Binding of Pseudomonas aeruginosa AlgZ to Sites Upstream of the algZ Promoter Leads to Repression of Transcription

Abstract

Mucoid variants of the opportunistic pathogen Pseudomonas aeruginosa produce the exopolysaccharide alginate and colonize the respiratory tracts of cystic fibrosis patients. The genes encoding the alginate biosynthetic enzymes are clustered in a single operon, which is under tight transcriptional control. One essential activator of the alginate operon is AlgZ, a proposed ribbon-helix-helix DNA binding protein that shares 30% amino acid identity with the Mnt repressor of Salmonella enterica serovar Typhimurium bacteriophage P22. In the current study, we examined the role of AlgZ as an autoregulator. Using single-copy algZ-lacZ transcription fusions, an increase in algZ transcription was observed in an algZ mutant compared to the isogenic wild-type strain, suggesting that AlgZ may have an additional role as a repressor. To identify the AlgZ binding site, overlapping regions upstream of algZ were incubated with AlgZ and analyzed by electrophoretic mobility shift assays. Specific binding activity was localized to a region spanning from 66 to 185 base pairs upstream of the algZ transcriptional start site. Two AlgZ binding sites were defined using copper-phenanthroline footprinting and deletion analyses, with one site centered at 93 base pairs and the other centered at 161 base pairs upstream of the algZ promoter. Deletion of both binding sites resulted in the loss of AlgZ binding. These results indicate that AlgZ represses algZ transcription, and this activity is mediated by multiple AlgZ-DNA interactions.

Cystic fibrosis (CF) is the most common autosomal recessive disorder of the Caucasian population, affecting approximately one in every 2,500 live births in the United States (30). Patients with CF bear mutations in the gene encoding the cystic fibrosis transmembrane conductance regulator (CFTR), which is expressed on the apical membranes of secretory epithelial cells and regulates the rate of chloride and bicarbonate movement across epithelia (30, 31). Mutations in CFTR lead to mislocalization of the protein or loss of function, resulting in defective chloride ion transport across epithelial cell surfaces. In the airway, this leads to highly viscous submucosal gland secretions and an impaired mucociliary escalator, which allow trapped bacteria and other microorganisms to successfully colonize the lung (15, 23).

Nonmucoid isolates of the opportunistic pathogen Pseudomonas aeruginosa colonize the respiratory tract surface of the CF lung following persistent infections by organisms such as Staphylococcus aureus and nontypeable Haemophilus influenzae (1, 32). As lung disease progresses, mucoid variants of P. aeruginosa emerge and dominate the population (6). Exopolysaccharide expression leads to the formation of an endobronchial biofilm, which encases the microorganisms in a dense polysaccharide matrix that allows the bacteria to escape the lung's abundant phagocytic cell population (10, 18, 33, 35). Once established, mucoid P. aeruginosa microcolonies are seldom eradicated, leading to chronic infection of the lungs and perpetuating progressive pulmonary disease (28).

Mucoidy is a descriptive term for the production of the exopolysaccharide alginate, which is a negatively charged linear copolymer of β-1,4-linked d-mannuronic acid (partially O acetylated), and its C-5 epimer, α-l-guluronic acid (11, 19). P. aeruginosa alginate has been shown to prevent complement activation and neutrophil chemotaxis, as well as to resist phagocytosis and scavenge free radicals released by activated macrophages (7, 27, 34). Expression of the mucoid phenotype is largely due to the acquisition of stable mutations in at least two regions of the chromosome, collectively referred to as the muc loci (12). Of the three classes of muc mutations that have been shown to lead directly to alginate production, the muc-2 and muc-22 mutations are within mucA, which encodes an anti-sigma factor responsible for sequestering the alternative sigma factor AlgT (AlgU; σ²²) to the inner membrane (4, 20, 22). Over 80% of mucoid P. aeruginosa isolates collected from CF patients demonstrated mutations in mucA (4), suggesting that the environment of the CF lung may select for variants that acquire stable mutations within this region of the chromosome.

Most of the P. aeruginosa genes encoding the alginate biosynthetic enzymes are clustered in a single operon, and the first gene of this operon, algD, encodes GDP-mannose dehydrogenase. GDP-mannose dehydrogenase catalyzes the rate-limiting step in alginate biosynthesis by converting GDP-mannose to GDP-mannuronic acid (37, 40). Transcription of algD is initiated by AlgT, which shares homology with the Escherichia coli extreme heat shock sigma factor σ^E and is responsible for the expression of several regulatory factors that positively enhance algD operon transcription (21, 24, 47). These include the response regulators AlgR and AlgB and the transcriptional activator AlgZ (2, 46).

AlgZ (GenBank accession no. AF139988) (3), PA3385 (39), is a proposed member of the ribbon-helix-helix (RHH) family of DNA binding proteins and shares 30% amino acid identity with the Mnt repressor of the Salmonella enterica serovar Typhimurium bacteriophage P22 (3). AlgZ binding activity was first observed while examining cell extracts from the mucoid P. aeruginosa CF isolate FRD1 for binding to sequences upstream of the algD promoter (2). AlgZ expression is conserved among mucoid CF isolates, and AlgZ is essential for algD transcription and alginate production (3). Recent studies have shown that algZ transcription is dependent on AlgT, suggesting that regulation of algD transcription is controlled by AlgT and AlgT-mediated expression of algZ (47).

Many DNA binding proteins in prokaryotes act as either activators or repressors of transcription (5, 14, 38). Most members of the RHH family of DNA binding proteins are repressors (29, 41). Some, such as PutA from E. coli and ParD of the broad-host-range plasmid RP4/RK2, also serve as autoregulators (13, 25). Here, we present evidence that AlgZ binds specifically to two sites upstream of its promoter and that subsequent protein-DNA interactions lead to repression of algZ transcription. Thus, AlgZ has a dual role as an activator of an essential P. aeruginosa virulence determinant and a repressor of its own synthesis. To our knowledge, this is the first report of an RHH protein exhibiting both activator and repressor functions in vivo.

MATERIALS AND METHODS

Bacterial strains, plasmids, chemicals, oligonucleotides, and growth conditions.

For all manipulations, E. coli and P. aeruginosa were cultured as previously described (47). All P. aeruginosa strains were derived from FRD1 (mucA22), a mucoid CF isolate (26). FRD1310 (mucA22 attB::algZ-lacZ) was generated as previously described (47). FRD1312 (mucA22 algZ::xylE aacC1 attB::algZ-lacZ) was generated as follows. The 1.8-kb algZ fragment from pDJW585 (47) was cloned into pEX18Ap to generate pDJW586. Plasmid pDJW586 was cut with XhoI and filled in with Klenow fragment, and a 2.2-kb SmaI fragment from pX1918G containing an xylE aacC1 cassette was ligated to give an algZ::xylE aacC1 insertion in pEX18Ap (pDJW588). algZ::xylE aacC1 was substituted for wild-type algZ in the FRD1 chromosome to generate FRD1200 (mucA22 algZ::xylE). A single-copy algZ-lacZ fusion was placed at the attB site as previously described (47), resulting in FRD1312. FRD2503 (mucA22 algZ19 attB::algZ-lacZ) was generated as follows. The 1.8-kb BamHI fragment from pDJW585 was subcloned into pALTER (pPJ136), and site-directed mutagenesis (Promega Altered Sites) was used with oligonucleotides algZ11 and algZ12 (Table 1) to insert NdeI and NotI restriction sites flanking algZ (pAB1). The BamHI-HindIII fragment from pAB1 was subcloned into pEX18Ap (pAB2). This vector was cut with NdeI and NotI, and the ends were filled in using Klenow fragment. An omega tetracycline cassette from pHP45ΩTc was cut with SmaI and ligated to generate pAB3. This construct was then substituted for wild-type algZ in the chromosome of FRD1 to generate FRD1224 (mucA22 ΔalgZ::Ωtet). Site-directed mutagenesis of pPJ136 was used with mutant oligonucleotide algZ19 (Table 1) to generate the algZ19 allele, which substituted an alanine residue for AlgZ residue Arg22 (pPJ150). The BamHI-HindIII fragment from pPJ150 was subcloned into pEX18Ap, resulting in pPJ157. This was then substituted for the ΔalgZ::Ωtet allele in FRD1224 to generate FRD2238, which expressed AlgZ R22A. A single-copy algZ-lacZ fusion was placed at the attB site of FRD2238, resulting in FRD2503. All manipulations in E. coli were performed in either the JM109 or BL21(DE3) strain (Promega). All oligonucleotides used in this study are listed in Table 1 and were synthesized by the DNA Synthesis Core Laboratory at Wake Forest University School of Medicine. Chemicals and molecular biology reagents were purchased from Sigma or Promega unless otherwise stated.

TABLE 1.

Oligonucleotides used in this study

Description	Oligonucleotide sequence	Reference
algB12	5′-GTCGATGACGAGTCGGCGAT	3
algB63	5′-CATCTGGGCGAGAACGTCGAGCC	This study
algB66a	5′-GGACTAGTCCACTTTCCGTTATTGCC	This study
algB67a	5′-CCGCTCGAGCGGGAAGCACAGGTCGAA	This study
algD5	5′-AAGGCGGAAATGCCATCTCC	3
algD7	5′-AGGGAAGTTCCGGCCGTTTG	3
algD40	5′-GAATGCATGGCGGCTGAAAGTTC	This study
algZ11a	5′-CTCTACAGGTTCAACATATGCGCCCACTG	This study
algZ12a	5′-CACACGATGCGGAGCTGGCCGCGGCCGCCTGAGCGCCGAAGCCCACG	This study
algZ14	5′-AACGACCGGTGGTCAGAAGG	This study
algZ15	5′-GTTGCCTGTTTCAGTGGGCG	47
algZ16	5′-GTCGGGCGATGATCTCGGAG	This study
algZ19a	5′-CAAATTCGTCGTTGCTCTGCCCGAGGG	This study
algZ24a	5′-GGTGTAGACCAAGCTTGAAGGAGACTG	47
algZ25	5′-GCCTATGACCACGATTTCCG	This study
algZ26	5′-CGCGGATCCATGCGCCCACTGAAACAGGCA	This study
algZ27	5′-CCGGAATTCTCAGGCCTGGGCCAGCTCCGC	This study
algZ32	5′-AGGAAGCGTTCATTGTTCTC	This study
algZ33	5′-AGACGCCAGCGGCGCCATCT	This study
algZ34	5′-GTCAATTGTGCGTTGCGTGC	This study
algZ35	5′-CGCACAATTGACGTCAACTA	This study
algZ36	5′-GAACGCTTCCTCGGCACCGC	This study
algZ37	5′-TTGTCAGCGGTACGGCTGGA	This study
algZ39	5′-ACCACCGGTCGTTCAAGAGA	This study
algZ47a	5′-TAGTTGACGTCACTCGAGCGTTGCGTGCCGGCG	This study
algZ48a	5′-TTTGCCAGTACCAGCTATCTCTCGAGCGAGCGATTCAGCAAAAGC	This study
algZ57a	5′-TGGCCTTCTGACCACCGGACTAGTAAGAGATTGTCACTCG	This study
algZ60a	5′-AACAACAATTAGTTGACTCGAGTTGTGCGTTGCGTGCC	This study
algZ62	5′-GAAGGAGACTGTGTCAGC	This study
algZ67	5′-TACTGATCGATTGACCACCGGTCGT	This study
algZ68	5′-TACTGATCGATCCCAGTAAA	This study
algZ69	5′-TACTGGAATTCGTTGCCTGTTTCAG	This study

^aOligonucleotides with nucleotide substitutions indicated in boldface.

Transcriptional reporter assays.

Transcriptional reporter assays were essentially performed as described previously (47), with the following differences. Liquid cultures of P. aeruginosa strains FRD1310, FRD1312, and FRD2503 were grown to mid-exponential phase and then plated for overnight growth on Luria agar lacking sodium chloride at 37°C. Cells were scraped from the plate, pelleted, and resuspended in working buffer (61 mM Na₂HPO₄, 39 mM NaH₂PO₄, 10 mM KCl, 10 mM MgSO₄ · 7H₂O, 400 μM dithiothreitol, pH 7.0). The cells were permeabilized by addition of 20 μl of chloroform and 10 μl of 0.1% sodium dodecyl sulfate (SDS). The substrate o-nitrophenyl-β-d-galactopyranoside was added to the permeabilized cells, and the reaction mixtures were incubated at room temperature until a color change was visible. Reactions were stopped by the addition of 1 M sodium carbonate. The absorbances of the samples at 420 nm, 550 nm, and 660 nm were read, and the β-galactosidase activity (amount of o-nitrophenyl-β-d-galactopyranoside hydrolyzed per minute as a function of cell density) was calculated. Values for six separate experiments were averaged, and data were compared using a one-way analysis of variance test from GraphPad InStat version 3.06 (GraphPad Software).

Overexpression of AlgZ and AlgZ R22A, extract preparation, and immunoblot analysis.

Plasmid pPJ145, which expresses native AlgZ (47), was transformed into E. coli BL21(DE3) cells. The bacteria were grown to mid-log phase, and protein expression was induced for 3 hours by the addition of 1 mM isopropyl-thio-β-d-galactopyranoside (IPTG). The cells were spun at 8,000 rpm and resuspended in fractionation buffer with glycerol (10 mM Tris-HCl, pH 8.0, 100 mM NaCl, 1 mM MgCl₂, 5% glycerol). Following cell rupture by French press, the crude lysate was spun at 35,000 rpm and passed through a 0.45-μm filter. To overexpress AlgZ R22A in E. coli, the algZ19 allele was subcloned into a pET29a vector (pPJ162). The procedure for protein expression and extract preparation for AlgZ R22A is the same as for AlgZ, as stated above. Protein levels were quantitated by Bradford assay (9), and the titrations used in EMSAs for AlgZ or AlgZ R22A proteins (Fig. 1) were visualized by immunoblot analysis using an AlgZ-specific polyclonal antibody as previously described (47).

FIG. 1.

AlgZ binds to the algZ promoter. EMSAs were used to test binding interactions between wild-type AlgZ or variants and end-labeled fragments (−278 to +202 bp) containing the algZ promoter. (A) Labeled DNA fragments were incubated with different amounts of an extract derived from E. coli BL21(DE3)/pPJ145 (AlgZ, lanes 3 to 10) or BL21(DE3)/pET29a (expression vector, lane 2) in the following amounts: lane 1, 0 μg; lane 2, 0.5 μg; lane 3, 0.25; lane 4, 0.5 μg; lane 5, 0.65 μg; lane 6, 1 μg; lane 7, 1.5 μg; lane 8, 2 μg; lane 9, 2.2 μg; lane 10, 2.5 μg. (B) Corresponding Western blot illustrating the amount of AlgZ that was added to each lane of the DNA binding assays shown in panel A. (C) EMSA containing labeled DNA incubated with an E. coli extract derived from BL21(DE3)/pPJ162 (AlgZ R22A, lanes 3 to 10) or BL21(DE3)/pET29a (expression vector, lane 2) in the following amounts: lane 1, 0 μg; lane 2, 0.5 μg; lane 3, 0.5 μg; lane 4, 1 μg; lane 5, 1.5 μg; lane 6, 2 μg; lane 7, 2.5 μg; lane 8, 5 μg; lane 9, 10 μg; lane 10, 15 μg. (D) Corresponding Western blot of AlgZ R22A illustrating the amount of AlgZ R22A that was added to each lane of the DNA binding assays shown in panel C. (E) EMSA with wild-type (lane 2) and His-tagged AlgZ (lanes 3 to 10) in the following amounts: lane 1, 0 μg; lane 2, 0.5 μg of wild-type AlgZ from BL21(DE3)/pPJ145 extract; lane 3, 0.092 μg of pure His-AlgZ; lane 4, 0.21 μg; lane 5, 0.30 μg; lane 6, 0.39 μg; lane 7, 0.51 μg; lane 8, 0.60 μg; lane 9, 0.69 μg; lane 10, 0.81 μg. (F) EMSA with native R22A (lane 2) and His-AlgZ R22A (lanes 3 to 10) in the following amounts: lane 1, 0 μg; lane 2, 0.5 μg of native R22A AlgZ from BL21(DE3)/pPJ162 extract; lane 3, 0.10 μg of pure His-AlgZ R22A; lane 4, 0.20 μg; lane 5, 0.30 μg; lane 6, 0.41 μg; lane 7, 0.51 μg; lane 8, 0.61 μg; lane 9, 0.71 μg; lane 10, 0.79 μg.

Construction and purification of N-terminal His-AlgZ and His-R22A AlgZ.

The algZ gene was cloned into pTrcHisA (Invitrogen) using primers algZ26 and algZ27 and template pPJ145, resulting in pDR2. Cleared extracts derived from E. coli JM109/pDR2 were prepared as described above. The cleared extract was added to a slurry containing Ni-nitrilotriacetic acid magnetic agarose beads (QIAGEN) and binding buffer (50 mM NaH₂PO₄, 300 mM NaCl, 20 mM imidazole, pH 8.0), and the sample was allowed to mix at 4°C for 1 hour. Residual proteins were removed with two wash steps, and the pure protein was eluted from the beads by the addition of binding buffer containing 250 mM imidazole and 0.005% Tween. Protein expression and purification steps were monitored by SDS-polyacrylamide gel electrophoresis (PAGE) and staining with Coomassie brilliant blue (Fischer). To generate His-R22A AlgZ, the algZ19 allele was cloned into pTrcHisA (Invitrogen) using primers algZ26 and algZ27 and template pPJ150, resulting in pHL22. Preparation of cleared extracts and protein purification were performed as described above.

Isolation and labeling of DNA.

DNA fragments were isolated and labeled as previously reported (2). In brief, fragments containing portions of algZ, algD, or algB were amplified by PCR using Taq polymerase and oligonucleotides listed in Table 1. Template DNA was contributed by pDJW585 (algZ) (47), pDJW221 (algD) (45), or pJG194 (algB) (3). For the labeled fragment used in Fig. 1, algZ sequences (−278 to +202 bp) were amplified with the algZ33 and algZ16 primers (Table 1). For the overlapping PCR assay (Fig. 2), end-labeled fragments were generated with primer pairs algZ33 and algZ37 (−278 to +109 bp), algZ33 and algZ39 (−278 to −173 bp), algZ14 and algZ34 (−185 to −66 bp), algZ32 and algZ35 (−77 to +20 bp), and algZ36 and algZ37 (+10 to +109 bp). One primer for each PCR was end labeled using T4 polynucleotide kinase and [γ-³²P]ATP, as described in the manufacturer's instructions. Following PCR amplification, excess [γ-³²P]ATP was removed using G-50 columns (Amersham). Labeled DNA was then extracted using phenol-chloroform-isoamyl alcohol, ethanol precipitated, dried, and resuspended in a suitable amount of deionized water. For the competition assays, the algZ14 primer was labeled as described above and incubated in a PCR with algZ25 to generate a 183-bp end-labeled fragment (−185 to −3 of algZ). Unlabeled competitor DNAs 183 bp in length were generated by PCR amplification of pDJW585 (algZ) using algZ14 and algZ25 primers, of pDJW221 (algD) using algD5 and algD40 primers, and of pJG194 (algB) using algB12 and algB63 primers.

FIG. 2.

AlgZ binding is localized to a region spanning from −185 to −66 bp upstream of algZ. (A) The sequence of algZ (−278 to +109 bp) was divided into four overlapping fragments (fragments A to D). Fragment A encompassed sequences from −278 to −173 bp. Fragment B encompassed sequences from −185 to −66 bp. Fragment C encompassed sequences from −77 to +20 bp. Fragment D encompassed sequences from +10 to +109 bp. (B) DNA fragments were end labeled and incubated either alone (lanes 1, 4, 7, 10, and 13), with 0.5 μg of an extract derived from E. coli BL21(DE3)/pET29a (expression vector, lanes 2, 5, 8, 11, and 14), or with 0.5 μg of an extract derived from BL21(DE3)/pPJ145 (AlgZ, lanes 3, 6, 9, 12, and 15) in an EMSA. The amount of AlgZ used in this experiment was determined empirically and verified by phosphorimaging densitometry to be the amount of protein required to shift 89% of the free DNA containing all four fragments (A to D) (lane 3). Arrows indicate AlgZ-DNA complexes.

EMSA.

DNA binding assays and conditions were similar to those previously reported (2). All DNA binding assays were performed at 25°C for 10 min in a reaction volume of 10 μl unless otherwise noted. Gels were dried at 80°C for 20 min and exposed to a phosphorimaging screen for variable amounts of time. For the electrophoretic mobility shift assays (EMSAs) in Fig. 1, the following methods were used. Increasing amounts of an E. coli extract enriched for either AlgZ (pPJ145) or AlgZ R22A (pPJ162) or purified His-AlgZ (pDR2) or His-R22A AlgZ (pHL22) were incubated in a DNA binding assay with an end-labeled algZ fragment. Reactions were processed as described above. To verify loss of binding activity of AlgZ R22A at the AlgZ binding site at algD (data not shown), an end-labeled algD fragment was incubated with increasing amounts of an E. coli extract enriched for AlgZ R22A in a DNA binding assay. Reactions were processed as described above. For the EMSA in Fig. 2, fragments were incubated with 0.5 μg of AlgZ-enriched extract (pPJ145) in a DNA binding assay with an overall reaction volume of 20 μl. Reactions were then processed as described above, except the gel was allowed to run for 2.5 h at 200 V and dried for 1 hour at 80°C. The dried gel was exposed to a phosphorimaging screen, and the amount of free DNA was quantitated using ImageQuant version 5.2 software (Molecular Dynamics).

Competition assay.

Quantitation of labeled and unlabeled fragments for the competition assay (see above) was performed using an ethidium bromide spot plate and Alpha Ease version 5.5 software (Alpha Innotech Corporation). Following standardization of the concentrations of all DNA fragments, approximately 4 pmol of labeled DNA was incubated in a DNA binding reaction (in a 40-μl overall volume) without protein or with 1 μg of an AlgZ-enriched extract, as well as 3 μg of poly(dI-dC) and binding buffer. Unlabeled competitor DNAs were then added to specified reactions at a ratio of 1:1-, 1:5-, 1:10-, 1:25-, 1:50-, or 1:100-fold excess of competitor. All reactions were incubated at 25°C for 10 minutes and separated on a 4% PAGE gel for 2 h at 200 V. The gel was dried at 80°C for 45 min and exposed to a phosphorimaging screen. The amount of free DNA was quantitated using ImageQuant software.

In situ copper phenanthroline footprinting.

In situ 1,10-phenanthroline-copper footprinting was performed as previously described (2). Primer algZ25 was end labeled with [γ-³²P]ATP and used in a PCR amplification of the algZ promoter (pDJW585) with unlabeled primer algZ14 (see Fig. 4). To footprint the coding strand (data not shown), primer algZ14 was end labeled with [γ-³²P]ATP and used in a PCR amplification of the algZ promoter (pDJW585) with unlabeled primer algZ25. These experiments were also repeated with His-AlgZ as the protein source (data not shown), and all reactions were processed as described previously (2). For another footprint (see Fig. 5), a similar protocol was followed, except primer algZ62 was end labeled with [γ-³²P]ATP and used in a PCR amplification of the algZ promoter (pDJW585) with unlabeled primer algZ34. End-labeled fragments were incubated in DNA binding assays containing 1.13 μg of purified His-AlgZ and footprinted as described above. For the noncoding strand (data not shown), primer algZ34 was end labeled with [γ-³²P]ATP and used in a PCR amplification of the algZ promoter (pDJW585) with unlabeled primer algZ62. The fragment was incubated with His-AlgZ and footprinted as described previously (2).

FIG. 4.

Copper-phenanthroline footprinting reveals an AlgZ binding site (algZ1) centered at −93 bp upstream of algZ. (A) AlgZ-algZ complexes subjected to copper-phenanthroline footprinting. DNA binding assays containing AlgZ and an end-labeled DNA fragment (−185 to −3 bp) were separated on a 4% acrylamide gel in an EMSA. A representative EMSA is shown, with lanes containing the following amounts of an extract derived from BL21(DE3)/pPJ145 (AlgZ): lane 1, 0 μg; lane 2, 0.13 μg; lane 3, 0.25 μg; lane 4, 0.5 μg; lane 5, 0.65 μg; lane 6, 1 μg; lane 7, 1.5 μg; lane 8, 2 μg; lane 9, 2.2 μg; lane 10, 2.5 μg. Free and complexed DNAs were subjected to in situ copper-phenanthroline footprinting, and cleaved DNAs from three higher-order complexes (1 to 3) were recovered from the gel, along with a control fragment (free algZ) lacking protein. These fragments are indicated by arrows. (B) Copper-phenanthroline footprinting of AlgZ-algZ complexes. Treated fragments (free algZ DNA [−] and complexes 1 to 3) were separated on a 6% PAGE gel and run adjacent to a sequencing ladder. Areas of footprint are indicated by solid boxes, and regions of hypersensitivity are indicated by striped boxes. (C) Sequence of the algZ1 binding site, centered approximately 93 bp upstream of the algZ promoter and covering a 38-bp region of DNA spanning from −112 to −74 bp. Footprints for both noncoding (B) and coding (data not shown) strands are shown. Footprinted regions are indicated by boldface type and solid boxes, and regions of hypersensitivity are indicated by striped boxes.

FIG. 5.

Identification of a second AlgZ binding site (algZ2) centered at −161 bp upstream of algZ using copper-phenanthroline footprinting. (A) AlgZ-algZ complexes subjected to copper-phenanthroline footprinting. DNA binding reactions were analyzed by EMSA and contained an end-labeled DNA fragment (−242 to −66 bp) and His-AlgZ in the following amounts: lane 1, 0 pmol; lane 2, 1.52 pmol; lane 3, 3.05 pmol; lane 4, 4.57 pmol; lane 5, 6.1 pmol; lane 6, 7.62 pmol; lane 7, 9.15 pmol; lane 8, 10.7 pmol; lane 9, 12.2 pmol; lane 10, 13.7 pmol. Free and complexed DNAs were subjected to in situ copper-phenanthroline footprinting, and cleaved DNAs from two higher-order complexes (1 and 2) were recovered from the gel, along with a control fragment (free algZ) lacking protein. These fragments are indicated by arrows. (B) Copper-phenanthroline footprinting of AlgZ-algZ complexes. Treated fragments (free algZ DNA [−] and complexes 1 and 2) were separated on a 6% PAGE gel and run adjacent to a sequencing ladder. Areas of footprint are indicated by solid boxes, and the region of hypersensitivity is indicated by a striped box. (C) Sequence of the algZ2 binding site, centered approximately 161 bp upstream of the algZ promoter and covering an 18-bp region of DNA spanning −170 to −153 bp. Footprinted regions for the coding strand (part B) are shown. Footprinted regions are indicated by boldface type and solid boxes, and the region of hypersensitivity is indicated by a striped box.

Mutagenesis of the AlgZ binding sites at algZ.

Promoter sequences from wild-type algZ were amplified from pDJW585 using primers algZ24 and algZ15 and cloned into the HindIII-SmaI sites of pALTER1 (Promega) to generate pPJ147. Site-directed mutagenesis of pPJ147 using primers algZ47 and algZ48 generated XhoI sites that flanked the algZ1 binding site (pDR8). This vector was cut with XhoI, and the ends were ligated to generate the algZ47 allele (pDR9), resulting in a 36-bp deletion (−111 to −76 bp). To generate algZ67, primers algZ24 and algZ67, as well as primers algZ68 and algZ69, were used in PCR to amplify two promoter fragments (−258 to −169 and −152 to +79) flanked by HindIII-ClaI and ClaI-EcoRI sites, respectively. The two PCR fragments were digested with ClaI and ligated and then digested with HindIII and EcoRI and cloned into pUCP18 (pHL33), resulting in a replacement of the algZ2 binding site (−170 to −153 bp) with an 8-bp substitution sequence (ATCGCGAT). Mutagenesis of pPJ147 using primers algZ57 and algZ60 generated SpeI and XhoI sites that flanked site B (−185 to −66 bp [Fig. 2A]), containing high-affinity AlgZ binding sites (pDR20). This vector was cut with SpeI and XhoI, and the ends were filled in with Klenow fragment and ligated together to generate algZ57 (pDR22), with a loss of 110 bp from the algZ promoter (−180 to −71 bp). To generate algZ58, a 110-bp fragment of algB DNA flanked with SpeI and XhoI restriction sites was amplified from pJG194 using algB66 and algB67. This fragment was digested and cloned into the SpeI-XhoI sites of pDR20 to generate a construct in which the high-affinity AlgZ binding sites were replaced with algB DNA (pDR29). End-labeled fragments (−258 to +79 of algZ) were amplified from pPJ147 (algZ), pDR9 (algZ47), pHL33 (algZ67), pDR22 (algZ57), and pDR29 (algZ58) using the algZ24 and algZ15 primers. Increasing amounts of purified His-AlgZ were incubated in DNA binding assays with these end-labeled fragments and processed in EMSA using the standard conditions described above.

RESULTS

AlgZ binds to the algZ promoter.

With evidence to suggest that members of the RHH family of DNA binding proteins can act as autoregulators (13, 25), we investigated whether AlgZ might also regulate its own transcription. First, we determined if AlgZ would bind to its own promoter in vitro. An end-labeled fragment containing sequences that flanked the mapped algZ transcriptional start point (47) was incubated with native AlgZ in an EMSA (Fig. 1A). An identical amount of AlgZ added to each DNA binding reaction was also processed by SDS-PAGE and probed with an AlgZ-specific antibody in an immunoblot assay (Fig. 1B). As the concentration of AlgZ increased (Fig. 1B), several AlgZ-algZ complexes were observed (Fig. 1A, lanes 3 to 10). These shifts were absent in the reaction mixture containing free DNA alone (Fig. 1A, lane 1) and in the reaction mixture with DNA incubated with an extract containing the expression vector (Fig. 1A, lane 2). The number of protein-DNA complexes increased from one (Fig. 1A, lane 6) to two (lane 7) with increasing concentrations of AlgZ (Fig. 1B, lanes 6 and 7) until the free DNA was completely shifted and two complexes were formed (Fig. 1A, lanes 8 and 9). These results suggest that AlgZ binds to the algZ promoter.

We then investigated whether the binding of AlgZ to the algZ promoter was specific. The Salmonella enterica serovar Typhimurium bacteriophage P22 repressor Mnt utilizes N-terminal amino acid residues Arg2, His6, Asn8, and Arg10 for high-affinity DNA contacts that are required for transcriptional regulation (16). AlgZ shares approximately 30% amino acid identity with Mnt, including conservation of an N-terminal arginine residue at position 22 that corresponds to Arg10 in Mnt (3). Site-directed mutagenesis was used to substitute an alanine residue for Arg22, and this variant (AlgZ R22A) was examined for its ability to bind to the algZ promoter. An end-labeled DNA fragment containing sequences flanking the algZ promoter was incubated with increasing concentrations of AlgZ R22A, and the protein-DNA complexes were separated using EMSA (Fig. 1C). Samples were also analyzed by immunoblot assay to detect the amount of AlgZ R22A added to each DNA binding reaction (Fig. 1D). In contrast to native AlgZ, AlgZ R22A failed to bind algZ DNA (Fig. 1C). This loss of binding activity was also observed when AlgZ R22A was tested for binding to the AlgZ binding site at algD (data not shown), suggesting that Arg22 is required for AlgZ DNA binding activity at multiple promoter regions and contributes specific contacts at each promoter.

The above-mentioned studies used native AlgZ or AlgZ R22A present in extracts of E. coli. A His-AlgZ variant was generated, purified, and tested for binding to the algZ fragment (see Materials and Methods). The binding activity of His-AlgZ for the algZ sequence was similar to the activity seen with native AlgZ (Fig. 1E), since multiple AlgZ-algZ complexes were formed with increasing His-AlgZ and the complexes exhibited mobilities similar to those formed with native AlgZ (Fig. 1E, compare lanes 2 and 3). We also generated a His-R22A AlgZ, which failed to bind DNA similarly to native AlgZ R22A (Fig. 1F). We concluded from these studies that the DNA binding activity of purified AlgZ was similar to the activity of the native protein. Both native and His-AlgZ proteins were used in the subsequent experiments.

AlgZ represses its own synthesis.

To directly test the role of AlgZ in autoregulation, we engineered single-copy algZ-lacZ transcription fusions at the neutral attB site in algZ⁺ (FRD1310) and algZ-null (FRD1312) P. aeruginosa strains. The β-galactosidase levels produced by each strain were measured in six separate experiments, and the averages are shown in Table 2. We observed a >2-fold increase in β-galactosidase activity over wild-type levels in the algZ-null strain FRD1312, suggesting that AlgZ functions as an autorepressor.

TABLE 2.

AlgZ represses transcription of the algZ promoter^a

Strain	Genotype	β-Galactosidase assay (Miller units)	P value	% FRD1310
FRD1310	mucA22 attB::algZ-lacZ	1,499 ± 143		100
FRD1312	mucA22 algZ::xylE aacC1 attB::algZ-lacZ	3,158 ± 130	<0.01	211
FRD2503	mucA22 algZ19 attB::algZ-lacZ	3,714 ± 207	<0.01	248

^aSingle-copy algZ-lacZ transcriptional fusions were placed at the neutral attB site in an isogenic algZ-plus (FRD1310), algZ-null (FRD1312), or AlgZ R22A (FRD2503) strain of P. aeruginosa. β-Galactosidase activity was recorded in Miller units, and results were averaged from six separate experiments. P values were calculated from a one-way analysis of variance test.

To determine if AlgZ-DNA contacts mediate repression of algZ transcription, we replaced wild-type algZ with an allelic variant that expressed AlgZ R22A (see Materials and Methods). The amount of algZ-lacZ transcription observed in the AlgZ R22A background (FRD2503) was similar to that seen in an algZ-null background (FRD1312), with a >2-fold increase in transcription over wild-type levels (Table 2). This result suggests that loss of AlgZ DNA binding to the algZ promoter (Fig. 1C) lifts the repression exerted by AlgZ on algZ transcription (Table 2). Collectively, the data in Fig. 1 and Table 2 support the hypothesis that AlgZ binds to the algZ promoter and contributes protein-nucleotide contacts that are required for repression of algZ transcription.

Localization of AlgZ binding at algZ.

To further define the AlgZ-algZ binding interaction, the sequence contained in the algZ-lacZ transcriptional fusion (−278 to +109 bp) was divided into four overlapping fragments designated A to D (Fig. 2A). These fragments were labeled, amplified, and incubated with AlgZ in an EMSA (Fig. 2B). Addition of AlgZ to the fragment with all four regions (A to D) (Fig. 2A) produced two protein-DNA complexes and shifted 89% of the DNA (Fig. 2B, lane 3) compared to controls containing either free DNA alone (lane 1) or DNA incubated with an extract containing an expression vector (lane 2). When this concentration of AlgZ was tested with fragments A to D individually, 81% of fragment B (−185 to −66 bp) was shifted upon addition of AlgZ (Fig. 2B, lane 9) and most closely resembled the binding activity observed with the fragment containing regions A to D (compare lanes 3 and 9). AlgZ bound weakly to fragments A and D individually, with approximately 2% of the free DNA of fragment A (−278 to −173 bp) shifted (Fig. 2B, lane 6) and 8% of the free DNA of fragment D (+10 to +109 bp) shifted upon addition of AlgZ (lane 15). Although a shift appeared to be present when AlgZ was incubated with fragment C (Fig. 2B, lane 12), there was not a significant percentage of free DNA lost upon addition of AlgZ (compare lanes 10 and 12). These results suggested that fragment B (−185 to −66 bp) contained high-affinity AlgZ binding site(s).

AlgZ binding upstream of algZ is a specific interaction.

Loss of DNA binding activity by AlgZ R22A (Fig. 1C) suggests that AlgZ binds specifically to the algZ promoter. To rigorously validate that AlgZ binds specifically to a site(s) found between −185 and −66 bp, we performed a competition assay. An end-labeled fragment which overlapped fragment B (Fig. 2A) was incubated with saturating amounts of AlgZ and increasing concentrations of unlabeled competitor DNA fragments. Free and complexed DNAs were analyzed by EMSA (Fig. 3). As increasing concentrations of the unlabeled specific competitor algZ fragment were added (lanes 2 to 7), the complexes were gradually titrated away. In contrast, when the AlgZ-algZ DNA binding assay mixtures were incubated in the presence of unlabeled nonspecific DNA, the protein-DNA complexes were not significantly titrated away until the unlabeled fragment was at 100-fold excess of the labeled fragment (lanes 9 to 14), suggesting that AlgZ interactions upstream of the algZ promoter are specific.

FIG. 3.

AlgZ binds specifically to a site(s) upstream of the algZ promoter. End-labeled DNA fragments encompassing sequences from −185 to −3 bp upstream of algZ were incubated with the indicated amounts of protein and competitor DNA in an EMSA. Lanes 1 to 7 and 9 to 20 contain 1 μg of an extract derived from BL21(DE3)/pPJ145 (AlgZ). The lanes contained the following molar excess concentrations of specific competitor DNA (lanes 2 to 7, algZ competitor; lanes 15 to 20, algD competitor) or nonspecific algB competitor DNA (lanes 9 to 14): lanes 2, 9, and 15 contained 1× concentration competitor DNA; lanes 3, 10, and 16 contained 5×; lanes 4, 11, and 17 contained 10×; lanes 5, 12, and 18 contained 25×; lanes 6, 13, and 19 contained 50×; lanes 7, 14, and 20 contained 100×. Lane 1 contained no competitor DNA, and lane 8 contained no protein or competitor DNA.

To compare the strength of binding to this region of algZ versus a previously published binding site, we used the mapped AlgZ binding site at algD (2) as a specific competitor (Fig. 3, lanes 15 to 20). Protein-DNA complexes were gradually titrated away, and some free DNA appeared as AlgZ was competed away by higher concentrations of specific algD competitor DNA (Fig. 3, lanes 19 to 20). However, the algZ fragment appears to be a more effective competitor than the algD fragment (Fig. 3, compare lanes 2 to 7 with lanes 15 to 20). Approximately 23% of the free DNA was unoccupied at 10× concentrations of the algZ competitor, but a 100× concentration of the algD competitor DNA was required to titrate away an equivalent amount of free DNA (Fig. 3, compare lanes 3 and 20). This suggests that the affinity of AlgZ for its promoter is higher than its affinity for algD.

Footprinting studies reveal two adjacent AlgZ binding sites positioned upstream of the algZ promoter.

To accurately identify the AlgZ binding site(s) in fragment B (−185 to −66 bp) (Fig. 2A), in situ copper-phenanthroline footprinting was performed. An end-labeled DNA fragment was incubated with increasing concentrations of AlgZ, and the protein-DNA complexes were subjected to EMSA. Three distinct complexes with differing mobilities were detected when increasing concentrations of AlgZ were added to the DNA binding reactions (Fig. 4A, compare lanes 4, 8, and 10), with complexes 1 and 2 producing signals of greater intensity than complex 3. Multiple DNA-binding reactions containing labeled DNA and AlgZ in the same amounts used in lane 4 or 10 were processed by EMSA, and complexes 1 to 3 (Fig. 4A) were subjected to in situ copper-phenanthroline footprinting. The individual complexes were excised from the gel, and the cleaved DNA was eluted and separated on a 6% sequencing gel adjacent to a sequencing ladder (Fig. 4B). The footprints were examined on both the noncoding (Fig. 4B) and the coding (data not shown) strands.

When complex 1 was analyzed by footprinting (Fig. 4B), an AlgZ binding site was identified along a string of A-T residues positioned 93 bp upstream of the transcriptional start site. Alternating regions of footprint and hypersensitivity covering an approximately 38-bp sequence from −112 to −74 bp were also noted (Fig. 4C). This sequence was designated algZ1. Interestingly, the footprinted region was not extended when complexes 2 and 3 were examined, suggesting that differing mobilities among AlgZ-algZ complexes were not caused by AlgZ tracking along the algZ1 binding site (Fig. 4A and B, complexes 2 and 3, respectively). Both DNA strands were footprinted, and the three complexes demonstrated identical patterns of footprint and hypersensitivity for each DNA strand examined (Fig. 4B and data not shown). This suggested that multiple-complex formation was not due to a binding preference of AlgZ for one DNA strand over another.

When we performed the footprinting assay described above, we were unable to resolve additional binding sites upstream of algZ1. To explore this possibility, an end-labeled fragment spanning −242 to −66 bp upstream of the algZ transcriptional start site was incubated with His-AlgZ in a DNA binding reaction. The sequence selected for this footprinting experiment covered both a portion of fragment A and all of fragment B (Fig. 2A), providing greater resolution of potential binding interactions immediately upstream of algZ1.

The protein-DNA complexes were separated by EMSA, and two protein-DNA complexes were readily observed (Fig. 5A, complexes 1 and 2). Following in situ copper-phenanthroline footprinting of the coding strand (Fig. 5B), a second AlgZ binding site, designated algZ2, was found to be centered at −161 bp upstream of the transcriptional start site (Fig. 5C). AlgZ bound to the coding strand along a region spanning approximately 18 bp (−172 to −155), which contained two footprinted areas separated by a region of hypersensitivity (Fig. 5B). The noncoding strand exhibited increased hypersensitivity upon addition of His-AlgZ, but no footprint was evident (data not shown). The algZ2 sequence covered a region of DNA that was approximately half of the algZ1 protection zone. At lower protein concentrations (complex 1) (Fig. 5A), the algZ2 region was unoccupied, as demonstrated by a lack of protection (Fig. 5B). It was only at higher protein levels (complex 2) (Fig. 5A) that the algZ2 site became occupied (Fig. 5B). This suggests an affinity difference between the two binding sites, with the algZ1 sequence occupied at both low and high AlgZ concentrations and the algZ2 binding site occupied only at higher concentrations. Another lower-affinity binding site(s) present in the sequence of fragment A (−278 to −173 bp) (Fig. 2B) was not detected in this experiment. From these studies, we conclude that AlgZ binds at least two sites upstream of the algZ promoter and exhibits a different affinity for each site.

Deletion of mapped AlgZ binding sites results in loss of DNA binding activity.

Based on the identification of two AlgZ binding sites, we questioned how the loss of these sites would affect the ability of AlgZ to bind to its promoter. We generated four mutant alleles by either deleting AlgZ binding sites (algZ47, algZ67, and algZ57) or substituting an identically sized nonspecific DNA in place of these binding sites (algZ58). The algZ47 allele contained a deletion of algZ1 (−111 to −76 bp), the algZ67 allele contained a deletion of algZ2 (−153 to −170), and the algZ57 allele contained deletions of both algZ1 and algZ2 (−180 to −71 bp) (Fig. 6A). The algZ58 allele contained a substitution of a 110-bp fragment of algB DNA in place of both algZ1 and algZ2 (Fig. 6A). End-labeled DNAs containing wild-type or allelic variants of the algZ promoter were incubated with increasing amounts of His-AlgZ and analyzed by EMSA (Fig. 6B to F).

FIG. 6.

Loss of mapped AlgZ binding sites at algZ results in loss of AlgZ binding to the algZ promoter. EMSAs were used to test binding interactions between His-AlgZ and end-labeled fragments (−258 to +79 bp of algZ) containing the wild-type algZ promoter or allelic variants. (A) Illustrations of the deletions and substitution of the 5′ regulatory region of the algZ promoter are shown. The wild-type algZ promoter contains two mapped AlgZ binding sites centered at −93 and −161 bp upstream of the transcriptional start site, designated algZ1 and algZ2, respectively. Deletion of the algZ1 binding site resulted in a loss of 36 bp (−111 to −76 bp) and was designated the algZ47 allele. Deletion of the algZ2 binding site resulted in a loss of 18 bp (−153 to −170 bp) of algZ and was designated the algZ67 allele. Removal of a 110-bp fragment (−180 to −71 bp) resulted in a deletion of both algZ1 and algZ2 binding sites and was designated the algZ57 allele. A fragment containing 110 bp of algB DNA was used to replace the 110-bp fragment (−180 to −71 bp) containing both mapped AlgZ binding sites, generating the algZ58 allele. (B to F) EMSAs containing labeled DNAs from the wild-type algZ promoter or allelic variants (algZ47, algZ67, algZ57, or algZ58) were incubated with His-AlgZ in the following amounts: lane 1, 0 pmol; lane 2, 0.61 pmol; lane 3, 1.22 pmol; lane 4, 1.83 pmol; lane 5, 2.44 pmol; lane 6, 3.66 pmol; lane 7, 4.88 pmol; lane 8, 6.10 pmol; lane 9, 7.32 pmol; lane 10, 8.54 pmol.

In comparison to the wild-type promoter (Fig. 6B), loss of either algZ1 or algZ2 reduced the number of AlgZ-algZ complexes (Fig. 6C and D, compare lanes 6 through 10 with identical lanes in Fig. 6B). Loss of both algZ1 and algZ2 (algZ57) completely eliminated AlgZ binding. The addition of an identically sized nonspecific sequence in place of algZ1 and algZ2 restored the phasing of the DNA but did not reestablish binding (Fig. 6B, algZ58). These results suggest that the algZ1 and algZ2 sites are bound specifically by AlgZ and contribute the necessary protein-DNA contacts that lead to repression of algZ transcription (Fig. 7).

FIG. 7.

Summary of regulation of algZ transcription and a proposed AlgZ consensus motif. (A) Relevant nucleotide sequences and features of the 5′ region of the algZ promoter (positions −218 to +74 relative to the transcription initiation site, designated +1) are shown. The −10/−35 boxes recognized by the alternative sigma factor AlgT (AlgU) are indicated in boldface type. Underlined sequences (from −112 to −74 for algZ1 and −170 to −153 for algZ2) indicate AlgZ binding sites as determined by copper-phenanthroline footprinting and deletion analyses. A portion of the AlgZ N-terminal amino acid residues is also listed. (B) Sequence comparison of AlgZ binding sites and proposed AlgZ consensus binding motif. The proposed consensus motif is shown with the following symbols: lowercase letters indicate wobble positions, uppercase letters indicate nucleotides conserved among at least three AlgZ binding sites, and arrows indicate a partial palindrome. algD refers to the previously reported AlgZ binding site at algD (2). The two AlgZ binding sites at algZ reported in this study are designated algZ1 and algZ2. The localization of the consensus motif within the binding sites, i.e., on the coding or noncoding DNA strands, is also listed.

Proposed AlgZ consensus motif.

Based on the identification of two AlgZ binding sites at algZ and comparison of these sites with the previously reported AlgZ binding site at algD (2), we propose that AlgZ recognizes a consensus motif with the partially palindromic sequence of 5′-gGCCAttACCagcc-3′, where lowercase letters indicate nucleotides conserved in at least two AlgZ binding sites and uppercase letters indicate nucleotides conserved in all three AlgZ binding sites (Fig. 7B). This consensus motif is observed twice in the algZ1 binding site, whereas the sequence is present in only one copy in the coding strands of algZ2 and the algD site (Fig. 7B).

DISCUSSION

In this study, we report that AlgZ binds at least two sites upstream of the algZ promoter, resulting in repression of transcription as mediated by protein-DNA contacts (Fig. 7). By comparing the AlgZ binding sites at algZ, designated in this study algZ1 and algZ2, and the AlgZ binding sequence at algD, a 14-bp consensus motif consisting of the partially palindromic sequence 5′-gGCCAttACCagcc-3′ emerged (Fig. 7B). We propose that AlgZ recognizes this site in a specific manner, leading to either activation or repression of transcription. Previously published mutagenesis studies of the algD site revealed that substitutions in the ATTAC portion of the AlgZ binding site resulted in decreased binding affinity and significantly decreased algD transcription (2). Conservation of these nucleotides among multiple binding sites suggests that they contribute contacts required for site-specific recognition and binding by AlgZ. In this study, we report that deletion of the algZ1 and algZ2 binding sites results in a loss of DNA binding activity, suggesting that these sites contribute contacts that are required for binding of AlgZ to its own promoter (Fig. 6B to F).

Based on sequence homology with the bacteriophage P22 repressor Mnt, AlgZ is proposed to be a member of the RHH family of DNA binding proteins, in which DNA binding activity is modulated by β-strand recognition and binding of nucleotides within the major groove of the DNA. Mnt is a tetramer in solution and binds a 17-bp operator, and each dimer of the tetramer is hypothesized to recognize one symmetric half-site (17, 41, 43). Although detailed structural information about soluble AlgZ or AlgZ bound to its operator DNA is currently unavailable, glutaraldehyde cross-linking studies demonstrate that AlgZ forms dimers and larger oligomeric species in solution (D. M. Ramsey and D. J. Wozniak, unpublished observations). Similar to other members of the RHH family, it is likely that AlgZ binds to its operator site as an oligomer, and the presence of two binding sites upstream of algZ would allow for multiple protein-DNA contacts that could effectively modulate algZ transcription. Binding of both the upper and lower DNA strands in algZ1 may strengthen the binding interaction, as evidenced by the observation that this binding site was occupied at both low and high protein concentrations (Fig. 4B). The algZ2 site is bound only at higher protein concentrations (Fig. 5B), and its reduced affinity may be due to the lack of an adjacent motif on the lower DNA strand that could stabilize AlgZ oligomer formation. The competition assay demonstrated that there is also an affinity difference of AlgZ for the binding sites at algZ versus algD (Fig. 3). This may be due to multiple binding sites present at algZ in comparison to the single site at algD, although differences in nucleotide sequence could contribute to the strength or weakness of individual binding interactions or binding interactions involving oligomers of AlgZ.

An alanine substitution for Arg22 of AlgZ results in a loss of binding to the algZ promoter (Fig. 1C and F) and at algD (data not shown), suggesting that AlgZ utilizes Arg22 for recognition of multiple binding sites. Because the DNA binding activity of the N-terminal portion of Mnt is distinct from the oligomerization properties of the C-terminal domain (44), the binding and oligomerization properties of AlgZ may also be mediated by separate regions. The oligomeric structure of AlgZ in solution and the boundaries of its DNA binding domain are questions that remain to be answered.

The ability of AlgZ to contact or affect the rate of transcription initiation by RNA polymerase may depend on the position of the operator sites in relation to the promoter, the sequence of the operator sites, their affinity for AlgZ, and the number of operator sites within the promoter region itself. These may also contribute to the ability of AlgZ to switch roles from an activator to a repressor. The ability of an RHH protein to assume dual regulatory roles has been illustrated in vitro by studies of Arc, which represses the P_ant and P_mnt promoters during lytic growth of bacteriophage P22 (36, 42). Upon binding to specific variants of P_ant, Arc acted to slow open-complex formation and accelerate promoter clearance, thus acting as both a repressor and an activator in vitro (36). Here, we report evidence that the transcriptional activator AlgZ is capable of acting as a repressor of algZ transcription in vivo by a currently uncharacterized mechanism (Table 2). This illustrates not only the complexity of transcriptional control by an RHH DNA binding protein, but also the versatility that allows one protein to serve as both activator and repressor of two different genes.

The presence of at least two AlgZ binding sites upstream of the algZ promoter and the subsequent repression exerted by AlgZ suggest a complex regulatory mechanism modulated by both the alternative sigma factor AlgT and now the transcriptional regulator AlgZ (Fig. 7). This repression, however, is modest (approximately twofold) (Table 2), suggesting that the ability of AlgZ to repress its own transcription could be modulated based on the occupation of the binding sites. At low concentrations of AlgZ, the algZ1 binding site is occupied and may activate algZ transcription. As protein levels increase, however, both the algZ1 and algZ2 sites are occupied. Cooperative protein-protein interactions between AlgZ oligomers bound to both sites could introduce a torsional strain in the DNA that reduces the affinity of RNA polymerase for the promoter. Because the affinity of AlgZ for algZ2 appears to be weaker than for algZ1 and highly dependent on the protein concentration, the algZ2 site may be transiently occupied and serve as the switch between activation and repression. The ability of AlgZ to weakly repress transcription guarantees that a measurable amount of AlgZ will be present in cells in which the alternative sigma factor AlgT is derepressed, so that its role as an activator of algD transcription and influence on subsequent alginate production is not greatly attenuated by its role as repressor of its own transcription. Thus, the expression of the mucoid phenotype is sustained through the dual activities of repression and activation by the transcriptional regulator AlgZ.