ABSTRACT
HetR is the master regulator of heterocyst differentiation in the filamentous cyanobacterium Anabaena sp. strain PCC 7120. Genetic selection was used to identify 33 amino acid substitutions in HetR that reduced the proportion of cells undergoing heterocyst differentiation to less than 2%. Conservative substitutions in the wild-type HetR protein revealed three mutations that dramatically reduced the amount of heterocyst differentiation when the mutant allele was present in place of the wild-type allele on a replicating plasmid in a mutant lacking hetR on the chromosome. An H69Y substitution resulted in heterocyst formation among less than 0.1% of cells, and D17E and G36A substitutions resulted in a Het− phenotype, compared to heterocyst formation among approximately 25% of cells with the wild-type hetR under the same conditions. The D17E substitution prevented DNA binding activity exhibited by wild-type HetR in mobility shift assays, whereas G36A and H69Y substitutions had no affect on DNA binding. D17E, G36A, and H69Y substitutions also resulted in higher levels of the corresponding HetR protein than of the wild-type protein when each was expressed from an inducible promoter in a hetR deletion strain, suggesting an effect on HetR protein turnover. Surprisingly, C48A and S152A substitutions, which were previously reported to result in a Het− phenotype, were found to have no effect on heterocyst differentiation or patterning when the corresponding mutations were introduced into an otherwise wild-type genetic background in Anabaena sp. strain PCC 7120. The clustering of mutations that satisfied the positive selection near the amino terminus suggests an important role for this part of the protein in HetR function.
Anabaena sp. strain PCC 7120 is a filamentous cyanobacterium capable of fixing nitrogen in differentiated cells known as heterocysts, where the oxygen-labile nitrogenase complex is sequestered from atmospheric oxygen and processes that evolve O2. When deprived of fixed nitrogen, heterocysts develop at regular intervals, approximately every 10th cell, along filaments (18). Along with spore formation by Bacillus subtilis and fruiting-body formation by Myxococcus xanthus, among others, heterocyst formation by Anabaena sp. has emerged as a paradigm system for the regulation of cellular differentiation and pattern formation in bacteria. HetR is the master regulator of heterocyst formation. The genetic network regulating differentiation consists of a series of interconnected positive and negative feedback loops intersecting at hetR (14, 19). Understanding the intricacies of HetR function is essential to understanding differentiation and patterning in heterocystous cyanobacteria.
The deletion of hetR results in an inability to form heterocysts, while overexpression promotes excessive differentiation even in media with a source of fixed nitrogen (3, 4). The induction of hetR transcription occurs early in the developmental process and is localized to clusters of cells around developing heterocysts, eventually resolving to those cells that will differentiate. hetR is positively autoregulated and may be posttranslationally modified upon nitrogen step-down (1, 20). HetR has been reported to be a novel, Ca2+-dependent serine protease capable of degrading itself and CcbP, a calcium binding protein, and substitutions at residues S152 and S179 have been reported to abolish protease activity and heterocyst differentiation (9, 16, 21). S152 has been reported to be the active site. HetR binds to DNA of its own promoter region as well as to regions in the patS and hepA promoters. DNA binding by HetR as a homodimer has been reported to require a disulfide bond between monomers at C48, and a C48A substitution in HetR has been reported to prevent dimer formation and DNA binding in vitro and to result in a Het− phenotype in vivo (11). Two inhibitors of heterocyst formation, patS and hetN, have been shown to work at the level of HetR. An R223W substitution mutant protein results in increased differentiation when cells are deprived of combined nitrogen and is insensitive to patS and hetN, suggesting that both the patS- and hetN-dependent pathways converge at HetR (12). In vitro, the PatS pentapeptide (RGSGR) was shown to inhibit the DNA binding capability of HetR (11). Although much has been discerned about HetR, its exact function in promoting cellular differentiation as well as the nature of any posttranslational modifications has yet to be elucidated.
A genetic approach was used to isolate hetR mutant strains with altered levels of heterocyst differentiation in order to further characterize HetR. Here we report the results of two screens used to identify amino acid substitutions resulting in altered HetR function. Residues D17, G36, and H69 were found to be essential for normal heterocyst differentiation and levels of HetR protein turnover. Residue D17 was required for DNA binding activity. Contrary to previous reports, the S152 and C48 residues of HetR were found not to be essential for heterocyst formation.
MATERIALS AND METHODS
Bacterial strains and growth conditions.The growth of Escherichia coli and Anabaena sp. strain PCC 7120 and its derivatives; concentrations of antibiotics; the induction of heterocyst formation in BG-110 medium, which lacks a combined nitrogen source; the regulation of Pnir expression; and conditions for photomicroscopy were as previously described with the exception that medium containing 10 mM NH4 was supplemented with 30 mM TES [N-tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid], pH 8.0, and E. coli cultures used for introducing the PCR mutagenesis library into PCC 7120 were diluted 1,000-fold to ensure that the selected exconjugant filaments arose from the conjugal transfer of a single plasmid (2, 8). To determine heterocyst percentages, 200 cells were counted and the number of heterocysts was recorded. All results are expressed as averages of results for three replicates, with error bars indicating 1 standard deviation, with the exception of data represented in Table 1, which are the percentages of heterocysts among 500 counted cells.
Mutational analysis of hetR
Plasmid construction.To create a library of plasmids containing mutations within the hetR coding frame, a 1,889-bp fragment containing the hetR coding region and 966 bp of upstream DNA was amplified via error-prone PCR (5) with 2 mM Mn2+, 10 mM dTTP and dCTP, 2 mM dGTP and dATP, and 5% dimethyl sulfoxide (using primers hetR-F-BamHI [5′-ATCCCGGATCCCCTGCCAATGCAGAAGGTTAAAC-3′] and hetR-R-SacI [5′-CATTAGAGCTCCTTTTATTCACTCTGGGTGC-3′]) and cloned directly into pAM504 (17) as BamHI-SacI fragments.
Plasmids pDR137, pDR138, pDR144, and pDR145 were used to express hetR carrying the mutation corresponding to S152A [hetR(S152A)], hetR, hetR(C48A), and hetR(S179N), respectively, from the native hetR promoter. A 966-bp fragment upstream of the hetR coding region was amplified via PCR (using primers PhetR-KpnI-F [5′-GGTACCCCTGCCAATGCAGAAGGTTAAAC-3′] and PhetR-NdeI-R [5′-CATATGACAAATAGTTGAATAGCACGCTTATTAG-3′]) to give PhetR and cloned into pGEM-T (Promega). PhetR was subsequently cloned into pBluescript SK(+) (Stratagene) as an ApaI-PstI fragment to create pDR133. hetR and hetR(S179N) were amplified from genomic DNA of strain PCC 7120 and strain 216 (3), respectively (using primers hetRcf-NdeI [5′-CATATGAGTAACGACATCGATCTGATC-3′] and hetR6H-r [5′-TTAGTGATGGTGATGGTGATGATCTTCTTTTCTACCAAACACCATTTG-3′]). hetR(S152A) and hetR(C48A) were created using overlap extension PCR (10) using primers hetR152A-F (5′-GTTAGTTACTGCCTTTGAGTTTTTGGAGTTGATC-3′) and hetR152A-R (5′-CAAAAACTCAAAGGCAGTAACTAACTGAGCGTCGGG-3′) and hetR48-F (5′-GCGGCTAAGGCCGCCATTTACATGACTTATCTAGAGC-3′) and hetR48-R (5′-GTAAATGGCGGCCTTAGCCGCCGTTGCTG-3′), respectively, with hetRcf-NdeI and hetR6H-r. The subsequent fragments containing alleles of hetR were cloned into pGEM-T and then moved to pDR133 as an NdeI-PstI fragment using NdeI sites introduced by the primer. The subsequent PhetR and hetR fragments were cloned into pAM504 as KpnI-SacI fragments.
Plasmid pDR147 was used to replace the wild-type allele of hetR in strain PCC 7120 with hetR(S152A). hetR(S152A) in pGEM-T was moved into pBluescript as a KspI-SacI fragment and subsequently cloned into pRL278 (1) as a BamHI-SacI fragment to create pDR147.
Plasmid pDR167 was used to replace the wild-type allele of hetR in strain PCC 7120 with hetR(C48A). hetR(C48A) fused to PhetR [PhetR-hetR(C48A)] was excised from pBluescript SK(+) and reintroduced into pBluescript SK(+) as a KspI fragment in order to create flanking SpeI sites. PhetR-hetR(C48A) was then moved into pRL277 (1) as a SpeI fragment to create pDR167.
Plasmid pDR203 is a shuttle vector used to express hetR(E254G) from its native promoter. hetR(E254G) was created via overlap extension PCR (using internal primers hetR254E-G F [5′-CGAGCCTTAGAAGGACTCGATGTGCCACCAGAG-3′], hetR254E-G R [5′-GCACATCGAGTCCTTCTAAGGCTCGCATAGCGTTTG-3′], hetR-F-BamHI, and hetR-R-SacI) and cloned into pAM504 as a BamHI-SacI fragment using the restriction sites introduced on the primers to create pDR203.
To create a library containing mutations within pDR203, this plasmid was introduced into the hypermutator E. coli strain XL1-Red (Stratagene). Plasmid was isolated from the resultant colonies and introduced into a ΔhetR strain, UHM103 (2), via conjugation.
To create conservative substitutions within the hetR coding frame, overlap extension PCR was used with the overlapping inner primers described in Table 1 and the outer primers hetR-BamHI-F and hetR-SacI-R, and the PCR fragments containing mutant coding regions were cloned into pAM504 as BamHI-SacI fragments, with the exception of hetR(G295A), which was amplified with hetR-BamHI-F and hetRG295A-R (Table 1).
For the overexpression of recombinant HetR carrying the D17E substitution (HetRD17E), HetRG36A, and HetRH69Y from E. coli, the appropriate hetR allele was amplified via PCR using the primers hetRcf-NdeI and hetRH6-BamHI-R, cloned into the SmaI site of pBluescript SK(+) as a blunt-end fragment, and then moved into pET21a as an NdeI-BamHI fragment. For the overexpression of recombinant HetR, HetRC48A, HetRS152A, and HetRS179N, the appropriate hetR allele was moved from pDR137, pDR138, pDR144, or pDR145 to pET21a (Novagen) as an NdeI-NotI fragment.
Plasmid pSMC188 was created to transcriptionally fuse hetR to the Pnir promoter. The EcoRI site of plasmid pAM504 was removed by cutting with EcoRI and blunt-ending with T4 polymerase followed by ligation. A fragment of DNA containing Pnir was generated by PCR as previously described (8) and cloned as a BamHI-SacI fragment into the resulting vector to create pSMC188. hetR mutant alleles were then amplified from appropriate template DNA (using the primers hetR-EcoRI-F [5′-CTATTGAATTCATGAGTAACGACATCGATCTGATC-3′] and hetRH6-BamHI-R [5′-TTACTGGATCCTTAGTGATGGTGATGGTGATGATCTTCTTTTCTACCAAACACCATTTG-3′] for carboxy-terminal six-His-tagged constructs or hetR-EcoRI-F and hetR-BamHI-R [5′-TTACTGGATCCTTAATCTTCTTTTCTACCAAACACCATTTG-3′]) and cloned into pSMC188 as EcoRI-BamHI fragments. All plasmid-borne constructs described herein were sequenced to verify integrity.
Strain construction.All strains and oligonucleotides not listed in the text are described in Table 1. The replacement of hetR with hetR(C48A) and hetR(S152A) was performed as previously described (2) with the following exceptions. Double recombinants were screened for the presence of mutant alleles of hetR via PCR using a pair of primers whose final one or two 3′- end nucleotides corresponded to those of either the wild-type or the mutant allele of hetR [hetR152wt-F (5′-GACGCTCAGTTAGTTACTT-3′) and hetR152A-F (5′-GACGCTCAGTTAGTTACTG-3′) for hetR(S152A) or hetR48wt-F (5′-AGCAACGGCGGCTAAGTG-3′) and hetR48A-F (5′-AGCAACGGCGGCTAAGGC-3′) for hetR(C48A)] along with hetR6H-r. The hetR region from isolates that tested positive for the mutant allele was amplified via PCR with two sets of primers, each containing at least one primer whose sequence was outside the region on the plasmid used to construct the corresponding strain (hetR-5′ F [5′-TTTAGATCTGTCGTTCTCAGCCACAGAGATTTGTCC-3′] and hetR6H-r or hetRcf-NdeI and hetR3′ R [5′-TTTCTGCAGATGTCTTGGCTCAGTCGCGGATGATGG-3′]), and the PCR products were sequenced to ensure the presence of only the intended mutation. The hetR region from these isolates was then amplified with primers corresponding to sequences inside the hetR coding frame (hetRD17E-F and hetRE254D-R [5′-GCACATCGAGATCTTCTAAGGCTCGCATAGCG-3′]) and sequenced to ensure that another copy of hetR was not present in the chromosome. Finally, Southern blot analysis was performed on chromosomal DNA from strains PCC 7120, R48, R152, and UHM103 (a ΔhetR strain), and the DNA was digested with SpeI or HindII and probed with the entire hetR coding region.
Western blot analysis.One hundred-milliliter cultures of the appropriate strain were concentrated and resuspended in 1 ml of lysis buffer (10 mM Tris, 300 mM NaCl, 1 mM phenylmethylsulfonyl fluoride, pH 7.5, containing 1 mg/ml lysozyme) at 4°C and lysed via sonication. The lysate was cleared by centrifugation at 10,000 × g for 30 min, and the protein concentration was determined using the DC protein assay (Bio-Rad Laboratories). Lysate containing 50 μg of protein was then subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and electrophoretic transfer onto polyvinylidene difluoride membrane. HetR was detected with Penta-His antibodies (QIAGEN) and chemiluminescent detection (Western Breeze; Invitrogen). Coomassie staining of the membrane was performed using Simply Blue safe stain (Invitrogen). Quantification of Western blot signals was done using Scion Image (Scion Corp.).
Overexpression and purification of recombinant HetR.Recombinant HetR from E. coli BL21(DE3) transformed with the appropriate pET21a derivative was obtained as previously described (20) with the exceptions that β-mercaptoethanol was used in place of dithiothreitol (DTT) and inclusion bodies were refolded in a solution containing 20 mM Tris base, 300 mM NaCl, 0.1 mM EGTA, and 20% glycerol, pH 8.0, at 4°C (TSEG), incubated several days at 4°C, and then purified using Ni-nitrilotriacetic acid affinity chromatography (QIAGEN). The column was washed three times with 1 column volume of TSEG containing 20 mM imidazole, and recombinant proteins were eluted in TSEG containing 250 mM imidazole.
DNA binding assays.Electrophoretic mobility shift assays were performed using the Light Shift chemiluminescent electrophoretic mobility shift assay kit (Pierce) and NuPAGE 6% DNA retardation gels (Invitrogen). The FhetR-1 region of the hetR promoter was amplified via PCR as previously described with the primers FhetR-1-F (biotin-5′-CAGATAAGTTCCGGATAATAGGG-3′) and FhetR-1-R (5-5TATAACTAGCAACAGTGTTGTTA-3′) (11). Ten-microliter binding reaction mixtures containing 20 fmol FhetR-1 or Epstein-Barr virus biotinylated DNA, 15% glycerol, 0.5 μg poly(dI-dC), 1× binding buffer (Pierce), and recombinant HetR or Epstein-Barr virus nuclear extract as indicated in Fig. 5 were incubated for 20 min at room temperature prior to electrophoresis.
RESULTS
S152 and C48 of HetR are not required for heterocyst differentiation.To search for alleles of hetR resulting in a Het− phenotype, error-prone PCR was used to generate a library of the hetR promoter and coding regions containing random point mutations. This library was cloned into a shuttle vector and introduced into a hetR deletion strain, UΗΜ103 (2). Exconjugants were selected and cultured in BG-11, then transferred into BG-110. Colonies unable to grow without combined nitrogen turned brown and were rescued to BG-11. One hundred five Het− isolates were identified. As controls, plasmids containing wild-type hetR (pDR138), hetR(C48A) (pDR144), hetR(S152A) (pDR137), and hetR(S179N) (pDR145) were also introduced into the ΔhetR strain UHM103. hetR(C48A), hetR(S152A), and hetR(S179N) have previously been reported to be unable to cause differentiation when introduced on a plasmid or into the chromosome of a hetR mutant strain (3, 9, 11). Surprisingly, two of the alleles used as negative controls for complementation did complement the mutant. hetR(S152A) and hetR(C48A) on a plasmid complemented UHM103 in BG-110, resulting in heterocyst frequencies comparable to that conferred by wild-type hetR. Only hetR(S179N) was found not to complement the ΔhetR strain. To determine if complementation of UHM103 by hetR(C48A) and hetR(S152A) was a result of extra copies of the genes on a plasmid and represented a limitation of our screen, hetR(C48A) and hetR(S152A) unmarked substitutions were created in the chromosome of the wild-type strain, PCC 7120. The resultant strains, R48 and R152, had double and single base pair substitutions in the chromosome, respectively, that replaced the cysteine or serine with an alanine residue in HetR. Each strain had a phenotype similar to that of the wild-type strain, PCC 7120 (Fig. 1A to C), with heterocyst frequencies of 8.6% (R48), 8.6% (R152), and 9.5% (PCC 7120). Strain 216 (3), which carries hetR(S179N) in the chromosome, was incapable of heterocyst formation (Fig. 1D). The reason for the discrepancy between our results and those of previous studies (9, 11) is unclear, but it appears that neither cysteine-48 nor serine-152 is essential for HetR function.
Phenotypes of strains carrying amino acid substitutions in HetR 48 h postinduction. Images correspond to Anabaena sp. strains PCC 7120 (A), R48 (B), R152 (C), and 216 (D). Carets indicate heterocysts.
Isolation of hetR(E254G) and positive selection for Het− alleles of hetR. Because the mutation rate generated by the error-prone PCR method was higher than expected (approximately 5/kb), alleles containing single mutations were subcloned or recreated using overlap extension PCR to identify those responsible for the Het− phenotype. One recreated allele of hetR with a glutamate-to-glycine substitution at residue 254 could not be introduced into the ΔhetR strain presumably because cells that receive the allele differentiate into heterocysts, which do not divide and so are not recovered. This mutation was initially isolated with another mutation resulting in an A41V substitution that was found to reduce heterocyst formation and was likely responsible for our ability to originally isolate the E254G mutation. A more detailed analysis of the hetR(E254G) mutant will be published elsewhere.
Because hetR(E254G) resulted in essentially lethal differentiation when introduced on a plasmid even in BG-11, it could be used in positive selection for mutations in hetR resulting in reduced HetR function. The plasmid containing PhetR-hetR(E254G) (pDR203) was introduced into XL1-RED, a hypermutator strain of E. coli, in order to introduce random mutations into the hetR promoter and coding frame. Plasmid was isolated from this strain and introduced into the ΔhetR strain UHM103. Neor exconjugants that arose on BG-11 medium represented isolates that had reduced or no heterocyst formation due to a mutation within pDR203 or a second site mutation within the genome. To confirm that the mutation responsible for the observed phenotype was carried on pDR203, plasmid was isolated from these strains and reintroduced into UHM103. Colonies of all mutants turned brown when transferred to BG-110 and showed no discernible growth when streaked onto solid BG-110. Sequencing revealed 33 unique amino acid substitutions that conferred a Het− or reduced heterocyst phenotype compared to pDR203 (Table 1). Of the 33 substitutions, 19 reduced the percentage of heterocysts to below 2% of cells (Table 1) and 14 resulted in no heterocyst formation (Table 1). Over 30% of the mutations occurred in one predicted α-helix near the N terminus of HetR, located at residues H33 to I50, which comprises only 6% of the protein (7) (Fig. 2). All of the sequenced mutant alleles of hetR differed from the wild-type allele by the substitution, deletion, or insertion of a single base pair.
Schematic diagram of HetR depicting mutation sites and secondary structure predictions. Mutations (thin, vertical black lines), α-helices (thick, horizontal black lines), and random coils and extended strands (gray lines) are represented.
Conservative substitution of amino acids identified by the hetR(E254G) screen.Although 33 unique amino acid substitutions were identified that alter the function of HetRE254G, it was unclear whether these were due to the replacement of a residue required for a specific function of HetR, such as DNA binding, or were simply the result of disrupted protein structure induced by a nonconservative substitution. To determine which residues were truly essential to HetR function, conservative substitutions were created in an otherwise wild-type copy of hetR at each of the 33 residues identified in the previous screen and introduced on a plasmid into the ΔhetR strain UHM103 (Table 1). While most of the conservative substitutions resulted in a lower frequency of heterocysts compared to wild-type hetR, two substitutions, D17E and G36A, resulted in alleles of hetR completely incapable of promoting heterocyst differentiation, and an H69Y substitution resulted in the differentiation of less than 0.1% of cells. Because glutamate has been reported to mimic the phosphorylated state of aspartate residues in some proteins (13), a D17N substitution was also tested to see if it would increase the rate of differentiation but was found to be incapable of promoting differentiation as well.
Dominant-recessive testing of Het− alleles of hetR.Extra copies of hetR(S179N) on a plasmid have been reported to suppress heterocyst formation in a wild-type background (11). To determine if hetR(D17E), hetR(G36A), or hetR(H69Y) has a similar effect, plasmids containing the hetR promoter and the respective alleles were introduced into PCC 7120 along with 17 nonsense and frameshift mutant alleles isolated from the XL1-Red library to test for dominance. PCC 7120 containing hetR(H69Y) on a plasmid had levels of differentiation similar to those in the negative control, PCC 7120 containing a plasmid bearing a fusion of PhetR to the gene encoding the green fluorescent protein (PhetR-gfp), under all conditions, and so hetR(H69Y) appears to have no effect on the level of differentiation (data not shown). Twenty-four hours postinduction with BG-110, PCC 7120 containing hetR(S179N), hetR(G36A), and hetR(883+T) [hetR(E254G) with a thymine insertion at nucleotide +883 with respect to the start codon, resulting in a G295W substitution, followed by the introduction of a stop codon] had reduced heterocyst frequency (<0.5%) compared to PCC 7120 containing a plasmid bearing PhetR-gfp (pSMC127) (8.3%) (6) (Fig. 3A). By 120 h postinduction, the strain with hetR(S179N) had reached a heterocyst frequency (8.67%) similar to that in the PhetR-gfp control (8.83%) while strains containing hetR(G36A) and hetR(883+T) continued to show reduced heterocyst frequencies (3.83% and 2%, respectively). Surprisingly, compared to PCC 7120 containing PhetR-gfp (Fig. 3A and B), PCC 7120 containing het(RD17E) (Fig. 3A and C) exhibited excessive differentiation and perturbed heterocyst patterning, including both multiple contiguous heterocysts and irregularly long vegetative cell intervals in some parts of filaments. Excessive heterocyst differentiation caused by extra copies of hetR(D17E) in PCC 7120 was not as severe as the effects of extra copies of wild-type hetR (Fig. 3A).
Heterocyst frequency versus time postinduction and the phenotype of PCC 7120 with extra copies of hetR(D17E). (A) Heterocyst frequency over time for PCC 7120 carrying extra copies of hetR alleles on a plasmid. PhetR-gfp, diamonds; hetR, rectangles; hetR(S179N), x's; hetR(833+T), plus signs; hetR(G36A), triangles; hetR(D17E), squares. (B) PCC 7120 carrying PhetR-gfp 48 h postinduction. Carets indicate heterocysts. (C) PCC 7120 carrying extra copies of hetR(D17E) on a plasmid 48 h postinduction.
HetR turnover is affected by D17E, G36A, and H69Y substitutions.HetR has been reported to be a serine protease capable of autodegradation, and a strain with an S179N substitution, which prevents protease activity and differentiation, has been shown to have higher levels of HetR protein than the wild-type strain (9, 21). To determine if hetR(D17E), hetR(G36A), or hetR(H69Y) also results in higher HetR protein levels, these alleles were ectopically overexpressed in PCC 7120 from the Pnir promoter to bypass any effects the mutations may have on transcription, and the level of HetR was assessed by Western blot analysis. Plasmids containing hetR, hetR(S179N), hetR(S152A), hetR(C48A), hetR(D17E), hetR(G36A), and hetR(H69Y) under the control of Pnir were introduced into the ΔhetR strain, UHM103. When transcription from Pnir was induced by the removal of ammonium and the addition of nitrate, the level of HetR in UHM103 containing hetR(D17E), hetR(G36A), hetR(H69Y), and hetR(S179N) was substantially higher than that in the same strain containing wild-type hetR, hetR(C48A), or hetR(S152A) (Fig. 4A and B). This suggests that HetRD17E, HetRG36A, and HetRH69Y have reduced turnover rates like HetRS179N. The fact that UHM103 containing hetR(S152A) had levels of HetR protein similar to those of UHM103 containing wild-type HetR suggests that the S152A substitution does not affect protein turnover. When hetR and hetR(S179N) were ectopically overexpressed from the PpetE promoter in BG-110, similar results were obtained, indicating that the presence of nitrate did not affect the assay (data not shown). hetR, hetR(C48A), and hetR(S152A) were capable of inducing heterocyst formation when overexpressed, while hetR(D17E), hetR(G36A), hetR(H69Y), and hetR(S179N) were not.
Characterization of HetR levels corresponding to different alleles. (A) Mean pixel densities in Western blot analysis of various hetR alleles in UHM103. (B) Representative results from Western blot analysis of HetR from different alleles in UHM103. (C) Coomassie-stained section of the membrane showing loading with equal amounts of protein. Lanes 1, 2, and 4 are controls containing recombinant HetR from E. coli with the addition of six histidine residues to the C terminus, protein from strain UHM103 carrying pSMC188, and wild-type HetR from UHM103 without the six-His tag, respectively. Remaining lanes contain variations of HetR from UHM103 with a six-His tag at the carboxy terminus: HetR (lane 3), HetRD17E (lane 5), HetRG36A (lane 6), HetRC48A (lane 7), HetRH69Y (lane 8), HetRS152A (lane 9), and HetRS179N (lane 10).
D17 but not C48 is necessary for DNA binding activity of HetR.It was previously reported that HetR binds to DNA from its own promoter region as well as that of the hepA and patS promoter regions and that under reducing conditions DNA binding and dimer formation by HetR are prevented (11). A C48A substitution and PatS-5 have also been reported to prevent DNA binding (11). To determine if any of the mutant alleles for HetR isolated in this study were deficient in DNA binding activity, recombinant HetR was overexpressed, purified from E. coli, and tested for DNA binding activity. Control mobility shift assays showed that wild-type HetR was capable of binding specifically to the previously defined FhetR-1 region of DNA and was inhibited by a fivefold excess of PatS-5, consistent with previous findings (11), and that neither PatS-5 nor DTT affected the DNA binding in a control reaction containing Epstein-Barr virus DNA and Epstein-Barr virus nuclear extract (Fig. 5A and B). HetRD17E had decreased binding activity, shifting only a small portion of DNA at a protein concentration sufficient to completely shift all the biotinylated FhetR-1 DNA for HetR, HetRG36A, HetRH69Y, HetRS179N, and surprisingly, HetRC48A (Fig. 5C). At lower protein concentrations, no shift was detected only for HetRD17E (data not shown). Contrary to previous findings (11), 50 mM DTT did not abolish DNA binding activity. Dimer formation could be detected for wild-type HetR, HetRD17E, HetRG36A, and HetRH69Y but not for HetRC48A (Fig. 5D and data not shown).
Results from mobility shift assays of recombinant HetR and SDS-PAGE analysis. Biotinylated Epstein-Barr virus DNA (EB Ext, Epstein-Barr virus nuclear extract) (A) or FhetR-1 (B) was incubated with additional factors as indicated prior to electrophoresis, demonstrating the specificity of HetR DNA binding and PatS-5 inhibition (HetR, PatS-5, and DTT were present at concentrations of 75 μM, 375 μM, and 50 mM, respectively). (C) FhetR-1 was incubated with the indicated variants of recombinant HetR at a concentration of 300 μM. (D) Result from nonreducing SDS-PAGE analysis of recombinant HetR, HetRD17E, and HetRC48A. Positions of the monomer and dimer forms of HetR are indicated.
DISCUSSION
Prior to this study, three amino acid residues had been reported to be important for HetR function. An S179N substitution mutant was initially isolated by Buikema and Haselkorn to implicate hetR in heterocyst differentiation and show that the substitution prevents differentiation (3). Subsequently, it was proposed that HetR is a protease that can degrade itself, that the S179N substitution prevents proteolysis in vitro, and that protein levels in a strain carrying the mutation are higher than those in the wild type (20, 21). Our results are consistent with these previous findings.
The active-site serine of HetR was later reported to be S152, and a hetR mutant strain, 884a, carrying the mutant allele on a plasmid was reported to be incapable of heterocyst differentiation (9). Strain 884a has a chromosomal copy of hetR with DNA corresponding to amino acid residues 35 to 168 of HetR replaced by DNA containing luxAB reporter genes, oriT and oriV from plasmid pMB1, and a gene encoding resistance to neomycin (1, 9). A strain with hetR(S152A) introduced into the chromosome of strain 884a by single recombination with a plasmid bearing the mutant allele was also reported to be incapable of differentiation (9). As described previously, the strain, designated S152A, should contain two copies of the hetR promoter, one promoting transcription of the mutant allele, followed by the pET3a vector and an Smr-Spr cassette, and the other promoting a fragment of hetR from residues 1 to 35, followed by luxA and luxB, the pMB1 oriV and oriT, a neomycin resistance cassette, and residues 168 to 299 of hetR. HetRS152A protein was also reported to lack autodegradation in vitro, and levels of the mutant protein were higher in strain S152A than were levels of the wild-type protein in the wild-type strain. Our results are inconsistent with most of these previous findings. The introduction of a hetR mutant producing an alanine residue at position 152 [hetR(S152A)] on a plasmid promoted heterocyst differentiation like the wild-type allele in strain UHM103, which has the hetR coding and promoter regions cleanly deleted from the chromosome. In addition, the replacement of a single nucleotide in the chromosome of the wild-type strain to give an S152A substitution had no discernible effect on heterocyst differentiation or patterning. In addition, HetR protein levels in a strain expressing hetR(S152A) from an inducible promoter in a ΔhetR background were similar to those in a strain expressing the wild-type allele under the same conditions. The simplest explanation for the discrepancy in the two sets of results is the difference in the strains that were used to express the mutant alleles and the methods for creating the chromosomal mutations.
HetR has DNA binding activity, and residue C48 has been reported to be necessary for this activity (11). It was also reported that incubation with a reducing agent prevents dimer formation by wild-type HetR in vitro and that the recreation of hetR(C48A) in the chromosome of the hetR mutant strain 884a did not restore heterocyst differentiation. Here we have confirmed that a reducing agent and a C48A substitution prevent the formation of a covalent linkage between two HetR monomers in vitro, suggesting the formation of a disulfide bond by the wild-type protein. However, in our hands, HetRC48A exhibited DNA binding activity, hetR(C48A) on a plasmid complemented a ΔhetR mutant, and the replacement of 2 bp in the chromosome of PCC 7120 to generate a C48A substitution did not affect heterocyst differentiation or patterning. Although differences in mutant construction between the work presented here and the previous work, which relied on strain 884a as well, could explain the different phenotypes of the mutant Anabaena strains, the difference in DNA binding results is enigmatic. Regardless of the cause of the discrepancy between our findings and previous results, it is clear that C48 and S152 are not necessary for HetR function in vivo.
In this work, genetic selection was used to identify 33 amino acid substitutions in HetR that affect heterocyst differentiation. Conservative substitutions at each site revealed that D17, G36, and H69 have important roles in HetR activity. Recombinant protein with a D17E substitution had substantially less DNA binding activity. Conservative substitutions at all three positions gave protein levels higher than those of wild-type HetR, suggesting that all three positions are involved in protein turnover in PCC 7120, but we believe that the effect at D17 may be indirect. We have found that the absence of patA leads to elevated levels of HetR protein in vivo (D. D. Risser and S. M. Callahan, unpublished data) and that hetR is required for transcriptional upregulation of patA (S. S. Young and S. M. Callahan, unpublished data). Because HetRD17E lacks DNA binding activity, the elevated level of HetR in vivo may be due to decreased levels of PatA instead of a direct effect of the D17E substitution on HetR turnover. We are currently investigating this possibility.
Both HetRD17E and HetRG36A were found to form dimers in vitro. Huang and coworkers proposed that the suppression of heterocyst formation by extra copies of hetR(S179N) in a wild-type background resulted from dimer poisoning, rendering functional molecules of HetR inert by coupling them to an excess of nonfunctional HetRS179N (11). This model could also be applied to the suppression of differentiation when extra copies of hetR(G36A) were present in a wild-type background but does not explain the excessive differentiation resulting from extra copies of hetR(D17E). It is possible that a heterologous dimer between HetR and HetRD17E may be functional while a HetRD17E homodimer is not. Thus, extra copies of hetR(D17E) in a wild-type background would increase the number of functional HetR dimers. An alternative hypothesis is that HetRD17E, which lacks DNA binding activity, may be functional in one or more other functions and that the addition of hetR(D17E) on a plasmid increased the activity of this function to promote differentiation.
Elevated levels of HetRD17E, HetRG36A, and HetRH69Y relative to that of the wild-type HetR extend the correlation between proper protein turnover and heterocyst differentiation, which was first demonstrated with HetRS179N (20). Unfortunately, we were unable to demonstrate protease activity resulting in autodegradation in vitro for any of the recombinant HetR proteins, including the wild-type version (data not shown) and, therefore, could not extend the correlation to HetR-protease activity. The turnover of HetR could be required to produce the active form of HetR from an apoprotein in developing cells. In this case, the degradation of HetR would be required for differentiation. Conversely, the protease activity of HetR could both limit the number of cells that differentiate by regulating the amount of HetR and also be required for the degradation of another protein, such as CcbP or an unidentified activator of differentiation, to promote differentiation, thus coupling the observance of low HetR protein levels and the ability to promote differentiation. Alternatively, both activities may be necessary for differentiation. From our results, it is not possible to tell if HetRG36A, HetRH69Y, and HetRS179N have lost protease activity or have lost the ability to be degraded. Separation of protease activity from HetR protein turnover or separation of turnover or protease activity from differentiation by mutation would help resolve questions regarding the regulation of differentiation by HetR.
ACKNOWLEDGMENTS
We thank Ann Garrido for help with cloning of the mutant hetR alleles.
This work was supported by grant MCB-0343998 from the National Science Foundation.
FOOTNOTES
- Received 7 August 2006.
- Accepted 5 January 2007.
- Copyright © 2007 American Society for Microbiology
