Resistance Escherichia to Penicillins: Identification of ... · experiment, P4xa2(ampA2,ampC+,metB)was used as donor and TP3 was used (ampA1, ampC3, ilv) as recipient. Alinkage analysis

JOURNAL OF BACTERIOLOGY, OCt. 1973, p. 123-130Copyright 0 1973 American Society for Microbiology

Vol. 116, No. 1Printed in U.S.A

Resistance of Escherichia coli to Penicillins:Identification of the Structural Gene for the

Chromosomal PenicillinaseLARS G. BURMAN,I JAMES T. PARK,2 E. BORJE LINDSTROM, AND HANS G. BOMAN

Department of Microbiology, University of Umed, S-901 87 Umed, Sweden

Received for publication 13 June 1973

A screening procedure was used to isolate a number of mutants of Escherichiacoli K-12 with low penicillinase activity. By co-transduction with purA, three ofthe mutants were found to map near 82 min. Penicillinase was purified from onemutant and from a transductant with a temperature-sensitive enzyme. Compari-son with wild-type penicillinase revealed similarities in the Ouchterlony im-munodiffusion test but differences in the catalytic properties. It is concluded thatthe mutations have occurred in the structural gene of the chromosomalpenicillinase (designated ampC). Purified enzyme and a temperature-sensitivemutant were used to investigate whether the penicillinase has a physiologicalfunction related to biosynthesis or breakdown of murein. No positive evidence forany such function was obtained.

High levels of chromosomally mediated peni-cillin resistance are the result of several consec-utive mutations. In Escherichia coli K-12 thefirst step found was a mutation in the ampAgene which increased the penicillinase activity(8). Further work showed that the enzyme waslocated in the periplasm (2) and that ampAmapped at 82 min on the chromosome (7).Purification of the penicillinase from differentstrains has shown that ampA strains containabout 10 times more enzyme than wild-typestrains (14). No qualitative differences could befound between enzyme purified from ampA andfrom wild-type strains. Taken together, thesedata suggest that ampA is a regulatory gene forthe penicillinase. Assuming this to be correct, itwould be desirable to find the structural genefor the chromosomal penicillinase.The present study was therefore initiated

with the following chief aims: (i) to isolate andmap penicillinase-negative mutants; (ii) tocompare the penicillinases purified from wild-type and mutant strains with respect to en-zymic and immunological properties; and (iii)to use both a mutant with a heat-labile penicil-linase and purified enzyme to investigatewhether a previously assumed natural function

1 Present address: Department of Clinical Bacteriology,University of Umel, S-901 87 Umed, Sweden.

2Present address: Department of Molecular Biology andMicrobiology, Tufts University School of Medicine, Boston,Mass. 02111.

12.

of the penicillinase (2, 5) is linked to thebiosynthesis of murein. For the last purpose itwas found necessary to modify our method ofpurification of the penicillinase (14) in such away that the use of lysozyme for making sphero-plasts was avoided.The results show that we obtained mutants in

the structural gene of the penicillinase. Thisnew locus, designated ampC, is closely linked toampA at 82 min on the chromosomal map.(ampB has been used for class II mutations, onplates giving a twofold increase in chromosom-ally as well as episomally mediated penicillinresistance [16].)

MATERIALS AND METHODSOrganisms. The bacterial strains used were all

derivatives of E. coli K-12 (see Tables 1 and 2). Thefollowing Hfr strains have been used: Gllal, whichcarries the original ampAl allele; P4xa2, which isassumed to carry the ampA2 allele; and R12, which isa transductant carrying the ampAl allele. These Hfrstrains transfer their chromosomal genes in the follow-ing order: for Gllal origin-purE-thr-ampA, for P4xa2origin-lac-thr-ampA, and R12 origin-metB-ampA-thr.

Phage Plbt used for transduction was described byGross and Englesberg (9). High titer stocks wereprepared as described before (7).Media and growth conditions. The minimal me-

dium used was medium E (23). It was supplementedwith 0.2% glucose, 1 gg of thiamine per ml, 25 jug ofthe L-isomer of the required amino acids per ml, and

on Novem

ber 22, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

TABLE 1. Properties of the reference strains of Escherichia coli K-12a

Strain ampA Sex Str Relevant markers Source orallele reference

D21 1 F- R pro, trp, hisGll + Hfr S metB, ilu G. StentGllal 1 Hfr S metB, ilv, 8Hi + F- R thr, leu, proA, purA86 P. G. de HaanKG20 1 F+ R pro, his, arg, purA 16P4xa2 2 Hfr S metB 8PA2004 + F- R thr, leu, his, pyrB R. LavalleR12 1 Hfr S metB 16

a All strains except HI carry X prophage. Gene abbreviations according to Taylor (22) with known relevantallele numbers. Str denotes response to streptomycin: R, resistant; S, sensitive.

TABLE 2. Parental and penicillinase-negative strains of Escherichia coli K-12

Genotype Resistance to

Strain Parent or Tusedp Penicillinase Commentsrecipient ampA ampC (C) PenG D-amp activityaampC (C) Agml (gIl

R12 Ri 1 + 37 200 20 3.30 See Table 1UM104 R12 1 12 30 7.5 2 0.50 Hfr lostUM1041 UM104 1 12 30 7.5 2 0.50 strAD21 D2 1 + 37 250 25 4.60 See Table 1TP1 D21 1 1 37 5 1 0.01TP3 D21 1 3 37 5 1 0.90 ilvHi - + + 37 10 2 0.40 See Table 1Hlt3 Hi 1 + 37 200 25 5.60 See textHltl4 Hlt3 1 12 30 15 2 0.55 See Table 3- - - - 1 12 42 2.5 1 0.01

a Expressed as units per milligram dry weight of organisms. One unit of penicillinase was defined as theamount of enzyme that hydrolyzed 1 Amol of benzyl penicillin per h in 0.05 M potassium phosphate buffer (pH7.4) at 37 C. The background due to spontaneous hydrolysis of benzyl penicillin in a standard incubationcorresponds to about 0.004 units.

25 tsg of adenine per ml, when necessary. The com-plete medium used was LB (1) supplemented withmedium E (23) and 0.2% glucose. All plates contained1.5% agar. LB plates, with or without ampicillin,contained LB supplemented with 0.2% glucose, 10-3M CaCl2, and a vitamin mixture (7).

For purification of penicillinase, 15- to 60-literbatches of the various strains were grown at 37 C asdescribed previously (14). Strain Hiti4 was grown at30 C and harvested during log phase. The latter wasnecessary to avoid the formation of capsular materialthat made harvesting difficult. Growth was followedby recording optical density by using a Klett-Sum-merson colorimeter with filter W66. A reading of 100Klett units corresponds to 4 x 108 log-phase cells perml of LB.

Materials. Benzyl penicillin (PenG) and D-ampicillin (D-amp) were kindly provided by ABAstra, Sodertalje, Sweden. Streptomycin sulfate wasdonated by AB Kabi, Stockholm, Sweden. Zul-kowsky's starch (used for penicillinase determination)was from Merck, Darmstadt, Germany. N-methyl-N-nitroso-N-nitroguanidine (NNG) was from K & KLaboratories, Inc., Hollywood, Calif. Materials forchromatography were as described previously (14).

Procedure for mutagenesis. Cells growing expo-nentially in LB, or starved, stationary-phase cells (4x 108 cells/ml) were chilled, harvested, and washedwith 0.9% NaCl. After resuspension to the originalvolume in a freshly prepared solution of mutagen(NNG, 400,g/ml in 0.05 M citrate buffer, pH 5.5), thebacteria were incubated at 37 C for 30 min and thenwashed twice with 0.9% NaCl before being spread onLB plates. This procedure yielded 10 to 20% survivors.

Analytical methods. Protein content was meas-ured at 280 nm. The penicillinase activity of wholecells or fractions was assayed in 0.05 or 0.1 Mphosphate buffer, pH 7.4, by the automated microio-dometric method (15). PenG was normally used assubstrate because it is hydrolyzed 20 to 30 timesfaster than is ampicillin (14). Values for Km were ob-tained graphically from Lineweaver-Burk plots of thereaction velocities at 37 C (also for strain Hltl4).When assaying activity of whole cells, 1,000 jug of sub-strate per ml was used to give a good saturation ofthe penicillinase (4).

Determination of antibiotic resistance. The bac-teria to be tested were harvested during exponentialgrowth in LB and diluted in 0.9% NaCl. Cells(100-400) were spread on LB plates containing differ-

124 BURMAN ET AL. J. BACTERIOL.

on Novem

ber 22, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

STRUCTURAL GENE FOR E. COLI PENICILLINASE

ent concentrations of the antibiotic. The resistancelevel was defined as the maximal antibiotic concen-tration permitting 100% survival of the cells. Refer-ence strains were included in each determination (16).

RESULTSIsolation of penicillinase-negative

mutants. Penicillinase-negative mutants havepreviously been isolated in strains of Bacillus(6) and Staphylococcus (17) that produce highlyactive exopenicillinases. The screening methodsemployed permitted an identification on theplates of clones lacking penicillinase. When thetechnique was employed on our E. coli strainswith a cell-bound penicillinase, no mutantswere detected. This was probably due to the lowtumover number of the enzyme (14) as well asto the fact that the substrates do not readilypenetrate the outer membrane barrier of E. coli(4, 21).

Penicillinase-negative mutants were isolatedfrom two different parental strains, D21 (F-)and R12 (Hfr Reeves 1), both of which carriedthe ampAl allele giving class I type of ampicil-lin resistance (growth on plates with 20-25,ug/ml). After NNG mutagenesis (see Materialsand Methods), the cells were spread on LBplates and incubated for about 16 h at 30 C (topermit also the isolation of temperature-sensi-tive mutants). Only those colonies which grewovernight and which then did not grow whenreplicated onto LB plates at 30 C with anampicillin concentration of 5 ,ug/ml were se-lected (or 2 jgg/ml for plates incubated at 42 C).The clones selected were grown up in LB andassayed for their penicillinase activity by usingthe automated iodometric method (15).

After screening of about 20,000 colonies inthis way, we obtained a few hundred withdecreased ampicillin resistance. Of these, only10 to 15 were classified as penicillinase negative.Those selected for further studies in this com-munication are listed in Table 2. The onlytemperature-sensitive mutant obtained,UM104, had lost the Hfr property of the paren-tal strain. This is in agreement with earlierobservations that the Hfr property of R12 issomewhat unstable (16). Strain TP3 is a doublemutant with a new requirement for isoleucineand valine (ilv). Many mutants (like TP1) hadpenicillinase activities that were significantlylower than in the wild type. However, of thoseselected for purification, TP3 had more enzymeactivity than the wild type, and UM104 pre-sumably a temperature-labile enzyme.Conjugation and transduction mapping of

the ampC gene. In a preliminary conjugationexperiment, P4xa2 (ampA2, ampC+, metB) was

used as donor and TP3 was used (ampA1,ampC3, ilv) as recipient. A linkage analysis ofilv +/strA recombinants showed that P4xa2transferred ampC before metB. In another in-terrupted mating experiment we used as donorGllal (ampAl, ampC+). The recipient,UM1041, carried the ampAl and the ampC12allele, the latter mediating a heat-labile penicil-linase activity and at 30 C reducing the am-picillin resistance and penicillinase activity towild-type levels (Table 2). As a control, asimultaneous mating experiment was per-formed with Gllal and PA2004 (ampA+,ampC+). From the result given in Fig. 1, it canbe concluded that ampC12 is linked to ampA.In both experiments the time of entrance wasalso in reasonable agreement with the knownmap position of ampA at 82 min (7).Transduction experiments were first per-

formed by using strain TP3 (ampAl, ampC3,purA+) as donor and strain KG20 (ampAl,ampC+, purA) as recipient. PurA+ transduc-tants were selected, purified, and tested forampicillin resistance. Since both donor andrecipient carried the same allele of ampA,ampicillin sensitivity (growth on plates with 1-2jAg/ml, no growth on 5 ,ug/ml) was taken asevidence for co-transduction of ampC andpurA+. The results in Table 3 show that, withKG20 as recipient, ampC3 was co-transducible

-St 160

x

i,, 12Go

CLEn0-

oc 8

E0to' 4

E

00 10 20 30 40

Time of sampling (min)FIG. 1. Parallel experiments with chromosome

transfer using strain Gl lal (ampAl, ampC+) as donorand strains PA2004 (ampA+, ampC+) and UM1041(ampAl, ampC12) as recipients. The same culture ofGllal (a derivative of Hfr Cavalli) was used as sourceof donor bacteria. Ampicillin-resistant (Amp-r) re-combinants were selected on LB plates with 10 pg ofampicillin/ml.

VOL. 116, 1973 125

on Novem

ber 22, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

BURMAN ET AL.

TABLE 3. Co-transduction of ampC and ampA + withpurA +a

No. of Amp-s, Pur+

Donor Recipient Pur+ clonesclonestested No. %

TP3 (ampAl, KG20 (purA) 60 16 27ampC3)

TP1 (ampAl, H1t3 (purA86) 78 17 22ampCl)

TP3 (ampAl, H1t3 (purA86) 65 23 35ampC3)

UM104 (ampAl, Hlt3 (purA86) 28 7 25ampC12)

Gll (ampA+, Hlt3 (purA86) 41 10 24ampC+)

a Both recipients contained ampAI, ampC+ and purA. Alltransductants were purified once and tested for their respec-tive amino acid requirements. Ampicillin sensitivity (Amp-s)was first scored by replica plating and then confirmed bysingle-cell resistance tests. Amp-s was defined as growth onampicillin concentrations of 1 to 2 lAg/ml and no growth at 5Ag/ml. The transductant Hltl4 (used for the isolation of theampC12 enzyme) was isolated in a pilot experiment withUM104 as donor and Hlt3 as recipient.

with purA+ with a frequency of 27%. However,KG20 and all strains with the same allele ofpurA show poor growth. Further co-transduc-tion experiments were therefore performed withstrain Hlt3, a transductant obtained fromstrain Hi (ampA+, ampC+, purA86), by usingGllal (ampAl, ampC+, purA+) as donor andselecting an ampicillin-resistant recombinantwhich was still purA86. Strain Hlt3 (ampAl,ampC+, purA86) was used as recipient in fur-ther transduction experiments in which purA+transductants were selected. The results inTable 3 show that the ampCl and ampC12alleles were also co-transducible with purA. Thefrequencies were of the same orders as found forpurA+ and ampA+, but the numbers wereconsidered too low to provide map distances.However, the results in Table 3 show a closelinkage between ampA, ampC, and purA, and avery close linkage between ampA and ampC.

Purification of penicillinases from strainswith different alleles of ampC. The penicillin-ase of E. coli is an entirely cell bound enzymethat is located in the periplasm (2, 14). There-fore, step 1 in our first method of penicillinasepurification was the release of the enzyme byformation of spheroplasts using ethylenedia-minetetraacetic acid and lysozyme. However,when testing a purified penicillinase prepara-tion for lytic activity against Micrococcus lyso-deikticus, considerable lysis occurred. This ac-tivity was found to be more heat stable than thepenicillinase activity and was therefore attri-buted to contaminating lysozyme. On this basis,

steps 1 and 3 in the purification procedure weremodified as follows.

Step 1. The bacteria were cultivated andharvested as described above. The cells werewashed once in ice-cold 0.01 M phosphatebuffer, pH 6.8, and in all subsequent steps thetemperature was 0 to 5 C. The bacterial paste(50-75 g wet weight) was resuspended in about100 ml of phosphate buffer and disrupted in aFrench press. After addition of 10 ml of 0.1 Methylenediaminetetraacetic acid, the suspen-sion was diluted to about 200 ml and cen-trifuged for 30 min at 48,000 x g. The superna-tant fluid was collected. The pellet was passedonce more through the French press and thesupernatant fluid was collected as above.Step 2. The total supernatant fluid was

dialyzed against distilled water overnight andthen equilibrated against 0.01 M phosphatebuffer, pH 6.8.

Step 3. The extract was applied to a column(30 by 850 mm) with 600 ml of sulfoethylcellulose (14). Elution was carried out with alinear gradient made up from 200 ml each of0.01 M and 0.25 M potassium phosphate buff-ers, pH 6.8. The material that showed penicil-linase activity was collected and dialyzedagainst 0.1 M potassium phosphate buffer, pH6.8 (fraction A).

Step 4. Fraction A was applied to a hydrox-ylapatite column (32 by 475 mm) and eluted bya gradient of phosphate buffer (14). The sam-ples that showed activity were collected,dialyzed against water, and lyophilized (frac-tion B).

Fraction B was isolated from strains Gllal(ampC+), TP3 (ampC3), and Hltl4 (ampC12)and tested for purity on polyacrylamide-gelelectrophoresis. The results in Fig. 2 show thatthe ampC3 protein from the mutant TP3 was aspure as the penicillinase from strain Gllal andthat both enzymes had the same electrophoreticmobility. Thus, the modified purification proce-dure gives as homogeneous a product as doesthe previous method (14). The material fromthe transductant Hltl4 (ampC12) gave a dis-tinct band with a slower electrophoretic mobil-ity than the protein from strain Gllal. With theampC12 enzyme (no. 3 in Fig. 2), material wasalso trailed from the origin. This protein maytherefore not be stable at pH 4.7.The yields of penicillinase were almost the

same for strains Gllal and TP3 (0.1 mg ofprotein/g of wet bacterial paste) but were fivetimes lower for strain Hltl4. Attempts were alsomade to purify the penicillinase from the mu-tant TP1. However, despite a very low butmeasurable penicillinase activity, the yields

126 J. BACTERIOL.

on Novem

ber 22, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

STRUCTURAL GENE FOR E. COLI PENICILLINASE

were found to be too low.Characterization of purified penicillinase.

Table 4 gives the Km values for two substratesusing the penicillinases purified from strainsTP3 and Hltl4. We have also included earlierKm values for wild-type enzyme from strainsGll and Gllal (14) as well as the ampicillinresistances at 37 C for all four strains. Com-pared to the penicillinases from strains Gllaland Hltl4, the ampC3 enzyme from the mutantTP3 showed significantly higher Km values forPenG and D-amp.

For the ampC3 enzyme from strain TP3,Vma, was about three times lower than the valuefor the wild-type enzyme from strain Gllal.The low Vmax obtained for the wild-type enzyme

Start

FIG. 2. Electrophoresis of penicillinase purifiedfrom the following strains: 1, Gllal (ampC+); 2, TP3(ampC3); 3, Hltl4 (ampC12); 4, a mixture of I and 2;5, a mixture of 1 and 3. To gels I to 3 was layered 100,gliters of the respective samples, each containingabout 100 Ag of protein and 8% (wt/vol) sucrose toincrease the density. To gels 4 and 5 was added 100ILliters of a mixture made up of equal amounts of therespective proteins including 8% (wt/vol) sucrose.Electrophoresis was performed in 7.5% acrylamide at4 C for 2 h at 100 Vand 6 to 8 mA per tube at pH4.7.TABLE 4. Enzymic characteristics of wild-type and

mutant penicillinasesa

Source of enzyme Resistance Km (uM)~~to D-amp 0UmgxStrain ampA ampC (ag/mi U/mg) PenG D-amp

Gll + + 1-2 26.7 12 5Gllal 1 + 15-20 47.9 12 6TP3 1 3 1 14.4 1,000 250Hltl4 1 12 2 61.3 8

a One unit of penicillinase was defined in footnote a, Table2. The specific activity of the peak fraction from step 4 wasdetermined at 25 C by using 0.5 mM penicillin G as sub-strate. Vm.ax was calculated from these data and the corre-sponding Km values. Protein was determined by using 21.0 asextinction coefficient (14).

from strain Gll is explained by the fact thatthis enzyme was never obtained in a homogene-ous form (14).

Antibodies against pure wild-type penicillin-ase from strain Gllal (14) were used to testwhether the enzymes isolated from strains TP3(ampC3) and Hltl4 (ampC12) were immuno-logically related to the wild-type, ampC+ en-zyme from Gllal. The Ouchterlony test in Fig.3 revealed basic similarities among all threepenicillinase preparations.

Pollock was the first to use antisera to demon-strate differences between penicillinases (19).Figure 4 shows such titrations of penicillinasepurified from strains Gllal, TP3, and Hltl4.To a constant amount of enzyme protein wereadded different amounts of antibodies whilefollowing the enzyme activity. The ampC3 en-zyme from the mutant TP3 was more sensitiveto antibody inhibition than was the wild-type,ampC+ enzyme from strain Gllal. The slope ofthe curve for the ampC12 enzyme from thetransductant Hltl4 was different from that ob-tained for the wild-type enzyme from strainGllal, but the level of maximum inhibition wasalmost the same.The temperature stability of an enzyme is a

relatively specific property. The transductantHltl4 was suspected to possess a heat-labilepenicillinase mediated by the ampC12 allele.Heat inactivation at 23 and 44 C was thereforetried with enzyme purified from Hltl4 by usingwild-type enzyme from Gllal as a control.Figure 5 shows that the ampC12 enzyme lostalmost all activity after 5 min of incubation at44 C. The wild-type enzyme from strain Gllalwas fairly stable at this temperature for at least27 min, as were both proteins at 23 C.

Fingerprinting (12) of enzymes from strainsGllal, TP3, and Hltl4 was performed to detectpossible amino acid replacements. However, noclear differences in the peptide patterns couldbe detected.Search for a physiological role of

penicihlinase. All the penicillinase-negativemutants studied appeared normal with respectto growth curves, generation times, rate ofautolysis, cell morphology, cell volume, andenvelope permeability, the latter measured asresistance to cholate lysis (4). Thus, loss ofpenicillinase activity did not alter the grossproperties of the bacteria.Except for penicillinase, all known enzymes

that interact with penicillins are connected withmurein (10, 13, 25). Strain Hltl4, which has aheat-labile penicillinase, was selected to furtherstudy its murein biosynthesis. These experi-ments were performed at various temperatures

VOL. 116, 1973 127

on Novem

ber 22, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

BURMAN ET AL.

FIG. 3. Immunodiffusion analysis of penicillinasepurified from different strains. The central well (A)contained rabbit antibodies against purified wild-type enzyme from strain Gllal. The diffusion wasperformed in 1% (wt/vol) Noble agar (Oxoid) contain-ing 0.05 M tris(hydroxymethyl)aminomethane-hydrochloride, pH 8.6, at 23 C for 16 h.

EC 100

Ht4(

o Gllal (ampC+)

>50

E TP3 (amDC3)N

° O 11 1

3 0 5 10

& Antibody concentration (mg/ml)FIG. 4. Antibody neutralization of penicillinase

purified from different strains. To samples with 25,uliters of enzvme (10 jg/ml) was added 25 uliters ofvarious dilutions of a stock solution of antibodies (100mg of protein/ml). After 10 min of incubation at 23 C,penicillin G was added, and after 5 min of additionalincubation at 23 C, the penicilloic acid formed wasassayed (15). Wild-type enzyme from strain Gllal(prepared by the previous method of purification) wasused for obtaining rabbit antisera from which penicil-linase antibodies were purified (14).

(30-45 C), and the almost isogenic strain, Hlt3,served as control. Murein synthesis, measuredas incorporation of 3H-diamino pimelic acid(20), was normal in Hltl4 during growth in LB,casein hydrolysate, and glucose minimal me-

dium, as well as during incubation in a "mureinmedium" containing Tris buffer, glucose,MgCl2, murein amino acids, and chlorampheni-col. The murein formed showed a normal, veryslow turnover and a normal, 50% degree ofcross-linkage (24).Murein sacculi with their lipoprotein append-

ages (3) were prepared from Hlt3 and Hltl4cultivated in LB at 30 and 42 C. A Beckmanamino acid analyzer was used to determine thecontent of amino acids and glucosamine. Nodifferences were detected with respect tomurein composition or the ratio of murein tolipoprotein (3).The possibility that penicillinase is an auto-

lytic enzyme was investigated by using penicil-linase purified without lysozyme (see above).The preparation had no lytic effect on M.lysodeikticus cells. When added to purified E.coli murein sacculi, no decrease in opticaldensity (660 nm) or appearance of reducing oramino groups was seen during long incubationsand various conditions. Thus, E. coli penicillin-ase did not hydrolyze glycosidic or peptidebonds in E. coli murein under the conditionsused.Lysozyme-produced muropeptides (24) and

uridine 5'-diphosphate-N-acetyl-muramyl-pen-tapeptide (18) purified from Bacillus cereuswere tested as possible substrates for penicillin-ase. None of the enzymic activities ascribed tothe E. coli autolytic system (lysozyme, en-dopeptidase, N-acetyl-muramyl-L-alanine ami-

In._c E

*_ ,

= U%

z, o= o

> 0ipC4,0

100

50

0 5 10Time of Pre-incubation (min)

FIG. 5. Heat inactivation of penicillinases purifiedfrom strains Gllal (ampC+) and H1t14 (ampC12).Solutions of enzyme containing 0.1% (wt/vol) ofgelatin (to stabilize the protein) were incubated at 23and 44 C. At the times indicated, samples werewithdrawn and their penicillinase activity was as-sayed.

128 J. BACTERIOL.

on Novem

ber 22, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

STRUCTURAL GENE FOR E. COLI PENICILLINASE

dase, D-alanine-carboxypeptidase I and II, re-ferences 13 and 24) was observed. Finally, inheated extracts of strain Hltl4, whose penicil-linase was >90% inactivated, the carboxypep-tidase I and II activities (13) were normal. Thus,no connection was found between penicillinaseand biosynthetic or degradative murein en-zymes.

DISCUSSIONThe mutants studied in the present investiga-

tion were selected because of their alteredpenicillinase activities (Table 2). The mappingdata presented (Table 3) show that, in strainsTP1, TP3, and UM104, the mutations wereco-transducible with purA+. Thus, they arelocated near ampA at 82 min on the chromo-some map. The penicillinases purified from themutant TP3 (ampC3) and from the transduc-tant Hltl4 (ampC12) had enough enzyme ac-tivity to allow their characterization. Determi-nations of Km and the response to antibodiesmade against the wild-type enzyme (Table 4and Fig. 4) show that the catalytic properties ofthe ampC3 and ampC12 enzymes have beenaltered. That the basic molecular properties ofthe three enzymes still are very similar isevident from the Ouchterlony test (Fig. 3) aswell as from the fact that the purificationprocedure was identical for the three prepara-tions. Taken together, these genetic and bio-chemical data show that strains TP3 andUM104 were mutated in the structural gene forthe chromosomal penicillinase. This claim isfurther strengthened by the demonstration thatthe ampC12 enzyme is more heat labile than iswild-type penicillinase (Fig. 5).The purification procedure for the penicillin-

ase yielded from strain Hltl4 only 20% of theexpected amount of protein. With the mutantTP1, hardly any protein was obtained. Thiscould indicate the presence of regulatory muta-tions in both of the strains. (If proven to be so,the allele of TP1 may have to be renamed.) ForHltl4, it could imply that the ampC12 allele ofthe structural gene simultaneously affects theregulation of the penicillinase synthesis or thepresence of an independent and very closelylinked regulatory mutation. Alternatively, thedecreased yields could be due to an increasedinstability to the conditions used during thepurification. Reversions of ampAl are unlikelysince this mutation was recovered in geneticexperiments with TP1 and Hltl4 (Burman etal., unpublished data). Further work is, how-ever, needed for the understanding of the con-trol of the penicillinase gene.

Bacterial mutants lacking a given enzymehave often proved helpful for elucidating thephysiological role of the enzyme. This has beenthe case of penicillin resistance, and we haverecently discussed the respective roles of thepenicillinase and the barrier function of theother membrane (H. G. Boman et al., Ann. N.Y.Acad. Sci., in press). We have also used purifiedpenicillinase and a temperature-sensitive mu-tant to investigate whether the enzyme has aphysiological function related to biosynthesis orbreakdown of murein. No positive evidence forany such function was obtained. However, therecent finding that the ampA gene is trans-dom-inant (Burman and Normark, personal commu-nication) and the isolation of a protein inhibitorof a penicillinase (11) can be taken as newindirect evidence for a function other than the,B-lactamase activity.

ACKNOWLEDGMENTSWe thank Eva (Matsson) Skogman and Anita Lindstrom

for valuable technical assistance.The work was supported by grants from The Swedish

Cancer Society (Project 157) and The Swedish Natural Sci-ence Research Council (Dnr 2453). J. T. P. was the recipientof a U.S. Public Health Service special fellowship.

LITERATURE CITED

1. Bertani, G. 1951. Studies on lysogenesis. I. The mode ofphage liberation by lysogenic Escherichia coli. J. Bac-teriol. 62:293-300.

2. Boman, H. G., K. G. Eriksson-Grennberg, J. Foldes, andE. B. Lindstrom. 1967. The regulation and possibleevolution of a penicillinase-like enzyme in Escherichiacoli, p. 366-372. In V. V. Koningsberger and L. Bosch(ed.), Regulation of nucleic acid and protein biosynthe-sis. Biochemiac et Biophysica Acta Library, vol. 10,Elsevier Publishing Co., Amsterdam.

3. Braun, V., and K. Rehn. 1969. Chemical characteriza-tion, spatial distribution and function of a lipoprotein(murein-lipoprotein) of the Escherichia coli cell wall.Eur. J. Biochem. 10:426-438.

4. Burman, L. G., K. Nordstr6m, and G. D. Bloom. 1972.Murein and the outer penetration barrier of Esche-richia coli K-12, Proteus mirabilis, and Pseudomonasaeruginosa. J. Bacteriol. 112:1364-1374.

5. Burman, L. G., K. Nordstr6m, and H. G. Boman. 1968.Resistance of Escherichia coli to penicillins. V. Physio-logical comparison of two isogenic strains, one withchromosomally and one with episomally mediatedampicillin resistance. J. Bacteriol. 96:438-446.

6. Dubnau, D. A., and M. R. Pollock. 1965. The genetics ofBacillus licheniformis penicillinase: a preliminaryanalysis from studies on mutation and inter-strain andintra-strain transformations. J. Gen. Microbiol.41:7-21.

7. Eriksson-Grennberg, K. G. 1968. Resistance of Esche-richia coli to penicillins. H. An improved mapping ofthe ampA gene. Genet. Res. 12:147-156.

8. Eriksson-Grennberg, K. G., H. G. Boman, J. A. T.Jansson, and S. Thoren. 1965. Resistance of Esche-richia coli to penicillins. I. Genetic study of someampicillin-resistant mutants. J. Bacteriol. 90:54-62.

9. Gross, J., and E. Englesberg. 1959. Determination of theorder of mutational sites governing I-arabinose utiliza-

VOL. 116, 1973 129

on Novem

ber 22, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

BURMAN ET AL.

tion in Escherichia coli B/r by transduction with phagePlbt. Virology 9:314-331.

10. Hartmann, R., J.-V. H6ltje, and U. Schwarz. 1972.Targets of penicillin action in Escherichia coli. Nature(London) 235:426-429.

11. Hata, T., S. Omura, Y. Iwai, H. Ohno, H. Takeshima,and N. Yamaguchi. 1972. Studies on penicillinaseinhibitors produced by microorganisms. J. Antibiot.25:473-474.

12. Ingram, V. M. 1957. Gene mutations in human haemo-globin: the chemical difference between normal andsickle cell haemoglobin. Nature (London) 180:326-328.

13. Izaki, K., M. Matsuhashi, and J. L. Strominger. 1968.Biosynthesis of the peptidoglycan of bacterial cellwalls. XIII. Peptidoglycan transpeptidase and D-ala-nine carboxypeptidase: penicillin sensitive enzymaticreaction in strains of Escherichia coli. J. Biol. Chem.243:3180-3192.

14. Lindstr6m, E. B., H. G. Boman, and B. B. Steele. 1970.Resistance of Escherichia coli to penicillins. VI. Purifi-cation and characterization of the chromosomally me-diated penicillinase present in ampA-containingstrains. J. Bacteriol. 101:218-231.

15. Lindstrom, E. B., and K. Nordstrom. 1972. Automatedmethod for determination of penicillins, cephalosporinsand penicillinases. Antimicrob. Ag. Chemother.1:100-106.

16. Nordstr6m, K., K. G. Eriksson-Grennberg, and H. G.Boman. 1968. Resistance of Escherichia coli to penicil-lins. III. AmpB, a locus affecting episomally andchromosomally mediated resistance to ampicillin and

chloramphenicol. Genet. Res. 12:157-168.17. Novick, R. P. 1963. Analysis by transduction of muta-

tions affecting penicillinase formation in Staphylococ-cus aureus. J. Gen. Microbiol. 33:121-136.

18. Park, J. T., and A. N. Chatterjee. 1966. Membraneassociated reactions involved in bacterial cell wallmucopeptide synthesis, p. 466-472. In E. F. Neufeldand V. Ginsberg (ed.), Methods in enzymology, vol. 8.Academic Press Inc., New York.

19. Pollock, M. R. 1956. An immunological study of theconstitutive and the penicillin-induced penicillinasesof Bacillus cereus, based on specific enzyme neutraliza-tion by antibody. J. Gen. Microbiol. 14:90-108.

20. Smith, D. H., and B. D. Davis. 1967. Mode of action ofnovobiocin in Escherichia coli. J. Bacteriol. 93:71-79.

21. Smith, J. T. 1963. Penicillinase and ampicillin resistancein a strain of Escherichia coli. J. Gen. Microbiol.30:299-306.

22. Taylor, A. L. 1970. Current linkage map of Escherichiacoli. Bacteriol. Rev. 34:155-175.

23. Vogel, H. J., and D. M. Bonner. 1956. Acetylornithinaseof Escherichia coli: partial purification and some

properties. J. Biol. Chem. 218:97-106.24. Weidel, W., and H. Pelzer. 1964. Bagshaped mac-

romolecules-a new outlook on bacterial cell walls.Advan. Enzymol. 26:193-232.

25. Wise, E. M., and J. T. Park. 1965. Penicillin: its basic siteof action as an inhibitor of a peptide cross-linkingreaction in cell wall mucopeptide synthesis. Proc. Nat.Acad. Sci. U.S.A. 54:75-81.

130 J. BACTERIOL.

on Novem

ber 22, 2020 by guesthttp://jb.asm

.org/D

ownloaded from