Pseudomonas aeruginosa AlgZ, a ribbon–helix–helix DNA-binding protein, is essential for alginate synthesis and algD transcriptional activation


  • Patricia J. Baynham,

    1. Department of Microbiology and Immunology, Medical Center Boulevard, Wake Forest University School of Medicine, Winston-Salem, NC 27157, USA.
    Search for more papers by this author
  • Amy L. Brown,

    1. Department of Microbiology and Immunology, Medical Center Boulevard, Wake Forest University School of Medicine, Winston-Salem, NC 27157, USA.
    Search for more papers by this author
  • Lisa L. Hall,

    1. Department of Microbiology and Immunology, Medical Center Boulevard, Wake Forest University School of Medicine, Winston-Salem, NC 27157, USA.
    Search for more papers by this author
  • Daniel J. Wozniak

    1. Department of Microbiology and Immunology, Medical Center Boulevard, Wake Forest University School of Medicine, Winston-Salem, NC 27157, USA.
    Search for more papers by this author

Daniel J. Wozniak. E-mail; Tel. (+1) 336 716 2016; Fax (+1) 336 716 9928.


The Pseudomonas aeruginosa algD gene is the first gene of an operon encoding most of the enzymes necessary for biosynthesis of the exopolysaccharide alginate. Transcriptional activation of algD results in the high-level synthesis of alginate, an important P. aeruginosa virulence factor with antiphagocytic and adherence properties. Previously, we have identified a protein(s), AlgZ, expressed in mucoid P. aeruginosa CF isolates that specifically bound to sequences located 280 bp upstream of the algD promoter. Mutagenesis of the AlgZ DNA binding site and transcription assays were used to show that AlgZ was an activator of algD transcription. In the current study, the monomeric size of AlgZ was estimated to be between 6 kDa and 15 kDa by electroelution of a protein preparation from an SDS–PAGE gel and analysis of the fractions via protein staining and electrophoretic mobility shift assays. A biochemical enrichment procedure, resulting in a 130-fold enrichment for AlgZ, was devised, the protein identified and a partial amino-terminal sequence obtained. Using the P. aeruginosa Genome Project database, a complete sequence was obtained, and algZ was cloned and expressed in Escherichia coli. Expression of algZ was sufficient for the observed AlgZ DNA binding previously observed from extracts of P. aeruginosa. A protein database search revealed that AlgZ is homologous to the Mnt and Arc repressors of the ribbon–helix–helix family of DNA-binding proteins. An algZ deletion mutant was constructed in the mucoid CF isolate FRD1. The resulting strain was non-mucoid and exhibited no detectable algD transcription. As an indirect role in transcription would probably result in some residual algD transcription, these data suggest that AlgZ is an integral activator of algD and support the hypothesis that both AlgZ and the response regulator AlgR are involved in direct contact with RNA polymerase containing the alternative sigma factor, AlgT. The cloning of algZ is a crucial step in determining the mechanism of algD activation.


Individuals with the genetic disease cystic fibrosis (CF) are predisposed to recurrent pulmonary infections. The respiratory tract of CF patients is colonized by a number of bacteria, including the opportunistic pathogen Pseudomonas aeruginosa (Fick et al., 1992). Initial colonizing strains of P. aeruginosa are non-mucoid. Eventually, these strains, which are impossible to eradicate completely, develop a mucoid colony morphology. This mucoid phenotype results from overproduction of the exopolysaccharide alginate (Govan and Deretic, 1996). This virulence factor has been credited with protecting the bacteria from phagocytes and reactive oxygen species as well as promoting adherence (Marcus and Baker, 1985; Govan and Harris, 1986; Govan and Deretic, 1996). The production of alginate by P. aeruginosa isolates is correlated with a worsening prognosis for the CF patient (Fick et al., 1992; Zielinski et al., 1994). Respiratory failure is the major cause of death in 95% of CF patients, and P. aeruginosa is the major pathogen involved (Wood et al., 1976).

The production of alginate is regulated in a highly complex manner that is not completely understood (Wozniak and Ohman, 1994; Govan and Deretic, 1996). One element required for alginate production is the utilization of an alternative sigma factor, σ22 (known as AlgT or AlgU), which is a homologue of σE, the stress response sigma factor of Escherichia coli (Martin et al., 1993; Devries and Ohman, 1994). σ22 and σE are members of the extracytoplasmic function sigma factor subfamily of sigma factors (Lonetto et al., 1994; Missiakas and Raina, 1998). The algT(U ) operon at 68 min on the P. aeruginosa chromosome also encodes the mucABCD genes, which appear to modulate the activity of this sigma factor (Govan and Deretic, 1996; Ohman et al., 1996; Mathee et al., 1997). AlgT is required for the expression of the algD operon, which encodes most of the enzymes necessary for alginate biosynthesis (Deretic et al., 1987; Chitnis and Ohman, 1993; Wozniak and Ohman, 1994). Transcriptional activation of the algD operon is a key regulatory event, as this results in a significant change in cell physiology and morphology (Govan and Deretic, 1996). Therefore, this operon is tightly controlled, and much study has been devoted to determining the mechanism of its activation.

Several proteins have been shown to bind algD sequences and are required for full transcriptional activation (Fig. 1A). AlgR, a member of the response regulator two-component superfamily (Parkinson, 1993), binds three sites within the algD promoter and is required for mucoidy and algD transcription (Kato and Chakrabarty, 1991; Mohr et al., 1991; 1992). There are also two sites for the binding of integration host factor (IHF) protein (Mohr and Deretic, 1992; Toussaint et al., 1993, Wozniak, 1994) as well as CysB and E. coli CAP binding sites (1Fig. 1A; Devault et al., 1991; Delic-Attree et al., 1997). Mutagenesis studies suggest that both IHF and CysB play a role in algD activation, albeit much reduced in comparison with AlgR (Wozniak, 1994; Delic-Attree et al., 1997). Furthermore, additional genes are required for algD activation, but their mechanism of action is not completely understood. These include algB (Wozniak and Ohman, 1991; Goldberg and Dahnke, 1992), algP (or algR3 ) (Kato et al., 1990; Konyecsni and Deretic, 1990; Deretic et al., 1992) and algQ (Konyecsni and Deretic, 1990). It is possible that there are additional elements in the algD promoter that are responsive to these genes or that these gene products play a more general role in the stress response of the bacterium, which induces a σ22 response. The algD promoter is unusual in that several of the proteins described bind and apparently activate transcription from a distance (Fig. 1A). The resulting picture is of a highly regulated promoter more reminiscent of eukaryotes than of prokaryotes. As all of these genes are required, but none appear to be sufficient for full algD activation, this suggests a complex three-dimensional nucleoid structure for this promoter.

Figure 1.

. A. Map of the algD promoter and locations of cis sequences to which various proteins bind. Numbers are positions from the start of algD transcription. Proteins represented are: alginate and twitching motility regulator AlgR (Kato and Chakrabarty, 1991; Mohr et al., 1992; Whitchurch et al., 1996); CAP, E. coli CAP binding site (Devault et al., 1991); algD activator, AlgZ (Baynham and Wozniak, 1996); IHF, integration host factor (Mohr and Deretic, 1992; Toussaint et al., 1993; Wozniak, 1994); RNAP, RNA polymerase (Wozniak and Ohman, 1994; Deretic et al., 1987); CysB, regulator of cysteine biosynthesis (Delic-Attree et al., 1997). The DNA fragment used in all EMSA experiments is indicated. B. Analysis of eluates from the electroelution of P. aeruginosa FRD810 extracts enriched for AlgZ from a preparative SDS–PAGE gel. An SDS–PAGE silver-stained gel containing the starting solution (1.3 μg, lane 2) or 30 μl fractions from the electroelution experiment (fractions 3–19, lanes 3–19). C. Analysis of AlgZ-binding activity of the eluates. Samples are: starting solution (1.2 μg, lane 1); 31 μl pooled fractions 5–7 (lane 2); or 31 μl individual fraction 5 (lane 3); fraction 6 (lane 4); and fraction 7 (lane 5); or no protein (lane 6).

We have observed previously an algD-specific DNA-binding activity present in cell-free extracts from the mucoid CF isolate FRD1 (Wozniak and Ohman, 1994; Baynham and Wozniak, 1996). This binding activity could not be explained by any of the previously identified regulators of algD transcription. This binding activity was found to be specific, was localized to a sequence positioned 280 bp upstream of the algD promoter and was absolutely dependent on the alternative sigma factor σ22. This protein(s), designated AlgZ, was found to be an activator of algD using transcriptional analysis of algD alleles with mutations in the AlgZ binding site and validated AlgZ as a critical regulator of algD transcription and alginate synthesis in P. aeruginosa.

The cloning of algZ is an essential step in the determination of the mechanism of algD activation. In this study, a biochemical approach was used to clone algZ. Expression of algZ in E. coli resulted in the characteristic AlgZ–DNA complexes, which is consistent with this interaction being caused by a single protein. Construction of an algZ deletion mutant in the wild-type mucoid CF isolate FRD1 resulted in a non-mucoid phenotype. Analysis of an algZ deletion strain using a transcriptional fusion construct revealed no detectable algD transcription. AlgZ is homologous to Mnt and Arc of bacteriophage P22 from Salmonella typhimurium, members of the ribbon–helix–helix family of DNA-binding proteins (Raumann et al., 1994a). These data confirm AlgZ as an essential activator of algD transcription and provide information critical for further analysis of this complex promoter.


Biochemical characterization of AlgZ

In a previous study, we identified a protein or protein complex, AlgZ, which specifically bound to algD sequences (ZCS; AlgZ cis sequence) located ≈280 bp 5′ of the promoter (Fig. 1A) (Wozniak and Ohman, 1994; Baynham and Wozniak, 1996). Mutagenesis of the ZCS revealed AlgZ as a novel activator of algD transcription (Baynham and Wozniak, 1996). To characterize the interaction of AlgZ with the ZCS as well as with other proteins that activate algD, a cloning strategy was undertaken. In this approach, a biochemical enrichment of AlgZ was used to isolate the protein for amino-terminal sequence determination.

To determine the monomeric size of AlgZ, a protein sample was resolved by preparative SDS–PAGE and electroeluted under partially renaturing conditions (Fig. 1B). AlgZ DNA-binding activity was detected in the pooled fractions 5–7 (Fig. 1C, lane 2). An electrophoretic mobility shift assay (EMSA) of the individual fractions revealed that AlgZ DNA-binding activity was confined to fraction 5 (Fig. 1C, lane 3). The corresponding lane on the SDS–PAGE gel (Fig. 1B, lane 5) suggested a monomeric molecular mass for AlgZ of between 6 kDa and 15 kDa.

Enrichment for AlgZ results in the identification of a novel 12.3 kDa protein

While the gel elution strategy indicated an approximate monomeric molecular mass for AlgZ, results with subsequent Tris–tricine gel electrophoresis and silver stain analysis of the above fractions were ambiguous, probably because of limited AlgZ recovery from SDS–PAGE. Therefore, a biochemical enrichment scheme was devised to identify a specific polypeptide with AlgZ DNA-binding activity. To accomplish this, a 15 L culture of P. aeruginosa strain FRD810 was harvested, and cytoplasmic extracts were prepared by disruption of cells in a French pressure cell. After enrichment of the cytoplasmic proteins for AlgZ DNA-binding activity (see Experimental procedures ), the protein preparation was applied to a heparin agarose column and proteins eluted with a linear NaCl gradient. To quantify the enrichment procedure, the specific activity of samples from each step of the enrichment was determined. One unit of AlgZ-binding activity was arbitrarily defined as the amount of protein required to bind half of the algD ZCS DNA in an EMSA (see Experimental procedures ). The enrichment scheme from crude extracts through heparin agarose column chromatography resulted in a 130-fold enrichment of AlgZ DNA-binding activity (data not shown).

The individual fractions from heparin agarose chromatography were assayed for AlgZ DNA-binding activity by EMSA (Fig. 2A) and for protein content by resolution using a Tris–tricine PAGE gel followed by silver staining (Fig. 2B). In this experiment, AlgZ DNA-binding activity in the EMSA was compared with the intensity of the protein bands in the 6–15 kDa range. While several proteins present in the range appropriate for AlgZ were visualized, a single polypeptide of ≈12 kDa (arrow in Fig. 2B) corresponded with the DNA-binding activity observed in the EMSA.

Figure 2.

. Analysis of the fractions obtained from heparin agarose chromatography. A. EMSA assay for AlgZ DNA-binding activity. Samples are fractions from heparin column chromatography in order of increasing NaCl (lanes 1–9), 1 μl of a previous heparin eluate preparation (lane 10) or no protein (lane 11). Heparin column eluates were assayed using 1 μl of a 1:10 dilution. B. Tris–tricine silver-stained SDS–PAGE gel of the eluates from heparin column chromatography. Lanes 1–9 correspond to the same fractions as in (A). Each lane contains 30 μl of the heparin eluate fraction indicated (18–36 μg of total protein).

A sample of the heparin agarose-fractionated AlgZ was subsequently separated on a Tris–tricine PAGE gel, the 12 kDa putative AlgZ protein was recovered (see Experimental Procedures) and the amino-terminal sequence determined. The amino-terminal sequence obtained (MRPLKQATPTYSSRTADKFVVRL) was used in a tblastn search of the unfinished microbial genome database ( This analysis obtained a perfect match corresponding to bases 51304–51242 within contig 77 of the P. aeruginosa genome (; 15 December 1998 release). The translated product of this sequence yielded a predicted 12.3 kDa protein of unknown function in P. aeruginosa.

Expression of the algZ gene in E. coli is sufficient for AlgZ DNA-binding activity

In previous studies (Wozniak and Ohman, 1994; Baynham and Wozniak, 1996) and in data outlined in 1Figs 1C and 22A, it was unclear whether the DNA-binding activity representing AlgZ consisted of a single polypeptide or of a protein complex. In addition, it was not known whether the amino-terminal sequence of the 12.3 kDa polypeptide identified in Fig. 2 indeed represented AlgZ. To address this, it was necessary to express the gene encoding the 12.3 kDa protein in the heterologous host E. coli and test for AlgZ DNA-binding activity. This approach was appealing, as previous studies revealed no DNA-binding activity when extracts from E. coli JM109 were tested in an EMSA for binding to the algD ZCS. In order to clone the gene encoding the 12.3 kDa polypeptide, primers were designed that flanked the predicted coding sequence obtained from the P. aeruginosa genome. The gene encoding the 12.3 kDa protein was cloned via polymerase chain reaction (PCR) into pUC18, resulting in pPJ120. This plasmid was transformed into E. coli JM109, and the 12.3 kDa protein was expressed under the control of the pUC18 lac promoter. When extracts prepared from E. coli JM109 harbouring pPJ120 were examined by EMSA (Fig. 3A, lanes 3–7), the typical AlgZ–ZCS complexes were visible. Significantly, these protein–DNA complexes were of identical mobility to those observed with extracts derived from P. aeruginosa (Fig. 3A, compare lane 2 with lanes 3–7). As described above, when extracts prepared from E. coli JM109 harbouring the vector pUC18 were tested, no protein–DNA complexes were apparent (Fig. 3A, lanes 8–12). In order to confirm that the activity observed was caused by AlgZ and not by a non-specific DNA-binding protein, a competition experiment was performed using unlabelled specific and non-specific DNA (Fig. 3B). The protein–DNA complexes were disrupted only when specific ZCS competitor DNA was used (Fig. 3B, lanes 2–4). Taken together, these results verify that the gene encoding the 12.3 kDa protein identified in Fig. 2 indeed represents algZ and suggests strongly that this gene is sufficient for the AlgZ DNA-binding activity described previously (Wozniak and Ohman, 1994; Baynham and Wozniak, 1996).

Figure 3.

. EMSA demonstrating algZ expression in E. coli. A. EMSA of extracts prepared from E. coli JM109 containing pPJ120 (lanes 3–7) or pUC18 (lanes 8–12) were examined for AlgZ DNA-binding activity by EMSA. Samples contained 0.5 μg (lanes 3 and 8), 1.0 μg (lanes 4 and 9), 1.5 μg (lanes 5 and 10), 2.0 μg (lanes 6 and 11) or 3.0 μg (lanes 7 and 11)of protein. An extract enriched for AlgZ-binding activity (see Experimental procedures ) derived from P. aeruginosa FRD810 (lane 2, 2.0 μg) or no protein (lane 1) served as controls. B. Competition EMSA analysing the extracts prepared from pPJ120 (as above) and various amounts of specific (lanes 2–4) and non-specific (lanes 5–7) unlabelled competitor DNA. Lane 1 contains no competitor DNA; lane 8 contains no protein. Competitor is fivefold (lanes 2 and 5), 25-fold (lanes 3 and 6) or 100-fold (lanes 4 and 7) with regard to labelled DNA.

Deletion of algZ abolishes mucoidy and algD transcription

Previous studies involving plasmid-borne algD–cat transcriptional fusions containing ZCS mutations suggested that AlgZ played a role in algD activation (Baynham and Wozniak, 1996). In order to determine the contribution of algZ to alginate production and to algD transcription, an algZ mutant (FRD1220) was constructed in the mucoid strain FRD1 by allele exchange using a sacB-containing plasmid with an algZΔ ::tet allele (see Experimental procedures ). FRD1220 was scored for mucoidy in order to determine the effect of an algZ deletion on the production of alginate. Compared with the parental mucoid FRD1 strain, the algZ deletion mutant was non-mucoid (Fig. 4B). FRD1220 could be complemented back to mucoidy using a plasmid-borne genomic clone containing algZ (data not shown). This indicated that the phenotype was caused by an algZ deletion and not by secondary mutations or polar effects on genes downstream of algZ. As anticipated, when protein extracts prepared from FRD1220 were assayed for DNA-binding activity (Fig. 4A), no detectable protein–DNA complexes were present (Fig. 4A, lanes 2–3).

Figure 4.

. Deletion of algZ results in abolished DNA-binding activity and a non-mucoid strain. A. EMSA of protein extracts (2 μg) prepared from the AlgZ-producing strain FRD810 (lane 1), the algZ deletion mutant FRD1220 (lanes 2–3, 0.5 μg or 2 μg respectively) or no protein (lane 4). B. Alginate phenotype of the mucoid FRD1 strain (mucA22 ) and the isogenic algZ deletion mutant FRD1220 (mucA22 algZΔ::Ωtet ) inoculated on a LANS plate.

In order to quantify the effect of an algZ deletion on algD transcription, the algZΔ ::tet allele was constructed in FRD875, a strain in which the chromosomal wild-type algD is replaced with an algD::xylE transcriptional fusion. The resulting strain, designated FRD1222, was assayed for algD-xylE transcriptional activity (Table 1). For comparison, algD-xylE levels were determined in the parental mucA22 strain FRD875, along with isogenic strains harbouring mutations in genes encoding the response regulators algB (FRD879) or algR (FRD880). As observed previously (Wozniak and Ohman, 1991; Woolwine and Wozniak, 1999), a mutation in algB resulted in an ≈20-fold reduction in algD transcription. However, no algD transcription was detectable in the algR or the algZ deletion strains. This verifies AlgZ as an essential activator of algD transcription.

Table 1. . AlgZ is essential for algD transcription. a. algD-xylE levels were determined as outlined in Experimental procedures. Data are expressed as the percentage of parental (FRD875) algD-xylE levels and represent the average of three independent experiments.Thumbnail image of

AlgZ is homologous to P22 phage repressors Mnt and Arc

In order to obtain the full-length, wild-type algZ gene, the PCR-cloned gene was used in a Southern hybridization experiment and revealed that algZ was contained on a 2.0 kb BamHI fragment (data not shown). The wild-type algZ gene from strain FRD1 was cloned into pUCP21T, the algZ insert was verified by PCR (see Experimental procedures ), and the region corresponding to algZ was sequenced and found to be identical to the corresponding locus within contig 77 of the P. aeruginosa genome database. The AlgZ protein sequence was used in a blastP search and revealed that AlgZ is homologous to Mnt and Arc, two repressor proteins of Salmonella typhimurium phage P22. These repressors contain a ribbon–helix–helix DNA-binding motif at their amino-terminus (Raumann et al., 1994a). Mnt and Arc are 40% identical and have the same tertiary fold (Burgering et al., 1994; Raumann et al., 1994b). A sim (local similarity program; Huang and Miller, 1991) alignment of AlgZ with Mnt is shown in Fig. 5. AlgZ was found to be 28% identical to Mnt. Many of the hydrophobic positions in AlgZ are conserved in Mnt. The carboxy-terminal region of Mnt, which does not appear to be as well conserved as AlgZ, plays a role in tetramer formation (Waldburger and Sauer, 1995). Biochemical and biophysical experiments currently under way in the laboratory will address AlgZ structure–function relationships and determine the significance of this homology.

Figure 5.

. Protein alignment of AlgZ with the Mnt protein of phage P22 of S. typhimurium. The portions of Mnt shown to function in DNA binding and dimerization compared with tetramerization are indicated (Waldburger, 1995). Identical amino acids are underlined. Numbers at the top indicate the amino acid position for AlgZ.


The majority of the genes required for alginate biosynthesis are in a large operon at 34 min on the P. aeruginosa chromosome. This operon is regulated by many factors at the level of the algD promoter (reviewed by Zielinski et al., 1994; Govan and Deretic, 1996). Previously, we identified a protein(s), designated AlgZ, which bound specifically to sequences upstream of the algD promoter and functioned as a novel activator of algD transcription (Baynham and Wozniak, 1996). In order to determine the precise role of algZ in algD transcription, isolation and characterization of the algZ gene was necessary.

In this study, a biochemical approach was used to clone algZ. This involved electroelution of a protein preparation from an SDS–PAGE gel in order to approximate the mass of AlgZ. After an enrichment procedure, AlgZ was identified and a partial amino-terminal protein sequence obtained. Using the P. aeruginosa Genome Project database, a complete sequence was obtained, and algZ was cloned and expressed in E. coli. Extracts prepared from E. coli expressing AlgZ exhibited a DNA-binding activity identical to that observed from extracts of P. aeruginosa. An algZ deletion mutant was constructed, and the resulting strain exhibited a non-mucoid phenotype and no detectable algD transcription. These data indicate that AlgZ is essential for algD activation.

A protein database search revealed that AlgZ belongs to a family of DNA-binding proteins containing the Mnt and Arc repressors of Salmonella typhimurium phage P22 (Waldburger, 1995). The interaction of Arc and Mnt with their respective operator DNA sequences has been well characterized (Knight et al., 1989; Knight and Sauer, 1989; Waldburger, 1995). Mnt and Arc bind to different operator DNA sequences (Vershon et al., 1987a,b). Likewise, there is no obvious similarity between the algD ZCS and either of these operator sequences (Baynham and Wozniak, 1996). Both Arc and Mnt repressors are tetrameric when bound to their respective operator sequences (Vershon et al., 1987a; Knight et al., 1989). Although the oligomeric state of AlgZ has not been determined, preliminary gel filtration chromatography data suggest a molecular mass of ≈40 kDa (data not shown). This would be consistent with a tetrameric form for AlgZ in solution, which is also the case for Mnt (Vershon et al., 1987a). In EMSA, each repressor forms a single protein–DNA complex with its cognate operator (Vershon et al., 1987a,b). In contrast, AlgZ forms multiple protein–DNA complexes, which vary depending on the AlgZ concentration. These complexes could represent different oligomeric forms of AlgZ or additional molecules of AlgZ binding the ZCS. Mnt and Arc make DNA contacts that span ≈21 bp, with the key contacts forming a palindromic sequence (Vershon et al., 1987a,b). Previous copper–phenanthroline footprinting of a single highly retarded AlgZ–ZCS complex identified a 36 bp region protected by AlgZ (Baynham and Wozniak, 1996). It is possible that each oligomeric form of AlgZ binds to a 12 bp sequence of DNA or that the AlgZ binding sites are overlapping. Resolution of the specific interactions of AlgZ with the algD ZCS will be greatly facilitated by the cloning and expression of algZ undertaken in this study.

As Mnt and Arc repress by binding to operator regions and blocking the transcriptional machinery, it is interesting that algZ shares homology with these proteins, yet functions as an activator. The action of AlgZ at the algD promoter may be architectural in establishing a precise nucleoprotein structure to enable activation to occur. Alternatively, AlgZ may act in a more direct fashion by interacting with the RNA polymerase–σ22 holoenzyme, perhaps via the carboxy-terminal region. Considering the complexity of the algD promoter, it is likely that there are multiple structural proteins that enhance transcription as well as one or more direct activators. Candidates for the former category include IHF, which bends DNA, and CysB, which is hypothesized to maintain the 3′ region of the algD promoter in a supercoiled state. Consistent with this hypothesis is the observation that deletion of the ihf or cysB genes still results in detectable algD transcription (Delic-Attree et al., 1996; 1997). As deletion of either algR or algZ resulted in no detectable algD transcription, it is possible that these proteins contact RNA polymerase directly, either separately or in concert. AlgR and AlgZ bind DNA independently (data not shown), so if these proteins do act together, it is after the point of binding to algD sequences.

Previously, it was determined that AlgZ DNA-binding activity was dependent on the extracytoplasmic function sigma factor σ22, also called AlgT or AlgU (Baynham and Wozniak, 1996). Moreover, in all P. aeruginosa CF isolates studied previously, AlgZ DNA-binding activity correlated with a mucoid phenotype (Baynham and Wozniak, 1996). Interestingly, expression of σ22 in the non-mucoid P. aeruginosa strain PAO1 resulted in a mucoid phenotype and expression of active AlgZ (data not shown). This suggests that most, if not all, P. aeruginosa strains have the genetic capacity to synthesize AlgZ. As AlgZ expression or activity requires the stress response sigma factor σ22, it is likely that AlgZ is part of the global stress response in P. aeruginosa. Although AlgZ DNA-binding activity is σ22 dependent, it is still unclear whether σ22 is involved in direct or indirect control of the algZ gene itself. Examination of sequences directly 5′ of algZ did not reveal a typical σ22-regulated promoter containing a GAACTT at the −35 region and a TC—-A/C at the −10 box (Deretic et al., 1994). Thus, it is likely that AlgZ is indirectly dependent on the alternative sigma factor σ22. This regulation may be in the form of post-transcriptional control, or perhaps σ22 is responsible for the expression of a gene that then allows transcription of algZ. In either case, algZ represents another member of the σ22 regulon, and studies regarding the mechanism of algZ control by σ22 are currently under way.

The proteins that function as activators of algD transcription frequently have roles beyond this promoter. In addition to activating algD, AlgR also plays a role in the type IV fimbriae-mediated form of surface translocation called twitching motility (Whitchurch et al., 1996). The sigma factor σ22, which functions in the expression of numerous alginate genes or operons, is also involved in regulating twitching motility, stress response, heat shock and probably other unknown cellular processes (Govan and Deretic, 1996; Whitchurch et al., 1996). Another protein that binds the algD promoter, CysB, is necessary for cysteine biosynthesis and, therefore, has a major role aside from the regulation of alginate production (Delic-Attree, 1997). Integration host factor (IHF), which binds the algD and algB promoters and is necessary for full algD activation, is a global regulator in other bacterial species and is therefore probably not dedicated to alginate gene expression (Friedman, 1988). Likewise, it is possible that AlgZ may have other functions in the cell in addition to the activation of algD. Consistent with this idea is the observation of a significant growth defect of the P. aeruginosa algZ mutant FRD1220 obtained in this study (data not shown).

An examination of the DNA sequence surrounding algZ (within contig 77, 15 December 1998 release, www. revealed three open reading frames (ORFs) oriented in the opposite direction to algZ. These ORFs encode homologues of the enteric PhnC, PhnD and PhnE proteins, which are involved in alkylphosphonate metabolism (Chen et al., 1990). Positioned ≈600 bp downstream of algZ are two previously described ORFs, including rhlG, involved in rhamnolipid synthesis, and rcsF, of unknown function. The promoter for these genes was mapped to an area downstream of algZ (Campos-Garcia et al., 1998). The spacing and orientation of the genes at this locus suggests that algZ is transcribed independently and is not likely to be part of an operon. Considering that some members of this family of DNA-binding proteins identified previously have been phage proteins, one possibility was that algZ was part of a phage genome present in P. aeruginosa, perhaps within a pathogenicity island. The high G + C ratio (63%), the strong bias for G or C in the third position of AlgZ codons (West and Iglewski, 1988), as well as the presence of biosynthetic genes and not phage morphogenesis genes in this region is consistent with algZ being bacterial and not phage derived.

The mechanism by which AlgR and AlgZ function in the activation of palgD is currently not clear. There are examples of activators that act to recruit RNA polymerase at a particular promoter or by influencing a post-recruitment step, such as isomerization or promoter clearance (Roy et al., 1998). To accomplish activation by either mechanism, multiple protein–DNA and protein–protein interactions may function synergistically. One well-studied transcription factor, the E. coli cyclic AMP receptor protein (CRP), can activate over 100 different promoters and is triggered by cyclic AMP. Activation at some promoters (e.g. gal ) is dependent on binding one CRP dimer close to the promoter (−41) and another dimer further upstream, contacting RNA polymerase at the amino-terminal and carboxy-terminal domains of the α-subunit respectively (Busby and Ebright, 1994; Ptashne and Gann, 1997). In this case, co-dependence on two activators results from simultaneous contacts between separate activators and RNA polymerase. This mechanism allows for flexibility, as any activator capable of contacting the C-terminal domain of the RNA polymerase α-subunits can act co-operatively with CRP. Similar contacts may occur at the algD promoter between AlgR, AlgZ and RNA polymerase containing σ22. As such, the algD promoter may serve as an excellent model system for studying complex promoters and will be useful in determining the requirements for gene activation in other systems

Experimental procedures

Bacterial strains and plasmids

The P. aeruginosa strains used in this study originated from FRD1, an alginate-producing (mucA22 Alg+) isolate obtained from a CF patient (Ohman and Chakrabarty, 1981). Strains FRD444 (algB ::Tn501 ) and FRD810 (algR ::ΩSmr) are described elsewhere (Wozniak and Ohman, 1994). P. aeruginosa FRD875 (mucA22 algD::xylE ) and FRD879 (mucA22 algD::xylE algB ::Tn501 ), which were used for monitoring algD transcription (Table 1), have been described elsewhere (Woolwine and Wozniak, 1999). Chromosomal algD–xylE fusions were generated in strain FRD810 (algR ::ΩSmr) or FRD1220 (algZΔ ::tet ) using pDJW530, as outlined previously (Woolwine and Wozniak, 1999), resulting in FRD880 or FRD1222 respectively. E. coli JM109 was used in routine cloning and expression experiments and has been described elsewhere (Maniatis et al., 1982). pDJW221 (Wozniak and Ohman, 1994) contains a wild-type algD-cat reporter construct cloned into pALTER-1 (Promega) and was used as a template for preparation of PCR products for EMSA. pUC18 was used in the cloning and expression of algZ (Yanisch-Perron et al., 1985). pEX100T (Schweizer and Hoang, 1995) or pEX18Ap were used for P. aeruginosa gene replacement experiments. pPJ120 and pPJ123 contain the algZ gene generated by PCR of genomic DNA cloned into pUC18 in opposite orientations. pPJ126 contains algZ from the 800 bp BamHI fragment of pPJ123 filled in using the Klenow fragment and cloned into the SmaI site of pEX100T. pPJ128 is the result of a 193 bp algZ deletion by cleavage of pPJ126 with XhoI and MluI, blunted using the Klenow enzyme, and the omega fragment from pHP45 on a SmaI fragment ligated into this vector (Fellay et al., 1987). pRK2013 was used to mobilize plasmids into P. aeruginosa via triparental mating (Figurski and Helinski, 1979). The plasmid used for complementation and sequencing is pPJ130, which contains a 2 kb BamHI fragment obtained by FRD1 genomic digest and cloned into pUCP21T (Schweizer et al., 1996).

Media and chemicals

L broth (10 g l−1 tryptone, 5 g l−1 yeast extract, 5 g l−1 NaCl, pH 7.5) and L agar (as above with 15 g l−1 agar) were used for culturing E. coli strains. L broth or L agar without NaCl (LBNS/LANS) were used to culture P. aeruginosa strains. The medium used to cultivate P. aeruginosa strains after triparental mating was VB minimal salts medium (Wozniak and Ohman, 1991) or LANS plates supplemented with Irgasan (Irgasan DP300; Ciba Geigy) at 25 μg ml−1. Antibiotics were used in the concentrations described elsewhere (Wozniak and Ohman, 1991). Incubation was carried out at 37°C except for sucrose counterselection (see below), which was carried out at 30°C. Total protein was quantified by the BCA method (Pierce) or by the method of Bradford (1976).

Nucleic acid manipulations

Routine genetic manipulations were performed as described elsewhere (Woods et al., 1991; Ausubel et al., 1992). Restriction enzymes, alkaline phosphatase, T4 ligase and T4 polynucleotide kinase were purchased from Promega. Plasmid DNA was isolated from E. coli using Qiagen columns and procedures. DNA sequences from plasmid DNA were obtained from the DNA Core Laboratory of Wake Forest University Medical Center using a cycle sequencing protocol. Gel purification of PCR products was accomplished using Qiaquick columns (Qiagen). DNA was quantified using the A260 method or an ethidium bromide spot plate using pUC18 as a standard. Radiolabelled DNA fragments for EMSA were generated using [α-32P]-dCTP (300 Ci mmol−1; ICN Radiochemicals) and Pwo polymerase (Boehringer Mannheim), as described elsewhere (Roesch and Blomfield, 1998). Fragments were then purified using Wizard PCR purification columns (Promega).

Electrophoretic mobility shift assay (EMSA)

The AlgZ DNA-binding assays were performed using DNA fragments prepared as described above and the indicated protein preparation. Oligonucleotides used in PCR amplification were synthesized by the DNA Core Laboratory at Wake Forest University Medical Center. The sequences of the primers used were algD5, 5′-AAGGCGGAAATGCCATCTCC-3′ and algD7, 5′-AGGGAAGTTCCGGCCGTTTG-3′, which yield a fragment of 299 bp spanning a region 25 bp to 324 bp upstream of the start of algD transcription. Most binding reactions were performed at 25°C for 10 min in a total volume of 10 μl. The reactions contained the indicated amount of protein, 0.75 μg of poly-(dI–dC), 4.0 mM Tris-HCl (pH 8.0), 4 mM MgCl2, 4% glycerol and ≈1–50 fmol of labelled DNA. The exception was that the electroeluted fractions used in Fig. 1 were assayed in a reaction volume of 40 μl containing 31 μl of electroeluted fraction eluate and 1.35 μg of poly-(dI–dC) (Boehringer Mannheim) with an incubation time of 20 min at 25°C.

EMSA as described above was used to demonstrate the specificity of AlgZ binding. The sequences of the primers used to obtain specific competitor DNA were algD5 (above) and algD9 (5′-CTTAATCTTCGACCCATGCA-3′), which yielded a 100 bp fragment spanning from 224 bp to 324 bp upstream of the start of algD transcription. The DNA fragment used as a non-specific competitor was generated by PCR amplification of pJG194 with the oligonucleotides (algB12) 5′-GTCGATGACGAGTCGGCGAT-3′ and (algB39) 5′-TGCAACAGGGCCTCCGCCTG-3′. This reaction resulted in the amplification of a 106 bp fragment of algB, which had previously been shown not to bind AlgZ in EMSA. The fragments were gel purified using Qiaquick columns (Qiagen). After quantification of all DNAs used by the OD260 method, EMSA was performed as above with 1.5 μg per sample of the AlgZ-enriched fraction from expression in E. coli, ensuring that the labelled specific DNA fragment was added last. The molar ratio of labelled to unlabelled DNA was 1:5, 1:25 and 1:100.

EMSA was also used to quantify the amount of AlgZ activity throughout the biochemical enrichment steps. For each step in the process, 10 pmol of DNA was combined with a titration of the sample yielding multiple lanes with both greater than and less than half of the DNA bound. The dried gels were subjected to direct image analysis on an AMBIS radioanalytical imager with version 2.0 software. Total counts from both free and AlgZ-bound DNA fragments were determined and used to calculate the fraction of bound and free DNA. Linear regression was then used to determine at what volume of sample half of the DNA was bound. One unit of AlgZ was arbitrarily determined to be the amount of protein required to bind half of the DNA.

PAGE, preparation of P. aeruginosa extracts and gel electroelution

All SDS–PAGE reagents were prepared according to the method of Maniatis et al. (1982). Tris–tricine gels with 16% acrylamide and 2 × cross-linker were prepared according to the method of Judd (1998). Gels were typically electrophoresed for 15–18 h at 8–15 mA. Partially enriched extracts of P. aeruginosa were prepared by growth of 1.5 l of FRD810 in LBNS at 37°C with shaking to an OD600 of 0.7. Cells were recovered by centrifugation (12 000 × g for 10 min), and the pellet was resuspended in 20 ml of fractionation buffer (10 mM Tris-HCl, pH 8.0, 100 mM NaCl and 1 mM MgCl2) supplemented with 0.1 mM phenylmethylsulphonyl fluoride (PMSF) and 0.1 mM dithiothreitol (DTT). Cells were disrupted using a French pressure cell with one passage at 10 000 psi (25°C), and whole cells and cellular debris were removed by centrifugation at 25 000 × g for 15 min. The resulting extract was then clarified by ultracentrifugation (Beckman L8-M ultracentrifuge) using a SW-28 rotor at 19 000 × g for 1 h at 4°C. Solid ammonium sulphate was added to a concentration of 45% and allowed to stir overnight. The precipitate was recovered by centrifugation at 12 000 × g for 15 min, resuspended in 7 ml of buffer T (10 mM Tris-HCl, pH 8.0, 1 mM EDTA, 0.1 mM PMSF, 0.1 mM DTT, 7% glycerol) and dialysed exhaustively against buffer T. Approximately 5 mg of this preparation was separated on a 12% SDS–polyacrylamide gel (1.5 mm × 16 cm × 18 cm) under reducing conditions without prior heating of the samples. After electrophoresis, the gel was equilibrated in electroelution buffer (50 mM Tris-HCl, pH 8.0) for 40–45 min with three changes of buffer as described previously (Anderson and Heron, 1993). Electroelution was performed at 200 mA for 30 min with a 15 s reversal of current using a Bio-Rad whole gel eluter. A 31 μl aliquot of each eluate fraction was tested for AlgZ DNA-binding activity by EMSA and analysed for protein content by resolution of 30 μl on a 15% SDS–PAGE gel and detection by silver staining (Bio-Rad silver staining kit).

Biochemical enrichment for AlgZ

For partial purification of AlgZ, a 15 l culture of P. aeruginosa FRD810 was grown and prepared as above. After disruption and ultracentrifugation, solid ammonium sulphate was added to the 304 ml supernatant (45% final concentration), and the preparation was allowed to stir overnight at 4°C. The precipitate was recovered by centrifugation as above, resuspended in 20 ml of buffer T and dialysed exhaustively against buffer T. This preparation was then dialysed against buffer S (50 mM sodium acetate, pH 5.0, 0.5 mM EDTA, 0.1 mM PMSF, 0.1 mM DTT, 7% glycerol), the precipitate recovered by centrifugation (3000 × g, 20 min) and resuspended in 20 ml of buffer T. To renature the proteins, the preparation was dialysed exhaustively against buffer T containing 0.2 M NaCl. The dialysed sample was applied to a heparin agarose column (Sigma Chemical) with a 15 ml bed volume at a flow rate of 0.5 ml min−1 in buffer T containing 0.2 M NaCl. After washing with five column-volumes of the identical buffer, proteins were eluted with a linear gradient of 0.2–1.5 M NaCl in buffer T, and 2 ml fractions were collected. These fractions were assayed for activity by EMSA and for protein content on Tris–tricine gels as described above.

Amino-terminal sequence analysis

The active fractions from the heparin eluates above were pooled, and 210 μg of protein was separated on a Tris–tricine gel prepared as above. After electrophoresis, proteins were transferred to an Immobilon−sq membrane (Millipore) using a Bio-Rad Trans-blot SD semi-dry transfer unit at 15 V for 1 h with reagents as described previously (Ma et al., 1998). The membrane was stained with Coomassie brilliant blue solution (Maniatis et al., 1982) for 1 min, destained briefly in 50% methanol and then placed in distilled water. The putative AlgZ band (Fig. 2) was excised and submitted to the Protein Core Laboratory of the Wake Forest University Medical Center for amino-terminal sequence determination. The resulting amino-terminal sequence was used in a tblastn search of the unfinished microbial genome database (www.ncbi.nlm.nih. gov/blast/unfinishedgenome.html).

Cloning, expression and DNA sequencing of algZ

Initially, algZ was cloned by amplification of FRD1 genomic DNA with Pwo polymerase (as above) and oligonucleotide primers designed using the P. aeruginosa genome database (, 12/15/98 release). The primers algZ7 (5′-CTGACTCTACAGGTTCAATG-3′) and algZ8 (5′-ATAGGCTTGTTCGCCCATGG-3′) were designed to amplify a 1442 bp fragment 20 bp 5′ and 1095 bp 3′ of the predicted algZ open reading frame. The PCR product was gel purified and ligated into pUC18, which had been cleaved with SmaI and alkaline phosphatase treated. Plasmids were isolated containing algZ in both orientations. One plasmid, pPJ120, which contained algZ oriented in the same direction as plac, was used for expression. Another plasmid, pPJ123, which had algZ oriented opposite to that of pPJ120, was used in the construction of the P. aeruginosa algZ mutant below. For expression, E. coli harbouring pPJ120 or the pUC18 vector were induced as follows: 1 ml of an overnight culture was inoculated into 100 ml of LB supplemented with ampicillin. The bacteria were grown at 37°C with shaking to an OD580 of 0.2 and then induced with a final concentration of 1 mM IPTG for 3 h. Cells were harvested by centrifugation (12 000 × g for 15 min) and resuspended in 10 ml of fractionation buffer. The cells were disrupted with a French pressure cell, and cellular debris was removed by centrifugation and ultracentrifugation as above. The extracts were further enriched for AlgZ activity by 45% ammonium sulphate precipitation as above. The resulting extracts were assayed for AlgZ DNA-binding activity by EMSA. Southern blot analysis (Southern, 1975) was used to localize algZ to a 2 kb BamHI fragment. This sized fragment was isolated from FRD1 genomic DNA by restriction digest and gel purification (Qiaquick column; Qiagen) and partially filled in with G and A. The vector, pUCP21T (Schweizer et al., 1996), which had been cleaved with Sal I and partially filled in with dTTP and dCTP was then ligated with the 2 kb BamHI fragments. After transformation into E. coli JM109, the presence of the insert containing algZ was monitored by PCR using oligonucleotides algZ7 and algZ8. This plasmid, designated pPJ130, was used in the complementation analysis and was the source of sequence data regarding algZ. Sequencing was performed by the DNA Core Laboratory of Wake Forest University. The blastP homology searches were performed using the National Center for Biotechnology Information web site ( Alignments described in the text and in Fig. 5 were performed using sim (Huang and Miller, 1991) on the ExPASy web site (

Construction of an algZΔ::tet P. aeruginosa strain

To generate an algZ mutant strain of P. aeruginosa, pPJ123 (above) was cleaved with BamHI, and the 800 bp fragment containing algZ was gel purified, blunted and ligated into pEX100T, which had been cleaved with SmaI and treated with alkaline phosphatase. The resulting plasmid, pPJ126, was cleaved with XhoI and MluI, thereby removing 179 bp of algZ, and treated with Klenow polymerase and alkaline phosphatase. The omega tet cassette encoding tetracycline resistance was recovered as a SmaI fragment from pHP45ΩTc and ligated into pPJ126 treated as above. The resulting plasmid, pPJ128, was transferred by conjugation into the mucoid strain FRD1 (mucA22 ) or the isogenic strain FRD875 (mucA22 algD::xylE ). Plasmid integration was selected using VB minimal salts media containing tetracycline. Sucrose-mediated counterselection of the plasmid was accomplished as described elsewhere (Woolwine and Wozniak, 1999). The resulting colonies were passed on LANS plates and scored for carbenicillin sensitivity to verify that the plasmid backbone had been excised.

Assay of algD::xylE transcriptional fusions

Overnight cultures of each of the algD::xylE strains indicated in Table 1 as well as their isogenic wild-type algD counterparts were grown overnight in LBNS at 37°C, diluted 1:500 in fresh prewarmed LBNS and cultured at 37°C with agitation until reaching an A540 of 0.4–0.5. A 1.0 ml sample of each culture was pelleted at 20 800 × g for 4 min and resuspended in 1.0 ml of assay buffer (50 mM potassium phosphate, pH 7.5, 10% acetone). Samples (100–500 μl) were assayed for XylE activity in 1.0 ml of assay buffer containing 1 mM catechol (Sigma). The corresponding isogenic wild-type algD strain was prepared in an identical fashion and served as a blank. After the addition of the catechol, the A375 was monitored at ambient temperature for 2 min, and the rate of formation of the reaction product (2-hydroxymuconic semialdehyde) was calculated from the molar extinction coefficient of the product (4.4 × 104 M−1). Values were normalized by dividing through with the A540 of the culture at the time of harvest. Data depicted in Table 1 were derived from three independent experiments and are expressed as the percentage of parental (FRD875) algD-xylE levels.

Nucleotide sequence accession numbers

The nucleotide sequence of algZ has been deposited in the GenBank database under accession number AF139988.


Many thanks to Dr Peter T. Chivers for helpful suggestions on this manuscript. This work was supported by Public Health Service grant 5R01 HL58334 (D.J.W.) from the National Heart, Lung, and Blood Institute. P.J.B. is a predoctoral student supported by NIH training grant AI07401. Oligonucleotides were provided by the DNA Synthesis Core Laboratory of the Cancer Center of Wake Forest University. DNA sequencing was performed by E. Jung of the DNA Sequencing Core Laboratory, and protein sequencing was accomplished by M. Morris of the Protein Core Laboratory. Both facilities are supported in part by NIH grant CA-12197.