ARTICLE IN PRESS

1075-9964/$ - se

doi:10.1016/j.an

�CorrespondE-mail addr

Anaerobe 11 (2005) 137–143

www.elsevier.com/locate/anaerobe

Antimicrobial susceptibility

Genetic basis of resistance in Propionibacterium acnes strains isolatedfrom diverse types of infection in different European countries

Cristina Opricaa,b, Sonja Lofmarka, Bodil Lunda, Charlotta Edlunda, Lennart Emtestamb,Carl Erik Norda,�

aDepartment of Laboratory Medicine, Division of Clinical Bacteriology, F 82, Karolinska Institutet, Karolinska University Hospital Huddinge,

SE-141 86, Stockholm, SwedenbDepartment of Medicine, Division of Dermatology and Venereology, Karolinska Institutet, Karolinska University Hospital Huddinge,

Stockholm, Sweden

Received 4 October 2004; received in revised form 8 December 2004; accepted 6 January 2005

Abstract

The purpose of the study was to characterize the resistance mechanism of 36 clindamycin (CL) and erythromycin (EM) resistant

Propionibacterium acnes strains and 27 tetracycline (TET) resistant P. acnes isolates, collected from nine European countries, both

from acne patients and from patients with different infections. PCR and sequencing of the genes encoding domain V of 23S rRNA

for CL and EM resistant strains and 16S rRNA for TET resistant strains were performed. Pulsed-field gel electrophoresis was used

as a typing method to establish the relationship between resistance genotypes and pulsed-field types.

Several unique resistant genotypes were found to be distributed throughout Europe. P. acnes CL and EM resistant strains

carrying one of the mutations within the 23S rRNA were predominantly isolated from Swedish acne patients (64%) compared to

other infections (43%), OR ¼ 2.33 [CI ¼ 1.16–4.69]. Of 36 P. acnes isolates tested, none was found to carry the erm(X) resistance

gene. Forty-four percent of TET resistant strains were found to carry a G–C transition in the 16S rRNA of the small ribosomal

subunit and all these strains were isolated from Swedish acne patients. MIC of TET among all strains carrying this G–C mutation

(n ¼ 12) was 32mg/L and the MIC range for the strains where no mutation was detected ranged from 2 to 8mg/L. The MIC values

of TET were unaffected by the presence of reserpine, a well-known inhibitor of efflux pumps. Those TET resistant strains

harbouring the mutation at 16S rRNA were clustered in one pulsotype. For TET resistant strains where no mutation was found,

greater variability was noticed. No correlation was noticed between different resistance genotypes of CL and EM resistant strains

and pulsed-field types.

r 2005 Elsevier Ltd. All rights reserved.

Keywords: P. acnes; Acne; Clinical isolates; Antibiotics; Resistance mechanism

1. Introduction

Propionibacterium acnes, a resident commensal bac-terium that colonizes the lipid-rich environment of thepilosebaceous unit of the skin, produces chemotacticfactors and proinflammatory molecules responsible forthe inflammatory phase of acne [1]. This microorganism

e front matter r 2005 Elsevier Ltd. All rights reserved.

aerobe.2005.01.005

ing author. Tel.: +468 585 878 38; fax: +468 7113918.

ess: carl.erik.nord@labmed.ki.se (C.E. Nord).

has also been implicated in severe infections, especiallyin compromised patients and newborn infants [2,3].Erythromycin (EM), clindamycin (CL) and tetracy-

cline (TET) are antibiotics commonly used for acnetreatment, with therapy courses lasting from weeks tomonths. This will result in a high pressure of resistancedevelopment among P. acnes strains [4,5]. It has recentlybeen shown that antimicrobial resistance against EM,CL and TET among P. acnes strains isolated fromsystemic infections has emerged [3].

ARTICLE IN PRESS

Table 1

CL and EM resistant isolates from seven countries (n ¼ 36) and TET resistant isolates from five countries (n ¼ 27)

CL and EM resistant isolates ðn ¼ 36Þ TET resistant isolates ðn ¼ 27Þ

Infection type (number of isolates)

Croatia Blood (3) —

Czech Rep. — Skin (1), soft tissue (1), bile (1)

Finland — Head–neck (1), faeces (1)

Italy Blood (2) —

Germany Soft tissue (1) —

Greece Skin (2) Skin (1)

Norway Blood (1) Blood (1)

Slovenia Blood (3), prosthetic device (1), head/neck (1) —

Sweden Acne (22) Acne (19), bone (1)

C. Oprica et al. / Anaerobe 11 (2005) 137–143138

Ross et al. [6] have shown the genetic base ofresistance against EM and CL in cutaneous propioni-bacteria. The most prevalent mechanism was due tothree different point mutations in genes encodingdomain V, the peptidyltransferase loop of the 23SrRNA, and the authors found that each of thesemutations is associated with a specific cross-resistancephenotype. An A–G transition at base 2058 determinedstrains highly resistant to EM and variable resistance toother macrolides and CL (phenotype I). A G–Atransition at base 2057 is associated with low levelresistance to EM (phenotype III) and an A–G transitionat base 2059 is associated with high level resistance to allmacrolides and elevated but variable resistance to CL(phenotype IV) [6].The corynebacterial transposon Tn 5432 carrying

erm(X) resistance gene has been detected in P. acnes

strains highly resistant to the MLS antibiotics andcorresponds to phenotypic resistance group II [7].Generally, resistance to TET has previously been

shown to be linked to resistance to EM and CL [3,4,8]. Itwas found that a single G–C transition in the 16S rRNAof the small ribosomal subunit at Escherichia coli

equivalent base 1058 was responsible for clinical TETresistance in P. acnes [9].The purpose of the present study was to characterize

the resistance mechanisms of 36 CL and EM resistant P.

acnes strains and 27 TET resistant P. acnes strainscollected from different European countries, and origi-nating from acne patients and from patients withsystemic infections.

2. Material and methods

2.1. Bacterial strains and antimicrobial susceptibility

determinations

A total number of 36 CL and EM resistant strains and27 TET resistant clinical strains were analysed. The

isolates were collected between 1996 and 2002 fromdifferent geographical areas, both from acne patientsand from different infections (Table 1) [3,4]. The strainshad previously been identified by Gram-staining,biochemical tests (Rapid-Ana II System; REMEL Inc.,Lenexa, KS, USA) and gas-chromatographic analysis[10]. The minimum inhibitory concentrations (MICs) ofCL, EM and TET were determined by the agar dilutiontechnique as recommended by the National Committeefor Clinical Laboratory Standards (NCCLS) [11], usingbrucella base-sheep blood agar supplemented with 5%sterile defibrinated horse blood. A final inoculum of105CFU per spot was employed using a multipointinoculator. According to the European Committee onAntimicrobial Susceptibility Testing (EUCAST) recom-mendations, resistance to TET was defined atMICX2mg/L, to CL at MICX0.25mg/L and to EMat MICX0.5mg/L [3,4,12]. For DNA extraction, thebacterial strains were cultured on blood agar platesunder anaerobic conditions (BBL GasPak AnaerobicSystem, Cockeysville, MD, USA) at 37 1C for 48 h. Oneloop of approximately 1 mL bacteria was suspended in100 mL milliQ water before heated at 95 1C for 15minand centrifugated (10 000� g, 7min, 4 1C). The DNAcontaining supernatant, free from cellular debris, wasstored at �20 1C until used.

2.2. CL and EM resistant strains

2.2.1. PCR and sequencing of CL and EM resistant

strains

PCR and sequencing of the genes encoding domain Vof 23S rRNA were performed in order to detect the basisof resistance among CL and EM resistant strains. P.

acnes ATCC 6919 and P. acnes ATCC 11828 were usedas negative controls. To locate mutations, amplificationof a section of the DNA, corresponding to E. coli

equivalents as described by Ross et al. [6] encoding the23S rRNA, including the conserved domain V, wasaccomplished with a pair of 20-bp primers. Primers

ARTICLE IN PRESSC. Oprica et al. / Anaerobe 11 (2005) 137–143 139

23SF 50 CGA AAT TCC TTG TCG GGT AA 30 and23SR 50 GTA TTT CAA GGT CGG CTC CA 30 weredesigned to obtain expected fragments of 241 bp.The reactions were performed in a 50 mL volume

containing 1 mL DNA, 10�PCR reaction buffer,1.5mM MgCl2, PCR nucleotide mix including 200 mMeach of dNTP (Promega Corporation, Madison, WI,USA), 0.5 mM of each primer (ThermoHybaid GmbH,Ulm, Germany) and 0.5 U of Taq DNA polymerase(Promega Corporation). All PCR analyses were sub-jected to an initial denaturation at 94 1C (10min),followed by 30 cycles of 94 1C (30 s), 52 1C (30 s) and72 1C (1min) and a final elongation at 72 1C (10min).Electrophoresis with 1.5% agarose gels was used fordetection of amplified products.PCR products were purified using a QIAquik PCR

purification kit (QIAGEN GmbH, Hilden, Germany) toremove free primers and nucleotides. The sequenceanalysis using ABI 310 Genetic Analyser was performedaccording to the manufacturer (Perkin Elmer, FosterCity, CA, USA). Alignment analysis was made using theClustalW interactive multiple sequence alignment atEuropean Bioinformatics Institute (http://www.ebi.a-c.uk) and the sequenced region of all the isolates wascompared. All 36 CL-EM resistant strains and 10susceptible P. acnes strains were analysed.

2.2.2. Extended PCR to investigate the association of

erm(X) in CL-EM resistant strains

To detect whether the erm(X) gene was carried by CLand EM resistant strains, a pair of primers was designed(Primer Premier version 5.0) based on the sequence ofthe P. acnes erm(X) gene, GenBank accession no.AF411029: ermXF 50 ATA ACG GCA GTT GAAGTG GA 30 and ermXR 50 CGA AGA ATG GCAGTG GTG 30 giving a 167 bp fragment. The PCRreactions were performed as described above, withexception of using 2mM MgCl2. All obtained fragmentswere visualized by ethidium bromide staining after gelelectrophoresis using 3% agarose gels. Susceptiblecontrol strains P. acnes ATCC 6919 and P. acnes ATCC11828 were used, and the strain 98-4277-2 of Arcano-

bacterium pyogenes [13], kindly provided from Dr. B.H.Jost, was used as a positive control.

2.2.3. PFGE of CL-EM resistant strains

PFGE using SpeI restriction enzyme (PromegaCorporation) was used as previously described [4] inorder to establish relationship between resistancegenotypes and pulsotypes. Calculation of similaritymatrices and creation of dendrograms were done bythe Molecular Analyst Software program (Bio-RadLaboratories, Hercules, CA, USA) using the unweightedpair group method with arithmetic averages (UPGMA).The similarity coefficients were calculated according tothe method of Dice, expressed as percentages [14]. P.

acnes ATCC 6919 was used as control for each gelexperiment to allow comparisons using the MolecularAnalyst Software program. Capital letters were used todesignate the main genetic lineages of P. acnes.

2.3. TET resistant strains

2.3.1. PCR and sequencing of TET resistant strains

Amplification of a 444 bp section of DNA encodingthe 16S rRNA was accomplished with a pair of primersdesigned with the program Primer Premier version 5.0:16SF 50GAC ATG GAT CGG GAG TGC 30 and 16SR50 TCG GGT GTT ACC AAC TTT CA 30. Thereactions were performed as above, with 1.5mMMgCl2.The PCR products were purified using a QIAquik PCRpurification kit (QIAGEN GmbH) and the sequenceanalysis using ABI 310 Genetic Analyser was performedaccording to the manufacturer (Perkin Elmer). Controlstrains P. acnes ATCC 6919 and P. acnes ATCC 11828were used. All 27 TET resistant strains and 10susceptible P. acnes strains were analysed.

2.3.2. PFGE of TET resistant strains

The same methodology as described in Section 2.2.3was used to analyse the 27 TET resistant strains.

2.3.3. Reserpine test of TET resistant strains

In order to verify the contribution of active efflux tothe TET resistance phenotype, the in vitro activity ofTET was compared for all 27 strains, with and withoutthe efflux pump inhibitor reserpine [15]. The hypothesiswas that the active efflux and target gene mutationcontributes independently to TET resistance.TET was obtained from Lederle Laboratories (Pearl

River, NY, USA) and reserpine from Sigma ChemicalCo. (St. Louis, MO, USA). Reserpine solution wasfreshly prepared before use by using DMSO as asolvent. MIC determination of TET was performedwith the agar dilution method [11] using brucella base-sheep blood agar, with and without addition ofreserpine (20 mg/mL) [16]. Control strains were P. acnes

ATCC 6919, P. acnes ATCC 11828 and three clinicalisolates susceptible to TET.

3. Results

3.1. Resistant mechanisms in CL and EM resistant

strains

3.1.1. Results of PCR and target gene sequencing of CL

and EM resistant strains

The resistance mechanisms among CL-EM resistantstrains, both from acne patients and clinical isolates,including three specific previously described mutations,are presented in Table 2.

ARTICLE IN PRESS

Table 2

Resistance genotypes identified in CL-EM resistant strains (n ¼ 36) and sequence alignments of part of peptidyl transferase region of domain V of

23S rRNA

23S rRNA alleles Number of

strains acne,

Stockholm

(n ¼ 22)

Number of

strains clinical

isolates,

Europe

(n ¼ 14)

MIC range

EM (mg/L)

Acne and other

infections

(n ¼ 36)

MIC range CL

(mg/L) Acne

and other

infections

(n ¼ 36)

Group I AGACGGGAAGACCCCGGGA 10 (46%) 4 (28.5%) 0.5a–4128 0.5a–64

A 2058 G mutation (14)

Group III AGACGAAAAGACCCCGGGA 2 (9%) — 0.5–128b 0.5–1

G 2057 A mutation (2)

Group IV AGACGGAGAGACCCCGGGA 2 (9%) 2 (14.2%) 128–4128 2–32

A 2059 G mutation (4)

No mutation AGACGGAAAGACCCCGGGA 8 (37%) 8 (57.1%) 0.5–4128 0.25–64

(16)

In italics and bold represent three different mutations. MIC ranges for CL and EM resistant P. acnes strains. EM resistant strains MICX0.5mg/L

and CL resistant strains MICX0.25mg/L.aOne strain carrying a type I mutation with low MIC values both for EM and CL.bOne strain carrying a type III mutation with high level of EM resistance.

Num

ber

of s

trai

ns

0.25 0.50 1 2 4 8 16 32 64MIC clindamycin (mg/L)

0

1

2

3

4

5

6

7

Type IV mutationNo mutation detectedType I mutationType III mutation

Num

ber

of s

trai

ns

0.50 1 2 4 8 16 32 64 128 256

MIC erythromycin (mg/L)

0

1

2

3

4

5

6

7

8

9

10TypeIV mutationNo mutation detectedTypeI mutationTypeIII mutation

(a)

(b)

Fig. 1. Distribution of MIC values for CL (a) and EM (b) among CL-

EM resistant P. acnes strains.

C. Oprica et al. / Anaerobe 11 (2005) 137–143140

Different resistance genotypes were distributedthroughout Europe. The majority of the strains carryingone of the described mutations was found to belong togenotype I. Type III was less prevalent and was foundonly in 9% of resistant strains collected from acnepatients in Stockholm. P. acnes strains carrying eitherone of the mutations within the 23S rRNA werepredominantly isolated from Swedish acne patients(64%) compared to other infections (43%), OR ¼ 2.33[CI ¼ 1.16–4.69]. For 44% of the resistant strains, nomutation was detected (Table 2). None of the suscep-tible strains was found to carry any mutation within theinvestigated region of 23S rRNA.Pairwise alignment and comparison using ClustalW/

EMBL did not reveal any other unknown mutation inthe resistant strains within the DNA sequence of theinvestigated fragment.Table 2 shows the MIC range for CL and EM in

relation to the different resistance genotypes. Group Iwas generally associated with high resistance to EM andvariable resistance to CL; group III was characterizedby variable MICs for EM and CL; type IV mutationconferred high level resistance to EM and variableresistance to CL. Resistance genotype distributions inrelation to MICs for CL (Fig. 1a) and EM (Fig. 1b)showed a broad range of distribution of differentgenotypes regarding MICs for CL and a bimodaldistribution of genotypes regarding MICs for EM.According to the latter, with the exception of one strain(MIC ¼ 8mg/L), all the other strains showed EM MICvalues of either 0.5mg/L or higher than 32mg/L.

3.1.2. Detection of erm (X) gene in CL-EM resistant

strains

Of 36 P. acnes isolates analysed, none was found tocarry this resistance gene.

3.1.3. Relationship of resistance genotype to PFGE

analysis data

Fig. 2 shows the dendrogram of similarities amongthe observed PFGE patterns and Fig. 3 describes the

ARTICLE IN PRESS

Fig. 2. Dendrogram of similarities and PFGE patterns of Sma I digest

of total DNA between 36 CL and EM resistant strains. The clonal

group level was set at X75% similarity (line). Thirty-six patterns of

DNA digests grouped into five main genetic lineages of P. acnes (A–E)

were detected. The cophenetic correlation coefficient of the UPGMA

dendrogram was 78%.

PFGEclone

Num

ber

of s

trai

ns

A C D E B0

2

4

6

8

10

12

Type IVmutationNo mutation detectedType I mutationType III mutation

Fig. 3. Correlations between clonal type (PFGE) and resistance

genotype among CL-EM resistant strains.

% similarity40 50 60 70 80 90 100

A

BC

DE F

G

**********

**

Fig. 4. Dendrogram analysis of 27 TET resistant strains divided into

seven major clonal groups (A–G). The line corresponds to the cut-off

level of 75% used to delineate PFGE types. The cophenetic correlation

coefficient of the UPGMA dendrogram was 94%. Marked with * are

TET resistant strains carrying the 1058 base mutation in 16S rRNA.

They are all clustered in clone A.

C. Oprica et al. / Anaerobe 11 (2005) 137–143 141

correlations between resistance genotypes and pulso-types. Using a cut-off value of 75% among strains, aband-based cluster analysis (Molecular Analyst; Bio-Rad) revealed five different clones. Seventy-nine percentof the strains carrying mutation type I belonged to asingle clone B. The only two strains with mutation typeIII also belonged to clone B.Among the tested strains, a higher diversity was

noticed among isolates from Sweden (n ¼ 22; clusteredin four different clones) and Slovenia (n ¼ 5; clustered infive clones). A unique pulsed-field pattern D was noticedin Slovenia.

3.2. TET resistant strains

3.2.1. Identification of point mutations in TET resistant

P. acnes strains

PCR and DNA sequencing of the 27 TET resistantstrains revealed that 44% of them were found to carry aG–C transition at E. coli equivalent base 1058 in the 16SrRNA of the small ribosomal subunit. These strainswere all isolated from Swedish acne patients. The MICvalue of TET among all strains carrying the mutation(n ¼ 12) was 32mg/L, while the MIC range for strainswhere no mutation could be detected was 2–8mg/L.

3.2.2. Clonal relatedness of TET resistant strains

The clonal group level was set at 475% similarity.Among the clinical isolates, those harbouring themutation at 16S rRNA were clustered in clone A (Fig.4). For strains where no mutation was found, a greatervariability was noticed; these strains were distributedamong all seven clones with a range of similaritybetween them of 75–35%.

3.2.3. Reserpine effect on TET resistant strains

The MIC values were unaffected by the presence ofreserpine in all tested strains.

4. Discussion

EM, a macrolide, and CL, a lincosamide, arechemically distinct but share a similar mechanism ofaction, inhibiting protein synthesis by their effect on 30Ssubunit of the ribosome function [17]. Bacteria generallybecome resistant to macrolides and lincosamides in

ARTICLE IN PRESSC. Oprica et al. / Anaerobe 11 (2005) 137–143142

three ways: through target site modification by methyla-tion or mutation that prevents antibiotic binding to itsribosomal target, through efflux of the antibiotic, and bydrug inactivation [17]. Modifications of the ribosomaltarget will confer resistance to both types of antibiotics,whereas the last two mechanisms will affect only one ofthese drugs.In the present study, CL-EM resistant strains were

collected from different areas and at different timeperiods, and this is the first study describing the P. acnes

resistance mechanisms in clinical isolates collected fromdiverse infection types and regions. Different resistancegenotypes were found to be distributed through Europe,the most prevalent ones were genotypes I and IV. Theseresults are in agreement with previous studies character-izing isolates from skin [5]. Type III mutation was onlyidentified in two Swedish strains from acne patients andthis resistance genotype has previously been identifiedamong skin isolates from UK [5]. Possible, the relativelysparse prevalence of the group III resistance genotypemight be due to the fact that this mutation renders a costfor the bacteria which is not balanced by the relativelylow level of EM resistance, which it is associated with [5].The correlations between CL and EM MIC values anddifferent mutation types presented a certain pattern withtwo exceptions: one strain from the phenotypic group Ishowed a lower level of resistance to EM and CL (0.5mg/L) and the other strain from group III showed a higherlevel of EM resistance (128mg/L) compared to strainsfrom the same group. The first result might be explainedby a compensatory mutation and the second by asecondary mutation, in another region, or by the presenceof another resistance mechanism.In pathogenic bacteria, erm (erythromycin ribosome

methylase) genes encode for methylases that methylatethe N6 position of E. coli A2058 in the 23S rRNA. Theoverlapping binding sites for macrolides, lincosamides,and streptogramins B in 23S rRNA are responsible forcross-resistance to these antibiotics [17]. Nearly 40 erm

genes have been reported so far, but only the erm (X)has been found to be present in P. acnes strains. All CL-EM resistant strains were tested for the presence of erm

(X) resistance gene, but the results were negative. Thisresult was not unexpected, since in a previous report itaccounted for less than 10% of resistance among thetested strains [7].For the remaining CL-EM P. acnes resistant strains

that had no identified mutation and different resistancephenotypes, some other mechanisms may be involvedsuch as mutations in ribosomal proteins (L4 or L22)[18], transition at position 2611 in domain V of 23SrRNA [19] or deletions within domain II of 23SrRNA[20]. These possible mechanisms are only able to confermacrolide resistance and they should be accompanied byother mechanisms in order to determine the doubleresistance to EM and CL.

Extensive clinical use of TETs in acne patients hasresulted in the resistance to this group of antimicrobialagents. Bacteria commonly use three strategies tobecome resistant to TET. The most common isenergy-dependent drug efflux; the second mechanisminvolves protection of the ribosome and the thirdconfers enzymatic inactivation by the product of thetet X locus [21].The single 16S rRNA base mutation, responsible for

TET resistance, was found in 12 of the 27 P. acnes

strains. Unlike a previous study [5], in the present studyall strains carrying this mutation were found to have theMIC values of X32mg/L. The strains carrying thismutation were all collected from acne patients. Tointerpret the epidemiological importance of this ob-servation, PFGE was used as a typing method and itwas noticed that all these isolates were clustered in oneclone. This clone was found only in acne patients,heavily treated with different antibiotics. Seven strainswith low level of TET resistance and no detectedmutation within the 16S rRNA investigated region wereshown to be closely genetically related to the strainscarrying the specific mutation, indicating that themutation associated with TET resistance is a relativelyrecent event.Unlike TET resistant strains, for CL-EM resistant

strains no correlation was found between differentresistance genotypes and pulsed-field types. The twostrains belonging to group III were found to be clusteredin the same clone, but more strains have to be analysedin order to validate this observation.TET efflux offers another way of limiting access of the

antibiotic to ribosomes, and represents the most familiarmechanism of TET resistance [21]. It has been shownthat reserpine, a multi-drug-resistance efflux pumpinhibitor, could increase the intra-cellular concentrationof antibiotic in bacteria exhibiting efflux pump mediatedresistance [15]. Our results do not support the idea of thecontribution of efflux pumps in TET resistant clinicalisolates of P. acnes.

The absence of both erm(X) gene in CL-EM resistantstrains and of active efflux pump in TET resistantisolates did not signify a non-expected finding. P. acnes

has been shown to carry few copies of the rRNA [9] andit is known that the fewer rrn operons a bacteriumpossesses, the greater the likelihood is that resistancewill be conferred by rRNA mutations, rather than othermechanisms [22].

5. Conclusion

P. acnes is an important contributor to the inflam-matory response in acne lesions, but is also isolated fromserious systemic infections, especially in immunocom-promised patients.

ARTICLE IN PRESSC. Oprica et al. / Anaerobe 11 (2005) 137–143 143

In the present study, different resistance genotypeswere found to be distributed throughout Europe, amongstrains isolated both from acne patients and fromdifferent infection types. The resistant strains showedwell-known mutations in the 23S rRNA or 16S rRNA,but apparently new mechanisms of resistance haveevolved. A clonal distribution was noticed for TETresistant cutaneous P. acnes strains, carrying the single16S rRNA base mutation and presenting high levelresistance.Surveillance of both the prevalence of resistant P.

acnes strains and associated resistance mechanisms isimportant due to the rapid variation in resistancepatterns. Monitoring the clonality by using PFGE couldoffer interesting data for a better understanding ofepidemiological findings.

References

[1] Leyden JJ. The evolving role of Propionibacterium acnes in acne.

Semin Cutan Med Surg 2001;20:139–43.

[2] Jakab E, Zbinden R, Gubler J, Ruef C, von Graevenitz A, Krause

M. Severe infections caused by Propionibacterium acnes: an

underestimated pathogen in late postoperative infections. Yale J

Biol Med 1996;69:477–82.

[3] Oprica C, Nord CE. European surveillance study on antibiotic

susceptibility of Propionibacterium acnes strains. Clin Microbiol

Infect 2004 in press.

[4] Oprica C, Emtestam L, Lapins J, Borglund E, Nyberg F, Stenlund

K, et al. Antibiotic-resistant Propionibacterium acnes on the skin

of patients with moderate to severe acne in Stockholm. Anaerobe

2004;10:155–64.

[5] Ross JI, Snelling AM, Eady EA, Cove JH, Cunliffe WJ, Leyden

JJ, et al. Phenotypic and genotypic characterization of antibiotic-

resistant Propionibacterium acnes isolated from acne patients

attending dermatology clinics in Europe, the USA, Japan and

Australia. Br J Dermatol 2001;144:339–46.

[6] Ross JI, Eady EA, Cove JH, Jones CE, Ratyal AH, Miller YW, et

al. Clinical resistance to erythromycin and clindamycin in

cutaneous propionibacteria isolated from acne patients is

associated with mutations in 23S rRNA. Antimicrob Agents

Chemother 1997;41:1162–5.

[7] Ross JI, Eady EA, Carnegie E, Cove JH. Detection of

transposon Tn5432-mediated macrolide-lincosamide-strepto-

gramin B (MLSB) resistance in cutaneous propionibacteria

from six European cities. J Antimicrob Chemother 2002;

49:165–8.

[8] Eady AE, Cove JH, Layton AM. Is antibiotic resistance in

cutaneous propionibacteria clinically relevant? Implications of

resistance for acne patients and prescribers. Am J Clin Dermatol

2003;4:813–31.

[9] Ross JI, Eady EA, Cove JH, Cunliffe WJ. 16S rRNA mutation

associated with tetracycline resistance in a gram-positive bacter-

ium. Antimicrob Agents Chemother 1998;42:1702–5.

[10] Celig DM, Schreckenberger PC. Clinical evaluation of the

RapID-ANA II panel for identification of anaerobic bacteria. J

Clin Microbiol 1991;29:457–62.

[11] National Committee for Clinical Laboratory Standards. Methods

for antimicrobial susceptibility testing of anaerobic bacteria.

Approved standard M11-A6. Wayne, PA: National Committee

for Clinical Laboratory Standards; 2004.

[12] Kahlmeter G, Brown D, MacGowan A, Goldstein F, Mouton

JW, Rodloff AC, et al. The EUCAST General Committee.

EUCAST the European committee on antimicrobial susceptibility

testing. Abstracts of the 13th European congress of clinical

microbiology and infectious diseases, Glasgow, UK. Clin Micro-

biol Infect 2003;9(Suppl. 1):422.

[13] Jost BH, Field AC, Trinh HT, Songer JG, Billington SJ. Tylosin

resistance in Arcanobacterium pyogenes is encoded by

an erm X determinant. Antimicrob Agents Chemother 2003;47:

3519–24.

[14] Dice LR. Measures of the amount of ecological association

between species. J Ecol 1945;26:297–302.

[15] Neyfakh AA, Bidnenko VE, Chen LB. Efflux-mediated multidrug

resistance in Bacillus subtilis: similarities and dissimilarities

with the mammalian system. Proc Natl Acad Sci USA

1991;88:4781–5.

[16] Oh H, Hedberg M, Edlund C. Efflux mediated fluoroquinolone

resistance in the Bacteroides fragilis group. Anaerobe 2003;8:1–6.

[17] Leclercq R. Mechanisms of resistance to macrolides and

lincosamides: nature of the resistance elements and their clinical

implications. Clin Infect Dis 2002;34:482–92.

[18] Gregory ST, Dahlberg AE. Erythromycin resistance mutations in

ribosomal proteins L22 and L4 perturb the higher order structure

of 23 S ribosomal RNA. J Mol Biol 1999;289:827–34.

[19] Vannuffel P, Di Giambattista M, Morgan EA, Cocito C.

Identification of a single base change in ribosomal RNA

leading to erythromycin resistance. J Biol Chem 1992;267:

8377–82.

[20] Dam M, Douthwaite S, Tenson T, Mankin AS. Mutations in

domain II of 23 S rRNA facilitate translation of a 23 S rRNA-

encoded pentapeptide conferring erythromycin resistance. J Mol

Biol 1996;259:1–6.

[21] Speer BS, Shoemaker NB, Salyers AA. Bacterial resistance to

tetracycline: mechanisms, transfer, and clinical significance. Clin

Microbiol Rev 1992;5:387–99.

[22] Vester B, Douthwaite S. Macrolide resistance conferred by base

substitutions in 23S rRNA. Antimicrob Agents Chemother

2001;45:1–12.