Eur. J. Biochem. 147,437-446 (1985) 0 FEBS 1985

The nucleotide sequences of the ponA and ponB genes encoding penicillin-binding proteins 1A and 1B of Escherichia coli K12 Jenny K. BROOME-SMITH, Alex EDELMAN, Samira YOUSIF and Brian G. SPRATT Microbial Genetics Group, School of Biological Sciences, University of Sussex, Brighton

(Received September 28/November 27, 1984) - EJB 84 1042

Penicillin-binding proteins 1A and 1 B of Escherichia coli are the major peptidoglycan transglycosylase- transpeptidases that catalyse the polymerisation and insertion of peptidoglycan precursors into the bacterial cell wall during cell elongation. The nucleotide sequence of a 2764-base-pair fragment of DNA that contained the ponA gene, encoding penicillin-binding protein 1 A, was determined. The sequence predicted that penicillin-binding protein 1A had a relative molecular mass of 93500 (850 amino acids). The amino-terminus of the protein had the features of a signal peptide but it is not known if this peptide is removed during insertion of the protein into the cytoplasmic membrane. The nucleotide sequence of a 2758-base-pair fragment of DNA that contained the ponB gene, encoding penicillin-binding protein 1 B, was also determined. Penicillin-binding protein 1 B consists of two major components which were shown to result from the use of alternative sites for the initiation of translation. The large and small forms of penicillin-binding protein 1B were predicted to have relative molecular masses of 94 100 and 88 800 (844 and 799 amino acids). The amino acid sequences of penicillin-binding proteins 1 A and 1 B could be aligned if two large gaps were introduced into the latter sequence and the two proteins then showed about 30% identitiy. The amino acid sequences of the proteins showed no extensive similarity to the sequences of penicillin-binding proteins 3 or 5, or to the class A or class C p-lactamases. Two short regions of amino acid similarity were, however, found between penicillin-binding proteins 1 A and 1 B and the other penicillin-binding proteins and fi-lactamases. One of these included the predicted active-site serine residue which was located towards the middle of the sequences of penicillin-binding proteins 1 A, 1B and 3, within the conserved sequence Gly-Ser- Xaa-Xaa-Lys-Pro. The other region was 19 - 40 residues to the amino-terminal side of the active-site serine and may be part of a conserved penicillin-binding site in these proteins.

fi-Lactam antibiotics kill Escherichia coli by inactivating a series of penicillin-binding proteins (PBPs) that are located in the cytoplasmic membrane and which catalyse the final stages of peptidoglycan synthesis [l, 21. In E. coli there are seven well characterised PBPs with relative molecular masses (Mr) between 40000 and 91 000 [3]. Genetic analysis of the role of each of the E. coli PBPs in the lethal effects of p-lactam antibiotics has shown that the lower-Mr PBPs 4, 5 and 6 are non-essential for bacterial growth and are therefore not of primary importance in the killing action of p-lactams [l , 4-71. The higher-M, PBPs 1A/lB, 2 and 3 have been shown to be essential enzymes and p-lactams exert their lethal effects by the inactivation of one or more of these killing targets [l , 81. Inactivation of PBP 3 results in the inhibition of cell division and the formation of long filamentous cells [9]; inhibition of PBP 2 results in the growth of E. coli as spherical cells [S]. The ability of p-lactams to cause rapid lysis of E. coli cells was shown to correlate with their binding to PBP 1 [S].

Abbreviations. PBP, penicillin-binding protein, SDS, sodium dodecyl sulphate.

Enzymes. Calf intestinal alkaline phosphatase (EC 3.1.3.1); DNA polymerase I (Klenow fragment) (EC 2.7.7.7); phage-T4-DNA polymerase (EC 2.7.7.7); T4-DNA ligase (EC 6.5.1.1); P-lactamase (EC 3.5.2.6); restriction endonucleases: BamHI (EC 3.1.23.6), BglII

(EC 3.1.23.24), Hind111 (EC 3.1.23.21), MluI (EC 3.1.23.), PstI (EC 3.1.23.31), PvuII (EC 3.1.23.33), SaA (EC 3.1.23.27), Sau3A (EC 3.1.23.27), SmaI (EC 3.1.23.44), SphI (EC 3.1.23.108), TaqI (EC 3.1.23.39) and XhoI (EC 3.1.23.42).

(EC 3.1.23.10), EcoRI (EC 3.1.23.13), EcoRV (EC 3.1.23.-)), H@I

Subsequently PBP 1 was shown to consist of two components, PBP 1A and PBP lB, and the lytic ability of /?-lactams cor- related with their affinity for PBP 1 B but not with their affinity for PBPlA [lo, 111. Mutant studies have shown that inactivation of either PBP 1A or PBP 1B is not lethal under laboratory conditions but double mutants, lacking both PBP 1A and PBP 1B activity, are non-viable [lo, 111 (and S. Y., J. K. B.-S. and B. G. S., unpublished results). Rapid cell lysis of E. coli therefore appears to require the simultaneous inactivation of both PBP 1A and PBP 1B and it has been proposed that these proteins have compensatory roles in peptidoglycan synthesis [l 11.

PBP 1 B has been identified as the major penicillin-sensitive peptidoglycan transpeptidase activity in crude cell envelopes [12]. The purified protein catalyses the synthesis of cross- linked peptidoglycan from lipid-linked peptidoglycan pre- cursors and the enzyme is therefore bifunctional, catalysing both the penicillin-insensitive peptidoglycan transglycosylase reaction and the penicillin-sensitive peptidoglycan trans- peptidase reaction [13, 141. PBP 1B migrates on SDS/poly- acrylamide gels as multiple components but each of the com- ponents have indistinguishable enzymatic activities [ 151. PBPlA has also been demonstrated to be a bifunctional penicillin-sensitive peptidoglycan transglycosylase-transpep- tidase but, in vitro, produces peptidoglycan with a higher level of peptide cross-linking than that synthesised by PBP 1 B [16].

PBP3 [17], and possibly PBP2 1181, also appear to be bifunctional peptidoglycan transglycosylase-transpeptidases that have specific functions in the synthesis of the cross-wall

438

at cell division and (perhaps) in the initiation of new sites of cell wall growth, respectively El, 81. The low-M, PBPs catalyse the D-alanine carboxypeptidase reaction and the precise role of these enzymes in peptidoglycan synthesis is still unclear

The amino acid sequences of one D-alanine carboxy- peptidase (PBP 5 of E. coli) and one high-M, PBP (PBP 3 of E. coli) have been reported [19, 201. These PBPs show no extensive sequence similarity to one another or to either the class A or class C p-lactamases. There are however small regions of similarity between the amino-terminal regions of D-alanine carboxypeptidases and class A p-lactamases [ 19, 21 1, and between the amino-termini of the above proteins and a region towards the middle of the primary sequence of PBP 3

We report here the subcloning and nucleotide sequencing of the genes encoding PBP 1A (ponA) and PBP 1B @onB) of E. coli and discuss the sequence similarities between PBP 1A and PBP 1B and other proteins that interact with p-lactam antibiotics .

[I, 21.

ti, 191.

MATERIALS AND METHODS

Bacterial strains, plasmids and bacteriophage

Escherichia coli C600 thr leu thi lac Y tonA supE was used in most experiments. SP61 trp tyr ilv sup-126 ponA6 and SP1026 his supFponB, which lack PBP 1 A and PBP 1 B activity respectively, were used for the detection of plasmids that ex- press theponA orponB genes. JMl01 was used for the growth of M13 phage derivatives [22].

The plasmid vectors pBR322 and pLG339 have been de- scribed 1231. The vector pPH126 is a high-copy-number (rop-) derivative of pBR322 in which the Ap‘ gene has been removed by replacing the region between the PstI and EcoRI sites of pBR322 with a 1696-base-pair fragment from Tn903 that encodes Km‘ (Hedge, P. J. and B. G. S., unpublished). pLC29-47 and pLC19-19 are from the collection of Clarke and Carbon [24] and carry the ponA and ponB genes re- spectively [25]. M13mp8 and M13mp9 bacteriophage have been described [22].

The presence of the tonA gene on plasmids was detected by their ability to make E. coli C600 (tonA) sensitive to phage T5.

Assay of penicillin-binding proteins

[3H]benzylpenicillin (27 Ci . mmol-’) as described 13, 71. PBPs were assayed using freeze-thawed bacteria and

Manipulation of DNA and nucleotide sequencing

Preparation of plasmid and M13 phage DNA, ligation, agarose gel electrophoresis, and transformation were carried out as described [23]. M13 phage clones for nucleotide sequencing were obtained by generating libraries of Sau3A, HpaII, TaqI and randomly sheared fragments of the ponA and ponB genes in M13mp8 or M13mp9. Randomly sheared fragments were produced by ultrasonic treatment and were end-tilled with T4-DNA polymerase, size-fractionated, and cloned into dephosphorylated SmaI-digested M13mp8 DNA as described by Deininger [26]. Additional M13 recombinants were obtained by cloning fragments using convenient restric- tion endonuclease cleavage sites in the ponA and ponB genes. Sequencing of M13 phage templates was carried out by the dideoxy method using deoxyadenosine 5’-(~[~’S]thio)tri-

phosphate, a 17-nucleotide universal primer, and 6% polyacrylamide gradient gels as described 1271 except that ‘gradients’ were generated by using wedge-shaped gels that were 0.25 mm thick at the top and 0.50 mm thick at the bottom.

Computer analysis

DNA sequence was assembled and analysed using the computer programs of Staden [28].

Amino acid sequences of proteins were compared using the computer graphics program DIAGON [29]. The program produces a similarity matrix by comparing segments of a chosen span length of the two proteins and allots them scores based on the amino acid residues at each position in the segments. The scores are calculated using the MDM78 amino acid matrix of Dayhoff [30] which takes into account accepted point mutations that occur in related proteins. The statistical significance of a particular score is assessed by the double matching probability [31] which gives the probability of that score being obtained in comparisons of infinitely long sequences of the same amino acid composition as the proteins being compared. A double matching probability of 2 x implies that the score would occur by chance in 1 in 5000 comparisons of the randomised sequences.

The method of Kyte and Doolittle [32] was used to analyse the hydropathy along the amino acid sequence of proteins. The average hydropathy of segments of 19 residues was plotted at the mid-point of the segment. Hydrophobic regions give positive scores and hydrophilic segments give negative scores.

Chemicals

Restriction endonucleases were obtained from Amersham International, Boehringer Mannheim, and Bethesda Research Laboratories. T4 DNA polynierase was purchased from Pharmacia PL Biochemicals, DNA ligase from Boehringer Mannheim, and the Klenow fragment of DNA polymerase I and SmaI-digested, dephosphorylated, M13mp8 DNA from Amersham International. The 17-mer universal M13 sequenc- ing primer d(G-T-A-A-A-A-C-G-A-C-G-G-C-C-A-G-T) was purchased from Celltech. Deoxyadenosine S’-(C(-[~’S]- thio)triphosphate (410 Ci . mmol-’) was obtained from Amersham International and [~henyl-4-~H] benzylpenicillin (27 Ci . mmol-’) was a generous gift of Merck Sharp and Dohme.

RESULTS

Subcloning and sequencing of the ponA gene

The PBP 1 A gene ( p o d ) has been located on the plasmid pLC29-47 from the Clarke and Carbon collection [24, 251. This plasmid was digested partially with Sau3A and fragments of 5 - I0 x lo3 base pairs were purified and inserted into the BumHI site within the tetracycline-resistance gene of the vector pLG339 [23]. Escherichia coli SP61 (ponA -) carrying recombinant plasmids were assayed for the production of PBP 1A. Comparison of the restriction maps of plasmids that expressed PBP 1A indicated the approximate location of the ponA gene. One of the latter plasmids (pSY42) was digested partially with BglII and completely with SphI and the 3.2 x 103-base-pair partial BglII-SphI fragment was subcloned into the plasmid pPH126 (Fig. 1) to produce pBS98. E. coli

439

which is in excellent agreement with the estimate of 91000 obtained from SDS/polyacrylamide gels [I, 31.

The end of the open reading frame is immediately followed by a 9-base-pair inverted repeat centred at position 2702 that may function as the transcription terminator of theponA gene.

PSY42 14.3Kb

Fig. 1. Subcloning ofthe ponA gene. B, BurnHI; Bg, BglII; E, EcoRI; H, HindIII; P, PstI; Sp, SphI; X, XhoI. Kb = x lo3 base pairs

PBP -1 A *l B Large ‘1 BS ma1 I

A B C D E F G H I J K Fig. 2. Assays of PBPs in E. coli currying recombinant plasmids. (A) SP61; (B) SP61 (pBS98); (C) SP1026; (D) C600 (pBS96); (E) C600 (pBS74); (F) C600 (pBS91); (G) SP1026 (pBS92); (H and K) SP1026 (pBS94); (I) SP1026 (pBS94C); (J) SP 1026 (pBS94B). Only the plasmid-encoded PBPs are clearly visible at this exposure

carrying pBS98 produced high levels of PBP 1A indicating that the ponA gene was carried on this fragment (Fig. 2B). The ponA gene appeared to extend across the XhoI, PstI and BgAI sites within the chromosomal DNA of pBS98 (Fig. 1) since insertion of DNA fragments (bacteriophage lambda DNA digested with XhoI + Sari, PstI, or BgTII, respectively) into each of these sites inactivated the gene.

The ponA gene was sequenced using the dideoxynucleotide method as described in Methods and the sequence from the junction BgAI site of pBS98 to an MluI site after the end of the gene is shown in Fig. 3. The sequence was fully overlapped and 98.5% of the sequence was determined on both strands. One major open reading frame extended from position 103 to a stop codon at positions 2686-2688. The triplet for AUG at positions 112-114 and that for GUG at positions 136- 138 are possible initiation codons and the method of Stormo et al. [33] was used to predict which of these was the more likely position for the start of translation of the ponA gene. The triplet for AUG at positions 112-114 gave very low scores (< - 100) on all three of the matrices of Stormo et al. [33] whereas that for GUG at positions 136 - 138 gave highly significant values of f 7 3 and +22 for the W71 and WlOl matrices respectively although it did not score as a significant gene start with the W51 matrix. Initiation of translation therefore probably occurs at nucleotides 136 - 138 and would result in an amino-terminal sequence with the characteristics of a signal peptide.

The size of PBP I A that is derived from the nucleotide sequence is 850 amino acids with a calculated M , of 93500

Subcloning and sequencing of the ponB gene

The PBPIB gene @onB) has been located on pLC19-19 from the Clarke and Carbon collection 1251 and can be obtained from this plasmid as two PstI fragments [25]. pLC19-19 was digested partially with PstI and the fragments were inserted into the PstI site within the ampicillin-resistance gene of pBR322. E. coli C600 carrying recombinant plasmids were assayed for the overproduction of PBP 1 B and a plasmid, pBS74, that expressed the PBP 1 B gene was obtained (Fig. 2E). Various fragments from pBS74 were subcloned into pPH126, or into a derivative of pBR328 from which the P-lac- tamase gene had been deleted, and the resulting plasmids were examined for the expression of the PBP 1B gene (Fig. 4). E. coli C600 carrying plasmids (e.g. pBS94) which contained the 3.2 x 103-base-pair EcoRV-MluI fragment of pBS74 overproduced both components of PBP 1 B whereas plasmids that contained the 2.41 x 103-base-pair EcoRV-SphI fragment (e.g. pBS92) produced truncated forms of the components of PBP 1B that were approximately 6 kDa smaller than the normal forms (Fig. 2G, H). The PBPIB gene is therefore transcribed from the EcoRV site towards and across the SphI site and the end of the coding region should be located about 200 base pairs after the SphI site. The position of the gene differs considerably from that reported by Takeda et al. 1261.

Analysis of the nucleotide sequence of PBP 1B (see below) showed that the end of the gene and the putative transcription terminator could be otained on a 347-base-pair SphI-Tug1 fragment. This fragment was contained within a 463-base- pair TaqI fragment (nucleotides 2294-2758 in Fig. 6) which was cloned into the replicative form of phage M13mp8, and was excised from the M13 phage recombinant as a 354-base- pair SphI-BamHI fragment. The PBP 3B gene was obtained on a convenient DNA fragment by ligating the 2.41 x lo3- base-pair EcoRV-SphI fragment from pBS74 with the 354- base-pair SphI-BamHI fragment and inserting the resulting 2.76 x 1O3-base-pair EcoRV-BamHI fragment into pPHl26 to produce pBS96 (Fig. 5) . E. coli C600 (pBS96) produced high levels of the two components of PBP 1B (Fig. 2D). TheponB gene was sequenced using the dideoxynucleotide method as described in Methods and the sequence of the 2.76 x lo3- base-pair EcoRV-BumHI fragment of pBS96 is shown in Fig. 6. The sequence was fully overlapped and 98.6% of the sequence was determined on both strands. In one region (nucleotides 379 - 381) the sequence was ambiguous on both strands. Sequencing reactions were carried out at increased temperature (37°C) but no substantial improvement of the gel readings was obtained. The sequence we give is d(A-A- C-T-C,G,C-T-G-C-G) with the order of the C,G,C triplet uncertain. Otherwise there were no ambiguities in the se- quence. An open reading frame extended from position 56 to a stop codon at position 2606. The method of Stormo et al. [33] was used to search for the predicted translation start of pop1B. The triplet for AUG at positions 74 - 76 was identified as a highly significant translation start giving scores of +274, +83 and +I79 respectively using the W51, W71 and WlOl matrices of Stormo et al. [33]. The stop codon at nucleotides 2606 - 2608 is followed by an 1 1 -base-pair inverted repeat

440

AGATCTTAAATGCCATTGn;ATGATCTCCPTATCCPTATCACCCGTCACTCTGACGGGTATATCAATGCGTCTGGCPTGCCPTTATACTACCGCGCGTTTGTTTATAAACTGCCCAAATG~CTA 1 2 0 BglII

1 2 0 M K F V K Y F L I L A V C C I L L G A G S I Y G L Y R Y I E P Q L P D

AATGGGAAATTTCCAGTAAGT~GTAAAGTATTTTTTGATCCPTGCAGTCTGTn;CATTCTGCTGGGGAGCAGGCPCGATTTATGGC~ATACCGCPACATCGAGCCACAACTGCCGGAT 240

40 60 V A T L K D V R L Q I P M Q I Y S A D G E L I A Q Y G E K R R I P V T L D Q I P

G ~ ; G C G A C A T T A A A A G A T G T T C G C A A A T T C C G A T G C G G T A G A A A C G T C G T A T T C C G G T T A C G T T G G A T C A A A T C C A

8 0 100 P E M V K A F I A T E D S R F Y E H H G V D P V G I F R A A S V A L F S G H A S

CCGGAGATGGTAAAGCCPTATCGCGACAGAAGACAGCCGCTGTCTACGAGCATCACGGCGTTGACCCGG~;GGGATCTTCCGTGCAGCAAGCGTGCGCTGTTCTCCGGTC~TCA

Q G A S T I T Q Q L A R N F F L S P E R T L M R K I K E V F L A I R I E Q L L T CAAGGGGCAAGTACCATTACCCAGCAGCTGGCGAGAAACTTCTTCCTCAGTC~GRACGCACG~GA~CGTAAGA~AAGGRAGTCTTC~CGCGATTCGCATTGAACAGCTGCTGACG

MluI 120 140

160 1 an -"" K D E I L E L Y L N K I Y L G Y R A Y G V G A A A Q V Y F G K T V D Q L T L N E

AAAGACGAGATCCTCGAGCTTTATCTG~AAGATTTACCTTGGTTACCGCGCCTATGGn;TCGGn;CTGCGGCACAAGTCTATTTCGG~CGG~CCAACTGACGCTGAACGAA

M A V I A G L P K A P S T F N P L Y S M D R A V A R R N V V L S R M L D E G Y I ATGGCGGTATAGCCGGGCPGCCGAAAGCGCGC~TCCACCTTCAACCCGCTCTACTCGATGGATCGTCCGTCGCGCGGCGTAACG~GTCTGTCGCGGATGCPGGATGAAGGGTATATC

XhoI SalI 2 0 0 2 2 0

2 40 26 0 T P Q Q F D Q T R T E A I N A N Y H A P E I A F S A P Y L S E N V R Q E M Y N R

ACCCAACAACAGTTCGATCAGACACGCACTGAGGCGATTAACGCPAACTATCACGCGCCGGAGATTGCPTTCTCTGCGCCGTACCTGAGCGAAATGGTCGCCAGGAGATGTATAACCGT

2 8 0 300 Y G E S A Y E D G Y R I Y T T I T R K V Q Q A A Q Q A V R N N V L D Y D N R H G TATGGCGAAAG~;CCPATGAGACGGTTATCG~TTTACAC~CCATCACCCGCAAAGTCAGCAGGCCGCGCAGCAGGCGGTACGTRATAACGTGCTGGACTACGACATGCGCCACGGC

Y R G P A N V L W K V G E S A W D N N K I T D T L K A L P T Y G P L L P A A V T 3 2 0 340

TATCGCGGCCCGGCAAATGTCTGTGGAAAGTG~CGAGTATAACAACAAGATTACCGATACGCPGAAGGCG~GCCAACCTATGGTCCGCPG~GCCTGCCGCAGTCACC

S A N P Q Q A T A M L A D G S T V A L S M E G V R W A R P Y R S D T Q Q G P T P 360 3 8 0

AGCGCCAATCCTCAGCAAGCGACGGCGATGCPGGCGGACGGG~CG~GCATTGAGTATGGARGGCGTTCGCTGGGCGCG~CTTACCGTrrGGATACTCAGCAAGGACCGACGCCG

R K V T D V L Q T G Q Q I W V R Q V G D A W W L A Q V P E V N S A L V S I N P Q 4 0 0 4 2 0

CGTAAAGTGACCGATGTTCTGCAAACGGGTCAGCAAATCTGGGTTCGTCAGGTTGGCGATGCATGGTGGCPGGCACAAGTGCCGGRAGTAACTCGGCGCTGGTGTCGATCAATCCGCAA

N G A V M A L V G G F D F N Q S K F N R A T Q A L R Q V G S N I K P F L Y T A A 440 4 6 0

AACGGTCCGTTATGGCGCTGGTCGGTGCTTTGATTTCAATCAGAGCAAGTTTRACCGCGCCACC~GGCACTGCGTCAGG~GGT~CAACATCAAACCGT~CTCTACACCGCGGCG

M D K G L T L A S M L N D V P I S R M D A S A G S D W Q P K N S P P Q Y A G P I 4 80 5 0 0

ATGGATAAAGCGGTCTGACGCTGGCAAGTATGT~A~GATGn;CCAATTTCTCGCPGGGATGCAAGTGCCGGTTCTGACTGGCAGCCGRAG~CTCACCACCGCAGTATGCTGGTC~ATT

Sal I

540 5 2 0 R L R Q G L G Q S K N V V M V R A M R A M G V D Y A A E Y L Q R F G F P A Q N I

CGCPTACGTCAGGGGCPGGGTCAGTCG~CGTGTATGGTACGCGCAATGCGGGCGATGGGCG~CTACG~~GAATATCTGCAACG~TCGG~TCCCGG~CA~CATT

V H T E S L A L G S A S F T P M Q V A R G Y A V M A N G G F L V D P W F I S K I GTCCACACCGAATCGCTGGCGCTGGGTTCAGCGTCCTTCACCCCAATGCAGGn;GCGCGCGG~ACGCGGTCATGGCGAACGGCGGCPTCCPGGTGACCCGTGGT~ATCAGCA~TT

Sal I PSt I 5 8 0 560

620 600 E N D Q G G V I F E A K P K V A C P E C D I P V I Y G D T Q K S N V L E N N D V

GAAAACGATCAGGGCGGCGTATTTTCGAAGCGAAACCGAAAGTAGCCTGCCCGGAATGCGATATTCCGGTGATPTACGGTGATACGCAGAAATCGARCGTGCPGGAAAATAACGATGTT

660 640 E D V A I S R E Q Q N V S V P M P Q L E Q A N Q A L V A K T G A Q E Y A P H V I

GAAGATGTCGCTATCTCCCGCGAGCAGCAGRATGTTTCTGTACCAATGCCGCAGCTGGAGCAGGCAARTCAGGCGTTAGTGGCGAAGACTGGCGCGCAGGAGTACGCACCGCACGTCATC

7 0 0 680 N T P L A F L I K S A L N T N I F G E P G W Q G T G W R A G R D L Q R R D I G G

ARCACTCCGCTGGCATTCCTGATPAAGAGTGCTTTGAACACCAATRTCTTTTGGTGAGCCAGGCPGGCAGGGTACTGG~GGCGTGCAGG~GTGA~TGCAGCGTCGCG~CGGCGGG

K T G T T N S S K D A W F S G Y G P G V V T S V W I G F D D H R R N L G H T T A AAAACCGGGACCACTAACAGTTCGAAAGATGCGTGGTTCTCGGGTTACGG~CGGGCGTTGTGACCTCGG~TGGA~GGCPTTGATGATCACCGTCGTAATCTCGGTCATACAACGGCT

EcoRV 7 4 0 7 2 0

7 80 760 S G A I K D Q I S G Y E G G A K S A Q P A W D A Y M K A V L E G V P E Q P L T P TCCGGAGCGATTAAAGATCAGATCTCAGGTTACGAAGGCGGTGCCAAGAGTGCCCAGCCTGCATGGGACGCTTATATGAAAGCCGTTCTTGARGGTGTGCCGGAGCAGCCGCTGACGCCG

-1 -~ ~ ~

8 2 0 800 P P G I V T - V N I D R S T G Q L A N G G N S R E E Y F I E G T Q P T Q Q A V H E

CCACCGGGTATTGTGACGGTAATATCGATCGCAGCACCGGGCAGTTAGCTAATGGTGGCAACAGCCGCGAAGAGTATTTCATCGAAGGTACGCAGCCGACACAACAGGCAGTGCACGAG

360

4 80

600

7 2 0

8 4 0

9 5 0

1080

1200

1 3 2 0

1 4 4 0

i 560

1680

1 8 0 0

1 9 2 0

2040

2160

2280

2400

2520

2640

2760

GCGT

Fig. 3. The nucleotide sequence of the ponA gene and the predicted amino acidsequence of PBPIA. The asterisk marks the UGA stop codon triplet. The arrows show the position of an inverted repeat. The single-letter code for amino acids is used [49]. The d representing deoxy and the hyphens representing phosphodiester bonds have been omitted

MluI

44 1

WnB tonA ' pons tonA t t PBS74

PBS65 - t

pE67 - t

PSSW t -

0' - pm92

d t m91 - w96 c_- t -

Fig. 4. Location ofthe ponB gene. The top line shows the restriction map of the chromosomal insert in pBS74. The position, and the direction of transcription of the ponB and tonA genes are shown. The horizontal lines indicate the chromosomal regions carried by the plasmids that were used to locate the ponB gene. pBS92 produced truncated forms of PBP 1B. pBS91 produced only the small form of PBP 1B. Kb = x lo3 base pairs

P E W M13taq3 Sphl Sphl 1 2 r 2 . 4 1 Kb fragment

\ligation

Purify 354bp fragment EcoRV 41

ECORV sphl BamHl

-A EcoRV,@amHI

Fig. 5. Subcloning of the ponB gene. A SphI-BamHI fragment containing the end of the ponB gene, and the putative transcription terminator, was obtained from the replicative form of an M13mp8 clone (M13taq3), and was joined to the EcoRV-SphI fragment from pBS94, and was ligated into pPH126. RV, EcoRV; M, M l d ; Sa, SaA; Sm, SmaI and other abbreviations as in the legend to Fig. 1

centred on position 2637 which may function as the transcrip- tion terminator of the ponB gene.

Multiple translation starts in the ponB gene

If translation of PBP 1B was initiated at position 74, the removal of DNA sequences on the upstream side of the MluI site at positions 119- 124 should abolish the synthesis of the protein. A plasmid (pBS91) was constructed which lacked the smallest MluI fragment of pBS74 (Fig. 4). C600(pBS91) overproduced only the smaller component of PBP 1B (Fig. 2F). PBP 1B therefore appeared to have at least two translation starts: one of these was between the EcoRV and MZuI sites and produced the-large form of PBP IB and the other was located after the MluI site and gave rise to the small form of PBP I B. If this interpretation is correct, alteration of the reading frame at the MluI site at nucleotides 119- 124 should prevent the synthesis of the large form of PBP 1 B but should have no effect on the synthesis of the small form of the protein. TheponB gene of pBS94 was cut at the MluI site, the cohesive ends were converted to flush ends with T4-DNA polymerase, and the ends were rejoined with T4-DNA ligase. Sequencing of two of the resulting plasmids showed that the

MluI site d(A-C-G-C-G-T) had been converted in one case (pBS94C) to d(A-C-G-C-G-C-G-T) and in the other case (pBS94B) to d(A-C-G-C-C-G-C-G-T). Fig. 2 shows that pBS94C, which has an alteration of the reading frame by the insertion of 2 base pairs at the MluI site, produced only the small form of PBP 1B whereas pBS94B, which has maintained the reading frame by the insertion of 3 base pairs at the MluI site, produced both forms of PBP 1 B. The translation starts for the large and small forms of PBP 1B are therefore re- spectively upstream and downstream of the MluI site.

The most likely candidate for the second translation start is the AUG at position 209 - 21 1 of the mRNA. This position was not identified as a translation start by the method of Stormo et al. [33] but the AUG is preceded by a sequence G-A- G-G-A-U at positions 194- 199 that may act as a ribosome- binding site. Translation from the postulated major and minor translation initiation sites would produce proteins of 844 and 799 amino acids with M, of 94100 and 88800 respectively. These values are in good agreement with the value of approximately 90000 obtained from SDS/polyacrylamide gels 11, 31.

Analysis of the amino acid sequence of PBP 1A and PBP 1B

The overall similarity between the amino acid sequences of PBP 1 A and PBP 1 B was low but could be observed clearly (data not shown) using the computer program DIAGON [29]. The amino acid sequence of PBP 1A between residues 84- 758 could be aligned with residues 231 -739 of PBPIB (Fig. 7). No satisfactory alignment of the amino-terminal or the carboxy-terminal regions of these proteins could be obtained. Within the regions that could be aligned there was 31 YO identity between the PBP IA and PBP 1B sequences and two large gaps of 106 and 66 residues had to be introduced into the PBP 1B sequence at positions 468 and 668 to obtain the alignment. Several regions of the PBPs showed extensive similarity but in other regions the similarity was low and the alignment of these regions was tentative (Fig. 7).

D-Alanine carboxypeptidases and the class A [34] and class C p-lactamases [35] are serine enzymes in which penicillin (and cell wall peptides in the case of D-alanine carboxypeptidases) acylates an active-site serine residue that occurs close to the amino-terminus within the sequence Ser-Xaa-Xaa-Lys [21,36- 381. We therefore examined the sequence of PBP lA, PBP 1 B, and PBP 3 for the occurrence of the tetrapeptide Ser- Xaa-Xaa-Lys. The sequence occurred once in PBPIA (at position 465 - 468), twice in PBP 3, and three times in PBP 1 B. However, the sequence around the tetrapeptide Ser-Xaa-Xaa- Lys at position 307-310 in PBP3 and position 510-513 in PEP 1 B showed convincing similarity to the sequence around the unique Ser-Xaa-Xaa-Lys of PBP 1A (Fig. 8B). It is strik- ing that the sequence Gly-Ser-Xaa-Xaa-Lys-Pro occurred near the middle of each of the high-M, PBPs and it is very likely that the serine residue within this sequence is the active-site amino acid that is acylated by penicillin and cell wall peptide substrates.

The sequences of D-alanine carboxypeptidases and TEM P-lactamase show some similarities at their amino-termini [19, 21, 391. The maximum similarity between these proteins occurs about 9-25 amino acids to the amino-terminal side of the active site-serine; the same region also shows some similarity to residues 261 - 277 that occur to the amino-ter- minal side of the putative active-site serine of PBP 3 [I , 191. A similar region can also be found on the amino-terminal side of the putative active-site serine residue of PBP 1A and PBP 1B.

442

1 d A G N D R E P J G R K k K P T

GATATCTTCTTCTGTCTTGTAACAGAAG~AGAAAATCGGGCT~GCGCCTGAATATTGCGGAGARAAAGCATGGCCGGG~TGACCG~AGCCAATTGGACGCAAA~~CCG~C ECORV

R P V K Q K V S R R R Y E D D D D Y D D Y D D Y E D E E P M P R K G K G K G K G GCGTCCGGTCAAACAAAAGGTARGCCGrrGTCGTTACGAAGATGACGATGATfACGACGATfATGATGACTATGAGGATGAAG~CCGATGCCGCGCAAAGGTAAGGGCAAAGGCAAAGG MluI

R K P R G K R G W L W L L L K L A I V F A V L I A I Y G V Y L D Q K I R S R I D GCGTAAGCCTCGn;GCAAACGCGGCPGGCTATGGCTACTGCT~CTGG~ATCGT~TGCCGTGCGATCGCCAT~ACGGCGTTTATCTCGATC~TTCGTAGCCGTATfGA

40 20

80 60

100 120 G K V W Q L A A A V Y G R M V N L E P D M T I S K N E M V K L L E A T Q Y R Q V

n;GCAAGGrrTGGCAACTCGffG~GCAGTWATGGCCG~TGG~MTCTTGAGCCAGACATGACCATCAG~AG~GAGAT~TG~GCTGCTGGAGGCGACCCAGTATCGrrAGGT

140 1 6 0 S K M T R P G E F T V Q A N S I E M I R R P F D F P D S K E G Q V R A R L T F D

GTCGAAAATGACCCGTCCTGGCGAA~TACCGTGCAGGCCAACAGCATTGAGATGATPCGCCGTCCGT~GA~CCCGGACAGTAAAGAAGGACAGGTGCGCGCGCGTCTGACCTTTGA

180 200 G D H L A T I V N M E N N R Q F G F F R L D P R L I T M I S S P N G E Q R L F V

n;GCGATCATCTGGCGACGATCGTCAATATGGAGAACAACCGrrAGT?CGG~TTCCGTCTTGATCCGCGTCTGATCACCATGATCTCTTCGCCA~CGGTGAGCAGCGTCTGTTTGT

220 240 P R S G F P D L L V D T L L A T E D R H F Y E H D G I S L Y S I G R A V L A N L

GCCGCGCAGTGGmCCCGGAmGCTGGn;GATACTTTGCTGGCGACAGAAGACCGrrATTT~ACGAGCATGATGG~TCAGTCTCTACTCAATCGGACGTGCGGTGCTGGCAAACCT

7r.n 280 --- T A G R T V Q G A S T L T Q Q L V K N L F L S S E R S Y W R K A N E A Y M A L I

GACCGCCGGACGCACGGTACAGGGn;CGAGTACGCTGACGCAA~GTGARAAACCTGT?CCTCTCCAGCGAGCGTTCTTACTGGCGTAAAGCGAACG~ACATGGCG~GAT PVUII Hind111

I n n 320 _ _ _ M D A R Y S K D R I L E L Y M N E V Y L G Q S G D N E I R G F P L A S L Y Y F G

CATGGACGCGCGTTACAGCAAAGACCGTATfCTTGAG~GTATATGAACGAG~TGTATCTCGGrrAGAGCGGCGACAACGAAATCCGCGGCTTCCCGCTGGCA~TGTATTACTTTGG Hind1 I I

340 360 ~ ~~

R P V E E L S L D Q Q A L L V G M V K G A S I Y N P W R N P K L A L E R R N L V T C G C C C G G T A G A A G A G C T A A G C C T C G A C C A G C A G G C G C P G C G T C C A T C T A C A A C C C G T G G C G T A A C C C A A A A C T G G C G C P G G A G C G A C G T A A T C T G G T

380 400 L R L L Q Q Q Q I I D Q E L Y D M L S A R P L G V Q P R G G V I S P Q P A F M Q

GCTGCGTCTGCTGCAACAGCACAGATTATTGATCAAGA~TCTATGACATGTTGAGTGCCCGTCCGCTGGGGGTTCAGCCGCGCGGTG~GTGATCTCTCCTCAGCWLGCCTTATGCA

420 440 L V R Q E L Q A K L G D K V K D L S G V K X F T T F D S V A Q D A A E K A A V E

ACTGGn;CGTCAGGAGCT~GCA~CTGGGCGATAAGGTAAAA~TCTCCGGCGTGAA~TTCACTACCTTTGACTCGGTGGCCCAGGACGCGGCRGARAAAGCCGCCGTGGA PSt I BglII B g l I I

460 4 80 G I P A L K K Q R K L S D L E T A I V V V D R F S G E V R A M V G G S E P Q F A

AGGCAWCCGGCACTGAAGAAACACCGTAAGTTGAGCGATC~G~CTGCGATfG~G~G~CCGCTTTAGTGGTGAAGT~GTGCGATGGTCGGAGGTTCTGAGCCGCAGTTfGC Sal I

500 520 G Y N R A M Q A R R S I G S L A K P A T Y L T A L S Q P K I Y R L N T W I A D A

GGGCTACAACCGn;CGATGCAGGCGCGCGTCGT~GA~GGT~CCTTGCAAAACCAGCGACTTATCTGACGGCCTAAGCCAGCCG~TCTATCGTCTGRATACGTGGA~GCGGATGC

540 560 P I A L R Q P N G Q V W S P Q N D D R R Y S E S G R V M L V D A L T R S M N V P

GCCAATTGCGffGCGTCAGCCGAATGGCCAGGCCAGGrrTGGTCACCGCRGAATGATGACCGTCGTTATAGCG~GCGGCAGAGTGATGCGGTGGATGCGTTGACCCGTTCGATGAACGTGCC

580 600 T V N L G M A L G L P A V T E T W I K L G V P K D Q L H P V P A M L L G A L N L

GACGGTkAATCTGGGGATGGCGCPGGGGCPGCCTGCGGTTACGGAGACCTGGAW~CTGGGCGTACCGAAAGArrAGTTGCATCCGGTrrCGGWLATGCPGCGGGGGCGTTGAACTT

620 640 T P I E V A Q A F Q T I A S G G N R A P L S A L R S V I A E D G K V L Y Q S F P

AACGCCAATCGAAGTGGCGCGGCATTCCAGACCATCGCCAGCGGTGGT~CGTG~CCGCPTrrTGCGffGCGTTCGGT~TCGCGG~GATGGCAAAGTG~GTATCAGAGCTrrCC

660 680 Q A E R A V P A Q A A Y L T L W T M Q Q V V Q R G T G R Q L G A K Y P N L H L A

GCAGGCGGAACGCGCTGTTCCGGCGCAGGCGGCGGCGTATCTGACACTATGGACCATGCAGCAGGTGGTAC~CGCGGTACGGGTCGTCAGCPTGGGGCGAAATACCCGAACCCGAACCTGCATCTGGC

700 720 G K T G T T N N N V D T W F A G I D G S T V T I T W V G R D N N Q P T K L Y G A

AGGGAAAACAGGGACTACCAACAATAACGTAGATAC~GGTTTGCGGGCATTGACGGCAGCACGGTGACCATCACCTGGGrrGGCCG~ATAACAACCAGCCGACCARACTGTATGGTGC

740 760 S G A M S I Y Q R Y L A N Q T P T P L N L V P P E D I A D M G V D Y D G N F V C

CAGCGGGGCAATGrrGAmATCAGCGTTATCTGGCT~CCAGACGCCAACGCCGCPGAATCTTGT~CGCCAGRAGATATTGCAGATATGGGCGTGGACTACGACGGCRACTTTGT'G

780 80 0 S G G M R I L P V W T S D P Q S L C Q Q S E M Q Q Q P S G N P F D Q S S Q P Q Q

CAGCGGTGGCATGCGTATCTTGCCGGTCTGGACCAGCGATCCGCAATCG~G~C~G~GAGCGAGATGCAGCAGCAGC~TCAGGCAATCCGT~GATCAGTCTTCTCAGCCGWLGCA

820 840 844 Q P Q Q Q P A Q Q E Q K D S D G V A G W I K D M F G S N *

ACAGCCGCAACAGCAACCGCTCAG~AGAGCAGGCGAAAGACAGCGACGGTGTAGCCGGTTGGArrAAGGATATGTTfGGTAGT~TTAACATCTAAGCG~AAATACCGGATGG~GAGTT~

CCATCCGGTAAAATAACATCCCATCTRAGATATTAACCCTTTCTTTTCATCTGGTTGTTTATTAACC~~~~AGGRACGCPCAGATTGCGTACCGCPTGCGAACCCGCCAGCGTTTCGACG_

GATCC BamHI

120

240

360

480

600

7 20

840

960

1080

1200

1320

1440

1560

1680

1800

1920

2040

2160

2280

2400

2520

2643

2760

Fig. 6. The nucleotide sequence of the ponB gene and the predicted amino acid sequence of PBP 1B. The asterisk marks the UAA stop codon triplet and the arrows mark an inverted repeat. Nucleotides after the TuqI site at positions 2755-2758 were added in the subcloning of the ponB gene to produce pBS96. Abbreviations as in Fig. 3

443

PBPlA

PBPlB

PBPlA

PBPlB

PBPlA

PBPlB

PBPlA

PBPlB

PBPlA

PBPlB

PBPlA

PBPlB

1 9 5 80 KAFIATEDSRFYEHHGVDPVGIFRAASVALFSGHASQGASTITQQLARNFFLSPERTLMRKIKEVFLAIRIEQLLTKDEILELYLNKIYLGYRA----YGVGAAAQWFGKTVDOLTLNE

DTLLATEDRHFYEHDGISLYSIGRAVLANLTAGRTV~ASTLTQQLVKNLFLSSERSY~~RKANEAYMALI~ARYSKDRILELYMNEWLGQSGDNEIRGFPLASLYYFGRPVEELSLDQ 346 2 2 1

t* ***** t **t t * * * * tt*t * * * * * * t t *tt*+ *.** * * * * f ** t

314 196 MAVIAGLPKAPSTFNPLYS~lDRAVARRNWLSRMLDEGYITQQQFDQTRTEAINANYHAPEIAFSAPYLSEMVRQEMYNRYGESAYE-DGYRIYTTITRKVQQAAQQAVRNNVLDYDMRH

OALLVGMVKGASIYl?PWRNPKLALERRNLVLRLLQQQQIIDQELYDMLSARPLGVOPRGGVI-SPQPAFMQLVRQEL~AKLGDKVKDLSGVKIFTTFDSVAODAAEKAAVEGIPALKKQR 46 5 3 41

ttt t t * * t **** * t t* * * t * * * * * *

43 4 3 1 5 GYRGPANVLWKVGESAWDNNKITDTLKALPTYGPLLPAAVTSAMPQQATAMLADGSTVALSMEGVRMRRPYRSDTQ~PTPRKVTDVLQTGQQIWVRQVGDAI.IWLAQVPCVNSALVS1h.P

DLETAIVVVDR t *

KLS---------------------------------------------------------------------------------------------------------- 466 47 9

550 43 5 QMGAVMALVGGFDFNQSKFNRATQALRQVGSNIKPFLYTAA--MDKGLTLASMLNDVPISRWDASAGSD~~QPKNSPPQYAGPIR--LRQGLGQSKNVVMV~MRAMGVDYAAEYLQRFGF

FSGEVRAEIVGGSEPOFAGYNRAMQARRSIGSLAKPATYLTALSQPKIYRLNTWIADAPIAL-RQPNGQVWSPQNDDRRYSESGRVMLVDALTRSMNVPTVNLG~lALGLPAVTE?WIKLGV 598 480

5 5 1 670 PAQNIVHTESLALGSASFTPMQVARGYAVMANGGFLVDPWFISKIENDQGGVIFEAKPKVACPECDIPVIYGDTQKSNVLENNDVEDVAISREOQNVSVPMPQLEQANQALVAKT~AQEY t * * * * ** * * * * PKDQLHPVPAI~LLGALNLTPIEVAQAFOTIASGGN~PLSALRSVIAEDGKVLYQSFPQAERAVPAOAAY-------------------------------------------------- 59 9

6 7 1 760 A P H V I N T P L A F L I K S A L N T N I F G E P G W Q G T G W R A G R D L Q R I G F D D H R R N L G H T T A S G A I K

- - - -__-_________ LTLWTMQQWQRGTGRQLGAKYPNLHLAGKTGTTNNNVDTWFAGIDGSTVTITWVGRDNNQPTKL-YGASGAMS 6 6 9 7 4 1

* t t +** * * * * * t t* t t * * * * * * * * * * * * * t t

*** * * * * * * * * * * * tt t * * * *

Fig. 7. Alignment of the amino acid sequences of PBPIA and PBPlB. Residues 84-158 of PBP 1A could be aligned with residues 231 -739 of PBP 1B. No convincing alignment could be obtained outside these regions. Asterisks mark the positions where the same amino acid occurs in the two sequences. Gaps introduced into the sequences are shown as dashes

A) P B P 1A P B P 1E PEP 3

P B P 5 PRP 6 P R P 5

I E . c o l i r e s i d u e s 420-4421 I E . c o l i r e s i d u e s 4 7 3 - 4 8 7 1 (E. c o l i r e s i d u e s 261-2751

I E . c o l l r e s i d u e s 1 9 - 3 3 ) I E . c a l l r e s i d u e s 1 4 - 2 8 ] (B. s u b t i l l s r e s i d u e s 1 1 - 2 5 ]

TEM p - l a c t a m a s e l r e s l d u e s 2 0 - 3 4 1 G Y U E I.

B) PR P 1A PBP 1 R P R P 3

PEP 5 P B P 6 P R P 5

T E M p - l a c t a r n a s e ampC p - l a c t a m a s e

c) P E P 1A PBP 1 8 P E P 3

IE. c o l 1 r e s i d u e s 4 5 4 - 4 7 6 ) I E . c o l i r e s i d u e s 4 9 9 - 5 2 1 1 (E. c o l l r e s l d u e s 2 9 6 - 3 1 8 1

IE. c o l i r e s i d u e s 3 3 - 5 5 ) (E. c o l i r e s i d u e s 2 8 - 5 0 ) I R . subtills r e s i d u e s 2 5 - 4 7 )

(resldues 3 4 - 5 6 ) ( r e s i d u e s 5 0 - 7 2 )

I E . c o l i r e s i d u e s 5 6 3 - 5 9 0 1 I E . c o l i r e s i d u e s 6 1 1 - 6 3 8 1 I E . c o l i r e s i d u e s 4 1 8 - 4 4 5 1

M M T S Y V I G I M T S Y V V G M M T Y E L L L w

Fig. 8. Regions of amino acid sequence similarity in PBPs and /3-lactamases. Sequences of E. coli PBPs 3 and 5, B. subtilis PBP5, TEM /3-lactamase and ampC /?-lactamase are from [19 - 21, 341 and (351 respectively. The sequence of E. coli PBP 6 is unpublished data from this laboratory. The single-letter code for amino acids is used [49]. PBPs IA, 1B and 3 are high-Mr bifunctional PBPs, PBPs 5 and 6 are D-alanine carboxypeptidases, TEM and ampC /3-lactamases belong to class A and class C respectively. (A) The region on the amino-terminal side of the active-site serine residue. Identical, or structurally related, amino acids that occur in a t least four of the proteins are boxed. (B) The active- site serine region. The asterisks mark the position of the acylated serine residue. Residues that are conserved in high-Mr PBPs, D-alanine carboxypeptidases, or /3-lactamases are boxed. (C) The region of maximum similarity in comparisons of PBP IA, or PBP IB, with PBP 3. Residues that occur in a t least two of the proteins are boxed. Structurally related amino acids are defined as Asp and Glu (D, E); Ile and Val (I, V); Ser and Thr (S, T); Asn and Gln (N, Q); Arg and Lys (R, K)

Fig. 8A shows the alignment of this region in the high-M, of - 2 x lo-’ using a span length of 15 residues) but the PBPs, D-alanine carboxypeptidases, and TEM B-lactamase. occurrence of identical, or structurally related, amino acids at

The similarities between the active-site serine regions of several positions in these PBPs gave increased significance to the PBPs and the regions on the amino-terminal side of the the alignments [40]. The overall similarity between PBP 1A/ serine were not statistically significant (or were only PBP 1B and PBP 3 was, however, very low and no extensive marginally significant) in painvise comparisons of PBP 1A or regions of similarity could be observed using the program PBP 1B with PBP 3 or PBP 5 (double matching probabilities DIAGON [29]. The most statistically significant match

444

I r 1801 2001 300' 4001 5001 6001 7001 8001

RESIOUE POSITION

PBPlB i w

iBa1 200' 3001 4001 5001 60al 7001 8001 RESIDUE POSITION

Fig. 9. Hydropathy along the amino acid sequences of PBP 1A and PBPlB. The average hydropathy of segments of 19 residues are plotted at the mid-point of the segment. Regions above the central line are hydrophobic in character whereas those below the line are hydrophilic

(Fig. SC) in the comparison of PBP 3 with both PBP 1A and PBP 1 B occurred between residues 41 8 - 445 of PBP 3 and residues 563 - 590 and 61 1 - 638 respectively in PBP 1A and PBP 1B (double matching probabilities of 1.3 x and 2.7 x

Comparisons of the amino acid sequences of PBP 1A and PBP 1B with PBP 5 of E. coli, and with class A and class C p- lactamases, showed no extensive regions of similarity.

Fig. 9 shows the average hydropathy [32] along the sequences of PBPIA and the large form of PBPlB. Both of the proteins were rather hydrophilic. The amino-terminus of the large form of PBP 1B was extremely hydrophilic (see Discussion): Apart from the regions close to the amino-termini, which may act as signal peptides for the insertion of the proteins into the cytoplasmic membrane, there were no ex- tensive hydrophobic stretches in either PBP 1A or PBP IB. The polarity indices [41] of the large and small forms of PBPlB were 45% and 44% respectively and the polarity index of PBP 1A was 43%. Although it seems likely that a substantial proportion of PBP LA and PBP 1B must be trans- located across the cytoplasmic membrane to catalyse the in- corporaton of peptidoglycan precursors into the cell wall, there is no information on the mode of insertion of these PBPs in the cytoplasmic membrane.

using a span length of 15 residues).

DISCUSSION

The nucleotide sequences of the ponA and p o d genes have allowed us to predict the amino acid sequences of PBPs

3 A and 1 B, the major penicillin-sensitive peptidoglycan trans- glycosylase-transpeptidases of Escherichia coli.

PBP 3, PBP 5 and PBP 6 of E. coli have been shown to be synthesised as pre-proteins with amino-terminal signal peptides [19, 421. The amino-terminus of PBPIA has the characteristics of a signal peptide but the size of the protein synthesised in vitro and in vivo appeared to be identical (Pratt, J. M., and B. G. S., unpublished experiments). Although this suggests that the amino-terminus of PBPIA is a non-pro- cessed signal peptide, the large size of the protein may have prevented the resolution of pre-protein and processed forms of PBP IA on SDS/polyacrylamide gels.

We have shown here that the multiple components of PBP 1B arise because of the use of alternative translation initiation codons, although additional post-translational modifications may also occur. The large form of PBP 1B must be translated from an initiation codon upstream of the MZuI site at nucleotides 119- 124. Initiation from the translation start predicted by the method of Stormo et al. [33] would result in a highly hydrophilic sequence at the amino-terminus of the large form of PBP 1B consisting of three alternating blocks of basic and acidic amino acids. The acidic region comprises a remarkable stretch of 17 amino acids consisting entirely of glutamate and aspartate residues interspersed with four tyrosine residues. The location of the translation start of the smaller form of PBP 1B was not clearly predicted by the method of Stormo et al. [33] but the most likely position at nucleotides 209 - 21 1 would result in the start of the smaller form at Met-46 in the sequence of the large form. The PBP 1B

445

sequence of residues 48 - 64 contains a run of basic amino acids interspersed with glycine and proline residues; this is followed by a stretch of hydrophobic amino acids (residues 65 - 87) broken only by a lysine at position 71. This region has some of the features of a signal peptide but the identical mobility of the large and small forms of PBPlB that are synthesised in vitro and in vivo (Pratt, J. M. and B. G. S., unpublished experiments) suggests that processing of the amino-terminus does not occur.

Several examples are known of genes that have an internal in-frame translation start and, in general, the resulting pairs of gene products perform related, but subtly different, physio- logical functions [43]. In the case of the cheA gene of E. coli, the smaller translation product is found in the cytoplasm whereas a proportion of the larger form is in the cytoplasmic membrane [44]. The two forms of PBP 1B appear to catalyse identical enzymatic reactions [15] but they may differ in their mode of membrane attachment (Hedge, P. J., unpublished).

PBP 1A and PBP 1B show substantial similarities in their amino acid sequences and the two proteins are clearly homolo- gous. They are not however closely related and major re- arrangements appear to have occurred during their diver- gence. Neither PBP IA nor PBP 1B showed any substantial similarity to PBP3 and the high-MI PBPs of E. coli are therefore not closely related in sequence. Similarly, the amino acid sequences of low-MI weight PBPs (D-alanine carboxypeptidases) and class A and class C 8-lactamases showed no extensive similarities to PBPs 1A or 1B or 3.

Enzymes that recognise the same inhibitor, or substrate, often show little amino acid sequence similarity except in short regions of the polypeptides that form the inhibitor/ substrate binding site in the three-dimensional structure 145, 461. Two possible regions of similarity were found amongst the sequences of PBPlA, PBPlB, PBP3, PBP5, PBP6 and TEM p-lactamase. One of these regions included the active-site serine residue (or the predicted active-site serine residue) and was characterised by the sequence Ser-Xaa-Xaa- Lys that has so far been found at the acitve site of all PBPs and class A and class C p-lactamases, with the apparent excep- tion of the D-alanine carboxypeptidase from Streptomyces strain R39 [47]. The second region occurs at a variable distance of 17 - 38 residues to the amino-terminal side of the active- site serine and this latter region may also constitute part of the binding site for penicillin. Apart from these two regions we could find no other sequences that were similar in all of the above proteins. Although the PBPs of E. coli show very little similarity in their amino acid sequence there may be similarities in their three-dimensional structures, but at pre- sent there is insufficient evidence to decide whether the low- M , PBPs, the high-MI PBPs, and the class A and class C p-lactamases will share common structural features.

The occurrence of the sequence Gly-Ser-Xaa-Xaa-Lys-Pro in PBP lA, PBP 1 B, and PBP 3 provides a strong candidate for the active-site serine in each of these high-MI PBPs. The assignment of Ser-465, Ser-510, and Ser-307 as the acylated residues of PBP IA, PBP lB, and PBP 3 has now been con- firmed directly by the isolation and characterisation of the peptides labelled with radioactive benzylpenicillin (Keck, W., Glauner, B., Schwarz, U., J. K. B.-S. and B. G. S., in the press).

The high-M, PBPs are therefore ‘serine enzymes’. The location of the active-site serine residue towards the middle of the bifunctional PBPslA, IB, and 3 contrasts with the position of the serine residue close to the amino-terminus in the D-alanine carboxypeptidases and p-lactamases. We have

previously suggested that the penicillin-sensitive trans- peptidase domain of the high-Mr PBPs is located towards the carboxy-termini of these proteins and that the penicillin- insensitive peptidoglycan transglycosylase activity is located at the amino-terminus [l, 191. This view is consistent with the finding that the amino-terminal 240 amino acids of PBP 3 can be removed without loss of the penicillin-binding activity of the protein [48]. The beginning of the region of similarity that precedes the active-site serine residue of the high-MI PBPs would be a likely location for the start of the penicillin- sensitive transpeptidase domain. The alignment betwen PBP 1 A and PBP 1B breaks down at this point and the dele- tion of 106 residues that appears to have occurred in the PBP 1B sequence at around residue 470 may have been tolerated because it is located at the junction between the transglycosylase and transpeptidase domains.

This work was supported by project grant G8106393CB from the Medical Research Council. S. Y. was supported by a scholarship from the University of Basra. We are most grateful to Dr P. J. Cassidy of Merck Sharp and Dohme for the gift of [3H]benzylpenicillin and Ms Lynda Robson for technical assistance.

REFERENCES 1. Spratt, B. G. (1983) J . Gen. Microbiol. 129, 1247-1260. 2. Waxman, D. J. & Strominger, J. L. (1983) Annu. Rev. Biochem.

3. Spratt, B. G. (1977) Eur. J. Biochem. 72.341 -352. 4. Iwaya, M. & Strominger, J. L. (1977) Proc. Natl Acad. Sci. USA

5. Matsuhashi, M., Takagaki, Y., Maruyama, I. N., Tamaki, S., Nishimura, Y., Suzuki, H., Ogino, U. & Hirota, Y. (1977) Proc. Natl Acad. Sci. USA 74,2976 - 2979.

52, 825 - 870.

74,2980 - 2984.

6. Spratt, B. G. (1980)J. Bacteriol. 144, 1190-1192. 7. Broome-Smith, J. K. & Spratt, B. G. (1982) J . Bacteriol. 152,

8. Spratt, B. G. (1975) Proc. Nut/ Acad. Sci. USA 72, 2999-3003. 9. Spratt, B. G. (1977) J . Bacteriol. 131,293 -305.

10. Spratt, B. G., Jobanputra, V. & Schwarz, U. (1977) FEES Lett.

11. Suzuki, H., Nishimura, Y. & Hirota, Y. (1978) Proc. Natl Acad. Sci. USA 75, 664-668.

12. Tamaki, S., Nakajima, S. & Matsuhashi, M. (1977) Proc. Natl Acad. Sci. USA 74, 5472- 5476.

13. Nakagawa, J . 4 , Tamaki, S. & Matsuhashi, M. (1979) Agric. Biol. Chem. 43, 1379- 1380.

14. Suzuki, H., Van Heijenoort, Y., Tamura, T., Mizoguchi, J., Hirota, Y. & Van Heijenoort, J. (1980) FEES Lett. 110,245- 249.

15. Nakagawa, J.-I. & Matsuhashi, M. (1982) Biochem. Biophys. Res. Commun. 105,1546-1553.

16. Ishino, F., Mitsui, S., Tamaki, S. & Matsuhashi, M. (1980) Bio- chem. Biophys. Res. Commun. 97,287 - 293.

17. Ishino, F. & Matsuhashi, M. (1981) Biochem. Biophys. Res. Commun. IOI,905-911.

18. Ishino, F., Tamaki, S., Spratt, B. G. & Matsuhashi,, M. (1982) Biochem. Biophys. Res. Commun. 109,689 - 696.

19. Broome-Smith, J . K., Edelman, A. & Spratt, B. G. (1983) in The target of penicillin (Holtje, J.-V., Hackenbeck, R. & Labi- schinski, H., eds), pp. 403-408, Walter de Gruyter, Berlin.

20. Nakamura, M., Maruyama, I. N., Soma, M., Kato, J.-I., Suzuki, H. & Hirota, Y. (1983) Mol. Gen. Genet. 191, 1-9.

21. Waxman, D. J. & Strominger, J. L. (1980) J. Biol. Chem. 255,

22. Messing, J. & Vieira, J. (1982) Gene 19, 269-276. 23. Stoker, N. G., Fairweather, N. F. & Spratt, B. G. (1982) Gene 18,

24. Clarke, L. & Carbon, J. (1976) Cell 9, 91 -99.

904 - 906.

79,374- 378.

3964 - 3976.

335 - 341.

446

25. Takeda, Y., Nishimura, A,, Nishimura, Y., Yamada, M., Yasuda,

26. Deininger, P. L. (1983) Anal. Biochem. 129,216-223. 27. Biggin, M. D., Gibson, T. J. & Hong, G. F. (1983) Proc. Natl

28. Staden, R. (1980) Nucleic Acids Res. 8 , 3673-3694. 29. Staden, R. (1982) Nucleic Acids Res. 10,2951 -2961. 30. Dayhoff, M. 0. (1978) in Atlas of protein sequence and structure,

vol. 5, pp. 356-358, National Biomedical Research Founda- tion, Washington, DC.

S., Suzuki, H. & Hirota, Y . (1981) Plasmid6, 86-98.

Acad. Sci. USA 80, 3963 - 3965.

31. McLachan, A. D. (1971) J . Mol. Biol. 61,409-424. 32. Kyte, J. & Doolittle, R. F. (1982) J . Mol. Biol. 157, 105-132. 33. Stormo, G. D., Schneider, T. D., Gold, L. & Ehrenfeucht, A.

34. Ambler, R. P. (1980) Phil. Trans. R . SOC. Lond. Ser. B289,321-

35. Jaurin, B. & Grundstrom, T. (1981) Proc. Natl Acad. Sci. USA

36. Frere, J.-M., Duez, C., Ghuysen, J.-M. & Vanderkerckhove, J.

37. Knott-Hunziker, V., Waley, S. G., Orlek, B. S. & Sammes, P. G.

(1982) Nucleic Acids Res. 10, 2997-3011.

331.

78,4897-4901.

(1976) FEBS Lett. 70,257-260.

(1979) FEBS Lett. 99, 59-61.

38. Knott-Hunziker, V., Petursson, S., Jayatilake, G. S., Waley, S. G., Jaurin, B. & Grundstrom, T. (1982) Biochem. J . 201,621 - 627.

39. Waxman, D. J., Amanuma, H. & Strominger, J. L. (1982) FEBS Lett. 139. 159- 163.

40. Doolittle, R. F. (1981) Science (Wash. DC) 214, 349-158. 41. Capaldi, R. A. & Vanderkooi, G. (1972) Proc. Nut1 Acud. Sci.

42. Pratt, J. M., Holland, I. B. & Spratt, B. G. (1981)Nature (Lond.)

43. Normark, S . (1983) Annu. Rev. Genet. 17, 499-525. 44. Smith, R. A. &Parkinson, J. S. (1980) Proc. Nut1 Acad. Sci. USA

45. Walker, J. E., Saraste, M., Runswick, M. J. & Gay, N. J. (1982)

46. Hempel, J., von Bahr-Lindstrom, H. & Jornvall, H. (1984) Eur.

47. Duez, C., Joris, B., Frere, J.-M., Ghuysen, J.-M. & Van Beeumen,

48. Hedge, P. J. & Spratt, B. G. (1984) FEBS Lett. 176, 179- 184. 49. IUPAC-IUC Commission on Biochemical Nomenclature 1984

USA 69,930-932.

293,307-309.

77, 5370 - 5374.

EMBO J . I , 945-951.

J. Biochem. 141,21-35.

J. (1981) Biochem. J . 193,83-86.

Eur. J . Biochem. 138,9 - 37.

J. K. Broome-Smith, A. Edelman, S. Yousif and B. G. Spratt, Microbial Genetics Group, School of Biological Sciences, University of Sussex, Biology Building, Farmer, Brighton, Sussex, England BN1 9QG