JOURNAL OF BACTERIOLOGY, Jan. 1991, p. 642-648 Vol. 173, No. 2 0021-9193/91/020642-07$02.00/0 Copyright X) 1991, American Society for Microbiology Mutations in DNA Gyrase Result in Novobiocin Resistance in Halophilic Archaebacteria MELISSA L. HOLMES* AND MICHAEL L. DYALL-SMITH Department of Microbiology, University of Melbourne, Parkville, Victoria 3052, Australia Received 10 September 1990/Accepted 9 November 1990 We have developed a cloning vector for use in halophilic archaebacteria which has a novobiocin resistance determinant as a selectable marker. The resistance determinant, which was derived from the genome of a resistant mutant strain, was mapped to a site within a 6.7-kb DNA clone by using a recombination assay and was sequenced. An open reading frame of 1,920 nucleotides (640 amino acids) was identified, with the predicted protein being highly homologous to the DNA gyrase B subunit (i.e., GyrB) of eubacteria. Three mutations were identified in the GyrB protein of the resistant mutant compared with the wild type (at amino acids 82, 122, and 137) which together enable Haloferax cells to grow in concentrations of novobiocin some 1,000 times higher than that possible for cells carrying only the wild-type enzyme. One base beyond the stop codon of gyrB was the start of gyrA, coding for the gyrase A subunit. The archaebacteria are a novel and relatively unexplored group of organisms, and their study has radically altered the picture of cellular evolution on this planet. Since their discovery in the 1970s, many surprising features have been reported, such as their unique membrane lipids and the presence of genes containing introns, and they remain of great biological interest (for a review, see reference 31). Within the archaebacterial kingdom, the halobacteria, and particularly members of the genus Haloferax, are the most convenient group for genetic study (6, 7), but until recently the analysis of gene structure and function had been frus- trated by the unavailability of genetic systems, with much of the work being limited to sequence comparisons with eubac- terial and eucaryotic homologs. We recently described the construction of a plasmid cloning vector for halobacteria which consisted of a halo- bacterial plasmid with a dominant selectable marker confer- ring resistance to the antibiotic novobiocin (15). The resis- tance determinant had been cloned from the genomic DNA of a novobiocin-resistant (Novr) mutant isolated in our laboratory. Using the polyethylene glycol transformation method described by Cline and Doolittle (8), the vector could be introduced into Haloferax cells at high efficiency (i.e., >106 transformants per ,ug of plasmid DNA), allowing the analysis of halobacterial genes to be carried out in vivo. For the effective use of this vector in such studies, it is important to have a thorough knowledge of the structure and function of relevant genes, particularly that of the resistance marker. Novobiocin is a naturally occurring antibiotic known to inhibit the activity of eubacterial DNA gyrase, specifically the B subunit (21, 27), and is also strongly inhibitory to many archaebacteria, including halobacteria. In the latter case, the concentration range at which novobiocin is inhibitory to- gether with evidence from several biochemical studies has led to the suggestion that the target of action may be the same as in eubacteria (26), but this has not been conclusively proven. In this study, we were able to test this hypothesis by locating and sequencing the gene responsible for conferring novobiocin resistance in halobacteria. * Corresponding author. MATERIALS AND METHODS Bacterial strains and plasmids. Haloferax phenon K isolate Aa 2.2 (from M. Torreblanca, University of Alicante, Ali- cante, Spain) and a Novr mutant of this strain (15) were used in all experiments. Wild-type (wt) Haloferax cultures could be transformed to novobiocin resistance by using plasmid pMDS2 (Fig. 1; 15) and the polyethylene glycol-mediated transformation system (15). Transformants were isolated on transformation medium containing 0.1 ,ug of novobiocin per ml. Plasmid pMDS2 was also the source of the Novr gene which had been included as a selectable marker in the construct. Single-stranded DNA templates used in sequencing exper- iments were isolated from Escherichia coli JM101 cells transformed with M13 phage vectors (33). E. coli XL1-Blue was transformed with the plasmid vector pBS(+) (Strata- gene, La Jolla, Calif.) and screened for ampicillin resistance. DNA and RNA isolation. E. coli plasmid DNA and single- stranded templates were isolated by standard procedures (17, 33); the extraction of plasmid DNA from Haloferax cultures has been described previously (15). Total DNA was isolated from wt and Novr mutant cultures by the following procedure. Cells were harvested from late-log-phase cul- tures by centrifugation at 2,800 x g for 15 min, resuspended in 1 M NaCl, and incubated at 37°C for 1 h in the presence of proteinase K (120 ,ug/ml). The cells were then lysed by the addition of a Triton-X solution (0.2% Triton X-100, 6 mM EDTA, 5 mM Tris hydrochloride, pH 8), incubated at 37°C for 1 h, and then treated with 8% hexadecyltrimethyl ammo- nium bromide (CTAB)-0.56 M NaCl at 65°C for 10 min. An equal volume of chloroform-isoamyl alcohol (24:1, vol/vol) was added; the solution was mixed well and then centrifuged at 2,800 x g for 10 min to separate the CTAB-protein/ polysaccharide complexes from the cleared lysate. Total t)NA was then precipitated at room temperature with 0.6 volume of isopropanol, and the pellet was spun down (2,800 x g for 10 min) and washed with 70% ethanol before being redissolved in a small volume TE buffer (10 mM Tris hydrochloride, 1 mM EDTA, pH 8). RNA was isolated from mid-log-phase Haloferax cultures by sodium dodecyl sulfate lysis in the presence of sodium azide (10mM), followed by phenol and then chloroform 642
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
JOURNAL OF BACTERIOLOGY, Jan. 1991, p. 642-648 Vol. 173, No. 20021-9193/91/020642-07$02.00/0Copyright X) 1991, American Society for Microbiology
Mutations in DNA Gyrase Result in Novobiocin Resistance inHalophilic Archaebacteria
MELISSA L. HOLMES* AND MICHAEL L. DYALL-SMITH
Department of Microbiology, University of Melbourne, Parkville, Victoria 3052, Australia
Received 10 September 1990/Accepted 9 November 1990
We have developed a cloning vector for use in halophilic archaebacteria which has a novobiocin resistancedeterminant as a selectable marker. The resistance determinant, which was derived from the genome of aresistant mutant strain, was mapped to a site within a 6.7-kb DNA clone by using a recombination assay andwas sequenced. An open reading frame of 1,920 nucleotides (640 amino acids) was identified, with the predictedprotein being highly homologous to the DNA gyrase B subunit (i.e., GyrB) of eubacteria. Three mutations wereidentified in the GyrB protein of the resistant mutant compared with the wild type (at amino acids 82, 122, and137) which together enable Haloferax cells to grow in concentrations of novobiocin some 1,000 times higherthan that possible for cells carrying only the wild-type enzyme. One base beyond the stop codon of gyrB wasthe start of gyrA, coding for the gyrase A subunit.
The archaebacteria are a novel and relatively unexploredgroup of organisms, and their study has radically altered thepicture of cellular evolution on this planet. Since theirdiscovery in the 1970s, many surprising features have beenreported, such as their unique membrane lipids and thepresence of genes containing introns, and they remain ofgreat biological interest (for a review, see reference 31).Within the archaebacterial kingdom, the halobacteria, andparticularly members of the genus Haloferax, are the mostconvenient group for genetic study (6, 7), but until recentlythe analysis of gene structure and function had been frus-trated by the unavailability of genetic systems, with much ofthe work being limited to sequence comparisons with eubac-terial and eucaryotic homologs.We recently described the construction of a plasmid
cloning vector for halobacteria which consisted of a halo-bacterial plasmid with a dominant selectable marker confer-ring resistance to the antibiotic novobiocin (15). The resis-tance determinant had been cloned from the genomic DNAof a novobiocin-resistant (Novr) mutant isolated in ourlaboratory. Using the polyethylene glycol transformationmethod described by Cline and Doolittle (8), the vectorcould be introduced into Haloferax cells at high efficiency(i.e., >106 transformants per ,ug of plasmid DNA), allowingthe analysis of halobacterial genes to be carried out in vivo.For the effective use of this vector in such studies, it isimportant to have a thorough knowledge of the structure andfunction of relevant genes, particularly that of the resistancemarker.Novobiocin is a naturally occurring antibiotic known to
inhibit the activity of eubacterial DNA gyrase, specificallythe B subunit (21, 27), and is also strongly inhibitory to manyarchaebacteria, including halobacteria. In the latter case, theconcentration range at which novobiocin is inhibitory to-gether with evidence from several biochemical studies hasled to the suggestion that the target of action may be thesame as in eubacteria (26), but this has not been conclusivelyproven. In this study, we were able to test this hypothesis bylocating and sequencing the gene responsible for conferringnovobiocin resistance in halobacteria.
* Corresponding author.
MATERIALS AND METHODS
Bacterial strains and plasmids. Haloferax phenon K isolateAa 2.2 (from M. Torreblanca, University of Alicante, Ali-cante, Spain) and a Novr mutant of this strain (15) were usedin all experiments. Wild-type (wt) Haloferax cultures couldbe transformed to novobiocin resistance by using plasmidpMDS2 (Fig. 1; 15) and the polyethylene glycol-mediatedtransformation system (15). Transformants were isolated ontransformation medium containing 0.1 ,ug of novobiocin perml. Plasmid pMDS2 was also the source of the Novr genewhich had been included as a selectable marker in theconstruct.
Single-stranded DNA templates used in sequencing exper-iments were isolated from Escherichia coli JM101 cellstransformed with M13 phage vectors (33). E. coli XL1-Bluewas transformed with the plasmid vector pBS(+) (Strata-gene, La Jolla, Calif.) and screened for ampicillin resistance.DNA and RNA isolation. E. coli plasmid DNA and single-
stranded templates were isolated by standard procedures(17, 33); the extraction of plasmid DNA from Haloferaxcultures has been described previously (15). Total DNA wasisolated from wt and Novr mutant cultures by the followingprocedure. Cells were harvested from late-log-phase cul-tures by centrifugation at 2,800 x g for 15 min, resuspendedin 1 M NaCl, and incubated at 37°C for 1 h in the presence ofproteinase K (120 ,ug/ml). The cells were then lysed by theaddition of a Triton-X solution (0.2% Triton X-100, 6 mMEDTA, 5 mM Tris hydrochloride, pH 8), incubated at 37°Cfor 1 h, and then treated with 8% hexadecyltrimethyl ammo-nium bromide (CTAB)-0.56 M NaCl at 65°C for 10 min. Anequal volume of chloroform-isoamyl alcohol (24:1, vol/vol)was added; the solution was mixed well and then centrifugedat 2,800 x g for 10 min to separate the CTAB-protein/polysaccharide complexes from the cleared lysate. Totalt)NA was then precipitated at room temperature with 0.6volume of isopropanol, and the pellet was spun down (2,800x g for 10 min) and washed with 70% ethanol before beingredissolved in a small volume TE buffer (10 mM Trishydrochloride, 1 mM EDTA, pH 8).RNA was isolated from mid-log-phase Haloferax cultures
by sodium dodecyl sulfate lysis in the presence of sodiumazide (10mM), followed by phenol and then chloroform
642
NOVOBIOCIN RESISTANCE MARKER IN ARCHAEBACTERIA
A1.4kb .
Sail SalSail Sall
Sail SphlSall
pni SphlKpnlpMDS2 Sall
Bgl 11 SadSacl x Nov KpnlEcoRSphlSa NsiI ph
FIG. 1. Restriction map of plasmid pMDS2. The construct in-cludes a 6.7-kb KpnI fragment (_) containing the Novr determi-nant (multiple Sall sites are not shown) cloned into the Kpnl site ofpHK2, a 10.5-kb multicopy plasmid from Haloferax strain Aa 2.2.
extractions (25). It was then precipitated with ethanol andresuspended in a small volume of TE buffer.DNA sequencing. A region extending some 2,550 nucleo-
tides (nt) (centered around the 1.4-kb Sall fragment of the6.7-kb clone) was sequenced on both strands by using thedideoxynucleotide chain termination method (24). T7 DNApolymerase (sequencing kit from Pharmacia), and [35S]dATP(Bresatec Ltd., Adelaide, South Australia). Restriction frag-ments were subcloned into M13 vectors by utilizing the SalIl,Sacl, and EcoRI sites, and sequencing was completed with12 synthetic primers (whose numbers are given in parenthe-ses) complementary to nt 695 to 711 (1), 773 to 755 (2), 976 to993 (3), 1216 to 1199 (4), 1379 to 1361 (5), 1669 to 1652 (6),1653 to 1670 (7), 538 to 521 (8), 2372 to 2355 (9), 256 to 239(10), 2170 to 2188 (11), and 239 to 257 (12) (see Fig. 3 and 4).
Isolation of the wt gyrB gene. Total DNA from wt Halo-ferax strain Aa 2.2 was restricted with NsiI and BglII, andfragments of 2 to 2.2 kb were isolated by electrophoresisonto dialysis tubing (17). These were cloned into M13vectors in both orientations, and 700 colorless plaques werethen picked and transferred to microtiter tray wells contain-ing 150 ,ul of 2YT (17) plus a 1/50 dilution of an overnightJM101 culture. Three Novr gyrB clones were included aspositive controls. After overnight incubation at 37°C, 20-,ulsamples were spotted onto Zeta-probe nylon membranes(Bio-Rad) and fixed with 0.4 NaOH for 2 h. Hybridizationswere performed as instructed by the membrane manufac-turer, using a 32P-labeled Novr gyrB NsiI-BglII fragmentprobe (nick translation kit; Bresatec). Approximately 25% ofthe isolates positively hybridized with this probe, and Ttracking revealed that these were gyrB clones.Primer extension. An 18-mer oligonucleotide primer com-
plementary to nt 538 to 521 (primer 8; see Fig. 3 and 4) wasextended with the Klenow fragment of DNA polymerase I(Bresatec) incorporating [35S]dATP. A defined 3' end wasobtained following restriction with XmnI, which cleavesafter position 442, resulting in a primer some 96 nt long. Thesingle-stranded labeled primer was then purified on a 6%(wt/vol) acrylamide sequencing gel (17). The primer exten-sion reactions, using 30 pg of total RNA and 5 x 104 cpm ofpurified primer, were performed as previously described(17), and the products were run on 8% (wt/vol) sequencing
O.99kb0.7kb -
B 0.44kb.0.4kb -
Transformsto novR + + +
FIG. 2. Summary of results of the recombination assay used tolocate the Novr determinant within the original clone. The 6.7-kbKpnI fragment of pMDS2 was partially restricted with Sall, andfragments were recloned into E. coli[pBS(+)]. These fragmentswere used to transform wt Haloferax cells to novobiocin resistance.(A) The Sall fragments present in each of seven subclones (minusthe 3.2-kb vector band) as they would appear after agarose gelelectrophoresis. (B) Indication of whether or not the subclone abovewas able to transform Haloferax cells to full resistance. The 1.4-kbSall fragment is common to all of the subclones capable of trans-forming Haloferax cells to Novr.
gels with sequence ladders from the same oligonucleotideprimer as standards (see Fig. 6).Northern (RNA) hybridization. RNA samples (20 ±g) were
electrophoresed on a formaldehyde-agarose gel alongsidedenatured 32P-end-labeled Hindlll-digested lambda sizestandards. They were then transferred by capillary blotting,over 5 h, in the presence of 50 mM NaOH onto Zeta-probenylon filters (Bio-Rad). An internal 1.4-kb Sall fragment ofgyrB was 32p labeled (nick translation kit; Bresatec), and theprobe was then included in hybridizations under the stan-dard conditions recommended by the membrane manufac-turer.
Nucleotide sequence accession number. The Genbank/EMBL nucleotide accession number for the sequence datapresented in this report is HLFGYRB.
RESULTS
Location of the Novr determinant. We have reported pre-viously the shotgun cloning of a Novr determinant from amutant Haloferax strain, producing the halobacterial plas-mid cloning vector pMDS2 (Fig. 1; 15). The exact location ofthe resistance gene within the 6.7-kb KpnI clone was notknown and was determined as follows. The isolated KpnIfragment was partially digested with Sall, which cuts atmultiple sites, and the pieces were recloned into the E. colivector pBS(+). We used the highly efficient recombinationsystem observed in halobacteria (9, 15, 20) and tested theability of specific subclones to transform wt Haloferax cells(by gene conversion) to novobiocin resistance (Fig. 2). Someof the plasmids contained more than one Sall fragment, butit was clear that a single fragment of 1.4 kb was sufficient fortransformation to full resistance, and we focused on thisregion of the original clone for detailed restriction mappingprior to full sequencing. Sall sites close to the 1.4-kb piecewere identified by progressive BAL 31 nuclease digestion ofthe KpnI fragment and then Sall digestion. In this way, the0.7- and 0.4-kb fragments shown in Fig. 2 were found to beadjacent to but at opposite ends of the 1.4-kb fragment (Fig.3).Novr gene sequence. By using the strategy outlined in Fig.
3, the three contiguous Sall fragments (1.4, 0.7, and 0.4 kb)described above were sequenced, and a total of 2,550 bp is
VOL. 173, 1991 643
644 HOLMES AND DYALL-SMITH
co 0~
I I
o) 0 0) 0
oo
cCe X) en
I I I
i4f --
0.4kb 1.4kb
I CDIc
z enI L (Kpnl)
1 1 9
1 0.7kb I
FIG. 3. Restriction map of the 2.5-kb region sequenced within the 6.7-kb KpnI fragment of pMDS2. Positions of the 1.4-, 0.7-, and 0.4-kbSall fragments shown in Fig. 2 and orientations of the synthetic primers (1 to 12) used in sequencing experiments are indicated.
presented in Fig. 4. Computer analysis of both strandsrevealed the presence of four long open reading frames(ORFs), which are discussed below.
(i) ORF1. ORFi extends 1,920 nt, from nt 384 to 2303, andcan code for a protein of 640 amino acids (aa). Comparisonwith published sequences of bacterial proteins revealed a
striking homology to GyrB, the gyrase B subunit protein ofeubacteria (1, 22). Overall, there is about 60% identitybetween GyrB of Bacillus subtilis and the predicted ORFiprotein (Fig. 5), but in some regions up to 91% identity isobserved (e.g., aa 11 to 55, 314 to 349, and 408 to 460). Onthe basis of its clear sequence homology to other publishedGyrB proteins, we have designated the Haloferax ORFigene gyrB. While these relationships were obvious at theprotein level, at the nucleotide level there is only weakhomology between the Haloferax and Bacillus genes, a
reflection of their very different base compositions (13, 14).(ii) ORF2. ORF2 begins just one base downstream from
the end of the gyrB coding sequence and extends at least 400nt (Fig. 4). Although the full sequence has not yet beendetermined, the predicted protein is strongly homologous tothe amino-terminal end of eubacterial GyrA, the A subunit ofDNA gyrase. No transcriptional terminator sequence (5)could be seen downstream from the end of gyrB, and theclose proximity of the gyrA gene suggests that these genes
may be cotranscribed, a pattern not found in the gyrase
genes of E. coli and B. subtilis (2, 16).(iii) ORF3. On the same strand as the gyrB gene, but in a
different frame, is a long ORF, ORF3. It begins with a GUGstart codon and extends from nt 806 to 2117, thus being ableto code for a protein of 437 aa. The predicted protein isunusually abundant in arginine, histidine, and proline. Thehigh content of positively charged residues is suggestive ofDNA-binding proteins, but extensive computer searchesfailed to find any obvious homology to sequenced proteins.At present, the status of this ORF remains unclear.
(iv) ORF4. A surprising finding was a very long ORF on
the opposite strand which encompasses all of the gyrBcoding sequence and part of the gyrA sequence also. Itbegins at nt 2402 and ends at nt 168 and can code for a
protein of 744 aa. The predicted protein has an abundance ofacidic amino acids but no obviously unusual features, itscodon frequency being similar to those of other halobacterialproteins (25). Computer searches of protein sequence databases failed to find any known homolog.Comparison of wt and Novr gyrB genes. To identify the
mutations within gyrB that were responsible for increasedresistance to novobiocin, the Haloferax wt gene was clonedand the sequence of the entire coding region was determined(Fig. 5). Comparison with gyrB from the resistant mutantrevealed three base changes, all producing amino acid sub-
stitutions at the highly conserved amino terminus of theprotein (Fig. 5). The amino acids at positions 123 and 138 ofGyrB are identical (i.e., are rigidly conserved) in E. coli, B.subtilis, and Haloferax cells. In the mutant GyrB, these havebeen substituted for amino acids with similar properties (Serand Arg in the wt substituted with Thr and His in the Novrstrain). However, the substitution of Asp with Gly at aa 83results in the replacement of a negatively charged side chainwith an amino acid with no side chain. The significance ofthis change is difficult to interpret, since glycine is normallyfound in this position in the wt enzymes of E. coli and B.subtilis.
Transcription of the gyrB gene. (i) Mapping of the start oftranscription. The start of transcription of the HaloferaxgyrB gene was determined by primer extension on cellularmRNA (Fig. 6) and found to be located at a G residue 115bases upstream from the predicted ATG start codon of gyrB(Fig. 4). Transcription of halobacterial genes usually beginsat a G within the sequence TGCPuA (box B; 18, 29), and our
result is consistent with this. The comparatively long non-
coding region contains a number of interesting features,including a small ORF (nt 294 to 327), repeated sequences
(TCGG repeated four times between nt 274 and 300), a
region of alternating purine-pyrimidine sequence (nt 330 to347) suggestive of formation of left-handed Z-DNA (23), andthree inverted repeat sequences (centered on nt 291, 337 to338, and 369 to 370) that could form hairpin loops in themRNA (Fig. 4). The function(s) of these sequences is not yetknown.Upstream from the transcription start point we could not
find a sequence motif that closely resembled the consensus
promoter (box A) of stable RNAs (29), but bases - 17 to -24matched closely the predicted promoter of the bop gene fromHalobacterium halobium (Fig. 7; 10). The low homology ofbop and gyrB promoters to the halobacterial consensus
sequence may be related to their tight regulation.(ii) Determination of the size of the gyrB mRNA. To test
whether gyrA and gyrB genes are cotranscribed, we exam-
ined the size of the gyrB transcript by Northern blot hybrid-ization. Total cell RNA was extracted, separated on dena-turing agarose gels, transferred to nylon membranes, andhybridized to a radiolabeled gyrB-specific DNA probe (Fig.8). A single RNA band, running just below 23S rRNA, was
detected in all cells and was estimated to be 2,800 nt long.This value is probably an underestimate since the bandappeared to be displaced downward because of its proximityto the large 23S rRNA band, but even this value means thatthe mRNA is considerably larger than gyrB, which is about2,033 nt from the start of transcription to the translationalstop codon. However, it is still much shorter than expectedif gyrA and -B genes are cotranscribed, assuming that the
(Kp nil 1 I I- I IL I I I
J. BACTERIOL.
III.
8
VOL.173,1991~~NOVOBIOCINRESISTANCE MARKER IN ARCHAEBACTERIA 64
R P L E EA G YV YA AOQPP L YRV R YR G NTYIDA M DE A E RD REI EEE E C NG N PTOQVOQR1960 1970 1980 1990 2000 2010 2020 2030 2040 2050 2060 2070 2080 2090 2100
F KG L GE MN POO LWID T T MNP EN R VL KR I T VE I A AA AODR MF NEI L MG D AVG P
2110 2120 2130 2140 2150 7 21W 2170 2180 2 '* -'
2 2210 7'2220 2230 2240 2250
R KO0F I KOD NA NODAE WVOD I *M S SODA P I SF E PG AG I A AE VKN A RI E ED E M EO0S Y
FIG. 4. Nucleic acid sequence of a 2,550-bp region overlapping the identified Nov' determinant (0.4-, 1.4-, and 0.7-kb Sall fragments).Four ORFs were identified. The start of transcription of ORFi was mapped to nucleotide 269 (*), and the putative translation start codon is
shown in bold (positions 384 to 386). Nucleotides shown in lowercase indicate positions where point mutations have occurred (position 628,
A--+G, position 747, T-+*A; position 793, G--*A; wt --* Nov' gyrB). The proteins encoded by ORFi (gyrB) and ORF2 (gyrA) are indicated in
one-letter code above the nucleotide sequence. Some direct and indirect repeats are underlined with arrows.
gyrA gene is similar in size to that of B. subtilis (22), i.e.,
about 1,800 nt. To resolve this matter, we probed the same
blot with a labeled gyrA-specific sequence, but the result
was inconclusive. Completion of the gyrA sequence and
further transcriptional mapping studies should clarify this
point.
(ii) mRNA levels in mutant and plasmid-transformed cells.
Since equal quantities of RNA were loaded onto the gel
tracks, the northern blot hybridization results can be used to
estimate the levels of gyrB mRNA in the various cell types.
The Nov' mutant and wt Haloferax cells showed no apparent
difference in mRNA level, indicating that resistance was not
due to overexpression. In contrast, plasmid-transformedcells appeared to have a much higher level of gyrB mRNA
than did wt cells, consistent with the cloned gyril gene being
carried on a multicopy plasmid estimated to be present at
about seven copies per cell (15).
DISCUSSION
We have identified the gene and, by analogy to the
eubacterial studies on novobiocin action, the mechanism
involved in novobiocin resistance in halobacteria. This iden-
tification is based on the following evidence. The location of
the resistance determinant was mapped to a 1.4-kb Saillsubclone of the original 6.7-kb Kpnl fragment carrying the
resistance determinant. When sequenced, this subclone was
found to contain an ORF with a predicted protein sequence
strongly homologous to GyrB of eubacteria. Comparison of
this sequence with the wt Haloferax sequence revealed three
FIG. 5. Comparison of GyrB protein sequences from B. subtilis (BS) and the Haloferax Novr mutant (HF). The sequences were alignedto maximize their homology. Gaps introduced are indicated by dots. Positions with identical amino acids in both proteins are indicated byasterisks. Amino acid substitutions found in the Haloferax Novr GyrB compared with the wt protein are shown in bold (positions 83, 123, and138 in this alignment), with the amino acid found in the wt protein given immediately above.
point mutations in gyrB (all within the 1.4-kb Sall fragment),which all produce amino acid substitutions. Since novobio-cin has been shown to bind and inhibit the function of theDNA gyrase B subunit in eubacteria, it is reasonable toconclude that it inhibits halobacteria by the same mecha-nism.Novobiocin inhibits DNA gyrase activity by binding to
GyrB and blocking the access of ATP to its binding site onthis subunit (21, 27). Inhibition is competitive, but novobio-cin shows little structural similarity to ATP (27). Resistantbacteria have been shown to produce a gyrase that bindsnovobiocin less avidly (28), yet to date, the precise muta-tions responsible for novobiocin resistance in eubacteriahave not been published. The GyrB mutations observed inour study show three significant features: they occur in theamino-terminal region; they cluster together, spanning a
distance of only 55 residues; and they lie in a region that is
highly conserved across eubacteria, eucaryotes, and archae-bacteria (32; this study). These features would be consistentwith the possibility that the region defined by these muta-tions overlaps, or lies close to, the residues involved in ATPbinding. Unfortunately, these residues have not been iden-tified either by direct experimentation or by homology toconsensus sequence motifs (11, 30). Using the sequence Adescriptor of Bradley et al. (4), i.e., ,B strand-GXXXX(GK ahelix)-(0-11 aa)-,B strand, we could not find any perfectmatches in the Haloferax and B. subtilis GyrB sequences orthe corresponding region of human topoisomerase II, butthose most closely resembling the descriptor and conservedbetween the three organisms occurred in or close to theregion identified by the mutations (i.e., centered on glycineresidues at positions 37, 77, and 121 in the GyrB sequencesaligned in Fig. 6).
In eubacteria, DNA supercoiling is known to modulate
J. BACTERIOL.
NOVOBIOCIN RESISTANCE MARKER IN ARCHAEBACTERIA
A B3
G A C T 1 2
FIG. 6. Primer extension mapping of the Haloferax gyrB tran-
scriptional start site. Total RNAs extracted from wt (lane 1) and
Novr (lane 2) cells were used as templates. The sequence ladder
standard (lanes 0, A, C, and T) was from the same oligonucleotideprimer (primer 8) as that used in primer extension reactions. The
type of dideoxynucleotide used in the DNA sequencing reactions is
indicated in capital letters above the lanes. The region shown in
lowercase letters extends from nt 266 to 274 in Fig. 4.
gene expression (12) and is rigorously controlled by the
opposing activities of DNA gyrase and topoisomerase I,
which are balanced to produce an optimal level of superheli-cal tension. This balance is largely due to the level of DNA
gyrase, relaxation of the DNA resulting in the increased
expression of both gyrA and gyrll genes (19). This was
clearly demonstrated when the promoter region of E. coligyrB was fused to the galactokinase gene and changes inisupercoiing were readily observed to affect the level of
transcription (1). We have not yet investigated the regulationof gyrB expression in Halooferax cells, but from our analysisof sequences in the 5' leader and the surrounding ORFs it
may be rather complex.The discovery of the gyrA gene immediately downstream
from gyrB was of considerable interest, since it raises the
possibility that regulated expression of these genes is
achieved in part by including them both on the same mRNA.
Biologically, this would be reasonable since equal numbers
of GyrA and GyrB subunits (two of each) are required for
functional DNA gyrase (21). However, in E. ccli the two
genes are not closely spaced on the chromosome, and in B.
subtilis, even though the genes are about 210 bases apart,
they are transcribed independently (16, 22). The HaloferaxDNA gyrase genes are very closely spaced, but this does not
necessarily mean that they are cotranscribed, since there are
examples of halobacterial genes in which the stop and start
codons of two adjacent genes overlap and yet they are
transcribed separately (3).The presence on the opposite strand of ORF4, which
overlaps all of gyrB as well as part of gyrA, is particularlyinteresting since it may well play a role in the regulation of
FIG. 8. Northern hybridization analysis of Haloferax gyrB tran-scripts. Total RNAs from wt lane (1), Novr mutant (lane 2), andpMDS2 transformant (lane 3) strains of Haloferax were electro-phoresed through a formaldehyde-agarose gel (A), transferred to anylon membrane, and probed with the 1.4-kb SalI fragment withinthe gyrB gene (B). The photograph of the rightmost lane wasunderexposed, since the band was far darker than those in the twoadjacent lanes. Denatured 32P-labeled HindIII-digested lambda frag-ments and rRNA bands were used as size standards (indicated innucleotides at the left).
DNA gyrase expression. A similar ORF does not occur in E.coli or B. subtilis gyrB, nor could we find any homology tobacterial or eucaryotic proteins in computer data bases.However, the predicted ORF4 protein has all of the hall-marks of a typical halobacterial protein and the ORF isextremely long, features which argue in favor of it being areal gene and not a chance occurrence. In contrast, by thesame criteria, the ORF3 protein does not appear to be atypical halobacterial protein.Using the information gained in this study, we are now
able to significantly reduce the size of the resistance deter-minant of our vector by removing the region upstream of thegyrB promoter. When we have completed the gyrA se-quence, a further reduction in size (i.e., 3' to gyrA) can becontemplated. We have also mapped a fragment withinpHK2 which is sufficient for plasmid maintenance in Halo-ferax cells (unpublished data). Together, these reductionswill remove many inconvenient restriction sites and substan-tially improve the utility of the vector for genetic experimen-tation in halobacteria.
ACKNOWLEDGMENTSThis research was financed by a grant from the Australian
Research Council. M. H. was supported by an Australian Postgrad-uate Research Award.
BoxB
bopmyrA
CTTA AGTA TGCPuA Halophile consensussequence
FIG. 7. Nucleotide sequences 5' of H. halobium bop and Haloferax gyrB genes. Transcriptional start sites as determined by primerextension analyses are indicated (*). The putative halobacterial promoter sequences, box A and box B, are shown underneath.
VOL. 173, 1991 647
648 HOLMES AND DYALL-SMITH
We thank J. Praszkier for reading the manuscript. We are gratefulto J. K. Tamura and M. Gellert for communicating their resultsbefore publication.
ADDENDUM
After submitting this manuscript for publication, we were
sent a copy of a manuscript by Tamura and Gellert (27a).The authors found that the reactive ATP analog pyridoxal5'-diphospho-5'-adenosine bound to lysines 103 and 110 insubunit B ofDNA gyrase, thus indicating the position of theATP-binding domain. Our results are in remarkable agree-
ment.
REFERENCES1. Adachi, T., M. Mizuuchi, E. A. Robinson, E. Appella, M. H.
O'Dea, M. Gellert, and K. Mizuuhi. 1987. DNA sequence of theE. coli gyrB gene: application of a new sequencing strategy.Nucleic Acids Res. 15:771-783.
2. Bachmann, B. J. 1983. Linkage map of Escherichia coli K-12,edition 7. Bacteriol. Rev. 47:180-230.
3. Betlach, M. C., R. F. Shand, and D. M. Leong. 1989. Regulationof the bacterio-opsin gene of a halophilic archaebacterium. Can.J. Microbiol. 35:134-140.
4. Bradley, M. K., T. F. Smith, R. H. Lathrop, D. M. Livingston,and T. A. Webster. 1987. Consensus topography in the ATPbinding site of the simian virus 40 and polyomavirus largetumour antigens. Proc. Natl. Acad. Sci. USA 84:4026-4030.
5. Brown, J. W., C. J. Daniels, and J. N. Reeve. 1989. Genestructure, organization, and expression in archaebacteria. Crit.Rev. Microbiol. 16:287-338.
6. Charlebois, R. L., and W. F. Doolittle. 1989. Transposableelements and genome structure in halobacteria, p. 297-307. InD. E. Berg and M. M. Howe (ed.), Mobile DNA. AmericanSociety for Microbiology, Washington, D.C.
7. Charlebois, R. L., W. L. Lam, S. W. Cline, and W. F. Doolittle.1987. Characterization of pHV2 from Halobacterium volcaniiand its use in demonstrating transformation of an archaebacte-rium. Proc. Natl. Acad. Sci. USA 84:8530-8534.
8. Cline, S. W., and W. F. Doolittle. 1987. Efficient transfection ofthe archaebacterium Halobacterium halobium. J. Bacteriol.169:.1341-1344.
9. Cline, S. W., W. L. Lam, R. L. Charlebois, L. C. Schalkwyk,and W. F. Doolittle. 1989. Transformation methods for halo-philic archaebacteria. Can. J. Microbiol. 35:148-152.
10. DasSarma, S., U. L. RajBhandary, and H. G. Khorana. 1984.Bacterio-opsin mRNA in wild-type and bacterio-opsin-deficientHalobacterium halobium strains. Proc. Natl. Acad. Sci. USA81:125-129.
11. Fry, D. C., S. A. Kuby, and A. S. Milvan. 1986. ATP-binding siteof adenylate kinase: mechanistic implications of its homologywith ras-encoded p21, F1-ATPase, and other nucleotide-bindingproteins. Proc. Natl. Acad. Sci. USA 83:907-911.
12. Geliert, M. 1981. DNA topoisomerases. Annu. Rev. Biochem.50:879-910.
13. Gibbons, N. E. 1974. Family V. Halobacteriaceae fam. nov., p.269. In R. E. Buchanan and N. E. Gibbons (ed.), Bergey'smanual of determinative bacteriology, 8th ed. The Williams &Wilkins Co., Baltimore.
14. Gibson, T., and R. E. Gordon. 1974. Family I. Bacillaceae, p.
529. In R. E. Buchanan and N. E. Gibbons (ed.), Bergey'smanual of determinative bacteriology, 8th ed. The Williams &Wilkins Co., Baltimore.
15. Holmes, M. L., and M. L. Dyall-Smith. 1990. A plasmid vector
with a selectable marker for halophilic archaebacteria. J. Bac-teriol. 172:756-761.
16. Lampe, M. F., and K. F. Bott. 1985. Genetic and physicalorganization of the cloned gyrA and gyrB genes of Bacillussubtilis. J. Bacteriol. 162:78-84.
17. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecularcloning: a laboratory manual. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.
18. Mankin, A. S., and V. K. Kagramanova. 1988. Complex pro-moter pattern of the single ribosomal RNA operon of anarchaebacterium Halobacterium halobium. Nucleic Acids Res.16:4679-4692.
19. Menzel, R., and M. Geliert. 1983. Regulation of the genes for E.coli DNA gyrase: homeostatic control of DNA supercoiling.Cell 34:105-113.
20. Mevarech, M., and R. Werczberger. 1985. Genetic transfer inHalobacterium volcanii. J. Bacteriol. 162:461-462.
21. Mizuuchi, K., M. H. O'Dea, and M. Gellert. 1978. DNA gyrase:subunit structure and ATPase activity of the purified enzyme.Proc. Natl. Acad. Sci. USA 75:5690-5963.
22. Moriya, S., N. Ogasawara, and H. Yoshikawa. 1985. Structureand function of the region of the replication origin of the Bacillussubtilis chromosome. III. Nucleotide sequence of some 10,000base pairs in the origin region. Nucleic Acids Res. 13:2251-2265.
23. Rich, A., A. Nordheim, and A. J-J. Wang. 1984. The chemistryand biology of left-handed Z-DNA. Annu. Rev. Biochem.53:791-846.
24. Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequenc-ing with chain-terminating inhibitors. Proc. Natl. Acad. Sci.USA 74:5463-5467.
25. Shimmin, L. C., and P. P. Dennis. 1989. Characterization of theLii, Li, L10, and L12 equivalent ribosomal protein genecluster of the halophilic archaebacterium Halobacterium cutir-ubrum. EMBO J. 8:1225-1235.
26. Sioud, M., 0. Possot, C. Elie, L. Sibold, and P. Forterre. 1988.Coumarin and quinolone action in archaebacteria: evidence forthe presence of a DNA gyrase-like enzyme. J. Bacteriol. 170:946-953.
27. Sugino, A., N. P. Higgins, P. O. Brown, C. L. Peebles, and N. R.Cozzareili. 1978. Energy coupling in DNA gyrase and themechanism of action of novobiocin. Proc. Natl. Acad. Sci. USA75:4838-4842.
27a.Tamura, J. K., and M. GelHert. J. Biol. Chem., in press.28. Thiara, A. S., and E. Cundliffe. 1988. Cloning and characteriza-
tion of a DNA gyrase B gene from Streptomyces sphaeroidesthat confers resistance to novobiocin. EMBO J. 7:2255-2259.
29. Thomm, M., and G. Wich. 1988. An archaebacterial promoterelement for stable RNA genes with homology to the TATA boxof higher eukaryotes. Nucleic Acids Res. 16:151-163.
30. Walker, J. E., M. Saraste, M. J. Runswick, and N. J. Gay. 1982.Distantly related sequences in the a- and f-subunits of ATPsynthase, myosin, kinases and other ATP-requiring enzymesand a common nucleotide binding fold. EMBO J. 1:945-951.
31. Woese, C. R., and G. J. Olsen. 1986. Archaebacterial phylog-eny: perspectives on the urkingdoms. Syst. Appl. Microbiol.7:161-171.
32. Wycoff, E., D. Natalie, J. M. Nolan, M. Lee, and T.-S. Hsieh.1989. Structure of the Drosophila DNA topoisomerase II gene.Nucleotide sequence and homology among topoisomerase II. J.Mol. Biol. 205:1-13.
33. Yanisch-Perron, C., J. Vieira, and J. Messing. 1985. ImprovedM13 phage cloning vectors and host strains: nucleotide se-quences of the M13mpl8 and pUC19 vectors. Gene 33:103-119.