Carbon-dependent control of electron transfer and central carbon pathway genes for methane biosynthesis in the Archaean, Methanosarcina acetivorans strain C2A

Gene redundancy in the M. acetivorans genome

The M. acetivorans genome contains many seemingly redundant copies of genes annotated with roles in methanogenesis [5]. These include two sets of genes annotated for a molybdate-type formylmethanofuran dehydrogenase (fmd), and two gene sets for a tunsten-type formylmethanofuran dehydrogenase (fwd), five heterodisulfide reductase-like hdrED and hdrABC gene clusters for reduction of Coenzyme M-Coenzyme B heterodisulfide, two sets of vht genes for F420 non-reducing hydrogenase, and two sets of genes for ATP synthesizing complexes [5]. Additional genes include frh hydrogenase-like genes, plus additional genes for rnf- and mrp-type membrane associated bacterial electron transfer complexes, plus genes needed for acetate metabolism (discussed below). Homologous and seemingly "redundant" genes/gene sets are also found in the genomes of M. mazei, and M. barkeri (Table 1). The reason for these genome makeups is currently unknown. M. acetivorans was used as a model microorganism to evaluate expression of over twenty sets of genes using gene specific primer pairs designed to eliminate cross-hybridization when DNA sequence similarity exists (Methods). RT-PCR, pPCR, and 5' analysis was then performed using RNA isolated from M. acetivorans cells grown with either acetate or methanol as the sole source of carbon and energy. In this study, a number of new M. acetivorans gene designations were established to distinguish among homologous orfs (Table 1, and described below).

Formylmethanofuran dehydrogenase (fmd, fwd) gene expression

Two of the four previously annotated sets of genes for formylmethanofuran dehydrogenasethese were designated as molybdenum-type enzymes and are named here as fmdE1F1A1C1D1B1 and fmdF2A2C2D2B2 (Figure 1A). Two additional gene sets were annotated as tungsten-type formylmethanofuran dehydrogenase, and are designated here as fwdD1B1A1C1 and fwdG2B2D2 (Figure 1B). Using qPCR analysis methods (Methods), the molybdenum-type operon reporter genes fmdE1 and fmdA1 (Figure 1A) were shown to be expressed at 14-fold higher levels during methanol growth conditions relative to acetate growth (Figure 1C). The second set of reporter genes (fmdF2, fmdA2, and fmdB2) were expressed about 2-fold higher during these conditions, but the maximal level of expression was less than 5% of that seen for the fmdE1 and fmdA1 genes. Noteworthy, the fmdE1 and fmdA1 gene expression values were within the same range observed for the fpoN and fpoL genes that encode subunits of the F420 H₂ dehydrogenase needed for central pathway electron transfer functions (described below). The high transcript abundance of the fmdE1F1A1C1D1B1 gene cluster implies a major role of this gene set during methanogenesis in contrast to that for the fmd2 gene set.

The annotated tungsten containing formylmethanofuran dehydrogenase gene cluster fwdD1B1A1C1 reporter genes designated fwdB1 and fwdA1 (Figure 1B) were also expressed 15-fold higher levels during methanol growth relative to acetate (Figure 1C). Interestingly, this was within the magnitude observed for the fmdE1F1A1C1D1B1 gene cluster. However, the second tungsten-type gene cluster (as reported by the fwdB2 gene), was constitutively expressed and at a level about one-half of that observed for either fwdA1 or fwdB1. These fmd/fwd transcript abundance measurements clearly demonstrate that two of the four fmd and fwd gene clusters (i.e., fmdE1F1A1C1D1B1 and fwdD1B1A1C1) are highly transcribed in response to substrate availability, and furthermore this suggests that two distinct formylmethanofuran dehydrogenase activities are concurrently utilized during methanol growth conditions (discussed below).

Heterodisulfide reductase gene expression

M. acetivorans genome analysis revealed five genes/gene clusters annotated as heterodisulfide reductase, an enzyme essential for electron transfer from methanogenic electron donors to methyl-CoM reductase (Table 1, Figure 2A). These include genes for a membrane-type protein designated here as hdrE1, hdrD1 and hdrD2 similar to those needed for methane formation in M. barkeri[7]. An additional six genes encoding soluble-type heterodisulfide reductase proteins are also present in the genome. They include the hdrA1 gene associated with a poly-ferredoxin-like gene (pfd), an unlinked set of hdrCB genes called hdrC1 and hdrB1, and a third hdr gene cluster designated hdrA2 hdrC2 hdrB2 (Figure 2B).

Quantitative gene expression experiments (Figure 2C) revealed that the membrane-type hdrD1 gene was most highly expressed during acetate cell growth conditions, and where methanol conditions gave slightly lower transcript abundance (ca. 0.7-fold). In contrast, hdrD2 gene expression was very low (i.e., at level of about one twentieth that seen for the hdrD1 gene Figure 2C), suggesting a minor or no direct function in methanogenesis. Interestingly, the abundance of the soluble-type hdrA1pfd and hdrC1B1 gene transcripts were also nearly as high as for the membrane type hdrED1 genes (Figure 2C). Here, acetate growth gave three-fold higher hdrA1 transcript levels versus methanol growth conditions. The participation of a soluble-type hdrABC enzyme in M. acetivorans metabolism is currently unknown but must now be considered. An orf following the hdrA1 gene is annotated as a polyferredoxin (pfd), and this suggests a role for this protein in electron transfer to couple the soluble-type Hdr complex with an appropriate electron donor complex. In contrast, hdrA2 and hdrB2 transcript abundance was about two to twenty-fold lower under the corresponding conditions. This suggests a minor role for the second set of HdrABC-type genes (i.e., hdrA2B2C2) in methanogenesis.

The hdrA1pfd and hdrC1B1 genes for the soluble-type enzyme subunits are located at different chromosomal loci, and are coordinately expressed since their mRNA abundance levels are alike (Figure 2C). Additionally, the PCR-based gene experiments also demonstrate that the hdrA1pfd and the hdrED1 genes are each expressed as operons (data not shown). Taken together, these data are consistent with a need for both a membrane-type and a soluble type Hdr enzyme for electron transfer/energy conservation under acetate and methanol cell growth conditions. This suggests that distinct electron transfer pathways are operating to service the alternative Hdr enzymes.

The vht and frh gene clusters

The M. acetivorans genome lacks an echABCDEF gene cluster encoding an Ech-type hydrogenase with described roles in hydrogen uptake and ion translocation in M. mazei[3, 5]. Since M. acetivorans cells do not exhibit significant hydrogenase activity [8, 9], some other mechanism must provide a means for electron transfer from cellular donor(s) to Hdr. Interestingly, the M. acetivorans genome contains two sets of genes (designated vhtG1A1C1D1, and vhtG2A2C2) for F420-nonreducing hydrogenase-types (Figure 3A, 3B, Table 1). It also contains a set of frhADGB genes for a coenzyme F420-type hydrogenase (Figure 3A). Quantitative RT-PCR assays (Figure 3C) established that the vhtG1 and vhtC1 genes were each expressed at four- to six-fold higher levels during methanol growth conditions, and this is within the range seen for the fpoL and fpoN genes needed for methyl group oxidation for methanol and acetate metabolism. In contrast, expression of the vhtG2 and vhtC2 genes was low under all conditions examined (Ca. about 17-20-fold lower than vhtG1 and vhtC1). Finally, the frhA and frhB gene expression levels were low relative to vhtG1 or fpoL (Figure 3C), and this suggests a minor role for the frhADGB and vhtG2A2C2 gene clusters in either methanol or acetate-dependent cell growth. Since vhtG1 transcript abundance was elevated and about half of that observed for the fpoL and fpoN genes that encode subunits of the F420 H₂ dehydrogenase (Figure 3C), this implies a significant physiological role for the vhtG1A1C1D1 gene products during methanol growth. The biochemical and physiological roles for the vhtG1 and vhtC1 hydrogenase-type genes in M. acetivorans are presently unknown.

The rnfXCDGEABY gene cluster is abundantly expressed

M. acetivorans contains a set of six genes (MA0659-0664) annotated as nqr123456[5] that are absent in the M. mazei, and M. barkeri genomes (Table 1). These genes were subsequently re-designated rnfCDGEAB based on sequence comparisons to the rnf and nqr-type genes in other microorganisms, [10]. This gene cluster also contains two additional genes of unknown function that we designate here as rnfX and rnfY (Figure 4A) whereby the first (MA0658) precedes rnfC and the second (MA0665) follows rnfB. We propose that these genes may encode unique input/output modules for membrane associated electron transfer since they are absent in other microbial genomes. During acetate cell growth relative to methanol growth conditions, the rnfX, rnfG, and rnfA reporter genes exhibited elevated transcript abundance (ca. 2.5 to 3.5-fold; Figure 4D). Each gene was also more highly expressed than many reference genes involved in central methanogenesis (e.g., fpoN, and fpoL that encode subunits of the F420 H2 dehydrogenase). Therefore, the rnfXCDGEABY gene expression data support the proposal that the products participate in electron transfer during acetate metabolism as proposed via methanophenazine [10]. In addition, they must also function during methanol culture conditions based on transcript abundance (Figure 4D). Other roles can be envisioned including participation in electron transfer to a soluble-type heterodisulfide reductase via a poly-ferredoxin (e.g., encoded by the hdrA1 pfd and hdrC1B1 gene complex, described below).

The mrpABCDEFG gene cluster is acetate induced

The M. acetivorans genome contains a set of seven genes called mrpABCDEFG (Figure 4B) with similarity to the gene clusters found in a variety of bacterial species but absent in either M. barkeri or M. mazei (Table 1) [5, 11–13]. The mrp- encoded protein complex in Bacillus subtilis was shown to confer a role in multiple drug resistance and/or pH regulation [13, 14]. As revealed by the M. acetivorans transcript analysis studies (Figure 4D), the mrpA and mrpF reporter genes were expressed more highly during acetate cell growth conditions (Ca. 11 to 12-fold) relative to methanol growth. These levels were above the expression levels observed for the ack, pta, and hdr genes needed for acetate utilization, and within the range seen for the rnf gene cluster. These findings imply a major role for the six mrp gene products in acetate metabolism versus methanol metabolism.

Expression of the atp and aha genes encoding ATP synthase complexes

M. acetivorans contains genes for a bacterial-type F₀F₁ synthase encoded by the MA2441 to MA2433 genes designated here as atpDCIHBEFAG, plus an archaeal-type A₀A₁ ATP synthase encoded by the ahaHIKECFABD genes (MA4152 to MA4160) (Figure 5). Although prior DNA microarray experiments [6] demonstrated that six of the nine genes in the archaeal-type A₀A₁ ATP synthase (ahaECFABD) encoding the ATP-hydrolysing/synthesizing domain (A₁) were expressed two-fold higher in acetate grown cells relative to methanol, the other genes were not [6]. It is still unknown how their expression varies quantitatively relative to atpDCIHBEFAG gene cluster expression. Corresponding DNA microarray studies with the atpDCIHBEFAG genes that encode a bacterial-like F₀F₁ complex revealed that only two of the nine genes (atpD and atpC) were expressed significantly higher in acetate by 3.2 and 1.8 fold, respectively: the remaining genes were either not detected or did not exhibit changes. Lastly, relative to central pathway genes for acetate and methanol utilization, it was unresolved how the aha and atp gene sets are expressed since the microarray data did not address this.

From the RT-PCT transcript abundance studies, three representative aha genes representing the archaeal-type A₀A₁ ATP synthase genes were highly expressed relative to the atp reporter genes (Figure 5C). Acetate cell growth conditions resulted in two-fold higher aha transcript levels relative to methanol cell growth. These genes were the most highly expressed in the cell regardless of the growth condition. In contrast, the bacterial-type F₀F₁atpD, atpA and atpG genes were expressed at less than 2% of the level seen for the ahaI, ahaC and ahaB genes: this suggests a minor role for the atp genes in methanogenesis in contrast to the aha gene cluster.

Acetate-induced genes

One M. acetivorans gene of unknown function (MA4008) was revealed by our prior DNA microarray studies to be more highly expressed during acetate growth conditions relative to methanol cell growth (Lars Rohlin, personal communication). Inspection of the amino acid sequence revealed six trans-membrane spanning regions reminiscent of a membrane solute uptake system (Additional file 1, Figure S1 and discussed below). To extend these MA4008 gene expression findings, quantitative PCR experiments were performed (Methods, Figure 6A). MA4008 was expressed at a 125-fold higher level during acetate versus methanol cell growth conditions. Interestingly, when methanol was also present in the culture medium in addition to acetate, MA4008 expression was suppressed to a level seen when only methanol was present (ca. by 215 fold). This indicates that the MA4008 gene is expressed only when the energetically superior carbon substrate is absent, consistent with a proposed role in acetate uptake. The M. acetivorans MA4008 orf is designated aceP for its role in an acetate-dependent membrane function. Two other genes required for acetate utilization are ack (MA3606) and pta (MA3607) that encode acetate kinase and phosphoacetyl transferase, respectively ([15] Table 1). Quantitative PCR experiments (Figure 6A) established that both genes were highly expressed and at levels similar to aceP when acetate was the sole substrate. The 11-18-fold differential pta and ack gene expression findings are similar to previous reports in M. acetivorans and M. thermophila[6, 16].

Location of the fpoP, hdrE, hdrA1, mrpA, pta, aceP, and ahaA promoters

The mRNA 5' ends of the fpoPABCDHIJJKLMNO, hdrED1, hdrA1-pfd, mrpABCDEFG, pta ack, aceP and ahaHIKECFABD genes/clusters were determined to locate their corresponding promoter elements. Using primer extension methods (Figure 7A), all but one of the promoter elements were demonstrated to have long un-translated regions (UTR's) that range from 51 to 137 nucleotides in length. For example, the aceP 5' mRNA end is located 104 nucleotides upstream of the translational start site. Similar findings were seen for the mrpA, fpoP, ahaH, hdrE, and hdrA genes. Only the pta gene had a relatively short UTR (i.e., 27 nt). We did not detect mRNA 5' ends for either rnfX or hdrC1. Alignment of all the upstream regions of these promoter elements (Figure 7A) revealed the highly conserved sequence present in other archaeal promoters, the TATA box (Figure 7B) located approximately 20-30 nt upstream of the +1 mRNA start site (discussed below). This site is bound by the TBP protein that aids RNA polymerase binding [17]. In contrast, the BRE box elements were not well conserved. When the UTR elements and the upstream regions were further examined using a suite of bioinformatics tools (Materials), no clearly discernable DNA sequence elements with either dyad symmetry or direct repeats were found. Similarly, no conserved regions within the RNA UTR's were seen for the coordinately expressed hdrA1pfd and hdrC1B1 genes sets.

Carbon control in the Archaea

Considerable information is available concerning carbon control of gene expression in bacterial and eukaryal systems, but little is yet known about related carbon control in the Archaea. Few studies have been reported for any archaeal species but include microarray studies in Pyrococcus furiosus[28], M. mazei[29, 30], and M. acetivorans[6]. The present experiments extend these studies to address a larger set of genes needed for carbon flow and electron transfer leading to methane formation from two key methanogenic substrates (Figure 8). It provides a foundation of RNA transcript abundance and 5' end data to begin exploring regulatory controls in this organism at the level of regulated mRNA synthesis and turnover. Little is known about the relative contributions of archaea transcription factors, translation factors, and/or small RNA's in gene regulation in the Methanosarcina species to provide the distinct patterns of gene expression observed here. M. acetivorans clearly maintains a cellular commitment to dynamically control transcript levels in response to methanogenic substrate type where two major gene families are further defined by this study.

Cell culture

Methanosarcina acetivorans C2A [1] was cultivated in a mineral medium that contained (in grams per liter): NaCl, 11.69 g; MgSO₄ 7H₂O, 12.32 g; KCl, 0.76 g; CaCl₂·2H₂O, 0.14 g; NH₄Cl, 0.5 g; Resazurin solution (10,000 × stock solution), 0.1 ml; trace metal solution (100×) 10 ml [29]; vitamin solution (100×) 10 ml [29]; HCl (12.1 N) 0.5 ml; Na₂HPO₄ 7H₂O, 1.12 g; cysteine-HCl H₂O, 0.25 g; Na₂CO₃, 3.0 g. An atmosphere (80:20) of nitrogen to carbon dioxide was used in the vessel headspace. Following sterilization, the medium was supplemented with filter-sterilized 0.1 ml 50% methanol or 0.2 ml 5 M acetate per 10 ml medium as previously described [30].

RNA purification

For RNA isolation, cultures of M. acetivorans C2A cells were grown on acetate or methanol with serial transfer of three times to mid-exponential phase before cell harvest. Total RNA was purified from 10 ml of cell samples using the RNAwiz (Ambion Austin, TX) following the manufacturer's instructions. The purified RNA was treated with DNase I as described [31, 32].

Quantitative RT-PCR

The real time reverse transcription (RT-PCR) reactions were performed using Superscript II reverse transcriptase (Invitrogen Carlsbad, CA) according to the manufacturers recommended protocol using random primers and 1 μg of total RNA. A mock reaction without Superscript was run to evaluate for the presence of genomic DNA contamination. To remove complementary RNA, 1 μl RNase H was added to mixture and incubated for 20 min at 37°C. The RNase was then heat inactivated at 70°C for 15 min. The cDNA from the RT reaction was diluted 10 fold, and 1 μl of the diluted cDNA was subsequently used in a 30 μl iQ SYBR green supermix according to the manufactures recommendations following addition of 1.5 μl DMSO. The real time PCR reactions were conducted on a Biorad iCycler (Biorad, Hercules, CA) or an Eppendorf Realtime² (Eppendorf, Westbury, NY) using a four-step program consisting of, denaturing, annealing, extension, and acquisition steps. The RT-PCR primers were created by a modified version of MyPROBES [32]. The PCR product lengths were in a range of 100-200 bp, the melting temperature was in the range of 55-66°C, the GC content was 55-65%, and the primer length was 17-22 bases (Additional file 4, Table S1). The primers were tested against serial dilution of genomic DNA (10⁶ to 10² copies) to generate a standard curve for each gene tested. The products were also visualized on agarose gels to establish the generation of a unique product of the correct size for each gene probe. Each gene expression value was then determined in triplicate for each of the three biological samples in conjunction with a genomic DNA serial dilution standard. Melting curves were analyzed to establish that non-specific amplification had not occurred (i.e., biphasic vs mono-phasic for a single product). The reported copy number was calculated from a total of nine data points. Each gene was also tested against the mock reaction. The gene expression data for each gene was compared to a reference gene (MA3998) that showed no significant up or down regulation in microarray experiments of Li, et al. [6]. In an independent approach, all qPCR signals were also normalized to the total amount of RNA used in the experiment, and in a separate analysis, to the RNA for the mcr genes (MA4546-4550) that encode methyl coenzyme M reductase. The results from the latter two approaches were in excellent agreement to the MA3998 normalization procedure. Values are reported in transcript copy number per 5 μg total RNA.

Primer extension analysis

To determine mRNA 5' ends, primer extension reactions were performed as described previously [33] using gene specific primers which were located approximately 60 bases downstream of the ATG start codons of the mrpA, hdrE, hdrA, aceP, ahaH, pta, and fpoP genes (see Additional file 4, Table S1 listing each primer). Total RNA was isolated described above. A total of 30 μg of RNA was used in each primer extension reaction: the primer and RNA was heated to 85°C for 10 min, and then slowly cooled to 45°C: ³³P-labeled dATP and unlabeled dCTP, dGTP, and dTTP were added to the mixture, and reverse transcription was then performed at 50°C using Superscript III Reverse Transcriptase (Invitrogen Carlsbad, CA) according to manufactures recommendations. The reaction was stopped by sequentially adding 5 μl 3 M sodium acetate (pH 5.2) and 150 μl 100% ice-cold ethanol followed by overnight incubation at -20°C. The cDNA's was precipitated at 13,000 rpm at 4°C for 35 min. For generation of fragments of the indicated regulatory region was cloned into TOPO-PCR4 vector (Invitrogen Carlsbad, CA). The Sequtherm Excel II Kit (Epicentre Madison, WI) was used to perform sequencing reactions of the DNA regions cloned into TOPO-PCR4 using the above primers to confirm the intended sequences. The extension and sequencing products were resolved on a 6.0% sequencing gel and exposed to a phosphorimager screen as previously described [32].

Informatics analysis and data visualization

Protein similarities were determined using BLAST [34], the alignment and the phylogentic tree of proteins were done with clustalw [35] and the visualization of the trees were done with splitTree4 [36]. Upstream DNA regions were searched for palindromic and repeated motifs using simple Perl script software written in house. Similar searches were also performed for conserved elements in the UTR regions.