sinI- and expR-Dependent Quorum Sensing in Sinorhizobium meliloti

Abstract

Quorum sensing (QS) in Sinorhizobium meliloti, the N-fixing bacterial symbiont of Medicago host plants, involves at least half a dozen different N-acyl homoserine lactone (AHL) signals and perhaps an equal number of AHL receptors. The accumulation of 55 proteins was found to be dependent on SinI, the AHL synthase, and/or on ExpR, one of the AHL receptors. Gas chromatography-mass spectrometry and electrospray ionization tandem mass spectrometry identified 3-oxo-C₁₄-homoserine lactone (3-oxo-C₁₄-HSL), C₁₆-HSL, 3-oxo-C₁₆-HSL, C_16:1-HSL, and 3-oxo-C_16:1-HSL as the sinI-dependent AHL QS signals accumulated by the 8530 expR⁺ strain under the conditions used for proteome analysis. The 8530 expR⁺ strain secretes additional, unidentified QS-active compounds. Addition of 200 nM C₁₄-HSL or C_16:1-HSL, two of the known SinI AHLs, affected the levels of 75% of the proteins, confirming that their accumulation is QS regulated. A number of the QS-regulated proteins have functions plausibly related to symbiotic interactions with the host, including ExpE6, IdhA, MocB, Gor, PckA, LeuC, and AglE. Seven of 10 single-crossover β-glucuronidase (GUS) transcriptional reporters in genes corresponding to QS-regulated proteins showed significantly different activities in the sinI and expR mutant backgrounds and in response to added SinI AHLs. The sinI mutant and several of the single-crossover strains were significantly delayed in the ability to initiate nodules on the primary root of the host plant, Medicago truncatula, indicating that sinI-dependent QS regulation and QS-regulated proteins contribute importantly to the rate or efficiency of nodule initiation. The sinI and expR mutants were also defective in surface swarming motility. The sinI mutant was restored to normal swarming by 5 nM C_16:1-HSL.

Quorum sensing (QS) is the population density-dependent regulation of gene expression in bacteria mediated by the exchange of small signal molecules between nearby bacterial cells (16, 74, 77). QS enables the individual bacterial cells in a local population to coordinate the expression of certain genes. Such coordinated gene expression is important to the ability of bacterial pathogens and symbionts to infect and be effective in their eukaryotic hosts (36, 73, 80). We have been interested in the role of QS in the symbiotic interaction between the nitrogen-fixing bacterium Sinorhizobium meliloti and its Medicago legume hosts. The establishment and maintenance of this intimate symbiosis are intricate. They begin with initial contacts via extracellular signals produced by each partner, and these are followed by the initiation of new root nodule meristems, the induced, tubular invagination of host cell walls that carry multiplying bacterial cells into the nodule tissue; release of bacteria into envelopes of host plasma membrane; their differentiation and multiplication to fill most of the cell volume of the infected host cells; the initiation and regulation of symbiotic N fixation in these cells; the subsequent senescence of many host and bacterial cells; and the renewed multiplication of the bacteria in older nodules (17, 37, 70).

QS in S. meliloti is complex and only partially characterized. Genomically sequenced strain 1021 has been reported to produce a diversity of N-acyl homoserine lactones (AHLs) (40, 68). Several AHLs produced by the SinI synthase have long acyl side chains, including C₁₂-HSL, C₁₄-HSL, 3-oxo-C₁₄-HSL, C₁₆-HSL, C_16:1-HSL, 3-oxo-C_16:1-HSL, and C₁₈-HSL. In addition, Marketon et al. (41) reported that this S. meliloti strain also used a second, unidentified AHL synthase to produce AHLs with shorter acyl side chains, including C₆-HSL, 3-oxo-C₆-HSL, and C₈-HSL, when cultured in rich media but not when cultured in defined media.

Two AHL receptors, SinR and ExpR, have been identified in S. meliloti (40, 55). The genomic sequence indicates the presence of at least four other proteins with equally good homology to classical LuxR AHL receptors (19, 55). The SinR receptor was found to modulate sinI expression and the level of long-chain AHLs (41) and was subsequently reported to affect the expression of several other genes (27). The ExpR receptor was initially shown to activate genes required for production of EPSII in a population density-dependent and AHL-responsive manner (39, 55) and subsequently reported to modulate the accumulation of transcripts of a diversity of other genes (27). In genomically sequenced laboratory strain 1021, the expR open reading frame (ORF) is interrupted by a native insertion sequence (55). Spontaneous excision of this native insertion sequence restored functional AHL receptor activity in the resulting strain, 8530. Most S. meliloti strains isolated independently of strain 1021 have a functional expR sequence (55).

The first global study of QS-regulated functions in S. meliloti was a proteomic analysis of the responses of low-density cultures of expR mutant strain1021 to C₁₄-HSL or 3-oxo-C_16:1-HSL (10). Peptide mass fingerprinting (PMF) identified more than 50 of the 100 AHL-responsive proteins. The putative functions of the identified proteins included a number essential to symbiotic N fixation, carbon and nitrogen metabolism, various transporters, protein turnover, and DNA synthesis (10). A subsequent proteome study identified an additional 25 proteins in strain 1021 that responded to crude extracts of the bacterium's own AHLs (67). Hoang et al. (27) recently used complete genome microarrays to identify more than 170 genes in S. meliloti that differed in expression between the 8530 expR⁺ strain and an expR mutant (strain 1021), a sinI mutant, or a sinR mutant. These expR-, sinI-, and sinR-dependent genes were related to a wide range of functions, including EPSII synthesis, motility, cell division, transposases, transporters, and various metabolic activities.

Since the sinI, sinR, and expR mutants of S. meliloti 8530 are capable of infecting and forming effective nodules on host plants (41, 55), QS regulation of genes and proteins via SinI, SinR, and ExpR is not required for effective symbiosis. Nonetheless, there are several reasons for thinking that QS regulation could be important to establishing and maintaining an effective symbiosis between S. meliloti and its legume hosts. Several of the QS-regulated genes and proteins identified thus far in S. meliloti have functions required for normal symbiotic N fixation. These include PII nitrogen regulatory proteins (10, 67) that are central regulators of N metabolism in most organisms; the FixS, FixK1, and FixK2 proteins required for fixation and transcriptional regulation (27); and a molybdenum cofactor biosynthesis protein required for nitrogenase function (67) and the synthesis of symbiotically important signals such as EPSII (27, 59). QS regulation could also be important to the interactions between S. meliloti and Medicago hosts at two additional levels. It has been recently shown that higher plants, including Medicago plants, synthesize and secrete a diversity of compounds that effectively mimic AHLs in the ability to activate (or inhibit) QS-regulated gene expression in reporter bacteria (20, 69). These “AHL mimic” compounds from plant hosts have the potential to disrupt, initiate, or reinforce symbiotically relevant, QS-regulated gene expression in bacteria such as S. meliloti (3). A further layer of QS-related interaction between the partners involves the ability of the Medicago truncatula host to “listen” to bacterial AHL conversations. Recent proteome studies have shown that M. truncatula can detect nanomolar levels of bacterial AHLs, including 3-oxo-C_16:1-HSL from S. meliloti, and make global and specific responses to those bacterial signals (42). Thus, the AHL QS signals produced by S. meliloti could elicit many symbiotically relevant changes in the host root while concurrently regulating various symbiotically relevant, QS-regulated activities in the bacterium.

In this paper, we describe the identification of more than 50 proteins in S. meliloti that are differentially accumulated in an expR- or sinI-dependent manner. The effects of added SinI AHLs on the accumulation of these proteins were determined in order to confirm QS regulation and investigate specificity. The effects of the sinI mutation and added SinI AHLs on EPS production, swarming motility, and nodule initiation were examined. Changes in the expression of genes corresponding to representative proteins were examined by using single-crossover chromosomal promoter fusions, as was the ability of the single-crossover mutants to initiate nodule formation. In addition, we determined the identities and relative levels of the SinI AHL signals produced by strain 8530 and expR and sinI mutants under the culture conditions used for proteome and transcriptional analysis and reexamined evidence for a second AHL synthase.

MATERIALS AND METHODS

Bacterial strains, media, and growth conditions.

The bacterial strains used in this study are listed in Table 1. The 1021 strain served as the DNA source for the S. meliloti Genome Sequencing Project (http://bioinfo.genopole-toulouse.prd.fr/annotation/iANT/). Cells were grown at either 30°C (S. meliloti, P. putida and Agrobacterium tumefaciens) or 37°C (Escherichia coli). Growth media were TA (8) or M9 (44) for S. meliloti and Luria-Bertani (LB) medium for E. coli. Antibiotics were used at the following final concentrations: streptomycin, 500 μg/ml; neomycin, 200 μg/ml; gentamicin, 50 μg/ml for S. meliloti and 5 μg/ml for E. coli; tetracycline, 10 μg/ml; kanamycin, 25 μg/ml.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant characteristic(s)	Reference or source
Sinorhizobium meliloti
1021	expR102::ISRm2011-1 expR Smr	19
8530	1021 expR+ (formerly expR101) Smr	55
MG32	8530 with in-frame sinI deletion (sinI expR+) Smr	This study
MG75	1021 with in-frame sinI deletion (sinI expR) Smr	This study
Escherichia coli
DH5α	hsdR17 endA1 thi-1 gyrA96 relA1 supE44 ΔlacU169 (φ80dlacZΔM15)	Invitrogen
DH10B	F′ mcrA (mrr-hsdRMS-mrcBC) φ80lacZDM15 lacX74 deoR recA1 endA1 araD139Δ(ara leu)7697 galU galK1	Gibco BRL
AHL reporters
Agrobacterium tumefaciens NTI(pZLR4)	TraR receptor, cognate AHL = 3-oxo-C8-HSL Gmr	63
Pseudomonas putida 117(pAS-C8)	CepR receptor; cognate AHL = C8-HSL Gmr	65
E. coli JM109(pSB1075)	LasR receptor; cognate = 3-oxo-C12-HSL Apr	78
Plasmids
pBluescript II SK(+/−)	Cloning vector; ColE1 origin Apr	Stratagene
pTH113	Vector pRK7813; intact sinI gene Tcr	14
pJQ200SK	Cloning vector; p15A origin sacB+ Gmr	56
pRK600	pRK2013 Nm::Tn9 Cmr	51
pVO155	pUC119-derived integrational vector	51
pVMG	pVO155 with stop codons upstream of gus in all ORFs	This study
pMG1	Internal 649 bp of Smb21121 (ivdI) in pVMG	This study
pMG2	Internal 354 bp of Smb21133 in pVMG	This study
pMG3	Internal 413 bp of Smc00154 (gor) in pVMG	This study
pMG4	Internal 456 bp of Smc00242 in pVMG	This study
pMG5	Internal 472 bp of Smc01907 in pVMG	This study
pMG6	Internal 354 bp of Smc01966 in pVMG	This study
pMG8	Internal 354 bp of Smc02562 (pckA) in pVMG	This study
pMG9	Internal 472 bp of Smc02689 in pVMG	This study
pMG10	Internal 295 bp of Smc03930 (soxG2) in pVMG	This study
pMG11	Internal 472 bp of Smc40330 (mttB) in pVMG	This study

In-frame deletion of the sinI gene.

An in-frame deletion mutant of the Smc00168 sinI gene was generated by homologous recombination with sacB selection. DNA flanking sequences upstream and downstream of the sinI gene were amplified with genomic DNA from strain 8530 as the template and the primers 5′ACGCGTCGACGTTGAGTGGTCCGCCTACCG, 5′CGCGGATCGCGAATTCTCCGTTCACTATC, 5′CGCGGATCCTGAAACGGCACGCGCCGCCTG, and 5′GCTCTAGATTGAGTTGCCGCAGACTCG. A BamHI site was introduced at the 3′ end of the upstream flanking region of the sinI ORF, and another BamHI site was introduced at the 5′ end of the downstream flanking region. The two PCR products were cloned separately into pBluescriptIISK as a 494-bp SalI-BamHI fragment (upstream flanking region) and a 511-bp BamHI-XbaI fragment (downstream region), resulting in plasmids p168L and p168R. The SalI-BamHI fragment was released from SalI-BamHI-treated plasmid p168L and joined with the downstream flanking region by ligation with SalI-BamHI-treated plasmid p168R to generate p168de, which contains a novel BamHI site at the join site and a copy of sinI lacking 95% of the coding region. DNA sequencing confirmed the in-frame deletion. The mutated sinI gene was recloned as a SalI-XbaI fragment from p168de into sacB suicide vector pJQ200SK (56) to yield plasmid pJQ168de. The wild-type (WT) sinI gene was replaced in the S. meliloti chromosome by mobilizing pJQ168de into strains 8530 and 1021 with helper plasmid pRK600 (51). Isolates corresponding to single-crossover events were selected by gentamicin resistance, purified, and streaked onto duplicate tryptone yeast plates with or without 5% sucrose to select for homologous recombinants. The presence of the expected in-frame sinI deletion in strains MG32 and MG75 was confirmed by PCR and Southern hybridization. For complementation, plasmid pTH113 (14) carrying an 8,500-bp genomic clone containing the sinRI genes was mobilized into MG32 by triparental mating.

Culture of S. meliloti for proteome analysis and AHL identification.

Two independent cultures of S. meliloti strains 8530, 1021, MG32, and MG75 were generated from glycerol stocks and shake cultured overnight at 30°C and 225 rpm in 3 ml of TA medium without antibiotics but buffered to pH 6.0 with 10 mM morpholinepropanesulfonic acid (MES) to prevent alkaline inactivation of AHLs. The early stationary-phase cultures (optical density at 600 nm [OD₆₀₀], ∼1.5) obtained were centrifuged, resuspended in fresh medium, and diluted so that overnight cultures reached an OD₆₀₀ of 0.4 to 0.8. These log-phase cultures were diluted with MES-buffered TA to give corresponding 100-ml cultures with an OD₆₀₀ of 0.015. Duplicate cultures of the MG32 strain were supplemented with 200 nM C₁₄-HSL or C_16:1-HSL during initial growth in tubes and in the 100-ml subcultures. The 100-ml subcultures were grown in 500-ml Erlenmeyer flasks at 30°C and 220 rpm until the OD₆₀₀ reached 1.0 (late log phase). Cells were collected by centrifugation, and the pellets were freeze-dried for later protein extraction. At harvest, the pH of the cultures was 6.2 to 6.5. No contaminating bacteria were found.

Protein extraction, separation, quantification, and identification.

Proteins were extracted from freeze-dried cells as previously described (8) and quantified based on a modified Bradford protein assay (23). Protein concentrations were normalized, and the samples were subjected to two-dimensional (2D) gel separation as previously described (9, 48). Preparative gels were stained with Coomassie brilliant blue in a stepwise colloidal staining procedure (49). Digitized images (600 dots/in.) of the stained gels were quantified with MELANIE 4 image analysis software (Bio-Rad, Hercules, CA). Protein spot locations were compared to 10 landmark proteins and matched against a specialized proteomic database for S. meliloti 1021 (76). Each differentially accumulated protein spot identified by computer analysis was manually checked and defined. The optical density of each spot over its area (volume) as a percentage of the relative OD of the gel image (percent volume) was used to quantify each spot. Digitized spot images were statistically analyzed with GenStat 4.2 software as previously described (42). A polypeptide was deemed differentially accumulated, relative to its accumulation in the 8530 expR⁺ strain, if the chi-square value was less than 0.05. Results are based on proteins extracted from two independent cultures and separated in four independent gels per treatment per strain. Proteins were identified by tryptic digestion of the polypeptides isolated from the Coomassie-stained gels, followed by PMF by matrix-assisted laser desorption ionization-time of flight mass spectrometry performed on a Micromass TofSpec 2E time of flight mass spectrometer (Bruker Daltonik GmbH, Leipzig, Germany). Peptide mass fingerprints were identified by comparison against the S. meliloti 1021 proteomic database with Mascot software (Micromass; Waters Corp.) (76).

AHL purification and identification.

Cultures of the 8530 strain and the sinI (MG32) and expR (1021) mutants were grown as described for proteome analysis. Culture supernatants for each strain were extracted with acidified ethyl acetate and the extracts dried and redissolved as previously described (68). After passage through a Supelco DSC-NH₂ column to remove fatty acids, AHLs were further purified by reverse-phase high-performance liquid chromatography (HPLC). Extracts from 2 liters of culture supernatants were dissolved in 600 μl of acetonitrile, 200 μl of water was added, the mixture was centrifuged, and the supernatant was injected onto an analytical C₁₈ column (Altima; C₁₈, 5 μm, 250 by 4.6 mm) equilibrated with 10:90 acetonitrile-water. The column was eluted at 0.5 ml/min with 90:10 water-acetonitrile for 5 min and then at 1 ml/min with a linear water-acetonitrile gradient to reach 100% acetonitrile after 40 min, followed by an additional 20 min in 100% acetonitrile. The retention times for AHL standards were determined with the same column and gradient to elute 800 μl of acetonitrile-water containing a mixture of 8 nmol each of synthetic C₈-HSL, C₁₄-HSL, and C_16:1-HSL. The CepR and LasR AHL reporters (Table 1) were used to assay 10-μl aliquots of the HPLC fractions from elution of the AHL standards and S. meliloti extracts as previously described (65, 68). Gas chromatography-mass spectrometry (GC/MS) and electrospray ionization tandem mass spectrometry (ESI/MS/MS) identification of AHLs in the HPLC fractions were done as previously described (68).

Thin-layer chromatography (TLC) analysis of S. meliloti extracts.

Cultures were grown to stationary phase on LB medium supplemented with Ca and Mg and extracted with dichloromethane essentially as described by Marketon et al. (41). Concentrated extracts were dried in 2-ml Microfuge tubes by SpeedVac and treated with 100 μl of water or 0.1 M NH₄OH for 2 to 3 h, dried again, re-extracted with 200 μl of dichloromethane, centrifuged, concentrated to 10 to 20 μl, and spotted onto Whatman KC18F TLC plates. Plates were developed with 60:40 methanol-water and then dried. The A. tumefaciens NTI(pZLR4) AHL reporter was cultured and used to overlay TLC plates essentially as previously described (63).

Construction of gus promoter fusion/gene disruptions in ORFs of QS-regulated proteins.

Plasmid pVO155 (51), with a polylinker containing restriction sites for insertions and a promoterless gus (uidA) reporter gene encoding GUS, was used as a base for promoter fusion and gene disruptions. pVO155 was modified by creating stop codons in all three reading frames to ensure transcriptional fusions. Modified plasmid pVMG was confirmed by sequencing. To make gus fusion/gene disruptions, a region of about 200 to 600 bp in the middle of the target ORF was PCR amplified with the following primers: MG1 (Smb21121), 5′-ATCGCGGATCCAGCAATGCCTTTCCCATGT and 5′-ATCGCGGATCCACGACCCGCTCATAATCGA; MG2 (Smb21133), 5′-ATCGCGGATCCATAGTGAAATTAGCGCTTGT and 5′-TACTCGGATCCAAGACGATGGTCGAAGTGTA; MG3 (Smc00154), 5′-ATCGCGGATCCCTCTATGTTTATGCCTCGCA and 5′-ATCTCGGATCCAAGATATTGGCGAATTCCA; MG4 (Smc00242), 5′-ATAGAGGATCCAAAGATCTGCTTCATCTAT and 5′-ATCGCGGATCCATACGACCTTGATCTTGAA; MG5 (Smc01907), 5′-ATCGCGGATCCACAGGTATTCAATCGCAT and 5′-ATCTCGGATCCTTCAGCGACAGGATCA; MG6 (Smc01966), ATAGAGGATCCTCCTTCGACAAGAACATCTT and 5′-ATTCGCCGGATCCCAGAA; MG7 (Smc02562), ATCGCGGATCCAGCATTTCGAGGTTCTTCGT and 5′-TACGCGGATCCTAATTGAGAACGGTGA; MG8 (Smc02689), ATCGCGGATCCCGAACATTCTGATGAAGAT and 5′-TACGCGGATCCTAAAAGCGATCTTGGCGATG; MG9 (Smc03930), 5′-ATCGCGGATCCGTCATGTTCTCTGGCTCGGT and 5′-TACGCGGATCCAGCAGCCCGAACACGAA; MG10 (Smc04330), 5′-ATCGCGGATCCTATATCCAACGCAGGATT and 5′-TACGCGGATCCTTTGTCGGAATAACGGATAT. The PCR-amplified fragments were cloned in front of the promoterless uidA gene in pVMG and the resulting plasmid mobilized into the 8530 strain and into the sinI, expR, and sinI expR (MG75) mutant strains by triparental mating and selection on neomycin plates, verifying the inserts by PCR with a primer inside of uidA facing toward the insert paired with a primer outside and upstream of the original PCR fragment to confirm that the insert was in the correct location and orientation.

Swarming assays.

Plates containing 20-fold-diluted LB medium (with or without AHL) and 0.3% agar (Sigma Chemical Co., St. Louis, MO) were inoculated on the surface with 2 μl of a washed early stationary-phase bacterial suspension.

Nodulation tests.

Cultures of the S. meliloti 8530 strain, sinI mutant, and single-crossover gus fusion/mutant strains were grown to mid-log phase in TA medium from early stationary-phase starter cultures and centrifuged, and the cells were resuspended in an equal volume of water or diluted 100-fold for inoculation onto seedlings. Seeds of M. truncatula A17 from the South Australian Research and Development Institute were surface sterilized with 95% ethanol and 6% hypochlorite, followed by extensive washing with sterile water. The seeds were kept at 4°C overnight and then transferred to water agar plates. When seedling roots were 1.5 to 2 cm long, the seedlings were transferred to (nonsterile) seedling growth pouches (Northrup King, Minneapolis, MN) wetted to saturation with 9 ml of sterile, fourfold-diluted, N-free Jensen's medium (31). Four holes of about 1 cm were made in the perforated bottom of the seed trough with sterile forceps and a seedling carefully inserted through each hole with the root oriented toward the bottom and the cotyledon in the trough. Pouches were kept in an upright position in a box with spacers between sets of 10 to 15 pouches to prevent bending, incubated in a growth chamber (16-h day, 24°C, 100 μE; 8 h darkness, 24°C), and restored to original moisture levels each day with sterile water. After 18 to 24 h of growth, an indelible marker pen was used to gently mark the pouch face to indicate the position of each root tip at the time of inoculation. Seedlings were inoculated along the length of the root and slightly below with 100 μl/seedling. After 24 h in the growth chamber, the position of each root tip was marked again. Three weeks after inoculation, the number of nodules in different zones of the primary root was determined for each plant that had a primary root that reached the bottom of the pouch.

RESULTS

Proteome analysis of sinI- and expR-dependent functions.

As illustrated in Fig. 1, 55 protein spots were observed to differ significantly in accumulation between the WT strain and either the sinI mutant, the expR mutant, or the sinI expR double mutant. The accumulation of another four proteins was significantly altered by the addition of C₁₄-HSL or C_16:1-HSL to cultures of the sinI mutant even though their accumulation was not significantly affected by the sinI or expR mutation. Thirty-five of these 59 proteins showed at least a twofold change in accumulation. As shown in Table 2, it was possible to tentatively identify all but 3 of the 59 differentially accumulated proteins by PMF against sequences predicted by the genomic database. The PMF match for each of the identified proteins had a high confidence rating of 3 on the scale of Weiller et al. (76).

FIG. 1.

Proteome map of the S. meliloti 8530 expR⁺ strain grown in TA medium. Extracted proteins were separated by 2D sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) with an isoelectric focusing (IEF) range of pH 4 to pH 7. Protein spots that were accumulated to significantly (P < 0.05) different levels in the sinI, expR, or sinI expR mutant or in response to added SinI AHLs are indicated here and listed in Table 2. Hexagons designate proteins affected by both the sinI and expR mutations. Circles designate proteins affected by the sinI mutation. Squares designate proteins affected by the expR mutation. Triangles designate proteins affected by the sinI expR double mutation. Stars designate proteins affected by the addition of AHLs but not by the sinI or expR mutation.

TABLE 2.

Identification of sinI- and expR-dependent, AHL-responsive proteins in S. meliloti

Spot no.a	Geneb	Predicted protein	Accumulation dependence	Mascot search results			Mol wt/pI		Protein accumulation (% of 8530 expR strain)c
				Score	No. of peptides matched	% of sequence covered	Experimental	Theoretical	sinI	sinI + C14	sinI + C16:1	expR	sinI expR
52	SMa1296	AdhA1 (EC 1.1.1.1), alcohol degydrogenase	expR only	79	9	38	38,794/6.09	36,212/5.68	—d	—	140	60	—
23	SMa1450	Probable thiolase	sinI, expR	66	16	47	40,346/6.37	40,823/5.87	70	60	80	60	70
50	SMa1993	Conserved hypothetical protein	AHLs only	55	4	27	25,370/5.66	24,881/5.65	—	450	690	—	—
16	SMa2037	Putative oxidoreductase	expR only	66	5	10	88,764/6.71	86,608/8.15	—	—	—	60	50
22	SMa2041	Probable oxidoreductase	sinI, expR	76	8	36	29,565/6.55	29,589/6.05	80	80	80	50	60
18	SMb20170	AdhC3 (EC 1.2.1.1), glutathione-dependent formaldehyde dehydrogenase	sinI, expR	74	14	42	40,398/6.56	39,834/6.01	60	70	120	20	30
49	SMb20712	MocB, putative rhizopine uptake ABC transporter periplasmic solute-binding protein precursor	sinI only	55	9	33	33,311/4.40	32,849/4.54	180	130	70	—	—
38	SMb20899	IdhA (EC 1.1.1.18), myo-inositol dehydrogenase	sinI, expR	57	7	31	37,361/5.55	35,125/5.38	170	210	—	240	260
24	SMb21121	IvdH (EC 1.3.99.10), putative isovaleryl-CoAe dehydrogenase	expR only	66	22	57	40,274/5.96	41,188/5.58	—	—	—	50	70
53	SMb21133	Putative sulfate uptake ABC transporter periplasmic solute-binding protein precursor	AHLs only	65	7	26	38,333/5.06	37,055/4.97	—	590	610	—	—
14	SMb21309	ExpE6, putative membrane protein	sinI, expR	94	10	22	46,468/6.14	79,356/5.98	Missing	Missing	—	Missing	Missing
20	SMb21321	ExpA4, putative membrane-anchored protein	sinI, expR	119	10	37	42,977/6.48	43,684/6.03	20	20	60	20	20
30	SMb21652	LacE, probable lactose uptake ABC transporter periplasmic solute-binding protein precursor	sinI, expR	86	8	32	42,005/4.43	45,830/4.50	50	30	40	Missing	Missing
35	SMc00153	Conserved hypothetical protein	expR only	51	4	41	15,521/4.93	18,912/5.49	—	—	—	170	160
57	SMc00154	Gor (EC 1.6.4.2), glutathione reductase	expR only	55	10	26	45,365/6.39	49,891/5.81	—	—	—	150	180
27	SMc00242	Signal peptide hypothetical/global homology	expR only	74	7	29	39,128/5.16	38,504/5.23	—	—	—	220	190
39	SMc00361	Hypothetical protein	expR only	92	8	45	34,061/5.55	30,703/5.26	—	170	—	170	230
37	SMc00365	PheS (EC 6.1.1.20), phenylalanyl-tRNA synthetase alpha chain	sinI, expR	67	7	52	40,532/5.39	40,199/5.13	130	—	—	260	290
41	SMc00576	Conserved hypothetical protein	expR only	61	7	25	30,675/5.47	32,384/5.18	—	—	—	150	—
45	SMc00594	LigE, beta-etherase (beta-aryl ether cleaving enzyme)	sinI, expR	59	4	32	26,947/6.64	25,486/6.17	40	30	60	50	50
46	SMc00777	Conserved hypothetical protein	expR only	73	6	43	16,025/6.48	15,910/5.97	—	—	—	350	340
40	SMc00910	Conserved hypothetical protein	sinI, expR	54	5	27	31,859/5.59	30,503/5.22	170	—	—	170	170
13	SMc01345	AccC (EC 6.3.4.14), biotin carboxylase	expR only	96	12	43	46,569/6.14	48,968/5.72	—	—	—	60	—
8	SMc01885	Putative aminopeptidase P	sinI, expR	80	9	28	61,931/5.71	65,974/5.38	80	—	—	60	60
1	SMc01907	Transmembrane hypothetical/partial homology	sinI, expR     double	258	22	51	81,536/4.51	77,127/4.51	—	150	70	—	210
7	SMc01920	NuoG1 (EC 1.6.5.3), NADH dehydrogenase 1 chain G	sinI, expR	104	12	26	69,553/5.55	73,962/5.26	New	New	New	New	New
54	SMc01966	Putative spermidine/putrescine-binding periplasmic ABC transporter protein	sinI, expR	57	7	33	36,728/5.12	39,366/5.33	190	150	—	150	—
25	SMc02047	GcvT (EC 2.1.2.10), aminomethyl transferase (glycine cleavage system protein)	expR only	59	16	40	40,241/5.79	40,344/5.41	—	—	80	200	210
12	SMc02049	GcvP (EC 1.4.4.2), glycine dehydrogenase decarboxylating	expR only	64	14	21	101,185/6.13	104,039/5.66	—	—	—	260	360
26	SMc02094	Putative tansmembrane outer membrane protein	sinI, expR	63	7	14	44,303/5.24	84,469/4.72	160	140	—	220	170
36	SMc02156	Conserved hypothetical protein	sinI, expR	55	4	32	31,453/4.66	28,535/4.80	280		240	230	190
4	SMc02365	DegPI (EC 3.4.21.−), protease precursor	sinI, expR	166	15	45	58,352/4.86	53,003/4.80	160	160	—	180	180
32	SMc02378	Periplasmic binding protein	sinI, expR	141	13	46	35,010/4.59	35,789/4.68	230	210	—	230	170
56	SMc02465	SdhA (EC 1.3.99.1), succinate dehydrogenase flavoprotein subunit	sinI expR     double	54	9	21	63,363/6.05	66,935/5.76	—	—	—	—	170
9	SMc02562	PckA (EC 4.1.1.49), phosphoenolpyruvate carboxykinase	sinI, expR	140	15	44	53,328/5.92	58,083/5.56	40	20	—	20	20
11	SMc02562	PckA (EC 4.1.1.49), phosphoenolpyruvate carboxykinase	sinI, expR	112	16	47	54,049/6.02	58,083/5.56	30	20	—	20	20
55	SMc02562	PckA (EC 4.1.1.49), phosphoenolpyruvate carboxykinase	sinI, expR	59	9	24	54,523/5.81	58,083/5.56	40	20	—	20	30
3	SMc02634	Transmembrane hypothetical	sinI only	68	10	23	67,942/4.76	73,515/4.78	120	—	140	—	—
51	SMc02689	Probable aldehyde dehydrogenase	AHLs only	87	14	43	47,617/5.82	55,216/5.43	—	130	190	—	—
5	SMc02857	DnaK, heat shock protein 70 (Hsp70) chaperone	sinI, expR	51	7	15	65,402/5.14	68,907/4.91	140	200	130	170	190
6	SMc02857	DnaK, heat shock protein 70 (Hsp70) chaperone	sinI, expR	54	9	17	65,254/5.11	68,907/4.91	160	220	130	200	140
48	SMc02896	IlvEI (EC 2.6.1.42), branched-chain amino acid aminotransferase	sinI only	95	9	33	38,794/5.78	39,587/5.43	190	240	—	—	270
21	SMc02898	KdsB (EC 2.7.7.38), 3-deoxy-manno-octulosonate cytidylyltransferase	sinI only	81	6	42	31,019/6.87	27,086/6.18	150	140	—	—	—
28	SMc03061	AglE, ABC transporter α-glucoside-binding periplasmic protein	sinI, expR	159	12	41	41,881/4.68	49,672/4.71	170	150	—	160	150
2	SMc03233	Transmembrane hypothetical	sinI, expR	111	8	32	70,786/4.57	45,017/4.53	140	140	—	150	120
34	SMc03786	Bfr, bacterioferritin	expR only	69	5	56	25,678/4.89	18,337/4.75	—	—	—	160	160
59	SMc03823	LeuC (EC 4.2.1.33), 3-isopropylmalate dehydratase large subunit	expR only	67	11	31	48,110/6.03	50,963/5.58	—	—	—	180	170
33	SMc03864	ABC transporter amino acid-binding periplasmic protein	sinI, expR	111	11	44	26,329/4.47	29,353/4.65	220	210	140	250	210
44	SMc03930	SoxG2 (EC 1.5.3.1), sarcosine oxidase gamma subunit	expR only	56	5	39	30,035/6.46	18,917/5.78	—	—	—	40	40
29	SMc03969	Conserved hypothetical protein	AHLs only	70	6	35	29,861/5.05	26,729/4.81	—	190	170	—	—
47	SMc03972	MexE2, transmembrane multidrug efflux system	sinI, expR	55	4	15	40,359/5.22	41,853/5.46	30	40	80	30	60
17	SMc03983	FbaB (EC 4.1.2.13), fructose-bisphosphate aldolase class I	sinI, expR	112	14	43	39,531/6.83	37,297/6.22	50	30	—	40	30
19	SMc03983	FbaB (EC 4.1.2.13), fructose-bisphosphate aldolase class I	sinI, expR	96	11	41	39,599/6.58	37,023/6.22	50	40	—	40	50
58	SMc04040	IbpA, heat shock protein	expR only	77	8	53	29,852/6.58	17,402/6.11	—	—	—	70	180
10	SMc04330	MttB, trimethylamine methyltransferase	expR only	50	10	27	49,259/5.85	56,934/5.35	—	—	—	30	50
15	SMc04347	Conserved hypothetical protein	expR only	53	17	33	81,117/6.51	73,870/6.11	—	—	—	Missing	Missing
31		No good match	sinI, expR				40,333/4.40		New	New	—	New	New
42		No good match	sinI, expR				16,294/5.02		80	80	90	60	40
43		No good match	sinI, expR				50,866/5.71		190	210	—	400	370

^aNumbers refer to protein spots in Fig. 1.

^bBolded genes were tested as single-crossover fusions (see Table 4).

^cAverages from two independent cultures and four independent gels per treatment or strain.

^dValue not significantly different from accumulation in 8530.

^eCoA, coenzyme A.

About 75% of the proteins tentatively identified as QS regulated in Table 2 showed significant changes in their accumulation upon the addition of either C₁₄-HSL or C_16:1-HSL to cultures of the sinI mutant strain. This serves to confirm that these proteins are indeed subject to AHL-mediated QS regulation. Of the half dozen known SinI AHLs, C₁₄-HSL and C_16:1-HSL were chosen for these tests because the synthetic, purified compounds were available. Previously, C₁₄-HSL was detected only in cultures of S. meliloti 1021 grown on a defined medium (68), whereas C_16:1-HSL was detected in cultures grown on both rich (41) and defined (68) media and is also produced by another S. meliloti strain, AK631 (68).

Multiple isoforms were detected for three of the proteins shown in Fig. 1: DnaK, PckA, and FbaB. In each case, the isoforms appeared to be separated by virtue of differences in pI rather than molecular weight. The sinI and expR mutations or addition of AHLs affected the individual isoforms of each protein in very similar ways (Table 2). Thus, there is no evidence that QS regulation affected the relative abundance of the different protein isoforms.

Responses at the transcriptional level.

To investigate whether the QS-regulated changes in protein accumulation shown in Table 2 were accompanied by similar changes in regulation at the transcriptional level, single-crossover gus promoter fusions corresponding to 10 of the 59 differentially accumulated proteins were constructed and placed in WT and mutant backgrounds as described in Materials and Methods. Seven of the 10 fusions tested were found to have significant changes in GUS activity in response to the sinI, expR, or sinI expR background or to added AHLs when tested at either late log or early stationary phase (Table 3). Fusions in Smc00154, Smc01907, and Smc01966 had no significant differences in activity in the different backgrounds or in response to the tested AHLs at either growth stage (data not shown).

TABLE 3.

Activity of single-crossover gus fusions in genes corresponding to QS-regulated proteins

Strain background for single crossover	GUS activity (Miller units)a [culture OD]b with the indicated protein accumulation dependence:
	ivdH, Smb21121 (expR only)c	SO4 transporter, Smb21133 (AHLs only)	Hypothetical signal peptide, Smc00242 (expR only)	pckA, Smc02562 (sinI, expR)	Aldehyde dehydrogenase, Smc02689 (AHLs only)	soxG2 Smc03930 (expR only)	mttB, Smc04330 (expR only)
WT	26† [0.7 ± 0.1]	1.4 × 102† [0.7 ± 0.1]	4.2 × 102† [0.8 ± 0.1]	3.9 × 103† [0.6 ± 0.1]	4.5 × 102† [0.6 ± 0.1]	18† [0.8 ± 0.1]	3.3 × 103† [1.3 ± 0.2]
expR	21‡	2.0 × 102‡	5.7 × 102‡	0.6 × 103‡	3.3 × 102‡	3.5‡	1.0 × 103‡
sinI expR	21‡	2.1 × 102‡	5.7 × 102‡	0.8 × 103‡	3.0 × 102‡	3.1‡	NAd†
sinI	18§	1.7 × 102†	5.4 × 102‡	0.9 × 103‡	2.9 × 102‡	2.9‡	3.2 × 103†
sinI + C14	16§	1.9 × 102‡	4.6 × 102†§	0.7 × 103‡	3.2 × 102‡	6.3‡	3.1 × 103†
sinI + C16:1	18§	1.9 × 102‡	3.9 × 102†	3.0 × 103§	3.8 × 102§	4.4‡	3.4 × 103†
WT	12 × 102† [1.8 ± 0.1]	4.8 × 102† [1.8 ± 0.1]	2.2 × 103† [1.8 ± 0.1]	25 × 103† [1.8 ± 0.1]	1.8 × 103† [2.0 ± 0.1]	520† [2.0 ± 0.1]	4.7 × 103† [2.0 ± 0.1]
expR	10 × 102†	9.0 × 102‡	6.2 × 103‡	20 × 103‡	2.9 × 103†	23‡	2.0 × 103‡
sinI expR	10 × 102†	8.6 × 102‡	6.2 × 103‡	21 × 103‡	2.5 × 103†	22‡	NA§
sinI	8.5 × 102†	7.7 × 102‡	6.2 × 103‡	19 × 103‡	2.2 × 103†	17‡	5.6 × 103§
sinI + C14	NA†	7.4 × 102‡	3.5 × 103†	23 × 103†‡§	4.2 × 103‡	18‡	5.3 × 103§
sinI + C16:1	8.1 × 102†	5.9 × 102†	5.9 × 103‡	26 × 103§	4.8 × 103‡	19‡	5.3 × 103§

^aValues are averages from three replicate cultures for each test strain in a representative experiment. Values followed by different symbols are significantly different from each other at the 0.05 level according to Tukey's studentized range (HSD) test.

^bOD values apply to all values in the section of the column in which they appear.

^cProtein accumulation dependence is indicated in Table 2. Fusions in Smc00154, Smc01907, and Smc01966 were tested similarly to those shown and had no significant changes in activity.

^dNA, value not available from this experiment. The value from an independent experiment was not significantly different from the value for the sinI mutant.

Identification of AHLs produced by the 8530 WT and sinI and expR mutants.

HPLC fractions from ethyl acetate extracts of each strain were assayed with two AHL reporter strains, P. putida F117(pAS-C8) (= CepR) and E. coli(pSB1075) (= LasR). As shown in Fig. 2A, the LasR reporter (cognate AHL = 3-oxo-C₁₂-HSL) was modestly activated by compounds in fractions peaking at 18 and 20 and activated to essentially saturating levels by compounds in almost all fractions from 24 through 43 in extracts of the WT. In contrast, no significant activity was detected by the LasR reporter in any fraction in extracts from the sinI mutant (Fig. 2A). HPLC fractions from the expR mutant activated the LasR reporter in much the same way as extract fractions from the sinI⁺ expR⁺ 8530 strain, although appreciable quantitative differences were seen in some fractions. As illustrated in Fig. 2B, the CepR reporter (cognate AHL = C₈-HSL) detected activity peaking in fractions 22, 34, and 38 in extracts from the 8530 expR⁺ strain. The sinI mutant strain showed activity only in fractions peaking at 22. Under the same conditions, the AHL standards C₈-HSL, C₁₄-HSL, and C_16:1-HSL eluted with peaks in fractions 20, 35, and 39, respectively.

FIG. 2.

Bioassay of HPLC fractions of extracts from the 8530, sinI, and expR strains of S. meliloti with the LasR and CepR AHL reporters. Ethyl acetate extracts of culture supernatants were separated by reverse-phase HPLC and QS-active compounds detected with AHL reporters. (A) Activities in 10-μl samples of 1-ml fractions detected with the LasR reporter relative to the reporter-only control. The C₈-, C₁₄-, and C_16:1-HSL standards eluted in fractions peaking at 21, 35, and 38, respectively (data not shown). (B) Activities detected with the CepR reporter relative to the reporter-only control. AcCN, acetonitrile:water gradient.

The AHLs present in pooled HPLC fractions from the 8530 expR⁺, expR, and sinI strains were identified by GC/MS and ESI/MS/MS. Both methods showed that the S. meliloti 8530 expR⁺ strain secreted and accumulated a complex mixture of five AHLs with long acyl side chains when cultured in the rich medium used for the proteome comparisons (Table 4). These include 3-oxo-C₁₄-HSL, C₁₆-HSL, 3-oxo-C₁₆-HSL, C_16:1-HSL, and 3-oxo-C_16:1-HSL. The accumulation of 3-oxo-C_16:1-HSL was substantially greater than the accumulation of the other four AHLs. No AHLs were detected in any of the pooled fractions from cultures of the sinI mutant, indicating that all of the AHLs detectable in extracts of the 8530 expR⁺ strain were synthesized by SinI. HPLC fractions 1 to 25 contained considerable amounts of substances from the culture medium that could potentially mask the detection of AHLs with short acyl side chains (C₄ to C₈). Among the compounds eluting in these fractions were cyclic dipeptides (diketopiperazines, [DKPs]) (A. Eberhard and M. R. Gronquist, unpublished data). DKPs are very stable, can be synthesized by bacteria or formed from amino acids during autoclaving, and are known to be capable of affecting AHL QS in bacteria (12, 28). Two (cVV and cPF) of the nine synthetic DKPs tested were able to stimulate the Agrobacterium reporter in TLC plate assays (A. Eberhard and M. Gao, unpublished data).

TABLE 4.

Mass spectrometric identification of AHLs secreted by S. meliloti 8530 and sinI and expR mutant strains

AHL	Relative responsea						HPLC fractions
	8530		expR		sinI
	GC/MS	ESI/MS/MS	GC/MS	ESI/MS/MS	GC/MS	ESI/MS/MS
C16	1.0	1.0	1.0	1.0	0	0	43-46
3-Oxo-C14	2	0.5	0.3	0.1	0	0	33-36
C16:1	2	1	0.5	2	0	0	37-39
3-Oxo-C16	0.5	0.2	0.3	1	0	0	37-39
3-Oxo-C16:1	80	8	30	7	0	0	33-36

^aAmount of AHL detected relative to C₁₆-HSL.

Attempts to detect AHLs with short acyl side chains.

In an earlier study, Marketon et al. (41) reported the ESI/MS/MS identification of C₈-HSL in whole culture extracts of the 1021 (expR mutant) strain and provided TLC evidence for the production of AHLs with short acyl side chains corresponding in mobility to C₈-HSL, C₆-HSL, and 3-oxo-C₆-HSL by both the 1021 strain and the sinI mutant. These observations led the authors to conclude that S. meliloti 1021 has a second AHL synthase that makes AHLs with short acyl side chains in a manner independent of the long acyl side chain AHLs produced by the SinI synthase. Since our analysis failed to detect any compounds corresponding to AHLs with long or short acyl side chains in extracts of our sinI mutant by either GC/MS or ESI/MS/MS (Table 4) and detected only one sinI-independent CepR-active compound (Fig. 2), we attempted to repeat the TLC studies of Marketon et al. (41).

Cultures of the 8530 WT, sinI mutant, and expR mutant (strain 1021) were grown and extracted essentially as described by Marketon et al. (41). The extracts were separated on C₁₈ TLC plates and compounds detected by overlays with the A. tumefaciens AHL reporter. As shown in lane 1 of Fig. 3, culture supernatant extracts from the 8530 sinI⁺ expR⁺ WT strain contained a number of substances capable of stimulating the Agrobacterium reporter, roughly resolved into six spots (a to f). Extracts of the whole cultures (cells plus supernatant) and extracts of the expR mutant strain, as tested by Marketon et al. (41), gave TLC patterns very similar to that shown in lane 1 (data not shown). The activities present in culture supernatants from the 8530 strain after treatment with 0.1 M NH₄OH are shown in lane 2. Alkali rapidly hydrolyzes the lactone rings of AHLs and greatly diminishes their ability to activate AHL receptors (81). Treatment with 0.1 M NH₄OH essentially eliminated spots a, b, d, and e and substantially reduced, but did not eliminate, the large spot (f) at the origin, where AHLs with long (C₁₄ to C₁₈) acyl side chains remain. NH₄OH treatment was able to inactivate >60-fold greater amounts of C_16:1-HSL than were added to the mixture of AHLs shown in lane 3 (data not shown). The residual activity at the baseline after NH₄OH treatment of WT extracts might thus correspond to one or more Agrobacterium-active, alkali-stable compounds. Lane 3 shows the AHL standards recovered after addition of C_16:1-HSL, C₈-HSL, C₆-HSL, and 3-oxo-C₆-HSL to autoclaved LB medium, followed by extraction with dichloromethane, treatment with water, and re-extraction. NH₄OH treatment eliminated all of these activities, as expected (lane 4). Culture supernatant extracts of the sinI mutant (lane 5) contained no active compounds corresponding to spots a, b, and f of the 8530 parent, indicating that the synthesis and/or secretion of these compounds is directly or indirectly dependent on SinI. The sinI mutant extracts did contain a compound corresponding to spot c and reduced amounts of compounds corresponding to spots d and e of the 8530 parent. Spots corresponding to d and e were not detectable in the sinI mutant extract after NH₄OH treatment, whereas spot c was alkali stable. Our in-frame sinI deletion mutant and the sinI::Tn5 mutant of Marketon et al. (41) gave quite similar TLC patterns (data not shown). Autoclaved LB medium contained substances capable of activating the Agrobacterium reporter, but only when added in two to three times greater amounts than the 5 ml of bacterial cultures tested in Fig. 3.

FIG. 3.

TLC analysis of QS-active compounds in extracts from the 8530, sinI, and expR strains of S. meliloti. Dichloromethane extracts from 5 ml of stationary-phase culture supernatants of the bacteria were dried, treated with water or 0.1 M NH₄OH for 3 h, dried, re-extracted with dichloromethane, and spotted onto C₁₈ TLC plates. The plates were eluted with 60:40 methanol-water and dried, and QS-active compounds were detected on the plates with an overlay of the Agrobacterium reporter. Lanes: 1, water-treated extract of 8530 parent; 2, NH₄OH-treated extract of 8530; 3, water-treated extract of 5 ml of LB medium containing C₈-, C₆-, 3-oxo-C₆-, and C_16:1-HSLs; 4, NH₄OH-treated extract of 5 ml of LB medium containing C₈-, C₆-, 3-oxo-C₆-, and C_16:1-HSLs; 5, water-treated extract of sinI mutant; 6, NH₄OH-treated extract of sinI mutant. Active spots resolved by TLC in lane 1 are indicated by the letters a to f.

EPS and swarming phenotypes of the sinI mutant.

Colonies of the sinI mutant strain were much less mucoid than WT colonies, as expected since previous studies established that SinI AHLs stimulated production of EPSII via the ExpR receptor (27, 39, 55). Consistent with these earlier studies, the addition of 200 nM C_16:1-HSL and 3-oxo-C_16:1-HSL was able to restore sinI colonies to normal mucoidy, whereas the addition of 200 nM C₄-, C₆-, C₈-, 3-oxo-C₆-, 3-oxo-C₁₂-, C₁₄-, or C₁₆-HSL did not (data not shown).

As shown in Fig. 4, strain 8530 was able to form large, mucoid “swarm” colonies that spread slowly over the agar surface and developed distinct patterns of high cell density after 2 days. Under the same conditions, the sinI mutant showed no swarming on the agar surface (Fig. 4), nor did the expR or sinI expR mutant (data not shown). Swarming of the sinI mutant was restored to WT levels by genetic complementation of the sinI mutant with the WT sequence for sinI (Fig. 4). Swarming of the sinI mutant, but not the sinI expR mutant, was restored to normal levels by 5 nM C_16:1-HSL but required a 200 nM concentration of C₁₆- or 3-oxo-C_16:1-HSL. Swarming was not restored by 200 nM C₈-HSL, C₁₄-HSL, or 3-oxo-C₁₄-HSL (data not shown). Swarming of the mutants, but not the 8530 expR⁺ strain, depended critically on the dryness of the agar surface at the time of inoculation. If the agar surface was allowed to partly dry in a laminar-flow hood while the agar was setting (15 to 20 min), the mutants were not able to generate surface swarming colonies, whereas the mutants swarmed like the WT if the plates were covered soon after pouring.

FIG. 4.

Surface swarming colonies of S. meliloti 8530 expR⁺, sinI mutant, and complemented sinI mutant strains. Plates containing 0.3% agar and 20-fold-diluted LB medium were inoculated with the 8530 strain (A), the sinI deletion mutant strain MG32 (B), or the sinI deletion mutant carrying a copy of the WT sinI gene in pTH113 (C) and incubated for 2 days.

Effect of disrupting sinI and QS-regulated genes on nodulation of host roots.

In order to compare the rate and efficiency of nodule initiation by strains defective in QS regulation or mutated in QS-regulated genes, the 8530 expR⁺ parent, the sinI mutant, and the single-crossover gus promoter fusion/mutants in the 8530 expR⁺ background (Table 5) were inoculated onto seedlings of M. truncatula in growth pouches. As shown in Table 5, the WT strain elicited an average of 0.5 nodule/plant on the primary root above RT1, the position of the root tip at the time of inoculation. The WT formed an average of 1.1 nodules between RT1 and RT2, the interval of root growth during the following 24 h, and an average of 1.5 nodules on the primary root between RT2 and the bottom of the pouch. Due to the acropetal development of root hair cells that are susceptible to infection by S. meliloti, nodules that develop in these three regions correspond to sustained infections that were initiated approximately 8 to 12 h after inoculation, 12 to 36 h after inoculation, and 36 to ∼110 h after inoculation, respectively (6, 7). Plants inoculated with a 100-fold dilution of the WT developed no nodules above RT2, corresponding to a delay in nodule initiation of at least 36 h. The sinI mutant had a delayed-nodulation phenotype, forming no nodules from infections initiated during the first 12 h after inoculation and very few nodules on the primary root from infections initiated between 12 and 36 h after inoculation. Plants inoculated with the single-crossover mutants in ORFs Smc00154, Smc00242, Smc01907, and Smc04330 had a pattern of delayed nodulation similar to the sinI mutant (Table 5). Bacteria recovered from representative nodules on these plants were confirmed as the inoculant strain by antibiotic resistance.

TABLE 5.

Nodule initiation on primary roots of M. truncatula by S. meliloti 8530 and mutant strains

Inoculum strain	Avg no. of nodules/planta with avg root zone length (cm)d of:			No. of plants
	∼1 (above RT1)b	1.98 (RT1-RT2)c	∼11 (below RT2)
8530	0.5	1.1	1.5	84
8530 diluted 100-fold	0**	0**	4.1**	20
MG32 (sinI)	0**	0.4**	1.8	91
MG32 diluted 100-fold	0**	0**	2.2	19
Smb21221::pVMG	0.4	1.2	1.0	29
Smb21133::pVMG	0.4	1.2	1.1	35
Smc00154::pVMG	0.07**	0.8	3.7**	44
Smc00242::pVMG	0.03**	0.9	1.4	33
Smc01907::pVMG	0.04**	0.6**	2.4**	44
Smc01966::pVMG	1.0*	1.0	1.5	30
Smc02689::pVMG	0.2**	0.5**	1.1	36
Smc03930::pVMG	0.4	1.0	1.5	35
Smc04330::pVMG	0**	0.4**	2.5*	41
Uninoculated control (H2O)	0	0	0	18

^aAverage number of nodules per plant that differ significantly from the corresponding value for plants inoculated with the 8530 strain according to multiple analysis of variance tests are indicated as follows: **, P < 0.01; *, P < 0.05.

^bAverage number of nodules per plant on the primary root above the position of the root tip at the time of inoculation (RT1).

^cAverage number of nodules per plant on the primary root between RT1 and the position of the root tip 24 h after inoculation (RT2).

^dFor the 84 plants inoculated with the S. meliloti 8530 strain, above RT1 is the average length of the zone of developing root hairs at the time of inoculation, RT1-RT2 is the distance between RT1 and RT2, and below RT2 is the distance between RT2 and the bottom of the pouch.

DISCUSSION

Functions of QS-regulated proteins.

Twelve of the 56 QS-regulated proteins with good matches in the genomic database were annotated as hypothetical proteins of unknown function. Among the 44 proteins with sufficient homology to suggest a probable function, there are several contributing to potential roles as oxidoreductases (e.g., AdhA1, AdhC3, IdhA, IvdH, NuoG1, GcvP, SdhA, and SoxG2), transporters (e.g., MocB, LacE, AglE, and MexE2), protein synthesis/processing components (e.g., PheS, DegP1, DnaK, and IbpA), and transferases (e.g., GcvT, IlvE1, KdsB, and MttB).

A number of the S. meliloti proteins identified in Table 2 may have QS-regulated, sinI- and/or expR-dependent functions that are relevant to symbiotic interactions with host plants. ExpE6 is one of the proteins most strongly affected by added AHLs or mutations in sinI and expR. This protein is involved in the synthesis of EPSII (5). EPSII can serve as a (redundant) signal to the host plant to allow infection (22). The accumulation of ExpE6 falls below detectable levels in the sinI and expR mutants but is restored to levels of the 8530 expR⁺ parent when the sinI mutant is supplemented with 200 nM C_16:1-HSL (Table 2). Thus, it appears that C_16:1-HSL or one of the other AHLs made by SinI (e.g., 3-oxo-C_16:1-HSL) normally activates the ExpR receptor and this directly or indirectly enhances accumulation of the ExpE6 protein and contributes to production of EPSII. Accumulation of ExpA4, another protein involved in production of EPSII, follows a similar pattern of QS regulation (Table 2). This relationship is consistent with earlier transcriptional studies (39, 55).

Two of the QS-regulated proteins, IdhA and MocB, have roles in the synthesis and uptake of rhizopines. Rhizopines are derivatives of inositol synthesized in host root nodules by the differentiated, N-fixing bacteroid form of rhizobia. Undifferentiated sibling cells of the rhizopine-producing strains have the ability to take up and utilize the rhizopines as nutrients, but other bacteria do not, so that rhizopines provide members of the producing strains with a selectively utilizable nutrient (45, 46). IdhA, myo-inositol dehydrogenase, is the first enzyme in inositol catabolism. idhA mutants of S. meliloti L5-30 are unable to catabolize the rhizopine 3-0-methyl scyllo-inositol (18). IdhA is required for efficient N fixation and symbiotic competitiveness by Sinorhizobium fredii (33) and is important to S. meliloti in the utilization of inositol, which is relatively abundant in the rhizosphere (18). The MocB protein contributes to rhizopine catabolism as a transporter (60). MocB was previously identified as one of the proteins that accumulated strongly in stationary-phase cultures of the expR mutant strain (10).

Other proteins with functions potentially relevant to host interactions include PckA, which is crucial to the utilization of dicarboxylic acids, the form of carbon supplied most abundantly to S. meliloti in symbiotic association with its hosts (53). Bacteroids inside nodules express pckA at high levels (51), and pckA mutants are defective in nitrogen fixation in planta (54). The putative glutathione reductase corresponding to Smc00154 may play a significant role in host interactions, as suggested by the important contribution of glutathione production to nodule initiation and symbiotic N fixation by S. meliloti (24). Mutants lacking LeuC (SMc03823) and related leucine/isovaline synthesis mutants of S. meliloti are known to be defective in entry into root nodules (13, 25). The S. meliloti homolog of PotA, a putative spermidine-putrescine transporter (Smc1966), is known to be part of an operon of ABC transporter components required for normal interactions of A. tumefaciens with plant cells (43). DegP (HtrA) is a small heat shock protein that is suppressed by SinI AHL-mediated QS via ExpR (Table 2). In Streptococcus pneumoniae, DegP functions as a periplasmic serine protease that can contribute importantly to virulence (30). degP mutants had no obvious nodulation phenotype in S. meliloti (21), but DegP levels are enhanced in Sinorhizobium medicae at acidic pH (57). DnaK is a chaperone complex that governs the activity of sigma factor 32 (rpoH), which is crucial to heat shock and stress responses in many alpha proteobacteria (47). RpoH1 mutants in S. meliloti are defective in both heat shock tolerance and symbiosis (52). AglE is a periplasmic binding protein component of a sugar ABC transporter system in S. meliloti (78). AglE is specific for binding sucrose, maltose, and trehalose and may thus be important in the utilization of these α-glycosides from host plants. There appears to be a second, semiredundant transport system for these α-glycosides in S. meliloti that is more responsive to trehalose than sucrose (32). Mutants in the α-glycoside transport systems show normal nodulation but defective root colonization (32).

The identification of gcvT as QS regulated in S. meliloti may be of interest in relation to “riboswitching.” The expression of certain genes involved in central metabolism, including gcvT, is dependent on the binding of relevant metabolite ligands (such as glycine) to “riboswitch” elements in their mRNAs (2, 50, 66). We have found sequences very similar to the conserved glycine-binding riboswitch elements of Bacillus subtilis and other bacteria (38) upstream of the gcvT gene (Smc02047) in S. meliloti. Thus, riboswitching and QS appear to converge on the regulation of the glycine cleavage system in S. meliloti, the first committed step toward using glycine as an energy source.

sinI- and expR-dependent QS regulation of protein accumulation.

Since only a modest fraction of the proteins corresponding to ORFs in the S. meliloti genome can be detected and resolved by 2D polyacrylamide gel electrophoresis, and since we only examined cytoplasmic proteins with isoelectric points between 4 and 7, our proteome analysis is clearly limited in defining the full range of sinI- and/or expR-dependent, QS-regulated functions. Nonetheless, the number of proteins showing sinI- and/or expR-dependent accumulation is substantial, representing about 5% of the ∼1,150 proteins resolved by the gels. In comparison, the microarray studies of Hoang et al. (27) indicated that about 2.4% of the 6,200 known and putative ORFs in S. meliloti exhibited sinI- and/or expR-dependent expression. A study done with P. aeruginosa showed that the percentage of QS-regulated proteins (23%) detected was likewise higher than the percentage of QS-regulated genes (6 to 10%) identified (1). This was attributed to extensive posttranscriptional QS regulation (1). Posttranscriptional QS regulation may also play important roles in S. meliloti, although the molecular mechanisms involved have not been identified.

About 50% (26/49) of the proteins with ExpR-dependent accumulation (Table 2) had significantly greater accumulation in the expR mutant than in the 8530 expR⁺ strain. Thus, ExpR-mediated responses to SinI AHLs resulted in reduction or suppression of protein accumulation as frequently as it resulted in greater accumulation.

A comparison of the proteome analysis of sinI- or expR-dependent QS reported here with the microarray analysis of Hoang et al. (27) reveals some major differences. Among the 150 sinI- or expR-dependent genes identified by microarray analysis, only two of the corresponding proteins (ExpA4 and ExpE6) were identified in Table 2. No clear explanation for this lack of overlap is available. The genes identified as QS regulated by Hoang et al. (27) were not confirmed by demonstrating their responsiveness to added AHLs. Earlier proteome and microarray studies of QS regulation in P. aeruginosa (1, 4, 26, 61, 71, 75) found that confirmation of differences between QS mutant and WT strains by responsiveness to added AHLs markedly narrowed identification. Several other factors, including posttranscriptional regulation, differences in the initial dilution of the test cultures, differences in growth phase at harvest, differences in subculturing methods, and differences in culture conditions may all contribute to the lack of correspondence.

For the 31 proteins showing sinI- and expR-dependent accumulation in Table 2, it is noteworthy that the sinI and expR mutations and the sinI expR double mutation each independently affected the levels of these proteins and did so in the same direction and to approximately the same extent. This correspondence provides further confirmation that proteins in this group are indeed subject to QS regulation, even in those cases where the changes in accumulation are modest (less than twofold). The correspondence between the effects of sinI and expR on the accumulation of these proteins also indicates that their QS regulation involves the interactions of SinI AHLs with the ExpR receptor. In agreement with this model, the levels of 28 of these 31 proteins were affected by the addition of C₁₄-HSL or C_16:1-HSL to the sinI mutant. SinI AHLs other than C₁₄-HSL or C_16:1-HSL may act through the ExpR receptor to effect regulation of the three proteins that did not respond to the two AHLs tested.

Levels of the four proteins showing sinI-only-dependent accumulation (MocB, IlvE1, KdsB, and the protein corresponding to Smc02634) were significantly affected by the addition of C₁₄-HSL and/or C_16:1 HSL (Table 2). These results are consistent with the possibility that accumulation of these sinI-dependent proteins depends on the interaction of AHLs synthesized by SinI with some AHL receptor other than ExpR (e.g., SinR or one of the other predicted AHL receptors).

Addition of C₁₄-HSL or C_16:1-HSL significantly affected the levels of only 3 of the 18 proteins identified in Table 2 as having expR-only-dependent accumulation, raising the question of which signal molecules normally interact with the ExpR receptor to affect their accumulation. As discussed below, promoter fusion studies provide evidence that most members of this expR-dependent class of proteins do show modest but significant sinI-dependent, AHL-responsive regulation when tested. Thus, it does not seem necessary to invoke the involvement of signals other than SinI AHLs at this time.

An earlier proteome study of QS in S. meliloti found that more than 100 proteins were differentially accumulated in the expR mutant strain in response to a 2- to 8-h exposure to C₁₄-HSL or 3-oxo-C_16:1-HSL during early log-phase growth (10). No overlap would be expected between the proteins with expR-dependent accumulation in Table 2 and the AHL-responsive S. meliloti proteins identified earlier in the expR strain (10). Nonetheless, three proteins were identified as QS regulated in both studies. These are the hypothetical signal peptide corresponding to Smc00242, PckA (Smc02562), and the hypothetical transmembrane protein corresponding to Smc03233. In each case, the responses to added SinI AHLs were quite different in the expR⁺ and expR mutant bacteria. For example, 3-oxo-C_16:1-HSL stimulated accumulation of the Smc00242 protein via some receptor other than ExpR (10), whereas the levels of this same protein were reduced by the presence of functional ExpR and were unaffected by added C₁₄- or C_16:1-HSL (Table 2). In further contrast, the activity of the corresponding promoter fusion was significantly increased by a functional ExpR receptor, and SinI AHLs restored activity to WT levels (Table 3). These results provide evidence that SinI AHLs mediate three different kinds or levels of QS regulation of Smc00242.

Transcriptional analysis.

Based on results from gus promoter fusion studies (Table 3), it seems likely that almost all of the genes and proteins identified in Table 2 are QS regulated in a sinI- and expR-dependent, AHL-responsive manner. Seven of the 10 fusions listed in Table 4 showed significant differences in activity in the sinI mutant background. This is consistent with sinI-dependent QS regulation but was unexpected since the accumulation of several of the corresponding proteins was not different in the sinI mutant (Tables 2 and 3). The sinI dependence of the fusions indicates that SinI AHLs contribute to QS regulation of transcription, even though this contribution may not be evident at the protein level for cells harvested in the late log phase. Only one (Smc00242) of the five fusions corresponding to proteins showing expR-only-dependent accumulation responded significantly to added C₁₄- or C_16:1-HSL. To determine whether other AHLs might be effective in stimulating these fusions, the fusions in Table 3 were tested for responses to 200 nM C₈-HSL and 200 nM 3-oxo-C_16:1-HSL. C₈-HSL had no effect on any of the fusions tested. However, the addition of 3-oxo-C_16:1-HSL fully restored the activities of the soxG2, ivdH, and mttB fusions to WT levels (not shown), indicating that SinI AHLs other than those initially tested can be effective.

Identification of signals responsible for sinI- and/or expR-dependent changes in protein accumulation.

Both GC/MS and ESI/MS/MS identified 3-oxo-C₁₄-HSL, C₁₆-HSL, 3-oxo-C₁₆-HSL, C_16:1-HSL, and 3-oxo-C_16:1-HSL as AHL QS signals made by the 8530 expR⁺ strain, but not by the sinI mutant (Table 2). It seems likely that these SinI AHLs are the signals used by S. meliloti for QS regulation of the levels of most or all of the proteins listed in Table 2. Four of these AHLs were previously identified as SinI AHLs synthesized by strain 1021 (41, 68). Those earlier studies did not detect 3-oxo-C₁₆-HSL, most likely because the media and times of harvest were different. It appears that 3-oxo-C₁₆-HSL is a novel QS signal that is unique to S. meliloti.

Based on MS analyses, 3-oxo-C_16:1-HSL accumulated to substantially (8- to 160-fold) higher concentrations in the culture supernatants than any of the other four SinI AHLs (Table 4). This suggests that 3-oxo-C_16:1-HSL could have a dominant role in regulating gene expression and protein accumulation in S. meliloti. However, when cultured in a defined glucose-nitrate medium, cultures of the expR mutant strain accumulated five times more C_16:1-HSL than 3-oxo-C_16:1-HSL (68). Thus, genetic background, culture medium, culture age, and other environmental conditions might substantially affect the relative ratios and biological importance of the different SinI AHLs. Based on GC/MS responses, the concentration of 3-oxo-C_16:1-HSL in the culture supernatants of the 8530 expR⁺ strain is estimated to be roughly 20 nM, assuming complete recovery and no masking of the mass ion signal by other compounds in the same fractions. The concentration of C_16:1-HSL was also estimated to be about 20 nM in culture supernatants of the expR mutant strain grown in the defined glucose-nitrate medium (A. Eberhard, unpublished data). At the behavioral level, the minimal concentration of C_16:1-HSL found to restore normal swarming was between 1 nM and 5 nM, which seems consistent with the estimated concentrations above.

We speculate that each of the half dozen known SinI AHLs may have separate, as well as overlapping, roles in QS regulation. In this regard, addition of C₁₄-HSL affected the accumulation of some proteins quite differently than did addition of C_16:1-HSL (Table 2). For example, accumulation of ExpE6, IdhA, DegP1, and PckA was restored to WT levels by addition of C_16:1-HSL but not by addition of C₁₄-HSL, whereas the conserved protein corresponding to Smc02156 was restored to WT levels by addition of C₁₄-HSL but not by addition of C_16:1-HSL. The levels of certain other proteins were not appreciably affected by addition of either C₁₄-HSL or C_16:1-HSL (e.g., LacE and LigE). These might be responsive to other, untested SinI AHLs. At the phenotypic level, the production of EPS by the sinI mutant was specifically restored by 200 nM C_16:1-HSL or 3-oxo-C₁₆:-HSL, but not by the other AHLs tested, while swarming was fully restored by 5 nM C_16:1-HSL or 200 nM 3-oxo-C_16:1-HSL and partially restored by 200 nM C₁₆-HSL but was not restored by the other AHLs.

Interestingly, the rather specific effects of certain SinI AHLs on these phenotypes and on the levels of the proteins mentioned above all depend on the ExpR receptor. It is not clear how ExpR is able to regulate different behaviors or the levels of different proteins in response to different SinI AHLs. One of several possibilities is that each of the half dozen known and putative AHL receptors in S. meliloti 8530 interacts preferentially with one of the SinI AHLs and forms not only homodimers to regulate specific sets of genes but also heterodimers with other receptors to regulate other sets of genes. In regard to this possibility, Ledgham et al. (35) and Ventre et al. (72) recently provided evidence that each of the three AHL receptors in P. aeruginosa (QscR, LasR, and RhlR) can form both homomultimers and heterodimers with the other two receptors.

AHL production.

It is still not clear whether the S. meliloti 8530/1021 genotype has a second AHL synthase capable of producing AHLs with short acyl side chains, as suggested earlier by Marketon et al. (41) and again by Hoang et al. (27). Some of the evidence from our studies is consistent with the possibility of a second AHL synthase, i.e., the sinI-independent peak of CepR activity (Fig. 2B) and Agrobacterium-active TLC spot e (Fig. 3). However, the substances responsible for these activities have not been identified and both could derive from a single compound. Other kinds of evidence do not support the presence of a second AHL synthase. In E. coli, SinI appears to synthesize both long-chain and short-chain AHLs (40), so there is no need to invoke a second AHL synthase to explain the secretion of short-chain AHLs. No AHLs with short acyl side chains (C₄ to C₈) were detected by GC/MS or by ESI/MS/MS in our culture extracts (Table 4). Our HPLC bioassay and TLC studies revealed the presence of compounds that have mobilities similar to, but not quite matching, those of the tested short-chain AHLs. However, in contrast to the TLC results of Marketon et al. (41), the active compounds we detected were much reduced or not detectable in extracts of the sinI mutant (Fig. 2 and 3). It is not clear whether these LasR- and Agrobacterium-active substances are AHLs, but if they are, their synthesis is dependent, either directly or indirectly, on SinI. Hoang et al. (27) mentioned preliminary results showing that Smc00714 encoded a second AHL synthase in S. meliloti 8530. The protein corresponding to Smc00714 has similarity to the HtdS AHL synthase of P. fluorescens (34). We tested two putative in-frame deletion mutants of Smc00714 and found no difference between the TLC patterns of these mutants and the 8350 strain (M. Connolly and J. B. Robinson, unpublished data). Nor did we find any evidence that acquisition of the Smc00714 gene enabled the P. putida CepR reporter to synthesize AHLs (Gao, unpublished). When a variety of transconjugants of the CepR and Agrobacterium reporter strains that contained positive clones from an S. meliloti genomic library were tested, the only transconjugants that produced detectably higher levels of AHLs were found to contain clones of the sinI gene, not the Smc00714 gene (D. Grau, M. Teplitski, and M. Gao, unpublished data). While none of these results exclude the possibility of a second AHL synthase in the 8530 strain, they do indicate the need for better genetic and chemical evidence to establish the existence of such an enzyme.

Regulation of swarming behavior.

Soto et al. (64) first observed surface swarming in S. meliloti. Their G4 WT strain did not swarm under the conditions tested, but a fadD mutant did. Our results show that swarming of the 8530 strain can be dependent on SinI- and/or ExpR-mediated QS. QS is known to regulate swarming in other bacteria (11, 15, 58), and swarming of the S. meliloti G4 fadD mutant was dependent on inoculum size (64), consistent with involvement of QS regulation. In S. meliloti 8530, the need for QS to initiate swarming colony development seems to be restricted to events after contact with a partially dried agar surface. The ability of C_16:1-HSL and other C₁₆ AHLs to stimulate swarming of the sinI mutant, but not the sinI expR mutant, indicates that swarming colony initiation on the partially dry surfaces involves these specific SinI AHLs acting as signals mediated by the ExpR receptor.

Role of QS in nodule initiation.

The sinI mutant and mutants with changes in QS-regulated genes corresponding to Smc00154, Smc00242, Smc01907, Smc02689, and Smc04330 were able to initiate few, if any, productive infections within the first 8 to 36 h after inoculation (Table 5). The delayed-nodulation phenotype of the sinI mutant leads to the conclusion that sinI-dependent QS regulation is not required for nodule initiation but does contribute in various ways to improving the rate of infection initiation or the efficiency with which the bacterium can initiate productive infections. Although not examined here, QS is also likely to contribute significantly to the efficiency of symbiotic functioning within developing nodules, as suggested by sinI-dependent regulation of N regulatory proteins (10), fix genes (27), Mo cofactor synthesis protein (67), and PckA (Table 2).

The delayed-nodulation phenotype of five of the nine single-crossover mutants (Table 5) provides evidence that the corresponding genes also contribute importantly to the rate or efficiency of nodule initiation. Presumably, QS helps to provide appropriate regulation of the corresponding genes and proteins during early interactions with the host plant. The delayed-nodulation phenotype of the Smc00154 mutant, presumably defective in glutathione reductase activity, is consistent with the delayed-nodulation phenotype recently reported by Harrison et al. (24) for a gshB glutathione synthesis mutant.

In competition with WT strains, delayed-nodulation mutants have been observed to form significantly fewer nodules (29, 62). Thus, SinI- and/or ExpR-mediated QS could be of considerable importance to the bacterium's competitiveness in natural rhizosphere interactions. This view is consistent with the observation here that 100-fold dilution of the WT inoculum resulted in delayed nodule initiation comparable to that of the sinI mutant.