Vol. 176, No. 2JOURNAL OF BACTERIOLOGY, Jan. 1994, p. 443-4490021-9193/94/$04.00+0

Site-Directed Mutagenesis of the mecA Gene from aMethicillin-Resistant Strain of Staphylococcus aureus

C. Y. ERNIE WU,' WILLIAM E. ALBORN, JR.,1 JANE E. FLOKOWITSCH,1 JOANN HOSKINS,'SERHAT UNAL,2 LARRY C. BLASZCZAK,1 DAVID A. PRESTON,1 AND PAUL L. SKATRUD1*

Infectious Disease Research, Eli Lilly and Company, Indianapolis, Indiana 46285,1 and Section of Infectious Diseases,Department of Medicine, Hacettepe University, Ankara, Turkey2

Received 2 August 1993/Accepted 11 November 1993

The mecA-27r gene from Staphylococcus aureus 27r encodes penicillin-binding protein 2a (PBP2a-27r), whichcauses this strain to be methicillin resistant. Removal or replacement of the N-terminal transmembranedomain had no effect on binding of penicillin, but removal of portions of the putative transglycosylase domain(144, 245, or 341 amino acids after the transmembrane region) destroyed penicillin-binding activity. TheSXXK, SXN, and KSG motifs, present in all penicillin-interacting enzymes, were found in the expected linearspatial arrangement within the putative transpeptidase region of PBP2a-27r. Alterations of amino acids in allthree of these motifs resulted in elimination of penicillin-binding activity, confirming their roles in theinteraction with penicillin.

The emergence of Staphylococcus aureus strains resistant tomethicillin and other beta-lactam antibiotics occurred shortlyafter the introduction of methicillin into clinical use (8).Methicillin-resistant S. aureus and Staphylococcus epidermidiscause serious nosocomial infections in patients whose resis-tance to colonization has been compromised. Such nosocomialinfections represent a significant cause of morbidity and mor-tality in hospitals and chronic care facilities (7, 17).

Resistance to methicillin in clinical isolates of staphylococciis mediated by a novel high-molecular-weight penicillin-bind-ing protein (PBP) referred to as PBP2a (1, 5, 6, 24). Thismembrane-bound protein is a unique cell wall-synthesizingenzyme possessing low affinity for beta-lactam antibiotics (21).Because of this uncharacteristically low affinity for beta-lac-tams, PBP2a can apparently compensate for the transpeptida-tion function of all other PBPs in staphylococci and allowsurvival in the presence of otherwise lethal concentrations ofbeta-lactams (14). Staphylococci growing in the presence ofmethicillin express PBP2a, which causes an alteration in thecell wall architecture by reducing the cross-linking in pepti-doglycan (2, 3, 15, 32, 35).PBP2a is encoded by the mecA gene, which is highly

conserved in all clinical isolates of staphylococci that aremethicillin-resistant (18, 25, 28, 33). Geographically distinctmethicillin-resistant staphylococcal isolates from Japan, theUnited States, and Europe (including S. aureus and S. epider-midis) were more than 99% identical in their respective mecADNA sequences (25, 33). The mecA gene is not present inclinical isolates of methicillin-susceptible staphylococci (31),suggesting that mecA was obtained by genetic transfer fromanother species rather than by alteration of an endogenousPBP-encoding gene. The results of a recent epidemiologicalstudy are consistent with a clonal origin of methicillin resis-tance in the staphylococci followed by rapid spread to relatedstaphylococcal species (12). The source of the mecA generemains unidentified.

Analysis of the amino acid sequence of PBP2a deduced from

* Corresponding author. Mailing address: Infectious Disease Re-search, Eli Lilly and Company, Lilly Corporate Center, Indianapolis,IN 46285. Phone: (317) 276-7081. Fax: (317) 276-9086. Electronic mailaddress: Skatrud_Paul@Lilly.com.

the DNA sequence revealed specific motifs common to peni-cillin-interacting enzymes (25, 28, 33). The functional signifi-cance of these motifs has not been tested in PBP2a. Song andcoworkers (28) suggested that PBP2a is composed of twodomains-an N-terminal transglycosylase domain followed bya transpeptidase domain. However, the actual function andinterdependence of the domains have not been explored.Functional analysis of conserved motifs in PBP2a, as well asinvestigation of the relationship between domains, may beuseful in a structure-based drug design program aimed atdeveloping inhibitors of PBP2a.

This communication describes a series of site-specific mu-tagenesis experiments that were designed to probe the effect ofthe transglycosylase domain on penicillin-binding activity aswell as the significance of conserved motifs in PBP2a-27r.

MATERIALS AND METHODS

Plasmids, genes, and bacterial strains. The plasmids used invarious aspects of this study are described briefly in Resultsand are listed in Fig. 3 and 4. The starting plasmid used forsoluble expression of PBP2a-27r in E. coli contained a Bacillussubtilis vegetative promoter controlling expression of themecA-27r gene (19) and a kanamycin resistance gene formaintenance in E. coli. The mecA-27r gene was described in aprevious study (33). E. coli DH5ot/pEWSA37 which containsthe wild-type mecA-27r gene has been deposited in the North-ern Regional Research Laboratories Culture Collection, Peo-ria, Ill. (accession no. B-18753). Mutant forms of the mecA-27rgene were constructed in this study. The numbering systemused to describe the PBP2a amino acid sequence is that of themecA gene cloned from strain TK784 (28). Transformation-competent E. coli K-12 DH5ot (maximum efficiency [BethesdaResearch Laboratories, Gaithersburg, Md.]) cells were used asa host for protein expression and subcloning procedures. E.coli CJ236 (Bio-Rad Laboratories, Hercules, Calif.) was usedto produce uracil-containing single-stranded template DNAfor in vitro mutagenesis. M13mpl8 and M13mpl9 were pur-chased from New England Biolabs.DNA manipulations. Routine procedures, including DNA

isolation, gel electrophoresis, Southern blotting, Western blot(immunoblot) analysis, DNA sequencing, hybridization, clon-

443

on March 23, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

444 WU ET AL.

Deletion

Site

B

Aa.a. 144 5'CGTTCTGATTTTAAATTTTICCATGGIGTATGCTTTGGTCTTTCTGC

a.a. 245 5'CCTAATAGATGTGAAGTCGCTTTTTIC1GGATAGTTACGACTTTCTG

CCTGAGCCATAATCATTTTCATGG GTTATAAATACTCTTTTGAAC

Motif

Primer

SXXK

S403 403K4N-+ R4

SXN

S403- C483

N4%-+ 046KSG

SW-+ Cm

5'

Ser LsCCAGATTACAACTTCACCAGG T ACTC AM TATTAACAGC

TGCCys - CGT

ArgSer Asn

CGACTTAAAACAAGCAATAGAATC4=TTTGC_______~~~~JG.Gin

Ser5' CTTATGCAAACTTAATTGGCAAGGTACTGCAGAACTCA

Cys

FIG. 1. Synthetic oligonucleotide primers used for site-specificmutagenesis of the mecA-27r gene. (A) Deletion primers (5'--3') usedto remove portions of the putative transglycosylase domain fromPBP2a-27r. The boxed nucleotides indicate the positions of the NcoIrestriction sites. a.a., amino acid. (B) Mutagenic primers used to

change the amino acid sequence of conserved motifs in the putativetranspeptidase domain of PBP2a-27r. The DNA sequence presented isthat found in the wild-type form of mecA-27r. The solid lines below theDNA sequence represent the sequence of the mutagenic primer whereit does not deviate from that of the wild type. The boxed codonsindicate the codons that were altered. Amino acids illustrated abovethe boxes are the wild type, while the amino acid alterations are listedbelow the boxes.

ing, and other molecular biology techniques, were performedin substantial accordance with the procedures outlined byManiatis and colleagues (16). Restriction enzymes were ob-tained from Boehringer Mannheim and used according to themanufacturer's recommendations. The site-specific mutagene-sis procedure of Kunkel was used to introduce changes in theDNA sequence of the mecA-27r gene (13). Single-strandedDNA template was generated with the bacteriophage M13 asdescribed by Wu and colleagues (33). Synthetic DNA primersused for site-specific mutagenesis were synthesized on anautomated DNA synthesizer (model 380B; Applied BiosystemsInc.). The DNA sequences of synthetic mutagenic primers,along with specific changes incorporated in the DNA sequenceor encoded protein, are listed in Fig. 1. All mutations inducedwere confirmed by DNA sequencing with the TaqTrack Se-

quencing System (Promega Biotec, Madison, Wis.).Detection of wild-type and modified forms of PBP2a-27r.

The presence of PBP2a-27r and modified forms of this enzymein recombinant E. coli cells was detected by Western blot

analysis with polyclonal antibodies directed against PBP2a-27r

(16). Anti-PBP2a-27r antibodies were prepared by HazeltonResearch Products, Inc. (Denver, Pa.) from rabbits injectedwith purified PBP2a-27r.

Assessment of penicillin-binding activity. Crude cell extractswere obtained from recombinant E. coli cells after digestion

with lysozyme and sonication as described by Wu and col-leagues (33). The ability of wild-type and modified forms ofPBP2a-27r to bind '25I-penicillin V, both alone and in compe-tition with other beta-lactam compounds (to estimate a 50%inhibitory concentration [IC50]) was performed as describedpreviously (23, 33).

Nucleotide sequence accession number. Nucleotide se-quencing information related to the mecA gene cloned fromstrain TK784 is available from the EMBL Data Library(accession no. Y00688) (28).

RESULTS

Alterations of the transmembrane region of PBP2a-27r. Ingeneral, high-molecular-weight PBPs are composed of threestructurally distinct regions: a membrane anchor, a transglyco-sylase domain, and a transpeptidase domain. We recentlyreported the cloning and production of a water-soluble form ofPBP2a-27r in which the N-terminal membrane anchor (21amino acids) of PBP2a-27r was removed by site-specific mu-tagenesis of mecA-27r (33). The modified gene was expressedin E. coli, and cell extracts were analyzed for the presence ofthe modified PBP2a. No perturbation in penicillin-bindingactivity was observed in this modified version of PBP2a-27rwhen analyzed for its ability to bind penicillin (33 [pEWSA30in Fig. 3]). In work to be published elsewhere (34), we replacedthe transmembrane region with a short amino acid sequence(MGHWHHH [chelating peptide in plasmid pEWSA31]) thatwould enable PBP2a-27r to bind metal ions such as zinc (10,27). Addition of the chelating peptide to PBP2a-27r did notsignificantly alter its ability to bind penicillin and provided auseful molecular handle for protein purification.Removal of the transmembrane region and addition of the

chelating peptide without detrimental effect on penicillin bind-ing suggested that the putative transglycosylase domain mightnot be critical to penicillin-binding activity. These observationsprompted an investigation into whether all or a portion of thetransglycosylase-like domain might be removed from PBP2a-27r without loss of penicillin-binding activity. If a shortenedversion of this protein retained penicillin-binding activity andcrystallized more readily than the full-length protein, thenuseful structural information on the active site of the putativetranspeptidase domain might be obtained more easily.

Effect of removal of the putative transglycosylase domain ofPBP2a-27r on penicillin-binding activity. The beginning of thetransglycosylase-like domain of PBP2a-27r was assumed to belocated adjacent to the membrane-spanning region and tocontinue to roughly amino acid residue 341, the presumedbeginning of the transpeptidase-like domain. The beginning ofthe transpeptidase region was estimated by counting back 60amino acids from the SXXK motif, which is typically found 60amino acids from the N terminus of a transpeptidase (9).

Site-specific in vitro mutagenesis was used to alter themecA-27r gene by inserting an NcoI restriction site in framewith the open reading frame of PBP2a-27r at four differentpositions within the putative transglycosylase domain. TheNcoI restriction site was particularly useful in the constructionof E. coli vectors that express the mutated mecA-27r genes.Initial expression vectors contained a thermal-inducible ver-sion of the APLpromoter for expression of foreign proteins(26, 33). Analyses of cell extracts from transformants carryingtruncated versions of mecA-27r in the APLexpression systemrevealed the presence of new proteins of the predicted sizes. Inmost cases, virtually all of the modified PBP2a-27r was foundin inclusion bodies (granules) when the thermally inducible APL expression system was used to express the shortened

I

J. BACTERIOL.

on March 23, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

MUTAGENESIS OF THE mecA GENE 445

FIG. 2. Analysis of mutant forms of PBP2a-27r possessing dele-tions in the putative transglycosylase domain. (A) Results of compe-tition assays. The autoradiogram on the left side displays resultsobtained with membrane proteins (including PBP2a-27r) extractedfrom S. aureus 27r cells. The arrow indicates the estimation of an 150

(IC50 [i.e., the concentration of cefamandole required to reduce"2I-penicillin V binding by 50%]). The right side of panel A presentsa similar analysis of protein extracts obtained from cells carryingplasmid pEWSA76 ( - 245 amino acids [a.a.]; XPL expression system).When plasmid pEWSA76 was used, virtually all recombinant proteinappeared in granules. (B) Western blot analysis and penicillin-bindingdata from extracts obtained from cells containing plasmids thatproduce a high level of soluble protein (i.e., protein not in granules).Lanes: 1, plasmid pEWSA86 (-144 amino acids [lac expressionsystem]); 2, plasmid pEWSA52 (-21 amino acids [veg expressionsystem]); 3, no transforming DNA.

versions of PBP2a-27r. The modified proteins from crude cellextracts, upon initial inspection, appeared to retain penicillin-binding activity on the basis of their interaction with 12511penicillin V. An example of this binding is illustrated in Fig. 2A(right side) for extracts containing PBP2a-27r with 245 aminoacids removed from the N terminus. However, even at highconcentrations of cefamandole, no IC50 was formed with thistruncated PBP2a-27r. A similar result was obtained whenextracts from cells producing PBP2a-27r shortened by 144 or341 amino acids were analyzed with the XPL expression system.In contrast, an IC50 of between 90 and 180 ,ug of cefamandoleper ml was estimated for wild-type PBP2a-27r obtained from S.aureus 27r (Fig. 2A, left side). The inability to detect an IC50for the truncated forms of PBP2a-27r within the concentrationrange of cefamandole employed was inconsistent with bindingof the radiolabeled penicillin to the active site of theseproteins. Other experiments snot described here) suggestedthat nonspecific binding of 1 I-penicillin V was frequentlyassociated with inclusion body or granule formation. Theamount of soluble active recombinant protein (i.e., not ingranules) varied between isolates. Thus, in some cases, it waspossible to assess penicillin binding for a particular PBP2a-27rderivative, while in others it was not. Granule formation wasalso problematic in protein purification. Only about 2% of the

PenicillilnPlasnid Muttion Transglycosylass ITranseRtidase Binding

I I' wr MSR +

PEWS"12 CP

pEWSA31 2 21 a.a. + CP +

pEWSA6 -144a.a.L -

2pEWSA76 -245aa C_

2pEWSA74 .341 a.. _

FIG. 3. Description of plasmids and mutations and summary ofresults of penicillin-binding experiments for deletions of the putativetransglycosylase domain. Plasmid designations are presented at the leftfollowed by a brief description of alterations induced in PBP2a-27r bymutations in the mecA-27r gene. Plasmids that have a box drawnaround their name have been described previously (33). To the right ofthe mutations is illustrated the extent of N-terminal deletions. Resultsof penicillin-binding assays are summarized at the right side: +, adetectable degree of penicillin binding; -, no detectable penicillinbinding. Abbreviations: WT, wild-type PBP2a-27r; MSR, membrane-spanning region (diagonally hatched area); CP, chelating peptide(MGHWHH [vertically hatched area]). The thin line represents theregion of the protein that was deleted. Superscripts: 1, plasmid carriesthe native S. aureus promoter; 2, plasmid carries the A PL thermal-inducible promoter; 3, plasmid carries the E. coli lac promoter.

protein present in inclusion bodies was full-length PBP2a-27r.Furthermore, purification of PBP2a-27r from inclusion bodieswould require solubilization and subsequent refolding of theprotein in an active form for structural studies. In an attemptto eliminate nonspecific background binding, achieve higherlevels of PBP2a-27r expression, and avoid the need for com-plex unfolding-refolding experiments, alternative expressionsystems were examined in which inclusion bodies would not beformed or would at least be minimized. Expression of PBP2a-27r in baculovirus produced correctly folded soluble protein;however, the amount of the desired recombinant proteinproduced was insufficient for our purposes.

Alternative expression systems, utilizing either a B. subtilisvegetative promoter (veg) (19) or the E. coli lac promoter,allowed constitutive expression of modified mecA-27r genes ata lower temperature (25°C). With these expression systems,inclusion body formation could be avoided and analysis oftruncated forms of PBP2a-27r was less complicated by thepresence of nonspecific background binding of 125I-penicillinV. An example of the results obtained when modified PBP2a-27r proteins were expressed in a soluble form is displayed inFig. 2B. In this example, results obtained from Western blotanalysis and a penicillin-binding assay with extracts from an E.coli transformant carrying plasmid pEWSA86 (lac promotersystem) are displayed. Western blot analysis confirmed produc-tion of the shortened version of PBP2a-27r ( - 144 amino acids[left side of Fig. 2B]), and the penicillin-binding assay shownon the right side of Fig. 2B indicated the inability of thisprotein to bind penicillin. Similar analyses were performed foreach of the shortened versions of PBP2a-27r. Figure 3 sum-marizes the extent of N-terminal deletions within the trans-glycosylase-like domain of PBP2a-27r and penicillin-bindingresults. Removal of the membrane-spanning region (i.e., - 21amino acids) and addition of a chelating peptide did notdisrupt penicillin-binding activity. In contrast, removal of

A Competition AssayCefamandole (5gg/mi) Cofamandole (g±g/ml)

° °O coaN D" °l

Wild TypePBP2a_o

_BP2aB P 2PBP2a-245 a.a.

B g Western Penicillin-bindingAnalysis Assay

1 2 3 1 2 3

PBP2a PBP2a PBP2a-144 a PBP1 _

-144 a.a

VOL. 176, 1994

on March 23, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

446 WU ET AL.

portions of the putative transglycosylase domain (- 144 and-245 amino acids) or the entire domain (i.e., -341 aminoacids) eliminated penicillin-binding activity. Apparently, theregion of the protein which spans the membrane is notessential for proper folding of PBP2a-27r. The amino acidsequence following the transmembrane region, as well as theremainder of the presumed transglycosylase domain, is appar-ently critical for penicillin-binding activity.

Characteristics of PBP2a-27r from methicillin-resistant S.aureus 27r. PBP2a-27r encoded by mecA-27r is composed of anN-terminal membrane-spanning region (amino acids 1 to 22),a putative transglycosylase domain (approximately amino acids23 to 341), and a presumed transpeptidase domain (approxi-mately amino acids 342 to 668). Key motifs found in thetranspeptidase regions of other beta-lactam-binding proteins(SXXK, SXN, and KSG) were found within the transpeptidasedomain of PBP2a-27r. In the PBPs 1A, IB, 2, 3, and 5/6 fromE. coli, the average number of amino acids between conservedmotifs is as follows: SXXK to SXN, 66 amino acids; SXXK toKSG, 201 amino acids. E. coli PBP4 was not included in theaverages because it contains an atypical insertion of 188 aminoacids between the SXXK and SXN motifs (20). By comparison,the distance between the SXXK and SXN motifs in PBP2a-27ris 61 amino acids, and 195 amino acids lie between the SXXKand KSG motifs. Thus, the linear spacing of the conservedmotifs within the primary structure of PBP2a-27r was consis-tent with those of other characterized PBPs.The presence and appropriate spacing of these motifs do not

verify their importance in PBP2a's interaction with beta-lactamantibiotics. Therefore, each of these motifs was altered bysite-specific in vitro mutagenesis of the mecA-27r gene, and theresulting modified PBP2a-27r proteins were tested for penicil-lin-binding activity.

Site-specific mutagenesis of the conserved motifs within themec4-27r gene altered penicillin-binding activity. To test theimportance of the SXXK motif (STQK in PBP2a-27r, aminoacid residues 403-406), the DNA sequence of mecA-27r wasaltered by site-specific mutagenesis to create an S-403-->C-403mutation. In a parallel experiment, a K-406--R-406 mutationwas induced. The modified mecA-27r genes were subclonedinto E. coli expression vectors utilizing the veg promoter system(pEWSA79, S-403-->C-403, and pEWSA80 K-406-->R-406[Fig. 4]). DNA sequence analysis confirmed that only theintended DNA sequence changes had been incorporated in themecA-27r gene. Expression vectors containing the modifiedmecA-27r genes were transformed into E. coli DH5a compe-tent cells. Transformants were grown under conditions thatpermitted soluble expression of recombinant protein, andproduction of PBP2a-27r was confirmed by Western blotanalysis (Fig. 5). Assessment of penicillin-binding activity,illustrated immediately above the Western blot analysis resultsshown in Fig. 5, revealed complete elimination of penicillinbinding in both modified forms of PBP2a-27r (i.e., S-403-->C-403 and K-406--R-406). These results were consistent with thenotion that serine 403 is the site of acylation by the beta-lactamand that lysine 406 is also critical for penicillin-binding activitywithin this motif.

In a similar manner, residues within the SXN motif (SDN inPBP2a27r, amino acid residues 463--*465) were examined fortheir influence on penicillin binding. S-463 was replaced byC463 (subcloned into plasmid pEWSA83 [Fig. 4]) and N-465was replaced by Q-465 (subcloned into plasmid pEWSA84[Fig. 4]). Cell extracts prepared from E. coli transformantscarrying these plasmids (three transformants with pEWSA83and one transformant with pEWSA84) were examined for thepresence of the modified forms of PBP2a-27r by Western blot

pEWSA522- 21 a.a.

pEWSA792 84C

pEVSM82 K406'

pEWS3 S4463C

2N4pWSA42 N4SSQ

pEWSA642 859|C

MSR Trmug"losys Tispeptd.

M

CP .sr-.

CP ' VII

CP 7wcp 7

._

PenicillinI ln+

FIG. 4. Description of plasmids and mutations and summary ofresults of penicillin-binding experiments for alterations within con-served motifs. Plasmid designations are presented at the left followedby a brief description of alterations induced in PBP2a-27r by mutationsin the mecA-27r gene. The plasmid in the box has been describedpreviously (33). The mecA gene in plasmid pEWSA37 carries thenative S. aureus promoter. All other plasmids carry the veg promoterfrom B. subtilis, which enhances soluble expression of PBP2a-27r. Adepiction of the modified forms of PBP2a-27r is shown to the right ofthe mutations. STQK, SDN, and KSG indicate the relative positions ofthese conserved motifs within the putative transpeptidase domain,which is illustrated by the solid black area. The putative transglycosy-lase domain is indicated by the stippled area. The results of thepenicillin-binding assays are shown at the right: +, a detectable degreeof penicillin binding; -, no detectable penicillin binding. Abbrevia-tions: WT, wild-type PBP2a-27r; MSR, membrane-spanning region(diagonally hatched area); CP, chelating peptide (MGHWHH) [verti-cally hatched area]). The thin line represents the portion of themembrane-spanning region deleted from each modified PBP2a-27r.Superscripts: 1, plasmid carries the native S. aureus promoter formecA-27r; 2, plasmid carries the veg promoter from B. subtilis, themecA-27r gene has been modified to remove the transmembraneregionn, and a chelating peptide has been added to PBP2a-27r.

analysis, and then penicillin-binding activity was assessed (Fig.6). Both mutations (i.e., S-463--->C-463 and N-465-->Q-465)rendered PBP2a-27r incapable of binding penicillin. It shouldbe noted that another serine residue is present immediatelybefore S-463. In another experiment (results not shown), S-462was changed to K-462. Lysine was chosen because PBP3 of E.coli has a lysine residue in the analogous position. Analysis ofPBP2a-27r carrying the S-462-->K-462 alteration revealed nochange in its capacity to bind penicillin.

Finally, the results of altering the KSG motif (amino acids597--599) are illustrated in Fig. 7. In this experiment, S-598 ofPBP2a-27r was changed to C-598. The mutated mecA-27r genewas subcloned into plasmid pEWSA64 (Fig. 4). Extracts fromtwo E. coli transformants carrying pEWSA64 were examinedfor the effect the mutations had on PBP2a-27r. Western blotanalysis, summarized at the bottom of Fig. 7, once againconfirmed expression of the modified PBP2a-27r in thesetransformants. Assessment of penicillin-binding activity re-vealed that the S-598-->C-598 alteration in the KSG motifeliminated penicillin binding in this modified form of PBP2a-27r. Penicillin-binding results for all five of the mutationsdescribed above are summarized on the right side of Fig. 4.

J. BAcrERIOL.

on March 23, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

MUTAGENESIS OF THE mecA GENE 447

-PBP2a

FIG. 5. Site-specific mutagenesis of the SXXK motif destroyedpenicillin-binding activity. Cell extracts obtained from recombinant E.coli cells carrying plasmids pEWSA31 (WT, wild type with respect toSXXK motif), pEWSA79 (S-403-->C-403), and pEWSA80 (K-406- R-406), as well as an untransformed E. coli (DH5a) extract and amembrane preparation from methicillin-resistant S. aureus 447, werefirst analyzed for the presence of PBP2a by Western blot analysis. Theresults are summarized at the bottom of the figure; +, presence ofPBP2a; -, absence of PBP2a. Immediately above the Western blotanalysis is the penicillin-binding assay (125I-PenV) performed asdescribed by Wu and colleagues (33). The amino acid altered in eachmodified form of PBP2a-27r is indicated in parentheses above theplasmid designation.

DISCUSSIONThe interactions of beta-lactam antibiotics with PBPs in-

volved in bacterial cell wall biosynthesis have been studiedextensively, yet a detailed understanding of these interactionsat the molecular level remains unavailable. Ultimately, theresolution of these problems may require elucidation of thethree-dimensional structural details of PBPs. Armed with suchstructural knowledge, significant advances could be made inunderstanding how PBPs interact with beta-lactams. PBP2arepresents an intriguing target in such studies because itspresence in the staphylococci causes significant clinical prob-lems. A detailed understanding of the interaction of PBP2awith beta-lactam antibiotics may provide critical informationfor the design of new antibiotics effective in combating therising incidence of methicillin resistance.Removal of the transmembrane region of PBP2a-27r, which

anchors this protein to the exterior of the cell membrane as anectoprotein, did not significantly alter penicillin-binding activ-ity (33). Several other modifications made to the mecA-27rgene to examine the role of various regions and motifs inpenicillin binding are described in this report. Substitution of ashorter metal-chelating peptide in the position previouslyoccupied by the transmembrane region also caused no detect-able alteration in penicillin-binding activity in the modifiedPBP2a-27r. This modified form of PBP2a-27r was rapidlyrecovered from E. coli extracts (-95% pure) by chelatingpeptide-immobilized metal affinity chromatography in a singlestep (unpublished result). Conventional protein purificationprocedures required several steps to achieve this purity. Rapidpurification of an active form of PBP2a-27r will facilitatestructural analysis of this protein by X-ray crystallography.DNA sequence comparisons with other PBPs have sug-

ID 10 0

'ITt0 e 0 e

0

- -CI

, U7 U_ (J c U

m a a a¢m

1251-Pen VBindingAssay

PBP2a-_ 4-PBP2a

WesternAnalysis I+ m' + + + +I

FIG. 6. Penicillin-binding activity was destroyed when the SXNmotif was altered by site-specific mutagenesis. Cell extracts wereharvested from recombinant E. coli cells harboring plasmidspEWSA52 (WT, wild type with respect to SXN motif), pEWSA83(three isolates, each containing the S-463-->C-463 alteration),pEWSA84 (one isolate containing the N-465--Q-465 alteration),untransformed E. coli DH5a cells, and the methicillin-resistant S.aureus 447. These extracts were analyzed for the presence of PBP2a-27r protein by Western blot analysis. The results are summarized at thebottom of the figure: +, presence of PBP2a-27r; -, absence ofPBP2a-27r. The results of a penicillin-binding assay (125I-Pen V) arepictured above the Western blot analysis.

gested that PBP2a is composed of two domains (25, 28), withthe penicillin-interactive domain occupying the carboxyl regionof the protein. Crystallization of only the penicillin-interactivedomain might prove more successful because of the large sizeof the intact protein (-76 kDa). However, deletion analyses ofthe putative transglycosylase domain demonstrated that thisregion was also essential for penicillin-binding activity. Conse-quently, crystallization of a shortened version of PBP2a-27rwas not an option.

Analyses of amino acid sequences from PBPs and beta-lactamases have revealed conservation of certain amino acidmotifs in the penicillin-interacting portion or putativetranspeptidase domains of these enzymes (9, 18, 20). Crystal-lography data indicated that when these penicillin-interactingenzymes assume their native structure, these motifs arebrought into close proximity within the active-site cleft, wheresubstrate-enzyme inhibitor interactions occur (22, 29). Suchobservations have led to the suggestion that beta-lactamasesand PBPs share a common ancestry (9). The motifs of partic-ular interest include SXXK, SXN, and KSG.PBP2a-27r contains key active-site motifs of beta-lactamases

and PBPs in the carboxyl half of the protein, which representsthe presumed transpeptidase domain. In addition to the pres-ence of these motifs and localized sequence conservationsurrounding them, the linear spacing between motifs in PBP2a-27r was consistent with the spacing of the same motifs in PBPsfrom E. coli. Analysis of PBP2a from E. coli verified that theserine residue found within the SXXK motif of that enzymewas the site of acylation by beta-lactams (30). On the basis ofsequence homology with other PBPs, Song and colleagues (28)suggested that S-403 of PBP2a, within the SXXK motif, may bethe site of acylation by beta-lactams. The site-specific in vitromutagenesis described herein illustrates the importance of the

oW

Cm

1-PenV BJ uJ u

Binding 0. 0. 'UAssay

PBP2a -

WesternAnalysis + ++n. .+

VOL. 176, 1994

on March 23, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

448 WU ET AL.

a- U'%. v- xC' C4 0

M wU2 W

a coC.00t tI I

(A en

I- C44

3 32m UJ(003(DOuww

C

1251-Pen V

Binding Assay

*4 PBP2aPBP2a-_

~IWestern

D+ E AnalysisFIG. 7. Alteration of the serine in the KSG motif eliminated

penicillin-binding activity. Cell extracts were harvested from recombi-nant E. coli cells harboring plasmids pEWSA31 (WT, wild type withrespect to KSG motif), pEWSA64 (two isolates, each containing theS-598-*C-598 alteration), untransformed E. coli DH5a cells, andmethicillin-resistant S. aureus 447. These extracts were analyzed for thepresence of PBP2a-27r protein by Western blot analysis. The resultsare summarized at the bottom of the figure: +, presence of PBP2a-27r;-, absence of PBP2a-27r. The results of the penicillin-binding assay

(25I-Pen V) are pictured above the Western blot analysis.

SXXK motif in PBP2a-27r. When the S-403-->C-403 mutationwas induced in PBP2a-27r, penicillin-binding activity was no

longer detectable, consistent with the suggestion that S-403may be the site of acylation by beta-lactams. Evidence fromX-ray crystallography of DD-carboxypeptidase and a class Cbeta-lactamase indicated that the lysine of the SXXK motifand the lysine of the KSG motif interact through hydrogenbonding with beta-lactams, perhaps coordinating its positionwithin the active-site cleft (11, 22). When lysine was changed toarginine in the SXXK motif in a class A beta-lactamase fromStreptomyces albus, the rate constant of acylation was greatlyreduced (4). When K-406 was changed to R-406 in PBP2a-27r,penicillin-binding activity was lost once again. Loss of penicil-lin-binding activity when K-406 was changed to another basicamino acid (R-406) in the SXXK motif of PBP2a-27r suggestsa very specific spatial requirement for the positive charge oflysine in its interaction with beta-lactams.

It has been argued that the KSG motif in combination withthe lysine residue of the SXXK motif forms a hydrogen-bonding network that coordinates the position of the C-3carboxyl group of beta-lactams (11, 22). The precise functionof the SXN motif remains to be elucidated, although homologywith beta-lactamases suggests that it is involved in deacylationof a bound beta-lactam (22). Alterations of the SXN and KSGmotifs of PBP2a-27r also resulted in a loss of penicillin-bindingactivity, providing evidence that these motifs are intimatelyinvolved in the interaction of beta-lactams with PBP2a-27r. Itshould be noted that a number of other mutations whichaltered the amino acid sequence of PBP2a, made in vitro or

observed in mecA genes of other clinical isolates of S. aureus,did not alter penicillin-binding activity (data not shown).These results demonstrate that the putative transpeptidase

domain of PBP2a-27r interacts directly with beta-lactam mol-

ecules at the anticipated sites predicted from previous work onpenicillin-interactive enzymes. However, unequivocal demon-stration of PBP2a's capacity to synthesize cell wall remains tobe seen, and its reduced affinity for beta-lactam antibioticsremains a mystery.

REFERENCES1. Brown, D. F. J., and P. E. Reynolds. 1980. Intrinsic resistance to

13-lactam antibiotics in Staphylococcus aureus. FEBS Lett. 122:275-278.

2. de Jonge, B., and A. Tomasz. 1993. The muropeptide compositionof the peptidoglycan of Staphylococcus aureus determined by highperformance liquid chromatography, p. 77-82. In M. A. de Pedro,J.-V. Holtje, and W. Loffelhardt (ed.), Bacterial growth and lysis.Plenum Publishing Corp., New York.

3. de Jonge, B. L. M., Y. S. Chang, D. Gage, and A. Tomasz. 1992.Peptidoglycan composition of a highly methicillin-resistant Staph-ylococcus aureus strain. J. Biol. Chem. 267:11248-11254.

4. Gibson, R. M., H. Christensen, and S. G. Whaley. 1990. Site-directed mutagenesis of 3-lactamase I. Biochem. J. 272:613-619.

5. Hartman, B. J., and A. Tomasz. 1984. Low-affinity penicillin-binding protein associated with 3-lactam resistance in Staphylo-coccus aureus. J. Bacteriol. 158:513-516.

6. Hayes, M. V., N. A. C. Curtis, A. W. Wyke, and B. J. Ward. 1981.Decreased affinity of a penicillin-binding protein for b-lactamantibiotics in a clinical isolate of Staphylococcus aureus resistant tomethicillin. FEMS Microbiol. Lett. 10:119-122.

7. Horan, T. C., W. R. White, T. G. Jarvis, and D. H. Emori. 1984.Nosocomial infection surveillance. Morbid. Mortal. Weekly Rep.35:17SS-19SS.

8. Jevons, M. P. 1961. Celbenin-resistant staphylococci. Br. Med. J.1:124-125.

9. Joris, B., J.-M. Ghuysen, G. Dive, A. Renard, 0. Dideberg, P.Charlier, J.-M. Frere, J. A. Kelly, J. C. Boyington, P. C. Moews,and J. R. Knox. 1988. The active-site-serine penicillin-recognizingenzymes as members of the Streptomyces R61 DD-peptidasefamily. Biochem. J. 250:313-324.

10. Kasher, M. S., M. Wakulchik, J. A. Cook, and M. C. Smith. 1993.One-step purification of recombinant human Papillomavirus type16 E7 oncoprotein and its binding to the retinoblastoma geneproduct. BioTechniques 14:630-641.

11. Kelly, J. A., J. R. Knox, Z. Haiching, J.-M. Frere, and J.-M.Ghuysen. 1989. Crystallographic mapping of beta-lactams boundto a D-alanyl-D-alanine peptidase target enzyme. J. Mol. Biol.209:281-295.

12. Kreiswirth, B., J. Kornblum, R. D. Arbeit, W. Eisner, J. N.Maslow, A. McGeer, D. E. Low, and R. P. Novick. 1993. Evidencefor a clonal origin of methicillin resistance in Staphylococcusaureus. Science 259:227-230.

13. Kunkel, T. A. 1985. Rapid and efficient site-specific mutagenesiswithout phenotypic selection. Proc. Natl. Acad. Sci. USA 82:488-492.

14. Labischinski, H. 1992. Consequences of the interaction of j-lac-tam antibiotics with penicillin-binding proteins from sensitive andresistant Staphylococcus aureus strains. Med. Microbiol. Immunol.181:241-265.

15. Labischinski, H., M. Hochberg, T. Sidow, H. Maidhof, U. Henca,B. Berger-Bachi, and J. Wecke. 1993. Towards an understandingof the fine structure of the sacculi from Escherichia coli andStaphylococcus aureus, p. 9-21. In M. A. de Pedro, J.-V. Holtje,and W. Loffelhardt (ed.), Bacterial growth and lysis. PlenumPublishing Corp., New York.

16. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecularcloning: a laboratory manual. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.

17. Maple, P. C., J. M. Hamilton-Miller, and W. Brumfitt. 1989.Worldwide antibiotic resistance in methicillin-resistant Staphylo-coccus aureus. Lancet i:537-540.

18. Matsuhashi, M., M. D. Song, F. Ishino, M. Wachi, M. Doi, M.Inoue, K. Ubukata, N. Yamashita, and M. Konno. 1986. Molecularcloning of the gene of a penicillin-binding protein supposed tocause high resistance to 1-lactam antibiotics in Staphylococcusaureus. J. Bacteriol. 167:975-980.

J. BACrERIOL.

on March 23, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

MUTAGENESIS OF THE mecA GENE 449

19. Miller, J. R., S. Kovacevic, and L. E. Veal. 1987. Secretion andprocessing of staphylococcal nuclease by Bacillus subtilis. J. Bac-teriol. 169:3508-3514.

20. Mottl, H., P. Terpstra, and W. Keck. 1991. Penicillin-bindingprotein 4 of Escherichia coli shows a novel type of primarystructure among penicillin-interacting proteins. FEMS Microbiol.Lett. 78:213-220.

21. Murakami, K., K. Nomura, M. Doi, and T. Yoshida. 1987.Production of low-affinity penicillin-binding protein by low- andhigh-resistance groups of methicillin-resistant Staphylococcus au-reus. Antimicrob. Agents Chemother. 31:1307-1311.

22. Oefner, C., A. I'Arcy, J. J. Daly, K. Guernator, R. L. Charnas, I.Heinze, C. Hubschwerien, and F. K. Winkler. 1990. Refined crystalstructure of beta-lactamase from Citrobacter freundii indicates amechanism for beta-lactam hydrolysis. Nature (London) 343:284-288.

23. Preston, D. A., C. Y. E. Wu, L. C. Blaszczak, D. E. Seitz, and N. G.Halligan. 1990. Biological characterization of a new radioactivelabeling reagent for bacterial penicillin-binding proteins. Antimi-crob. Agents Chemother. 34:718-721.

24. Reynolds, P. E. W., and C. Fuller. 1986. Methicillin-resistantstrains of Staphylococcus aureus: presence of identical additionalpenicillin-binding protein in all strains examined. FEMS Micro-biol. Lett. 33:251-254.

25. Ryffel, C., W. Tesch, I. Birch-Machin, P. E. Reynolds, L. Barberis-Maino, F. H. Kayser, and B. Berger-Bachi. 1990. Sequencecomparison of mecA genes isolated from methicillin-resistant S.aureus and S. epidermidis. Gene 94:137-138.

26. Samson, S. M., J. E. Dotzlaf, M. L. Slisz, G. W. Becker, R. M. VanFrank, L. E. Veal, W.-K. Yeh, J. R. Miller, S. W. Queener, andT. D. Ingolia. 1987. Cloning and expression of the fungal expan-dase/hydroxylase gene involved in cephalosporin biosynthesis.Bio/Technology 5:1207-1214.

27. Smith, M. C., J. A. Cook, P. A. Smanik, M. Wakulchik, and M. S.

Kasher. 1992. Kinetically inert Co(III) linkage through an engi-neered metal binding site: specific orientation of recombinanthuman papillomavirus type 16 E7 protein on a solid support.Methods (Orlando) 4:73-78.

28. Song, M. D., M. Wachi, M. Doi, F. Ishino, and M. Matsuhashi.1987. Evolution of an inducible penicillin-target protein in methi-cillin-resistant Staphylococcus aureus by gene fusion. FEBS Lett.221:167-171.

29. Spratt, B. G., and K. D. Cromie. 1988. Penicillin-binding proteinsof Gram-negative bacteria. Rev. Infect. Dis. 10:699-711.

30. Takasuga, A., H. Adachi, F. Ishino, M. Matsuhashi, T. Ohta, andH. Matsuzawa. 1988. Identification of the penicillin-binding site ofpenicillin-binding protein 2 of Escherichia coli. J. Biochem. 104:822-826.

31. Unal, S., J. Hoskins, J. E. Flokowitsch, C. Y. E. Wu, D. A. Preston,and P. L. Skatrud. 1992. Detection of methicillin-resistant staph-ylococci by using the polymerase chain reaction. J. Clin. Microbiol.30:1685-1691.

32. Wilkinson, B. J., K. J. Dorian, and L. D. Sabath. 1978. Cell wallcomposition and associated properties of methicillin-resistantStaphylococcus aureus strains. J. Bacteriol. 136:976-982.

33. Wu, C. Y. E., J. Hoskins, L. C. Blaszczak, D. A. Preston, and P. L.Skatrud. 1992. Construction of a water-soluble form of penicillin-binding protein 2a from a methicillin-resistant Staphylococcusaureus isolate. Antimicrob. Agents Chemother. 36:533-539.

34. Wu, C. Y. E., L. C. Blaszczak, M. C. Smith, and P. L. Skatrud.Construction of a modified penicillin-binding protein 2a frommethicillin-resistant Staphylococcus aureus and purification byimmobilized metal affinity chromatography. Submitted for publi-cation.

35. Wyke, A. W., J. B. Ward, and M. V. Hayes. 1982. Synthesis ofpeptidoglycan in vivo in methicillin-resistant Staphylococcus au-reus. Eur. J. Biochem. 127:553-558.

VOL. 176, 1994

on March 23, 2021 by guest

http://jb.asm.org/

Dow

nloaded from