The Twin Arginine Transport System Appears To Be Essential for Viability in Sinorhizobium meliloti

Bacterial strains, plasmids, media, and growth conditions.The bacterial strains, plasmids, and primers are listed in Tables 1 and 2. E. coli strains were routinely grown at 30°C or 37°C in Luria-Bertani (LB) medium. S. meliloti strains where routinely grown at 30°C in either LB medium, LB medium containing 2.5 mM magnesium and calcium (LB^MC), or Vincent's minimal medium (VMM), as previously described (47, 51, 56, 61). For defined medium, carbon sources were filter sterilized and added to medium at a final concentration of 15 mM. When required, final concentrations of antibiotics were as follows: tetracycline (Tc), 5 μg ml⁻¹; gentamicin (Gm), 20 μg ml⁻¹ or 60 μg ml⁻¹; neomycin (Nm), 100 μg ml⁻¹; streptomycin (Sm), 200 μg ml⁻¹; kanamycin (Km), 40 μg ml⁻¹; ampicillin (Amp), 200 μg ml⁻¹; and chloramphenicol (Cm), 60 μg ml⁻¹.

Genetic manipulations.Triparental matings for plasmid transfer between strains of S. meliloti and E. coli were carried out as previously described using MT616 as a helper strain (51). Transductions were carried out essentially as previously described (19). Constructed mutants were recombined essentially as previously described (23, 53).

DNA manipulations and plasmid constructions.All DNA isolation, restriction digestion, gel electrophoresis, ligations, transformations and PCRs, and Southern blot analyses were performed essentially as previously described using standard protocols (56).

Construction of pRK7813sac.Construction of pRK7813sac was carried out as follows. The sacB gene of pEX18Tc was PCR amplified using primers 1 and 2 and cloned into a unique NcoI site found in pRK7813. Transformants were screened onto LB medium supplemented with Tc (LB^Tc) and LB^Tc medium plus 5% sucrose. Sucrose-sensitive plasmids were purified, restriction mapped, and retested for sucrose sensitivity.

Construction of pBP79.pBP79 was constructed to harbor tatABC as well as the entire 5′ upstream intergenic region such that the transcription of tatABC could be controlled by its native promoter. In addition, the orientation of tatABC was such that it was in the opposite orientation to the P_lac promoter found in pRK7813sac. To generate this construct, a 2,674-bp PCR amplification product was generated using primers 9 and 10 and Rm1021 DNA as a template; the product was isolated, restricted, and ligated into pRK7813sac. A transformant that appeared to be correct on the basis of restriction was sequenced to confirm that the entire insert was identical to the wild type. The construct was designated pBP79.

Construction of pBP121.The construction of pBP121 was carried out in three steps. First, the region immediately flanking tatA was PCR amplified using primers 5 and 6 (Table 2 and Fig. 1). This 601-bp product was ligated into pKnock-Gm as a NotI/SmaI fragment and designated pBP103. Second, the downstream region flanking tatC was amplified using primers 7 and 8 (Table 2 and Fig. 1). The 486-bp PCR product was restricted with SmaI/KpnI and ligated into pBP103 that was cut with the same restriction enzymes. This construct was named pBP105. The plasmid pBP121 was ultimately constructed by cutting pBP105 with SmaI and ligating a gel-isolated 1.4-kb SmaI fragment containing the Kan^r cassette from pMM22 into this site.

• Open in new tab
• Download powerpoint

FIG. 1.

Genetic region containing tatABC of S. meliloti. A schematic diagram showing the relative positions of the relevant primers used to construct and verify constructs in the tat region. Numbered arrows correspond to the primers listed in Table 2. The plasmid above corresponds to pBP121, whereas the plasmid below corresponds to pBP79. Details of the construction and verification are in the text. The arrow above the P_lac promoter indicates the direction of transcription nptII, neomycin phosphotransferase cassette.

Construction of pBP91 and pBP92.To construct pRK7813tatC::nptII, primers 3 and 4 were used to PCR amplify tatC. The 950-bp fragment was then cloned into pBluescript SK to yield pMM4. A gel-isolated nptII cassette from pMM22 was cloned into pMM4 as a SmaI fragment to produce pMM13. The tatC::nptII construct was subsequently cut out of pMM13 as an EcoRI fragment and ligated into pRK7813. Transformants were isolated, and the plasmids were screened for the orientation of the EcoRI fragment relative to P_lac by restriction digest. The insert in pBP91 was oriented such that tatC::nptII was in the same direction as P_lac, whereas the insert in pBP92 tatC::nptII was in the opposite orientation to P_lac.

Plasmid stability experiments.Cultures of Rhizobium strains containing either pRK7813 or pBP79 were grown in LB^Tc medium (2.5 μg ml⁻¹) to ensure that all bacteria contained the appropriate plasmid. These cultures were then subcultured into LB broth that did not contain tetracycline. Cultures were then allowed to grow overnight at 30°C; they were subcultured the next day into fresh LB medium and again incubated overnight. The process was repeated until the experiment was completed. At each subculture, cells were serially diluted and plated onto LB agar, and 100 of the resultant colonies were screened for the presence of the plasmid on the basis of resistance to tetracycline.

Microscopy and plate reader analysis.Cultures of Rhizobium strains were grown in LB medium to mid-log phase. Cells were pelleted, washed three times, resuspended in saline, and incubated for 1 h at room temperature without shaking. When required, killed cells were incubated in 70% ethanol for 1 h in the place of saline. Cells were then pelleted, resuspended, and incubated with an appropriate amount of stain from a Molecular Probes BacLight detection kit. Cells were then incubated for 15 min in the dark. For microscopy, cells were added to an agarose-covered slide and analyzed using a Zeiss fluorescent microscope; for plate reader assays, cells were aliquoted into a 96-well plate and read using a fluorescent plate reader. In order to set up a calibration curve to assess the percentage of live cells, different proportions of live cells were mixed with killed cells. Stained cultures were directly compared to this curve. A log-phase culture of Rm1021 was arbitrarily set as a theoretical maximum (100%).

Plasmid promoter activity is required for chromosomal recombination.To characterize the twin arginine transport pathway in S. meliloti, we set out to create a gene disruption using a Kan^r cassette targeted to the tatC gene. The tatC gene was PCR amplified, cloned into pBluescript, disrupted with an antibiotic cassette, recloned into pRK7813, and used for an allelic exchange experiment, as previously described (47). Although putative recombinants were isolated, none of these was a bona fide double-crossover event; they appeared to be single-crossover events in which the vector was also integrated into the genome. As a result, in addition to having a mutant copy of tatC, the recombinants also contained a wild-type copy of the tat gene under the control of the vector promoter.

Since these experiments were carried out using LB medium as a complex medium, we hypothesized that the medium could be limiting the growth of a tat mutant, potentially providing a false-negative result. To address this hypothesis, the selection was attempted on various medium types. Defined media that were used included VMM (61), rhizobial minimal medium (RMM) (9), and M9A (13). For each of these, medium trials were carried out with glucose, mannitol, or succinate as a carbon source, and ammonia, nitrate, or glutamate were used as a nitrogen source. Complex media that were tried included tryptone-yeast extract (TY) (6), yeast extract-mannitol (YEM) (61), and LB media (37). In every case we were unable to isolate tat mutants that had lost the wild-type copy of the tat gene.

Since Rm1021 is known to be compromised with respect to phosphate utilization (64, 65) and since it is known that the strain carries an insertion element within a major regulator (44), we reasoned that perhaps our inability to isolate a tatC mutant was because of the Rm1021 background. To address this, we attempted a similar strategy using the closely related wild-type strains Rm2011 and Rm5000, which are independently derived from the parental strain SU47 and not compromised with respect to phosphate utilization (65), as well as SRmA363, which was constructed to be expR⁺ (38). The results of these experiments were consistent with our previous findings. Taken together, these data suggested that the tat genes may play an essential role under laboratory growth conditions and that S. meliloti could not grow in the absence of tatC.

To address the hypothesis that we were preferentially selecting for single-crossover recombinants that reconstructed the wild-type tatC gene in vivo, we reoriented the fragment in pBP92 such that tatC was in an opposite orientation to the P_lac promoter on the vector. We reasoned that if this, indeed, was the case, there should be a marked difference in the frequencies with which we could isolate single-crossover recombinants based on which construct was used in the allelic exchange experiment.

Consistent with this hypothesis, we were able to easily isolate single-crossover mutants using pBP91 (Table 3). When pBP92 was used, even after repeated attempts, we were unable to isolate any colonies. When the single-crossover events were expressed as a frequency, this represents a difference that is greater than 4 orders of magnitude (Table 3).

To determine if our inability to generate single-crossover recombinants was a function of using S. meliloti, we decided to attempt a similar experiment in the sequenced Sinorhizobium medicae strain WSM419. Since WSM419 is a wild-type strain that is sensitive to many commonly used laboratory antibiotics, we first isolated a streptomycin-resistant variant. This variant was designated SmD100 (Table 1). The plasmids pBP91 and pBP92 were subsequently conjugated into SmD100 and used for allelic exchange experiments. The results were comparable to the S. meliloti data (Table 3). Our ability to generate recombinants in these two closely related species correlated with the orientation of the construct with respect to the P_lac promoter in the vector (Table 3).

A tat deletion can be generated in a strain that is merodiploid for tatABC.The inabilities to construct a mutation in the tat region led us to hypothesize that the Tat system may be essential for either the growth or viability of S. meliloti. Alternately, the region may represent a region of the chromosome that is poorly recombinogenic. However, since we were able to isolate single-crossover events in this region, the latter possibility seemed to be unlikely.

A classic method for demonstrating that a gene is essential utilizes the isolation of temperature-sensitive alleles of the gene in question (15, 32). The isolation and characterization of temperature alleles are often time-consuming and difficult strategies to pursue. To address the possibility that tatABC could be essential for growth or viability in S. meliloti, a strategy was devised utilizing a host that would be merodiploid for tatABC (SRmA938, containing tatABC on a replicating plasmid) and a suicide vector (pBP121) that carried the regions immediately flanking this region (Fig. 1). The suicide vector was chosen such that the vector could easily be conjugated into S. meliloti, would not replicate in Rm1021, and did not contain any region of homology with pBP79. Due to problems generating a PCR product, we note that pBP79 and pBP121 do have approximately 200 bp of common DNA sequence (Fig. 1).

To generate a chromosomal deletion, pBP121 was mobilized into SRmA938 using pRK600 as a helper plasmid. Sm^r Nm^r transconjugants were selected, and approximately 200 were screened for Gm^r and Tc^r. Two colonies were found to be Gm^s, Tc^r, and Nm^r. These colonies putatively represented strains in which a double-crossover event had occurred where the chromosomal tatABC region was replaced with the insert carried on pBP121, and the plasmid pBP79 was retained. These were designated SRmA946 and SRmA947.

Subsequent Southern analysis of SRmA946 and SRmA947 using pKnock-Gm as a probe revealed that SRmA946 contained vector sequences, whereas SRmA947 did not contain sequences derived from this plasmid. SRmA946 was subsequently shown to be an extrachromosomal cointegration of pBP79 and pB121.

To ensure that SRmA947 contained a tatABC chromosomal deletion, PCR products were generated using primers (the primer pair 11 and 14 and the primer pair 12 13) outside the region in pBP121 and primers specific to the Nm resistance cassette. These PCR products were sequenced to confirm integration into the chromosome. The sequence data clearly show that SRmA947 contains a chromosomal deletion of tatABC.

Transduction of the tatABC deletion is host dependent.To ensure that the ability to delete tatABC was not due to a second site mutation that might have occurred in SRmA947 and allowed this genetic event to occur, a lysate of SRmA947 was prepared and used to transduce Rm1021. Since it was possible to have very few or no transductants on the selection plates, if the incorporation of a tat deletion was deleterious to the wild type, a control transduction in which the transduction of plasmid pBP79 into Rm1021 from SRmA947 was also carried out.

The results clearly show that after repeated independent attempts, we were unable to isolate Rm1021 transductants with the tat deletion (<10⁻⁸ recombinants/recipient). In contrast, Rm1021 transductants carrying pBP79 were readily isolated at frequencies similar to those previously published (2.2 × 10⁻⁷ recombinants/recipient [19]). To ensure that these were real transductants, five were single-colony purified, and the Tc^r plasmid was conjugated into E. coli DH5α, isolated, and analyzed by restriction analysis. Each of these plasmids was identical to pBP79.

Since SRmA938 was merodiploid for tatABC, we wished to determine if the transduction of the tat deletion was dependent upon the presence of an additional copy of tatABC. To carry out this experiment a lysate of SRmA947 was used to transduce SRmA938. Utilizing SRmA938, it was possible to generate Nm^r transductants at frequencies similar those found for other markers (3.0 × 10⁻⁶ recombinants/recipient [19]).

pBP79 is completely stable in an S. meliloti strain containing a tat deletion.Through routine growth and maintenance of SRmA947, the strain was observed to grow more slowly than the wild type. Since SRmA947 contains a plasmid and was routinely grown in broth cultures in the presence of tetracycline, it was thought that the slow growth might be a consequence of growing in the presence of the antibiotic. To address this, growth rates of Rm1021 containing pBP79 were directly compared to the growth of SRmA947. It was found that SRmA947 had a growth rate that was lower than that of the wild type. Whereas a typical doubling time for S. meliloti grown on complex is medium is about 3 h when it is incubated at 30°C (38), the doubling time for SRmA947 was found to be 5 h (data not shown).

Initially, it was unclear why SRmA947 would have a lower growth rate. One possible reason for the slow growth could be the presence of tatABC in multicopy. An alternate explanation might be that cells that lose pBP79 become nonviable or cease to grow. Since the plasmid vector pRK7813 is inherently unstable (27), it seemed plausible that a portion of a population of SRmA947 could be losing the plasmid and thus ceasing to grow.

To address the latter possibility, it was hypothesized that all cells carrying a chromosomal tat mutation must retain the tatABC genes expressed in trans. To assess this, we performed an assay to measure stability of pBP79 in either a wild-type background (SRmA938) or a tat deletion background (SRmA947). In addition we included SRmA946, which contains a cointegrate of pBP79 and pBP121.

Strains were first grown to an equivalent optical density in LB medium containing tetracycline and then subcultured into LB broth without tetracycline. Cells were removed after each subculture and plated for viability on LB medium. Colonies that grew on LB plates were subsequently screened for tetracycline resistance. It was reasoned that this assay could help determine whether the slow growth that was observed for SRmA947 was due to a cessation of growth or a loss of viability since plasmid portioning is not associated with any putative or known tat substrates.

The results show that under our growth conditions, approximately 50% of the population of Rm1021 that harbored pBP79 would lose the plasmid at each subculture step. After 5 days there was a complete loss of the pBP79 in SRmA938 (Fig. 2). Similarly, SRmA946, which contains a cointegrate of pBP79 and pBP121, was also lost at a similar rate (data not shown). In addition, Rm1021 lost the plasmid pRK7813 at a rate similar to that of pBP79, suggesting that the loss of the plasmid was not due to overexpression of tatABC (data not shown). In contrast, pBP79 was completely stable in SRmA947 even after more than a week of subcultures (Fig. 2 and data not shown).

• Open in new tab
• Download powerpoint

FIG. 2.

Stability of pBP79 containing the tatABC operon. Presence of pBP79 was measured by the proportion of Tc-resistant colonies found at each subculture. Error bars represent standard deviations of three independent experiments. Where not shown, error bars are smaller than the symbol. ▴, SRmA947; ⧫, SRmA938.

A proportion of SRmA947 cells display aberrant morphologies.Since the plasmid stability experiments showed that pBP79 was unstable in S. meliloti, it would be unlikely that every cell of SRmA947 contained a plasmid. We predicted that a proportion of cells in a growing culture of SRmA947 would lose pBP79, in essence yielding a genotypic S. meliloti tat mutant. Although these cells would be tat deficient, TatABC would deplete as the cells continued to grow. Therefore, observing the cells microscopically might give some insight into the role that TatABC might play in the overall physiology of S. meliloti.

It is known that in E. coli a twin arginine transport mutant undergoes improper cell division, resulting in elongated cells (59). Cell morphology is an important developmental indicator of a cell's adaptation to its environment. Rapid cell division during logarithmic growth indicates a healthy population attuned to its environment. Delay of cell division or changes to cell appearance can indicate that there may be one problem or many problems occurring within the cell.

Using both phase-contrast and differential interference contrast (DIC) microscopy, we were able to see that, like E. coli, a proportion of the S. meliloti mutants exhibited aberrant cell morphologies. Whereas the wild-type cells appeared predominantly as short rods, a higher proportion of SRmA947 cells appeared to be elongated, reaching lengths approximately two to three times that of wild-type cells. In addition, the cells appeared to be typically swollen or branched or were found joined together with cells of heterogeneous size, possibly indicating an inability to undergo proper cell separation or perhaps defects in cell wall assembly (Fig. 3).

• Open in new tab
• Download powerpoint

FIG. 3.

Effect of tatABC mutation on cell morphology and viability in S. meliloti. (A and C) Rm1021. (B and D) SRmA947. Panels A and B show DIC images of S. meliloti cells. Panels C and D show cells that were stained using a BacLight LIVE/DEAD bacterial viability kit (Molecular Probes, Invitrogen) and viewed using a fluorescent microscope. Live cells fluoresce green, whereas dead cells fluoresce red. (E) Percentage of live cells in cultures of Rm1021, SRmA938, and SRmA947, stained using a BacLight LIVE/DEAD kit and quantitated using fluorescence. Data are presented as the mean of three independent replicates. Values in parentheses represent standard deviations.

A growing culture of SRmA947 contains a significantly higher proportion of dead cells than the wild type.Bacterial populations typically contain a number of live and dead cells. Microscopic observation suggested that SRmA947 contained a large number of misshapen cells. From these observations, we are unable to ascertain whether these cells are viable. We wished to determine whether a culture of SRmA947 contained a higher number of dead cells and if these cells are correlated with the altered cell morphology observed.

To determine cell viability, we utilized a BacLight LIVE/DEAD bacterial viability kit (Molecular Probes, Invitrogen). When visualized, Rm1021 cultures were predominantly seen as small rod-shaped cells. The majority of these stained green, indicating that they were live cells (Fig. 3C). In contrast, cultures of SRmA947 had a marked increase in the number of dead cells that were observed microscopically (Fig. 3D). It was not uncommon to find many of the large cells associated in chains or abnormally shaped cells stained red, suggesting that these were cells with compromised membranes that were either dead or dying (Fig. 3D). We note that similar shapes were also described for Agrobacterium tumefaciens cells with tat mutations (17).

Direct microscopic analyses to determine the proportion of viable cells within each culture were complicated by the fact that with SRmA947 it was possible to find many areas where there were high proportions of dead cells. In an effort to remove possible sources of bias in determining the proportion of dead cells in each culture, viability was assayed using fluorescence measurements. Consistent with what was observed microscopically, it was found that the proportion of dead cells in a culture of SRmA947 was significantly higher than that of Rm1021 (Fig. 3). To ensure that the plasmid pBP79 was not skewing our quantitation, SRmA938 was also assayed in a similar manner. The data show that the proportion of dead cells is not significantly different from that of the wild type, suggesting that pBP79 expressed in Rm1021 did not account for the differences that were observed between SRmA947 and Rm1021 (Fig. 3E).