Sinorhizobium meliloti Mutants Lacking Phosphotransferase System Enzyme HPr or EIIA Are Altered in Diverse Processes, Including Carbon Metabolism, Cobalt Requirements, and Succinoglycan Production

Sinorhizobium meliloti is a member of the Alphaproteobacteriathat fixes nitrogen when it is in a symbiotic relationship.Genes for an incomplete phosphotransferase system (PTS) havebeen found in the genome of S. meliloti. The genes present codefor Hpr and ManX (an EIIA^Man-type enzyme). HPr and EIIA regulatecarbon utilization in other bacteria. hpr and manX in-framedeletion mutants exhibited altered carbon metabolism and otherphenotypes. Loss of HPr resulted in partial relief of succinate-mediatedcatabolite repression, extreme sensitivity to cobalt limitation,rapid die-off during stationary phase, and altered succinoglycanproduction. Loss of ManX decreased expression of melA-agp andlac, the operons needed for utilization of ${alpha}$ - and β-galactosides,slowed growth on diverse carbon sources, and enhanced accumulationof high-molecular-weight succinoglycan. A strain with both hprand manX deletions exhibited phenotypes similar to those ofthe strain with a single hpr deletion. Despite these strongphenotypes, deletion mutants exhibited wild-type nodulationand nitrogen fixation when they were inoculated onto Medicagosativa. The results show that HPr and ManX (EIIA^Man) are involvedin more than carbon regulation in S. meliloti and suggest thatthe phenotypes observed occur due to activity of HPr or oneof its phosphorylated forms.

INTRODUCTION

Top

INTRODUCTION

Bacteria belonging to the genera Sinorhizobium, Rhizobium, andBradyrhizobium are members of the Alphaproteobacteria, a groupof organisms that contains many intracellular symbionts andpathogens. Sinorhizobium meliloti can grow in soil as a free-livingorganism or as a nitrogen-fixing symbiont inside root nodulesof alfalfa and several other plants belonging to the familyLeguminosae (9, 19, 34, 46, 53, 66).

Free-living rhizobia are like many heterotrophic bacteria inthat they can utilize a large number of compounds as sourcesof carbon, including sugars, amino acids, and tricarboxylicacid cycle intermediates. Although S. meliloti can utilize avariety of compounds, succinate and other C₄-dicarboxylic acidsplay an especially important role in metabolism during boththe free-living and symbiotic states. For example, C₄-dicarboxylicacids are used to fuel and provide reducing equivalents fornitrogen fixation by bacteroids (7, 28, 30, 62). Mutants unableto transport C₄-dicarboxylic acids are able to nodulate plants,but the bacteroids do not fix nitrogen (26, 63). In additionto its role in symbiotic nitrogen fixation, succinate is a favoredcarbon and energy source for free-living rhizobia. It supportsone of the highest rates of growth in laboratory medium andalso exerts catabolite repression on the utilization of othercarbon sources. In S. meliloti, succinate is used in preferenceto the secondary carbon sources glucose (36), fructose (36),galactose (36), lactose (39, 69), myo-inositol (57), and severalpentoses and polyols (36). This preference for succinate canbe manifested as diauxie when S. meliloti is grown on succinateplus a secondary carbon source (36). When S. meliloti is grownon succinate plus lactose, succinate downregulates transcriptionof the lac operon (39, 69). It has been shown that succinatealso prevents the accumulation of lactose and raffinose by inducerexclusion or inducer expulsion (11).

Glucose-mediated catabolite repression in Escherichia coli isprobably the best-studied system of global regulation of carbonutilization (Fig. 1). It consists of two mechanisms of control:inducer exclusion and gene regulation by cyclic AMP (cAMP) (12,32, 38, 40, 43). The glucose phosphotransferase system (PTS)transports and phosphorylates glucose, which makes it readyfor metabolism. EIIA^Glc is phosphorylated by HPr and donatesits phosphate to glucose molecules transported into the cell;thus, EIIA^Glc exists mainly in the unphosphorylated state whenglucose is being utilized (Fig. 1). Unphosphorylated EIIA^Glcinhibits transporters required for the uptake of secondary carbonsources when these transporters are bound with ligand. Whenglucose is absent, EIIA^Glc-P predominates and activates adenylatecyclase, the enzyme responsible for cAMP synthesis. In conjunctionwith the Crp protein, cAMP activates transcription of genesneeded to catabolize secondary carbon sources (65, 67). Thus,the EIIA^Glc enzyme plays two important roles in carbon metabolismin E. coli: it prevents secondary carbon sources from enteringthe cell when a primary carbon source is present, and it participatesin the induction of the genes responsible for catabolism ofsecondary carbon sources when the primary carbon source is absent,by activating adenylate cyclase.

View larger version (16K):
[in this window]
[in a new window]

FIG. 1. Glucose-mediated regulation of secondary carbon source (lactose) metabolism in E. coli (A) and B. subtilis (B).

Gram-positive bacteria regulate carbon utilization through thePTS enzyme HPr rather than through EIIA (12, 42, 60, 71) (Fig.1). Bacillus subtilis and other low-G+C-content gram-positivebacteria possess a complete PTS for uptake and utilization ofcarbohydrates. As it does in gram-negative bacteria, phosphateflows from phosphoenolpyruvate to the transported carbohydrate.During this process HPr is phosphorylated at the histidine-15residue by PTS enzyme I. Additionally, an HPr kinase/phosphatase(HPrK) can phosphorylate the serine-46 residue of HPr usingeither ATP or pyrophosphate (49, 54). HPr-Ser46-P is a poorsubstrate for additional phosphorylation at HPr-His15. Highlevels of HPr-Ser46-P accumulate under energy-replete conditionsdue to high levels of ATP and PP_i and high HPrK activity. HPr-Ser46-Pcan associate with the LacI-type transcriptional regulator CcpA,and the complex can activate or repress transcription of numerousgenes (2, 60, 71). The commonly repressed genes are associatedwith the catabolism of secondary carbon sources. There is evidencethat in some gram-positive bacteria HPr-Ser46-P interacts withsugar transport systems, resulting in inducer exclusion or inducerexpulsion (51, 71). Thus, the ratio of HPr-Ser46-P to HPr-His15-Pis important in regulating carbon metabolism and may also mediateinducer exclusion or expulsion (8).

The PTS not only regulates utilization of sugars but also regulatesother bacterial processes, often in a manner linked to sugarutilization. For example, in Listeria monocytogenes severalvirulence genes are repressed in the presence of glucose bya CcpA-independent mechanism (21, 29). In contrast, in bacteriabelonging to the group A streptococci, expression of the virulenceregulator Mga is stimulated by CcpA. In addition to virulencegenes, Mga regulates many metabolic genes, thereby linking metabolismto virulence (1, 35). The Bacillus anthracis virulence transcriptionfactor AtxA is also controlled in a carbon-source-dependentmanner. There is evidence that phosphorylation by HPr and EIIBmodulates the activity of AtxA (68).

In Pseudomonas putida the PTS orthologs EI^Ntr, NPr, and EIIA^Ntrare distinct from homologs involved in carbohydrate transportand are usually involved in nitrogen regulatory networks (20).In P. putida they are involved in the control of polyhydroxyalkanoatesynthesis, and in E. coli EIIA^Ntr regulates the potassium transporterTrkA and affects growth on organic nitrogen (44, 58, 70).

It appears that S. meliloti is missing genes required to assemblea complete PTS. The genome of S. meliloti contains smc02754(hpr), a gene similar to ptsH (coding for HPr in E. coli) andsmc02753 (manX) (coding a protein similar to the N-terminalhalf of E. coli ManX, which contains an EIIA^Man PTS domain).However, the S. meliloti genome has no genes encoding proteinssimilar to the EIIB or EIIC components of PTS transport systems(4, 14, 25). The genes encoding S. meliloti EI^Ntr and EIIA^Ntr(smc02437 and smc01141, respectively) are at a chromosomal siteseparate from hpr and manX.

The manX and hpr genes are adjacent to each other on the S.meliloti chromosome, are oriented the same direction, and arelikely cotranscribed because the start codon of hpr, the downstreamgene, overlaps the stop codon of the upstream manX gene. Importantly,S. meliloti HPr contains both the critical histidine that isphosphorylated in other bacteria during carbohydrate transportand the serine residue that is phosphorylated by HPrK in gram-positivebacteria during catabolite control (22, 72). In addition, theS. meliloti EIIA homolog, ManX, contains the histidine residuethat is phosphorylated by HPr (24). Upstream of the S. melilotimanX gene is an open reading frame (smc02752) that encodes aprotein similar to HPrK.

No genes encoding proteins similar to the gram-positive transcriptionalregulator CcpA, which also contain known Hpr-binding motifs,are present in the S. meliloti genome (41; D. J. Gage, unpublishedresults). A search of the genome for genes encoding Crp-likeproteins revealed genes encoding two potential proteins belongingto the Crp/FNR family of regulators. Members of this familyare versatile transcription factors that are involved in processessuch as nitrogen fixation, photosynthesis, sensing of oxidativestress, metabolism, and catabolite repression.

It is clear that succinate-mediated catabolite repression (SMCR)in S. meliloti is mechanistically different (at least partially)from catabolite repression in the well-studied models describedabove. This is because preferred carbon sources in the modelbacteria are transported via the PTS. In contrast, succinateenters S. meliloti via a very different system, the DctA permease.This paper describes results showing that some of the PTS genesneeded for PTS-dependent catabolite repression in well-studiedmodels affect SMCR and other aspects of carbon metabolism inS. meliloti. The results also show that in S. meliloti the PTSproteins are involved in regulation of other cellular processesas well.

MATERIALS AND METHODS

Top

MATERIALS AND METHODS

Media, strains, and plasmids. S. meliloti strain Rm1021 was used as the parent for constructionof all mutants. S. meliloti strains were grown in TY mediumor in M9 mineral salts medium supplemented with biotin (5 µg/ml),cobalt (5 ng/ml CoCl₂), and carbon sources at the concentrationsindicated below. E. coli XL1B MRF' (Stratagene) was the hostused for construction of all plasmids and was grown in LB mediumsupplemented with ampicillin, isopropyl-β-D-thiogalactopyranoside(IPTG) (48 µg/ml), and 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside(X-Gal) (40 µg/ml). Media were supplemented with antibioticsat the following concentrations when necessary: streptomycin,500 µg/ml; ampicillin, 100 µg/ml; spectinomycin,100 µg/ml; tetracycline, 10 µg/ml; and gentamicin,30 or 10 µg/ml.

Construction of hpr and manX in-frame deletion mutants. For construction of an unmarked, in-frame deletion of the hprgene (smc02754), the 5' end of hpr and the upstream sequencewere amplified from strain Rm1021 genomic DNA using primers192 (5'-GAGCTCGACGGCGTCATCATCCTGA-3') and 193 (5'-TCTAGATGCATGCGCGTCAGAGCCGTGTC-3').In the same manner, a piece of DNA including the 3' end of hprand downstream sequences was amplified using primers 194 (5'-TCTAGATGCATGCGGTTCGCGACAGGTTCGG-3')and 195 (5'-CTCGAGCATTTCCGTTTCGGCGA-3'). The primers were designedso that when the fragments were cut with NsiI and ligated together,they would remain in frame and a central 228-bp region of thehpr gene would be left out. After the fragments were ligated,the product was cloned into pJQ200SK, a suicide vector thatcarries sacB and a gene for gentamicin resistance, creatingplasmid pRB88. After confirmation of the in-frame deletion bysequencing, plasmid pRB88 was moved into strain Rm1021 by triparentalmating. The suicide plasmid was integrated by homologous recombinationinto the chromosomal hpr gene, resulting in strain RB108. Thisstrain was then grown in TY broth without antibiotics to allowa second recombination that would eliminate the suicide vectorand leave the ${Delta}$ hpr allele in the chromosome. The culture wasspread plated onto TY medium with 5% sucrose, and sucrose-resistantisolates were screened for gentamicin sensitivity to selectfor isolates that had lost the suicide vector. Isolates werescreened by PCR for the presence of the ${Delta}$ hpr allele. One strain,RB111 ( ${Delta}$ hpr) having an in-frame hpr deletion, was purified andused for further study. A manX (smc02753) in-frame deletionmutant was constructed in a similar manner. The primers usedto construct the manX in-frame deletion mutant were primers186 (5'-ATCTAGAGCGGGATGAGCAAAAGTGTTT-3') and 187 (5'-TGATCATGCATGCCGAAACTCATCCGCC-3')for amplification of the 5' fragment and primers 188 (5'-TGATCATGCATGCAGGCGTGCGCGG-3')and 189 (5'-CTGCAGGCCGCGTTTGTTGACG-3') for amplification ofthe 3' fragment. The primers were designed so that when thefragments were cut with SphI and ligated together, they wouldremain in frame and a central 243-bp region of the manX genewould be left out. The final S. meliloti strain having the unmarkeddeletion in manX was designated RB105. An hpr-manX double mutantwas constructed using primers 186 and 187 for amplificationof the 5' end of manX and primers 194 and 195 for amplificationof the 3' end of hpr. The strain with the deleted manX-hpr fragmentwas designated CAP32.

Complementation of deletion mutants. The unmarked deletion mutants were complemented by insertinga single copy of each gene into the chromosome under the controlof its native promoter. The DNA fragment for the hpr-manX operon,including the intergenic region upstream of manX, was amplifiedby PCR and cloned into pGEM-T Easy (Promega), and using thisconstruct plasmids were constructed by "outward PCR" (by amplifyingthe plasmid with primers directed outward from the boundariesof the fragment to be deleted and then ligating the PCR product).One resulting plasmid had an in-frame deletion of the manX geneand an intact hpr gene (pDG127). The other had an in-frame deletionof the hpr gene and an intact manX gene (pDG126). The promotergene sequences from pDG127 and pDG126 were cloned into pCAP77as EcoRI fragments, creating plasmids pCAP81 (hpr) and pCAP80(manX), respectively. pCAP77 is a pMB439-based suicide plasmid(5) that contains a fragment of the S. meliloti smc02324 genefollowed by a trp terminator upstream of a multiple cloningsite. This plasmid recombines into smc02324 and renders cellsunable to grow on rhamnose as the sole carbon source, and itinserts into the chromosome any gene that has been cloned intothe multiple cloning site (C. Arango Pinedo and D. J. Gage,unpublished). Plasmids pCAP80 and pCAP81 were moved into the ${Delta}$ manX and ${Delta}$ hpr mutant strains, respectively. Insertion into thesmc02324 gene was confirmed by the inability of the strainsto grow on M9 medium plates containing rhamnose and by the resultsof PCR screens with appropriate primers.

Construction of a melA promoter reporter plasmid. To construct a reporter plasmid for the melA promoter, a 338-bpfragment containing the 5' end of agpT (21 bp), the melA promoterregion (195 bp), and the 5' end of melA (122 bp) (10) was amplifiedfrom plasmid pRB27 with primers 180-RB27 (5'-CCGCTCGAGCCATCGATTGGCGGCCGCTCTAGAACTA-3';XhoI restriction site underlined) and 181-melA (5'-ACGCGTCGACTCAGATCCGTCAACGCGAA-3';SalI restriction site underlined). The product was cut withXhoI and SalI and cloned into the SalI site of pDG64, whichis between the trp terminator and gfp(mut1), creating pCAP03.pDG64 is the same as pDG65 described by Rosado and Gage (64),but pDG64 contains gfp(mut1) instead of gfp(mut3) (16, 23).The terminator, promoter, and gfp were excised from pCAP03 withHindIII and KpnI and cloned into plasmid pMB393 to obtain pCAP11.pCAP11 was electroporated into strain Rm1021. Green fluorescentprotein (GFP) expression patterns during growth on raffinoseand melibiose ( ${alpha}$ -galactosides), on succinate, and on succinateplus ${alpha}$ -galactosides confirmed that all PmelA regulatory sequenceswere present and functional in the reporter plasmid.

Growth curves and growth rates. Growth curves were obtained by removing 100-µl samplesfrom liquid cultures growing in flasks with constant shakingat 30°C and determining the optical density at 595 nm (OD₅₉₅)with a microwell plate reader (Microplate Reader 550; Bio-Rad).Alternatively, growth curves were obtained using 48-well plates(culture volume, 200 µl) and a Biotek Synergy HT-1 multidetectionmicroplate reader at 30°C with automatic reading and shakingevery 10 min. Growth rates were calculated from the data usinga least-squares fit method and data points for the exponentialphase of growth. Cultures used for growth curve analyses wereinoculated to obtain an initial OD₅₉₅ of 0.005 using startercultures in M9 medium containing glycerol in the exponentialphase of growth that had been rinsed in carbon-free M9 salts.

GFP fluorescence measurement. GFP fluorescence was measured while 200 µl cultures weregrowing in 48-well plates at 30°C with shaking (excitationwavelength, 480 nm; emission wavelength, 520 nm). Carbon sourceswere used at a concentration of 0.4% (succinate alone or raffinosealone) or (when succinate and raffinose were supplied together)at concentrations of 0.05% (succinate) and 0.1% (raffinose).Specific fluorescence was calculated by dividing the background-correctedfluorescence by the OD₅₉₅ of the sample.

Determination of β-galactosidase activity. Strains were grown in tubes with succinate or lactose as thesole carbon source (0.4%) or with succinate plus lactose (0.2and 0.1%, respectively). The optical density was monitored,and samples were retrieved during the early exponential to mid-exponentialphase. Culture samples were pelleted and frozen. At the timeof the assay, the samples were thawed and resuspended in Z-buffer.β-Galactosidase assays were conducted as previously described(50). OD₄₁₅ values were used to calculate the slope of the linearportion of the curve by a least-squares fit method. The valueswere normalized by dividing by the cell mass of the sample (initialOD₄₁₅) to obtain arbitrary β-galactosidase units.

Evaluation of succinoglycan production. Succinoglycan production and low-molecular-weight succinoglycanproduction were evaluated as previously described (75). Briefly,strains were pregrown in M9 broth with glycerol as the carbonsource, rinsed, and resuspended to an optical density of 0.03,and 5 µl of each cell suspension was placed on MGS platesthat contained calcofluor (0.02%). Succinoglycan productionwas visualized as bright fluorescence of the bacterial massunder UV light. Production of low-molecular-weight succinoglycanwas indicated by the presence of a halo of fluorescence aroundthe bacterial mass under UV light.

Nodulation assays. Alfalfa (Medicago sativa) seeds were sterilized and sproutedas described previously (27). Briefly, seedlings were placedon Nod3 agar slants in 18-mm glass tubes and inoculated with100 µl of a suspension of S. meliloti. Suspensions wereprepared by washing 1 ml of a TY medium-grown culture with 1ml of Nod3 medium and resuspending the pellet in 10 ml of Nod3medium. Plant growth tubes were loosely capped and incubatedin a growth chamber with a cycle consisting of 16 h of lightand 8 h of darkness at 26°C. Thirty-six days after inoculationnodules were counted, and plant shoots were cut above the cotyledons,dried, and weighed.

RESULTS

Top

RESULTS

Construction of in-frame hpr and manX deletions. HPr and EIIA-type proteins have been shown to be involved incatabolite repression control in both high-G+C-content gram-positiveand enteric gram-negative bacteria and could be expected tobe involved in catabolite repression in S. meliloti as well.Therefore, we constructed in-frame, unmarked deletions to investigatethe role of these two proteins in SMCR in S. meliloti.

Strain RB111 ( ${Delta}$ hpr) is especially sensitive to cobalt limitation. S. meliloti strain Rm1021 requires cobalt for proper growth(17, 73). Cobalt deficiency results in a decrease in the B₁₂coenzyme content in the cells and consequently a reduction inthe growth rate (17). The effect of cobalt limitation is notcommonly noted in the laboratory, even during growth in minimalmedium lacking added cobalt, because cobalt impurities in mediumcomponents are enough to support growth (17). Strain RB111 ( ${Delta}$ hpr)frequently exhibited premature growth arrest in minimal medium.We investigated if adding cobalt could alleviate the inconsistentgrowth of the ${Delta}$ hpr strain. Strains Rm1021 (wild type) and RB111( ${Delta}$ hpr) were grown to exponential phase in a no-cobalt M9 minimalmedium with glycerol as the carbon source. Cultures were dilutedinto fresh minimal medium with and without added cobalt (5 ng/mlCoCl₂). After repeated dilution in medium without added cobalt,the final yield of strain Rm1021 (wild type) was reduced around20%, but usually there was not a significant reduction in thegrowth rate compared to that of the cobalt-amended cultures.For the ${Delta}$ hpr strain variability in growth was observed withoutcobalt, and often the growth rate was lower or there was growtharrest in early exponential phase resulting a 25 to 50% decreasein the growth yield. A typical growth curve demonstrating thelower yields of the ${Delta}$ hpr strain in medium without cobalt is shownin Fig. 2. When cobalt was added early in stationary phase,growth resumed, and the culture yield was the same as the cultureyield of the cobalt-amended culture, confirming that the effectwas due to cobalt limitation (Fig. 2). Since the enzymes involvedin methionine synthesis in S. meliloti require B₁₂ cofactorsand thus cobalt for proper function (6, 48, 73), we investigatedthe ability of methionine to rescue cobalt-limited cells. Cellsexhibited the wild-type growth rate and yield when methioninewas supplied (data not shown). Because of the dramatic effectof cobalt on the growth phenotype of strain RB111 ( ${Delta}$ hpr), cobaltwas added to all minimal media used, unless otherwise noted.

View larger version (25K):
[in this window]
[in a new window]

FIG. 2. Growth of wild-type strain Rm1021 and strain RB111 ( ${Delta}$ hpr) in M9 minimal medium with and without added cobalt (5 ng/ml) following two dilutions in cobalt-free medium. The arrow indicates when cobalt was added to the cobalt-deficient cultures. The data are representative of three independent experiments. w.t., wild type.

${Delta}$ hpr mutation affects diauxic growth on succinate plus galactosides. Catabolite repression can be manifested as diauxic growth whenprimary and secondary carbon sources are provided to bacteria.To investigate if the lack of HPr affected SMCR phenotypes,strains Rm1021 (wild type) and RB111 ( ${Delta}$ hpr) were grown in liquidM9 medium with succinate (0.05%) plus either raffinose or lactose(0.1%) as carbon sources. When grown in medium containing succinateplus raffinose, the wild-type strain utilized succinate first,and growth lagged for about 30 h before growth on raffinosestarted (Fig. 3A). Strain RB111 ( ${Delta}$ hpr) exhibited a shorter lagphase (only 15 h) and resumed growth sooner than the wild type.When lactose (0.1%) was provided as the secondary carbon source,the wild type exhibited a very short diauxic lag (3 to 4 h),while strain RB111 ( ${Delta}$ hpr) showed almost no lag, and diauxie wasevident only because of the abrupt change in the growth ratethat resulted from switching from succinate utilization to lactoseutilization (Fig. 3C). These results showed that cataboliterepression was weaker in the ${Delta}$ hpr strain and suggest that theHPr protein has a role in succinate-mediated catabolite repression.

View larger version (16K):
[in this window]
[in a new window]

FIG. 3. Diauxie phenotypes of the ${Delta}$ hpr strain (A and C) and the ${Delta}$ manX strain (B and D). Data for the corresponding complemented strains are also shown. (Inset in panel C) Detail of the diauxic curve showing no diauxic lag in the ${Delta}$ hpr strain. Strains were grown in M9 minimal medium with succinate (0.05%) plus raffinose (0.1%) (A and B) or with succinate plus lactose (0.1%) (C and D). The data are representative of three independent experiments. Curves were time shifted so that the points of succinate exhaustion in each panel were aligned. The arrows indicate the OD₅₉₅ attained by strain Rm1021 when it was grown on 0.05% succinate (0.06 ± 0.003 [average ± standard error; n = 4]). w.t., wild type.

The lower level of repression in the ${Delta}$ hpr strain could have beendue to constitutive or higher levels of expression of the melA-agpand lac operons, which code for proteins necessary for utilizationof ${alpha}$ - and β-galactosides, respectively.

To test for constitutive or elevated levels of transcriptionfrom the melA-agp operon, the expression from a PmelA::gfp fusionplasmid (pCAP11) in the ${Delta}$ hpr strain was measured and found tobe comparable to the wild-type expression when the strains weregrown in M9 medium containing raffinose alone, succinate alone,or succinate plus raffinose (Fig. 4A). These results showedthat the ${Delta}$ hpr mutation did not result in constitutive or elevatedlevels of melA-agp expression in these media.

View larger version (31K):
[in this window]
[in a new window]

FIG. 4. (A) Expression of the melA-agp operon as monitored with the PmelA::gfp fusion in plasmid pCAP11. Samples from mid-exponential-phase growth were used, and specific fluorescence was calculated. For dual-carbon-source growth, samples from early to mid-exponential phase during growth on succinate were used. (B) β-Galactosidase activity in PTS mutant strains. Samples from mid-exponential growth on the indicated carbon sources were assayed to determine the β-galactosidase activity. For dual-carbon-source growth, samples from the mid-exponential phase were used, when succinate was still being utilized. The bars indicate the averages of three to eight independent experiments. The error bars indicate standard errors. An asterisk indicates that a value is statistically different from the wild-type value (P < 0.01). w.t., wild type.

Expression of the lac operon was investigated by measuring endogenousβ-galactosidase levels in strains Rm1021 (wild type) andRB111 ( ${Delta}$ hpr) (Fig. 4B). The levels of activity of β-galactosidasein strain RB111 ( ${Delta}$ hpr) were equivalent to the levels in the wild-typestrain during growth on lactose alone and on lactose plus succinate.However, the activity of β-galactosidase in cultures growingon succinate alone was slightly, but significantly, higher inthe ${Delta}$ hpr strain than in the wild-type strain. In all cases theactivity levels were 5- to 20-fold lower than the fully inducedlevels, showing that SMCR was not relieved in strain RB111 ( ${Delta}$ hpr),as succinate still efficiently suppressed expression of thelac operon in the presence of lactose.

Thus, the ${Delta}$ hpr mutation did not result in high-level expressionof the lac or melA-agp operon in the presence of single carbonsources, nor did it result in relief of succinate-mediated repressionin medium containing succinate plus lactose or raffinose.

${Delta}$ manX mutation affects diauxic growth on succinate plus galactosides. The growth of strain RB105 ( ${Delta}$ manX) on succinate (0.05%) plusraffinose (0.1%) or on succinate (0.05%) plus lactose (0.1%)was compared to the growth of strain Rm1021 (wild type). Duringgrowth on succinate plus raffinose there was earlier releasefrom succinate repression for the ${Delta}$ manX strain, manifested bya slightly shorter diauxic lag (Fig. 3B). This phenomenon wasnot observed during growth on succinate plus lactose (Fig. 3D).In this case, strain RB105 ( ${Delta}$ manX) lagged longer than the wildtype before it started to grow on lactose. The absence of ManX(EIIA^Man) may have altered regulation of raffinose and lactoseutilization and not SMCR directly. This was tested by measuringmelA-agp expression and β-galactosidase levels during growth(Fig. 4). The level of expression of the melA-agp operon duringgrowth on raffinose, as measured by using the PmelA::gfp reporterfusion (pCAP11), was twofold lower in the ${Delta}$ manX strain than instrain Rm1021 (wild type) (Fig. 4A). The expression of the PmelA::gfpreporter fusion during growth on succinate and during growthon succinate plus raffinose was not significantly differentfrom the expression in strain Rm1021 (wild type) (Fig. 4A).These experiments showed that the ${Delta}$ manX mutation resulted inlower levels of melA-agp expression in the presence of raffinose,conditions in which the genes should be fully induced, and ithad no effect on melA-agp promoter activity under conditionsin which full expression was not expected.

Levels of endogenous β-galactosidase were measured duringgrowth on lactose, on succinate, and on succinate plus lactose,and the results paralleled the results for melA-agp expression(Fig. 4B). The enzyme levels were twofold lower in the ${Delta}$ manXstrain during growth on lactose but were not significantly differentfrom the wild-type levels during growth on succinate or on succinateplus lactose. These results indicate that expression of boththe lac and melA-agp genes was lower in the absence of ManX(EIIA^Man) under inducing conditions.

hpr deletion affects growth on specific carbon sources. Strain RB111 ( ${Delta}$ hpr) showed alterations in diauxic growth. Todetermine if this strain also showed altered growth on singlecarbon sources, the growth rates of strains Rm1021 (wild type)and RB111 ( ${Delta}$ hpr) in M9 minimal medium with succinate, raffinose,lactose, maltose, glucose, and glycerol as carbon sources weredetermined (Table 1).

View this table:
[in this window]
[in a new window]

TABLE 1. Growth rate constants in M9 minimal media containing cobalt or in rich medium

Strain RB111 ( ${Delta}$ hpr) exhibited wild-type growth rates in mediumcontaining succinate, glucose, and glycerol, but the growthrates on raffinose and maltose were significantly lower. Growthof the ${Delta}$ hpr strain on lactose was noticeably slower than growthof the wild type, but the difference was not statistically significantbecause of high experimental variation. Strains Rm1021 (wildtype) and RB111 ( ${Delta}$ hpr) exhibited comparable growth rates on rich,complex TY medium (Table 1). Thus, the ${Delta}$ hpr strain had lowergrowth rates on carbon sources that are used as secondary sourcesafter succinate or glucose.

manX deletion affects growth rates on single carbon sources. Since the induced levels of β-galactosidase and melA-agpexpression were lower in the ${Delta}$ manX strain, we investigated theeffect of these lower levels on growth rates in M9 minimal mediumwith various single carbon sources. With all carbon sourcestested (succinate, glycerol, lactose, raffinose, maltose andglucose) the growth rates of strain RB105 ( ${Delta}$ manX) were lowerthan the growth rates of strain Rm1021 (wild type) (Table 1).The growth rate of the deletion mutant in a rich medium (TYmedium) was also lower. The low growth rates with many carbonsources, along with the low levels of expression of catabolicand transport genes needed for utilization of at least somecarbon sources, suggest that ManX (EIIA^Man) may have a generalrole in the regulation of carbon metabolism or in other importantprocesses that impact the growth rate. The slow growth of the ${Delta}$ manX mutant is not likely to be the cause of its low levelsof melA::gfp and β-galactosidase gene expression becausethe ${Delta}$ hpr mutant, which grew as slowly in raffinose and lactose(Table 1), did not show decreased expression of melA::gfp orβ-galactosidase (Fig. 4). In fact, in the ${Delta}$ hpr mutant thelevel of expression of these genes is slightly higher than thatin strain Rm1021 (wild type). Therefore, slow growth, by itself,is not sufficient to explain the altered expression seen inthe ${Delta}$ manX mutant.

hpr deletion results in poor survival during stationary phase. The stationary-phase survival of strains Rm1021 (wild type)and RB111 ( ${Delta}$ hpr) was evaluated in four different media: richmedium (TY medium), M9 medium containing succinate, M9 mediumcontaining glycerol, and M9 medium containing glycerol pluscobalt. Die-off in TY medium cultures started as soon as thestationary phase was reached for both the wild type and ${Delta}$ hprstrains. The number of viable wild-type cells in the TY mediumculture decreased by 1 order of magnitude and then stabilized(Fig. 5A). In TY medium, the number of strain RB111 ( ${Delta}$ hpr) cellsdecreased by 3 orders of magnitude, and then recovery was observed,perhaps resulting from the growth of ${Delta}$ hpr suppressor mutantsin the culture (Fig. 5A).

View larger version (25K):
[in this window]
[in a new window]

FIG. 5. Survival of strains during stationary phase. (A) TY medium. (B) M9 minimal medium plus glycerol (0.4%) without cobalt. (C) M9 minimal medium plus glycerol (0.4%) with cobalt (5 ng/ml). The data are representative of three independent experiments. w.t., wild type.

In cobalt-deficient M9 media, strain Rm1021 (wild type) reachedthe stationary phase, and then the viable counts slowly decreasedtwofold during the experiment. Strain RB111 ( ${Delta}$ hpr) exhibitedgrowth arrest in early exponential phase, which is typical ofcobalt limitation for this strain, and die-off started at theonset of this arrest (Fig. 5B). In contrast to the results forstrain Rm1021, the viable counts in the strain RB111 ( ${Delta}$ hpr) culturedecreased 2 to 4 orders of magnitude. The effect of hpr deletionon stationary-phase survival was not as striking when cobaltwas added to the minimal medium. The number of cells was reducedby 1 order of magnitude for strain RB111 ( ${Delta}$ hpr) and by 50% forRm1021 (wild type) (Fig. 5C). In contrast to these findings,deletion of manX did not affect the die-off in TY or minimalmedium (Fig. 5A and data not shown).

Effects of the ${Delta}$ hpr and ${Delta}$ manX mutations on succinoglycan production and root nodule formation. Expression of succinoglycan, the most abundant exopolysaccharidein S. meliloti strain Rm1021, can be detected by plating bacteriaon MGS plates with calcofluor, which fluoresces when it bindsto succinoglycan (45). Fluorescence in strain RB111 ( ${Delta}$ hpr) wasapparent sooner than fluorescence in strain Rm1021 (wild type)(Fig. 6). The mutant strain also exhibited a fluorescent halo1 day before the wild type exhibited a halo, which indicatedthat there was earlier production of low-molecular-weight succinoglycan(75).

View larger version (114K):
[in this window]
[in a new window]

FIG. 6. Production of succinoglycan on MGS plates with calcofluor. Strains were grown in M9 broth with glycerol as the carbon source and rinsed before plating. All photographs were taken with an exposure time of 0.1 s and identical camera settings, and the images were not level or contrast adjusted. The arrows indicate the diffuse fluorescence in the low-molecular-weight succinoglycan halos. The nondiffuse rings around all of the colonies resulted from preferential cell growth at the edge of the spot (arrowhead) The fluorescence inside the spots and in the nondiffuse rings resulted from accumulated high-molecular-weight succinoglycan.

When strain RB105 ( ${Delta}$ manX) was plated on MGS plates with calcofluor,it exhibited a calcofluor-bright phenotype compared to strainRm1021 (wild type) after day 1, indicating that there was enhancedaccumulation of succinoglycan (Fig. 6). Production of the low-molecular-weightsuccinoglycan halo in the ${Delta}$ manX strain was delayed, and the halowas smaller than that observed with strain Rm1021 (wild type).

Some S. meliloti mutants with altered succinoglycan synthesisare symbiotically defective, typically forming nodules thatdo not fix nitrogen (15, 55). Both strain RB111 ( ${Delta}$ hpr) and strainRB105 ( ${Delta}$ manX) elicited nodules on alfalfa grown in nitrogen-freemedium, and the numbers were not different from the numbersobserved with the wild type (Fig. 7). The nodules were functional,as judged from the presence of leghemoglobin, by the shoot weightof plants, and by the normal appearance of the internal structureof the nodules (Fig. 7 and data not shown).

View larger version (41K):
[in this window]
[in a new window]

FIG. 7. Symbiotic capabilities of PTS mutant strains with M. sativa (alfalfa) plants: average number of nodules per plant (A), weight of plant shoots (B), and percentage of nodules that were pink (C). Solid bars, Wild-type strain Rm1021; bars with horizontal lines, ${Delta}$ hpr strain RB111; bars with diagonal lines, ${Delta}$ manX strain RB105; open bars, uninoculated control. The error bars indicate standard errors. An asterisk indicates that a value is significantly different from the wild-type value (P < 0.00001).

Phenotypes of a ${Delta}$ manX ${Delta}$ hpr strain are the same as those of the ${Delta}$ hpr mutant. An in-frame deletion that removed both manX and hpr was constructed,and the phenotypes were examined to investigate epistatic relationshipsbetween the mutations. Strain CAP32 ( ${Delta}$ manX ${Delta}$ hpr) was evaluatedfor the SMCR phenotype on succinate plus raffinose and on succinateplus lactose, for growth rates on different carbon sources,for enzyme expression and activity levels, for sensitivity tocobalt limitation, for survival in stationary phase, for succinoglycanproduction, and for symbiotic capabilities. In all of theseexperiments the ${Delta}$ manX ${Delta}$ hpr strain exhibited phenotypes similarto those of the ${Delta}$ hpr strain (Fig. 4 to 6 and Table 1; data notshown). This suggests that the phenotypes associated with themanX deletion are mediated by HPr.

Complementation of the hpr and manX deletions. To confirm that the observed phenotypes of the ${Delta}$ hpr and ${Delta}$ manXstrains were a result of the gene deletions rather than secondarymutations, a single wild-type copy of each gene under the controlof its native promoter was inserted into the rhamnose operonon the chromosome of strains RB111 ( ${Delta}$ hpr) and RB105 ( ${Delta}$ manX). Thewild-type genes complemented all phenotypes tested (Fig. 3 anddata not shown), confirming that the altered phenotypes wereindeed due to deletion of the hpr and manX genes.

DISCUSSION

Top

DISCUSSION

S. meliloti lacks the EII components EIIB and EIIC, as wellas a classical PTS EI. It does have genes coding for HPr, ManX(EIIA^Man), EIIA^Ntr, and EI^Ntr. The last two proteins are partof an orthologous PTS system that is often involved in nitrogenmetabolism regulation. In addition, this organism contains agene encoding an HPrK, which is typically found in gram-positivebacteria, although it is also present in many other alphaproteobacteria(3, 37). HPrK phosphorylates HPr on a serine residue ratherthan at the histidine residue targeted by EI.

The S. meliloti Hpr protein is interesting because it displayscharacteristics of both the Hpr and Npr proteins. hpr is ina gene cluster with manX and hprK. This gene cluster is highlyconserved in alphaproteobacteria, implying that the interactionof Hpr with HprK and ManX (a typical EIIA) is important andhas been maintained by selection (37). This suggests that theprotein has properties characteristic of classic Hpr proteins.In addition, the S. meliloti Hpr protein has sequence similarityto Npr proteins which interact with the EI^Ntr and EIIA^Ntr proteins.Also, it likely is phosphorylated at His15 by EI^Ntr, the onlyEI in S. meliloti.

In order to shed light on the role of HPr and ManX (EIIA^Man)in S. meliloti, we constructed mutants with unmarked, in-framedeletions of these genes and characterized the resulting phenotypes.

Carbon source utilization. Deletion of manX in S. meliloti resulted in strong phenotypes.The ${Delta}$ manX strain grew slower than the wild-type strain on allcarbon sources tested (succinate, glucose, glycerol, raffinose,lactose, and maltose). It exhibited lower levels of expressionof the PmelA::gfp fusion during growth on raffinose, as wellas lower β-galactosidase activity during growth on lactose.When cells were tested for SMCR on succinate plus raffinoseor on succinate plus lactose, the length of the diauxic lagwas altered in a secondary-carbon-source-dependent fashion.The lag was shorter than that of the wild type on succinateplus raffinose and longer than that of the wild type on succinateplus lactose.

In contrast, deletion of hpr did not result in altered expressionof the melA-agp or lac operon as measured by PmelA::gfp fusionand by endogenous β-galactosidase activity. Growth of the ${Delta}$ hpr strain on succinate, glycerol, and glucose was normal. Growthon lactose, maltose, and raffinose, which are carbon sourcessubject to repression by succinate, was slower. The diauxicgrowth curves for succinate plus raffinose and for succinateplus lactose showed a shorter diauxic lag than the diauxic growthcurves obtained with the wild-type strain, indicating that therewas earlier derepression of genes needed for utilization ofthe secondary carbon sources. Deletion of manX and hpr togetherresulted in phenotypes identical to those of the ${Delta}$ hpr strain.

One hypothesis is that the low levels of expression of raffinoseand lactose catabolic genes in the ${Delta}$ manX strain were a directconsequence of the absence of one of the forms of ManX (EIIA^Man),i.e., that ManX or ManX-P was required for normal expressionof these catabolic genes. Normal PmelA::gfp expression and enzymelevels in the ${Delta}$ hpr strain, which should lack ManX-P, suggestedthat this form of ManX was not required. Additionally, the ${Delta}$ manX ${Delta}$ hpr strain, which lacked both ManX-P and ManX, also exhibitednormal levels of enzyme activity and gene expression, rulingout the possibility that there was direct involvement of ManX(EIIA^Man) in the observed ${Delta}$ manX phenotype.

Given the results described above, an alternative hypothesisis that one of the forms of HPr is responsible for the phenotypesobserved for the ${Delta}$ manX strain. Removal of HPr in a ${Delta}$ manX background( ${Delta}$ manX ${Delta}$ hpr strain) restored normal levels of expression, providingstrong evidence that supports this hypothesis. It is likelythat the absence of ManX (EIIA^Man) results in accumulation ofhistidine-phosphorylated HPr (HPr-His-P), as ManX (EIIA^Man)dephosphorylates this form of HPr. Therefore, elevated levelsof HPr-His-P in the ${Delta}$ manX strain may have resulted in downregulationof the melA-agp and lac operons. As the levels of HPr-His-Pincrease, the levels of HPr-Ser-P should decrease because HPr-His-Pis less amenable to phosphorylation by HprK at the serine residue(61). The same phenotype (downregulation of catabolic genes)should be observed for the ${Delta}$ manX strain if HPr-Ser-P were requiredfor relief from catabolite repression. The two strains lackinghpr ( ${Delta}$ hpr and ${Delta}$ manX ${Delta}$ hpr mutants) have no HPr-Ser-P, and they donot exhibit low levels of melA-agp or lac expression. Thesestrains lack HPr-His-P as well, and this strongly suggests thatHPr-His-P is directly or indirectly responsible for the lowlevels of expression of melA-agp and lac observed in the ${Delta}$ manXstrain. This does not rule out the possibility that it is thebalance between HPr-His-P and HPr-Ser-P which regulates thelevels of expression of melA-agp and lac. It has been suggestedthat in some gram-positive organisms a high proportion of HPr-His-Pplus HPr relative HPr-Ser-P slows the metabolism of specificcarbon sources (56). This may also be the case for S. meliloti.

The SMCR phenotypes of the ${Delta}$ hpr and ${Delta}$ manX ${Delta}$ hpr strains also suggestedthat one of the forms of HPr is involved. Both of these mutantslack HPr and exhibited early release from the diauxic lag. However,the levels of expression of melA-agp and lac in these strainsduring the first phase of diauxic growth on succinate plus raffinoseor on succinate plus lactose were similar to the levels of expressionin the wild-type strain. The unaltered levels of expressionwere a surprise given the early release from the diauxic lag.This indicated that the time necessary for melA-agp or lac inductionand the resulting increase in the enzyme level, after succinatewas depleted, was shorter in the ${Delta}$ hpr strains than in the wild-typestrain. We do not currently know how this might occur.

When the SMCR phenotypes of the ${Delta}$ manX mutant were considered,the diauxic lag on succinate plus raffinose was shorter thanthe diauxic lag for the wild-type strain, while its diauxiclag was longer than that of the wild type on succinate pluslactose. This was in spite of the fact that the expression ofmelA-agp and lac was not different from the expression in thewild-type strain during the first phase of growth in eithercondition. This suggests that differences in the molecular mechanismsof induction of melA-agp and lac may be important for how ManXinfluences gene activation. It is known that inducer exclusionis involved in SMCR in S. meliloti (11). It is possible thatManX (EIIA^man) or HPr mediates carbon metabolism at least partiallythrough inducer exclusion. HPr and ManX (EIIA^man) could alsointeract with the nitrogen-PTS proteins encoded by genes inthe S. meliloti genome and influence carbon metabolism in conjunctionwith them.

There is no evidence that there is a homologue of CcpA in S.meliloti, and a regulator with the role of Crp has not beenfound yet, although there are two genes that could encode Crp-likeproteins. However, it is possible that there is a mechanismsimilar to one present in Pseudomonas spp. P. putida and Pseudomonasaeruginosa are also subject to SMCR; they possess the Crc (cataboliterepression control) global regulator that acts at the translationallevel, preventing the synthesis of activators necessary forexpression of genes for secondary carbon source utilizationin the presence of succinate (33, 47, 52). Analysis of the S.meliloti genome yields at least three genes coding for probableexodeoxyribonucleases with similarity to Crc. Whether such proteinsinfluence SMCR in S. meliloti is currently unknown.

Cobalt requirements and survival in stationary phase. Deletion of hpr from S. meliloti resulted in increased sensitivityto cobalt deficiency. In the absence of cobalt, growth of themutant was frequently arrested, and the growth rates were oftenlower than those on minimal medium plus cobalt. The ${Delta}$ hpr and ${Delta}$ manX ${Delta}$ hpr strains also exhibited rapid die-off during stationaryphase, which was greater in rich medium or in cobalt-deficientminimal medium than in cobalt-amended minimal medium. In minimalmedium the inability to survive in stationary phase seemed tobe linked to the extreme sensitivity of the mutants to cobaltdeficiency. However, in rich medium there was steep die-offduring stationary phase when, presumably, cobalt was not limiting.The fact that the ${Delta}$ manX mutant did not exhibit altered survivalin stationary phase or sensitivity to cobalt deficiency suggeststhat the phenotype of the ${Delta}$ hpr mutant is due to the absence ofone or more forms of HPr and not due to its inability to phosphorylateManX (EIIA^Man).

Cobalt is the central atom in the tetrapyrrole ring structureof vitamin B₁₂ cofactors. These cofactors are critical for methyltransferases,ribonucleotide reductases, and other enzymes (17, 48). S. melilotipossesses several B₁₂ (cobalt)-dependent enzymes involved inDNA synthesis, methionine synthesis, and propionate metabolism(6, 17, 18). We did not investigate which of the cobalt-dependentenzymes mentioned above, if any, is involved in the responseto cobalt limitation or how these enzymes could be related tothe function of HPr.

Exopolysaccharide production and symbiotic characteristics. The ${Delta}$ hpr strain and the ${Delta}$ manX ${Delta}$ hpr strain exhibited calcofluor-brightphenotypes and a fluorescent halo sooner than the wild type,indicating that deletion of HPr resulted in an alteration inthe control of succinoglycan synthesis. In contrast, the ${Delta}$ manXstrain showed accumulation of high-molecular-weight succinoglycan,as judged from the calcofluor-bright phenotype and the verylate appearance of a halo. Succinoglycan is critical for symbiosis;mutants unable to synthesize succinoglycan or low-molecular-weightsuccinoglycan cannot colonize their host or fix nitrogen, whilesome mutants that overproduce the polymer are also defectivein colonization or nitrogen fixation (15, 31, 45, 59, 74). Thethree mutant strains ( ${Delta}$ hpr, ${Delta}$ manX, and ${Delta}$ manX ${Delta}$ hpr) were symbioticallyunimpaired. Clearly, the succinoglycan-related phenotypes exhibitedby the mutants in this study were mild enough not to alter symbiosis.It is interesting that a recently isolated mutant defectivein B₁₂ cofactor synthesis was identified by a succinoglycanoverexpression screen, establishing a link between regulationof succinoglycan synthesis and the cobalt-containing cofactor(13).

We show here that deletion of the genes encoding the PTS proteinsHPr and ManX (EIIA^Man) resulted in profound physiological changesand affected carbon catabolism, succinoglycan production, andthe ability to cope with nutrient stresses. We also presentevidence that HPr-His-P is the factor most likely to be directlyor indirectly responsible for the observed phenotypes. We speculatethat HPr acts as a regulator for a variety of processes in thecell and that the phenotypes observed in the absence of ManX(EIIA^Man) are due to the accumulation of HPr-His-P.

ACKNOWLEDGMENTS

This work was supported by Department of Energy contracts DE-FG02-01ER15175and DE-FG02-06ER15805 to D.J.G., by a University of ConnecticutResearch Foundation grant to D.J.G., and by a Heinz HerrmannGraduate Fellowship in Cell Biology from the University of Connecticut'sGraduate School to R.M.B.

FOOTNOTES

* Corresponding author. Mailing address: Department of Molecular and Cell Biology, University of Connecticut, 91 N. Eagleville Rd., U-3125, Storrs, CT 06269-3125. Phone: (860) 486-3092. Fax: (860) 486-1784. E-mail: daniel.gage{at}uconn.edu

Published ahead of print on 15 February 2008.

Present address: Molecular, Cellular, & Developmental Biology,University of Arizona, 1007 E. Lowell Street, Tucson, AZ 85721-0106.

REFERENCES

Top