Cloning, Functional Organization, Transcript Studies, and ...

JOURNAL OF BACTERIOLOGY, Jan. 1996, p. 143–148 0021-9193/96/$04.0010 Copyright q 1996, American Society for Microbiology

Vol. 178, No. 1

Cloning, Functional Organization, Transcript Studies, and Phylogenetic Analysis of the Complete Nitrogenase Structural Genes (nifHDK2) and Associated Genes in the Archaeon Methanosarcina barkeri 227 YUEH-TYNG CHIEN

AND

STEPHEN H. ZINDER*

Section of Microbiology, Wing Hall, Cornell University, Ithaca, New York 14853 Received 4 August 1995/Accepted 30 October 1995

Sibold et al. (25) described two sets of nifH genes in M. barkeri: nifH1, which is a member of cluster II, and nifH2, which was in cluster III. In our previous study (4), we demonstrated that the N-terminal sequence of the purified nitrogenase component 2 from M. barkeri cells grown diazotrophically in Mo-containing growth medium corresponded to that for the predicted gene product of the nifH2 gene. Moreover, DNARNA dot blot analysis demonstrated the presence of transcripts of nifH2 and nifDK2, indicating that this set of genes is expressed under standard conditions. We also determined the sequence of nifD2 and the first 45 amino acids of nifK2 and showed that they also were most similar to their homologs in C. pasteurianum. The only study of the functional organization and transcription of nif genes in methanogens was that of Souillard and Sibold (27). They showed that between nifH1 and nifD1 in Methanococcus thermolithotrophicus were two small open reading frames (ORFs) encoding predicted polypeptides with lengths of 105 and 125 amino acids. ORFs resembling these have been found between the nifH and nifD genes in all known functional methanogen nitrogenase clusters, and the predicted products have been shown to resemble the glnB gene product in enteric bacteria (25), also called the PII protein, which can be reversibly uridylylated and plays a role in nitrogen-regulatory cascades including adenylylation of glutamine synthetase and phosphorylation of the ntrC gene product (16, 21). Northern (RNA) blot analysis of M. thermolithotrophicus RNA by using a nifH probe indicated a transcript 1.8 kb long, which was not long enough to harbor the nifDK genes. Primer extension analysis showed a transcription start for nifH1 preceded by a

Nitrogen fixation is known to be widespread in both eubacteria and methanogenic archaea. All known nitrogenases consist of two components: component 1, dinitrogenase, or the MoFe protein (except in alternative nitrogenases lacking Mo), an a2b2 tetramer encoded by nifD and nifK, and component 2, dinitrogenase reductase, or the Fe protein, a homodimer encoded by nifH. Related to the nifDK genes are the nifEN genes, which encode a protein resembling component 1 which is believed to serve as a ‘‘scaffold’’ for synthesis of the FeMo cofactor, which is then inserted into component 1 (5). Phylogenetic analysis by ourselves (4) and others (11, 17, 18, 30) has indicated that functional nitrogenase genes form three families, cluster I, consisting of most conventional eubacterial MoFe nitrogenases; cluster II, containing eubacterial alternative nitrogenases lacking Mo and several methanogen nitrogenases; and cluster III, containing the Mo nitrogenase (nif-1) genes from the gram-positive eubacterium Clostridium pasteurianum and the nif-2 genes from the archaeon Methanosarcina barkeri. Another recently recognized member of this cluster is nifH from Desulfovibrio gigas (29a). A fourth cluster contains genes from methanogens which are likely to serve in functions other than nitrogen fixation (4). Also related to nifH are bclX, bclL, and chlL genes, which encode Fe proteins involved in bacteriochlorophyll or chlorophyll ring reduction (3). These groupings raise interesting questions about the roles of horizontal gene transfer and ancient gene duplications in nitrogenase evolution. * Corresponding author. Phone: (607) 255-2415. Fax: (607) 2553904. Electronic mail address: [email protected]. 143

Downloaded from http://jb.asm.org/ on September 5, 2017 by guest

Determination of the nucleotide sequence of the nitrogenase structural genes (nifHDK2) from Methanosarcina barkeri 227 was completed in this study by cloning and sequencing a 2.7-kb BamHI fragment containing the 3* end of nifK2 and 1,390 bp of the nifE2-homologous genes. Open reading frame nifK2 is 1,371 bp long including the stop codon TAA and encodes a polypeptide of 456 amino acids. Phylogenetic analysis of the deduced amino acid sequences of the nifK2 and nifE2 gene products from M. barkeri showed that both genes cluster most closely with the corresponding nif-1 gene products from Clostridium pasteurianum, consistent with our previous analyses of nifH2 and nifD2. The nifE gene product is known to be homologous to that of nifD, and our analysis shows that the branching pattern for the nifE proteins resembles that for the nifD product (with the exception of vnfE from Azotobacter vinelandii), suggesting that a gene duplication occurred before the divergence of nitrogenases. Primer extension showed that nifH2 had a single transcription start site located 34 nucleotides upstream of the ATG translation start site for nifH2, and a sequence resembling the archaeal consensus promoter sequence [TTTA(A/T)ATA] was found 32 nucleotides upstream from that transcription start site. A tract of four T’s, previously identified as a transcription termination site in archaea, was found immediately downstream of the nifK2 gene, and a potential promoter was located upstream of the nifE2 gene. Hybridization with nifH2 and nifDK2 probes with M. barkeri RNA revealed a 4.6-kb transcript from N2-grown cells, large enough to harbor nifHDK genes and their internal open reading frames, while no transcript was detected from NH41-grown cells. These results support a model in which the nitrogenase structural genes in M. barkeri are cotranscribed in a single NH41-repressed operon.

144

CHIEN AND ZINDER

typical archaeal promoter [TTTA(A/T)ATA] (20) and a transcription start before the nifD1 sequence which did not have a detectable promoter sequence upstream from it. Thus, these results suggest that nifH and nifD are not cotranscribed, but the mode of transcription of nifD is unclear. In this publication, we complete the sequence of nifK2 from M. barkeri, making this the first complete set of structural genes from a methanogen sequenced, and show that these genes are transcribed as an operon preceded by an archaeal promoter. We also define potential genes upstream and downstream of the structural genes. Finally, we examine the phylogenetic relationships for several of these genes, including a possible rooting of the nifD gene tree by nifE genes. MATERIALS AND METHODS

Primer extension experiments. Primer extension experiments were performed as described by Sambrook et al. (23). The primers were a 21-mer oligonucleotide (59-TCCAATTCCCACCCTTTTCCG-39) complementary to bp 21 to 41 downstream of the 59 end of nifH2 and a 21-mer oligonucleotide (59-CGGGAG CACTTTTAACATATC-39) complementary to bp 52 to 72 downstream of the 59 end of nifD2. The primers were 59 end labeled with [g-32P]ATP (3,000 Ci/mmol) and T4 polynucleotide kinase according to the manufacturer’s instructions. RNA (30 to 50 mg) was mixed with 104 to 105 cpm of the labeled primers, denatured at 858C for 10 min, and then annealed to the RNA template by slow cooling of the mixture to 308C in 1 h. The suspension was mixed with a solution (33 ml) containing 6 ml of 53 reverse transcriptase buffer, 5 ml of 0.1 M dithiothreitol, 20 ml of a 2.5 mM deoxynucleosidetriphosphate mix, 1 ml of RNasin (40 U), and 1 ml of avian myeloblastosis virus reverse transcriptase (8 U/ml). The reaction mixture was incubated at 378C for 1 h and precipitated with 1/10 volume of sodium acetate and 2.5 volumes of ethanol. The pellet was resuspended in 4 ml of TE (pH 7.4)–6 ml of formamide loading buffer (Sequenase stop buffer; U.S. Biochemicals). The samples were denatured at 858C for 5 min before loading, and 4 ml of the mixture was loaded on a 6% polyacrylamide sequencing gel containing 8 M urea. A sequencing ladder, with the same primer in a Sequenase reaction, provided the markers for accurate determination of the transcript initiation site. Phylogenetic analysis. Amino acid sequences were used instead of DNA sequences to eliminate biases due to different G1C ratios (18) and were obtained from the combined GenBank-Swissprot-PIR database. Amino acid sequences were aligned by using ClustalW (15), and then a few minor changes were made manually. The PHYLIP 3.5c phylogeny inference software package (10), in the form of compiled executable programs for Macintosh computers, was used for comparison of the protein sequences. The primary sequence analysis was done by using the PROTDIST program with a Dayhoff amino acid comparison matrix. This program gave distances expressed in expected changes per amino acid position, including back mutations. The resulting distance matrices were converted to phylogenetic trees by using the program FITCH (11) which uses the Fitch-Margoliash least-squares distance matrix method and does not assume a constant evolutionary clock. Similar tree topologies were obtained by neighbor-joining and parsimony methods. The program TREEDRAW was used to draw the unrooted phylogenetic trees presented. Nucleotide accession numbers. The nucleotide sequences for nifKE2 and ORF186 from M. barkeri have been submitted to GenBank and have received the accession numbers U32665 and U32666, respectively.

RESULTS Structure of nifHDK2 and associated genes in M. barkeri. We cloned and sequenced a 2.7-kb BamHI fragment (pYTC006, see Materials and Methods) of M. barkeri DNA which was downstream from the previously cloned 4.7-kb BamHI fragment pYTC001 (4), which contained nifH2, ORFs 105 and 128, nifD2, and 135 bases of nifK2 (Fig. 1). We also sequenced the region of pYTC001 upstream of nifH2. As shown in Fig. 1, downstream from nifK2 is an ORF with clear resemblance to nifE genes. The ORF was 1,390 bp long and apparently continued past the BamHI site terminating the fragment. At a site 184 bp upstream of nifH2 we found a potential ORF encoding a predicted polypeptide 186 amino acids long and transcribed in the opposite direction from nifH2. A BLAST search of this sequence showed some regions of similarity with several flavoprotein oxidoreductases, but not flavodoxin. Downstream from this was another potential ORF, the predicted product of which showed no resemblance to known proteins, which apparently continued past the BamHI site terminating the fragment. Nucleotide sequence analysis of nifK- and nifE-homologous genes. The nucleotide sequences of nifK2 and a 1,390-bp stretch of nifE2 were determined. The predicted nifK2 ORF is 1,368 bp long, encoding 456 amino acids with a predicted molecular mass near 49,315 Da. A potential ribosome-binding site is located 7 bp upstream (59-AGAGG-39) of the translation start site (AUG). The G1C content of nifK2 was 46.8%, slightly higher than the values of 43.7% for the nifD2 gene and 44.8% for the nifH2 gene. An alignment (not presented) of the amino acid sequences for the deduced nifK2 gene product in M. barkeri and nifK sequences from other diazotrophs shows that the residues (three cysteines and one serine) that are involved in the coordination of the P cluster are present in the


Bacterial strains and plasmids. M. barkeri 227 (ATCC 43241, DSM 1538, and OCM 35) was obtained from our own culture collection. Escherichia coli DH5a was used for transformation and DNA-sequencing experiments and was obtained from S. Winans, Cornell University. Plasmids pBSK1 and pUC19 were used for subcloning and DNA sequencing. Cloning and nucleotide sequencing. M. barkeri genomic DNA was digested with the restriction enzyme PstI, diluted, and then incubated with T4 DNA ligase to produce a heterogeneous population of circles. Primers (59-GAAGTAACCA GAGAAGC-39 from the 0.6-kb BamHI-PstI fragment of clone pYTC001 [see Fig. 1] and reverse primer 1233 from the pUC19 vector) within the known DNA sequence were added to a concentration of 1 mM, and the solution was subjected to 30 cycles of PCR (denaturing, 2 min at 948C; annealing, 1 min at 548C; and extension, 1 min at 728C). The 710-bp PCR products were digested with PstI and BamHI and cloned into the PstI- and BamHI-digested pBSK1, and the identity of the fragment was verified by sequencing from both ends. The PstI-BamHI fragment was labeled with digoxigenin (DIG) by using the Genius random priming labeling system (Boehringer Mannheim Biochemicals, Indianapolis, Ind.) and was used for Southern blot and colony hybridization experiments. The hybridized DNA in these experiments was detected by the Genius chemiluminescence detection kit (Boehringer Mannheim) as described in the accompanying instructions. An M. barkeri BamHI library of 2,000 clones was constructed as described previously (4). A 2.7-kb positive clone, termed pYTC006, was detected by colony hybridization performed in the presence of 50% formamide at 428C by using the DIG-labeled probe and chemiluminescent detection. DNA sequencing was performed by the dideoxy chain termination method (24) with the Sequenase DNA sequencing kit (United States Biochemical Corporation, Cleveland, Ohio) and by ABI 373A automatic sequencing performed at the Cornell Biotechnology Institute. DNA sequences were determined from both strands by extension from vector-specific (T3 and T7 primers from pBSK1) priming sites and by primer walking. RNA isolation. M. barkeri cells (20 ml) from exponential-phase cultures grown with N2 or NH41 were anaerobically harvested inside an anaerobic chamber and mixed rapidly with 20 g of crushed ice and 15 mM azide. After the ice melted, cells were sedimented at 20,000 rpm for 1 min. The pellet was frozen with liquid N2, and the resulting brittle pellet was ground with a sterile mortar and pestle. The ground cells were resuspended in 6 ml of Tris-EDTA (TE) buffer containing 250 mM sucrose and were treated with a solution containing 600 ml of 10% sodium dodecyl sulfate (SDS), 660 ml of 5 M NaCl, and 800 ml of CTAB (hexadecyltrimethylammonium bromide)–0.7 M NaCl (12). The lysate was extracted twice with hot phenol-chloroform (1:1) and once with chloroformisoamyl alcohol (24:1) and precipitated with ethanol. The precipitated nucleic acids were redissolved in (per milliliter) 20 mM Tris-HCl (pH 8.0)–10 mM MgCl2–2 mM CaCl2–100 mg of RNase-free DNase. After incubation for 60 min at 378C, SDS, EDTA, and NaCl were added to final concentrations of 1%, 50 mM, and 0.2 M, respectively, and the remaining nucleic acids (RNA) were purified by phenol-chloroform extraction and ethanol precipitation. Pellets were dried in a speedVac SC110 vacuum concentration system (Savant Instruments, Inc., Farmingdale, N.Y.) for 10 min, and the pellet was carefully resuspended in 100 ml of diethylpyrocarbonate (Sigma Chemical Co., St. Louis, Mo.)-treated H2O. RNA samples were immediately frozen at 2708C. Northern blot experiments. The RNA (10 to 30 mg) was separated according to size by electrophoresis through a denaturing agarose gel containing 10% formaldehyde and was then transferred to Stratagene nylon membranes by capillary blotting according to the manufacturer’s directions. The membrane was baked at 808C in a vacuum oven for 1 h. The RNA of interest was then located by hybridization with DIG-labeled nifH2 and nifD2 probes overnight at 428C. After hybridization, the membrane was washed twice for 5 min each time in 23 SSC (13 SSC is 0.15 M NaCl plus 0.015 M sodium citrate)–0.1% SDS solution at room temperature and then washed twice more for 15 min each with 0.53 SSC–0.1% SDS at 658C. The hybridized mRNA was then detected by using the Genius chemiluminescent detection kit as described in the Genius (Boehringer Mannheim) protocol booklet.

J. BACTERIOL.

VOL. 178, 1996

NITROGENASE GENES IN M. BARKERI

145

deduced M. barkeri nifK2 product. As we showed previously (4), both the M. barkeri and C. pasteurianum sequences lack a long N-terminal region present in the Azotobacter vinelandii nifK product and other nifK products from cluster I. The sequence identity for the nifK polypeptides of M. barkeri and C. pasteurianum is 49%, and areas of identity and similarity could be detected throughout these proteins. An intergenic region of 83 nucleotides is present between the start of the putative nifE2 gene and the end of the nifK2 gene, and an AGAGG sequence 12 bp upstream of the initial AUG codon of nifE2 provides a potential ribosome-binding site. An alignment to nifD and other nifE gene products (not presented) demonstrates the presence of cysteine and other residues known to bind P clusters and provide FeMo-cofactor binding (5). Similar to nifE from C. pasteurianum, the predicted nifE product from M. barkeri lacked the 50-amino-acid insert which the corresponding nifD products had. The sequence identity for the nifE polypeptides of M. barkeri and C. pasteurianum is 54%, and areas of identity and similarity could also be detected throughout these proteins. Phylogenetic analysis of the nif gene products from M. barkeri. We have already performed phylogenetic analysis of nifH and nifD genes from M. barkeri (4). Figure 2A shows that the nifK genes form a tree with topology very similar to that for nifD, including relatively close clustering of the M. barkeri and C. pasteurianum nifK products. The nifK1 product from M. thermolithotrophicus is more divergent from cluster III than was nifH2, similar to its nifD1 product, as has been noted previously (1, 27) (Fig. 2A). Since nifE is considered to have arisen from a duplication of nifD (5), we performed a phylogenetic analysis of both gene

products together (Fig. 2B). The nifE products from M. barkeri and C. pasteurianum clustered together and showed divergence similar to that of their corresponding nifD products. Overall, the topologies of the nifD and nifE branches resembled each other. A notable exception is vnfE from A. vinelandii, which showed closest resemblance to nifE1 from the same organism rather than being in cluster II. No nifE representative from cluster II is presently known. We also analyzed the glnB-homologous ORF105 and ORF12X (which encode polypeptides of 105 and 122 to 128 amino acids, respectively) genes from methanogens (Fig. 2C). The gene products formed three distinct clusters, and there was hardly any greater similarity between the two ORFs than there was with glnB. The two M. barkeri sequences in each cluster were hardly more similar to each other than they were to the corresponding ORFs from other methanogens. Identification of the ORFnifHDK2 mRNA. The size of the ORFnifHDK2 transcript was determined by Northern blot analysis of M. barkeri total RNA isolated from N2- and NH41grown cells (Fig. 3). A band estimated to be 4.6 kb long was detected when the nifH2 fragment or the nifDK2 fragment was used as a hybridization probe against RNA from N2-grown cells. A smear containing indistinct bands of smaller size was also detected. A 4.6-kb fragment is large enough to harbor ORFnifH2, ORF105, ORF125, nifD2, and nifK2, indicating that they are cotranscribed. No mRNA was detected in the total RNA prepared from the NH41-grown culture, suggesting that NH41 repressed the transcription of ORFnifHDK2. Location of the nifH2 and nifD2 transcription start sites. The transcription start site for nifH2 was mapped at a T residue 34 nucleotides 59 to the nifH2 coding region by a primer


FIG. 1. Restriction map of BamHI fragments from M. barkeri cloned into pYTC001 and pYTC006. The original clone described by Sibold et al. (25) is defined by the internal HindIII fragment of clone pYTC001. The DNA fragments termed nifH2 and nifK2 probes, used as probes in the Southern, colony, and Northern hybridizations, are indicated by arrows. The nucleotide sequence preceding the transcription start site of nifH2 is shown. Sequences are numbered relative to the mRNA initiation site (indicated by an asterisk) determined by primer extension analysis. The sequence TATAAATA, closely matching the archaeal consensus promoter, is underlined. Abbreviations: B, BamHI; H, HindIII; E, EcoRI; P, PstI; A, AccI.

146

CHIEN AND ZINDER

J. BACTERIOL.

extension experiment using mRNA from M. barkeri grown with N2 (Fig. 1). The sequence TATAAATA located 32 nucleotides upstream of the putative transcription start site (Fig. 1) closely matches the archaeal promoter consensus sequence [TTTA(A/ T)ATA] (19, 20). A putative transcription start site that corresponded to a G residue was also mapped by primer extension for nifD2 (data not shown). However, transcription was initiated only 12 bp upstream of the translation start site for nifD2, which is within the upstream ORF125 (there is a 1-bp overlap between ORF125 and nifD2). In the upstream region of nifD2, no sequence similar to the consensus archaebacterial promoter was found. A stretch of 4 T’s, previously identified as the transcription termination site (20), was found 6 nucleotides downstream of the nifK2 gene. No inverted-repeat or stemloop structures were evident downstream of the nifK2 gene. DISCUSSION The nucleotide sequence was determined for a 7.4-kbp region of the genome of M. barkeri 227 which contains ORF186

(this study), nifH2, ORF105, ORF125 (25), nifD2 (4), nifK2, and nifE2 (this study). The nifK2 gene encodes 456 amino acids with a predicted molecular mass near 50 kDa and is the first complete nifK gene from a methanogen to be reported. The molecular masses for the purified M. barkeri component 1 subunits were estimated to be 62 and 57 kDa by polyacrylamide gel electrophoresis (PAGE) (14). While the predicted molecular weight for the a subunit (4) of 60,146 is close to the molecular mass of 62 kDa, the difference between the predicted subunit molecular mass (50 kDa) and that detected by PAGE for the purified b subunit of component 1 (57 kDa) from M. barkeri suggests atypical migration of the nifK2 product during gel electrophoresis. Downstream of the nifK2 gene in M. barkeri was an ORF homologous to nifE from a variety of eubacteria. Similar to the nifHDK2 genes from M. barkeri, nifE2 showed closest similarity to nifE1 from C. pasteurianum. The similarities of topologies of the nifD and nifE branches shown in Fig. 2B are consistent with the proposition that nifE arose from a duplication of nifD and suggest that this duplication occurred before the nif genes diverged into the various clusters (i.e., nifD and nifE are paralogous) and even suggest a root for nifD between clusters II and III. A notable exception to the similarities between nifD and nifE phylogenies is seen with nifE2 (vnfE) from A. vinelandii. It shows closest resemblance to nifE1 in cluster I from A. vinelandii, while nifD2 (vnfD) from the same organism is clearly in cluster II and is highly divergent from nifD1. This


FIG. 2. Unrooted trees for nifK (A), nifDE (B), and glnB, ORF105, and ORF12X (X 5 2 to 8) (C) amino acid sequences analyzed by using PROTDIST and FITCH programs from the PHYLIP phylogenetic package (see Materials and Methods). For consistency, we’ve adopted the convention that all methanogen sequences which cluster with alternative nitrogenases (cluster II) are numbered 1. Abbreviations: Metba1, M. barkeri 1; Metba2, M. barkeri 2; Clopa1, C. pasteurianum 1; Metth1, M. thermolithotrophicus 1; Clopa2, C. pasteurianum 2; Azovi3, A. vinelandii 3 (anfK); Azobr, Azospirillum brasiliense; Braja, B. japonicum; Frasp, Frankia sp.; Anasp1, Anabaena sp. 1; Anasp2, Anabaena sp. 2; Azovi1, A. vinelandii 1 (nifK); Azovi2, A. vinelandii 2 (vnfK); Klepn, Klebsiella pneumoniae; Thife, Thiobacillus ferrooxidans; PARArhi, Parasponia rhizobium; Rhoca, Rhodobacter capsulatus; Metiv, Methanobacterium ivanovii; Rhile, Rhizobium leguminosarum; Synsp., Synechococcus sp.

VOL. 178, 1996

situation is analogous to that for vnfH from A. vinelandii, which is closely related to the corresponding nifH1. It has been proposed that vnfH2 represents a fairly recent duplication of nifH1 recruited by the vanadium system (1, 22), perhaps after its genetic transfer into an ancestor of A. vinelandii. This hypothesis can now be extended to nifE2 in A. vinelandii. We have shown that the deduced M. barkeri nifD2 polypeptide contains a 50-amino-acid insert which was previously found only in nifD1 of C. pasteurianum (4). The absence of this insert in nifE2 suggests that the ancestor of the nifD2 gene does not contain the insert, which was later added to nifD2, or, alternatively, that the nifE2 gene lost the insert after the duplication from nifD2. The nifEN genes are immediately downstream from nifDK in C. pasteurianum, as well as in Bradyrhizobium japonicum (6), but in many other eubacteria nifK and nifE are not adjacent. Recently, it has been found that the nifEN genes are immediately downstream of nifDK in Methanococcus maripaludis (13). Thus, the arrangement of the nifKE2 genes in M. barkeri resembles that in some eubacteria and other archaea, but no general conclusions can be drawn. In eubacteria, the nitrogenase structural genes (nifH, nifD, and nifK) are usually contiguous. However, in all known functional methanogen nifHDK genes, two ORFs (ORF105 and ORF12X) resembling glnB are located between nifH and nifD (25, 27). The phylogenetic analysis of these ORFs shows that glnB, ORF105, and ORF12X genes all form separate clusters, suggesting an early divergence for these genes. Moreover, the considerable distances between ORF105 2 and ORF105 1 and ORF12X 2 and ORF12X 1 in M. barkeri are inconsistent with a relatively recent duplication of these genes, as is apparently the case for vnfH and vnfE from A. vinelandii, and argues against a relatively recent transfer of nifHDK genes from an ancestor of C. pasteurianum followed by recruitment of duplicates of the two ORFs. We have also found an ORF 184 bp upstream from nifH2 which is transcribed in the opposite direction from nifH2 (Fig. 1) and which has an acceptable ribosome-binding site (AGGA) and promoter (TTAATAAA) upstream from it. A BLAST search for this ORF indicated a significant similarity with a variety of electron-transferring proteins, including bacterio-

147

chlorophyll synthase, fixC proteins, NADH oxidases, lipoamide reductases, and other flavin adenine dinucleotide-containing flavoproteins, with the bacteriochlorophyll synthase being the closest match (probability, 6.6 3 10220). Whether the product of this ORF is involved in electron transfer related to nitrogen fixation or serves another function is not known. In terms of expression of the nitrogenase structural genes, the 4.6-kbp mRNA detected in the Northern blot experiment indicated that ORFnifH2, ORF105, ORF125, ORFnifD2, and ORFnifK2 are cotranscribed. A putative archaeal promoter sequence for nifH2 (TATAAATA) was detected from 32 to 39 nucleotides upstream of the transcription start site. The putative transcription start site detected for nifD2 by primer extension is more problematical, since it occurs only 12 bp from the initiation codon at the first G of the putative ribosome-binding site (which is encoded within ORF125) and we could detect no acceptable upstream archaeal promoters. Furthermore, no archaeal transcription terminators were found downstream of ORF125. The 4.6-kb transcript detected in the Northern blot experiment indicated that the identity of the sequence as the transcription start site for nifD2 was very unlikely and may be an artifact or the result of RNA processing, which has been reported for archaea (9, 26, 29). Since a transcription start with no detectable promoter was detected upstream of nifD1 in M. thermolithotrophicus (27), this type of phenomenon may be widespread in methanogen nif genes. Transcription termination sites in methanogens are usually identified as one of two motifs (20). In some examples, transcription is terminated following an inverted-repeat sequence that, common at rho-independent terminators in eubacteria, probably forms a stem-loop structure in the transcript to direct termination (9, 10). The second transcription terminator appears to be an oligo-T sequence, and in some cases, several tandemly arranged oligo-T sequences reside immediately downstream of methanogen genes (2, 8, 31). In M. barkeri 227, the transcription terminator of the nifHDK2 operon apparently resembles the second motif, with a stretch of 4 T’s being found immediately downstream of the nifK2 gene and no potential stem-loop structures being detected. The presence of a transcription start with an acceptable promoter sequence upstream from nifH2, a transcript size of 4.6 kb, and a potential terminator 6 bp downstream of the nifK2 gene and an acceptable promoter 68 bp upstream of nifE2 all support the argument for a cotranscribed operon consisting of nifH2, ORF105, ORF125, nifD2, and nifK2. The nifE2 gene is apparently the start of another transcription unit and is transcribed in the same direction, while ORF186 upstream of the nifH2 gene is transcribed in the opposite direction from the nifH2 gene. Transcript analysis has been performed on only a few methanogen genes thus far, with only a very few in Methanosarcina spp. Recently, the expression of the gene encoding carbon monoxide dehydrogenase, an enzyme central in acetate metabolism in Methanosarcina thermophila, a fairly close relative of M. barkeri, was shown to be regulated at the level of transcription, being repressed by the presence of the preferred methanogenic substrate, methanol (28), as were the acetate-activating enzymes acetate kinase and phosphotransacetylase (26). Our previous results showed that ammonium-grown cells of M. barkeri do not reduce acetylene or show bands cross-reacting with antibody to eubacterial component 2 in Western immunoblots (14), indicative of repression. Dot blot (4) and Northern blot (this study) hybridizations suggested that the expression of the nitrogenase structural genes was repressed by ammonia and the repression was at the level of transcription. Thus, nitrogen fixation, an energetically costly process, is


FIG. 3. Northern blot hybridization of total RNA isolated from NH41- or N2-grown M. barkeri 227 cells by using DIG-labeled nifH2 and nifDK2 probes (Fig. 1). The estimated size of the hybridizing mRNA species is indicated on the left.

NITROGENASE GENES IN M. BARKERI

148

CHIEN AND ZINDER

highly regulated in M. barkeri 227. In studies by Gohl et al. (7), RNA polymerase preparations from N2- and NH41-grown cells were equally effective at transcribing nifH1 from M. thermolithotrophicus, so that the nature of the regulation of nif gene expression in methanogens remains obscure. We are presently further investigating this phenomenon in M. barkeri. ACKNOWLEDGMENTS This research was supported by grant DE-FG02-85ER13370 from the U.S. Department of Energy and by USDA Hatch funds. We thank J. P. Shapleigh and J. D. Helmann for their helpful advice and for sharing their facilities with us. REFERENCES

Abstracts of the 95th General Meeting of the American Society for Microbiology 1995. American Society for Microbiology, Washington, D.C. 14. Lobo, A. L., and S. H. Zinder. 1990. Nitrogenase in the methanogenic archaebacterium Methanosarcina barkeri 227. J. Bacteriol. 172:6789–6796. 15. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. 16. Merrick, M. J. 1992. Regulation of nitrogen fixation genes in bacteria, p. 835–876. In G. Stacey, R. H. Burris, and H. J. Evans (ed.), Biological nitrogen fixation. Chapman and Hall, New York. 17. Normand, P., and J. Bousquet. 1989. Phylogeny of nitrogenase sequences in Frankia and other nitrogen-fixing microorganisms. J. Mol. Evol. 29:436–447. 18. Normand, P., M. Gouy, B. Cournoyer, and P. Simonet. 1992. Nucleotide sequence of nifD from Frankia alni strain ARI3: phylogenetic inferences. Mol. Biol. Evol. 9:495–506. 19. Palmer, J. R., and C. J. Daniels. 1995. In vivo definition of an archaeal promoter. J. Bacteriol. 177:1844–1849. 20. Reeve, J. N. 1992. Molecular biology of methanogens. Annu. Rev. Microbiol. 46:165–191. 21. Reitzer, L. J., and B. Magasanik. 1987. Ammonia assimilation and the biosynthesis of glutamine, glutamate, aspartate, asparagine, L-alanine, and D-alanine, p. 302–320. In F. C. Neidhardt, J. L. Ingraham, K. B. Low, B. Magasanik, M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli and Salmonella typhimurium: cellular and molecular biology. American Society for Microbiology, Washington, D.C. 22. Robson, R. L., P. R. Woodley, R. N. Pau, and R. R. Eady. 1989. Structural genes for the vanadium nitrogenase from Azotobacter croococcum. EMBO J. 8:1217–1224. 23. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. 24. Sanger, F. 1981. Determination of nucleotide sequences in DNA. Science 214:1205–1210. 25. Sibold, L., M. Henriquet, O. Possot, and J.-P. Aubert. 1991. Nucleotide sequence of nifH regions from Methanobacterium ivanovii and Methanosarcina barkeri and characterization of glnB-like genes. Res. Microbiol. 142:5– 12. 26. Singh-Wissmann, K., and J. G. Ferry. 1995. Transcriptional regulation of the phosphotransacetylase-encoding and acetate kinase-encoding genes (pta and ack) from Methanosarcina thermophila. J. Bacteriol. 177:1699–1702. 27. Souillard, N., and L. Sibold. 1989. Primary structure, functional organization and expression of nitrogenase structural genes of the thermophilic archaebacterium Methanococcus thermolithotrophicus. Mol. Microbiol. 3:541–551. 28. Sowers, K. R., T. T. Thai, and R. P. Gunsalus. 1993. Transcriptional regulation of the carbon monoxide dehydrogenase gene (cdhA) in Methanosarcina thermophila. J. Biol. Chem. 268:23172–23178. 29. Stolt, P., and W. Zillig. 1993. Structure specific ds-ss-RNase activity in the extreme halophile Halobacterium salinarium. Nucleic Acids Res. 21:5595– 5599. 29a.Ueda, T., Y. Suga, N. Yahiro, and T. Matsuguchi. 1995. Remarkable N2fixing bacterial diversity detected in rice roots by molecular evolutionary analysis of nifH gene sequence. J. Bacteriol. 177:1414–1417. 30. Young, J. P. W. 1992. Phylogenetic classification of nitrogen-fixing organisms, p. 43–86. In G. Stacey, R. H. Burris, and H. J. Evans (ed.), Biological nitrogen fixation. Chapman and Hall, New York. 31. Zillig, W., P. Palm, W.-D. Peiter, F. Gropp, G. Puhler, and H.-P. Klenk. 1988. Comparative evaluation of gene expression in archaebacteria. Eur. J. Biochem. 173:473–482.


1. Bishop, P. E., and R. Premakumar. 1992. Alternative nitrogen fixation systems, p. 736–762. In G. Stacey, R. H. Burris, and H. J. Evans (ed.), Biological nitrogen fixation. Chapman and Hall, New York. 2. Brockl, G., M. Behr, S. Fabry, R. Hensel, and H. Kaudewitz. 1991. Analysis and nucleotide sequence of the genes encoding the surface-layer glycoproteins of the hyperthermophilic methanogens Methanothermus ferridus and Methanothermus sociabilis. Eur. J. Biochem. 199:147–152. 3. Burke, D. H., J. E. Hearst, and A. Sidow. 1993. Early evolution of photosynthesis—clues from nitrogenase and chlorophyll iron proteins. Proc. Natl. Acad. Sci. USA 90:7134–7138. 4. Chien, Y.-T., and S. H. Zinder. 1994. Cloning, DNA sequencing, and characterization of a nifD-homologous gene from the archaeon Methanosarcina barkeri 227 which resembles nifD from the eubacterium Clostridium pasteurianum. J. Bacteriol. 176:6590–6598. 5. Dean, D. R., J. T. Bolin, and L. M. Zheng. 1993. Nitrogenase metalloclusters: structures, organization, and synthesis. J. Bacteriol. 175:6737–6744. 6. Dean, D. R., and M. R. Jacobson. 1992. Biochemical genetics of nitrogenases, p. 763–834. In G. Stacey, R. H. Burris, and H. J. Evans (ed.), Biological nitrogen fixation. Chapman and Hall, New York. 7. Gohl, H. P., W. Hausner, and M. Thomm. 1992. Cell-free transcription of the nifH1 gene of Methanococcus thermolithotrophicus indicates that promoters of archaeal nif genes share basic features with methanogen consensus promoter. Mol. Gen. Genet. 231:286–295. 8. Haas, E. S., C. J. Daniels, and J. N. Reeve. 1989. Genes encoding 5S rRNA and tRNA in the extremely thermophilic archaebacterium Methanothermus fervidus. Gene 77:253–263. 9. Hennigan, A. N., and J. N. Reeve. 1994. mRNA in the methanogenic archaeon Methanococcus vannielii: numbers, half-lives and processing. Mol. Microbiol. 11:655–670. 10. Higgins, D. G., and P. M. Sharp. 1988. Clustal: a package for performing multiple alignment on a microcomputer. Gene 73:237–244. 11. Hirsch, A. M., H. I. McKhann, A. Reddy, J. Liao, Y. Fang, and C. R. Marshall. 1995. Assessing horizontal transfer of nifHDK genes in eubacteria: nucleotide sequence of nifK from Frankia strain HFPCcI3. Mol. Biol. Evol. 12:16–27. 12. Jarrell, K. F., D. Gaguy, A. M. Herbert, and M. L. Kalmokoff. 1992. A general method of isolating high molecular weight DNA from methanogenic archaea. Can. J. Microbiol. 38:65–68. 13. Kessler, P. S., C. Blank, and J. A. Leigh. 1995. Transposon insertion mutagenesis of nif genes in Methanococcus maripaludis, abstr. I-22, p. 321. In

J. BACTERIOL.