Genes and pathways for CO2 fixation in the obligate, chemolithoautotrophic acidophile, Acidithiobacillus ferrooxidans, Carbon fixation in A. ferrooxidans

Background Acidithiobacillus ferrooxidans is chemolithoautotrophic γ-proteobacterium that thrives at extremely low pH (pH 1-2). Although a substantial amount of information is available regarding CO2 uptake and fixation in a variety of facultative autotrophs, less is known about the processes in obligate autotrophs, especially those living in extremely acidic conditions, prompting the present study. Results Four gene clusters (termed cbb1-4) in the A. ferrooxidans genome are predicted to encode enzymes and structural proteins involved in carbon assimilation via the Calvin-Benson-Bassham (CBB) cycle including form I of ribulose-1,5-bisphosphate carboxylase/oxygenase (RubisCO, EC 4.1.1.39) and the CO2-concentrating carboxysomes. RT-PCR experiments demonstrated that each gene cluster is a single transcriptional unit and thus is an operon. Operon cbb1 is divergently transcribed from a gene, cbbR, encoding the LysR-type transcriptional regulator CbbR that has been shown in many organisms to regulate the expression of RubisCO genes. Sigma70-like -10 and -35 promoter boxes and potential CbbR-binding sites (T-N11-A/TNA-N7TNA) were predicted in the upstream regions of the four operons. Electrophoretic mobility shift assays (EMSAs) confirmed that purified CbbR is able to bind to the upstream regions of the cbb1, cbb2 and cbb3 operons, demonstrating that the predicted CbbR-binding sites are functional in vitro. However, CbbR failed to bind the upstream region of the cbb4 operon that contains cbbP, encoding phosphoribulokinase (EC 2.7.1.19). Thus, other factors not present in the assay may be required for binding or the region lacks a functional CbbR-binding site. The cbb3 operon contains genes predicted to encode anthranilate synthase components I and II, catalyzing the formation of anthranilate and pyruvate from chorismate. This suggests a novel regulatory connection between CO2 fixation and tryptophan biosynthesis. The presence of a form II RubisCO could promote the ability of A. ferrooxidans to fix CO2 at different concentrations of CO2. Conclusions A. ferrooxidans has features of cbb gene organization for CO2-assimilating functions that are characteristic of obligate chemolithoautotrophs and distinguish this group from facultative autotrophs. The most conspicuous difference is a separate operon for the cbbP gene. It is hypothesized that this organization may provide greater flexibility in the regulation of expression of genes involved in inorganic carbon assimilation.


Background
Acidithiobacillus ferrooxidans is a mesophilic, obligately chemolithoautotrophic, γ-proteobacterium that gains energy and reducing power from the oxidation of ferrous iron and reduced inorganic sulfur compounds (RISCs) [1]. It grows optimally at pH 2, although growth as low as pH 1 has been reported [2]. The microorganism is a key player in the solubilization of copper in industrial bioleaching operations and makes an important contribution to the biogeochemical cycling of nutrients and metals in pristine and manmade acidic environments. In such environments, CO 2 would be expected to exist preferentially as a dissolved gas in equilibrium with the atmosphere and not in the bicarbonate form typically found at circum-neutral pHs [3].
A. ferrooxidans has previously been shown [4,5] to have candidate genes (cbbL and cbbS) for the large and small subunits of ribulose-1,5-bisphosphate carboxylase/ oxygenase (RubisCO, EC 4.1.1.39) that catalyses CO 2 fixation by the Calvin-Benson-Bassham (CBB) cycle in many organisms [6]. cbbL and cbbS are linked to genes predicted to encode carboxysome shell proteins [7] and a divergently transcribed gene encoding the LysR-type transcription regulator CbbR [4]. The intergenic region between cbbR and cbbL is predicted to harbor binding sites for CbbR [4]. In addition, microarray transcript profiling experiments have detected differential expression of several genes in A. ferrooxidans potentially involved in the CBB cycle depending on the growth substrate used [8].

Bacterial strains and culture conditions
Information regarding bacterial strains and plasmids used in this study is provided in Table 1. A. ferrooxidans was cultured in 9 K medium (adjusted to pH 3.5 with H 2 SO 4 ) containing 5 g/l elemental sulfur at 30°C under aerobic conditions on a rotary shaker at 150 rpm as described previously [21]. Escherichia coli harboring plasmids was grown at 37°C in LB broth with ampicillin (Amp: 100 μg/ml).

General DNA techniques and sequencing of DNA
A. ferrooxidans cultures were centrifuged at 800 × g to remove solid sulfur precipitates prior to cell harvest. Unattached cells were pelleted at 8000 × g for 10 min. The cell pellet was resuspended in 9 K salt solution for washing and washed cells were collected by centrifugation at 8000 × g for 10 min as described previously [21]. Standard procedures [22] were employed to isolate genomic and plasmid DNA from bacteria, to transform plasmid DNA into E. coli, and for general DNA handling. Restriction endonucleases and DNA-modifying enzymes were used as recommended by the manufacturers. Plasmid DNA was prepared by means of the QIAprep Spin Mini Kit (Qiagen). Polymerase chain reaction (PCR) products were amplified with Taq DNA polymerase (Fermentas) and purified from agarose gels using the QiaEx DNA Purification Kit (Qiagen). Each PCR reaction contained in a volume of 25 μl 1 ng of template DNA, 0.5 μM of required primers and 0.2 mM of each deoxyribonucleotide in 1× PCR buffer containing 1.5 mM MgCl 2 (Fermentas). PCR conditions were as follows: initial denaturing step at 95°C for 5 min followed by 30 amplification cycles (denaturation at 95°C for 30 s, annealing at the appropriate temperature depending on the specific primer pairs for 20 s, elongation at 72°C) and a final elongation step at 72°C for 10 min. DNA sequencing of pBAD-cbbR was carried out by the Göttingen Genomics Laboratory (Göttingen, Germany).

Isolation of RNA and RT-PCR
Total RNA was isolated from cells of A. ferrooxidans grown to mid-log phase in 9 K medium supplemented with sulfur, as described previously [23]. The RNA preparation was treated with DNase I (Fermentas) before proceeding with the cDNA synthesis step. One microgram of total cellular RNA was used for each reaction. Reverse transcription-PCR (RT-PCR) was performed on purified RNA using the One-Step RT-PCR kit (Qiagen). The sequences of the RT and PCR primers used are provided in Table 2. As controls, reactions were carried out that included RNA but lacked reverse transcriptase to assess genomic DNA contamination and that lacked RNA but contained 1 ng of genomic DNA.

Cloning and expression of cbbR
A DNA fragment corresponding to the coding region of cbbR of A. ferrooxidans was amplified by PCR using primers (Integrated DNA Technologies) cbbRfw and cbbRrev ( Table 2). The amplified product was cloned into the expression vector pBAD-TOPO (Invitrogen) according to the manufacturer's instruction. The resulting plasmid pBAD-cbbR was introduced by electroporation into E. coli TOP10 (Invitrogen) competent cells [22]. E. coli was grown at 37°C in 10 ml LB containing 100 μg/ml ampicillin to an OD 600 of 0.8. Overproduction of the recombinant His 6 -tagged CbbR protein was initiated by adding arabinose to a final concentration of 0.1% (w/v) with continued shaking at 200 rpm for 12 h.

Purification of CbbR
Cells from 1.5 l of induced culture were harvested by centrifugation (8,000 × g for 10 min at 4°C) and at -20°C. After thawing the cell pellet was resuspended in 40 ml denaturing buffer containing 6 M guanidine-HCl, 100 mM NaH 2 PO 4 and 10 mM Tris-HCl, pH 8.0, and incubated at room temperature with continuous stirring for about 30 min until inclusion body proteins were solubilized. Any remaining insoluble material was removed by centrifugation at 18,000 × g and 7°C for 20 min. The resulting supernatant was filtered through a 0.45-μm membrane and the recombinant protein subsequently purified by affinity chromatography on a 2.5ml Ni-nitrilotriacetic acid column under amalgam conditions (denaturing conditions-native conditions  [24], with bovine serum albumin as a standard. CbbR was stored at -20°C.

Production of antisera to CbbR
Multiple intradermal injections were applied to immunize a female Californian giant rabbit (3.0 kg) as described by [25]. A fresh CbbR preparation (0.5 ml; 1 mg/ml) was emulsified in one volume of complete Freund adjuvant (Commonwealth Serum Laboratories, Melbourne, Australia). The emulsion was prepared under aseptic conditions and 1.0 ml was initially injected into four sites on the back of the animal. Booster injections were given in the same way 75 days after the primary immunization, except that incomplete Freund adjuvant was used. The immune response was monitored by Western Blotting assays with serum separated from test blood samples (1.0 to 2.0 ml) that were obtained from an ear vein every 15 to 20 days after each immunization.

Electrophoretic mobility shift assays (EMSA)
DNA fragments containing the four potential cbb operon promoter regions were amplified by PCR and simultaneously biotinylated using the biotin 5'-labelled primers (

Bioinformatic analyses
Metabolic pathways involved in CO 2 assimilation were retrieved from KEGG http://www.genome.ad.jp/kegg/. Protein sequences derived from known genes involved in CO 2 assimilation were used as query sequences to search the genome sequence of A. ferrooxidans ATCC 23270, using TBlastN and BlastP, respectively, with default parameters. When a prospective candidate gene was identified, its predicted protein sequence was then used to formulate a BlastP http://www.ncbi.nlm.nih.gov search of the nonredundant database at NCBI. Only bidirectional best hits were accepted as evidence for putative orthologs. Candidate genes and their translated proteins were further characterized employing the following bioinformatic tools: ClustalW [26] for primary structure similarity relations, PSI-PRED [27] for secondary structure predictions, Prosite [28] for motif predictions, ProDom [29] and Pfam [30] for domain predictions. Information regarding the organization of genes in A. ferrooxidans was obtained from [2]. Logos were generated using the web-based application available at http://weblogo.berkeley.edu/logo.cgi. The height of each letter in bits corresponds to its relative abundance at each position. Promoters of the σ 70 -type and rho-independent transcriptional stops were predicted for operons cbb1-4 using the programs BPROM http://www.softberry.com and Transterm [31], respectively. The organization of gene clusters in facultative and obligate autotrophs involved in the CBB cycle was derived from information available in IMG-JGI http:// www.jgi.doe.gov/ and MicrobesOnline http://www. microbesonline.org/, with additional information added for H. marinus [18] and A. ferrooxidans, Acidithiobacillus caldus and Acidithiobacillus thiooxidans (this study). The phylogenetic cladogram of these bacteria was constructed from 16 S rRNA sequences obtained from KEGG Orthology K01977 http://www.genome.jp/kegg/ko.html and from GenBank http://www.ncbi.nlm.nih.gov/ for A. caldus (GI454888), A. thiooxidans (GI454888) and H. marinus (GI3882094). 16 S rRNA alignments were carried out using ClustalW and the cladogram was constructed by the NJ method using the program MEGA 4.0 [32]. The robustness of the tree was evaluated by bootstrapping using 1000 replicas. The tree was rooted using the 16 S rRNA of the ε-proteobacterium Helicobacter pylori.
The cbbR-cbbL1 intergenic region of A. ferrooxidans strain Fe1 has been shown to contain divergent σ 70 -type promoters and to exhibit two CbbR binding sites that partially overlap these promoters ( [4], Figure 1A). The binding sites conform to the pseudo-palindromic motif TNA-N 7 -TNA [13] that is a subset of the consensus LysR-type transcription factor binding site T-N 11 -A [37]. Logos were derived from a multigenome comparison of the cbbR-cbbL1 intergenic region of a number of bacteria (Additional file 3) and were aligned with the CbbR sites of A. ferrooxidans strain Fe1, allowing the prediction of the CbbR binding sites of A. ferrooxidans ATCC 27230 ( Figure 1B and 1C).
Organization and expression of gene clusters predicted to be involved in CO 2 fixation and associated pathways of central carbon metabolism A cluster of 16 genes, termed cbb1, was predicted to be involved CO 2 fixation. RT-PCR experiments showed that cbb1 is transcribed as a single unit and thus can be considered to be an operon (Figure 2A). Operon cbb1 consists of cbbL1 and cbbS1, potentially encoding the large and small subunits of form IAc RubisCO, seven cso genes predicted to be involved in α-carboxysome formation, two genes (cbbQ1 and cbbO1) presumed to be involved in RubisCO activation and cbbA, potentially encoding a fructose-1,6-bisphosphate aldolase. Gene descriptions are provided in Table 3.
Three additional gene clusters termed cbb2 (four genes), cbb3 (twelve genes) and cbb4 (five genes) were identified that are predicted to encode functions related to CO 2 fixation and central carbon metabolism (Table  3). RT-PCR experiments revealed that gene clusters cbb2, cbb3 and cbb4 are transcribed as single units, respectively, and thus constitute operons ( Figure 2B-D). cbb2 contains genes (cbbL2 and cbbS2) encoding additional copies of the large and small subunit of form IAq RubisCO and associated RubisCO activation genes (cbbQ2 and cbbO2) ( Figure 2B). The deduced amino acid sequences of these genes are similar but not identical to the corresponding proteins encoded in the cbb1 operon; CbbL1 and CbbL2 exhibit 84% amino acid sequence identity, whereas CbbS1 and CbbS2 share 56% identity and CbbQ1 and CbbO1 have 84% and 59% identity with CbbQ2 and CbbO2, respectively.
Genes predicted to be encoded by operons cbb3 and cbb4 are listed in Table 3 and their organization within these operons is shown in Figure 2.
The two enzymes that are unique to the CBB cycle are RubisCO (encoded by operons cbb1 and cbb2) and phosphoribulokinase (encoded by operon cbb4). RuBisCO catalyzes the first step of the cycle, the carboxylation of ribulose-1,5-bisphosphate (RuBP) with CO 2 . Phosphoribulokinase catalyzes the last step of the cycle which is the regeneration of the CO 2 acceptor molecule, RuBP, by phosphorylation of ribulose 5-phosphate with ATP. Other steps of the cycle, encoded in operon cbb3, are catalyzed by enzymes common to glycolytic and gluconeogenic pathways in central carbon metabolism [8,36].
Promoters of the σ 70 -type and rho-independent transcriptional stops were predicted for operons cbb1-4 ( Figure 2). In addition, potential CbbR-binding sites were identified in the four operons based on the detection of conserved TNA-N 7 -TNA and T-N 11 -A motifs described above for operon cbb1 (Figure 2).
CbbR binds in vitro to the predicted s 70 -like promoter regions of operons cbb1-4 Binding of CbbR to DNA fragments containing the predicted promoters of the four operons cbb1-4 was evaluated in vitro by electrophoretic mobility shift assays (EMSAs). For this purpose the cbbR gene was cloned and expressed in E. coli. Purified CbbR was used to prepare antisera (anti-CbbR antibodies) whose activity was checked by Western blotting against purified CbbR (data not shown). Biotin-labeled promoter DNA for the EMSA assays was prepared by PCR using primers specified in Table 2 and whose locations within the four operons are shown in Figure. 2. Results show that CbbR was able to retard the promoter regions of the cbb1, cbb2 and cbb3 operons but not the cbb4 operon (Figure 3). When a 50-fold molar excess of unlabelled fragment was included in the binding assay retardation of the labelled fragments was abolished. Furthermore, the addition of anti-CbbR antibodies to the reaction produced a supershift in migration, indicating that the shift was caused specifically by the binding of CbbR.
Binding of CbbR to the predicted promoter regions of operons cbb1-3 suggests that it is involved in their regulation. The reason for the failure of CbbR to retard the DNA fragment containing the predicted promoter of the cbb4 operon is not known. Perhaps this fragment requires the presence of additional factors for CbbR binding that are not present in the in vitro cocktail used for the EMSA analysis. Alternatively, the predicted CbbR binding site is not functional.

Gene organization of the cbb operons
The cbb3 operon includes not only genes involved in carbon assimilation but also harbors genes with similarity to trpE and trpG that are predicted to encode the components I and II of anthranilate synthase, the first enzyme of the tryptophan biosynthesis pathway. Anthranilate synthase catalyzes the conversion of chorismate to anthranilate with the concomitant release of pyruvate [38,39]. In some cases, this conversion can be accomplished by TrpE alone [40].
In order to determine if the association between trpEG and the cbb genes is restricted to A. ferrooxidans, an examination of gene organization was carried out in all sequenced genomes of facultative and obligate autotrophic proteobacteria. Twenty-six proteobacterial organisms (11 α-, 7 βand 8 γ-) were analyzed, including 10 obligate autotrophs. Linkage between trpE/G and cbbE and/or cbbZ was found in all sequenced obligate autotrophs, all of which belong to the βor  Table 3.   γ-proteobacteria divisions ( Figure 4, Table 4), whereas only 4 out of 14 facultative heterotrophs were detected with this clustering. These four exceptions are found only in the βor γ-proteobacteria and none in the αproteobacterial division (Figure 4, Table 4). This suggests a previously unreported linkage between genes encoding CBB cycle associated enzymes and trpEG or trpE that is most conserved in obligate autotrophs of the βand γ-proteobacteria. We hypothesize that in A. ferrooxidans production of pyruvate via anthranilate synthase activity provides a novel network connection between the CBB cycle on the one hand and general central carbon metabolism including the incomplete ("horseshoe"-like) TCA [2] on the other hand. Consistent with this idea is the presence of a predicted pykA upstream of trpEG in the cbb3 operon. PykA is predicted to encode pyruvate kinase that catalyzes the conversion of phosphoenol pyruvate (PEP) to pyruvate. In addition to supplying pyruvate, PykA could also reduce the level of intracellular PEP. PEP has been shown to be a ligand of CbbR in Ralstonia eutropha H16, promoting its binding to target DNA sites and consequently effecting the regulation of cbb genes [40]. If PEP carries out a similar function in A. ferrooxidans, the depletion of PEP via PykA activity could provide a means for feedback control of operons that are regulated by CbbR, including the auto-regulation of operon cbb3.
The organization of cbb genes in A. ferrooxidans exhibits similarities with obligate autotrophs that distinguish this group from facultative autotrophs. For example, A. ferrooxidans, contains three or more gene clusters dedicated to carbon assimilation. This is similar to other obligate autotrophic γ-proteobacteria including A. caldus, A. thiooxidans, Hydrogenovibrio marinus, Nitrosococcus oceani and Thiomicrospira crunogena, and obligate autotrophic β-proteobacteria such as Nitrosomonas europaea, Nitrosomonas eutropha, and Nitrosospira multiformis and Thiobacillus denitrificans. This contrasts with facultative autotrophs that contain only one or two cbb clusters (Figure 4, Table  4), with some exceptions, e.g. the α-proteobacteria Bradyrhizobium sp., N. hamburgensis, N. winogradski. R. sphaeroides and R. palustris and the β-proteobacterium R. eutropha, which contain unique, but duplicated, cbb clusters). Multiple cbb clusters could provide obligate autotrophs with a greater flexibility in regulating CO 2 fixation compared to facultative autotrophs. For example, this flexibility may be necessary to adjust carbon assimilation in response to changing environmental concentrations of CO 2 [18], whereas facultative autotrophs might be able to circumvent this need by exploiting organic carbon sources in times of low CO 2 concentrations.
Another characteristic of cbb gene organization in A. ferrooxidans is the lack of linkage of the phosphoribulokinsae gene, cbbP, with other cbb genes ( Figure 4, Table 4) as has previously been reported for the deepsea vent obligate chemolithoautotroph T. crunogena XCL-2 and for several other obligate autotrophs [20,41]; we now extend this list to include A. ferrooxidans ATCC 23270 and ATCC 53993, A. caldus, A. thiooxidans H. marinus, N. europaea and Thiomicrospira crunogena ( Figure 4, Table 4). In contrast, in all sequenced facultative autotrophs cbbP is associated with other cbb genes ( Figure 4, Table 4).
In obligate autotrophs, the contextual disconnection of cbbP from cbbLS could provide greater flexibility for CO 2 fixation by allowing RubisCO to be differentially expressed according to environmental and/or metabolic requirements without simultaneously expressing the remaining CBB cycle genes, many of which carry out functions in addition to carbon fixation. This is in sharp contrast to the organization found in most facultative autotrophs, where cbbP is usually juxtaposed to cbbLS and other genes of the CBB cycle facilitating their coordinate repression during heterotrophic growth [13,20,34,36,41].

Model for predicted enzymes and pathways involved in CO 2 fixation
A model is proposed for C i fixation in A. ferrooxidans based on the predicted roles of the genes encoded in the cbb operons ( Figure 5). In contrast to most facultative autotrophs, the main focus of regulation of the CBB cycle in A. ferrooxidans may be the CO 2 fixation reaction itself catalyzed by RubisCO, rather than at the level of the other CBB cycle enzymes. This hypothesis is supported by the observation that the genes encoding RubisCO and RubisCo accessory proteins, are clustered in operons that do not contain cbbP nor cbb that encode the main CBB enzymes. cbbP is also separated from the rest of the cbb genes in the cbb4 operon, with an apparent absence of CbbR binding to its promoter. We suggest that the promoters for the cbb1, cbb2 and cbb3 operons have different affinities for CbbR and may thus exhibit different regulation patterns, possibly associated with the environmental availability of CO 2 . The cbb1 operon, containing cbbLS-cso, is predicted to serve at low CO 2 concentrations because carboxysomes have been shown to improve RubisCO catalytic efficiency by concentrating CO 2 [6,13]. In contrast, the cbb2 operon, containing cbbLSQO, is predicted to be used when higher concentrations of CO 2 are available since carboxysome synthesis is energetically and materially expensive [18].
The cbb3 operon, containing genes for most CBB cycle enzymes and pyruvate kinase, is proposed to be Figure 4 Organization of gene clusters involved in the CBB cycle of facultative and obligate autotrophic a-, band g-proteobacteria presented as a phylogenetic cladogram based on 16 S RNA. Numbers refer to bootstrapping results from 1000 trees. Organism names are provided in the text. The asterisk indicates that the respective organism is an obligate autotroph. responsible for connecting CO 2 fixation with the rest of central carbon metabolism. Except for cbbG and cbbK encoding glyceraldehyde-3-phosphate dehydrogenase, type I and phosphoglycerate kinase respectively, genes of the cbb3 operon have duplicated copies in the genome (data not shown), potentially allowing regulation of the CBB cycle independently of the remaining pathways of central carbon metabolism. For example, some CBB cycle intermediates also form part of gluconeogenesis and glycolysis resulting in the production of pyruvate that is channeled, via the pyruvate dehydrogenase complex, into the incomplete TCA "horseshoe" where the flux of intermediates serves for amino acid biosynthesis (e.g. glutamate). The pyruvate dehydrogenase also provides acetyl-CoA used in fatty acid biosynthesis. In addition, the presence of cbbZ in the cbb3 operon is associated with phosphoglycolate phosphatase activity, responsible for removal of phosphoglycolate, an undesirable product of the oxygenase activity of RubisCO, that must be detoxified preferentially by rechanneling to 3-phosphoglycerate [13,36].
The co-transcriptional connection between the cbb, pykA and trpEG genes in the cbb3 operon may reflect the substrate requirement of anthranilate phosphoribosyltransferase for an activated pentose (5-phosphoribosyl 1-pyrophosphate) in order to proceed to the next step of tryptophan biosynthesis [42]. The production of the activated pentose would be stimulated by the activity of No. copies cbbR Figure 5 Proposed roles of the (A) predicted enzymes and pathways involved in CO 2 fixation in A. ferrooxidans linked to (B) gene evidence. Genes are color-coded to match the predicted function of their products. RPI, ribose phosphate isomerase; G-3-P, glyceraldehyde-3phosphate; DHAP, dihydroxyacetone phosphate; 3-PG, 3-phosphoglycerate; PEP, phosphoenolpyruvate.
the operon. An alternate hypothesis is that the co-transcriptional connection represents a means for pyruvate regeneration since both pykA and trpE/G produce pyruvate.
In addition to the four cbb operons described herein, a fifth gene cluster has recently been detected in A. ferrooxidans that includes genes cbbM, cbbQ3 and cbbO3 predicted to encode form II of RubisCO and its associated chaperons, respectively [43]. The cluster also contains another putative cbbR divergently transcribed from cbbMQO. Future work will evaluate the role of this cluster in CO 2 fixation.