Vol. 35, No. 3
ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, Mar. 1991, p. 471-476
0066-4804/91/030471-06$02.00IO Copyright © 1991, American Society for Microbiology
Relationship between the Clostridium perfringens catQ Gene Product and Chloramphenicol Acetyltransferases from Other Bacteria TRUDI L. BANNAM AND JULIAN I. ROOD* Department of Microbiology, Monash University, Clayton 3168, Australia Received 18 September 1990/Accepted 28 December 1990
The nucleotide sequence of the Clostridium perfringens chloramphenicol acetyltransferase (CAT)-encoding resistance determinant, catQ, was determined. An open reading frame encoding a protein of 219 amino acids with a molecular weight of 26,014 was identified. Although catQ was expressed constitutively, sequences similar in structure to those found upstream of inducible cat genes were observed. The catQ gene was distinct from the C. perfringens catP determinant. The deduced CATQ monomer had considerable amino acid sequence conservation compared with CATP (53% similarity) and other known CAT proteins (39 to 53%). Phylogenetic analysis revealed that the CATQ monomer was as closely related to CAT proteins from Staphylococcus aureus and Campylobacter coli as it was to CAT monomers from the clostridia.
The enzyme chloramphenicol acetyltransferase (CAT) mediates the inactivation of the antibiotic chloramphenicol, a potent inhibitor of prokaryotic peptidyltransferase activity. The active CAT enzyme, which catalyzes the acetyl coenzyme A-dependent acetylation of chloramphenicol, is a trimer of identical subunits of approximately 25 kDa (21, 34). Numerous CAT variants have been isolated from a diverse range of bacterial species (32, 34). Although the DNA sequences of various cat genes have little similarity, each of the enzymes functionally is very similar, and their primary structures are partially conserved, especially in the region that surrounds the histidyl residue which forms part of the catalytic site (22, 34). Expression of many of the cat genes, such as cat-86 from Bacillus pumilus (38) and the cat genes located on the Staphylococcus aureus plasmids pUB112 (7) and pC194 (18), is induced by subinhibitory concentrations of chloramphenicol. In contrast, cat genes of gram-negative origin generally are expressed constitutively. Research in this laboratory has for several years involved the genetic analysis of antibiotic resistance determinants from the gram-positive, anaerobic bacterium Clostridium perfringens (1-4). Although chloramphenicol resistance is not common in C. perfringens, several resistant isolates have been identified (1, 6), and resistance has been shown to be due to the production of CAT enzymes (30, 40). Hybridization analysis has demonstrated the existence of two distinct determinants in C. perfringens. The catP determinant, from the conjugative resistance plasmid pIP401 (3, 6), is located on the transposable element Tn4451 (2), whereas the catQ gene is chromosomally determined (29). The nucleotide sequence of the catP gene has been determined (36) and is almost identical to that of the catD gene from Clostridium difficile (39). In this paper we report the determination of the nucleotide sequence of the C. perfringens catQ determinant and describe the relationship of the deduced amino acid sequence to the sequences of CAT monomers from other bacteria.
*
MATERIALS AND METHODS
Bacterial strains, plasmids, and media. All strains were derivatives of Escherichia coli DH5cc and were grown in 2YT medium (25) supplemented with ampicillin (100 pg/ml) and/or chloramphenicol (30 ,ug/ml) where appropriate. Transformation methods were as previously described (24). Recombinant plasmids (pJIR240, pJIR241, and pJIR260) carrying the catQ determinant from C. perfringens CW531 were as previously described (29). The plasmid pJIR370 was constructed by ligating the 0.5-kb DraI-EcoRI fragment from pJIR260 (29) into Smal-EcoRI-digested pUC18. DNA isolation and sequence analysis. Plasmid DNA templates for sequencing were isolated from 100-ml cultures of spectinomycin (400 ,ug/ml)-amplified E. coli cells by using a modified alkaline lysis procedure (5). The DNA was purified by centrifugation in a CsCl-ethidium bromide gradient for 20 h at 53,000 rpm in a Beckman Ti7O rotor (24). The oligonucleotides used as primers for DNA sequencing were produced on an Applied Biosystems (Foster City, Calif.) DNA synthesizer as specified by the manufacturer. All sequence data were obtained by the double-stranded dideoxy-chain termination method by using either Pharmacia T7 or USB Sequenase 2.0 sequencing kits. Newly synthesized strands were labeled with "S-dATP (Amersham). Sequencing data were obtained from the available subclones (pJIR240, pJIR241, pJIR260, and pJIR370) by using both universal and reverse sequencing primers (Pharmacia). Synthetic oligonucleotide primers were utilized to extend the sequence data derived from the subclones. Unambiguous sequence was obtained on both strands and extended across all restriction sites. Computer analysis. The nucleotide sequence was manipulated and assembled by using the MELBDBSYS program which was developed at the Walter and Eliza Hall Institute, the Ludwig Institute for Cancer Research, and the Howard Florey Institute (Melbourne, Australia). Sequence alignments and phylogenetic relationships were determined by using the TREEALIGN (February 1990) program developed by Jotun Hein (14-16). To optimize the phylogenetic tree of CAT monomers, the gap penalty function of gk = 8 + (3 k) was utilized, where k is the length of the insertions-deletions. Nucleotide sequence accession number. The GenBank ac-
Corresponding author. 471
472
ANTIMICROB. AGENTS CHEMOTHER.
BANNAM AND ROOD TABLE 1. Sources of cat sequences
Organism
Clostridium perfringens Clostridium perfringens Clostridium difficile Campylobacter coli Staphylococcus aureus Staphylococcus aureus Staphylococcus aureus Staphylococcus aureus Proteus mirabilis Bacillus pumilus Escherichia coli Escherichia coli Streptomyces acrimycini
Designation or plasmid
CATQ CATP CATD CCOLI pC221 pUB112 pC223 pC194 PMIR cat-86 ECO I ECO III SACR
Accession no.
CTGCGTACAC ATCCAGACAT CGCTTTAGAG
Reference
TATGGTGALT TAAAGATGGA GCOGGCTTAT
100
N-1
GAAAAGGLTG CTTTGCGATG
I
GTGGCAAACT GTCTGTTAGG AGGTTATTCT CAAAGGATTG CALGCAG TTGAGOATAL TCCOTATAAC
M55620a M28717a X15100a
M35190a A00569b A24362b A00568b
A24651b A00570b A00566b Y00723a
This paper 36 38 37 33 7 36 17 8 12 35 27 26
aGenBank/EMBL nucleotide databases. b NBRF amino acid database.
200
150
TAACTATTAC ACATTCTTAA CATTGCTGGT TTGTATCGGT AGALTAACAC GAATTAACAA AGGATATATT 250
cession number of the DNA sequence of the catQ determinant is M55620. Accession numbers of other cat genes are indicated in Table 1. RESULTS Determination and analysis of catQ sequence. The chloramphenicol resistance determinant, catQ, from C. perfringens CW531, was cloned previously into the E. coli vector pUC18 and was localized by subcloning and BAL31 deletion analysis. The plasmid pJIR235 contains a 1.25-kb insert and is the smallest recombinant which confers chloramphenicol resistance (29). The nucleotide sequence of almost the entire pJIR235 insert (1,158 bp) was determined on both strands. A single open reading frame (ORF) of 657 bp beginning at nucleotide 459 was identified (Fig. 1). There is a consensus ribosome binding site located 8 bp upstream of the putative ATG start codon. In addition, there are consensus -35 and -10 promoter regions, with an interval gap of 17 nucleotides, located at positions 346 and 323, respectively. The sequence of these regions is consistent with both E. coli and C. perfringens consensus promoter sequences (11, 13, 31). The catQ gene sequence contains internal DraI, HaeIII, Hindll, and Sau3A sites, which is consistent with previous mapping data (29). The G+C content of the catQ ORF is 35%, which is similar to that of the C. perfringens catP (34%) (36) and C. difficile catD (34%) (39) genes. These values are somewhat higher than the G+C content of the C. perfringens genome, which is 24 to 27% (11, 19). Comparison of the catQ, catP, and catD nucleotide sequences confirmed the previous hybridization results (29), since it revealed that there was no significant sequence similarity between catQ and either catP or catD. In addition, the DNA sequence of catQ was not homologous to any of the known cat genes from other genera
Comparison of the deduced CATQ amino acid sequence to those of other CAT variants. The amino acid sequences of 12 CAT monomers, including the clostridial CATP and CATD proteins, have been determined or elucidated (Table 1). The deduced amino acid sequence of the catQ gene product, CATQ, corresponded to a protein of 219 amino acids with a molecular weight of 26,014. The size is consistent with the sizes of previously identified CAT monomers which range from 212 to 220 amino acids (34). We have aligned the sequences of each of these monomers to give the maximum
-10
-30
TTGTAGTAOC AAOTGTATTT GTTTTATATT CTATGALCCT ATTTOLTTTA CATAATTCTA TGTGOTATAA 350
300
TGGLCTTGCT TATAATATAL ATGCTTGOAT AOC? TOCAGO Lou TTA
Ala Ph. GCC TTC TAL AATTAAAGAA
TAAAATTTTWGATAAT
met Nat Met Ala Va1 Lys ATG ATG ATG OCA GTT AAL 400
Mat Lys Ph.
ALA
TT ATG
TTT
Asn Lou AAT TTG
Ile ATA
450 Va1 Arg Asp Ile Glu Asp Trp Asn Arg Lys Pro Tyr Ph. Glu HIs Tyr Lou Asn Ala OTT TTT GAG CAT TAT TTA AAT GCM CCA AGG OLG ATT TAC AGA AAG GAT TGG AAT GAT Dr& I
500
Cys Thr Tyr S r Mat Thr Ala Asn Ie Glu Ile Thr Gly Lau Lou Arg Glu Ile Lys TOC ACT TAC AGT ATG ACT GCA AAT ATA GAG ATA ACT GGT TTA CTO CGT GAL ATT AAA 550
Lsu Lys Gly Lou CTT AAG GGC CTG
Lyn Lou Tyr ALA CTG TAC
Pro Thr Lsu Ile Tyr Ile Ile Thr Thr CCT ACG CTT ATT TAT ATC ATC ACL ACT
Va1 Va1 Asn GTG GTT AAC HInd II
Arg His Lys Glu Ph. Arg Thr Cys Ph. Asp Gln Lys Gly Lys Lou Gly Tyr Trp Asp CGT CAC AOG GAG TTC CGC ACC TGT TTT GAT CAA AAA GGT AAG TTA GGA TAC TGG GAT 700
Oau 3A
650
(29).
CGATTCTCAC
Taq I
50
AGGATATTGA AGGCTACTGC ACTGGTAAGG ATGCATTTGT AAAGCAACTA
Thr Va1 Ph. His AGT ATG AAC CCA AGT TAT ACT GTC TTT CAT
Oar Mat Asn
Pro
SOr Tyr
Lys Asp Asn Glu Thr Ph. Sar SOr Ile AAG GAT AAC GOA ACT TTT TCA AGT ATT 750
Trp Thr Glu Tyr Asp Glu Asn Ph. Pro Arg Ph. Tyr Tyr Asn Tyr Lou Glu Asp I1e TGO ACA GAG TAT GAC GAG AAC TTC CCA CGT TTT TAC TAT ALT TAC CTT GAG OAT ATT Arg Asn Tyr Oar Asp Va1 Lou Asn Ph. Mat Pro Lys Thr Gly Glu Pro Ala Asn Thr AGA AAC TAT Mff GAC GTT TTG AAT TTC ATG CCT AAG ACA GGT GAA CCT GCT ALT ACA Has
~~~~~~~850
Ile
Asn
Va1
ATT AAT GTG
SOr SOr TCC
AGC
Ile Pro ATT CCT
Ans Asp
Ala Thr Tyr Lou Ile ALT GAT GCA ACA TAT CTA LTC 950
Asn Lys ALL AAT
Ile Lou Lou Pro Mat ATT TTA TTA CCT ATG
Asn Lou Asn AOC CTG ALT
Trp Val Asn Ph. Thr Gly Ph. TOG GTG AAT TTT ACC GGA TTC 900 Pro Ile Pha Thr Lou CCT ATT TTT ACT TTO
Ile ATA
Tyr TAC
Gly Lys Tyr Ph. Gln Gln Asp G5T AAL TAT TTT CAC CAC GAT
Oar Va1 Gln Va1 01l Hlis TCT GTA
Ala Va1 Cys Asp Gly Tyr CAG GTG CAT CAT OCG GTT TGC GAC GOT TAT
1000
His Ile Arg Ph. CAT ATA AOC AGA TTT 1050
Oer
Ph. Asn Glu TTT AAT
GOG
Ala
Gln
GCA CAG
Oer GAL TTA GCG TCA
Glu Lou Ala
Lou Gly Glu Lys TTA GGA GOA AAA TAA TATAAALTT GCTTGTGTTT
GGGOGCTATA
Asn Tyr Glu Thr Trp ALT TAT GAG AOL TOO 1100
OLLTC 1158
FIG. 1. Nucleotide sequence of the catQ determinant. The open reading frame beginning at nucleotide position 459 corresponds to the predicted CATQ monomer of 219 amino acids. The open reading frame beginning at nucleotide position 397 corresponds to a potential leader peptide of 9 amino acids. Symbols: *, stop codons; Li1, ribosome binding sites; -* and MI >
SE
AIN-QDGE
SEA
E
PI0.1
FEKDM CKGYEI FKDMX
KGYELP
LKS LAI
I
GYEIP
EKKL
IN-QDGE
Z IFjEMI
Q...I.
CKSDDSE 108
ii jTF
S
108
IIISENG CKSD
*FA
-N
108 112
111 111 111 111 109 117 116 116 111
ECOI ECOIII
168 168 168 172 171 171 171 171
169 177 175 175
LKTMS]
LHIMS(
MVNML'
171
IIKKDI
ICR.-------F
ELI-IV--TQVCL ----
IIKKDI YYEEG(
ICR-------F
ELI-IV--TQVCL ----
VCR-------FL ISR-------FF
DLL-N-----
pC194 pC221 pUB112 pC223 CAT-86
NGL-------F
ELS-DR--PNDWL --- L
SACR
TAR-------LT
CATP CATD CCOLI CATQ
58 58 58 53
LEPK SAY LKKS
PMIR
PMIR ECOI ECOIII
51
NGYFP
LLK
pC194 pC221 pUB112 pC223 CAT-86 SACR
CATP CATD CCOLI CATQ pC194 pC221 pUB112 pC223 CAT-86 SACR
52 52 52 52
EGYKF
YRN
VKS
SACR PMIR ECOI
CATP CATD CCOLI CATQ
WIITE
NQ-Q TLTRC
TFNIINL-------ET
KGLKL?
ILLRE
I PKE M;I NQ-QTI
TFNIIKL-----E
KQI--D-------
GKKL
LNAVRcrY
D
NKIDL-------
KGMKL
GNTPC
-------
NLIDI-------ED
49 49 49 53
KGIKLP
ASVPCY
473
PMIR ECOI ECOIII
212 212 207 219 216 215
K ---
ELA-SN--YETWLGEK-
WSL-------F
DII-HK--VDDWI----I]DII-NK--VDDWI-----
ASL-------FIl VSL-------F
NII-DN--VNEWI-----
rGQYVEYLRWLI
EWL-ND--SLHIT ----
215 215 220 220 217
TLLADP--AW--L ----
GR-------LL
QYC-DE--GC--K ----
219 213
QYC-DEWQGG--A----
GR-------L AR------FI
JELC-NS--KL--K
----
of FIG. 2. Amino acid alignment of CAT monomers. The shaded regions represent conserved positions and positions in which at least 10 designated are monomers CAT site. active coli E. the of formation the in are involved which acids Amino *, 13 amino acids are conserved. as shown in Table 1.
DISCUSSION In many gram-positive bacteria chloramphenicol resistance is an inducible phenotype. The genetic basis for the inducibility of the B. pumilus cat-86 gene and the S. aureus cat gene from pUB112 has been the subject of extensive investigation (7, 23). The mechanism that has been implicated is known as translational attenuation. It involves the gene for a small leader peptide which is encoded directly upstream of the cat gene and an inverted repeat sequence which is able to form a stem and loop structure which sequesters the cat gene ribosome binding site. In the presence of chloramphenicol, the ribosome translating the leader peptide stalls, which leads to destabilization of the RNA stem and loop, thereby making the ribosome binding site of the cat gene available for the initiation of translation (23). Examination of the coding region 5' to the translational start point of catQ revealed the presence of a possible leader peptide of 9 amino acids, which is the same length as the leader peptide from both cat-86 and pUB112 and is preceded
TABLE 2. Relationships between CAT monomersa Amino acid sequence
Determinant
1. CATQ 2. CATP 3. CATD 4. CCOLI 5. pC221 6. pUB112 7. pC223 8. pC194 9. PMIR
1
2
3
4
100 53 53 50 100 98 58 100 59 100
similarity (%)
5
6
7
8
9
10
11
12
13
51 44 44 45 100
51 44 44 45 98 100
51 46 46 46 80 80 100
48 46 46 44 54 53 57 100
45 43 44 44 40 39 40 41 100
44 40 40 41 42 43 41 41 38 100
41 44 45 44 38 38 40 39 76 39 100
41 43 43 41 37 37 38 40 47 41 47 100
39 38 39 40 31 31 33 33 35 32 38 40 100
10. cat-86 11. ECO I 12. ECO III 13. SACR a
Determinants are designated as shown in Table 1.
474
ANTIMICROB. AGENTS CHEMOTHER.
BANNAM AND ROOD
X
X
M
X
[FJUC X
cat-86
S
AAA UCA
pUB112
catQ
X
V
IICjGG
31
A
N
T
Y
S
S
*
GAG G&U UGU UCC UCC GAL AUA CAA AGL GUG AAA A
V
AUG GCA GUC
x
D
D
X
L
A
r
UEG ibA
*
AAA UGA GCC UUC UAA AAU GAA AGA AUA AAA UJU
X
I
S
S
GUU AGAUA
+C
UG AAU GD [
]
*
AAA ACAU AAA AUC UCC UCC UG
AUA CA
GAL AUC ACA UC [
A" |GAL AA [AE
FIG. 3. Comparison of putative cat leader sequences. For pUB112 and cat-86 the arrows delineate inverted repeats which, when formed into a stem and loop structure, sequester the CAT ribosome binding site. The 9-amino-acid leader peptides are indicated, as are the start codons of the CAT monomers AUG and UUG. The upstream region of the catQ gene also encodes a possible 9-amino-acid leader peptide and contains an inverted repeat of 20 bp (->,