Search

Corynebacterium glutamicum, like Mycobacterium tuberculosis, is a member of the Corynebacteriales, which have linear fatty acids and as branched fatty acids the mycolic acids. We identified accD1 and fasA as key genes of fatty acid synthesis, encoding the β-subunit of the acetyl-CoA carboxylase and a type-I fatty acid synthase, respectively, and observed their repression during growth on minimal medium with acetate. We also identified the transcriptional regulator FasR and its binding sites in the 5′ upstream regions of accD1 and fasA. In the present work we establish by co-isolation and gel-mobility shifts oleoyl-CoA and palmitoyl-CoA as effectors of FasR, and show by DNA microarray analysis that in presence of exogeneous fatty acids accD1 and fasA are repressed. These results are evidence that acyl-CoA derivatives derived from extracellular fatty acids interact with FasR to repress the genes of fatty acid synthesis. This model also explains the observed repression of accD1 and fasA during growth on acetate, where apparently the known high intracellular acetyl-CoA concentration during growth on this substrate requires reduced accD1 and fasA expression for fine control of de novo fatty acid synthesis. Consequently, this mechanism ensures that membrane lipid homeostasis is maintained when specific nutrients are available resulting in increased acetyl-CoA concentration, as is the case with acetate, or when fatty acids are directly available from the extracellular environment. However, the genes specific to mycolic acid synthesis, which are in part shared with linear fatty acid synthesis, are not controlled by FasR, which is in agreement with the fact that they can not be supplied from the extracellular environment but that their synthesis fully depends on a constant supply of linear fatty acid chains.

L-Isoleucine is an essential amino acid, which is required as a pharma product and feed additive. Its synthesis shares initial steps with that of L-lysine and L-threonine, and four enzymes of L-isoleucine synthesis have an enlarged substrate specificity involved also in L-valine and L-leucine synthesis. As a consequence, constructing a strain specifically overproducing L-isoleucine without byproduct formation is a challenge. Here, we analyze for consequences of plasmid-encoded genes in Corynebacterium glutamicum MH20-22B on L-isoleucine formation, but still obtain substantial accumulation of byproducts. In a different approach, we introduce point mutations into the genome of MH20-22B to remove the feedback control of homoserine dehydrogenase, hom, and threonine dehydratase, ilvA, and we assay sets of genomic promoter mutations to increase hom and ilvA expression as well as to reduce dapA expression, the latter gene encoding the dihydrodipicolinate synthase. The promoter mutations are mirrored in the resulting differential protein levels determined by a targeted LC-MS/MS approach for the three key enzymes. The best combination of genomic mutations was found in strain K2P55, where 53 mM L-isoleucine could be obtained. Whereas in fed-batch fermentations with the plasmid-based strain, 94 mM L-isoleucine with L-lysine as byproduct was formed; with the plasmid-less strain K2P55, 109 mM L-isoleucine accumulated with no substantial byproduct formation. The specific molar yield with the latter strain was 0.188 mol L-isoleucine (mol glucose)(-1) which characterizes it as one of the best L-isoleucine producers available and which does not contain plasmids.

The wild type of Corynebacterium glutamicum does not excrete l-lysine, but mutants are developed to use C. glutamicum for the large-scale production of this amino acid. l-lysine is derived from oxaloacetate and pyruvate. A reduced entry of these precursors into the citric acid cycle triggers l-lysine synthesis, but there is simultaneously an unfavorable reduction in growth rate. Here, we further investigated strains with reduced citrate synthase activity and observed that proline supplementation partly reversed the growth rate reduction, and simultaneously increased l-lysine accumulation by 10%. The combined effect of faster growth and higher l-lysine formation lead to an increase in l-lysine productivity of about 14%. Small scale bioreactor cultivations showed an increase of the glucose uptake rate from 3.96 to 5.12 mmol gCDW−1 h−1 as the effect of proline addition. Deletion of the proline-utilization gene putA demonstrated that proline did not serve to refill the citric acid cycle. The combined data show that proline addition has a favorable effect on l-lysine formation and sugar uptake, which indicates that engineering sugar uptake is a future option for increased product formation.

The pentanol isomers 2-methyl-1-butanol and 3-methyl-1-butanol represent commercially interesting alcohols due to their potential application as biofuels. For a sustainable microbial production of these compounds, Corynebacterium glutamicum was engineered for producing 2-methyl-1-butanol and 3-methyl-1-butanol via the Ehrlich pathway from 2-keto-3-methylvalerate and 2-ketoisocaproate, respectively. In addition to an already available 2-ketoisocaproate producer, a 2-keto-3-methylvalerate accumulating C. glutamicum strain was also constructed. For this purpose, we reduced the activity of the branched-chain amino acid transaminase in an available C. glutamicuml-isoleucine producer (K2P55) via a start codon exchange in the ilvE gene enabling accumulation of up to 3.67g/l 2-keto-3-methylvalerate. Subsequently, nine strains expressing different gene combinations for three 2-keto acid decarboxylases and three alcohol dehydrogenases were constructed and characterized. The best strains accumulated 0.37g/l 2-methyl-1-butanol and 2.76g/l 3-methyl-1-butanol in defined medium within 48h under oxygen deprivation conditions, making these strains ideal candidates for additional strain and process optimization.

We here developed a series of Corynebacterium glutamicum strains with gradual decreased specific citrate synthase (CS) activity and quantified in a multifaceted approach the consequences of residual activity on the transcriptome, metabolome, and fluxome level as well as on L-lysine formation and growth. We achieved an intended gradual L-lysine yield increase and recognized and overcame further new limitations in the L-lysine biosynthesis pathway to result in a strain with the highest yield reported so far when assayed under comparable conditions. As a non-intended outcome, a detailed flux analysis revealed an almost constant flux through CS at 10% remaining CS activity, whereas the metabolome data revealed an increase in the oxaloacetate and acetyl-CoA concentrations. Hence reduced CS activity is apparently efficiently buffered by increased concentrations of CS substrates, implying a certain robustness of the central metabolism in response of the imposed gene expressions.

Citrate synthase (CS) is located at a major branch point in the metabolism and is required for both tricarboxylic acid and glyoxylic acid cycle activity. Here we show that the CS gene gltA of Corynebacterium glutamicum is monocistronic, but that two transcripts are formed with their transcript initiation sites located 121 bp and 357 bp upstream of the translational start codon, respectively. Northern blot analyses revealed that during growth on acetate the short transcript prevails, whereas during growth on glucose the long transcript is dominant. Further Northern blots, reporter gene fusions, and CS activity measurements in mutants devoid of the transcriptional regulators RamA and RamB or with the global regulator GlxR overexpressed revealed a complex involvement of these regulators in gltA transcription. This was confirmed by demonstrating the direct interaction of isolated RamA, RamB and GlxR proteins with specific gltA promoter regions in vitro. The combined analyses point to an elaborate control of gltA transcript formation, which is possibly required as a dedicated mechanism to balance the total CS activity according to the physiological requirements.

Two unrelated protein families catalyse the oxidative decarboxylation of 2-oxoacids, i.e. the 2-oxoacid dehydrogenase complexes (OADHCs) and the 2-oxoacid ferredoxin oxidoreductases (OAFORs). In halophilic archaea, OAFORs were found to be responsible for decarboxylation of pyruvate and 2-oxoglutarate. Nevertheless, two gene clusters encoding OADHCs were found previously in Haloferax volcanii, but their biological function remained obscure. Here a third oadhc gene cluster of H. volcanii is presented. To characterize the function, the genes encoding the E1 subunit were inactivated in all three gene clusters by in-frame deletions. Under aerobic conditions none of the three mutants showed any phenotypic difference from the wild-type in various media. However, growth yields of two mutants were considerably lower than that of wild-type under nitrate-respirative conditions in complex medium. Northern blot analyses revealed (1) that polycistronic transcripts are formed and all three gene clusters are bona fide operons and (2) that transcription of all three operons is induced under anaerobic conditions compared to aerobic conditions. Taken together, the three H. volcanii enzymes do not fulfil one of the 'usual' aerobic functions of typical OADHCs, but decarboxylate an as-yet-unidentified novel substrate under anaerobic conditions. A survey of all 28 fully sequenced archaeal genomes revealed that nearly all archaea contain several OAFORs (three to four on average), suggesting that this protein family was already present in their last common ancestor. In contrast, only nine archaea encode one or two OADHCs, indicating that this protein family entered archaea by lateral transfer of the cognate genes from bacteria. This view is underscored by a phylogenetic tree of 33 archaeal and bacterial OADHCs.

Peptidyl-prolyl cis / trans isomerases (PPIases) catalyze the rate-limiting protein folding step at peptidyl bonds preceding proline residues and were found to be involved in several biological processes, including gene expression, signal transduction, and protein secretion. Representative enzymes were found in almost all sequenced genomes, including Corynebacterium glutamicum , a facultative anaerobic Gram-positive and industrial workhorse for the production of amino acids. In C. glutamicum , a predicted single-domain FK-506 (tacrolimus) binding protein (FKBP)-type PPIase (FkpA) is encoded directly downstream of gltA , which encodes citrate synthase (CS). This gene cluster is also present in other Actinobacteria . Here we carried out in vitro and in vivo experiments to study the function and influence of predicted FkpA in C. glutamicum . In vitro , FkpA indeed shows typical PPIase activity with artificial substrates and is inhibited by FK-506. Furthermore, FkpA delays the aggregation of CS, which is also inhibited by FK-506. Surprisingly, FkpA has a positive effect on the activity and temperature range of CS in vitro . Deletion of fkpA causes a 50% reduced biomass yield compared to that of the wild type when grown at 37°C, whereas there is only a 10% reduced biomass yield at the optimal growth temperature of 30°C accompanied by accumulation of 7 mM l -glutamate and 22 mM 2-oxoglutarate. Thus, FkpA may be exploited for improved product formation in biotechnical processes. Comparative transcriptome analysis revealed 69 genes which exhibit ≥2-fold mRNA level changes in C. glutamicum Δ fkpA , giving insight into the transcriptional response upon mild heat stress when FkpA is absent.

Using metabolic engineering, an efficient L-leucine production strain of Corynebacterium glutamicum was developed. In the wild type of C. glutamicum, the leuA-encoded 2-isopropylmalate synthase (IPMS) is inhibited by low L-leucine concentrations with a K(i) of 0.4 mM. We identified a feedback-resistant IMPS variant, which carries two amino acid exchanges (R529H, G532D). The corresponding leuA(fbr) gene devoid of the attenuator region and under control of a strong promoter was integrated in one, two or three copies into the genome and combined with additional genomic modifications aimed at increasing L-leucine production. These modifications involved (i) deletion of the gene encoding the repressor LtbR to increase expression of leuBCD, (ii) deletion of the gene encoding the transcriptional regulator IolR to increase glucose uptake, (iii) reduction of citrate synthase activity to increase precursor supply, and (iv) introduction of a gene encoding a feedback-resistant acetohydroxyacid synthase. The production performance of the resulting strains was characterized in bioreactor cultivations. Under fed-batch conditions, the best producer strain accumulated L-leucine to levels exceeding the solubility limit of about 24 g/l. The molar product yield was 0.30 mol L-leucine per mol glucose and the volumetric productivity was 4.3 mmol l⁻¹ h⁻¹. These values were obtained in a defined minimal medium with a prototrophic and plasmid-free strain, making this process highly interesting for industrial application.

Fast growth of industrial microorganisms, such as Corynebacterium glutamicum, is a direct amplifier for the productivity of any growth coupled or decoupled production process. Recently, it has been shown that C. glutamicum when grown in a novel picoliter bioreactor (PLBR) exhibits a 50% higher growth rate compared to a 1 L batch cultivation [Grunberger et al. (2012) Lab Chip]. We here compare growth of C. glutamicum with glucose as substrate at different scales covering batch cultivations in the liter range down to single cell cultivations in the picoliter range. The maximum growth rate of standard batch cultures as estimated from different biomass quantification methods is mu = 0.42 ± 0.03 h(-1) even for microtiter scale cultivations. In contrast, growth in a microfluidic perfusion system enabling analysis of single cells reproducibly reveals a higher growth rate of mu = 0.62 ± 0.02 h(-1). When in the same perfusion system cell-free supernatant from exponentially grown shake flask cultures is used the growth rate of single cells is reduced to mu = 0.47 ± 0.02 h(-1). Likewise, when fresh medium is additionally supplied with 5 mM acetate, a growth rate of mu = 0.51 ± 0.01 h(-1) is determined. These results prove that higher growth rates of C. glutamicum than known from typical batch cultivations are possible, and that growth is definitely impaired by very low concentrations of byproducts such as acetate.

The addition of fatty acids to either Escherichia coli or Bacillus subtilis elicits an elaborate cellular response of the lipid metabolism. We found that in Corynebacterium glutamicum the expression of accD1 encoding the β-subunit of the essential acetyl-CoA carboxylase is repressed in acetate-grown cells without the addition of fatty acids. The TetR-type transcriptional regulator NCgl2404, termed FasR, was identified and deleted. During growth on acetate, but not on glucose, 17 genes are differentially expressed in the deletion mutant, among them accD1, and fasA and fasB both encoding fatty acid synthases, which were upregulated. Determination of the 5' ends of accD1, fasA, fasB and accBC together with the use of isolated FasR protein identified the FasR binding site, fasO, which is located within the accD1 and fasA transcript initiation site thus blocking transcription by RNA polymerase binding directly. The identified fasO motif is present in C. efficiens or C. diphtheriae, too, and it is actually similarly positioned in these bacteria within the 5' ends of the accD1 and fasA transcripts, and a fasR orthologue is also present. The identification of the FasR-fasO system in Corynebacteriaceae might indicate a conserved transcriptional control of the unique lipid synthesis in these mycolic acid-containing bacteria.

Oxoglutarate dehydrogenase (ODH) and pyruvate dehydrogenase (PDH) complexes catalyze key reactions in central metabolism, and in Corynebacterium glutamicum there is indication of an unusual supercomplex consisting of AceE (E1), AceF (E2), and Lpd (E3) together with OdhA. OdhA is a fusion protein of additional E1 and E2 domains, and odhA orthologs are present in all Corynebacterineae , including, for instance, Mycobacterium tuberculosis . Here we show that deletion of any of the individual domains of OdhA in C. glutamicum resulted in loss of ODH activity, whereas PDH was still functional. On the other hand, deletion of AceF disabled both PDH activity and ODH activity as well, although isolated AceF protein had solely transacetylase activity and no transsuccinylase activity. Surprisingly, the isolated OdhA protein was inactive with 2-oxoglutarate as the substrate, but it gained transsuccinylase activity upon addition of dihydrolipoamide. Further enzymatic analysis of mutant proteins and mutant cells revealed that OdhA specifically catalyzes the E1 and E2 reaction to convert 2-oxoglutarate to succinyl-coenzyme A (CoA) but fully relies on the lipoyl residues provided by AceF involved in the reactions to convert pyruvate to acetyl-CoA. It therefore appears that in the putative supercomplex in C. glutamicum , in addition to dihydrolipoyl dehydrogenase E3, lipoyl domains are also shared, thus confirming the unique evolutionary position of bacteria such as C. glutamicum and M. tuberculosis .