ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, Dec. 1996, p. 2769–2774
Copyright ? 1996, American Society for Microbiology
Vol. 40, No. 12
Borderline Methicillin-Susceptible Staphylococcus aureus Strains
Have More in Common than Reduced Susceptibility to
ORIETTA MASSIDDA,* MARIA PIA MONTANARI, MARINA MINGOIA,
AND PIETRO EMANUELE VARALDO
Institute of Microbiology, University of Ancona Medical School, 60131 Ancona, Italy
Received 22 February 1996/Returned for modification 2 May 1996/Accepted 17 September 1996
Ten epidemiologically unrelated Staphylococcus aureus isolates with borderline levels of susceptibility to
antistaphylococcal penicillinase-resistant penicillins (PRPs) were investigated together with appropriate S.
aureus control strains. By a nitrocefin microplate assay, all borderline PRP-susceptible test strains were found
to produce comparable amounts of ?-lactamase. Hydrolytic activity against another chromogenic substrate
(PADAC) and against the PRPs was also demonstrated in membrane preparations from induced cells of 9 of
the 10 borderline test strains. When bacterial membranes were analyzed by sodium dodecyl sulfate-polyacryl-
amide gel electrophoresis, two methicillin-inducible bands of about 32 and 31 kDa were detected, after
Coomassie blue staining, in the membrane protein patterns of the same nine borderline test strains. By gel
renaturation and zymographic detection of ?-lactamase activity, both methicillin-inducible membrane proteins
were detected with nitrocefin as a substrate, whereas only one band (presumably the smaller protein) was
detected with PADAC. With the remaining borderline test strain (a40), only the larger band was detected in the
renatured gels with nitrocefin as a substrate. Plasmid DNA analysis revealed that the borderline susceptible
test strains, again with the exception of a40, shared a 17.2-kb plasmid yielding four HindIII fragments of 7.0,
5.3, 3.5, and 1.4 kb. In Western blot (immunoblot) experiments using rabbit antiserum to penicillin-binding
protein (PBP) 2a, test strain a40, which did not share a number of features characteristically associated with
the other borderline test strains, was eventually shown to produce PBP 2a. Five other S. aureus strains,
belonging to phage group 94/96, were found to display the borderline phenotype, including such distinguishing
features as the membrane-associated PRP- and PADAC-hydrolyzing activity, the two methicillin-inducible
membrane proteins, and the 17.2-kb plasmid. These results suggest that borderline susceptible S. aureus
strains share more common features than reduced susceptibility to PRPs.
Over the past few years, special attention has been focused
on Staphylococcus aureus strains with borderline levels of sus-
ceptibility or resistance to methicillin and related antistaphy-
lococcal penicillinase-resistant penicillins (PRPs) (30). Thorns-
berry and McDougal first reported the occurrence of S. aureus
strains with intermediate susceptibility to PRPs (17, 27) and
then defined the concept of borderline susceptibility (or low-
level resistance) by postulating that in these S. aureus strains—
which were neither heteroresistant nor multidrug resistant,
produced large amounts of ?-lactamase, and became fully sus-
ceptible to PRPs in the presence of ?-lactamase inhibitors—
the borderline MICs were due to hyperproduction of ?-lacta-
mase (18). The nonheterogeneous expression of this reduced
susceptibility to PRPs was confirmed by population analysis
(20, 31). Its nonintrinsic nature was confirmed by the experi-
mental findings that borderline susceptible S. aureus strains did
not contain DNA that hybridized with probes specific for the
methicillin resistance determinant (mec) (4) and did not pro-
duce the low-affinity penicillin-binding protein (PBP) 2a (4,
20). It was argued that there is no apparent reason for border-
line S. aureus isolates to be subject to the current practice of
considering methicillin-resistant staphylococci as resistant to
all other ?-lactams (20) and that infections caused by these
strains can probably be safely and effectively treated with ?-lac-
tam antibiotics (4).
However, the matter is probably more complex than origi-
nally believed. In particular, there is increasing evidence that
the borderline phenotype cannot be defined solely on the basis
of ?-lactamase hyperproduction: in fact, both borderline sus-
ceptible S. aureus strains which do not hyperproduce ?-lacta-
mase and fully PRP-susceptible strains which do hyperproduce
it have been reported (2, 5, 15, 26, 32). Moreover, mechanisms
involving PBP modification may also lead to reduced staphy-
lococcal susceptibility to PRPs. One such mechanism is char-
acterized by the presence of the mec gene and PBP 2a (5, 28)
and is similar, except for the lower PRP MICs, to conventional
high-level resistance. Another mechanism, best documented in
homogeneous, ?-lactamase-negative strains (3, 28, 29), is as-
sociated with altered binding capacity of normal PBPs (i.e., in
the absence of PBP 2a) possibly resulting from point mutations
in the PBP 2 gene (7).
We have recently shown that, in addition to the classical
penicillinase, borderline PRP-susceptible S. aureus strains pro-
duce a second, methicillin-hydrolyzing ?-lactamase (methicil-
linase) having a molecular mass slightly smaller than that of the
penicillinase (15). By plasmid DNA analysis, we subsequently
noted that the methicillinase-producing strains possessed a
common 17.2-kb ?-lactamase plasmid (14) apparently identical
to a penicillinase-encoding plasmid described in borderline
PRP-susceptible S. aureus isolates involved in widespread nos-
ocomial infections (19). Moreover, a close correlation between
the borderline phenotype and phage group 94/96 has been
reported in studies concerning S. aureus strains associated with
nosocomial infections (19, 34). In the present study, these
* Corresponding author. Mailing address: Institute of Microbiology,
University of Ancona Medical School, Via Ranieri, Monte d’Ago,
60131 Ancona, Italy. Phone: 39 71 2204694. Fax: 39 71 2204693.
features associated with borderline S. aureus strains have been
further investigated with the purpose of better characterizing
these organisms and providing more reliable markers to define
MATERIALS AND METHODS
Bacterial strains. Ten S. aureus strains (a3, a15, a28, a40, a56, a82, a217, a263,
a404, and a577)—independently isolated from a variety of clinical specimens and
representing geographic and chronologic diversities—were selected on the basis
of borderline PRP susceptibility levels (MIC of methicillin, 4 to 8 ?g/ml; MIC of
oxacillin, ?2 ?g/ml); PRP MICs proved to be reproducible in repeated assays.
Five additional S. aureus strains belonging to phage group 94/96 (the propagating
strains for the basic-set typing phages 94 and 96 and three unrelated strains
examined for routine phage typing) were obtained from the phage typing refer-
ence laboratory in Italy (National Institute of Health, Rome, Italy). Two S.
aureus control strains from our institute’s collection included a53 (a borderline
isolate previously characterized for PBPs , ?-lactamase activity [14, 15], and
plasmid DNA ) and a90 (a fully PRP-susceptible isolate with high-level
resistance to penicillin G). S. aureus ATCC 25923 and ATCC 29213 were used as
?-lactamase-negative and -positive controls, respectively.
Penicillins and ?-lactamase substrates. Penicillin G, methicillin, oxacillin,
cloxacillin, dicloxacillin, and nafcillin were purchased from Sigma Chemical Co.,
St. Louis, Mo. The chromogenic cephalosporins nitrocefin and PADAC were
purchased from Oxoid Ltd., Basingstoke, United Kingdom, and Calbiochem, San
Diego, Calif., respectively.
Susceptibility tests. MICs were determined by the broth microdilution method
as recommended by the National Committee for Clinical Laboratory Standards
(21). As recommended in particular for PRP susceptibility tests, the inoculum
was prepared directly from an overnight agar plate, and 2% NaCl was added to
the cation-adjusted Mueller-Hinton broth (Difco Laboratories, Detroit, Mich.).
Penicillins were tested at final concentrations (prepared from serial twofold
dilutions) ranging from 0.125 to 128 ?g/ml for penicillin G and from 0.25 to 32
?g/ml for methicillin and oxacillin. S. aureus ATCC 29213 was used for quality
Preparation of membrane fraction. Bacteria were grown to the mid-logarith-
mic phase in tryptic soy broth (Difco) at 37?C with shaking. The medium was
used unsupplemented or supplemented with methicillin (0.5 ?g/ml) as a ?-lac-
tamase inducer. The cultures were harvested after 3 h of exposure to the inducer.
Bacterial membranes were prepared as previously described (15) and resus-
pended in 10 mM phosphate buffer (pH 7.0) at a concentration of 2 mg/ml. The
protein content was determined with a DC protein assay kit (Bio-Rad Labora-
tories, Richmond, Calif.).
?-Lactamase assays. A nitrocefin microplate assay was similar to the method
originally applied by McDougal and Thornsberry (18). Aliquots of 50 ?l of a
500-?g/ml nitrocefin solution were added to two selected wells of broth microdi-
lution plates grown overnight: the growth control well (to test uninduced ?-lac-
tamase) and the well containing 0.5 ?g of methicillin per ml (to test induced
?-lactamase). On the basis of color change and reaction time, the results were
scored as ?, 1?, 2?, or 3? according to a previously described scheme (15, 32).
?-Lactamase in the membrane fraction, prepared as described above, was
tested by spotting 20 ?g of membrane protein on 1% agarose plates as reported
previously (14). Briefly, with the PRPs or penicillin G used as a substrate (1%,
wt/vol), hydrolytic activity was tested by an iodometric technique and revealed by
the appearance of a zone of clearing in the dark background. With the chromo-
genic substrates (100 ?M), hydrolytic activity was directly indicated by color
change (red for nitrocefin or yellow for PADAC). A crude extract of Escherichia
coli C600(R46), which produces OXA-2 ?-lactamase (10), was prepared as
described elsewhere (14) and used as a positive control for all substrates.
Analysis of membrane proteins by SDS-PAGE and detection of ?-lactamase
activity by gel renaturation. Bacterial membranes (aliquots of 50 to 100 ?g of
total protein) were subjected to sodium dodecyl sulfate (SDS)-polyacrylamide
gel electrophoresis (PAGE). The stacking and separating gels were 5 and 15%
polyacrylamide at pHs 6.8 and 8.8, respectively. The gels were run at 10 mA for
24 h in a continuous SDS-Tris-glycine buffering system. After electrophoresis,
the gels were either stained with Coomassie brilliant blue R250 or renatured in
phosphate buffer containing 1% (vol/vol) deionized Triton X-100 (16). In the
latter case, after 6 h of incubation, necessary to renature staphylococcal ?-lac-
tamases (15), the gels were overlaid with 1% agarose gel containing nitrocefin or
PADAC (100 ?M). Hydrolytic activity was detected after incubation up to 30
min at room temperature when nitrocefin was used as a substrate and overnight
at 37?C when PADAC was used.
Preparation of plasmid DNA and agarose gel electrophoresis. Plasmid DNA
was isolated with a Wizard Minipreps DNA Purification System (Promega Corp.,
Madison, Wis.). Lysostaphin (Sigma) was added to the cell resuspension buffer
at a final concentration of 0.1 ?g/ml. Plasmid DNA was digested with HindIII
(Boehringer Mannheim, Mannheim, Germany), and restriction fragments were
separated by electrophoresis on horizontal 0.8% agarose gels with TAE buffer
(40 mM Tris acetate [pH 8.5], 2 mM EDTA) and stained with ethidium bromide
Immunodetection of PBP 2a. PBP 2a was obtained from S. aureus 27R (me-
thicillin resistant, producing PBP 2a constitutively [9, 29]). After separation of
membrane proteins by SDS-PAGE and staining with Coomassie blue, a slice of
the gel containing PBP 2a was taken, washed, homogenized, and used to immu-
nize New Zealand White rabbits (8). The antiserum was stored at ?20?C in the
presence of 0.02% azide and, before use in immunoblot experiments, pread-
sorbed to membranes from S. aureus 27S (methicillin susceptible, isogenic with
27R [9, 29]). Samples of membrane fraction (150 ?g of total protein) were
prepared from S. aureus cells grown at 30?C in tryptic soy broth containing 4%
NaCl and separated by SDS-PAGE. Proteins were transferred to polyvinylidene
difluoride membranes (Immobilon-P; Millipore, Bedford, Mass.) in a semidry
system (Millipore) for 1 h at 0.8 mA/cm2(8) and immunoblotted with anti-PBP
2a antiserum (1:2,500). After washing, a second antibody consisting of alkaline
phosphatase-conjugated goat anti-rabbit immunoglobulin G (Boehringer Mann-
heim) was applied. This was followed by extensive washing and visualization with
an alkaline phosphatase substrate kit (Bio-Rad).
?-Lactamase activity of borderline S. aureus strains. The 10
borderline-PRP-susceptible S. aureus test strains were assayed
for ?-lactamase activity with different substrates. S. aureus a53
(previously shown to produce an inducible methicillinase ac-
tivity ), a90, ATCC 29213, and ATCC 25923 were tested as
controls. ?-Lactamase activity was scored from uninduced and
induced cultures by the microplate nitrocefin assay. By this
method, all of the borderline test strains appeared to produce
large amounts of ?-lactamase, particularly after methicillin
induction. Hydrolytic activity against nitrocefin, PADAC, pen-
icillin G, and different PRPs (methicillin, oxacillin, cloxacillin,
dicloxacillin, and nafcillin) was then tested by agarose plate
assays from cytoplasmic membranes prepared from uninduced
and induced cells. No activity on any substrate was shown by
the ?-lactamase-negative control (ATCC 25923), whereas ni-
trocefin and penicillin G were hydrolyzed by all other 13 S.
aureus strains. Of the latter, 10 (a53 and 9 of the 10 test strains)
also hydrolyzed the PRPs and PADAC, provided that mem-
brane preparations from induced cells were used. The remain-
ing borderline test strain (a40) and the fully PRP-susceptible
controls a90 and ATCC 29213 did not hydrolyze either the
PRPs or PADAC under any condition. A summary of the
susceptibility and ?-lactamase activity data is reported in Table
1. The results obtained with membrane fractions from induced
cells with nitrocefin, PADAC, penicillin G, methicillin, and
oxacillin as substrates are shown in Fig. 1.
Analysis of membrane proteins by SDS-PAGE. For each
strain, cytoplasmic membrane proteins obtained from cells
grown both in the absence and in the presence of methicillin
(0.5 ?g/ml) were separated by SDS-PAGE and stained with
Coomassie blue. Two closely migrating but distinct methicillin-
inducible bands, with molecular masses of about 32 and 31
kDa, were detected in the membrane fractions of 9 of the 10
borderline-susceptible test strains. The results obtained with
three such strains (a3, a28, and a56) are shown in Fig. 2. The
two bands appeared to be the same as those detected in the
membrane fraction of the borderline S. aureus strain a53 (Fig.
2, lane 12) and previously shown to possess ?-lactamase activ-
ity (15). With the remaining borderline test strain (a40) and
control strains a90 and ATCC 29213, as well as with the ?-lac-
tamase-negative control ATCC 25923, such methicillin-induc-
ible bands were not observed in the Coomassie blue-stained
gels, even after induction (a40 and a90 are shown in Fig. 2).
Renaturation assays and zymographic detection of ?-lacta-
mase activity. A set of similar gels with cytoplasmic membrane
proteins were run in parallel, renatured with Triton X-100, and
assayed for ?-lactamase activity. When nitrocefin was used as
a substrate, the 32-kDa methicillin-inducible band—previously
identified as the classical staphylococcal penicillinase (15)—
was detected in all strains except the ?-lactamase-negative
control ATCC 25923. The second band—i.e., the 31-kDa pro-
2770MASSIDDA ET AL.ANTIMICROB. AGENTS CHEMOTHER.
tein previously shown to have hydrolytic activity against me-
thicillin (15)—was detected in the same borderline strains
showing this protein in the Coomassie blue-stained gels, pro-
vided that the membranes were obtained from methicillin-
induced cultures. Even after induction, this second band was
not detectable in the membrane fraction of either the remain-
ing borderline test strain (a40) or control strains a90, ATCC
29213, and ATCC 25923. When PADAC was used, a band
denoting hydrolysis (presumably the 31-kDa protein) was de-
tected only in the membranes from methicillin-induced cul-
tures of the same borderline strains showing both the 32- and
the 31-kDa bands, under the same induction conditions, with
nitrocefin as a substrate. The results obtained with a90 and
four borderline test strains (a3, a28, a40, and a56) are shown in
Fig. 3A (nitrocefin) and B (PADAC).
Plasmid DNA analysis. As well as the control strain a53, the
borderline PRP-susceptible test strains (again with the excep-
tion of a40) showed the presence of a common 17.2-kb plasmid
with the same HindIII restriction pattern (four fragments of
7.0, 5.3, 3.5, and 1.4 kb). Different plasmid patterns were
shown by the borderline test strain a40 and the other control
strains (data not shown).
Immunodetection of PBP 2a. Further studies of test strain
a40 showed the presence of a 78-kDa band corresponding to
PBP 2a as in the control strain 27R but not in all the other
borderline susceptible test strains (data not shown).
Relationships to phage typing. The phage typing reference
laboratory in Italy stopped phage typing with staphylococci in
1994 but provided us with five S. aureus strains previously
assigned to phage group 94/96. These five strains were all
found to display the borderline phenotype, as previously sug-
gested (19, 34), with identical methicillin MICs (4 ?g/ml) and
oxacillin MICs ranging from 0.5 to 2 ?g/ml. All showed strong
?-lactamase activity, particularly after induction, and all shared
additional distinguishing features, including membrane-associ-
ated PRP- and PADAC-hydrolyzing activities, the two close
methicillin-inducible membrane proteins (32 and 31 kDa), and
the 17.2-kb plasmid yielding the characteristic four-fragment
pattern after digestion with HindIII.
Of the 10 epidemiologically independent S. aureus test
strains selected on the basis of their borderline PRP suscepti-
TABLE 1. Susceptibility to penicillin G, methicillin, and oxacillin and ?-lactamase activity in 14 S. aureus strains
MIC (?g/ml) of:
Inducible in the membrane fraction against:
Penicillin G MethicillinOxacillinUninduced InducedPenicillin G PRPsPADAC
FIG. 1. ?-Lactamase activity in cytoplasmic membranes from methicillin-
induced S. aureus cells against nitrocefin (A), PADAC (B), penicillin G (C),
methicillin (D), and oxacillin (E). With the chromogenic substrates nitrocefin
and PADAC, hydrolytic activity was directly indicated by color change. With the
other substrates, the hydrolytic activity was assayed by an iodometric technique
and revealed by the appearance of a zone of discoloration in the dark back-
ground. S. aureus strains (spots 1 to 14): 1, ATCC 29213; 2, ATCC 25923; 3, a90;
4, a53; 5, a28; 6, a3; 7, a56; 8, a577; 9, a404; 10, a217; 11, a82; 12, a15; 13, a263;
and 14, a40. A crude extract of E. coli C600(R46), which produces OXA-2
?-lactamase, was used as a positive control (spot 15).
VOL. 40, 1996 BORDERLINE METHICILLIN-SUSCEPTIBLE S. AUREUS 2771
bility, 9 behaved quite homogeneously and proved to be very
similar both to one another and to a53, a previously investi-
gated borderline S. aureus isolate (14, 15). Besides borderline
PRP susceptibility, these strains shared a particular 17.2-kb
?-lactamase plasmid, two distinct inducible ?-lactamase bands
in their membrane protein patterns, and membrane-associated
hydrolytic activity against PRPs and PADAC. By contrast, the
remaining borderline test strain (a40) did not share any of
these particular characteristics and, unlike the other borderline
test strains, was ultimately shown to produce PBP 2a. Although
we were not able to determine the phage type of our test
strains, investigations of strains belonging to the 94/96 complex
confirmed reported relationships between this particular phage
group and the borderline phenotype (19, 34), with the associ-
ated 17.2-kb ?-lactamase plasmid (2, 14, 19), and suggested
that this correlation also applies to methicillinase production.
It is worth noting that the group V phage complex (which
includes bacteriophages 94 and 96) is regarded as a quite
distinct and homogeneous class of S. aureus strains, worldwide
in occurrence, with a possible common origin (1, 22), and
correlating with the production of large amounts of ?-lacta-
mase (23). Such a homogeneity may reflect the function of the
restriction and modification systems in these strains that pre-
vent the acquisition of genetic material from strains outside the
complex (1, 22).
According to the original hypothesis by McDougal and
Thornsberry (18), some S. aureus strains that produce large
amounts of ?-lactamase may show borderline susceptibility or
resistance to PRPs because the staphylococcal ?-lactamase,
when hyperproduced, would succeed in partially hydrolyzing
these penicillins. It was soon realized, however, that ?-lacta-
mase hyperproduction cannot be regarded as the only mecha-
nism involved and that production of large amounts of ?-lac-
tamase may be neither sufficient nor necessary to determine
the borderline phenotype (2, 15, 26, 32). The recent discovery
of an inducible methicillinase—produced in addition to the
classical penicillinase—in the membrane fraction of borderline
S. aureus strains suggested that another, more specific PRP-
hydrolyzing ?-lactamase is more likely to account for the bor-
derline phenotype than an increased amount of the conven-
tional penicillinase (15). The slow hydrolysis of methicillin by
this methicillinase, suggested by the overnight incubation nec-
essary to detect the reaction in renatured gels (15), is consis-
tent with the fact that its presence results in only reduced
susceptibility rather than true resistance. On the other hand,
the staphylococcal penicillinase, even if hyperproduced—as is
the case with E. coli DH5?(pAH12), which expresses the prod-
uct of the blaZ gene (i.e., penicillinase) derived from the
17.2-kb plasmid of S. aureus a53 (14)—appears not to be per se
capable of hydrolyzing methicillin. This indirectly indicates
that penicillinase hyperproduction is unlikely to be the true
mechanism responsible for the borderline phenotype. That
increased production of ?-lactamase which is nevertheless of-
ten observed in borderline S. aureus strains might result from
the combined effect of the second ?-lactamase (methicillinase)
in addition to the conventional one (penicillinase) on some
common substrates such as nitrocefin and penicillin G.
The occurrence of such a methicillin-hydrolyzing ?-lacta-
mase in borderline S. aureus strains was first documented by
means of SDS-PAGE of membrane proteins followed by gel
renaturation (15). In the present study, the new ?-lactamase
activity associated with borderline S. aureus strains has been
investigated by both gel renaturation and agarose plate assays.
Methicillinase had been shown to be inducible and active
against penicillin G and nitrocefin, in addition to methicillin
(15); now, it has proved to be active, too, against isoxazolyl
penicillins and nafcillin and against PADAC, again only after
induction. While the ability of methicillinase to hydrolyze
PRPs other than methicillin was expected, the hydrolysis of
PADAC appears to deserve additional attention. In fact, this
chromogenic cephalosporin is regarded as a possible substrate
of the ?-lactamases produced by some gram-negative bacteria
but virtually insensitive to the ?-lactamase activity of staphy-
In borderline S. aureus strains, production of methicillinase
FIG. 3. Detection of ?-lactamase activity in the membranes of five S. aureus
strains by SDS-PAGE followed by gel renaturation. The chromogenic cephalo-
sporins nitrocefin (A) and PADAC (B) were used as substrates. Lanes: 1, 3, 5, 7,
and 9, strains a90, a3, a28, a40, and a56, respectively, uninduced; 2, 4, 6, 8, and
10, the same strains grown in the presence of methicillin (0.5 ?g/ml).
FIG. 2. Membrane protein patterns of six S. aureus strains (Coomassie blue-
stained gel). Lanes: 1, 3, 5, 7, 9, and 11, strains a90, a3, a28, a40, a56, and a53,
respectively, uninduced; 2, 4, 6, 8, 10, and 12, the same strains grown in the
presence of methicillin (0.5 ?g/ml); M, molecular mass markers (prestained
SDS-PAGE standards, low range [Bio-Rad]).
2772MASSIDDA ET AL.ANTIMICROB. AGENTS CHEMOTHER.
is consistently associated with the presence of a 17.2-kb ?-lac-
tamase plasmid, containing the blaZ gene encoding penicilli-
nase in the largest of the four HindIII fragments (14). An
apparently identical ?-lactamase plasmid, having the same size
and the same HindIII restriction pattern, has been described
by McMurray et al. (19) in borderline S. aureus strains. Barg et
al. (2) reported transfer and curing of this plasmid from a
borderline S. aureus strain belonging to phage group 94/96. In
curing experiments, loss of the plasmid resulted in loss of
borderline susceptibility. In transformation experiments, how-
ever, only the transformants obtained using a recipient of the
94/96 complex met the borderline susceptibility criteria, sug-
gesting that the strain genetic background is important for the
expression of the borderline phenotype. Considering its
smaller size and other features, the 17.2-kb plasmid appears to
be different from other ?-lactamase plasmids of S. aureus (13,
25) and peculiar to borderline strains. Recently, we suggested
that the 17.2-kb ?-lactamase-encoding plasmid is the most
likely candidate to encode methicillinase as well (14); in par-
ticular, we suggested that methicillinase could be the product
of a mutated blaR1 gene. In fact, the blaR1 product (a signal-
transducing membrane protein required for ?-lactamase in-
duction [6, 12]) has been shown to have high sequence simi-
larity in S. aureus (24) as well as in Bacillus licheniformis (33) to
the class D ?-lactamases.
In the absence of official reference guidelines, determination
of the limits of the borderline phenotype may be a problem
(30). The relative lack of homogeneity in the criteria which
have been adopted to define borderline S. aureus strains may
cause difficulty in comparing the results of different microbio-
logical and clinical studies involving these organisms. Of
course, the reliability and the meaning itself of given limits for
the borderline phenotype depend on the degree of correlation
between the existence of particular levels of PRP MICs and the
presence of particular biological characteristics. From this
point of view, however, reduced staphylococcal susceptibility to
PRPs may result from at least two categories of biological
mechanisms, some extrinsic (depending on ?-lactamase activ-
ity) and some intrinsic (depending on altered PBPs). The latter
may be further associated with either the presence of PBP 2a
(as reported previously [5, 28] and exemplified in this study by
strain a40) or altered binding capacity of normal PBPs (as
described in laboratory-derived [3, 29] and clinical 
strains). According to the original description by McDougal
and Thornsberry (18) and following studies (19, 20), including
the present findings, extrinsic mechanisms are probably more
common than intrinsic mechanisms among clinical isolates of
borderline PRP-susceptible S. aureus. These extrinsic border-
line strains appear to share, in addition to particular PRP
MICs, a unique phage group, a unique ?-lactamase plasmid,
and a unique PRP-hydrolyzing ?-lactamase (other than peni-
cillinase and produced in addition to it as a membrane-asso-
ciated enzyme). This methicillinase, in particular, is likely to be
the direct cause of the reduced PRP susceptibility in these
strains, so far regarded as ?-lactamase hyperproducers. All this
may give the concept of borderline PRP susceptibility a more
precise and concrete biological meaning, rather than reducing
it to a mere occurrence of given (more or less strict and pre-
This study was supported in part by the National Research Council,
targeted projects FATMA and BTBS.
We thank Gerald D. Shockman (Temple University, Philadelphia,
Pa.) for helpful discussion and comments on the manuscript and Gra-
ziella Orefici (National Institute of Health, Rome, Italy) for kindly
providing five S. aureus strains of phage group 94/96.
1. Asheshov, E. H., A. W. Coe, and A. Porthouse. 1977. Properties of strains of
Staphylococcus aureus in the 94, 96 complex. J. Med. Microbiol. 10:171–178.
2. Barg, N., H. Chambers, and D. Kernodle. 1991. Borderline susceptibility to
antistaphylococcal penicillins is not conferred exclusively by the hyperpro-
duction of ?-lactamase. Antimicrob. Agents Chemother. 35:1975–1979.
3. Berger-Ba ¨chi, B., A. Stra ¨ssle, and F. H. Kayser. 1989. Natural methicillin
resistance in comparison with that selected by in-vitro drug exposure in
Staphylococcus aureus. J. Antimicrob. Chemother. 23:179–188.
4. Chambers, H. F., G. Archer, and M. Matsuhashi. 1989. Low-level methicillin
resistance in strains of Staphylococcus aureus. Antimicrob. Agents Che-
5. Gerberding, J. L., C. Miick, H. H. Liu, and H. F. Chambers. 1991. Compar-
ison of conventional susceptibility tests with direct detection of penicillin-
binding protein 2a in borderline oxacillin-resistant strains of Staphylococcus
aureus. Antimicrob. Agents Chemother. 35:2574–2579.
6. Hackbarth, C. J., and H. F. Chambers. 1993. blaI and blaR1 regulate ?-lac-
tamase and PBP 2a production in methicillin-resistant Staphylococcus au-
reus. Antimicrob. Agents Chemother. 37:1144–1149.
7. Hackbarth, C. J., T. Kocagoz, S. Kocagoz, and H. F. Chambers. 1995. Point
mutations in Staphylococcus aureus PBP 2 gene affect penicillin-binding
kinetics and are associated with resistance. Antimicrob. Agents Chemother.
8. Harlow, E., and D. Lane. 1988. Antibodies: a laboratory manual. Cold Spring
Harbor Laboratory, Cold Spring Harbor, N.Y.
9. Hartman, B. J., and A. Tomasz. 1984. Low-affinity penicillin-binding protein
associated with ?-lactam resistance in Staphylococcus aureus. J. Bacteriol.
10. Jacoby, G. A., and L. Sutton. 1985. ?-Lactamases and ?-lactam resistance in
Escherichia coli. Antimicrob. Agents Chemother. 28:703–705.
11. Jorgensen, J. H., S. A. Crawford, and G. A. Alexander. 1982. Pyridinium-2-
azo-p-dimethylaniline chromophore, a new chromogenic cephalosporin for
rapid beta-lactamase testing. Antimicrob. Agents Chemother. 22:162–164.
12. Kobayashi, T., Y. F. Zhu, N. J. Nicholls, and J. O. Lampen. 1987. A second
regulatory gene, blaR1, encoding a potential penicillin-binding protein re-
quired for induction of ?-lactamase in Bacillus licheniformis. J. Bacteriol.
13. Lyon, B. R., and R. Skurray. 1987. Antimicrobial resistance of Staphylococ-
cus aureus: genetic basis. Microbiol. Rev. 51:88–134.
14. Massidda, O., M. P. Montanari, M. Mingoia, and P. E. Varaldo. 1994.
Cloning and expression of the penicillinase from a borderline methicillin-
susceptible Staphylococcus aureus strain in Escherichia coli. FEMS Microbiol.
15. Massidda, O., M. P. Montanari, and P. E. Varaldo. 1992. Evidence for a
methicillin-hydrolyzing ?-lactamase in Staphylococcus aureus strains with
borderline susceptibility to this drug. FEMS Microbiol. Lett. 92:223–227.
16. Massidda, O., G. M. Rossolini, and G. Satta. 1991. The Aeromonas hy-
drophila cphA gene: molecular heterogeneity among class B metallo-?-lac-
tamases. J. Bacteriol. 173:4611–4617.
17. McDougal, L., and C. Thornsberry. 1984. New recommendations for disk
diffusion antimicrobial susceptibility tests for methicillin-resistant (heterore-
sistant) staphylococci. J. Clin. Microbiol. 19:482–488.
18. McDougal, L., and C. Thornsberry. 1986. The role of ?-lactamase in staph-
ylococcal resistance to penicillinase-resistant penicillins and cephalosporins.
J. Clin. Microbiol. 23:832–839.
19. McMurray, L. W., D. S. Kernodle, and N. Barg. 1990. Characterization of a
widespread strain of methicillin-susceptible Staphylococcus aureus associated
with nosocomial infections. J. Infect. Dis. 162:759–762.
20. Montanari, M. P., E. Tonin, F. Biavasco, and P. E. Varaldo. 1990. Further
characterization of borderline methicillin-resistant Staphylococcus aureus
and analysis of penicillin-binding properties. Antimicrob. Agents Che-
21. National Committee for Clinical Laboratory Standards. 1993. Methods for
dilution antimicrobial susceptibility tests for bacteria that grow aerobically,
3rd ed. Approved standard M7-A3. National Committee for Clinical Labo-
ratory Standards, Villanova, Pa.
22. Parker, M. T. 1983. The significance of phage-typing patterns in Staphylo-
coccus aureus, p. 33–62. In C. S. F. Easmon and C. Adlam (ed.), Staphylo-
cocci and staphylococcal infections. Academic Press Inc., London.
23. Rosdahl, V. T., and K. Rosendal. 1983. Correlation of penicillinase produc-
tion with phage type and susceptibility to antibiotics and heavy metals in
Staphylococcus aureus. J. Med. Microbiol. 16:391–399.
24. Rowland, S. J., and K. G. H. Dyke. 1990. Tn552, a novel transposable
element from Staphylococcus aureus. Mol. Microbiol. 4:961–975.
25. Shalita, Z., E. Murphy, and R. P. Novick. 1980. Penicillinase plasmids of
Staphylococcus aureus: structural and evolutionary relationships. Plasmid
26. Sierra-Madero, J. G., C. Knapp, C. Karaffa, and J. A. Washington. 1988.
Role of ?-lactamase and different testing conditions in oxacillin-borderline-
VOL. 40, 1996BORDERLINE METHICILLIN-SUSCEPTIBLE S. AUREUS2773
susceptible staphylococci. Antimicrob. Agents Chemother. 32:1754–1757. Download full-text
27. Thornsberry, C., and L. McDougal. 1983. Successful use of broth microdi-
lution in susceptibility tests for methicillin-resistant (heteroresistant) staph-
ylococci. J. Clin. Microbiol. 18:1084–1091.
28. Tomasz, A., H. B. Drugeon, H. M. de Lencastre, D. Jabes, L. McDougall, and
J. Bille. 1989. New mechanism for methicillin resistance in Staphylococcus
aureus: clinical isolates that lack the PBP 2a gene and contain normal
penicillin-binding proteins with modified penicillin-binding capacity. Anti-
microb. Agents Chemother. 33:1869–1874.
29. Tonin, E., and A. Tomasz. 1986. ?-Lactam-specific resistant mutants of
Staphylococcus aureus. Antimicrob. Agents Chemother. 30:577–583.
30. Varaldo, P. E. 1993. The ‘borderline methicillin-susceptible’ Staphylococcus
aureus. J. Antimicrob. Chemother. 31:1–4.
31. Varaldo, P. E., F. Biavasco, and M. P. Montanari. 1987. Heteroresistant and
nonheteroresistant methicillin-resistant Staphylococcus aureus. J. Clin. Mi-
32. Varaldo, P. E., M. P. Montanari, F. Biavasco, E. Manso, S. Ripa, and F.
Santacroce. 1993. Survey of clinical isolates of Staphylococcus aureus for
borderline susceptibility to antistaphylococcal penicillins. Eur. J. Clin. Mi-
crobiol. Infect. Dis. 12:677–682.
33. Zhu, Y. F., I. H. A. Curran, B. Joris, J. M. Ghuysen, and J. O. Lampen. 1990.
Identification of BlaR, the signal transducer for ?-lactamase production in
Bacillus licheniformis, as a penicillin-binding protein with strong homology to
the OXA-2 ?-lactamase (class D) of Salmonella typhimurium. J. Bacteriol.
34. Zierdt, C. H., I. K. Hosein, R. Shively, and J. D. MacLowry. 1992. Phage
pattern-specific oxacillin-resistant and borderline oxacillin-resistant Staphy-
lococcus aureus in U.S. hospitals: epidemiological significance. J. Clin. Mi-
2774 MASSIDDA ET AL.ANTIMICROB. AGENTS CHEMOTHER.