INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Azotobacter vinelandii is a soil bacterium that undergoes a process of cellular differentiation to form metabolically dormant cysts resistant to desiccation (for a review, see reference 38). Wild-type strains of A. vinelandii produce the extracellular polysaccharide alginate, a linear copolymer of D-mannuronic acid and its C-5 epimer L-guluronic acid. Alginate is essential for the encystment process, since it is a component of the intine and exine layers of the cysts (31) and since nonmucoid strains fail to form cysts (5, 26, 29).

The pathway for alginate biosynthesis in A. vinelandii has been elucidated (36); fructose-6-P is converted by four enzymatic reactions to GDP-mannuronic acid, which is the substrate of the polymerase. The resultant polymannuronic acid is secreted and modified by an O-acetylase and several extracellular C-5 epimerases to give the final product, alginate (9, 36, 37).

The algD gene codes for GDP-mannose dehydrogenase, which converts GDP-mannose, a metabolite used for the synthesis of different saccharides, to GDP-mannuronic acid. algD is located in a biosynthetic gene cluster which contains most of the genes that code for the enzymes involved in the synthesis and modification of alginate (5, 23, 26, 27, 37, 42). A correlation between alginate production and algD transcription was found for three A. vinelandii strains producing different alginate levels (5).

In Pseudomonas aeruginosa, the ^E factor AlgU (also known as AlgT) controls transcription of the alginate biosynthetic operon and is negatively regulated by MucA, an inner membrane protein that has been shown to act as an antisigma factor, and MucB, a periplasmic protein proposed to sense denatured proteins in the periplasm (45). In A. vinelandii the algUmucABCD genes have been characterized and have been shown to regulate alginate production (25, 29). Transcription of the A. vinelandii algD gene is initiated from three sites (5, 29, 30); initiation from one of them, p2, requires AlgU (5, 29). Transcription of algD is also under the control of the two-component regulatory system GacSA (6).

The bacterial cell wall is composed of a heteropolymer known as murein or peptidoglycan. The peptidoglycan consists of glycan chains of alternating units of N-acetylglucosamine and N-acetylmuramic acid that are frequently cross-linked by short peptides to each other (34). Escherichia coli and presumably most other gram-negative bacteria degrade up to 50% of their murein per generation; however, most of the liberated murein fragments (1-6-anhydro-N-acetylmuramyl [MurNAc]-L-Ala-D-Glu-mA2pm) are transported from the periplasm into the cytoplasm to reutilize the tripeptide L-Ala-D-Glu-A2pm to form new murein (12, 13, 32).

In enterobacteria the ampD and ampE genes form an operon (15, 22). The ampE gene encodes an inner membrane protein with an ATP-binding motif whose function is unknown (15, 22). The ampD gene encodes a cytosolic N-acetylmuramyl-L-alanine amidase essential for murein tripeptide recycling. AmpD is specific for anhydro-MurNAc-tripeptides (anhydro-MurNAc-L-Ala-D-Glu-mA2pm) (17); thus, it distinguishes the newly synthesized peptidoglycan precursors which lack the anhydrous bond from the peptidoglycan-derived anhydromuropeptides transported back into the cell for recycling (17).

In many gram-negative bacteria, the inducible -lactamase gene ampC is transcriptionally controlled by a regulator encoded by ampR, which belongs to the LysR family of transcriptional regulators (21). Mutations in ampD result in constitutive expression of AmpC -lactamase even in the absence of -lactam antibiotics. AmpR is inhibited in the presence of the main murein precursor UDP-MurNac-pentapeptide, and this inhibition is reversed by anhydro-MurNac-tripeptide (18). In ampD mutants, the substrate (anhydro-MurNac-tripeptide) accumulates and acts as an intracellular positive effector for ampC expression (16). The presence of ampD, even in bacteria lacking an inducible ampC -lactamase gene, suggests that cell wall recycling may have a signaling role in other cellular processes (33). In agreement with this suggestion, in Bacillus subtilis, transport of wall-derived peptides into the cell has been proposed to have a signaling role in the initiation of sporulation (35).

In this study, we describe the isolation and characterization of a Tn5 insertion mutation which increases transcription of algD and, consequently, alginate production in A. vinelandii. This was accomplished by isolating a mutant derivative of strain WI12 carrying an algD::lacZ fusion that showed a deep blue color on plates containing X-Gal (5-bromo-4-chloro-3-indolyl--D-galactopyranoside). The Tn5 insertion was shown to interrupt an operon homologous to E. coli ampDE. The results presented here indicate that A. vinelandii ampD and ampE gene products contribute to normal cell morphology and suggest that cell wall recycling has a signaling role for the control of alginate synthesis and cyst formation in A. vinelandii.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Microbiological procedures. Bacterial strains and plasmids used are listed in Table 1. Media and growth conditions were as follows. A. vinelandii was grown at 30°C in Burk's nitrogen-free salts supplemented with 2% sucrose (BS) (19). E. coli DH5 was grown on Luria-Bertani (LB) medium (28) at 37°C. Antibiotic concentrations used for A. vinelandii and E. coli, respectively, were as follows: tetracycline, 20 and 20 µg/ml; kanamycin, 2 µg/ml (not used for E. coli); ampicillin, 100 µg/ml (not used for A. vinelandii); nalidixic acid, 20 and 10 µg/ml; spectinomycin, 100 and 100 µg/ml; streptomycin, 2 and 20 µg/ml; gentamicin, 1.5 and 10 µg/ml.

Transposon mutagenesis. Random transposon mutagenesis of WI12 was carried out as described previously with a pUT derivative containing the mini-Tn5-luxAB transposon (8). Matings were done on Burk's sucrose (BS)-LB medium plates overnight at 30°C. The cells were resuspended in 10 mM MgSO₄ and plated on BSTcNal plates, and 200 WI12 Tc^r derivatives were isolated from each mating.

Cloning of the AC28 transposon insertion. Southern blot analysis was used to determine suitable restriction fragments on which the transposon-inactivated locus in strain AC28 was to be recovered. We identified a 3-kb KpnI fragment which contained the Tc^r cassette derived from the transposon (data not shown). KpnI-digested genomic DNA from AC28 was then size fractionated by agarose gel electrophoresis, purified from the gel, and ligated into pBluescript KS+ (Stratagene), resulting in plasmid pK28. Ligations were transformed into E. coli DH5, and transformed cells were plated on LB medium containing tetracycline.

Nucleic acid procedures. RNA and DNA isolation, cloning, Southern blotting, and nick translation procedures were carried out as described previously (39). Plasmids pS194.5, pS195, pS196, and pS196.2 (Fig. 1; Table 1) were used to determine the nucleotide sequence reported in this study. DNA sequencing was done with the Taq FS DNA polymerase and fluorescent dideoxy terminators using a cycle sequencing method. The precise site of transposition in AC28 was determined by nucleotide sequencing of pK28 across the transposon insertion junction. Primer extension of algD was carried out as previously described (5) using a primer extension kit (Amersham) as instructed by the manufacturer.

View larger version (21K): [in this window] [in a new window]   FIG. 1.   Physical map of the A. vinelandii ampDE region and the plasmids constructed in this study. Arrows, direction of transcription; triangles, antibiotic resistance cassettes. Abbreviations: Kp, KpnI; RV, EcoRV; S, SmaI; Sh, SphI.

Construction of plasmid pKT194.5 carrying the A. vinelandii ampDE operon. Cosmid clone pSM194, derived from an A. vinelandii genomic library, was shown to contain the ampDE operon in a 2-kb SmaI fragment. This fragment was cloned into the pBluescript KS+ vector to yield plasmid pS194.5 (Fig. 1). The ampDE operon was released from this construction as a BamHI-HindIII fragment and cloned into the pKT230 vector, previously digested with BamHI and HindIII endonucleases, producing plasmid pKT194.5.

Construction of ampD::Gm, ampE::Sp, and ampDE::Gm mutants. Plasmid pS196 was used to introduce into the unique SphI site of ampD a 0.8-kb SmaI fragment with a gentamicin resistance (Gm^r) cassette (1), and plasmid pSD1 with the Gm^r cassette ligated in the same orientation as that for the ampD transcription was isolated (Fig. 1), generating an ampD::Gm nonpolar mutation. Plasmid pS196.2 contains a unique EcoRV site within ampE. Therefore, in order to construct an ampE::Sp mutation, plasmid pS196.2 was cleaved with EcoRV and the 2-kb SmaI fragment containing an  spectinomycin resistance cassette (10) was inserted to create pSE1. To generate an ampDE::Gm deletion mutant, plasmid pS196 was cleaved with SphI and EcoRV (releasing the entire ampD gene as well as the 5' end of ampE), blunt ended, and ligated to a 0.8-kb SmaI fragment containing a gentamicin resistance cassette, resulting in pSDE1. Plasmids pSD1, pSE1, and pSDE1 (Fig. 1), which were unable to replicate in A. vinelandii, were used to introduce the ampD::Gm, ampE::Sp, and ampDE::Gm mutations into strains ATCC 9046, AEIV, and WI12. Transformants were selected using the corresponding antibiotic and were confirmed by Southern blot analysis to carry the desired mutations (data not shown).

Microscopic analysis. A. vinelandii cells, grown for 48 h on BS plates, were resuspended, mounted between two coverslips, and visualized under a microscope (Eclipse 330; Nikon) with a ×60 air lens (numerical aperture, 0.7) (Nikon Inc., Melville, N.Y.). Cell morphology was observed using differential interference contrast microscopy, and image acquisition was made by using a Sensys chip-cooled device (CCD) camera (Photometrics). The CCD camera was driven with PMIS processing software (Photometrics), and acquired images were processed using Adobe Photoshop (version 3.0).

Nucleotide sequence accession number. The A. vinelandii ampDE operon sequence reported here has been assigned GenBank accession no. AF237388.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Isolation of strain AC28. In A. vinelandii the synthesis of alginate correlates well with the transcription of the gene algD (5). We hypothesized that mutations in genes whose products regulate transcription of algD can be identified by a loss or increase in the -galactosidase activity of a strain carrying an algD::lacZ transcriptional fusion such as WI12. Random Tn5 mutagenesis of strain WI12 was carried out as described in Materials and Methods. Tc^r derivatives (3,200) were isolated and screened for a deep blue color on BS plates containing X-Gal. Five derivatives that were bluer than WI12 were identified (data not shown). Derivative AC28, which showed the deepest blue phenotype, was chosen for further analysis.

Cloning and DNA sequence. The sequence contiguous to the Tn5 insertion in AC28 was cloned to produce plasmid pK28. The DNA sequence of pK28 revealed homology to those of the P. aeruginosa and E. coli ampD genes. Southern blot analysis with pK28 as the probe led to the identification, in an A. vinelandii genomic library, of cosmid pSM194 carrying a 2-kb SmaI fragment with the ampD wild-type sequence. Sequence analysis of this DNA fragment revealed two open reading frames, one encoding a 187-amino-acid polypeptide (AmpD) and the second encoding a 280-amino-acid protein (AmpE). Potential Shine-Dalgarno sequences (AGGAG and GGGAG) are present upstream of both the ampD and ampE start codons. As in other bacteria, such as E. coli and P. aeruginosa, the ampE start codon overlaps the ampD TGA termination codon, suggesting that, as in P. aeruginosa and E. coli, these two genes form an operon (15, 20, 22). The predicted A. vinelandii AmpD and AmpE proteins exhibited 70 and 57% identity, respectively, to their homologues in P. aeruginosa.

-Galactosidase activity of strain AC28. To confirm and evaluate the effect of the ampD::Tn5 mutation on the transcription of algD, we compared the -galactosidase activity of strain WI12 with that of strain AC28 during growth in BS. As shown in Fig. 2, transcription of algD in AC28, measured as -galactosidase activity, was threefold higher than that in strain WI12.

View larger version (32K): [in this window] [in a new window]   FIG. 2.   algD transcription measured as -galactosidase activity in strains WI12 (wild type [WT]), WID1 (ampD), WIE1 (ampE), WIDE1 (ampDE), and AC28 (ampD::Tn5). Cells were grown on BS solid medium and were harvested at the indicated times.

Transcription of algD in ampD and ampE mutants. Since ampD and ampE genes seem to be arranged as an operon transcribed from a promoter located upstream ampD, it seemed likely that the ampD::Tn5 mutation in AC28 was polar on ampE expression. To investigate whether polarity on ampE was contributing to the increase in algD transcription, we constructed WI12 derivatives carrying a nonpolar ampD::Gm mutation (WID1) and an ampE::Sp mutation (WIE1) and strain WIDE1 with an ampDE deletion. Similar to strain AC28, strain WIDE1 showed a deep blue color on X-Gal plates and had increased levels of -galactosidase activity. In contrast, the ampD and ampE single mutations did not affect the transcription of algD (Fig. 2). These results indicate a cooperative effect of both mutations on algD transcription.

Production of alginate. In strain WI12 alginate production is reduced 75% compared to that in its parental strain, ATCC 9046; this reduction is likely due to the fact that the lacZ insertion is located on the 3' end of the algD gene, resulting in a GDP-mannose dehydrogenase with a reduced activity (5). To determine if the increased AC28 -galactosidase activity shown above correlated with the production of alginate, ATCC 9046 and AEIV derivatives carrying the ampD::Gm, ampE::Sp, and ampDE::Gm mutations were constructed. The ampD::Tn5 insertion from AC28 was transferred by transformation to strain AEIV, thus creating an AE28 mutant. The resulting strains were found to produce more alginate than their parental strains (Fig. 3); however, the extent of such an increase depended on the strain tested. Several attempts were made to transfer the ampD::Tn5 mutation to the highly mucoid strain ATCC 9046 with no success. The AE28 strain exhibited a higher alginate production level than strain AEDE1, which carries a nonpolar ampDE::Gm mutation, whereas AE28 carries a polar Tn5 mutation; therefore, polarity on transcription of a gene downstream of ampE also affecting alginate production could explain this difference. Plasmid pKT194.5, which carries a copy of the A. vinelandii ampDE operon, was transferred to strain AE28. As shown in Fig. 3, strain AE28/pKT194.5 produced alginate levels similar to that produced by wild-type AEIV, indicating that the high level of alginate production observed in strain AE28 was due to the absence of ampDE gene products.

View larger version (27K): [in this window] [in a new window]   FIG. 3.   Alginate production in ampD, ampE, ampDE, and ampD::Tn5 derivatives of strains ATCC 9046 (top) and AEIV (bottom). Cells were grown on solid BS medium for 48 h. Alginate was separated from the cells and measured as described in Materials and Methods. WT, wild type.

The ampD::Tn5 mutation increases transcription of algD from its three promoters. As in strain ATCC 9046, transcription of algD in strain WI12 is initiated from three sites (Fig. 4). We tested whether the increase in algD transcription in strain AC28 was caused by specifically increasing initiation from one of these promoters. Primer extension analysis using RNA isolated from strains WI12 and AC28 was carried out. As shown in Fig. 4, the ampD::Tn5 mutation increased transcription of algD from its three promoters.

View larger version (37K): [in this window] [in a new window]   FIG. 4.   Primer extension analysis of algD transcription in strains WI12 (lane 1) and AC28 (lane 2). RNA (50 µg) isolated from bacterial cultures grown for 48 h in Burk's medium supplemented with 2.0% sucrose was used. Arrows, transcription initiation sites from the three algD promoters.

Effect of ampD and ampE mutations on the morphology of A. vinelandii. The bacterial cell wall determines the shape of the cell (34). As ampD mutants are impaired in peptidoglycan recycling, the effect of the ampD and ampE mutations on cell shape was investigated. Microscopic examination of strains WID1, WIE1, WIDE1, and AC28 grown on BS plates was carried out. As shown in Fig. 5, the parental WI12 strain cells had a normal rod shape and were 3 to 5 µm long. The nonpolar ampD mutation produced cells similar to those found in WI12, whereas the ampE mutation produced spherical cells. Cultures derived from AC28 and WIDE1 mutants showed a heterogeneous cell morphology; they produced filamented and spherical cells. In addition they produced cells whose poles have an arrow shape, and the septs between dividing cells are abnormal, suggesting a failure to produce a normal septum (Fig. 5). The effect of the ampDE mutations on cell morphology was less dramatic when the mutants were grown on liquid medium (data not shown).

View larger version (55K): [in this window] [in a new window]   FIG. 5.   Effect of mutations in the ampDE operon on cell morphology in A. vinelandii. All photographs were captured with a ×60 air lens. Bar, 10 µm. WT, wild type.

Effect of the ampDE mutation on encystment. Transport of wall-derived peptides into the cell has been proposed to have a signaling role in the initiation of sporulation in B. subtilis (35). We tested the effect of the ampDE mutation on encystment. The formation of desiccation-resistant cells in WI12 and AC28 cultures was measured. Strain WI12 has a frequency of cyst formation of about 5.0%, while mutant AC28 has a frequency of less than 10⁶. After induction, vegetative cells divide to give small rounded cells that eventually become mature cysts. As expected, WI12 cells induced for encystment on n-butanol plates produced small rounded cells; in contrast, AC28 cells induced for encystment had the abnormal morphology found in vegetative cultures of this mutant (data not shown).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The aim of this study was to identify genes whose products participate in the control of alginate production. The isolation of mutants in which algD transcription was increased led us to the identification of the A. vinelandii ampD and ampE genes. In P. aeruginosa and other gram-negative bacteria, the ampDE operon participates in a pathway for recycling the cell wall. This recycling pathway was proposed to regulate other functions besides -lactamase induction (33, 34).

This study presents evidence suggesting that, in A. vinelandii, peptidoglycan recycling participates in the regulation of alginate production. Similar to the effect of the ampDE mutation on transcription of ampC, inactivation of the ampDE operon in A. vinelandii resulted in an increase in algD transcription accompanied by an increase in alginate production. Neither ampD nor ampE single mutations seemed to affect algD transcription although they resulted in an increased alginate production. Similarly, in an algR mutant, alginate production was reduced 50% despite the fact that algD transcription was not affected (30). This suggests the presence of another regulated point, besides the algD promoter, which also determines the extent of alginate production. The ampD and ampE mutations could affect this regulation point, resulting in higher alginate levels.

In A. vinelandii the increase in algD expression observed in the ampDE mutants might be caused by accumulation of the AmpD substrate. We propose that inactivation of ampE also contributes to the accumulation of the anhydromuropeptide in the cytoplasm. This hypothesis implies that AmpE also participates in the peptidoglycan recycling pathway. The fact that ampE and ampD genes constitute an operon supports this hypothesis. AmpE might be involved in regulating the extent of cell wall breakdown, slowing down the lytic process by controlling the activity of lytic transglycosylases. In E. coli three lytic transglycosylases and two endopeptidases are present in the periplasm or bound to the cytoplasmic membrane with the active sites oriented toward the periplasm (40). If AmpE was involved in cell wall breakdown, lysis of the cell wall would increase in the absence of AmpE, with the subsequent increase of anhydromuropeptides in the cytoplasm, which in the absence of AmpD accumulate and serve as a regulatory signal.

The ampDE mutation increased the transcription of algD from its three initiation sites. We have recently reported that the two-component global regulatory system GacSA also controls algD transcription from its three sites (6); thus, it is possible that the GacSA system participates in the signal transduction pathway turned on in the ampDE mutants. Accumulation of murein precursors in the A. vinelandii ampDE mutants might increase algD transcription directly or indirectly by activating an AmpR-like regulator.

This study shows that mutations in the ampDE operon also affected cell shape. We found that the ampE mutation produced rounded cells. In E. coli, strains with mutations in rodA-pbpA or the mre region (41, 43) are defective in the elongation process and grow as spherical cells, suggesting that in the ampE mutant cell elongation may be affected. The ampDE mutations produced, in addition to rounded cells, filamented cells of different lengths, a phenotype similar to those of E. coli pbpB mutants. In E. coli, PBP3, the product of gene pbpB, is involved in septum formation (4). Temperature-sensitive pbpB mutants stop dividing at high temperature and form long filaments (44). Nothing is known about the regulatory mechanisms controlling cell wall elongation and septum formation; however, it was suggested that changes in the cytoplasmic concentration of anhydromuropeptides could play a role (33). In this context, the morphological aberrations observed in the double mutant ampDE could be due to the presence of high levels of anhydromuropeptide which, in turn, deregulates the events of cell wall elongation and septum formation. Since the cell wall determines the shape of a cell (34), this result indicates that cell wall recycling has important implications in determining cell shape.

Upon induction of encystment, A. vinelandii vegetative cells divide to give two small rounded cells, which are later surrounded by an inner coat, the intine, and a thick laminated exine (38). Alginate is a component of the intine and exine layers. A. vinelandii algU or algK mutants are unable to produce alginate and therefore to form a mature cyst; upon encystment, these mutants form rounded cells lacking the intine and exine layers (26, 29). A mutation in algR did not impair alginate production but caused a cyst-defective phenotype, producing cysts lacking the intine layer (30). The ampDE mutant reported here is impaired in encystment. Upon conditions for induction of encystment, cells of this mutant remained as vegetative cells. Thus, encystment seems to be blocked at the first step. This result suggests that the level of muropeptides in the cytoplasm could be a signal necessary for A. vinelandii cells to undergo the differentiation process.

Finally, alginate synthesis also occurs independently of encystment; this polysaccharide is normally produced by vegetative cells not undergoing this differentiation process. This indicates that these two cellular processes can be separated and is in agreement with the finding that the ampDE mutation enhanced alginate synthesis but abolished encystment.