Eur. J. Biochem. 207,943-949 (1992)
0FEBS 1992
Penicillin-binding protein 2x of Streptococcus pneumoniae Expression in Esckerichiu coli and purification of a soluble enzymatically active derivative Gotz LAIBLE’, Wolfgang KECK’, Rudi LURZ’, Harald MOTTL’, Jean-Marie FRERE3, Marc JAMIN3 and Regine HAKENBECK’ Max-Planck Institut fur molekulare Genetik, Berlin, Federal Republic of Germany University of Groningen, BIOSON Research Institute, Department of Biochemistry, The Netherlands Universitk de Liige, Centre d’IngCnierie des Proteines, Laboratoire d’Enzymologie, Institut de Chimie B6, Belgium (Received February 4/May 25, 1992) - EJB 920149
A 2.5-kb DNA fragment including the structural gene coding for the penicillin-binding protein 2x (PBP 2x) of Streptococcus pneumoniue has been cloned into the vector pJDC9 and expressed in Escherichia coli. Mapping of RNA polymerase binding sites by electron microscopy indicated that the pbpX promoter is well recognized by the E . coli enzyme. However, high-level expression occurred mainly under the control of the lac promoter upstream of the pJDC9 multiple cloning site. After induction with isopropyl P-d-thiogalactopyranoside, PBP 2x was expressed as one of the major cellular proteins. PBP 2x produced in E. coli corresponded to the pneumococcal PBP 2x in terms of electrophoretic mobility, fractionation with the cytoplasmic membrane, and penicillin-binding capacity. Deletion of 30 hydrophobic N-terminal amino acid residues at positions 19 -48 resulted in high-level expression of a cytoplasmic, soluble PBP 2x derivative (PBP 2x*) which still retained full p-lactam-binding activity. A two-step procedure involving dye affinity chromatography was established for obtaining large amounts of highly purified enzymatically active PBP 2x*.
Penicillin-binding proteins (PBPs), the target enzymes of p-lactam antibiotics, and p-lactamases constitute the family of penicillin-recognizing active-site serine enzymes (Ghuysen, 1991). Comparison of the deduced amino acid sequences of PBPs from different bacterial species and functional analysis mainly performed with Escherichia coli PBPs revealed three classes of PBPs. Low-MI PBPs contain DD-carboxypeptidase activity and appear to be dispensable for cellular growth. High-M, PBPs acting in essential late steps of murein biosynthesis are either homologous to the bifunctional E. coli PBPs l a and l b (Ghuysen, 1991; and unpublished results) or are related to E. coli PBP 2 and 3 , like Streptococcus pneumoniue PBPs 2x (Laible et al., 1989) and 2b (Dowson et al., 1989). E. coli PBP 2 and 3 are important components of the morphogenetic apparatus; the in vivo function of related PBPs from other bacterial species remains largely unknown. High-MI PBPs are mainly periplasmic enzymes that are bound to the membrane via a short N-terminal hydrophobic peptide. In addition to the penicillin-binding domain that includes the active center responsible for penicillin-sensitive transpeptidase activity, they usually carry an N-terminal domain and a long C-terminal extension. In E. coli PBPs l a and 1b, penicillin-insensitive transglycosylase activity is located on the N-terminal domain (Ishino et al., 1980; Suzuki et al., 1980; Tamaki et al., 1980). Correspondence to R. Hakenbeck, Max-Planck Institut fur molekulare Genetik, Ihnestr. 73, W-1000 Berlin 33, Germany Fax: +49 30 830 7385. Abbreviations. PBP, penicillin-binding protein; S2d, b e n z o y h alanyl thioacetyl ester.
PBPs are inhibited by p-lactam antibiotics by enzymatically forming a covalent penicilloyl (or cephalosporoyl) complex via the active-site serine residue which is part of the conserved SXXK sequence. Further conserved boxes in PBPs are the SXN sequence and the triad KTG or KSG that are part of the active-site cavity as deduced from the known three-dimensional structures of several p-lactamases and the Streptomyces R61 low-MI PBP (Dideberg et al., 1987; Herzberg and Moult, 1987; Moews et al., 1990; Oefner et al., 1990). Although not related in amino acid sequence, the polypeptide folding of the R61 PBP and the class A plactamases is amazingly similar with the conserved boxes representing critical components of the active-site cavity. It is assumed that the principal polypeptide arrangement of the R61 enzyme has been conserved in the penicillin-binding domain of high-Mr PBPs. N o experimental data on the threedimensional structure of high-MI PBPs is available yet, and nothing is known about the structure of their N-terminal domain nor their C-terminal extension. The importance of high-MI PBPs in S. pneumoniue was deduced from their role in intrinsic penicillin resistance. In pneumococci, penicillin resistance is mediated by the production of low-penicillin-affinity variants of four high-M, PBPs (Laible et al., 1991). PBP 2x appears to be the major target for p-lactams since it is the first PBP altered in cefotaxime-resistant laboratory mutants (Laible and Hakenbeck, 1987), and since transformation with DNA fragments encoding low-affinity variants of PBP 2x from clinical isolates or from laboratory mutants yield low-level-resistant transformants (Laible et al., 1989, 1990).
944 Dral
Haelll
I
+
c
pCGl9
pCG8 I
Pstl Haelll EwRV
Hind111
1’
+c
EwRl
100 bp
I
2‘
I
I
(Hindlll)
pCG28
I
I
I (EmRI)
I
pCGl0
I
Pstl
pCG29 (PBP 2x)
I
,
I
I
.
, I. ,, .) ,
pCG31 (PBP 2@)
I
iI :i
Pdl
I
I
,
I
Xmal
Fig. 1. Cloning of the PBP 2x and PBP 2x* gene. A map of the S. pneumoniae pbpX gene and its flanking regions is given at the top. The hashed segment marks the structural encoding gene pbpX with the arrow indicating the promoter. The cloned DNA fragments are indicated below. Plasmid pCG19 was used for mapping of E. coli RNA polymerase binding sites in the promoter region. pCG19 and pCG28 are derivatives of pCG9 (pCG9 contains the same insert as pCG8 but in the rcverse orientation). pCG29 containing the structural PBP 2x gene and flanking regions was obtained from pCG28 and pCG10. pCG31 encoding the soluble PBP 2x* was constructed using two DNA fragments obtained from pCG19 b y the polymerase chain reaction (the positions of the primers are indicated. 1, primer 365; l’, 2020; 2, 2021 ; 2’, 623), and the PstI - EcoRI fragment of pCG29 as described in the text.
Point mutations in the PBP 2x gene responsible for reduction of penicillin affinity from several laboratory mutants have been mapped (Laible and Hakenbeck, 1991). A prerequisite for studying the impact of the mutations in more detail and for ultimately resolving the three-dimensional polypypetide arrangement of wild-type and mutant proteins is the availability of large amounts of soluble PBP 2x. Like other highM , PBPs, PBP 2x contains a short N-terminal membraneanchoring domain. This hydrophobic peptide is believed not to be important for enzymatic function, since deletion of this region in E. coli PBP 1b, 2 or 3 had no effect on penicillinbinding activity (Adachi et al., 1988; Spratt et al., 1988). The construction and high-level expression of a soluble PBP 2x derivative (PBP 2x*) in E. coli that has retained full penicillinbinding activity will be described. Furthermore, a protocol was established enabling purification of enzymatically active PBP 2x*. MATERIALS AND METHODS Bacterial strains and plasmids
StreptococcuspneumoniueR6 is an unencapsulated derivative of the Rockefeller University strain R36A (Avery et al., 1944). Pneumococci were grown in caseine hydrolysate broth (Morrison et al., 1983). Plasmids were propagated in E. coli DH5, DH5cr (Hanahan, 1985) or JM103 (Messing et al., 1981) as stated in the text. Cells were grown in Luria-broth; recombinant clones were grown in the presence of 0.5 mg/ml erythromycin. For induction of the lac promoter, isopropyl /3-d-thiogalactopyranosidewas added to exponentially growing cultures at a final concentration of l mM. Plasmids used in this study are described in Fig. 1. Most of the plasmids harbouring different fragments ofpbpX(Laib1e et
al., 1989) are derivatives of the E. coli vector pJDC9 constructed for cloning pneumococcal DNA (Chen and Morrison, 1988). pCG19 contains the Hue111 fragment of pCG9 (Laible et al., 1989); pCG9 carries a Hind11 -EcoRV fragment in the reverse orientation as pCG8 which has been described (Laible et al., 1989); pCG28 is a derivative of plasmid pCG9 from which the 5‘ DraI -Hind11 fragment had been deleted including parts of the multiple cloning site of the vector pJDC9 by Hind11 restriction. pCGlO is a derivative of plasmid pR28 (Mkjean et al., 1981) containing the 1.8-kb PstI-EcoRI fragment of the R6 pbpX. Construction of recombinant plasmids
pCG29 contains a 2.5-kb Hind111 - EcoRI fragment covering the structural gene of PBP 2x and the promoter region. It was constructed by replacing the 3’ PstI - EcoRI fragment of pCG28 with the 1.8-kb PstI-EcoRI fragment of pCGl0 that covers the 3’-region of pbpX including the terminator (Fig. 1). pCG31 encoding the soluble PBP 2x* was constructed by introducing a site-directed deletion into pbpX by polymerase chain reaction technology. Two DNA fragments were amplified from plasmid pCG19. (a) The 5’-DNA fragment was obtained after amplification with the polymerase chain reaction of pCG19 with the two primers, 365 (5’-CTCAATTGGACCAGCG683) and 2020 (5’-TCCCCGGGCGATTTCCGATTTTTGGTCGCA282), the latter containing an endogeneous XmuI site. Restriction of the 1010-bp fragment with Hind11 and XmuI resulted in a 301-bp fragment covering the 5’-flanking region of pbpX and sequences encoding the first 18 amino acids. (b) A 395-bp DNA fragment was amplified from pCG19 with the two primers 2021 (5’-TTCCCGGGACAGGCACTCGCTTTGGAACAG421) carrying an
945 endogenous XmaI site, and 623 (5'-AATCCCCTTGACCTCTGCAG767) ; after restriction with XmaI/PstI, a 375-p fragment encoding a peptide starting with Xaa49 was obtained. Ligation of both fragments via the XmaI site results in deletion of amino acids 19 -48 representing the hydrophobic N-termind domain of PBP 2x. Both DNA fragments, as well as the 8.7-kb Hind11-PstI fragment of pCG29 were isolated from agarose gels and ligated. After transformation into E. coli DHSfa, plasmid pCG31 was obtained.
t
1-
l
Localization of RNA polymerase binding sites Binding of E. coli RNA polymerase (Promega) to pCG19 was performed according to Lurz et al. (1987). A total of 264 DNA molecules were analyzed, using three different restriction endonucleases with single restriction sites each for orientation of bound RNA polymerase: EcoRI, CluI and BumHI. Preparation of the DNA-protein complexes for electron microscopy and evaluation of micrographs were performed as described elsewhere (Perez-Martin et al., 1989). Manipulation of DNA Preparation of DNA and transformation were carried out according to standard protocols (Sambrook et al., 1989). Isolation of DNA fragments from agarose gels was performed as described (Vogelstein and Gillespie, 1979). Restriction endonucleases were obtained from Boehringer (Mannheim, FRG) if not otherwise stated and used according to the manufacturer's specifications. DNA sequencing was carried out using a T7 polymerase sequencing kit (Pharmacia) with a set of oligonucleotides that primed from intervals along the inserted DNA fragment of plasmid pCG31. Amplification of DNA by polymerase chain reaction was performed during 30 cycles consisting of 2.5 min extension at 7 7 T , 1.5-min denaturation at 95 "C, and 1.5-min annealing at 55 "C using a Pharmacia LKB Gene ATAQ Controller. The 50-p1 reaction mixture contained 0.3 ng plasmid DNA, 1 pM of each primer, and 1 U Tuq polymerase (Stratagene). The amplified fragments were purified from agarose gels before further manipulations. Cell fractionation A 40-ml culture of exponentially growing E. coli ( A 5 6 0= 1) was rapidly chilled on ice and cells were harvested by centrifugation. Cells were resuspended in 5 ml ice-cold 20 mM sodium phosphate pH 1.2 and disintegrated by sonification (M. S. E. ultrasonic disintegrator) for four 10-s intervals. Alternatively, 50 mM Tris/HCl pH 8, 100 mM NaCl, 1 mM EDTA was used during the procedure and cells disrupted in a French pressure cell (Amicon) at 137.8 MPa. Remaining intact cells were removed by low-speed centrifugation. The membrane fraction was recovered after ultracentrifugation (4 h, 150000 g). The supernatant after ultracentrifugation served as soluble fraction. No differences between the two methods in terms of localization of PBP 2x and its derivative were detected. Membranes were suspended in buffer with the addition of 0.2% Triton X-100, and stored at -80°C prior to detection of PBPs on SDS gels. Purification of PBP 2x" A 4-1 culture of E. coli DH5a harbouring pCG31 was grown to an A560 = 1.3. Cells were harvested by centrifuga-
EcoRl Fig. 2. RNA polymerase binding sites in pCGl9. The graphic representation summarizes results obtained from three independent data sets derived from linearized pCG19 using three different restriction endonucleases. RNA polymerase binding sites were determined by electron microscopy as described in the text. The map is aligned to the EcoRI site in the multiple cloning site of pJDC9. Positions of the inserted Hue111 pbpX fragment (1.73 kb) and the vector portion (7.2 kb) are indicated. The position of peak 1 corresponds to the position of the proposedpbpXpromoter (Laible et al., 1989). The arrow shows the direction of transcription.
tion, washed once in 50 mM Tris/HCl pH 8.0,lOO mM NaC1, and broken in a French pressure cell press after resuspension in 25 ml of the same buffer. Cell wall material was removed at 3 50000 g (1-h centrifugation); two 0.5-ml portions of the supernatant (28.5 mg) were applied to Procion blueherd mix 3967 coupled to Fractogel TSK HW-65 (F) (Merck, Darmstadt, FRG) dye affinity column (6 ml; 1 cm diameter) at a flow rate of approximately 0.15 ml/min connected to an FPLC system (Pharmacia, Uppsala). Coupling of the dye to Fractogel was performed essentially as described (Mottl and Keck, 1991). The column was washed with 24 ml10 mM Tris/ HCl pH 8.0, 100 mM NaC1, and PBP 2x* was eluted with a linear gradient (10 ml 100 mM NaCl and 1 M NaCl buffer each) followed by 16 ml 1 M NaCl buffer; 2-ml fractions were collected. Peak fractions (8 ml final volume) were concentrated by centrifugation through Centricon-30 microconcentrators (Amicon) with buffer exchange to 10 mM Tris/HCl pH 8; 1.6 mg protein was recovered. Final purification was obtained with a Mono Q 5/5 column equilibrated with 10 mM Tris/HCl pH 8; a maximum of 500 pg protein was applied. PBP 2x* was eluted with a stepwise gradient (20 ml 10 mM Tris/HCl pH 8, 0.4 M NaCl); 500-p1 fractions were collected at a flow rate of 60 ml/h. PBP 2x* eluted at approximately 250 mM NaCl; 560 pg PBP 2x* was obtained in 5.6 ml. Detection of PBPs PBPs were routinely labeled with [3H]propionylampicillin and visualized on fluorograms as described (Laible et al., 1987). For quantification of bound penicillin, 269 pmol [3H]benzylpenicillin(NEN Dupont, Dreiech, FRG; 28.2 Ci/ mmol) was incubated with PBP 2x* (5.38 pmol) in 20 p1 for 30 min at 37°C. After SDSjPAGE and fluorography, dried gel slices were incubated with 250 ~ 1 3 0 % H 2 0 2for 2 h at 50 "C and radioactivity measured after addition of 5 ml Instagel (Packard Instruments, Illinois) in a Beckmann LS 9800 scintillation counter. Counting efficiency was determined by spotting different amounts of [3H]benzylpenicillinon gel slices and measuring radioactivity as above.
946
6
A
1’ 2’ 3’
1 2 3
1
2
3
4
1’2’3’4‘
Fig.5. Expression of PBP 2x” from pCG31. The ligation mixture containing pCG31 was transformed into E. coli DH5u. Several independent transformants were analyzed, two examples being shown in lanes 3 and 4. For comparison, E. coli DH5 harbouring pCG29 (lane 1) and the vector pJDC9 (lane 2) are included. Total cellular protein (A) and PBPs (B) were visualized as described in the legend to Fig. 4.
Fig. 3. Expression of pbpX in E. cob. Cell lysates were incubated with [3H]propionylampicillin, proteins separated by SDSjPAGE and PBPs visualized after fluorography. Lanes 1 -3, protein pattern after Coomassie blue stain; lanes 1’-3’, fluorogram of 1-3 (5-days exposure). Lanes l,l’, E. coli DH5 harbouring pJDC9; lanes 2,2’, E. coli DH5 harbouring pCG29; lanes 3,3’, S. pneumoniue R6. The position of the R6 PBP 2x is indicated in the fluorogram; the arrow marks the additional protein band in DH5/pCG29.
A
1
2
3
4
5
6
7
6
Fig. 4. Induction of PBP 2x in E. coli. To an exponentially growing E. coli JM103/pCG29 culture, isopropyl j-d-thiogalactopyranoside was added at time zero (lane 1) and samples taken after 2 h (lane 3), 4 h (lane 5) and 6 h (lane 7). The control culture remained without isopropyl j-d-thiogalactopyranoside (lanes 2, 4 and 6). Cell lysates were prepared and PBPs labeled with [3H]propionylampicillin. (A) Total cellular protein after SDSjPAGE and Coomassie blue stain; (B) fluorogram of A after a 6-h exposure.
Enzymatic activity of PBP 2x* Activity of the purified PBP 2x* was determined using the compound S2d as substrate. S2d, a thioester compound (C6H,-CO-D-Ala-S-CH2-COOH),has been described (Adam et al., 1990). Spectrophotometric measurements were performed with a Perkin-Elmer 554 UVjVIS spectrophotometer. PBP 2x* was incubated with a 1 mM solution of S2d in 10 mM sodium phosphate pH 7.0 at 3 7 T , and the hydrolysis of substrate was monitored at 250 nm.
RESULTS Mapping of RNA polymerase binding sites in the 5’-flanking region of pbpX The pbpX gene was cloned in the vector pJDC9 which is protected by bidirectional transcriptional terminator signals in order to overcome plasmid instability frequently observed when pneumococcal genes are inserted in common E. coli plasmid vectors. In several cases, high promoter activity has been traced as the cause of such instability. The pbpX promoter has tentatively been identified by comparison to E. coli promoter consensus sequences (Laible et al., 1989). In order to see whether it is indeed potentially functional in E. coli, binding of E. coli RNA polymerase to pCG19 harbouring a 1.7-kb HaeIII DNA fragment with the regulatory region of pbpX was tested. Bound RNA polymerase molecules were visualized by electron microscopy and their location determined on a large number of linearized plasmid molecules. Three different restriction endonucleases were used for linearization in independent experiments and the relative locations of RNA-polymyerase -DNA complexes determined for each sample (Fig. 2). Two signals were detected in thepbpX insert. A very weak signal did not correspond to a promoterlike region. The position of the strong signal differed by a maximum of nine nucleotides in the three sets of experiments. The mean value obtained by measuring a total of 264 DNA molecules corresponded to position 228 of the publishedpbpX sequence (Laible et al., 1989), i.e. between the putative - 10 and - 35 regions of the PBP 2x promoter, demonstrating that the E. coli enzyme recognized this structure. Expression of pbpX in E. coli E. coli DH5 and DH5a cells harbouring plasmid pCG29 which contains the structural gene as well as the regulatory region of pbpX express a new protein which was identified as a PBP with the apparent M , of S.pneumoniae PBP 2x (Fig. 3). In order to decide whether expression of PBP 2x from pCG29 is controlled by the lac promoter being located directly adjacent to the 5’-cloning site, or whether the pbpX promoter itself is responsible for PBP 2x production, pCG29 was transformed into the l a d q strain E. coli JM103 and expression tested with and without addition of isopropyl P-d-thiogalactopyranoside to the growth medium. Without it, only low levels of PBP 2x were produced, whereas its presence resulted in a large increase of the PBP 2x product (Fig. 4).
947
A
1
2
3
4 5
B
Fig. 6. Fractionation of PBP 2x and PBP 2x* in E. coli. From E. coli DH5a harbouring pCG29 (left three lanes) or pCG31 (right three Fig.7. Purification of PBP 2x*. Aliquots of different fractions were lanes), cell lysates (L), soluble fraction (s),and membranes (M) were analyzed after SDSjPAGE and Coomassie blue stain. Lane 1, soluble prepared as described in the text and labeled with [ 3 H l p r o ~ i o n ~ l -fraction of E. coli DHSa/pCG31; lane 2, PBP 2x* after dye-affinity ampicillin. (A) Total protein; (B) fluorography of A. The arrows mark chromatography (7 pg); lanes 3 - 5 , PBP 2x* after MonoQ chromathe positions of PBP 2x (left) and PBP 2x* (right). tography (lane 3, 10 pg; lanes 4 and 5 , 1 pg).
Construction and expression of a water-soluble PBP 2x derivative The only stretch of highly hydrophobic amino acids is located at the N-terminal of PBP 2x between Leu30 and Ile48 and is believed to represent the trans-membrane anchor (Laible et al., 1989). Deletion of that portion of the gene should result in a hydrophilic PBP derivative. A site-directed deletion in pbpX was constructed using the polymerase chain reaction technology as described in Materials and Methods, resulting in plasmid pCG31. In this construct 90 bp, encoding amino acids 19-48, are lacking. In order to verify the expected sequence and to identify possible mutations introduced by DNA amplification via polymerase chain reaction, the 5'region of pbpX in two independently obtained recombinant plasmids was sequenced. One plasmid (pCG31-25) contained a silent mutation at nucleotide position CGA498 to CGG, and an A754AA to GAA exchange resulting in a Lys168 to Glu mutation according to the published sequence (Laible et al., 1989). E. coli DH5a cells containing the recombinant plasmids pCG31-26 (which does not carry a mutation) or pCG31-25 grew conspicuously slower as long filaments compared to control cells carrying pJDC9 (not shown). When analyzed on SDS/PAGE, a PBP slightly smaller than the intact PBP 2x was synthesized as the major cellular component in both cases (Fig. 5). No apparent reduction in terms of penicillin binding was apparent when compared to the 19-amino-acid-longer PBP 2x, independent of the mutations in one of the proteins. However, the construct appeared to be somewhat unstable : after restreaking of the initial transformant, two out of twelve cultures grown from single colonies produced only a low level of PBP 2x*. In contrast to the membrane-associated PBP expressed from pCG29, 95% of the pCG31 product PBP 2x* was recovered in the soluble fraction in the absence of detergents (Fig. 6).
1.25
1.15
10
20
30
time (rnin)
Fig. 8. Interaction between substrate S2d and PBP 2x*. Shown is the time course of disappearance of the substrate. The substrate solution (0.5 ml) was incubated at 37°C and after 8 min, PBP 2x* was added (4 pl containing 3 pg protein). d-alanine (50 p1 of a 100 mM solution) was added at time = 15 min (the shift in absorbance is due to the dilution) and the reaction finally stopped at 20 min by addition of 8 pl cefotaxime (22 pM).
and Procion blueheard). Further analysis of the dyes revealed that only Procion blueherd was highly selective for PBP 2x* and could be used for preparative dye-affinity chromatography. After chromatography of the soluble fraction of E. coli DHSa/pCG31,1.6 mg protein was collected in the pooled fraction containing PBP 2x* (5.6% of the total protein applied to the column) and final purification was achieved after MonoQ chromatography (Fig. 7). 0.56 mg of > 99% pure PBP 2x* was obtained in the main peak fraction. Enzymatic activity of PBP 2x*
Purification of PBP 2x* Screening of a set of 98 textile dyes (Mottl and Keck, 1991) led to the identification of five different dyes showing high affinity for PBP 2x* (Dylan A23 Bahama blue, Dylan A29 koala brown, Dylan A30 turquoise saga, Dylon 8 ebony black
Scanning of fluorograms of PBP 2x* indicated that at least 70% of the protein is active in terms of penicillin-binding (not shown). The recent availability of thioester compounds such as S2d provides an easier and more precise way of determining enzymatic activity of low-M, and high-Mr PBPs since its
948 hydrolysis can easily be monitored directly in a spectrophotometer (Adam et al., 1990, 1991). Fig. 8 shows the time course of hydrolysis of S2d by PBP 2x*. Hydrolysis time courses allowed the determination of k,,,/K, = 4200 f 500 M-1 s - l w'ith K,,, > 1 mM. The fact that no acceleration occurred upon addition of d-alanine is a further indication that the conditions used were below the K , of the thioester substrate. The activity was completely abolished upon addition of a 4.6 molar excess (0.4 pM) of cefotaxime over PBP 2x*. No loss of activity was observed after storing the enzyme at 4°C for several months, or in the presence of 10% glycerol at - 20 "C.
in the range of the values determined for E. hirae high-M, PBP derivatives, and approximately tenfold or more higher than those reported for soluble derivatives of S. pneumoniae PBP 2b or E. coli PBP 3 (Adam et al., 1991). Addition of cefotaxime completely inhibited PBP 2x activity. This system provides a suitable basis for studying the enzymatic reactivity of low-affinity PBP 2x derivatives. We thank Beate Dobrinski for skilful technical assistance, C. Damblon for the synthesis of the thioester substrate, and Mike Hearne for synthesis of oligonucleotides. This work was supported by the European Economic Community Contract EEC SC1*/0141-C, and the Fonds der Chemischen Zndustrie im Verband der Chemischen Zndustrie e. V . During these investigations, M. Jamin was on a shortterm EMBO fellowship for three months.
DISCUSSION The very efficient purification of low-M, PBPs 4 and 5 from E. coli by dye-affinity chromatography on Cibacrom navyblue 2GE or Procion rubine MX-B modified resins, respectively, has been reported (Mottl and Keck, 1991; van der Linden et al., 1992). It was further demonstrated that, with the exception of PBP 3, all E. coli PBPs showed affinity to Cibacrom navyblue 2GE. This paper reports for the first time on the purification of a high-Mr PBP by dye-affinity chromatography. Screening of the collection of 98 textile dyes revealed the PBP 2x* did not bind to Cibacrom navyblue but to five different dyes. A high affinity of other high-M, PBPs from S. pneumoniae for blue Sepharose has been reported (Hakenbeck, 1983). However, based on further unpublished observations (W. Keck) that PBP 2b from S. pneumoniae does not bind with sufficient selectivity to any of the 98 dyes, we suspect that the observed affinity of PBPs for the dyes is not a general characteristic of this protein family and is not based on the occurrence of a PBP-typical tertiary structural element as described for dehydrogenases (Thompson et al., 1975). The construction of a water-soluble active PBP 2x derivative (PBP 2x*) proves once more that the only membraneanchoring region of high-Mr PBPs is the N-terminal hydrophobic peptide domain, and that this peptide is apparently not necessary for enzymatic activity with the possible exception of E. coli PBP 3 (Hayashi et al., 1988). That the water-soluble PBP 2x* appeared perfectly stable over several months is further support for the structural independence of the hydrophilic main part of the PBP. Purification of the enzyme was greatly facilitated by the fact that it represented one of the major proteins synthesized in the E. coli host. The higher expression level of PBP 2x* compared to PBP 2x in E. coli as shown in Fig. 5 is most likely due to the different localization of the proteins (cytoplasm versus membrane). Somewhat surprising was the observation that although the pneumococcal promoter which is still present in the recombinant plasmid is recognized by E. coli RNA polymerase in vitro, the in vivo expression is apparently controlled via the lac promoter of the plasmid vector. When cloned in S. pneumoniae, no overproduction of PBP 2x was observed, indicating that the PBP 2x promoter is not a strong one in S. pneumoniae or that the amount of PBP 2x is controlled (unpublished observation). We do not know whether PBP 2x exhibits physiological activity in E. coli. Presumably, the filamentation of E. coli cells may simply reflect the abundance of the protein itself, since this phenomenon was also observed with the cytoplasmic PBP 2x*. Enzymatic activity has been determined using a thioester compound as substrate whose modification by the PBP could easily be monitored. The activity observed for PBP 2x* was
REFERENCES Adachi, H., Ohta, T. & Matsuzawa, H. (1988) FEBS Lett. 226,150154. Adam, M., Damblon, C., Plaitin, B., Christiaens, L. & FrZre, J.-M. (1990) Biochem. J . 270, 525 - 529. Adam, M, Damblo, C., Jamin, M., Z o r ~ i W., , Dusart, V., Galleni, M., El Kharroubi, A,, Piras, G., Spratt, B. G., Keck, W., Coyette, J . , Ghuysen, J.-M., Nguyen-Disteche & Frere, J.-M.(1991) Biochem. J . 279,601 -604. Avery, 0. T., McLeod, C. M. & McCarty, M. (1944) J. Exp. Med. 79, 137-158. Chen, J.-D. & Morrison, D. A. (1988) Gene 64,155-164. Dideberg, O., Charlier, P., WCry. J.-P., Dehottay, P., Dusart, J., Erpicum, T., Frere, J.-M. & Ghuysen, J.-M. (1987) Biochem. J . 245,911 -913. Dowson, C. G., Hutchison, A. & Spratt, B. G . (1989) Mol. Microbiol. 3,95 - 102. Ghuysen, J.-M. (1991) Annu. Rev. Microbiol. 45, 37-67. Hakenbeck, R. (1983) in The target of penicillin (Hakenbeck, R., Holtje, J.-V. & Labischinski, H., eds) pp. 415-420. Walter de Gruyter & Co., Berlin, New York. Hanahan, D. (1985) in DNA cloning (Glover, D. M., ed.) pp. 109135, IRL Press, Oxford. Hayashi, S., Hara, H., Suzuki, H. & Hirota, Y. (1988) J. Bacteriol. 170, 5392-5395. Herzberg, 0. & Moult, J. (1987) Science 236, 694-701. Ishino, F., Mitsui, S., Tamaki, S. & Matsuhashi, M. (1980) Biochem. Biophys. Res. Commun. 97, 287-293. Laible, G. & Hakenbeck, R. (1987) Mol. Microbiol. I , 355-363. Laible, G., Hakenbeck, R., Sicard, M. A,, Joris B. & Ghuysen, J.-M. (1989) Mol. Microbiol. 3, 1337-1348. Laible, G., Spratt, B. G. & Hakenbeck, R. (1991) Mol. Microbiol. 5, 1993-2002. Laible, G. & Hakenbeck, R. (1991) J. Bacteriol. 173, 6986-6990. Lurz, R., Heisig, A., Velleman, M., Dobrinski, B. & Schuster, H. (1987) J . Biol. Chem. 262, 16575-16579. Mejean, V., Claverys, J.-P., Vasseghi, H. & Sicard, A. M. (1981) Gene 15,289-293. Messing, I., Crea, R. & Seeburg, P. H. (1981) Nucleic Acids Res. 9, 309 - 321. Moews, P. C., Knox, J. R., Dideberg, O., Charlier, P. & Frere, J.-M. (1990) Proteins Struct. Funct. Genet. 7, 156-371. Morrison, D. A., Lacks, S. A., Guild, W. R. & Hageman, J. M. (1983) J . Bacteriol. 156, 281 -290. Mottl, H. & Keck, W. (1991) Eur. J . Biochem. 200, 767-773. Oefner, C., D'Arcy, A,, Daly, J. J., Gubernator, K., Charmas, R. L., Heinze, I., Hubschwerlen, C. & Winkler, F. K. (1990) Nature 343, 284- 288. Perez-Martin, J., del Solar, G. H., Lurz, R., de la Campa, A. G., Dobrinski, B. & Espinoza, M. (1989) J . Biol. Chem. 264,2133421 339. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular cloning: a laboratory manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor NY.
949 Spratt. B. G., Bowler, L., Edelman, A. & Broome-Smith, J. K. (1988) in Antibiotic inhibition of bacterial cell surface assembly and function (Actor, P., Daneo-Moore, L., Higgins, M. L., Salton, M. R. J. & Shockman, G. D., eds) pp. 292-300, American Society for Microbiology, Washington DC. Suzuki, S., van Heijenoort, Y., Tamura, T., Mizoguchi, J., Hirota, Y. & v a n Heijenoort, Y. (1980) FEBS Lett. 110,245-249.
Tamaki, S., Nakajima, S. & Matsuhashi, M. (1980) Proc. Natl Acad. Sci. USA 74, 5472 - 5416. Thompson, s. T., Cass, K. H. 8~Stellwagen, E. (1975) Proc. Nut( Acad SCi. USA 72,669-672. van der Linden, M. P. G., Mottl, H. & Keck, W. (1992) Eur. J . Biochem. 204, 197-202. Vogelstein, B. & Gillespie, D. (1979) Proc. Natl Acad. Sci. USA 76, 615- 619.
0FEBS 1992
Penicillin-binding protein 2x of Streptococcus pneumoniae Expression in Esckerichiu coli and purification of a soluble enzymatically active derivative Gotz LAIBLE’, Wolfgang KECK’, Rudi LURZ’, Harald MOTTL’, Jean-Marie FRERE3, Marc JAMIN3 and Regine HAKENBECK’ Max-Planck Institut fur molekulare Genetik, Berlin, Federal Republic of Germany University of Groningen, BIOSON Research Institute, Department of Biochemistry, The Netherlands Universitk de Liige, Centre d’IngCnierie des Proteines, Laboratoire d’Enzymologie, Institut de Chimie B6, Belgium (Received February 4/May 25, 1992) - EJB 920149
A 2.5-kb DNA fragment including the structural gene coding for the penicillin-binding protein 2x (PBP 2x) of Streptococcus pneumoniue has been cloned into the vector pJDC9 and expressed in Escherichia coli. Mapping of RNA polymerase binding sites by electron microscopy indicated that the pbpX promoter is well recognized by the E . coli enzyme. However, high-level expression occurred mainly under the control of the lac promoter upstream of the pJDC9 multiple cloning site. After induction with isopropyl P-d-thiogalactopyranoside, PBP 2x was expressed as one of the major cellular proteins. PBP 2x produced in E. coli corresponded to the pneumococcal PBP 2x in terms of electrophoretic mobility, fractionation with the cytoplasmic membrane, and penicillin-binding capacity. Deletion of 30 hydrophobic N-terminal amino acid residues at positions 19 -48 resulted in high-level expression of a cytoplasmic, soluble PBP 2x derivative (PBP 2x*) which still retained full p-lactam-binding activity. A two-step procedure involving dye affinity chromatography was established for obtaining large amounts of highly purified enzymatically active PBP 2x*.
Penicillin-binding proteins (PBPs), the target enzymes of p-lactam antibiotics, and p-lactamases constitute the family of penicillin-recognizing active-site serine enzymes (Ghuysen, 1991). Comparison of the deduced amino acid sequences of PBPs from different bacterial species and functional analysis mainly performed with Escherichia coli PBPs revealed three classes of PBPs. Low-MI PBPs contain DD-carboxypeptidase activity and appear to be dispensable for cellular growth. High-M, PBPs acting in essential late steps of murein biosynthesis are either homologous to the bifunctional E. coli PBPs l a and l b (Ghuysen, 1991; and unpublished results) or are related to E. coli PBP 2 and 3 , like Streptococcus pneumoniue PBPs 2x (Laible et al., 1989) and 2b (Dowson et al., 1989). E. coli PBP 2 and 3 are important components of the morphogenetic apparatus; the in vivo function of related PBPs from other bacterial species remains largely unknown. High-MI PBPs are mainly periplasmic enzymes that are bound to the membrane via a short N-terminal hydrophobic peptide. In addition to the penicillin-binding domain that includes the active center responsible for penicillin-sensitive transpeptidase activity, they usually carry an N-terminal domain and a long C-terminal extension. In E. coli PBPs l a and 1b, penicillin-insensitive transglycosylase activity is located on the N-terminal domain (Ishino et al., 1980; Suzuki et al., 1980; Tamaki et al., 1980). Correspondence to R. Hakenbeck, Max-Planck Institut fur molekulare Genetik, Ihnestr. 73, W-1000 Berlin 33, Germany Fax: +49 30 830 7385. Abbreviations. PBP, penicillin-binding protein; S2d, b e n z o y h alanyl thioacetyl ester.
PBPs are inhibited by p-lactam antibiotics by enzymatically forming a covalent penicilloyl (or cephalosporoyl) complex via the active-site serine residue which is part of the conserved SXXK sequence. Further conserved boxes in PBPs are the SXN sequence and the triad KTG or KSG that are part of the active-site cavity as deduced from the known three-dimensional structures of several p-lactamases and the Streptomyces R61 low-MI PBP (Dideberg et al., 1987; Herzberg and Moult, 1987; Moews et al., 1990; Oefner et al., 1990). Although not related in amino acid sequence, the polypeptide folding of the R61 PBP and the class A plactamases is amazingly similar with the conserved boxes representing critical components of the active-site cavity. It is assumed that the principal polypeptide arrangement of the R61 enzyme has been conserved in the penicillin-binding domain of high-Mr PBPs. N o experimental data on the threedimensional structure of high-MI PBPs is available yet, and nothing is known about the structure of their N-terminal domain nor their C-terminal extension. The importance of high-MI PBPs in S. pneumoniue was deduced from their role in intrinsic penicillin resistance. In pneumococci, penicillin resistance is mediated by the production of low-penicillin-affinity variants of four high-M, PBPs (Laible et al., 1991). PBP 2x appears to be the major target for p-lactams since it is the first PBP altered in cefotaxime-resistant laboratory mutants (Laible and Hakenbeck, 1987), and since transformation with DNA fragments encoding low-affinity variants of PBP 2x from clinical isolates or from laboratory mutants yield low-level-resistant transformants (Laible et al., 1989, 1990).
944 Dral
Haelll
I
+
c
pCGl9
pCG8 I
Pstl Haelll EwRV
Hind111
1’
+c
EwRl
100 bp
I
2‘
I
I
(Hindlll)
pCG28
I
I
I (EmRI)
I
pCGl0
I
Pstl
pCG29 (PBP 2x)
I
,
I
I
.
, I. ,, .) ,
pCG31 (PBP 2@)
I
iI :i
Pdl
I
I
,
I
Xmal
Fig. 1. Cloning of the PBP 2x and PBP 2x* gene. A map of the S. pneumoniae pbpX gene and its flanking regions is given at the top. The hashed segment marks the structural encoding gene pbpX with the arrow indicating the promoter. The cloned DNA fragments are indicated below. Plasmid pCG19 was used for mapping of E. coli RNA polymerase binding sites in the promoter region. pCG19 and pCG28 are derivatives of pCG9 (pCG9 contains the same insert as pCG8 but in the rcverse orientation). pCG29 containing the structural PBP 2x gene and flanking regions was obtained from pCG28 and pCG10. pCG31 encoding the soluble PBP 2x* was constructed using two DNA fragments obtained from pCG19 b y the polymerase chain reaction (the positions of the primers are indicated. 1, primer 365; l’, 2020; 2, 2021 ; 2’, 623), and the PstI - EcoRI fragment of pCG29 as described in the text.
Point mutations in the PBP 2x gene responsible for reduction of penicillin affinity from several laboratory mutants have been mapped (Laible and Hakenbeck, 1991). A prerequisite for studying the impact of the mutations in more detail and for ultimately resolving the three-dimensional polypypetide arrangement of wild-type and mutant proteins is the availability of large amounts of soluble PBP 2x. Like other highM , PBPs, PBP 2x contains a short N-terminal membraneanchoring domain. This hydrophobic peptide is believed not to be important for enzymatic function, since deletion of this region in E. coli PBP 1b, 2 or 3 had no effect on penicillinbinding activity (Adachi et al., 1988; Spratt et al., 1988). The construction and high-level expression of a soluble PBP 2x derivative (PBP 2x*) in E. coli that has retained full penicillinbinding activity will be described. Furthermore, a protocol was established enabling purification of enzymatically active PBP 2x*. MATERIALS AND METHODS Bacterial strains and plasmids
StreptococcuspneumoniueR6 is an unencapsulated derivative of the Rockefeller University strain R36A (Avery et al., 1944). Pneumococci were grown in caseine hydrolysate broth (Morrison et al., 1983). Plasmids were propagated in E. coli DH5, DH5cr (Hanahan, 1985) or JM103 (Messing et al., 1981) as stated in the text. Cells were grown in Luria-broth; recombinant clones were grown in the presence of 0.5 mg/ml erythromycin. For induction of the lac promoter, isopropyl /3-d-thiogalactopyranosidewas added to exponentially growing cultures at a final concentration of l mM. Plasmids used in this study are described in Fig. 1. Most of the plasmids harbouring different fragments ofpbpX(Laib1e et
al., 1989) are derivatives of the E. coli vector pJDC9 constructed for cloning pneumococcal DNA (Chen and Morrison, 1988). pCG19 contains the Hue111 fragment of pCG9 (Laible et al., 1989); pCG9 carries a Hind11 -EcoRV fragment in the reverse orientation as pCG8 which has been described (Laible et al., 1989); pCG28 is a derivative of plasmid pCG9 from which the 5‘ DraI -Hind11 fragment had been deleted including parts of the multiple cloning site of the vector pJDC9 by Hind11 restriction. pCGlO is a derivative of plasmid pR28 (Mkjean et al., 1981) containing the 1.8-kb PstI-EcoRI fragment of the R6 pbpX. Construction of recombinant plasmids
pCG29 contains a 2.5-kb Hind111 - EcoRI fragment covering the structural gene of PBP 2x and the promoter region. It was constructed by replacing the 3’ PstI - EcoRI fragment of pCG28 with the 1.8-kb PstI-EcoRI fragment of pCGl0 that covers the 3’-region of pbpX including the terminator (Fig. 1). pCG31 encoding the soluble PBP 2x* was constructed by introducing a site-directed deletion into pbpX by polymerase chain reaction technology. Two DNA fragments were amplified from plasmid pCG19. (a) The 5’-DNA fragment was obtained after amplification with the polymerase chain reaction of pCG19 with the two primers, 365 (5’-CTCAATTGGACCAGCG683) and 2020 (5’-TCCCCGGGCGATTTCCGATTTTTGGTCGCA282), the latter containing an endogeneous XmuI site. Restriction of the 1010-bp fragment with Hind11 and XmuI resulted in a 301-bp fragment covering the 5’-flanking region of pbpX and sequences encoding the first 18 amino acids. (b) A 395-bp DNA fragment was amplified from pCG19 with the two primers 2021 (5’-TTCCCGGGACAGGCACTCGCTTTGGAACAG421) carrying an
945 endogenous XmaI site, and 623 (5'-AATCCCCTTGACCTCTGCAG767) ; after restriction with XmaI/PstI, a 375-p fragment encoding a peptide starting with Xaa49 was obtained. Ligation of both fragments via the XmaI site results in deletion of amino acids 19 -48 representing the hydrophobic N-termind domain of PBP 2x. Both DNA fragments, as well as the 8.7-kb Hind11-PstI fragment of pCG29 were isolated from agarose gels and ligated. After transformation into E. coli DHSfa, plasmid pCG31 was obtained.
t
1-
l
Localization of RNA polymerase binding sites Binding of E. coli RNA polymerase (Promega) to pCG19 was performed according to Lurz et al. (1987). A total of 264 DNA molecules were analyzed, using three different restriction endonucleases with single restriction sites each for orientation of bound RNA polymerase: EcoRI, CluI and BumHI. Preparation of the DNA-protein complexes for electron microscopy and evaluation of micrographs were performed as described elsewhere (Perez-Martin et al., 1989). Manipulation of DNA Preparation of DNA and transformation were carried out according to standard protocols (Sambrook et al., 1989). Isolation of DNA fragments from agarose gels was performed as described (Vogelstein and Gillespie, 1979). Restriction endonucleases were obtained from Boehringer (Mannheim, FRG) if not otherwise stated and used according to the manufacturer's specifications. DNA sequencing was carried out using a T7 polymerase sequencing kit (Pharmacia) with a set of oligonucleotides that primed from intervals along the inserted DNA fragment of plasmid pCG31. Amplification of DNA by polymerase chain reaction was performed during 30 cycles consisting of 2.5 min extension at 7 7 T , 1.5-min denaturation at 95 "C, and 1.5-min annealing at 55 "C using a Pharmacia LKB Gene ATAQ Controller. The 50-p1 reaction mixture contained 0.3 ng plasmid DNA, 1 pM of each primer, and 1 U Tuq polymerase (Stratagene). The amplified fragments were purified from agarose gels before further manipulations. Cell fractionation A 40-ml culture of exponentially growing E. coli ( A 5 6 0= 1) was rapidly chilled on ice and cells were harvested by centrifugation. Cells were resuspended in 5 ml ice-cold 20 mM sodium phosphate pH 1.2 and disintegrated by sonification (M. S. E. ultrasonic disintegrator) for four 10-s intervals. Alternatively, 50 mM Tris/HCl pH 8, 100 mM NaCl, 1 mM EDTA was used during the procedure and cells disrupted in a French pressure cell (Amicon) at 137.8 MPa. Remaining intact cells were removed by low-speed centrifugation. The membrane fraction was recovered after ultracentrifugation (4 h, 150000 g). The supernatant after ultracentrifugation served as soluble fraction. No differences between the two methods in terms of localization of PBP 2x and its derivative were detected. Membranes were suspended in buffer with the addition of 0.2% Triton X-100, and stored at -80°C prior to detection of PBPs on SDS gels. Purification of PBP 2x" A 4-1 culture of E. coli DH5a harbouring pCG31 was grown to an A560 = 1.3. Cells were harvested by centrifuga-
EcoRl Fig. 2. RNA polymerase binding sites in pCGl9. The graphic representation summarizes results obtained from three independent data sets derived from linearized pCG19 using three different restriction endonucleases. RNA polymerase binding sites were determined by electron microscopy as described in the text. The map is aligned to the EcoRI site in the multiple cloning site of pJDC9. Positions of the inserted Hue111 pbpX fragment (1.73 kb) and the vector portion (7.2 kb) are indicated. The position of peak 1 corresponds to the position of the proposedpbpXpromoter (Laible et al., 1989). The arrow shows the direction of transcription.
tion, washed once in 50 mM Tris/HCl pH 8.0,lOO mM NaC1, and broken in a French pressure cell press after resuspension in 25 ml of the same buffer. Cell wall material was removed at 3 50000 g (1-h centrifugation); two 0.5-ml portions of the supernatant (28.5 mg) were applied to Procion blueherd mix 3967 coupled to Fractogel TSK HW-65 (F) (Merck, Darmstadt, FRG) dye affinity column (6 ml; 1 cm diameter) at a flow rate of approximately 0.15 ml/min connected to an FPLC system (Pharmacia, Uppsala). Coupling of the dye to Fractogel was performed essentially as described (Mottl and Keck, 1991). The column was washed with 24 ml10 mM Tris/ HCl pH 8.0, 100 mM NaC1, and PBP 2x* was eluted with a linear gradient (10 ml 100 mM NaCl and 1 M NaCl buffer each) followed by 16 ml 1 M NaCl buffer; 2-ml fractions were collected. Peak fractions (8 ml final volume) were concentrated by centrifugation through Centricon-30 microconcentrators (Amicon) with buffer exchange to 10 mM Tris/HCl pH 8; 1.6 mg protein was recovered. Final purification was obtained with a Mono Q 5/5 column equilibrated with 10 mM Tris/HCl pH 8; a maximum of 500 pg protein was applied. PBP 2x* was eluted with a stepwise gradient (20 ml 10 mM Tris/HCl pH 8, 0.4 M NaCl); 500-p1 fractions were collected at a flow rate of 60 ml/h. PBP 2x* eluted at approximately 250 mM NaCl; 560 pg PBP 2x* was obtained in 5.6 ml. Detection of PBPs PBPs were routinely labeled with [3H]propionylampicillin and visualized on fluorograms as described (Laible et al., 1987). For quantification of bound penicillin, 269 pmol [3H]benzylpenicillin(NEN Dupont, Dreiech, FRG; 28.2 Ci/ mmol) was incubated with PBP 2x* (5.38 pmol) in 20 p1 for 30 min at 37°C. After SDSjPAGE and fluorography, dried gel slices were incubated with 250 ~ 1 3 0 % H 2 0 2for 2 h at 50 "C and radioactivity measured after addition of 5 ml Instagel (Packard Instruments, Illinois) in a Beckmann LS 9800 scintillation counter. Counting efficiency was determined by spotting different amounts of [3H]benzylpenicillinon gel slices and measuring radioactivity as above.
946
6
A
1’ 2’ 3’
1 2 3
1
2
3
4
1’2’3’4‘
Fig.5. Expression of PBP 2x” from pCG31. The ligation mixture containing pCG31 was transformed into E. coli DH5u. Several independent transformants were analyzed, two examples being shown in lanes 3 and 4. For comparison, E. coli DH5 harbouring pCG29 (lane 1) and the vector pJDC9 (lane 2) are included. Total cellular protein (A) and PBPs (B) were visualized as described in the legend to Fig. 4.
Fig. 3. Expression of pbpX in E. cob. Cell lysates were incubated with [3H]propionylampicillin, proteins separated by SDSjPAGE and PBPs visualized after fluorography. Lanes 1 -3, protein pattern after Coomassie blue stain; lanes 1’-3’, fluorogram of 1-3 (5-days exposure). Lanes l,l’, E. coli DH5 harbouring pJDC9; lanes 2,2’, E. coli DH5 harbouring pCG29; lanes 3,3’, S. pneumoniue R6. The position of the R6 PBP 2x is indicated in the fluorogram; the arrow marks the additional protein band in DH5/pCG29.
A
1
2
3
4
5
6
7
6
Fig. 4. Induction of PBP 2x in E. coli. To an exponentially growing E. coli JM103/pCG29 culture, isopropyl j-d-thiogalactopyranoside was added at time zero (lane 1) and samples taken after 2 h (lane 3), 4 h (lane 5) and 6 h (lane 7). The control culture remained without isopropyl j-d-thiogalactopyranoside (lanes 2, 4 and 6). Cell lysates were prepared and PBPs labeled with [3H]propionylampicillin. (A) Total cellular protein after SDSjPAGE and Coomassie blue stain; (B) fluorogram of A after a 6-h exposure.
Enzymatic activity of PBP 2x* Activity of the purified PBP 2x* was determined using the compound S2d as substrate. S2d, a thioester compound (C6H,-CO-D-Ala-S-CH2-COOH),has been described (Adam et al., 1990). Spectrophotometric measurements were performed with a Perkin-Elmer 554 UVjVIS spectrophotometer. PBP 2x* was incubated with a 1 mM solution of S2d in 10 mM sodium phosphate pH 7.0 at 3 7 T , and the hydrolysis of substrate was monitored at 250 nm.
RESULTS Mapping of RNA polymerase binding sites in the 5’-flanking region of pbpX The pbpX gene was cloned in the vector pJDC9 which is protected by bidirectional transcriptional terminator signals in order to overcome plasmid instability frequently observed when pneumococcal genes are inserted in common E. coli plasmid vectors. In several cases, high promoter activity has been traced as the cause of such instability. The pbpX promoter has tentatively been identified by comparison to E. coli promoter consensus sequences (Laible et al., 1989). In order to see whether it is indeed potentially functional in E. coli, binding of E. coli RNA polymerase to pCG19 harbouring a 1.7-kb HaeIII DNA fragment with the regulatory region of pbpX was tested. Bound RNA polymerase molecules were visualized by electron microscopy and their location determined on a large number of linearized plasmid molecules. Three different restriction endonucleases were used for linearization in independent experiments and the relative locations of RNA-polymyerase -DNA complexes determined for each sample (Fig. 2). Two signals were detected in thepbpX insert. A very weak signal did not correspond to a promoterlike region. The position of the strong signal differed by a maximum of nine nucleotides in the three sets of experiments. The mean value obtained by measuring a total of 264 DNA molecules corresponded to position 228 of the publishedpbpX sequence (Laible et al., 1989), i.e. between the putative - 10 and - 35 regions of the PBP 2x promoter, demonstrating that the E. coli enzyme recognized this structure. Expression of pbpX in E. coli E. coli DH5 and DH5a cells harbouring plasmid pCG29 which contains the structural gene as well as the regulatory region of pbpX express a new protein which was identified as a PBP with the apparent M , of S.pneumoniae PBP 2x (Fig. 3). In order to decide whether expression of PBP 2x from pCG29 is controlled by the lac promoter being located directly adjacent to the 5’-cloning site, or whether the pbpX promoter itself is responsible for PBP 2x production, pCG29 was transformed into the l a d q strain E. coli JM103 and expression tested with and without addition of isopropyl P-d-thiogalactopyranoside to the growth medium. Without it, only low levels of PBP 2x were produced, whereas its presence resulted in a large increase of the PBP 2x product (Fig. 4).
947
A
1
2
3
4 5
B
Fig. 6. Fractionation of PBP 2x and PBP 2x* in E. coli. From E. coli DH5a harbouring pCG29 (left three lanes) or pCG31 (right three Fig.7. Purification of PBP 2x*. Aliquots of different fractions were lanes), cell lysates (L), soluble fraction (s),and membranes (M) were analyzed after SDSjPAGE and Coomassie blue stain. Lane 1, soluble prepared as described in the text and labeled with [ 3 H l p r o ~ i o n ~ l -fraction of E. coli DHSa/pCG31; lane 2, PBP 2x* after dye-affinity ampicillin. (A) Total protein; (B) fluorography of A. The arrows mark chromatography (7 pg); lanes 3 - 5 , PBP 2x* after MonoQ chromathe positions of PBP 2x (left) and PBP 2x* (right). tography (lane 3, 10 pg; lanes 4 and 5 , 1 pg).
Construction and expression of a water-soluble PBP 2x derivative The only stretch of highly hydrophobic amino acids is located at the N-terminal of PBP 2x between Leu30 and Ile48 and is believed to represent the trans-membrane anchor (Laible et al., 1989). Deletion of that portion of the gene should result in a hydrophilic PBP derivative. A site-directed deletion in pbpX was constructed using the polymerase chain reaction technology as described in Materials and Methods, resulting in plasmid pCG31. In this construct 90 bp, encoding amino acids 19-48, are lacking. In order to verify the expected sequence and to identify possible mutations introduced by DNA amplification via polymerase chain reaction, the 5'region of pbpX in two independently obtained recombinant plasmids was sequenced. One plasmid (pCG31-25) contained a silent mutation at nucleotide position CGA498 to CGG, and an A754AA to GAA exchange resulting in a Lys168 to Glu mutation according to the published sequence (Laible et al., 1989). E. coli DH5a cells containing the recombinant plasmids pCG31-26 (which does not carry a mutation) or pCG31-25 grew conspicuously slower as long filaments compared to control cells carrying pJDC9 (not shown). When analyzed on SDS/PAGE, a PBP slightly smaller than the intact PBP 2x was synthesized as the major cellular component in both cases (Fig. 5). No apparent reduction in terms of penicillin binding was apparent when compared to the 19-amino-acid-longer PBP 2x, independent of the mutations in one of the proteins. However, the construct appeared to be somewhat unstable : after restreaking of the initial transformant, two out of twelve cultures grown from single colonies produced only a low level of PBP 2x*. In contrast to the membrane-associated PBP expressed from pCG29, 95% of the pCG31 product PBP 2x* was recovered in the soluble fraction in the absence of detergents (Fig. 6).
1.25
1.15
10
20
30
time (rnin)
Fig. 8. Interaction between substrate S2d and PBP 2x*. Shown is the time course of disappearance of the substrate. The substrate solution (0.5 ml) was incubated at 37°C and after 8 min, PBP 2x* was added (4 pl containing 3 pg protein). d-alanine (50 p1 of a 100 mM solution) was added at time = 15 min (the shift in absorbance is due to the dilution) and the reaction finally stopped at 20 min by addition of 8 pl cefotaxime (22 pM).
and Procion blueheard). Further analysis of the dyes revealed that only Procion blueherd was highly selective for PBP 2x* and could be used for preparative dye-affinity chromatography. After chromatography of the soluble fraction of E. coli DHSa/pCG31,1.6 mg protein was collected in the pooled fraction containing PBP 2x* (5.6% of the total protein applied to the column) and final purification was achieved after MonoQ chromatography (Fig. 7). 0.56 mg of > 99% pure PBP 2x* was obtained in the main peak fraction. Enzymatic activity of PBP 2x*
Purification of PBP 2x* Screening of a set of 98 textile dyes (Mottl and Keck, 1991) led to the identification of five different dyes showing high affinity for PBP 2x* (Dylan A23 Bahama blue, Dylan A29 koala brown, Dylan A30 turquoise saga, Dylon 8 ebony black
Scanning of fluorograms of PBP 2x* indicated that at least 70% of the protein is active in terms of penicillin-binding (not shown). The recent availability of thioester compounds such as S2d provides an easier and more precise way of determining enzymatic activity of low-M, and high-Mr PBPs since its
948 hydrolysis can easily be monitored directly in a spectrophotometer (Adam et al., 1990, 1991). Fig. 8 shows the time course of hydrolysis of S2d by PBP 2x*. Hydrolysis time courses allowed the determination of k,,,/K, = 4200 f 500 M-1 s - l w'ith K,,, > 1 mM. The fact that no acceleration occurred upon addition of d-alanine is a further indication that the conditions used were below the K , of the thioester substrate. The activity was completely abolished upon addition of a 4.6 molar excess (0.4 pM) of cefotaxime over PBP 2x*. No loss of activity was observed after storing the enzyme at 4°C for several months, or in the presence of 10% glycerol at - 20 "C.
in the range of the values determined for E. hirae high-M, PBP derivatives, and approximately tenfold or more higher than those reported for soluble derivatives of S. pneumoniae PBP 2b or E. coli PBP 3 (Adam et al., 1991). Addition of cefotaxime completely inhibited PBP 2x activity. This system provides a suitable basis for studying the enzymatic reactivity of low-affinity PBP 2x derivatives. We thank Beate Dobrinski for skilful technical assistance, C. Damblon for the synthesis of the thioester substrate, and Mike Hearne for synthesis of oligonucleotides. This work was supported by the European Economic Community Contract EEC SC1*/0141-C, and the Fonds der Chemischen Zndustrie im Verband der Chemischen Zndustrie e. V . During these investigations, M. Jamin was on a shortterm EMBO fellowship for three months.
DISCUSSION The very efficient purification of low-M, PBPs 4 and 5 from E. coli by dye-affinity chromatography on Cibacrom navyblue 2GE or Procion rubine MX-B modified resins, respectively, has been reported (Mottl and Keck, 1991; van der Linden et al., 1992). It was further demonstrated that, with the exception of PBP 3, all E. coli PBPs showed affinity to Cibacrom navyblue 2GE. This paper reports for the first time on the purification of a high-Mr PBP by dye-affinity chromatography. Screening of the collection of 98 textile dyes revealed the PBP 2x* did not bind to Cibacrom navyblue but to five different dyes. A high affinity of other high-M, PBPs from S. pneumoniae for blue Sepharose has been reported (Hakenbeck, 1983). However, based on further unpublished observations (W. Keck) that PBP 2b from S. pneumoniae does not bind with sufficient selectivity to any of the 98 dyes, we suspect that the observed affinity of PBPs for the dyes is not a general characteristic of this protein family and is not based on the occurrence of a PBP-typical tertiary structural element as described for dehydrogenases (Thompson et al., 1975). The construction of a water-soluble active PBP 2x derivative (PBP 2x*) proves once more that the only membraneanchoring region of high-Mr PBPs is the N-terminal hydrophobic peptide domain, and that this peptide is apparently not necessary for enzymatic activity with the possible exception of E. coli PBP 3 (Hayashi et al., 1988). That the water-soluble PBP 2x* appeared perfectly stable over several months is further support for the structural independence of the hydrophilic main part of the PBP. Purification of the enzyme was greatly facilitated by the fact that it represented one of the major proteins synthesized in the E. coli host. The higher expression level of PBP 2x* compared to PBP 2x in E. coli as shown in Fig. 5 is most likely due to the different localization of the proteins (cytoplasm versus membrane). Somewhat surprising was the observation that although the pneumococcal promoter which is still present in the recombinant plasmid is recognized by E. coli RNA polymerase in vitro, the in vivo expression is apparently controlled via the lac promoter of the plasmid vector. When cloned in S. pneumoniae, no overproduction of PBP 2x was observed, indicating that the PBP 2x promoter is not a strong one in S. pneumoniae or that the amount of PBP 2x is controlled (unpublished observation). We do not know whether PBP 2x exhibits physiological activity in E. coli. Presumably, the filamentation of E. coli cells may simply reflect the abundance of the protein itself, since this phenomenon was also observed with the cytoplasmic PBP 2x*. Enzymatic activity has been determined using a thioester compound as substrate whose modification by the PBP could easily be monitored. The activity observed for PBP 2x* was
REFERENCES Adachi, H., Ohta, T. & Matsuzawa, H. (1988) FEBS Lett. 226,150154. Adam, M., Damblon, C., Plaitin, B., Christiaens, L. & FrZre, J.-M. (1990) Biochem. J . 270, 525 - 529. Adam, M, Damblo, C., Jamin, M., Z o r ~ i W., , Dusart, V., Galleni, M., El Kharroubi, A,, Piras, G., Spratt, B. G., Keck, W., Coyette, J . , Ghuysen, J.-M., Nguyen-Disteche & Frere, J.-M.(1991) Biochem. J . 279,601 -604. Avery, 0. T., McLeod, C. M. & McCarty, M. (1944) J. Exp. Med. 79, 137-158. Chen, J.-D. & Morrison, D. A. (1988) Gene 64,155-164. Dideberg, O., Charlier, P., WCry. J.-P., Dehottay, P., Dusart, J., Erpicum, T., Frere, J.-M. & Ghuysen, J.-M. (1987) Biochem. J . 245,911 -913. Dowson, C. G., Hutchison, A. & Spratt, B. G . (1989) Mol. Microbiol. 3,95 - 102. Ghuysen, J.-M. (1991) Annu. Rev. Microbiol. 45, 37-67. Hakenbeck, R. (1983) in The target of penicillin (Hakenbeck, R., Holtje, J.-V. & Labischinski, H., eds) pp. 415-420. Walter de Gruyter & Co., Berlin, New York. Hanahan, D. (1985) in DNA cloning (Glover, D. M., ed.) pp. 109135, IRL Press, Oxford. Hayashi, S., Hara, H., Suzuki, H. & Hirota, Y. (1988) J. Bacteriol. 170, 5392-5395. Herzberg, 0. & Moult, J. (1987) Science 236, 694-701. Ishino, F., Mitsui, S., Tamaki, S. & Matsuhashi, M. (1980) Biochem. Biophys. Res. Commun. 97, 287-293. Laible, G. & Hakenbeck, R. (1987) Mol. Microbiol. I , 355-363. Laible, G., Hakenbeck, R., Sicard, M. A,, Joris B. & Ghuysen, J.-M. (1989) Mol. Microbiol. 3, 1337-1348. Laible, G., Spratt, B. G. & Hakenbeck, R. (1991) Mol. Microbiol. 5, 1993-2002. Laible, G. & Hakenbeck, R. (1991) J. Bacteriol. 173, 6986-6990. Lurz, R., Heisig, A., Velleman, M., Dobrinski, B. & Schuster, H. (1987) J . Biol. Chem. 262, 16575-16579. Mejean, V., Claverys, J.-P., Vasseghi, H. & Sicard, A. M. (1981) Gene 15,289-293. Messing, I., Crea, R. & Seeburg, P. H. (1981) Nucleic Acids Res. 9, 309 - 321. Moews, P. C., Knox, J. R., Dideberg, O., Charlier, P. & Frere, J.-M. (1990) Proteins Struct. Funct. Genet. 7, 156-371. Morrison, D. A., Lacks, S. A., Guild, W. R. & Hageman, J. M. (1983) J . Bacteriol. 156, 281 -290. Mottl, H. & Keck, W. (1991) Eur. J . Biochem. 200, 767-773. Oefner, C., D'Arcy, A,, Daly, J. J., Gubernator, K., Charmas, R. L., Heinze, I., Hubschwerlen, C. & Winkler, F. K. (1990) Nature 343, 284- 288. Perez-Martin, J., del Solar, G. H., Lurz, R., de la Campa, A. G., Dobrinski, B. & Espinoza, M. (1989) J . Biol. Chem. 264,2133421 339. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular cloning: a laboratory manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor NY.
949 Spratt. B. G., Bowler, L., Edelman, A. & Broome-Smith, J. K. (1988) in Antibiotic inhibition of bacterial cell surface assembly and function (Actor, P., Daneo-Moore, L., Higgins, M. L., Salton, M. R. J. & Shockman, G. D., eds) pp. 292-300, American Society for Microbiology, Washington DC. Suzuki, S., van Heijenoort, Y., Tamura, T., Mizoguchi, J., Hirota, Y. & v a n Heijenoort, Y. (1980) FEBS Lett. 110,245-249.
Tamaki, S., Nakajima, S. & Matsuhashi, M. (1980) Proc. Natl Acad. Sci. USA 74, 5472 - 5416. Thompson, s. T., Cass, K. H. 8~Stellwagen, E. (1975) Proc. Nut( Acad SCi. USA 72,669-672. van der Linden, M. P. G., Mottl, H. & Keck, W. (1992) Eur. J . Biochem. 204, 197-202. Vogelstein, B. & Gillespie, D. (1979) Proc. Natl Acad. Sci. USA 76, 615- 619.
