RNA polymerase. J Bacteriol 171:5512±5522 Rohde BH, Pohlack B, Ullrich MS (1998) Occurrence of thermoregulation of genes involved in coronatine biosynthesis among various Pseudomonas syringae strains. J Basic Microbiol 38:41± 50 Sambrook J, Fritsch EF, Maniatis T (1989) Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY

Ste€en P, Goyard S, Ullmann A (1996) Phosphorylated BvgA is sucient for transcriptional activation of virulence-regulated genes in Bordetella pertussis. EMBO J 14:633±641 Tse-Dinh YC, Qi H, Menzel R (1997) DNA supercoiling and bacterial adaptation: thermotolerance and thermoresistance. Trends Microbiol 5:323±326 Ullrich M, Bender CL (1994) The biosynthetic gene cluster for coronamic acid, an ethylcyclopropyl amino acid, contains genes homologous to amino acid-activating enzymes and thioesterases. J Bacteriol 176:7574 ±7586 Ullrich M, Guenzi AC, Mitchell RE, Bender CL (1994) Cloning and expression of genes required for coronamic acid (2-ethyl-1aminocyclo-propane-1-carboxylic acid), an intermediate in the biosynthesis of the phytotoxin coronatine. Appl Environ Microbiol 60:2890±2897 Ullrich M, PenÄaloza-VaÂzquez A, Bailey AM, Bender CL (1995) A modi®ed two-component regulatory system is involved in temperature-dependent biosynthesis of the Pseudomonas syringae phytotoxin coronatine. J Bacteriol 177:6160±6169 Venkatesan M, Buysse JM, Kopecko DJ (1988) Characterization of invasion plasmid antigen (ipaBCD) genes from Shigella ¯exneri: DNA sequence analysis and control of gene expression. Proc Natl Acad Sci USA 85:9317±9321 Walker MS, DeMoss JA (1994) NarL-phosphate must bind to multiple upstream sites to activate transcription from the narG promoter of Escherichia coli. Mol Microbiol 14:633±641 Weiler EW, Kutchan TM, Gorba T, Brodschelm W, Niesel U, Bublitz F (1994) The Pseudomonas phytotoxin coronatine mimics octadecanoid signaling molecules of higher plants. FEBS Lett 345:9±13 Wiebe WL, Campbell RN (1993) Characterization of Pseudomonas syringae pv. maculicola and comparison with P.s. tomato. Plant Dis 77:414±419 Xiao Y, Lu Y, Heu S, Hutcheson SW (1992) Organization and environmental regulation of the Pseudomonas syringae pv. syringae 61 hrp cluster. J Bacteriol 174:1734±1741