INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Quorum sensing is a type of intercellular communication found in many gram-negative bacteria. Communication between cells by quorum sensing is mediated by the production and release of specific N-acyl homoserine lactone (AHL) signal metabolites. The individual cells in a population are able to sense the population density through extracellular accumulation of these signals (7, 8, 10, 21). Two families of transcriptional regulators, called the LuxI and LuxR protein families, control this communication system. LuxI proteins catalyze the synthesis of the AHL signals, which then interact with LuxR proteins to activate or repress specific target genes. This process was first identified as a process that controls bioluminescence in the marine bacterium Vibrio fischeri but has since been found to regulate a variety of functions in many different bacteria (21).

Many bacteria produce more than one type of AHL signal, thus creating the opportunity for interactions among different quorum sensing systems in the same organism (1, 16). One such organism is the symbiotic nitrogen-fixing bacterium Rhizobium leguminosarum, which produces at least six different AHLs (13, 17). These molecules have been identified as N-hexanoyl homoserine lactone (C₆HSL), N-(3-oxo)hexanoyl homoserine lactone [(3-O)C₆HSL], N-heptanoyl homoserine lactone (C₇HSL), N-octanoyl homoserine lactone (C₈HSL), N-(3-oxo)octanoyl homoserine lactone [(3-O)C₈HSL], and N-(3-hydroxy-7-cis)tetradecanoyl homoserine lactone [(3-OH)C_14:1HSL] (9, 13, 17, 19).

The abilities to produce these different signals are encoded by at least four luxI homologs in this species. The chromosomal gene cinI encodes a protein that synthesizes (3-OH)C_14:1HSL, while the rhiI gene of the symbiosis (Sym) plasmid pRL1JI encodes an enzyme that produces both C₆HSL and C₈HSL (13, 17). Two additional loci for AHL synthesis, one each on pRL1JI and the chromosome, have been identified but have not been described yet (13).

The only R. leguminosarum genes known to be regulated directly at the transcriptional level by quorum sensing are those composing the rhiABC operon. This operon is found on the Sym plasmid pRL1JI and is controlled by the LuxR homolog RhiR. The RhiR protein activates rhiABC transcription in response to the presence of C₆HSL and other AHLs (4, 17). It was previously found that addition of the chromosomally encoded signal (3-OH)C_14:1HSL also resulted in increased expression of the rhiABC operon (9), but subsequent work showed that the activation is indirect, possibly a result of increased synthesis of other AHLs (13, 17). Because (3-OH)C_14:1HSL also activates genes that inhibit the growth of cells containing pRL1JI (9), uncharacterized regulatory functions encoded by the Sym plasmid typically suppress (3-OH)C_14:1HSL synthesis to levels that are about 5% of normal levels (13). These low levels of (3-OH)C_14:1HSL still appear to be required for normal production of the other AHLs made by cells containing pRL1JI (13).

The complex effects of (3-OH)C_14:1HSL suggest that there may be communication among different quorum sensing loci and thus a hierarchy of AHL signals within R. leguminosarum. The purpose of this research was to examine possible signal interactions in R. leguminosarum by measuring the temporal pattern of AHL synthesis at a global level and the response of this pattern to exogenous addition of individual AHLs.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and culture conditions. The bacterial strains used in this research are shown in Table 1. All cultures were maintained on TY medium (2) at 30°C. Cultures to be radiolabeled were grown at 30°C in RK minimal medium (22) supplemented with a 10% (vol/vol) addition of TY broth. In vivo labeling experiments were performed in the same supplemented RK minimal medium at 25°C. When appropriate, antibiotics were added at the following final concentrations: chloramphenicol, 25 µg/ml; kanamycin, 50 µg/ml; streptomycin, 50 µg/ml; and tetracycline, 5 µg/ml.

In vivo labeling of AHLs. The radiolabeling procedure used to measure de novo synthesis of AHLs was adapted from the method of Schaeffer et al. (18). Cultures to be labeled were grown overnight in TY broth to an optical density at 660 nm (OD₆₆₀) of more than 1.0. In order to downregulate any genes expressed in an overnight culture as a result of quorum sensing, the cells were diluted to an OD₆₆₀ of 0.1 in fresh TY medium and grown for 1 h at 30°C with shaking. After this, the culture was diluted to an OD₆₆₀ of 0.01 in supplemented RK minimal medium and incubated at 30°C with shaking. At each labeling time, the OD₆₆₀ of the culture was recorded, and 1.5 ml was removed and incubated with 1.5 µCi of [-¹⁴C]methionine (specific activity, 55 mCi/mmol; American Radiolabeled Chemicals, Inc., St. Louis, Mo.) for 3 h at 25°C with shaking. After this, the labeled sample was centrifuged at a high speed in a microcentrifuge, and the cell-free supernatant was collected. This supernatant was sequentially extracted with two equal volumes of ethyl acetate containing 0.01% (final concentration) glacial acetic acid. The ethyl acetate phase, which contained the radiolabeled AHLs, was collected and stored at 20°C.

Analysis of labeled AHLs. Different AHLs labeled by the method described above were separated and quantified by reverse-phase high-performance liquid chromatography (HPLC). Samples of labeled extracts were dried under a stream of sterile air, and each residue was redissolved in 50% methanol. The labeled AHLs were separated on a C₁₈ column (0.46 by 25 cm; Phase Separations, Norwalk, Conn.) by using a 20 to 100% methanol gradient at a flow rate of 0.5 ml/min for 133 min. One-milliliter fractions were collected, scintillation fluid was added to each fraction, and the fractions were counted with a liquid scintillation counter.

AHL purification and bioassay. AHLs were isolated from conditioned medium extracts of stationary-phase R. leguminosarum 14482Sr and 8401(pIJ1089) cultures by using HPLC as described above. Samples of individual HPLC fractions were assayed for AHL activity by one of four methods having the following AHL specificities: R. leguminosarum 8401(pRL1JI/pIJ1769) responded to (3-OH)C_14:1HSL, R. leguminosarum 8401(pIJ1642) responded to C₆HSL, Chromobacterium violaceum CV0blu(pSB403) responded to C₆HSL, C₇HSL, and C₈HSL and more weakly to (3-O)C₆HSL and (3-O)C₈HSL, and Agrobacterium tumefaciens NT1(pDCI41E33) responded to C₈HSL, (3-O)C₆HSL, (3-O)C₈HSL, and (3-OH)C_14:1HSL. Bioassays with R. leguminosarum 8401(pRL1JI/pIJ1769) and 8401(pIJ1642) and C. violaceum CV0blu(pSB403) were performed as previously described (3, 9). Bioassays with A. tumefaciens NT1(pDCI41E33) were performed by diluting overnight cultures of the strain to an OD₆₆₀ of 0.01 in TY medium and incubating the cultures with samples from individual HPLC fractions for 3 h at 30°C with shaking. Activation of the traG::lacZ reporter in response to AHLs was then measured by performing a -galactosidase assay (14). HPLC fractions that tested positive for AHL activity were dried by rotary evaporation, resuspended in acidified ethyl acetate, and stored at 20°C. The molecular structures of the individual compounds isolated by this method and by ¹⁴C labeling were inferred based on comigration with known AHL structures normally produced by R. leguminosarum (9, 13, 17, 19) and by the previously described specificities of the different bioassays for individual AHLs.

Addition of purified AHLs to R. leguminosarum. Samples of C₆HSL, (3-O)C₈HSL, C₈HSL, and (3-OH)C_14:1HSL, prepared as described above, were added to sterile flasks and dried under a stream of sterile air. R. leguminosarum cells were inoculated into supplemented RK minimal medium as described above and added to the flasks so that the final volume of each culture equaled the volume of supernatant medium from which the dried AHL sample was originally purified. The cultures were then incubated, and individual samples were pulse-labeled with [¹⁴C]methionine as described above. HPLC analysis of the resulting labeled AHLs was performed as described above.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Temporal patterns of AHL synthesis in R. leguminosarum. Bacterial cells incorporate [-¹⁴C]methionine into protein and into the homoserine lactone portion of any AHL molecules that are synthesized (18, 20). As a result, diverse AHL structures are all equally labeled with a single ¹⁴C atom. Organic extraction of cell-free supernatants of labeled cultures removes the labeled proteins and unincorporated [¹⁴C]methionine from the labeled AHLs. The AHLs present in the extracts can then be separated by HPLC, and the relative amounts of these AHLs can be quantified by liquid scintillation counting.

View larger version (28K): [in this window] [in a new window]   FIG. 1.   HPLC analysis of AHLs produced by R. leguminosarum 8401 and 8401(pRL1JI). (A) AHL-containing extracts from 14C-labeled cultures with pRL1JI () and without pRL1JI () were eluted by using a 20 to 100% methanol gradient. Individual AHLs collected in 1-ml fractions were quantified by liquid scintillation counting. (B) Unlabeled culture extracts were similarly eluted, and the fractions were bioassayed for AHL activity. Bioassays with R. leguminosarum 8401(pIJ1642) () and A. tumefaciens NT1 (pDCI41E33) () are shown for the 8401 (pRL1JI) extract, and bioassays with R. leguminosarum 8401(pRL1JI/pIJ1769) () are shown for the 8401 extract. Individual peaks correspond to C6HSL (centered at fraction 22), (3-O)C8HSL (fraction 26), C8HSL (fraction 38), and (3-OH)C14:1HSL (fraction 54). AHL structures were inferred from comigration with known standards and from the specificities of individual bioassays.

View larger version (25K): [in this window] [in a new window]   FIG. 2.   AHL biosynthesis in R. leguminosarum with and without pRL1JI. Cells of R. leguminosarum 8401 (A) and 8401(pRL1JI) (B) were pulse-labeled with [14C]methionine at various times, and the labeled AHLs were separated by HPLC and quantified by liquid scintillation counting. The amounts of C6HSL (), C8HSL (), (3-O)C8HSL (), and (3-OH)C14:1HSL () produced per cell were determined. Culture growth is indicated by the dashed lines. All values are averages based on three replicate experiments. The standard deviations were 11% for all values.

Addition of purified AHLs to R. leguminosarum 8401(pRL1JI). The patterns of AHL synthesis observed in cells containing pRL1JI might have resulted from some form of hierarchical control in which the production of each signal was dependent on prior synthesis of a different AHL. Separate additions of purified AHLs to newly inoculated cultures thus allowed us to determine the effect of each of the signals on the overall pattern of AHL synthesis. In every case, the quantities of purified compounds added were equivalent to those naturally produced by other R. leguminosarum cultures. This was accomplished by adding AHLs isolated from a culture sample whose volume was equal to that of the experimental cultures.

View larger version (18K): [in this window] [in a new window]   FIG. 3.   Effect of exogenous (3-OH)C14:1HSL on AHL biosynthesis in R. leguminosarum 8401. AHLs synthesized by cells grown in the presence (open symbols) and in the absence (solid symbols) of (3-OH)C14:1HSL were quantified as described in the legend to Fig. 2. The amounts of (3-O)C8HSL ( and ) and (3-OH)C14:1HSL ( and ) synthesized per cell were determined. The values are averages based on three experiments. The standard deviations were 11% for all values.

View larger version (16K): [in this window] [in a new window]   FIG. 4.   Effect of exogenous (3-OH)C14:1HSL on AHL biosynthesis in R. leguminosarum 8401(pRL1JI). AHLs synthesized by cells grown in the presence (open symbols) and in the absence (solid symbols) of (3-OH)C14:1HSL were quantified as described in the legend to Fig. 2. (A) Amount of (3-O)C8HSL synthesized per cell; (B) amount of C8HSL synthesized per cell; (C) amount of C6HSL synthesized per cell. All values are averages based on three experiments. The standard deviations were 11% for all values.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

This work represents the first direct, quantitative comparison of the relative amounts of different AHLs synthesized by a bacterial culture. Previous investigations relied on bioassays that can detect only the net accumulation of AHLs during culture growth. In addition, different bioassays have unique affinities for specific AHL structures, thus precluding direct comparisons of the different AHLs that may be present. By contrast, the radiolabeling protocol used here allows the accurate determination of short-term rates of AHL synthesis. Because a single molecule of labeled substrate is incorporated equally into the homoserine lactone portions of diverse AHL structures, it is also possible to determine the relative amounts of different AHLs produced. With this method we attempted to address questions of temporal control and hierarchical regulation of AHL synthesis in R. leguminosarum on a global level.

We found that the Sym strain R. leguminosarum 8401 produced two major AHLs, (3-O)C₈HSL and (3-OH)C_14:1HSL, while cells containing Sym plasmid pRL1JI produced (3-O)C₈HSL, C₈HSL, and C₆HSL (Fig. 2). These data are at odds with the results of Lithgow et al. (13), who identified C₆HSL, C₇HSL, and C₈HSL in Sym strain 8401 but did not detect either of the two signals found in this study. Similarly, when they used pRL1JI-containing cells, Lithgow et al. identified the same three AHLs described in this paper, but they also reported the presence of two more compounds, (3-O)C₆HSL and C₇HSL.

These discrepancies could reflect differences in the methods used for AHL preparation, separation, and detection. First, the overall amounts of AHLs produced might differ as a result of the different growth conditions employed. Previous studies were performed with cultures grown in TY medium (2), while in our experiments we used cultures grown in TY medium diluted 10-fold with minimal medium. Thus, certain AHLs might have been produced in these experiments at levels below our level of detection as a result of the dilute nutrient conditions. When supernatant extracts from cultures grown in full-strength TY medium were subjected to HPLC, however, no new signals were found by any of the bioassays employed.

Alternatively, Lithgow et al. separated their AHL-containing extracts by thin-layer chromatography (TLC) rather than by HPLC and then used a Chromobacterium-based bioassay to detect active signal compounds on the TLC plates. This bioassay, which is optimal for unsubstituted short-chain acyl homoserine lactone structures (21), responds only weakly to (3-O)C₈HSL and does not detect (3-OH)C_14:1HSL at all (Blosser Middleton, unpublished data). Thus, the absence of these two signals in the results obtained by Lithgow et al. with the Sym strain may be attributable to the bioassay itself. Thinking that our inability to detect the additional compounds which Lithgow et al. found in pRL1JI-containing cells might be explained in a similar way, we retested HPLC fractions from cultures grown in both dilute and full-strength media by using a quantitative Chromobacterium-based bioassay (3). The new assays, however, still failed to identify any new peaks of signal activity in any of the cultures tested.

We therefore propose that the remaining discrepancies between our results and those of Lithgow et al. are due either to a greater sensitivity of the TLC method compared to the sensitivity of separations made by HPLC or to the much higher concentrations of AHL-containing extracts being tested in their work. In light of the differences, our data should be interpreted as reflecting only the major AHLs produced by R. leguminosarum cultures and not the signals that may be synthesized at minor levels.

Analysis of the labeled AHLs demonstrated that there was a unique temporal pattern for synthesis of each signal in the bacteria. Interestingly, synthesis of (3-OH)C_14:1HSL by Sym cells began when the numbers of cells were quite low but increased steadily on a per cell basis during most of the culture growth period (Fig. 2A). While we have no explanation for this unusual pattern of regulation, it stands in sharp contrast to the model system of V. fischeri, in which AHL synthesis is rapidly activated at high cell densities through a mechanism of positive autoregulation (8). None of the R. leguminosarum signals appeared to be autoregulated, however, as prior addition of these signals to the culture medium had no effect (Fig. 3 and data not shown).

This last finding is at odds with previous reports that described positive autoregulation of the R. leguminosarum AHL synthase genes cinI and rhiI (13, 17). There are two possible reasons for the differences. First, the amounts of AHLs added may have been very different. The previous investigators added synthetic AHLs at arbitrary concentrations that may have had little relationship to natural levels, whereas we endeavored to add undefined but biologically relevant concentrations of naturally produced AHLs. More importantly, positive autoregulation was demonstrated previously by using plasmid constructs transferred into heterologous hosts in order to avoid the confusing effects of additional regulators or competing signals. Our work, in contrast, was performed in the native host with all other regulatory systems intact specifically so that we could observe global patterns of control. Our results thus suggest that while cinI and rhiI have a demonstrated capacity to autoregulate when they are isolated in heterologous hosts (13, 17), they do not appear to do so in their native hosts when grown in laboratory culture. Instead, the primary control of AHL synthesis in R. leguminosarum seems to be due to other, as-yet-undescribed regulatory mechanisms.

A third possible explanation for the apparent lack of autoregulation in the experiments performed is that the strains to which the AHLs were added were already producing the same signals. It is thus possible that the receptor proteins in these cells were already responding to endogenously produced AHLs, so that the exogenous signals had no additional effect. The weakness of this hypothesis is that because the purified AHLs were added immediately upon transfer of the cultures, endogenous AHLs would have to be abundant enough to saturate the cellular response even in very dilute suspensions. However, bioassays of supernatant media from newly inoculated cultures revealed undetectable levels of AHL activity (data not shown). In addition, labeling experiments revealed extremely low levels of de novo AHL synthesis in the first few hours after culture transfer (Fig. 2).

Alternatively, one might hypothesize that AHLs already bound to the receptors could be carried over during culture transfer as stable complexes that are not removed by simple dilution of the cell suspensions. This was why all cultures were routinely preincubated for 1 h at a low cell density before the cells were diluted further at the beginning of each experiment. Although the possibility of retained activity cannot be ruled out completely, the time allowed should have been more than adequate for the turnover of any active signal-receptor complexes.

While none of the AHLs exhibited positive autoregulation in our experiments, addition of (3-OH)C_14:1HSL to cultures of R. leguminosarum 8401(pRL1JI) cells did promote synthesis of all three other AHLs normally produced by these cells (Fig. 4). This finding confirms previous suggestions that (3-OH)C_14:1HSL activates the synthesis of other AHLs (13, 17), but with the unexpected twist that this activation appears to occur only in cells containing pRL1JI. Insertional inactivation of the cinI gene, which is solely responsible for (3-OH)C_14:1HSL synthesis, has been shown to reduce the synthesis of all other AHLs in R. leguminosarum 8401(pRL1JI) (13). This has led to the hypothesis that while cells containing pRL1JI repress (3-OH)C_14:1HSL production to very low levels, these levels are nonetheless required for the synthesis of other AHLs (13).

Activation of AHL synthesis to levels above the normal level by exogenous (3-OH)C_14:1HSL may be a simple extension of the same process. Our hypothesis is that increased amounts of (3-OH)C_14:1HSL promote AHL synthesis to an even greater extent than normal while simultaneously activating growth inhibition functions. Presumably, both growth inhibition and enhanced AHL synthesis are mediated by the interaction of (3-OH)C_14:1HSL with LuxR-like receptors encoded by pRL1JI. Such a model accounts for the absence of any changes in growth or AHL synthesis when (3-OH)C_14:1HSL is added to cells that lack pRL1JI. It is interesting that while (3-OH)C_14:1HSL appears to play a dominant role in regulating the synthesis of pRL1JI-encoded AHLs, at low concentrations its effects are seemingly modulated by other regulatory functions so that each signal is produced in a unique pattern. Whether this modulation can be accounted for by differential receptor and/or promoter affinities or whether it involves additional regulatory networks remains to be seen.

In summary, this study provided the first quantitative comparisons of the global rates of synthesis of multiple AHL signals in a single bacterial species. The results indicate that of the six AHLs reported to be synthesized by R. leguminosarum, only two or three are produced in abundance. The predominant AHLs that we identified for R. leguminosarum were synthesized with distinct and occasionally unexpected patterns that require further investigation to explain. Finally, our data suggest that primary control of quorum sensing signal generation in R. leguminosarum may not be due to autoregulation or AHL signal hierarchies. Discrepancies between some of our results and the results of previous studies provide a valuable perspective on the differences that may exist between the global patterns of gene expression and the measured responses of isolated genes.