Effective pH homeostasis is correlated with optimal growth in C. glutamicum
We first quantified the efficiency of pH homeostasis in C. glutamicum. We performed growth assays in shaken micro titer plates (MTP) in minimal medium in presence of optimized buffers at a pH of 4 to 11 with subsequent determination of growth rates. As seen in Fig. , optimal growth rates were observed at a pH of 7 to 8.5. At an external pH below 6 and above 9 growth rates decreased drastically and at a pH of 4 as well as 10.5 and 11 no significant growth was observed.
Figure 1 Comparison of growth rate and internal pH of C. glutamicum exposed to different external pH values. Growth experiments were performed in selected buffer systems in shaken microtiter plates (pH 4-5.5 black circles: Homopipes, pH 5.5-6.5 white triangles (more ...)
Subsequently, we determined the internal pH of C. glutamicum cells grown at pH 7.5 after exposure to different external pH values. At a pH of 7.5 the internal pH value was found to be 7.5. This value was kept constant (± 0.5 pH units) after lowering the external pH down to 6 or increasing the pH up to 9. Below or above these external pH values the internal pH decreased respectively increased much faster in response to an external pH shift. We concluded that C. glutamicum can perform effective pH homeostasis in a range of external pH values from 6 to 9.
The failure of effective pH homeostasis at low or high external pH values could result from an impaired energy metabolism. The pH gradient across the cytoplasmic membrane (ΔpH) is important for generation of the proton motive force (pmf = ΔΨ - 2.3RT/F × ΔpH) which is essential for ATP synthesis by the F1F0ATPase. In order to prove whether the pmf is affected in C. glutamicum we determined ΔpH as well as the membrane potential ΔΨ in cells exposed to different pH values and calculated the pmf. The results are shown in Fig. . As expected, the pH gradient is zero at pH 7.5. At lower external pH values ΔpH increased up to 60 mV, whereas at higher external pH values, ΔpH was found to be reverted and decreased to -80 mV. The values for ΔΨ were found to be 110 mV at an external pH of 4.5, increased to 200 mV at pH 8, and at very high pH values (pH 10.5) 245 mV were measured. As a consequence of the decreasing ΔpH and increasing ΔΨ values the resulting pmf was kept relatively constant at a surprisingly broad pH range of 4.5 to pH 11, varying between 150 to 200 mV. At the most acidic pH of 4 the membrane potential ΔΨ collapsed and the resulting pmf value was 40 mV only.
Figure 2 The pH dependent bioenergetic homeostasis in C. glutamicum. Membrane potential (triangles) and pH gradient (circles) across the cytoplasmic membrane of C. glutamicum exposed to different external pH values and values for the resulting proton motive force (more ...)
Transcriptome and proteome analyses of pH acclimatization in C. glutamicum
In order to unravel components and processes involved in pH homeostasis we performed transcriptome analyses by DNA microarrays and proteome studies by 1D-nLC-ESI-MS/MS. For this purpose, two independent batch fermentations were carried out at a pH of 6, 7.5, and 9 in stirred bioreactors under continuous pH control and samples were drawn during the exponential phase in order to focus on the steady state pH homeostasis and to prevent additional perturbations by short term responses. We observed growth rates of 0.14 ± 0.01 at pH 6, 0.32 ± 0.02 at pH 7.5, and 0.14 ± 0.01 at pH 9. Cells were harvested and immediately frozen for metabolic inactivation. Transcriptome patterns were analyzed by co-hybridization of cDNA derived from cells grown at pH 6 vs. pH 7.5 and from cells grown at pH 9 vs. pH 7.5. The data analysis was performed as described previously [16
], using a m-value (log2
of the relative change in the respective mRNA ratio) cut-off of ± 1 which corresponds to transcription changes equal or greater than twofold. For the comparative proteome analysis we performed identification and quantification of peptides in the enriched soluble and membrane fractions as well as in the cell envelope fraction. Relative quantification of protein abundance and its change was performed by the spectral counting technique and the results are also given as log2
]. The complete set of data is available as supplementary material.
The transcriptome analyses revealed 42 genes with increased expression at pH 9 in comparison to pH 7.5 (Table ). In the respective proteome studies, 19 corresponding proteins were found with an increased peptide number in at least one protein fraction whereby four of them were present at significantly higher levels at pH 9 in comparison to pH 7.5 (Table ). For 39 genes we found a decreased mRNA level at pH 9 in comparison to pH 7.5 whereby for 26 corresponding proteins a lower content was indicated by lower peptide numbers in at least one fraction and for four of them a significant lower abundance at pH 9 was observed (Table ). For 10 genes differentially expressed at pH 9 we did not find a corresponding change of the peptide number and for 18 proteins we did not find any corresponding peptides at all indicating a low abundance of these proteins in C. glutamicum cells. The comparison of cells grown at pH 6 and pH 7.5 revealed higher mRNA pools for 88 genes whereby for 49 corresponding proteins (for 10 of them significantly) increased peptide numbers were found (Table ). A lower mRNA content at pH 6 was found for 91 genes whereby for 52 corresponding proteins (for 16 of them significantly) decreased peptide numbers were found at least in one protein fraction at acidic pH (Table ). In case of 17 proteins alterations of mRNA and protein content do not match at acidic pH. For 35 genes differentially expressed at pH 6 in comparison to pH 7.5 no peptide was found at all. In summary, we found many overlaps of transcriptome and proteome data for C. glutamicum grown at different pH values.
Differential expression pattern at pH 9 in comparison to pH 7.5
Differential expression pattern at pH 6 in comparison to pH 7.5
In addition, a number of proteins with changed abundance was detected for which no change in gene transcription was observed. We identified for 43 proteins increased and for 30 proteins decreased peptide numbers at pH 6 (Additional file 1
). The same held for 32 proteins with increased and for 20 proteins with decreased peptide numbers under alkaline conditions, (Additional file 2
). An example is the gene cg1111
encoding enolase. The mRNA content was neither significantly changed at pH 6 (m-value 0.24) nor at pH 9 (m-value 0.09) but 229 peptides were found in the cytoplasmic fraction at pH 7.5, 334 at pH 6, and 104 at pH 9 (Additional file 2
). Other examples with stable mRNA level and varying peptide numbers include the porines of the outer membrane PorA and PorH (decreased amounts of peptides found at pH 9 and 6 in comparison to pH 7.5 in the membrane fraction) as well as MetE (increased peptide numbers at pH 6 in the cytoplasmic fraction, Additional file 2
). This indicates that posttranscriptional or posttranslational control might be involved and that the regulation of protein stability is important during pH acclimatization. Furthermore, because of the (putative) function of many genes that are differentially expressed in a pH dependent manner the rearrangement of the cell wall might take place and influence the gene expression response.
Subsequently, we checked whether transcriptional regulators are known to be involved in expression control of genes that were found to be regulated. This was done using the CoryneRegNet data base which provides information on 72 regulators in C. glutamicum
]. For 21 of the 39 genes found to be repressed at pH 9, predictions were made or experimental evidence was obtained, for regulation by particular transcription factors (Table ). Accordingly, for approx. 50% of the genes found to be induced at pH 9 or differentially expressed at pH 6 the transcriptional regulator was proposed or identified (Table , ).
Iron homeostasis of C. glutamicum is affected by the external pH
The iron availability is monitored in C. glutamicum
by the binding of ferrous iron to the transcription factor DtxR [19
]. At high internal concentrations of ferrous iron, the regulator binds to operator sites in the promoter regions of target genes, including RipA, the second regulator of iron homeostasis. Whereas DtxR can act both as repressor and activator, RipA acts as repressor only [20
]. The combined transcriptome and proteome data suggest that the external pH value influences the availability of iron. At alkaline pH, DtxR-repressed genes like cg0925-28
, encoding a siderophore ABC transporter, or cg0767
, encoding a siderophore interacting protein, are found to be repressed, while the mRNA levels respectively peptide numbers of DtxR-activated genes like ftn
(encoding a ferritin-like protein involved in iron storage) and dps
) are increased. We found RipA-regulated genes encoding iron containing enzymes like succinate dehydrogenase (cg0446-0447
), aconitase (cg1737
), or catalase (cg0310
) to be (slightly) repressed at pH 6, whereas the same genes were found to be induced at alkaline pH (Table ). Additionally, higher peptide numbers were found for SdhA, SdhB, Acn, and KatA under alkaline conditions (Table ). Furthermore, genes of the SufR regulon, cg1759-65
, including the genes nifS2
, and sufB
which encode components of the FeS cluster assembly machinery, as well as the regulator SufR itself are induced at pH 9 (Table ). In contrast, the ABC type transporter for ferric iron uptake encoded by cg0508-0506
is not under the control of DtxR and no change of the transcript or protein level was detected (data not shown). In summary, we found a pH-dependent regulation of genes of the RipA and DtxR regulon indicating the activation of the iron starvation response at pH 6 and iron excess conditions at pH 9.
At neutral and acidic pH H2O2 can be detected in C. glutamicum cultures
The induction of iron starvation response at pH 6 was surprising because the solubility of iron is increased at low pH values and the availability should be increased at pH 6. Therefore, we speculated that activation of iron starvation could be caused by an impaired function of the cytoplasmic regulators. By oxidation of the cytoplasmic ferrous iron to ferric iron, the co-activator of DtxR, DtxR-mediated regulation might be triggered. Such a process could be induced by the endogenous formation of reactive oxygen species as described for the Fur protein in E. coli
]. In order to test for the pH dependent occurrence of oxidative stress in C. glutamicum
cells, we performed again batch fermentations in bioreactors in minimal medium under continuous pH control. During the exponential phase we detected significantly higher levels of H2
at pH 6 (6.5 μM, OD600
4) than under neutral (2.2 μM, OD600
12) or alkaline pH conditions (0.9 μM, OD600
6). Additionally, in cultures grown in buffered minimal medium in Erlenmeyer flasks we could detect H2
during exponential growth in C. glutamicum
cultures. We measured 3 μM H2
in cultures grown at pH 9 but in cultures grown at pH 7.5 and pH 6 we measured unexpected high concentrations of H2
, namely 20 μM after eight hours of incubation. The results indicate the increased occurrence of oxidative stress in C. glutamicum
and/or suggest that the defense against oxidative stress is impaired in a pH dependent manner. In order to assess an effect of H2
production at low pH we applied a well established method for the measurement of protein carbonylation by using the OxyBlot assay. Total proteins of cells grown at pH 6, 7.5 and 9 were extracted and subjected to 1D SDS PAGE before and after the OxyBlot treatment (Additional file 3
). Interestingly, a high number of proteins can be detected to harbour carbonyl groups in C. glutamicum
protein extracts of cells grown at every pH. We could not find a significant increase in carbonylation at low pH.
Furthermore we performed growth experiments in Erlenmeyer flasks at pH 7.5 and pH 6 in presence of external catalase enzyme (Fig. ). Interestingly, for C. glutamicum cells grown at pH 7.5 in presence of catalase (16 KU/ml) a higher growth rate was observed (μ = 0.393 ± 0.005) in comparison to the absence of external catalase (μ = 0.343 ± 0.006). At pH 6 addition of catalase had no significant beneficial effect because the growth rates in presence or absence of catalase were comparable (Fig. ). Catalase was also added after every hour of incubation in order to prevent loss of enzymatic activity and to provide continuous catalase activity but comparable results were obtained (data not shown). In conclusion, elimination of H2O2 by addition of external catalase enzyme facilitates growth of C. glutamicum at neutral pH but not at acidic pH conditions.
Figure 3 Impact of externally added catalase enzyme on growth of C. glutamicum. Wild type cells were exposed to pH 6.0 (white symbols) and 7.5 (black symbols) in buffered medium in Erlenmeyer flasks and growth was determined in absence (circles) or presence (squares) (more ...)
Metabolic alterations during response to acidic pH
The amounts of several enzymes were found to be affected by the changed external pH including succinate dehydrogenase and aconitase. In order to unravel metabolic alterations caused by the differing protein content, we performed GC-MS based metabolic profiling of cells grown at pH 6 and pH 7.5 under continuous pH control. Thereby, we identified numerous amino acids, intermediates of TCA, glycolysis, pentose phosphate pathway, and methionine pathway to be present at significantly different levels (Table ). For example, pyruvate was found at pH 6 at an eleven fold higher concentration than at pH 7.5. Within the TCA, citrate, which is the substrate of aconitase, was found to accumulate like malate and fumarate. In contrast the metabolites 2-oxoglutarate and succinate were found in significantly lower concentrations at pH 6 (Table ). Among the amino acids accumulation of phenylalanine, valine, glutamine, and alanine was observed, and proline and β-alanine were found in lower concentrations at pH 6. The pool size of methionine was slightly decreased, but we identified intermediates of the methionine pathway to be present in high concentrations like cystathionine and cysteine (Table ). On the other hand nearly all enzymes of the methionine pathway were found to be induced at the mRNA and/or protein level.
Differential metabolite pattern at pH 6 in comparison to pH 7.5
The McbR regulon is induced at acidic pH
At pH 6 we observed induction of genes encoding proteins of the methionine and cysteine pathway (Table , Fig. ). Intermediates of these pathways are involved in essential cellular functions including the assembly of iron sulfur clusters (cysteine), the de novo
synthesis of proteins (cysteine, methionine) or the metabolism of C1
-adenosyl-methionine, methyltetrahydrofolate; Fig. ). Many of the genes are under control of McbR and the ancillary regulators CysR and SsuR [23
]. Among them are, e.g., the fpr2
cluster and cysK
, encoding the sulfate permease CysZ, the complete set of enzymes involved in sulfate reduction to sulfide (CysDN, CysH, CysIX, and Fpr2) as well as the serine-O
-acetylserine sulfhydrylase CysK, involved in cysteine synthesis (Fig. ). Furthermore, the genes hom
, metB, metH, metK, metXY
, and metQN
, encoding enzymes of the methionine pathway and subunits of the primary methionine uptake system MetQNI, were found to be induced [8
]. The genes encoding the cysteine synthase (cysK
), the homocysteine methyltransferase (metE
), the β-C-S lyase (aecD
), and the S
-adenosyl-homocysteine hydrolase (sahH
) were not found to be induced at the mRNA level (Table , Fig. ). Corresponding to the unaffected mRNA level of aecD
no differential peptide numbers were found (Fig. ). For the AecD enzyme we determined unchanged cystathionine lyase activities in cells grown at pH 7.5 and at pH 6 using an enzymatic assay (data not shown). However, a higher protein level was found for MetE and a lower amount for CysK in spite of the unaffected mRNA levels (Additional file 1
, Fig. ). This might be an indication for increased protein stability of MetE and CysK at low pH. In contrast to pH 6 the McbR and CysR regulon were not found to be differentially expressed at pH 9 (Table , Fig. ). It should be noted that we are not able to report on genes under the control of the transcription factor SsuR, because no transcription data were obtained for these genes and no peptides were found representing the corresponding proteins.
Figure 4 The pH dependent regulation of the methionine and cysteine metabolism in C. glutamicum. The metabolite pool sizes at pH 6 in comparison to pH 7.5 are indicated below the intermediates. The involved proteins as well as the encoding genes are given in circles (more ...)
At the metabolite level we observed the accumulation of intermediates of the methionine pathway upstream of the AecD enzyme including L-homoserine, O-acetyl-L-homoserine, L-cysteine, and L,L-cystathionine (Table , Fig. ). Furthermore, the content of the McbR effector S-adenosyl-homocysteine was increased at low pH. In contrast, the pool sizes of homocysteine and methionine, representing metabolites downstream of AecD, were found to be slightly reduced.
From the observed metabolic imbalance we inferred that accumulation of intermediates of the methionine pathway upstream of AecD or the lower pool size of the downstream intermediates could contribute to the growth defect of C. glutamicum cells at acidic pH. In order to test this hypothesis we performed growth experiments at pH 7.5 and 6 in absence or presence of 10 mM cystathionine, cysteine, homocysteine, or methionine. Based on these assumptions, the addition of cystathionine or cysteine should increase pH dependent growth inhibition whereas homocysteine and methionine should supplement a putative demand for these compounds at pH 6. The addition of cystathionine, homocysteine and methionine had no significant effect on C. glutamicum growth at pH 6 (data not shown). However, addition of cysteine significantly decreased growth rates of cells exposed to acidic pH values. Further experiments revealed that the extent of growth inhibition by cysteine was indeed pH dependent. Whereas at pH 9 and 7.5 cysteine addition had no effect on the growth rate, at pH 7 growth was retarded and at pH 6.5 and 6 cells were hardly able to grow (Fig. ).
The pH dependent impact of cysteine on growth of C. glutamicum. Wild type cells were exposed to different pH values in MTP and growth rates were determined in absence (black bars) or presence 10 mM cysteine (white bars).
Differential expression of further regulatory modules
Beside the induction of methionine and cysteine synthesis, the complete arg cluster was found to be induced at pH 6. The expression of the arg genes, encoding all enzymes for synthesis of arginine from glutamate via the urea cycle, was proven to be under the control of the two repressors ArgR and FarR (Table ). The investigation of the metabolite pattern revealed, however, a lower pool size for ornithine, citrulline, and/or arginine, represented by only one signal in the GC-MS analysis (Table ).
The transcription factors RamA and RamB as well as GlxR are major regulators of the carbon flux in C. glutamicum
]. At pH 9 and pH 6 we observed the repression of several genes indicating alterations of the carbon metabolism. Among them are aceA
, encoding isocitrate lyase, aceB
, encoding malate synthase, and mctC
, encoding an uptake system for pyruvate, acetate and propionate [8
]. All of these genes are under the control of RamA and RamB. Additionally, alternative oxidases were found to be induced, like the FMN containing lactate dehydrogenase LldD and the pyruvate quinone oxidase Pqo. Whereas lldD
expression was proposed to be under the control of GlxR, pqo
expression was found to be part of the sigma factor SigB regulon [28
Another regulatory module found to be induced at pH 6 comprises the genes cg1214-18
. These genes encode the NadAC proteins, involved in NAD synthesis, a putative cysteine desulfurase, possibly involved in maturation of FeS clusters necessary for function of the NadAC complex, and the regulator NrtR [8
]. All genes of the operon were induced at pH 6 and for NadC (cg1215
), an increased number of peptides (33) was found at acidic conditions in comparison to pH 9 (12 peptides, Table ). Under alkaline conditions genes of the NrtR regulon were not induced. Subsequently, we determined the cellular concentration of all NAD derivatives in C. glutamicum
cells grown at pH 7 and pH 6. Whereas the NADP and NADPH concentrations at pH 6 were only half of those observed at pH 7 (NADP pH 6: 0.11 ± 0.01 mM, pH 7: 0.18 ± 0.03 mM; NADPH pH 6: 0.25 ± 0.03 mM, pH 7 0.48 ± 0.07 mM) the NAD and NADH concentrations were only one third at pH 6 in comparison to pH 7 (NAD pH 6: 0.57 ± 0.05 mM, pH 7: 1.53 ± 0.14 mM; NADH pH 6: 0.46 ± 0.05 mM, pH 7 1.49 ± 0.39 mM). Calculation of the ratios of the oxidized and reduced forms revealed that the reduction state of the cell was not affected by acidic pH. However, significantly lower levels of NAD derivatives were found under acidic conditions accompanied by the induction of genes encoding enzymes involved in the first steps of their synthesis.