ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, Apr. 1988, p. 420-4250066-4804/88/040420-06$02.00/0Copyright © 1988, American Society for Microbiology

Novel Mechanisms of Resistance to Lincosamides inStaphylococcus and Arthrobacter spp.

L. M. QUIROS, S. FIDALGO, F. J. MENDEZ, C. HARDISSON, AND J. A. SALAS*

Departamento de Microbiologia, Universidad de Oviedo, 33006 Oviedo, Spain

Vol. 32, No. 4

Received 17 August 1987/Accepted 21 December 1987

Clinical isolates of Staphylococcus and Arthrobacter spp. were screened for lincosamide resistance. Sixdifferent patterns of resistance were found. Strains designated SF27 and SF28 showed low-level resistance tolincosamides: one was susceptible to erythromycin (SF27) and the other was resistant (SF28). Analysis ofribosomes from the resistant strains in an in vitro poly(U)-dependent protein-synthesizing system showed thatribosomes of both strains were sensitive to lincomycin and clindamycin. Four patterns of high-level resistanceto lincosamides were observed (strains SF4, SF19, SF30, and SF31). All of these except SF30 had ribosomeswhich were highly resistant in vitro to the antibiotics and showed a close correlation with results of the in vivoexperiments. In vivo protein synthesis by strain SF30 was resistant to lincomycin and sensitive to clindamycin,whereas the ribosomes were sensitive when assayed in vitro. Lincosamide-inactivating enzymes were notdetected in cell extracts of the six resistant strains. Strains SF19 and SF31 demonstrated two ribosome-mediated lincosamides resistance mechanisms that were not previously reported. Both strains were highlyresistant to lincosamides and susceptible to erythromycin, but SF19 was also highly resistant to oleandomycinand partially resistant to various macrolides.

Lincosamide antibiotics efficiently inhibit protein synthe-sis in gram-positive bacteria but have low activity againstmany gram-negative bacteria. Lincomycin is among the bestknown representatives of the group, together with its deriv-atives clindamycin (7-chloro-7-deoxylincomycin) and ce-lesticetin. Lincomycin has been shown to bind to the 50Sribosomal subunits (5) and to act on aminoacyl-tRNA bind-ing and the peptidyl transferase reaction (3, 20, 26). Thestructure of lincosamides is chemically distinct from that ofmacrolide and streptogramin B-type antibiotics. Resistanceto lincosamide antibiotics is nearly always associated withcoresistance to macrolides and streptogramin B-type antibi-otics in the so-called MLS phenotype. The MLS antibioticsare closely related in their modes of action, and they appearto share a common or overlapping binding site in theribosome. The MLS resistance phenotype is found widelyamong Staphylococcus (4, 27, 28), Streptococcus (7, 10),Corynebacterium (12), Bacteroides (29, 36), Clostridium(17), and Bacillus (15) clinical isolates, as well as among theerythromycin producers (18, 30). The molecular basis forresistance in clinical isolates has been attributed to N6-dimethylation of adenine in 23S rRNA (22, 23), whichrenders the ribosomes insensitive to these antibiotics. Asimilar situation was found in the erythromycin producerStreptomyces eythraeus (32, 33).Few reports exist in the literature about resistance to

lincosamides in organisms that are simultaneously suscepti-ble to erythromycin and other macrolides. However, twosuch resistance phenotypes have been described in Staphy-lococcus species: one is associated with antibiotic degrada-tion and the other is unassociated with any apparent druginactivation (14). In addition, a clinical isolate of Staphylo-coccus haemolyticus has been reported to inactivate lin-comycin and clindamycin while being susceptible to macro-lides and streptogramin B-type antibiotics (24).

In this communication, we discuss several Staphylococ-

* Corresponding author.

cus and nonpathogenic Arthrobacter clinical isolates whichpresent novel phenotypes of resistance to lincosamides.

MATERIALS AND METHODSBacterial strains, culture conditions, and determination of

MICs. Clinical isolates were obtained from Hospital NuestraSeniora de Covadonga (Oviedo) and selected for resistance tolincomycin and/or clindamycin in agar-disk diffusion tests (5,ug per disk) with Mueller-Hinton agar plates. On the basis ofmorphological, physiological, and biochemical tests, theisolates were identified as staphylococci and coryneformbacteria. Strain designations for these isolates and for anti-biotic-susceptible standard strains are shown in Table 1.The organisms screened by disk diffusion assays were

grown overnight at 37°C in 2 ml of Luria broth, transferredonto Mueller-Hinton agar plates containing different antibi-otic concentrations by using a Steers replicator (34), andincubated for 16 h at 37°C.

All the cultures were grown aerobically in Luria broth at37°C and shaken at 250 rpm in a Gallenkamp orbital incuba-tor. For small cultures (10 to 25 ml), cells were grown in 100-or 250-ml Erlenmeyer flasks. For batch cultures (300 to 500ml), cells were grown in 1-liter Erlenmeyer flasks. Growthwas monitored by observing the A6. in a Unicam SP 1700spectrophotometer.

In vivo protein synthesis. Cultures were grown overnight at37°C in M9 medium (25) supplemented with 0.1% (wt/vol)yeast extract and 0.025% (wt/vol) casein hydrolysate. A10-ml sample of fresh medium was inoculated with 0.5 ml ofthe overnight culture and incubated until the A6. wasbetween 0.4 and 0.6. Samples (0.5 ml) were incubated for 5min at 37°C with different concentrations of the antibioticsand labeled for 15 min after the addition of 1 ,uCi of[3H]leucine per ml. The labeling was stopped by the additionof 1 ml of 15% (wt/vol) ice-cold trichloroacetic acid (TCA).After standing at 4°C for at least 30 min, the samples wereheated at 90°C for 15 min, filtered through Whatman GF/Cfilters, and dried, and the radioactivity in the filters wascounted.

420

RESISTANCE TO LINCOSAMIDES IN CLINICAL ISOLATES

TABLE 1. Strains used in this study and MICs of MLS antibiotics

MIC (pLg/ml)aOrganism

ERY OLE SPI CAR TYL JOS ROS LIN CLI CEL VER B

S. aureus ATCC 25923 <1 <1 2.5 <1 <1 5 <1 <1 <1 <1 2.5S. epidermidis ATCC 12228 <1 <1 <1 2.5 <1 <1 <1 <1 <1 <1 <1S. epidermidis SF4 >200 >200 >200 >200 >200 >200 >200 >200 >200 >200 >200Arthrobacter sp. strain SF19 <1 >200 75 50 50 50 50 >200 >200 >200 <1S. haemolyticus SF27 <1 <1 2.5 <1 <1 <1 <1 40 5 10 <1S. haemolyticus SF28 25 5 2.5 <1 <1 <1 <1 30 5 5 <1S. epidermidis SF30 <1 <1 <1 <1 <1 <1 <1 100 1 <1 <1Arthrobacter sp. strain SF31 <1 <1 <1 <1 2.5 <1 <1 >200 >200 >200 <1

aAbbreviations: ERY, erythromycin; OLE, oleandomycin; SPI, spiramycin; CAR, carbomycin; TYL, tylosin; JOS, josamycin; ROS, rosaramicin; LIN,lincomycin; CLI, clindamycin; CEL, celesticetin; VER B, vemamycin B.

Isolation of ribosomes. Ribosomes were isolated by a pre-viously described procedure (31) with some modifications.Thus, cells were broken by disintegration with glass beads(diameter, 0.10 to 0.11 mm) in a Braun model MSK mechan-ical homogenizer (Braun, Melsungen, Federal Republic ofGermany) for two periods of 2 min with intermittent cooling insolid CO2. The concentration of ribosomes in clean prepara-tions (A26/A280, .1.8) was calculated assuming that 1 A260contains 29.4 pmol of ribosomes. These were divided into100-,ul aliquots and kept frozen at -70°C until used.Poly(U)-dependent protein synthesis. To test the sensitivity

of ribosomes to the set of test antibiotics, an in vitrotranslation system dependent on the presence of a syntheticmessenger was used. As a source of soluble factors, S-100from a lincosamide-susceptible strain, Streptomyces antibi-oticus ATCC 11891, was used. The S-100 represents thesupernatant obtained by layering a cell extract over 2 vol-umes of high-salt buffer (1 M NH4Cl) followed by centrifu-gation at 100,000 x g and further dialysis against low-saltbuffer. Ribosomes (20 to 50 pmol) were incubated withdifferent concentrations of the antibiotics at room tempera-ture for 5 min. The reaction was initiated by adding (in a60-,u final volume) 12 ,u1 of the S-100 and 30 ,u1 of assaymixture. The assay mixture contained 40 mM HEPES (N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid)-KOH (pH7.6) at 20°C, 100 mM KCl, 17 mM MgCl2, 5 mM ATP, 0.7mM GTP, 15 mM phosphoenolpyruvate, 4.8 ,ug of pyruvatekinase, 30 ,ug of poly(U), 5 ,ug of tRNAPh,, and 1.6 ,uCi of[3H]phenylalanine. The mixture was incubated at 37°C, andsamples (10 p.1) were removed at intervals and added to 1 mlof 10% (wt/vol) TCA. After being heated at 90°C for 15 min,the samples were filtered through fiber glass disks (WhatmanGF/C), washed with 5 volumes of 5% (wt/vol) TCA, anddried, and the radioactivity was counted.

Antibiotic-inactivation assays. The presence of lincosa-mide-inactivating enzymes in the lincosamide-resistantstrains was tested with both intact cells and cell extracts. (i)Intact cells. Cells from 150 ml of an overnight culture in Lbroth were collected by centrifugation, washed once in 0.01M sodium phosphate buffer (pH 7.0), and suspended in 3 mlof the same buffer. Lincomycin or clindamycin (40 ,ug/ml)was added, and the cells were incubated at 37°C. At zerotime and after 8 h of incubation, the samples were centri-fuged in an Eppendorf minifuge and 10 p.1 of the supernatant(0.4 p.g) was spotted onto a paper disk. Residual activity ofthe antibiotic was assayed by placing the disk on a lawn ofMicrococcus luteus ATCC 10240 as test organism. (ii) Cellextracts. The S-100 of the different strains was incubated at37°C for 3 h with lincomycin or clindamycin (40 p.g/ml)together with 1 mM ATP or acetyl coenzyme A as the

potential cofactor for drug inactivation. A 10-,ul sample ofthe assay mixture was spotted onto a paper disk, and theantibiotic activity was measured by the microbiologicalassay described in i.

Chemicals. L-[4,5-3H]leucine (specific activity, 141 Ci/mmol; 5.21 TBq/mmol) and L-[2,3,4,5,6-3H]phenylalanine(specific activity, 101 Ci/mmol; 3.73 TBq/mmol) were fromAmersham International. Poly(U), tRNA (phenylalaninespecific), pyruvate kinase, ATP, GTP, sodium salt phos-phoenolpyruvate, HEPES, and acetyl coenzyme A werefrom Sigma Chemical Co. The sources of the antibioticswere as follows: erythromycin, spiramycin, carbomycin,and oleandomycin, Sigma; lincomycin, clindamycin, andcelesticetin, gifts from The Upjohn Co.; vernamycin B, giftfrom E. R. Squibb & Sons; tylosin, gift from Eli Lilly & Co.;josamycin, gift from Yamanouchi International; and rosara-micin, gift from Schering Corp. Each antibiotic was dis-solved in 50% (vol/vol) methanol at 9 mg/ml. All the otherchemicals were of analytical grade.

RESULTS

Isolation of lincosamide-resistant strains. A total of 1,213clinical isolates, mainly staphylococci and coryneform bac-teria, were screened by antibiotic disk tests for resistance tolincomycin and/or clindamycin. Of these, 289 isolates(23.8%) were found to be resistant to one or both antibiotics(each disk contained 5 p.g of antibiotic), and several isolates,each corresponding to one of six patterns of resistance tolincosamides and macrolides, were selected. Representa-tives of these groups, strains designated SF4, -19, -27, -28,-30, and -31, are shown in Table 1. We did not find anyrelationship between species or pathogenicity of bacteriaand resistance pattern. Two types of low-level resistance tolincomycin were found: one of susceptibility to erythromy-cin (strain SF27) and the other of resistance to low levels(MIC, 25 ,ug/ml) of erythromycin (strain SF28). Both strainswere also partially resistant to clindamycin and celesticetinand very susceptible to the other macrolides. Four types ofhigh-level resistance to lincomycin were detected: (i) high-level resistance to all the MLS antibiotics, the so-calledMLS phenotype (strain SF4); (ii) high-level resistance tolincosamides and oleandomycin; partial resistance to spi-ramycin, carbomycin, tylosin, josamycin, and rosaramicin;and susceptibility to erythromycin (strain SF19); (iii) high-level resistance only to a lincosamide antibiotic (lincomycin)and susceptibility to clindamycin, celesticetin, and macro-lides (strain SF30); and (iv) high-level resistance to all thelincosamides and susceptibility to macrolides (strain SF31).A characteristic of the four types of high-level resistance to

421VOL. 32, 1988

ANTIMICROB. AGENTS CHEMOTHER.

lincomycin was susceptibility to erythromycin and vernamy-cin B (except in SF4).

Resistance to lincosamides resulting from the MLS resis-tance phenotype was the most abundant pattern (243 iso-lates; 20.0%). On the other hand, the resistance patternshowed by strain SF19 was found only in Arthrobacterspecies (18 isolates), being absent in the Staphylococcusspecies screened. Resistance phenotypes corresponding tostrains SF27, SF28, SF30, and SF31 were not very frequentamong the isolates, being detected in 17, 4, 1, and 6 isolates,respectively. Two standard susceptible strains (Staphylo-coccus aureus and Staphylococcus epidermidis) were simi-larly susceptible to lincomycin (and also to the other lincos-amides, macrolides, and streptogramins) and, therefore, wechose S. aureus ATCC 25923 as a control susceptible strainfor further experiments.

Susceptibility of the different strains to other groups ofantibiotics was also tested by disk diffusion tests. While thestandard susceptible strain (S. aureus ATCC 25923) wassusceptible to most of the antibiotics, the different lincosa-mide-resistant strains were resistant to at least four antibi-otics. In some cases (strains SF4, SF28, and SF31), resis-tance to seven antibiotics (two P-lactam antibiotics and atleast three aminoglycosides) was found.

Effect of antibiotics on cell growth. Addition of erythromy-cin, lincomycin, or clindamycin (each at 50 ,ug/ml) to S.aureus ATCC 25923 growing exponentially in L broth at37°C completely stopped growth of the culture (Fig. 1).However, the growth of strains SF19 and SF31 (Fig. 1) wasblocked only by erythromycin, whereas lincomycin or clin-

2. Saureus SF19

1..

0380.4 e

* 2 S SF31

0.4.

2 4 6 8 10 2 4 6 8 10

TIME , hours

FIG. 1. Effect of antibiotics on cell growth of several lincosa-mide-resistant strains. Cells (0.5 ml) of an overnight culture in Lbroth medium were inoculated into 20 ml of fresh medium andincubated at 37°C with shaking. At the time indicated by the arrows,four 5-ml aliquots were removed to prewarmed Erlenmeyer flasksand challenged with 50 ,ug of erythromycin (0), lincomycin (A), orclindamycin (U) per ml or no antibiotic (0). Growth was recorded asthe increase in A6w of the culture. The S. aureus strain is ATCC25923.

_Z 60

LINCOZYCIN , Wl-20

10 20 30 40 s0 100LINCOMYCIN , aimI

FIG. 2. Effect of lincomycin on in vivo protein synthesis bydifferent strains. Cells (0.5 ml) in the exponential phase of growthwere incubated for 5 min at 37°C in M9 medium (supplemented asdescribed in Materials and Methods) with different lincomycinconcentrations. Then they were pulse-labeled for 15 min after theaddition of [3H]leucine (1 uCi/ml). The labeling was stopped byadding 1 ml of 15% (wt/vol) TCA, and the samples were thenprocessed as described in Materials and Methods. Symbols: 0, S.aureus ATCC 25923, 10,882 cpm; 0, strain SF19, 26,580 cpm; A,strain SF27, 8,540 cpm; A, strain SF28, 10,500 cpm; O, strain SF30,14,727 cpm; U, strain SF31, 12,300 cpm. The counts per minuteindicate the radioactive incorporation in the absence of antibioticafter the labeling period; this value was considered to be 100%7. Eachpoint value is the average of three determinations, and the standarderror was less than 4%.

damycin caused only a reduction in the growth rate. Similargrowth inhibition was found with strain SF4 (data notshown). In contrast, the growth of strain SF30 (Fig. 1) wasabolished by erythromycin and clindamycin and completelyunaffected by the presence of lincomycin.

Sensitivity of in vivo protein synthesis to antibiotics. Havingestablished the resistance patterns of the different strains,we decided to determine the effect of the antibiotics onprotein synthesis by intact cells. Therefore, we measured theincorporation of [3H]leucine into TCA-insoluble material inthe presence and absence of the antibiotics. For this andsubsequent experiments, we discarded strain SF4 since themolecular basis of the MLS resistance phenotype has beendescribed. Three of the five lincosamide-resistant strains(SF19, SF30, and SF31) were unaffected by lincomycinconcentrations of up to 100 ,ug/ml (Fig. 2); the other twostrains (SF27 and SF28) were partially affected by lincomy-cin but were less susceptible to the antibiotic than thecontrol strain. Similar results (data not shown) were ob-tained with clindamycin.

Strain SF19 was selected to test the correspondencebetween MICs and the sensitivity of in vivo protein synthe-sis. This strain was challenged with different concentrationsof lincomycin, clindamycin, erythromycin, oleandomycin,spiramycin, and carbomycin. The results (Fig. 3) show aclose correlation between both in vivo determinations. Pro-tein synthesis by SF19 was highly resistant to lincomycin,clindamycin, and oleandomycin (MICs, >200 ,ug/ml), par-tially resistant to spiramycin and carbomycin (MICs, 75 and

422 QUIROS ET AL.

RESISTANCE TO LINCOSAMIDES IN CLINICAL ISOLATES

z

MA 40

X 20

10 20 30 40 50 100ANTIBIOTIC, agml

FIG. 3. Effect of different antibiotics on in vivo protein synthesisby strain SF19. Conditions were the same as those described in thelegend to Fig. 1. Symbols: 0, lincomycin; 0, clindamycin, A,erythromycin; A, oleandomycin; O, spiramycin; U, carbomycin.After the labeling period, the incorporation of [3H]leucine in theabsence of antibiotic was 26,580 cpm (designated as 100%). Eachvalue is the average of three determinations, and the standard errorwas less than 4%.

50 ,tg/ml, respectively), and sensitive to erythromycin (MIC,'1 ,ug/ml).

In conclusion, in vivo protein synthesis by strains SF19,SF30, and SF31 was totally resistant to lincomycin, whilethat of strain SF27 and SF28 was partially resistant.

Sensitivity of ribosomes to antibiotics. The possibility re-mained that in vivo resistance of protein synthesis displayed

by the different strains might have been due to the inabilityof the antibiotics to penetrate the cell wall and the cytoplas-mic membrane of the resistant strains. To check this, 70Sribosomes were isolated from the different resistant strainsand their activity was measured in an in vitro translationsystem in the absence and presence of antibiotics. Compar-ison of the in vitro and in vivo resistance patterns of thedifferent strains revealed a close correspondence; thoseantibiotics which did not inhibit growth failed to affectpoly(U)-dependent polyphenylalanine synthesis in vitro. Ri-bosomes of strains SF19 and SF31 were insensitive tolincomycin (Fig. 4A) and clindamycin (Fig. 4B) but could bedifferentiated in their response to macrolides. Thus, thosefrom strain SF31 were very sensitive to spiramycin andcarbomycin, while ribosomes from strain SF19 were onlypartially affected by these drugs (Fig. 5). Thus, resistance toantibiotics in these strains is ribosomal in nature. In con-trast, ribosomes of strains SF27, SF28, and SF30 weresensitive to low concentrations of lincomycin (Fig. 4A) andclindamycin (Fig. 4B). Therefore, resistance to lincosamidesin strain SF27, SF28, and SF30 is not located in the ribo-somes.Assay of lincosamide-inactivating enzymes. Leclercq et al.

(24) reported a strain of S. haemolyticus which inactivatedlincomycin and clindamycin when whole cells were incu-bated with the antibiotics. We could not detect such inacti-vation of lincomycin or clindamycin in our lincosamide-resistant strains. Moreover, we also failed to detectinactivation of the antibiotics when cell extracts of thesestrains were supplemented with ATP and acetyl coenzyme Aas potential cofactors for drug modification.

DISCUSSION

Two types of resistance to lincosamides in clinical isolateshave been reported. The first one, the MLS resistancephenotype, is ribosomal in nature and is widely distributed

z120

100~~~~~~~~~~~~~~~~~~9

0.04 04 4 40 004 0.4 4 40LWJCOMIYCIN.Mgmi CLINDAMIYCIN. Mgmi

FIG. 4. Effect of lincomycin (A) and clindamycin (B) on in vitro poly(U)-dependent protein synthesis by ribosomes of different strains.Ribosomes (25 pmol) were incubated for 5 min at 37°C with different concentrations of lincomycin or clindamycin. Then the assay of proteinsynthesis was initiated by adding the remaining components of the reaction mixture, as described in Materials and Methods. After 45 min at37°C, the assay was stopped by removing 10-,ul samples and adding them to 1 ml of 10%o (wt/vol) TCA. The radioactivity incorporated intoTCA-insoluble material was then determined after processing of the samples as described in Materials and Methods. Symbols: 0, S. aureusATCC 25923, 19,320 cpm;@*, strain SF19, 18,750 cpm; A, strain SF27, 18,500 cpm; *, strain SF28, 16,540 cpm; O, strain SF30, 17,500 cpm;*, strain SF31, 16,500 cpm. The counts per minute indicate the radioactive incorporation after the labeling period in the absence of antibiotic;this value was considered to be 100%. Each value is the average of three determinations, and the standard error was less than 4%.

VOL. 32, 1988 423

ANTIMICROB. AGENTS CHEMOTHER.

100

WII

z

80

so

40

2(*

0.04 0.4 4 40

ANTIBIOTIC, jag/mI

FIG. 5. Effect of carbomycin and spiramycin on in vitro poly(U)-dependent protein synthesis by ribosomes of strain SF19. Ribo-somes (25 pmol) were incubated for 5 min at 370C with differentconcentrations of carbomycin or spiramycin. Then the assay wasdone as described in the legend to Fig. 3. Symbols: 0, S. aureusATCC 25923 against carbomycin, 19,320 cpm; 0, strain SF19against carbomycin, 18,750 cpm; A, S. aureus ATCC 25923 againstspiramycin, 19,320 cpm; A, strain SF19 against spiramycin, 18,750cpm. Each value is the average of two determinations, and thestandard error was less than 4%.

among clinical isolates of different organisms (for reviews,see references 11, 16, and 35). Many of the genes mediatingresistance to MLS antibiotics in clinical isolates are borne byplasmids, and the molecular basis for the resistance isunderstood (22, 23). The second one, inactivation of theantibiotic, has been described only in staphylococci fromdifferent origins (14, 24), and, at least in one case, theinactivating enzyme has been shown to be encoded by aplasmid (24). In both cases, the biochemical basis of theinactivation is not known. On the other hand, in nonclinicalisolates, high-level-lincomycin-resistant mutants have beenisolated from Escherichia coli which had altered ribosomalproteins (either in the S7 protein of the 30S subunit or in L14or L15 of the 50S subunit) (19). However, in these cases, arelation between the altered proteins and lincomycin resis-tance could not be proved. In addition, lincosamide-modi-fying enzymes have been described in Streptomyces spp.which phosphorylate (1, 8, 9) or adenylylate (2).The strains we describe here (SF19, SF27, SF28, SF30,

and SF31) present at least two lincosamide resistance mech-anisms that were not previously described. Firstly, consti-tutive ribosomal resistance to all the lincosamides is shownby Arthrobacter species (strains SF19 and SF31). Thesestrains possess different patterns of resistance to macrolides,SF19 being resistant to several macrolides (but not toerythromycin) and SF31 being resistant only to lincosa-mides. We do not know whether strain SF19 has a uniquemodification in the 50S subunit conferring resistance to bothmacrolides and lincosamides or two independent ribosomalmodifications, each making the ribosomal particle insensi-tive to a specific group of antibiotics. However, whateverthe basis is, it represents modification(s) in the lincosamideribosome-binding site which confers resistance to lincosa-mides independently of the MLS resistance phenotype.

Another interesting point is the finding that both strains weresusceptible to erythromycin. Lincomycin and erythromycinbind to the ribosome in partially or overlapping binding sites(but with different affinities), and cross-resistance to theseantibiotics is therefore commonly observed. However,cross-resistance to lincomycin and erythromycin does notexist in strains SF19 and SF31, indicating the high selectivityof the modification in the lincosamide-binding site.

Secondly, there may be a permeability barrier to lincosa-mides in three strains of staphylococci (namely, SF27, SF28,and SF30). This hypothesis is supported by two lines ofexperimental evidence. (i) 70S ribosomes of these strainswere sensitive in vitro to lincosamides. (ii) We could not findany specific lincosamide-inactivating enzyme either in cellextracts or in intact cells of these strains. Many, if not all,hydrophilic antibiotics with intracellular target sites areprobably actively accumulated by the cells using nutrienttransport systems (6). This could be achieved by uptakesystems for different solutes, as has been reported forfosfomycin (21, 37). In addition, it has been proposed (13)that aminoglycoside antibiotics enter cells by various trans-port systems (uptake systems for sugars, amino sugars,polyamines, and amino acids) which have slightly differentaffinities for different aminoglycosides. The alteration ofsome solute-uptake systems able to transport lincosamidesinside the cells could explain the growth of strains SF27,SF28, and SF30 in the presence of lincosamides. Interest-ingly, strain SF30 was highly resistant only to lincomycinand was susceptible to clindamycin and celesticetin, whilestrains SF27 and SF28 showed resistance to the threelincosamides, although only at a low level. Therefore, atleast in strain SF30, the alteration must be very selective. Apossible explanation is the existence of independent uptakesystems for lincosamides with different affinities for theseantibiotics. Such a situation could explain the difference inthe patterns of resistance to lincosamides observed in SF30compared with strains SF27 and SF28.

ACKNOWLEDGMENTS

We thank Eric Cundliffe for critical reading of the manuscript.This work was supported by grant 87/1426 of the Fondo de

Investigaciones Sanitarias de la Seguridad Social, Spain.

LITERATURE CITED1. Argoudelis, A. D., and J. H. Coats. 1969. Microbial transforma-

tion of antibiotics. II. Phosphorylation of lincomycin by Strep-tomyces species. J. Antibiot. 22:341-343.

2. Argoudelis, A. D., and J. H. Coats. 1971. Microbial transforma-tion of antibiotics. V. Clindamycin ribonucleotides. J. Am.Chem. Soc. 93:534-535.

3. Celma, M. L., R. E. Monro, and D. Vazquez. 1970. Substrateand antibiotic binding sites at the peptidyl transferase centre ofEscherichia coli ribosomes. FEBS Lett. 6:273-277.

4. Chabbert, Y. 1956. Antagonisme in vitro entre l'erythromicineet la spiramicine. Ann. Inst. Pasteur (Paris) 90:787-790.

5. Chang, F. N., C. J. Sih, and B. Weisblum. 1966. Lincomycin, aninhibitor of aminoacyl sRNA binding to ribosomes. Proc. Natl.Acad. Sci. USA 55:431-438.

6. Chopra, I., and P. Ball. 1982. Transport of antibiotics intobacteria. Adv. Microb. Physiol. 23:183-240.

7. Clewell, D. B., and A. E. Franke. 1974. Characterization of aplasmid determining resistance to erythromycin, lincomycin,and vernamycin Bat in a strain of Streptococcus pyogenes.Antimicrob. Agents Chemother. 5:534-537.

8. Coats, J. H. 1975. Clindamycin phosphotransferase. MethodsEnzymol. 43:755-759.

9. Coats, J. H., and A. D. Argoudelis. 1971. Microbial transforma-tion of antibiotics: phosphorylation of clindamycin by Strepto-

424 QUIROS ET AL.

RESISTANCE TO LINCOSAMIDES IN CLINICAL ISOLATES

myces coelicolor Muller. J. Bacteriol. 108:459-464.10. Courvalin, P., C. Carlier, and Y. Chabbert. 1972. Plasmid-

linked tetracycline and erythromycin resistance in group DStreptococcus. Ann. Inst. Pasteur (Paris) 123:755-759.

11. Courvalin, P., H. Ounissi, and M. Arthur. 1985. Multiplicity ofmacrolide-lincosamide-streptogramin antibiotic resistance de-terminants. J. Antimicrob. Chemother. 16:(Suppl. A):91-100.

12. Coyle, M. B., B. H. Minshew, J. A. Bland, and P. C. Hsu. 1979.Erythromycin and clindamycin resistance in Corynebacteriumdiphtheriae from skin lesions. Antimicrob. Agents Chemother.16:525-527.

13. Damper, P. D., and W. Epstein. 1981. Role of the membranepotential in bacterial resistance to aminoglycoside antibiotics.Antimicrob. Agents Chemother. 20:803-808.

14. Devriese, L. A. 1980. Two new types of resistance to lincomycinin pathogenic staphylococci from animals. Ann. Microbiol.(Paris) 131B:261-266.

15. Docherty, A., G. Grandi, R. Grandi, T. J. Gryczan, A. G.Shivakumar, and D. Dubnau. 1981. Naturally occurring macro-lide-lincosamide-streptogramin B resistance in Bacillus licheni-formis. J. Bacteriol. 145:129-137.

16. Dubnau, D. 1984. Translational attenuation: the regulation ofbacterial resistance to macrolide-lincosamide-streptogramin Bantibiotics. Crit. Rev. Biochem. 16:103-132.

17. Dutta, G. N., and L. A. Devriese. 1981. Macrolide-lincosamide-streptogramin resistance patterns in Clostridium perfringensfrom animals. Antimicrob. Agents Chemother. 19:274-278.

18. Graham, M. Y., and B. Weisblum. 1979. 23S ribosomal ribonu-cleic acid of macrolide-producing streptomycetes containsmethylated adenine. J. Bacteriol. 137:1464-1467.

19. Hummel, H., W. Piepersberg, and A. Bock. 1979. Analysis oflincomycin resistance mutations in Escherichia coli. Mol. Gen.Genet. 169:345-347.

20. Igarashi, K., H. Ishitsuka, and K. Akira. 1969. Comparativestudies on the mechanism of action of lincomycin, streptomycinand erythromycin. Biochem. Biophys. Res. Commun. 37:499-504.

21. Kahan, F. M., J. S. Kahan, P. J. Cassidy, and H. Kropp. 1974.The mechanism of action of fosfomycin (phosphonomycin).Ann. N.Y. Acad. Sci. 235:364-385.

22. Lai, C. J., and B. Weisblum. 1971. Altered methylation ofribosomal RNA in an erythromycin-resistant strain of Staphy-lococcus aureus. Proc. Natl. Acad. Sci. USA 68:856-860.

23. Lai, C. J., B. Weisbium, S. R. Fahnestock, and M. Nomura.1973. Alteration of 23S ribosomal RNA and erythromycin-induced resistance to lincomycin and spiramycin in Staphylo-coccus aureus. J. Mol. Biol. 74:67-72.

24. Leclercq, R., C. Carlier, J. Duval, and P. Courvalin. 1985.Plasmid-mediated resistance to lincomycin by inactivation inStaphylococcus haemolyticus. Antimicrob. Agents Chemother.28:421-424.

25. Miller, J. H. 1972. Experiments in molecular genetics. ColdSpring Harbor Laboratory, Cold Spring Harbor, N.Y.

26. Monro, R. E., and D. Vazquez. 1967. Ribosome-catalyzedpeptidyl transfer: effects of some inhibitors of protein synthesis.J. Mol. Biol. 28:161-165.

27. Novick, R. P., and E. Murphy. 1985. MLS-resistance determi-nants in Staphylococcus aureus and their molecular evolution.J. Antimicrob. Chemother. 16:(Suppl. A):101-110.

28. Parisi, J. T., J. Robbins, B. C. Lampson, and D. W. Hecht. 1981.Characterization of a macrolide, lincosamide, and streptogra-min resistance plasmid in Staphylococcus epidermidis. J. Bac-teriol. 148:559-564.

29. Privitera, G., A. Dublanchet, and M. Sebald. 1979. Transfer ofmultiple antibiotic resistance between subspecies of Bacte-roides fragilis. J. Infect. Dis. 139:97-101.

30. Roberts, A. N., G. S. Hudson, and S. Brenner. 1985. Anerythromycin-resistance gene from an erythromycin-producingstrain of Arthrobacter sp. Gene 35:259-270.

31. Skeggs, P. A., J. Thompson, and E. Cundliffe. 1985. Methylationof 16S ribosomal RNA and resistance to aminoglycoside antibi-otics in clones of Streptomyces lividans carrying DNA fromStreptomyces tenjimariensis. Mol. Gen. Genet. 200:415-421.

32. Skinner, R. H., and E. Cundliffe. 1982. Dimethylation of adenineand the resistance of Streptomyces erythraeus to erythromycin.J. Gen. Microbiol. 128:2411-2416.

33. Skinner, R. H., E. Cundliffe, and F. J. Schmidt. 1983. Site ofaction of a ribosomal RNA methylase responsible for resistanceto erythromycin and other antibiotics. J. Biol. Chem. 258:12702-12706.

34. Steers, E., E. L. Foltz, B. S. Gravies, and J. Riden. 1959. Aninocula replicating apparatus for routine testing of bacterialsusceptibility to antibiotics. Antibiot. Chemother. (Basel)9:307-311.

35. Weisblum, B. 1985. Inducible resistance to macrolides, lincosa-mides and streptogramins type B antibiotics: the resistancephenotype, its biological diversity and structural elements thatregulate expression-a review. J. Antimicrob. Chemother.16:(Suppl. A):101-110.

36. Welch, R. A., K. R. Jones, and F. L. Macrina. 1979. Transfer-able lincosamide-macrolide resistance in Bacteroides. Plasmid2:261-268.

37. Winkler, H. H. 1973. Distribution of an inducible hexose-phosphate transport system among various bacteria. J. Bacte-riol. 116:1079-1081.

VOL. 32, 1988 425