Heterologous complementation of the exopolysaccharide synthesis and carbon utilization phenotypes of Sinorhizobium meliloti Rm1021 polyhydroxyalkanoate synthesis mutants


  • Punita Aneja,

    1. Department of Biology, University of Waterloo, 200 University Avenue West, Waterloo, ON Canada N2L 3G1
    2. Department of Natural Resource Sciences, McGill University, 21,111 Lakeshore Road, Ste-Anne-de-Bellevue, QC Canada H9X 3V9
    Search for more papers by this author
    • 1

      Agribiotics Inc., 135 Turnbull Court, Cambridge, ON Canada N1T 1C6.

  • Meixue Dai,

    1. Department of Biology, University of Waterloo, 200 University Avenue West, Waterloo, ON Canada N2L 3G1
    2. Biology Department, Life Sciences College, Shandong Normal University, Jinan 250014, PR China
    Search for more papers by this author
  • Delphine A. Lacorre,

    1. Department of Biology, University of Waterloo, 200 University Avenue West, Waterloo, ON Canada N2L 3G1
    2. Département de Génie Biochimique et Alimentaire, L'Institut National des Sciences Appliquées, Toulouse, France
    Search for more papers by this author
    • 2

      Laboratoire de Biologie Vasculaire, IPBS – CNRS UMR 5089, 205 rte de Narbonne, 31400 Toulouse, France.

  • Brent Pillon,

    1. Department of Biology, University of Waterloo, 200 University Avenue West, Waterloo, ON Canada N2L 3G1
    Search for more papers by this author
  • Trevor C. Charles

    Corresponding author
    1. Department of Biology, University of Waterloo, 200 University Avenue West, Waterloo, ON Canada N2L 3G1
    Search for more papers by this author

  • Edited by A. Steinbuchel

*Corresponding author. Tel.: +519 888 4567 ext. 5606; fax: +519 746 0614, E-mail address: tcharles@uwaterloo.ca


A reduced exopolysaccharide phenotype is associated with inability to synthesize polyhydroxyalkanaote (PHA) stores in Sinorhizobium meliloti strain Rm1021. Loss of function mutations in phbB and phbC result in non-mucoid colony morphology on Yeast Mannitol Agar, compared to the mucoid phenotype exhibited by the parental strain. This phenotype is attributed to reduction in succinoglycan synthesis. We have used complementation of this phenotype and the previously described d-3-hydroxybutyrate/acetoacetate utilization phenotype to isolate a heterologous clone containing a Bradyrhizobium japonicum phbC gene. Sequence analysis confirmed that this clone contains one of the five predicted phbC genes in the B. japonicum genome. The described phenotypic complementation strategy should be useful for isolation of novel PHA synthesis genes of diverse origin.


The polyhydroxyalkanoates (PHAs) are important storage compounds that accumulate as intracellular deposits in several different bacteria. These polymers have generated significant commercial interest due to their promise as biodegradable plastics that can be produced using renewable resources [1]. The physiological and ecological roles of these compounds are also of important consideration, as they likely have a large influence on the ability of the bacterial cells to adapt to fluctuating nutrient conditions and to compete with other organisms. The central enzyme in the PHA synthesis pathway is PHA synthase, which mediates the polymerization of d-3-hydroxyacyl-CoA. PHA synthases have been divided into four classes based on their amino acid sequence, subunit composition, and substrate specificity [2]. The provision of substrate within the cell and the substrate specificity of the PHA synthase enzyme contribute to determination of the characteristics of the synthesized PHA [2].

The commercial potential of biotechnological production of novel polymers has driven the search for PHA synthesis systems. Although gene hybridization methods have been used in some of these efforts, phenotypic complementation strategies are preferred if diverse genes are desired. Phenotypic screens include colony opacity, staining with lipophilic dyes such as Sudan black B [3], Nile blue A [4] and Nile red [5], and fatty acid detoxification [6]. Not all of these screens work in all bacterial species, however.

In Sinorhizobium meliloti, the PHB synthesis genes are located in two separate regions of the chromosome, encoded by the phbC and phbAB transcripts, respectively. While the isolation of S. meliloti phbC mutants has been reported [7,8], there is yet no report of a phbA or phbB mutant. Here we describe the construction of a S. meliloti PHB synthesis mutant lacking NADP-acetoacetyl-CoA reductase activity (EC, encoded by phbB. This mutant is unable to produce PHB. We also demonstrate that this mutant, and a previously described phbC mutant, are defective in synthesis of the exopolysaccharide succinoglycan under growth conditions that favour both PHB and succinoglycan synthesis in the wild type strain. As reported previously for the phbC mutant [9], the phbB mutant also exhibits poor growth on the PHB cycle intermediates acetoacetate and 3-hydroxybutyrate. Furthermore, we demonstrate the isolation of a functional PHA synthase encoding gene from the nitrogen fixing soybean symbiont Bradyrhizobium japonicum by heterologous complementation of the S. meliloti phbC mutant.

2Materials and methods

2.1Bacterial strains and plasmids

A list of bacterial strains, plasmids and transposons used in this study is provided in Table 1. Culture methods in LB, TY, YM modified M9-minimal medium with various carbon sources, and antibiotic concentrations, were as described previously [10,11].

Table 1.  Strains and plasmids
S. meliloti  
RCR2011=SU47 wild typeRothamsted Experimental Station
Rm1021SU47 str-21, Smr[34]
Rm5000SU47 rif-5, Rfr[35]
Rm11105Rm1021 phbC1::Tn5[10]
Rm11107Rm1021 bdhA::Tn5[10]
Rm11134Rm1021 acsA2::Tn5[9]
Rm11144Rm1021 phbC1::Tn5-233[10]
Rm11172Rm1021 age-1::Tn5-Tp[36]
Rm11345Rm5000 phbB::ΩSmSpThis study
Rm11347Rm1021 phbB::ΩSmSpThis study
Rm11359Rm1021 phbB::ΩSmSp, age-1::Tn5-TpThis study
Rm7055Rm1021 exoF::Tn5[37]
E. coli  
DH5αFendA1 hsdR17 (rk, mk) supE44 thi-1 recA1 gyrA96 relA1Δ(argF lacZYA)U169 Φ80dlacZΔ M15, λ[38]
MT607pro-82 thi-1 hsdR17 supE44 recA56[39]
MT616MT607 (pRK600); mobilizing strain[39]
MT614MT607ΩTn5T.M. Finan, unpublished
pJQ200uc1Suicide vector, P15A ori, RP4 mob, sacB, Gmr[40]
pJQ200SKSuicide vector, P15A ori, RP4 mob, sacB, Gmr[40]
pUC7ColE1 cloning vector, Apr[41]
pUC19ColE1 cloning vector, Apr[42]
pTZ/PCT-vector, AprD. Tessier, NRC-Biotechnology Research Institute, Montreal
pBP7pLAFR1 cosmid clone containing B. japonicum genomic DNA, complements S. meliloti phbC mutantThis study
pMX354.8 kb Kpn1-BamH1 subclone from pBP7 in pSP329This study

2.2Genetics and molecular biology techniques

Bacterial conjugations, ΦM12 transductions, Tn5 mutagenesis, homogenotizations and transposon replacements were carried out as described previously [11,12]. The pLAFR1 cosmid [13] library of USDA110 genomic DNA was obtained from Laura Green [14]. DNA manipulations were performed using standard methods [15]. Oligonucleotide primers were purchased from Life Technologies GIBCO BRL (Gaithersburg, MD, USA). DNA amplification by PCR was performed in a MJ Research PTC-100 thermocycler.

2.3Enzyme assays and biochemical determinations

Cell-free extracts were prepared by sonication as described previously [10], from stationary phase TY cultures. Assays were carried out in triplicate. Assay for acetoacetyl-CoA reductase (EC activity was performed using the method described by Haywood et al. [16] for the oxidation reaction. The assay mixture (1.0 ml) consisted of glycine–NaOH buffer (pH 9.0; 50 μmol), dl-3-hydroxybutyryl-CoA (80 nmol) and NADP or NAD (80 nmol). The reaction was initiated with the addition of 50 μl of cell free extract. Thiolase activity was measured as previously described [9].

Cultures for PHB assays were obtained by growing the strains in 125 ml Erlenmeyer flasks containing 50 ml of YMB (g l−1: K2HPO4, 0.5; MgSO4· 7H2O, 0.2; NaCl, 0.1; mannitol, 10; yeast extract, 0.4) [17] and shaking at 160 rpm for 48 h. Following a saline wash and resuspension in 50 ml saline, PHB was extracted from a 2 ml fraction of culture and assayed using the method of Law and Slepecky [18]. The remaining 48 ml of culture was used for dry weight determination after incubation of the wet pellet at 37 °C until no further decrease in weight was noted. Values shown are means of two independently grown cultures.

2.4Nucleotide sequence accession number

The nucleotide sequence data reported in this paper have been submitted to GenBank and assigned Accession No. AY077580.

3Results and discussion

3.1Construction and characterization of a NADP-acetoacetyl-CoA reductase (phbB) mutant

An insertional mutagenesis strategy was employed to generate a S. meliloti phbB mutant. Based on the available sequence of the S. meliloti Rm41 phbB gene (GenBank Accession No. U17226) [17], two PCR primers were designed corresponding to 7–24 bp (5′-AGGGTAGCACTGGTAACG-3′ and 579–595 bp (5′-GAATGATCCGCTCGTTG-3′) of the coding sequence. The PCR product generated using RCR2011 genomic DNA as template was cloned into the vector pTZ/PC (which yields T-overhangs when digested with XcmI), generating recombinant plasmid pPA108, and then excised as a BamHI fragment and ligated to BamHI digested pUC7. The resulting plasmid pPA125 was cleaved at the unique SphI site, blunt ended with T4 DNA polymerase, and ligated to the SmaI ΩSmSp fragment [19]. Plasmid pPA131 with an ΩSmSp insertion in phbB was thus obtained. For gene replacement the mutagenized phbB was excised from pPA131 as an EcoRI fragment, blunt ended and ligated to the sac vector pJQ254 digested with SmaI giving pPA132. Plasmid pPA132 was introduced into Rm5000 by conjugation with selection for RfrGmr transconjugants, followed by selection on TY medium containing 5% sucrose Sm Sp to generate the double crossover recombinant, Rm11345. The insertion in phbB in strain Rm11345 was then transduced into Rm1021 to give strain Rm11347. The recombinant was confirmed by Southern blot analysis (data not shown).

Further confirmation of the phbB mutant was obtained by assaying for enzyme activities in cell extracts. Strain Rm11347 lacked NADP acetoacetyl-CoA reductase activity (Table 2). Both the wild type and the phbB mutant had comparable levels of NAD-acetoacetyl-CoA reductase activity, while the phbB mutant exhibited slightly reduced ketothiolase activity. Although the phbC gene has been shown to be essential for PHB synthesis, the requirement of the phbB gene product in PHB synthesis has not been demonstrated in S. meliloti due to the absence of a specific mutant strain. No PHB was detected in YMB grown cells of either Rm11347 (Rm1021 phbB mutant) or Rm11105 (Rm1021 phbC mutant) while the wild type Rm1021 strain grown under the same conditions accumulated 52% (w/w) of the total cellular dry weight in the form of PHB.

Table 2.  Enzyme activitya profile of phbB mutant compared to wild type
Strain3-Ketothiolase (thiolytic)3-Ketothiolase (condensation)NADP acetoacetyl-CoA reductaseNAD acetoacetyl-CoA reductase
  1. aSpecific activity (nmol/min/mg of protein; mean ± SD [n≥ 3]).

Rm10211015 ± 20212 ± 1.735 ± 1.14.3 ± 0.4
Rm11347623 ± 1713.7 ± 0.305.4 ± 0.1

The inability of the phbB mutant to accumulate PHB is in contrast to the phaB mutants of Rhodobacter capsulatus which are able to synthesize PHB [20,21]. This suggests that the remaining NAD+-acetoacetyl-CoA reductase activity in S. meliloti is unable to compensate for the absence of NADP+-acetoacetyl-CoA reductase activity towards the synthesis of d-hydroxybutyryl-CoA. Under the growth conditions tested, the sole pathway for generation of substrate for PHB production is via NADP+-acetoacetyl-CoA reductase.

We had earlier reported [9] that PHB synthesis mutants of S. meliloti exhibit reduced growth rate on the PHB cycle intermediates acetoacetate and 3-hydroxybutyrate as sole carbon source. The ability of the phbB mutant strain to utilize different carbon sources was compared on minimal medium containing glucose, acetate, acetoacetate or 3-hydroxybutyrate as sole carbon source. Although growth on glucose or acetate was not affected, the growth rate of the phbB mutant on acetoacetate or 3-hydroxybutyrate was considerably reduced. Similar reductions in growth rate on acetoacetate and 3-hydroxybutyrate have also been observed for the phbC mutant [9].

3.2PHB synthesis mutants are deficient in succinoglycan synthesis

Colonies of Rm1021, the wild type strain used in these studies, exhibit characteristically mucoid morphology on YM agar, while mutants of the succinoglycan synthesis pathway are non-mucoid on the same medium. Examples of the dry and mucoid phenotype are shown in Fig. 1. We observed that phbC and phbB mutant strains (e.g. Rm11105, Rm11347), in which the PHB synthesis pathway is disrupted and thus PHB deposits do not accumulate, were non-mucoid on YM agar (Fig. 1). The inability to synthesize PHB apparently results in a reduction in the levels of succinoglycan produced under these conditions. Mutants affected in PHB degradation, such as those with lesions in bdhA and acsA2, exhibited the same colony characteristics as the parental strain on YMA (Table 3). Furthermore, the succinoglycan deficiency of the PHB synthesis mutants cannot be suppressed by the overexpression of the acsA2 gene in the age-1 mutant background, as demonstrated by the YM agar dry phenotype of the double mutant Rm11359. It is therefore not the inability to utilize PHB cycle intermediates, but rather the inability to synthesize PHB itself, that is associated with the non-mucoid colony appearance.

Figure 1.

Colony phenotype of S. meliloti strains on YMA.

Table 3.  Carbon utilization and colony morphology phenotypes of the indicated S. meliloti strains
StrainRelevant characteristicsCarbon source utilizationColony morphology on YMA
Rm1021Wild type++++Mucoid
Rm11359phbB::Ω, age-1::Tn5-Tp++++++Dry

The reduced succinoglycan synthesis in the PHB synthesis mutants suggests that the pathways that lead to the synthesis of these polymers share a common regulatory circuit. Both of these polymers are produced when carbon is available in excess and nitrogen is limiting, such as is the case with the YM growth medium used in our studies. Reduction in NAD+/NADH ratio has been well documented in PHB synthesis mutants [22,23], as the important intracellular polymer pool for storage of reductant is abolished. In Azospirillum brasilense, however, this is accompanied by increased levels of exopolysaccharide synthesis [24], presumably as a result of channeling of excess reductant to exopolysaccharide polymer rather than intracellular polymer. An example of co-regulation of exopolysaccharide and PHB is the GacS control of both alginate and PHB synthesis in Azotobacter vinelandii[25]. When drawing comparisons between these studies and ours, it is important to bear in mind that the Rm1021 strain used in our work has an insertion mutation in the expR gene, which is a positive regulator of galactoglucan synthesis [26]. The reduced succinoglycan phenotype that we have described would probably not be detectable in a expR+ genetic background due to the increased galactoglucan production that renders the colonies mucoid even in the absence of succinoglycan synthesis.

3.3Phenotypic complementation of S. meliloti phbC mutant

We have described two phenotypic characteristics of S. meliloti PHB synthesis mutants that might be useful for isolation of heterologous PHB synthesis genes by functional complementation. To demonstrate this, the B. japonicum pLAFR1 genomic library was conjugated en masse into the S. meliloti phbC mutant Rm11144, and selection was carried out on M9 acetoacetate. Colonies that could be distinguished above the background growth on these plates were streak purified on the same medium, and tested for the acquisition of pLAFR1-encoded tetracycline resistance (Tcr). The selected plasmids were then transferred by conjugation into E. coli MT607. The plasmids thus isolated were able to complement the acetoacetate growth phenotype upon transfer back into the phbC mutant Rm11105. They also restored the YMA mucoid colony phenotype. As a final confirmation, it was demonstrated that one of the representative plasmids, pBP7, restored PHB synthesis ability to the phbC mutant Rm11144 (data not presented). This plasmid was also introduced into the phbB mutant Rm11347, and it did not restore the YMA mucoid colony phenotype.

3.4Transposon mutagenesis, subcloning and sequence analysis of phbC complementing clone

To localize the phbC gene on the complementing clone, in vivo Tn5 mutagenesis was carried out by conjugation of pBP7 from strain MT614 to Rm11144, and transconjugants were screened for colony phenotype on YM agar. Non-mucoid colonies were identified as carrying putative insertions in the plasmid-carried gene encoding PHB synthase. It was confirmed that such insertions abolished the ability to produce PHB, as well as reduced the ability to grow with acetoacetate as sole carbon source. Fragments containing four representative insertions and adjacent DNA were sub-cloned into pUC19 as BamH1 fragments. Initial sequence analysis was carried out using IS50 complementary primers, confirming that the inserts disrupted a gene with strong identity to known PHB synthase genes. Subcloning of pBP7 identified a 4.8-kb KpnI–BamH1 fragment (pMX35) that was sufficient for complementation. The portion of this fragment corresponding to the open reading frame disrupted by the Tn5 insertions was sequenced (GenBank Accession No. AY077580). This open reading frame corresponds to bll4360 in the B. japonicum genome sequence [27]. The ability of the clone to complement the S. meliloti phbC mutant confirms that the bll4360-encoded phbC is functional. The lack of restoration of the YMA mucoid colony phenotype to the phbB mutant strain by this clone is consistent with the absence of a phbB gene adjacent to the bll4360-encoded phbC.

3.5Multiple PHA synthesis encoding genes in the B. japonicum genome

Analysis of the B. japonicum genome sequence revealed the presence of five putative PHB/PHA synthase encoding genes, in different regions of the chromosome. The predicted proteins range in size from 360 to 600 amino acids, which is within the range of known PHA synthase proteins. BLASTP analysis [28] was done to identify the closest matches to each of these sequences in GenBank, and the percent amino acid sequence identity was then determined using Clustal pairwise analysis [29]. A neighbour joining tree constructed using the alignment of the more conserved carboxyl portion of the protein sequences is shown in Fig. 2. The gene that we cloned by heterologous complementation in this study, bll4360, is predicted to encode a 600 amino acid protein with 72% identity to a 601 amino acid predicted PHA synthase in the genome of Rhodopseudomonas palustris (NP_947843, RHOPA2) [30]. It also shares 57% identity with the 611 amino acid PHA synthase of S. meliloti (P50176, SINME) [7,17].

Figure 2.

Neighbour joining tree of the five B. japonicum PHA synthases and their closest matches. Alignments and trees were done with ClustalX [43] and PHYLIP [44]. For each polypeptide, only the core, more conserved portion was used, as indicated. blr3732, 199–494; blr2885, 44–398; bll6073, 241–535; bll4548, 54–342; bll4360, 229-523; ALCEU, 212-512; BACME, 45–340; RHOPA1, 213–507; RHOPA2, 230–524; SINME, 242–536; AZOVI, 80-372. Bootstrap analysis (1000 trials) gave support values of 100% in each case except between the blr3732-ALCEU and bll4360-RHOPA2-SINME branches, for which support was 67.8%.

The 572 amino acid sequence of blr3732 is 49% identical to the Class I PHA synthase of Wautersia eutropha (P23608, ALCEU) [31], and is the only one of the five predicted PHA synthase encoding genes to be located near other PHA synthesis genes. Although not immediately adjacent to phbB and phbA genes, the three genes are separated by less than 15 kb. This is an interesting region of the genome that also includes genes predicted to encode enzymes involved in C4-dicarboxylate transport and metabolism and pyruvate metabolism. It is tempting to speculate that these genes might be important in nodule invasion and bacteroid metabolism.

The sequence of blr2885, the shortest polypeptide of the five at 360 amino acids, is 38% identical to the Class IV 362 amino acid PHA synthase of Bacillus megaterium (AAD05260, BACME) [32], but no phaR gene was detected in the genome of B. japonicum, so presumably the blr2885-encoded protein does not require activation by a phaR-encoded protein as its B. megaterium homolog does [33]. Another short encoded sequence, bll4548, at 371 amino acids, is 44% identical to a PHA synthase in A. vinelandii (ZP_00092820.1, AZOVI). The 613 amino acid bll6073 is 67% identical to a PHA synthase in Rhodopseudomonas palustris (NP_949579.1, RHOPA1) [30]. Whether each of these genes encodes a functional PHA synthase must be tested experimentally. They all contained the modified lipase box motif GXCXG and the conserved catalytically important residues that are found in all PHA synthase enzymes [2]. The exhibited sequence diversity suggests multiple origins. Notably, none of the enzymes falls within the cluster associated with the pseudomonad Class II [2]. It will be interesting to investigate the functions and substrate specificities of the B. japonicum PHA synthase enzymes. The presence of multiple copies of PHA synthesis genes in the B. japonicum genome suggests an importance of intracellular carbon stores in either the symbiotic or saphrophytic aspects of the lifestyle of the cell.

3.6Concluding statement

We have demonstrated colony morphology and growth phenotypes in S. meliloti strain Rm1021 that are associated with the ability to synthesize PHB. Through the heterologous functional complementation of these phenotypes, we were able to isolate a novel PHB synthesis encoding gene from a B. japonicum gene library. Although we have so far demonstrated only the ability to isolate a heterologous Class I synthase gene, we foresee that this system could be further developed for isolation of diverse novel polymer biosynthesis genes from both cultivated and uncultivated bacterial genomes.


We are grateful to Laura Green for providing the USDA110 cosmid library, Michael Hynes for the plasmid pJQ200, Daniel Tessier for the plasmid pTZ/PC, and Turlough Finan for the exopolysaccharide mutants.

Research was supported by a grant to TCC from NSERC Canada. M-X.D. received support from China Scholarship Council during his stay as a visiting scholar in the Charles Laboratory.