ABC Transporter for Corrinoids in Halobacterium sp. Strain NRC-1

ABSTRACT We report evidence for the existence of a putative ABC transporter for corrinoid utilization in the extremely halophilic archaeon Halobacterium sp. strain NRC-1. Results from genetic and nutritional analyses of Halobacterium showed that mutants with lesions in open reading frames (ORFs) Vng1370G, Vng1371Gm, and Vng1369G required a 105-fold higher concentration of cobalamin for growth than the wild-type or parent strain. The data support the conclusion that these ORFs encode orthologs of the bacterial cobalamin ABC transporter permease (btuC; Vng1370G), ATPase (btuD; Vng1371Gm), and substrate-binding protein (btuF; Vng1369G) components. Mutations in the Vng1370G, Vng1371Gm, and Vng1369G genes were epistatic, consistent with the hypothesis that their products work together to accomplish the same function. Extracts of btuF mutant strains grown in the presence of cobalamin did not contain any cobalamin molecules detectable by a sensitive bioassay, whereas btuCD mutant strain extracts did. The data are consistent with the hypothesis that the BtuF protein is exported to the extracellular side of the cell membrane, where it can bind cobalamin in the absence of BtuC and BtuD. Our data also provide evidence for the regulation of corrinoid transport and biosynthesis. Halobacterium synthesized cobalamin in a chemically defined medium lacking corrinoid precursors. To the best of our knowledge, this is the first genetic analysis of an archaeal corrinoid transport system.

Corrinoids belong to the family of cyclic tetrapyrroles that includes hemes, chlorophylls, and coenzyme F 430 (16,48). A complete corrinoid (also called cobamide) has upper and lower ligands that play important biochemical roles (16). The upper ligand forms a labile, covalent bond with the cobalt ion of the corrin ring (Co-C), while the lower ligand interacts with the cobalt ion via a coordination bond. The best-known cobamide is cobalamin (Cbl), which in its biologically active form has a 5Ј-deoxyadenosyl group as an upper ligand, hence the name adenosylcobalamin or coenzyme B 12 . Cobamides are distinguished from one another by the nature of the lower ligand nucleotide base (39), which is 5,6-dimethylbenzimidazole (DMB) in the case of Cbl.
Because of the complex structure of corrinoids, biosynthesis of the complete Cbl molecule requires at least 24 genes (48). Only prokaryotes synthesize corrinoids, although many eukaryotes, including humans, require corrinoids for their metabolism (39,40,48). Active transport of corrinoids is a process found in both prokaryotes and eukaryotes. Because the levels of corrinoids in the environment are low, transport of corrinoids requires specific systems with high affinity. In prokaryotes, most of the work on corrinoid transport has been performed with the gram-negative bacteria Escherichia coli and Salmonella enterica (9,13,37,40,47). Corrinoid transport is a special problem to these bacteria because the molecule must pass through both an outer and an inner membrane and the periplasm (13). Transport across the outer membrane requires both the BtuB and TonB proteins (18,36). Active transport across the inner membrane is achieved via an ATP-binding cassette (ABC) transport system encoded by the btuC, btuD, and btuF genes, which encode the membrane permease, ATPase, and periplasmic-binding protein components, respectively (4,9,12,47). ABC transporters are widely distributed in all domains of life and drive the translocation of substrates across membranes by the hydrolysis of ATP.
No corrinoid transport systems have been described for archaea, although some archaea synthesize and require corrinoids for survival. For example, methanogenic archaea require cobamides for methanogenesis from H 2 and CO 2 , acetate, or methanol (14). Active cobamide-dependent (class II) ribonucleotide reductases have been purified from both Thermoplasma acidophilum (45) and Pyrococcus furiosus (38), suggesting cobamides are used by these organisms. We recently showed that the extremely halophilic archaeon Halobacterium sp. strain NRC-1 requires corrinoids under certain growth conditions, although the reasons for their corrinoid requirement remain unknown (50). We also showed that Halobacterium salvages Cbl and several of its precursors when present in the medium at subnanomolar concentrations, suggesting that this archaeon possesses a high-affinity corrinoid transport system (51).
Based on genome sequence analyses, ABC transporters appear to be as ubiquitous in archaea as in bacteria (2). Therefore, we hypothesized that, like bacteria, archaea use an ABC transporter for the utilization of corrinoids. Substrate uptake systems of this type have been studied in Sulfolobus solfataricus, P. furiosus, and Thermococcus litoralis and have been shown to be composed of a permease, an ATPase, and an extracellular substrate-binding protein that is anchored to the cell membrane (1,3,15,20,22,52). Presumably because archaea only have a single membrane and no periplasm, no orthologs of outer membrane transporters have been found.
Using Halobacterium sp. strain NRC-1 as a model system, we report genetic evidence of an ABC-type corrinoid transporter in archaea. The Vng1370G, Vng1370Gm, and Vng1369G genes were predicted to encode the archaeal orthologs of the bacterial BtuC, BtuD, and BtuF proteins, respectively. These functions were required to salvage low nanomolar levels of corrinoids from the environment of this archaeon. We also report evidence for the regulation of this transport system and demonstrate that Halobacterium synthesizes Cbl (with the lower ligand DMB) de novo.

MATERIALS AND METHODS
Strains and plasmids. The genotypes of the Halobacterium and S. enterica strains and plasmids used in this work are described in Table 1.
Halobacterium growth studies. Strains were grown in liquid rich peptone (RP) medium (Oxoid, Hampshire, England) (28) lacking trace metals. Halobacterium cultures were grown for 4 days to stationary phase at 37°C with shaking. Cells were added to 5 ml chemically defined (CD) medium (17) at a dilution of 1:100, and cultures were grown at 37°C with shaking. Briefly, the CD medium contains the amino acids A, R, C, E, G, I, L, K, M, F, P, S, T, Y, and V; 11 mM glycerol; salts; and trace metals. Growth was monitored every 24 h for 6 days by measuring the absorbance of the culture at 650 nm with a Spectronic 20D spectrophotometer (Milton Roy, Rochester, NY). To determine cell viability (calculated as CFU), cells were plated onto solid RP medium (6.6% [wt/vol] agar). In all cases, media were supplemented with uracil (450 M).
Halobacterium plasmid constructions. Plasmids were propagated in E. coli strain DH5␣, except where noted otherwise. Unless stated otherwise, Halobacterium strain MPK414 (⌬ura3) genomic DNA was used as the template for PCR and was prepared as previously described (50). The high-fidelity enzyme Pfu (Stratagene) was used for PCR amplification. All DNA fragments were digested with the appropriate restriction enzymes (indicated by the underlined portion in the name of the primers) and then gel purified using a QIA quick gel extraction kit (QIAGEN). Plasmid pMPK424 was prepared from the E. coli dam mutant strain GM2163 (New England Biolabs). All primers were purchased from Integrated DNA Technologies. Underlined portions of the primer sequences (see below) indicate introduced restriction sites. All plasmids were subsequently sequenced for verification. A diagram of the Halobacterium sp. strain NRC-1 DNA included in the most relevant plasmids is included in Fig. 1. Plasmid pVNG1370-2. The primer sets VNG1370DelXbaI5Ј (TCTAGATCT AGACGTCGGCAGCGATGTTGTGG)-VNG1370DelNcoI3Ј (CCATGGCCA TGGGCGCAGCGTGATCGGTTCC) and VNG1370DelNcoI5Ј (CCATGGCC ATGGGCTGTCGTGTCCGCAGTCG)-VNG1370DelHindIII3Ј (AAGCTTA AGCTTACGAGCGTGATGGTCTGTCC) were used to PCR amplify 890-bp and 760-bp fragments, respectively. The former fragment was cloned into the XbaI/NcoI restriction sites of plasmid pMPK428 (32). The second fragment was then cloned into the NcoI/HindIII restriction sites of the resulting plasmid to create plasmid pVNG1370-2, which contains an in-frame deletion of btuC and btuD replacing bases 112 to 1110 of btuC and bases 1 to 1143 of btuD with a 6-bp NcoI restriction site. The gene product of this construct should be a nonfunctional peptide with amino acid residues 1 to 37 of BtuC fused to amino acid resides 382 to 398 of BtuD.
Plasmid VNG1370-4. The primer set VNG1370-Comp-XbaI-5Ј (TCTAGTTC TAGATGTGATCGCGGTGTTGCTGG)-VNG1371-Comp-BglII-3Ј (AGATC AAGATCTTGGCTGCCGTGCGACCCATG) was used to PCR amplify a 2,620-bp fragment that was cloned into the XbaI/BglII restriction site of plasmid pT7-7 (44). This cloned insert was excised with an XbaI/BglII restriction enzyme digest from a plasmid prepared from strain GM2163. The DNA fragment was then cloned into the XbaI/BglII sites of plasmid pMPK424 (31). The resulting plasmid, pVNG1370-4, contains the putative operon containing Vng1370G and Vng1371Gm. In addition to the two open reading frames (ORFs), 225 bp of genetic material 5Ј of Vng1370G and 70 bp 3Ј of Vng1371Gm were included to preserve any transcriptional regulation.
Plasmid pHsBTUC5. The primer set VNG1370-Comp-XbaI-5Ј-Hs-BTUC-Comp-BglII-3Ј (ACATCAAGATCTAAAAGCCGCGCCGGTTGCCAACTCC ACGTCGAGG) was used to PCR amplify a 1,370-bp fragment that was cloned into the XbaI/BglII sites of pMPK424 (31). The resulting plasmid contains the entire btuC ORF, 225 bp 5Ј of btuC to preserve transcriptional regulation, and a 16-bp sequence derived from the bop transcriptional terminator (11) (included in the 3Ј reverse primer) to ensure termination of the btuC mRNA transcript.
Plasmid pHsBTUD1. The primer sets VNG1370-Comp-XbaI-5Ј-Hs-BtuD-Comp-EcoRI-3Ј (TGTTCTGAATTCAACGGTGCGCAGCGTGATC) and Hs-BtuD-Comp-EcoRI-5Ј (TGTTCAGAATTCATCATCACCGCCCTGATCG)-Hs-BtuD-Comp-BglII-3Ј (ACATCAAGATCTAAAAGCCGCGCCGGTTGTC CACGTAATACGTTCC) were used to PCR amplify 340-bp and 1,310-bp DNA fragments, respectively. Both DNA products were cut with EcoRI restriction enzyme and ligated together with T4 ligase (MBI Fermentas, Amherst, NY). Using the ligated DNA as the template, the primer set VNG1370-Comp-XbaI-5Ј-Hs-BtuD-Comp-BglII-3Ј was used to PCR amplify a 1,650-bp DNA fragment, which was cloned into the XbaI/BglII sites of plasmid pMPK424 (31). The resulting plasmid contains a wild-type allele of btuD, as well as the 225 bp 5Ј of btuC to include the transcription start site. To include this sequence, as well as 70 bp 5Ј of the btuD ORF (to include the ribosome-binding site), part of btuC was included but as an in-frame deletion. A 6-bp EcoRI restriction site replaced bases 118 to 1041 of btuC. This construct should not encode a functional BtuC peptide but should ensure the production of a btuD mRNA transcript. As described for plasmid pHsBTUC5, a transcriptional terminator sequence was included 3Ј of the btuD ORF.
Halobacterium corrinoid extraction assays. Ten milliliters of dense Halobacterium culture was used to inoculate 1 liter of liquid RP or CD medium supplemented with various concentrations of Cbl. The cultures were grown to full density (4 days in RP medium and 6 days in CD medium) at 37°C with shaking at 180 rpm. Serial dilutions of the cells were plated on solid medium to calculate total CFU. Cells were harvested at 4,300 ϫ g for 10 min in a Beckman-Coulter J21 centrifuge, washed by gently resuspending them in 200 ml of medium salts  (5) supplemented with glycerol and MgSO 4 . Two microliters of Halobacterium corrinoid extract or 2 pmol of authentic Cbl was spotted onto the agar overlay. The inoculated plates were incubated aerobically at 37°C for 24 h. The last step of cobamide biosynthesis in strain JE873 is blocked, making growth dependent on complete cobamides. Cell growth around the application site on overlays containing strain JE873 would indicate the presence of Cbl or another cobamide in the extract. Strain JE2243 was used as a negative control because it cannot transport Cbl (due to a lesion in the gene encoding the outer membrane corrinoid transporter BtuB) and will not respond to its presence in the extracts.
High-performance liquid chromatography (HPLC) analysis of corrinoids. Halobacterium corrinoid extracts were filtered using Corning Spin-X centrifuge filters. Corrinoids were separated by using a Beckman-Coulter HPLC system equipped with a Luna (Phenomenex) 5-m C 18 column (150 by 4.6 mm) developed with a modification of the system reported elsewhere (6) at a flow rate of 1 ml/min. The column was equilibrated with a buffer system containing 98% A and 2% B. For quantification of Cbl in the extracts, 2 min after injection, the column was developed for 10 min with a linear gradient until the final composition reached 100% B. For the purification of corrinoids for mass spectrometry analysis, the column was developed for 55 min to 100% B 5 min after injection. The solvents used were as follows: A, 100 mM phosphate buffer (pH 6.5)-10 mM KCN; B, 100 mM phosphate buffer (pH 8.0)-10 mM KCN-acetonitrile (1:1). Corrinoids were detected using a Beckman-Coulter photodiode array detector. Authentic Cbl was used as the standard.
Mass spectrometry. The HPLC-purified corrinoid in Halobacterium extracts was prepared for mass spectrometry analysis as previously described (49). This sample, as well as authentic Cbl, was submitted for analysis to the mass spectrometry facility at the University of Wisconsin-Madison Biotechnology Center. The mass spectrum was obtained using a Bruker Daltronics (Billerica, MA) BILFLEX III matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometer.

RESULTS AND DISCUSSION
Halobacterium synthesizes Cbl de novo. To identify the cobamide synthesized by Halobacterium, we extracted corrinoids from strain JE7681 (⌬btuF cbiP ϩ ) grown in CD medium. Strain JE7681 synthesized corrinoids de novo but did not salvage exogenous corrinoids due to the lack of BtuF function (see below). Corrinoids were resolved using HPLC as previously described (see Materials and Methods). A peak with the diagnostic UV-visible spectrum of corrinoids eluted 35.1 min after injection, the same retention time as authentic Cbl (data not shown). The material under this peak was used in bioassays.
S. enterica strain JE873 (metE ⌬cobUST) was the indicator strain used in bioassays to detect the presence of cobamides in Halobacterium corrinoid extracts (26). S. enterica strain JE2243 (metE btuB) lacking the outer membrane corrinoid transporter BtuB protein (19) was used as a negative control. A fraction containing the putative Halobacterium cobamide supported the growth of strain JE873, but not JE2243, consistent with the presence of a cobamide (data not shown). The MALDI-TOF mass spectrum of the Halobacterium cobamide contained a molecular ion signal with an m/z of 1,330.3 that was consistent with Cbl lacking an upper ligand. Authentic Cbl was used as a control, and its spectrum was strikingly similar to that of the Halobacterium cobamide ( Fig. 2A). Other peaks were observed, but these were also present in the mass spectrometry profile of authentic Cbl (Fig. 2B). On the basis of these results, we concluded that Halobacterium sp. strain NRC-1 synthesizes a cobamide with DMB as the lower ␣ ligand (i.e., Cbl).
Methanothermobacter marburgensis strain Marburg is the only other archaeon shown to produce Cbl, but it only did so when DMB was supplied in the medium (43). Cobamides have been isolated with the lower ligand 5-methylbenzimidazole from Archaeoglobus fulgidus and T. acidophilum (23), adenine from Methanosarcina barkeri (35), and 5-hydroxybenzimidazole from M. marburgensis strain Marburg (24). It appears that the differences between the cobamides made by different species of prokaryotes correlate not with the biological functions of the cobamide but with the metabolic conditions of the organism in its natural habitat (23). Halobacterium has an efficient corrinoid transport system. Strain JE6738 (⌬cbiP) was used to assess the ability of Halobacterium to assimilate low concentrations of various corrinoids. Strain JE6738 cannot synthesize corrinoids de novo due to the lack of CbiP (Cby synthase enzyme) and is dependent on exogenous cobamides or corrinoid precursors for growth (50). Strain JE6738 did not grow without added corrinoids, as opposed to strain JE6735 (cbiP ϩ ), which did not require corrinoids to grow (Fig. 3). The growth response of strain JE6738 was assessed as a function of the concentration of incomplete cobamides (i.e., Cby, Cbi, Cbi-GDP) and Cbl. The growth responses of strain JE6738 to all corrinoids tested were very similar. At least 1 nM corrinoid was required for growth equivalent to a functional de novo pathway, while a 100 pM concentration of every corrinoid tested allowed intermediate growth. Concentrations of Cbl of up to 100 M did not significantly increase growth any further and may have a slight inhibitory effect (Fig. 4A, inverted solid triangles). The concentration of corrinoid needed to support optimal growth of Halobacterium was very similar to the one needed for S. enterica and E. coli (4,29). This result strongly suggested the existence of a transport system for corrinoids in Halobacterium.
Bioinformatic analysis of the B 12 utilization (btu) genes of Halobacterium. In bacteria, the btuC, btuD, and btuF genes encode the corrinoid ABC transporter permease, ATPase, and corrinoid-binding periplasmic protein, respectively.
BtuD. The putative Vng1371Gm protein shared 29% identity and 45% similarity with the E. coli BtuD protein.
BtuF. ORF Vng1369G is encoded divergently from the putative btuCD operon and had 28% identity and 44% similarity to the E. coli BtuF protein. The btuF gene did not appear to be part of an operon. The Halobacterium BtuF protein is predicted to have an N-terminal signal peptide that directs it to the extracellular side of the cellular membrane via the Sec pathway (33,34,53). The signal (residues 1 to 21) is predicted to be cleaved after residue Ala24 (8). The BtuF protein has a C-terminal hydrophobic domain (amino acid residues 348 to 367) preceding a stretch of six hydroxylated amino acid residues, which may indicate anchoring to the extracellular side of the cell membrane (2).
A structure-based sequence alignment between Halobacterium BtuF and E. coli BtuF (the latter's crystal structure bound to B 12 has been solved [21]) was used to identify possible B 12 -biding residues. Of the 11 residues in E. coli BtuF that make direct contacts with the B 12 molecule, Halobacterium BtuF has three that are identical (E. coli BtuF residues S8, P9, and A10) and two are similar substituted hydrophobic residues (Y28F and W174Y). Based on this comparison, it is unclear if Halobacterium BtuF plays the same role or binds B 12 like E. coli BtuF.
btuC (Vng1370G), btuD (Vng1371Gm), and btuF (Vng1369G) of Halobacterium are required for corrinoid utilization. Strains JE7084 (⌬cbiP ⌬btuCD) and JE7862 (⌬cbiP ⌬btuF) were used to determine if Halobacterium btuC, btuD, and btuF functions are required for corrinoid utilization. In these strains, the lesion in cbiP blocks de novo cobamide synthesis, thus rendering cell growth dependent on corrinoid transport (50). Strains JE7084 and JE7682 were grown in CD liquid medium with various concentrations of Cbl. These strains failed to grow when provided with 10 nM Cbl, a concentration that was sufficient for growth of JE6738 (⌬cbiP btuCD ϩ btuF ϩ ) (Fig. 4A [open triangles versus inverted closed triangles] and C [closed diamonds versus closed triangles]). No significant growth was observed until the medium was supplemented with 100 M Cbl. Because cells are unlikely to encounter 100 M corrinoid in nature, this observed growth was most likely due to nonspecific transport. These strains were also tested for the ability to assimilate incomplete cobamides. At 10 nM, Cby, Cbi, and Cbi-GDP did not support the growth of either strain, while 1 M Cbl supported wild-type growth (data not shown). The observed block of corrinoid transport in strains JE7084 and JE7682 was corrected when wild-type alleles of Halobac-terium btuC and btuD or btuF were introduced into the chromosome. Strains JE7495 (⌬cbiP ⌬btuCD ura3::btuCD ϩ ) (Fig.  4B, inverted open triangles) and JE7797 (⌬cbiP ⌬btuF ura3::btuF ϩ ) (Fig. 4D, open squares) grew when 1 nM Cbl was added.

Absence of BtuF but not BtuCD functions prevents Cbl-cell association in Halobacterium.
To test if BtuC, BtuD, and BtuF are required for the assimilation of exogenous Cbl, Halobacterium ⌬cbiP mutants were grown in liquid RP medium with and without 100 nM Cbl. RP medium allows cobamide-independent growth, and all of the strains in these studies grew at similar rates in this medium regardless of exogenous corrinoids (data not shown). Possible reasons for why corrinoid auxotrophs grew in RP medium are discussed below.
S. enterica strain JE873 (metE ⌬cobUST) was used to test for Cbl presence in Halobacterium cells. None of the corrinoid extracts from Halobacterium cells grown without Cbl supported the growth of strain JE873 (Fig. 5), indicating that the de novo corrin ring biosynthetic pathway was not functional and that there was no contaminating Cbl in the RP medium. When 100 nM Cbl was added to the medium, Cbl was found in the extracts of strains JE6738 (⌬cbiP), JE7084 (⌬cbiP ⌬btuCD), JE7495 (⌬cbiP ⌬btuCD ura3::btuCD ϩ ), and JE7797 (⌬cbiP ⌬btuF ura3::btuF ϩ ) but not in strain JE7682 (⌬cbiP ⌬btuF) (Fig. 5). Collectively, these results suggested that the btuF gene is required for the presence of Cbl in extracts but that the btuCD genes are not. This phenotype of strain JE7682 was corrected by reintroduction of a wild-type copy of the btuF gene into the chromosome (Fig. 5). None of the spotted extracts supported the growth of S. enterica strain JE2243 (metE btuB), indicating that the growth of JE873 was due specifically to Cbl (results not shown).

5906
WOODSON ET AL. J. BACTERIOL. ml) of a culture grown in RP medium and they accumulated 490 pmol Cbl per g protein (Fig. 6A). Cbl-cell association reached a maximum level when CD medium was supplemented with 1 nM Cbl (1,620 pmol Cbl per g protein), consistent with growth data that showed maximum growth at this concentration (Fig. 4, open triangles). The relationship between CFU and total cell protein did not vary significantly between growth media, suggesting that levels of Cbl can be compared. Strain JE7681 (cbiP ϩ ⌬btuF) was used to determine how much Cbl was synthesized de novo by Halobacterium. When grown in RP medium, strain JE7681 synthesized 70 pmol Cbl per g protein compared to 480 pmol Cbl per g protein when grown in CD medium (Fig. 6A). These data suggest regulation of the corrinoid transport system as a function of nutrient availability. At this point, no specific nutrient(s) to which the cells may respond has been identified. Additionally, no obvious transcriptional regulatory sequences in the DNA sequences 5Ј of either the Cbl biosynthetic or transport genes have been identified (data not shown).
To determine how much a lesion in the btuCD or btuF locus would affect the assimilation of Cbl, strains JE7084 (⌬cbiP ⌬btuCD) and JE7682 (⌬cbiP ⌬btuF) were tested. These strains were grown in RP medium supplemented with 100 nM Cbl. Compared to strain JE6738 (⌬cbiP), the presence of Cbl in strain JE7084 was reduced 69% to 820 pmol Cbl per g protein, whereas no Cbl was detected in strain JE7682 extract (Ͻ30 pmol Cbl per g protein) (Fig. 6B). The phenotypes of strains JE7084 and JE7682 were corrected by reintroduction of the wild-type alleles of btuCD and btuF, respectively (Fig. 6B).
During the quantification studies, Cbl was the only corrinoid detected by HPLC in strains lacking de novo capabilities, suggesting that the Cbl molecules did not have to be modified for usage.
It is possible that in the btuCD mutant strain, Cbl is associated with the cells but is inaccessible to metabolism. The Cbl molecules may still be associated with the BtuF protein, which is predicted to be anchored to the extracellular side of the membrane by its C-terminal hydrophobic domain. Without the BtuC and BtuD proteins in the membrane, BtuF may be binding Cbl molecules but not releasing them. In vitro binding studies with Halobacterium BtuF and Cbl, as well as localization studies, are needed to test if BtuF may be binding Cbl on the outer surface of the cell membrane. The genotype of the strain from which Cbl was extracted is indicated above each column. Under each column, the type of liquid medium (RP, RP medium; CD, CD medium) and the concentration of Cbl added are indicated. Asterisks above a column indicate that the values are below the detection limit of the assay. The mean values of duplicated experiments are reported Ϯ the standard deviations. The strains used were JE6738 (⌬cbiP btuCD ϩ btuF ϩ ), JE7084 (⌬cbiP ⌬btuCD btuF ϩ ), JE7495 (⌬cbiP ⌬btuCD btuF ϩ ura3::btuCD ϩ ), JE7681 (cbiP ϩ ⌬btuF btuCD ϩ ), JE7682 (⌬cbiP ⌬btuF btuCD ϩ ), and JE7797 (⌬cbiP ⌬btuF btuCD ϩ ura3::btuF ϩ ).

VOL. 187, 2005
TRANSPORT OF CORRINOIDS IN ARCHAEA 5907 Conclusions. We have identified a corrinoid transport system in the hyperhalophilic archaeon Halobacterium sp. strain NRC-1. Genes encoding this system were annotated as hemU, hemV2, and hemV1 (27). We suggest a change in their nomenclature to btuC, btuD, and btuF, respectively, to reflect their role in corrinoid transport.
Most other available archaeal genome sequences are predicted to contain orthologs to the btuC, btuD, and btuF genes. Two notable exceptions lacking btuC, btuD, and btuF were Methanothermobacter thermautotrophicus strain ⌬H (42) and Methanopyrus kandleri AV19 (41). Both of these archaea appear to contain genetic information for an entire cobamide de novo biosynthetic pathway. The latter may have evolved to rely on endogenously synthesized corrinoids. However, M. marburgensis strain Marburg, a close relative of M. thermautotrophicus strain ⌬H, has been shown to assimilate exogenous corrinoids (43), suggesting that a nonorthologous transport system exists in this archaeon and thus may exist in other archaea.
Identification of btuC, btuD, and btuF orthologs in other archaea based on sequence analysis alone may be problematic. Corrinoid transport systems have amino acid sequences very similar to ABC-type Fe 3ϩ , siderophore, and heme transport systems. Many archaea have several putative orthologs to these systems, and they are not always encoded in close proximity to each other or cobamide biosynthetic genes, making it difficult to identify or match the components of transport systems. Identification of these transport systems may have to rely on more classical genetic and biochemical approaches, like the ones used in the work reported here.
Putative Cbl-dependent enzymes in Halobacterium. It is unknown why Halobacterium requires corrinoids to grow in CD medium. The ability of RP medium to allow growth of Halobacterium corrinoid mutants suggests that these strains are auxotrophic for a nutrient present in this medium. Analysis of the genome sequence predicts that Halobacterium synthesizes at least three cobamide-dependent enzymes, methylmalonyl-coenzyme A mutase (encoded by ORFs Vng0481G, Vng0653G, and Vng0673G), glutamate mutase (encoded by ORFs Vng2286G and Vng2288G), and class II ribonucleotide reductase (encoded by ORF Vng1644G) (27). These enzymes would likely require adenosylcobalamin as the coenzyme; thus, Cbl would have to be adenosylated after transport by an ATP: co(I)rrinoid adenosyltransferase (CobA in S. enterica). Halobacterium contains a putative cobA ortholog (Vng1574G in Halobacterium), but its function has not been demonstrated experimentally. Nutritional analyses of mutants defective for these enzymes may determine if growth in CD medium requires Cbl.