Isolation of methicillin-resistant coagulase-negative staphylococci from patients undergoing continuous ambulatory peritoneal dialysis (CAPD) and comparison of different molecular techniques for discriminating isolates of Staphylococcus epidermidis

2.1. Isolates 

CNS isolates were obtained from different sites: groin, nares, catheter and axilla of 11 CAPD patients attending a public hospital located in Rio de Janeiro City. The swabs were plated individually on trypticase soy agar (TSA) and the plates incubated at 37°C for 18–24 h. Five isolated colonies from each plate, with morphologic characteristics of staphylococci, were transferred to 5 individual tubes containing trypticase soy broth (TSB) and incubated for 18h at 37°C. Staphylococci were identified presumptively by Gram-stain and by free-coagulase and catalase tests. The Gram-positive, coagulase negative and catalase positive cultures were submitted to biochemistry identification using an Autoscan, Microscan System (Dade Behringer, Sacramento, USA). The strains were stored in TSB containing 10% glycerol (v/v) at −70°C.

2.2. Screening test for methicillin resistance 

This test was performed as described previously using TSA plates, containing 25 μg/mL methicillin (de Lencastre et al., 1991). After an incubation period of 24h to 48h at 35°C, any growth was taken as an indication of resistance. The methicillin-susceptible S. aureus (MSSA) strain ATCC 25923 and the methicillin-resistant S. aureus (MRSA) strain Col, obtained from Dr. Alexander Tomasz, from The Rockefeller University, NY, USA, were used as controls in this test.

2.3. Antibiotic susceptibility test 

The susceptibility test was carried out as recommended by the National Committee for Clinical Laboratory Standards (NCCLS, 1999). The disks used contained the following antibiotics: Pn, penicillin G (10 U), SxT, trimethoprim (1.25 μg) plus sulfamethoxazole (23.75 μg); Gn, gentamicin (10 μg); Tc, tetracycline (30 μg); Er, erytromycin (15 μg); Cl, clindamycin (2 μg); Co, chloramphenicol (30 μg). The strain ATCC 25923 was used for controlling the test.

2.4. DNA preparation 

All procedures for chromosomal or plasmid DNA preparations were carried out as described previously (Sambrook et al., 1989), except that bacterial cell lysis was achieved following pre-incubation of the cells with mutanolysin (103 U/ml), lysozime (7 × 103 U/ml) and Brij 35% (0,5% v/v) for 18 h at 37°C.

2.5. Pulsed-field gel electrophoresis 

The preparation of low-melting agarose inserted genomic DNA and the digestion with the restriction endonuclease SmaI was described previously (de Lencastre et al., 1994). Pulsed-field gel electrophoresis (PFGE) was performed in a CHEF-DRIII apparatus (BioRad), with a voltage of 5.5 V/cm, ramped with an initial forward time of 1s and a final forward time of 30s. Agarose DNA gels (1%; w/v) were run during 23h at 11°C, in 0.5× TBE (50 mM Tris buffer containing 50 mM boric acid and 0.2 mM EDTA, pH 8.0).

Gel staining and photography were carried out as described previously (Sambrook et al., 1989). The DNA fragments were blotted on to nylon membranes (H-Hybond, Amersham) using a vacuum transfer system, as recommended by the manufacturer (Pharmacia). After that, the DNA was fixed by baking at 80°C for 2h, and hybridized with the mecA-DNA probe. The criteria for interpreting the PFGE patterns were described previously (Tenover et al., 1995).

2.6. mecA gene polymorphism 

The agarose inserted DNA was digested with the restriction endonuclease ClaI, as recommended by the manufacturer. The DNA fragments were separated by conventional electrophoresis using a 0.8% w/v agarose DNA gel (40V/18h) in TAE buffer. The procedures for Gel staining and photography were described previously (Sambrook et al., 1989). The DNA fragments blotted on to nylon membranes by vacuum were fixed by baking and hybridized with the mecA-DNA probe.

2.7. Preparation of the mecA DNA probe 

The fluorescein-labeled mecA DNA probe was the 1.196kb Pst1-Xba1 fragment (MF13) from the mecA gene of the Australian type strain ANS46, cloned in pTZ219 (Matthews, Reed & Stewart, 1987). The digested plasmid DNA was separated by 0.8% w/v agarose gel electrophoresis. The correspondent band was excised from and purified using the Sephaglas Bandprep Kit (Pharmacia). The fragment was labeled by chemiluminescence using the ECL Random Prime Labeling System (Amersham International), as recommended by the manufacturer.

2.8. Arbitrarily primed polymerize chain reaction (AP-PCR) 

The AP-PCR reaction was performed using 50 ng DNA, 0.2 mM dNTPs (Pharmacia), 2.5U Taq DNA polymerize (Pharmacia), 1× PCR buffer (Gibco, BRL), 2.5 mM MgCl2 (Pharmacia), 50 pmol primer 1 (5′-GGTTGGGTGAGAATTGCACG-3′); van Belkum et al., 1995a). Cycling was performed in Perkin Elmer machine (Gene Amp® PCR System 9700), as described by van Belkum et al., 1995a. To assure reproducibility, all PCR reactions were carried out in at least three independent PCR amplifications, using different DNA preparations but the same batch of PCR reactions. In addition, DNA of a strain tested in previous PCR experiments (having a known AP-PCR type) was also amplified with each set of PCR reactions, for controlling the amplification conditions. Also, only one person was responsible for carrying out all the procedures for the AP-PCR.

Because most of the MRSE populations display a considerable large genetic diversity, to test the ability of the AP-PCR, using the primer 1, to identify as identical isolates having the same genetic background, we used the DNA of isolates belonging to a close related bacterial species, the S. aureus. Thus, we prepared genomic DNA from 10 well characterized MRSA isolates belonging to the Brazilian epidemic clone (BEC) displaying exactly the same genotype: PFGE type (A1), the same mecA location, ClaI/mecA polymorphism type (III), ClaI/Tn554 polymorphism type (B) and the same insertion pattern for the IS257 (Teixeira et al., 1995).

The criterion of one-band difference was used to discriminate the AP-PCR patterns. Differences due to band intensity and thickness were not taking in consideration.

3.1. Bacterial identification 

Seventy-nine CNS isolates were recovered from groin, nares, axilla or catheter of the 11 CAPD patients studied. Thirty isolates of the 79 CNS strains (38%) were methicillin-resistant. The identification of the MRCNS isolates showed that 22 (73%) were MRSE, 7 (23%) methicillin-resistant S. haemolyticus (MRSHae) and one (3%) methicillin-resistant S. hominis (MRSHom). All patients analyzed carried MRCNS in their flora (Table 1).

Table 1. Origin, molecular characterization and antibiotic resistance of methicillin-resistant coagulase negative staphylococci isolated from patients undergoing continuous ambulatory peritoneal dialysis

	Strain	Specie	Patient	Site	PFGE pattern	mecA location(Kb)/PFGE band	ClaI/mecA polymorphism	Antibiotypea
	BMB15UR	S. epidermidis	1	groin	B	253/c	2	Pn, Tc, Er
	BMB43UR	S. epidermidis	1	axilla	B1	260/c	3	Pn, Tc
	BMB31UR	S. epidermidis	3	groin	D	73/e	5	Pn, Er, Gn, Sxt
	BMB41UR	S. epidermidis	4	groin	E	231/e	6	Pn, Gn, Sxt
	BMB42UR	S. epidermidis	4	groin	E	231/e	7	Pn, Gn, Sxt
	BMB63UR	S. epidermidis	4	groin	E	231/e	7	Pn, Tc, Gn, Sxt
	BMB64UR	S. epidermidis	4	groin	E1	233/e	6	Pn, Er
	BMB73UR	S. epidermidis	4	axilla	E2	231/e	7	Pn, Gn, Sxt
	BMB48UR	S. epidermidis	5	axilla	F	301/c	8	Pn, Gn, Sxt
	BMB55UR	S. epidermidis	5	catheter	F	301/c	8	Pn, Tc, Sxt
	BMB59UR	S. epidermidis	6	groin	G	348/b	9	Pn
	BMB62UR	S. epidermidis	6	axilla	H	159/e	10	Pn, Sxt, Co
	BMB65UR	S. epidermidis	8	ND	J	358/b	9	Pn, Gn, Sxt
	BMB68UR	S. epidermidis	8	ND	J	358/b	9	Pn, Er, Gn, Sxt
	BMB69UR	S. epidermidis	8	ND	J1	368/a	13	Pn, Gn, Sxt
	BMB76UR	S. epidermidis	9	groin	L	88/h	14	Pn, Gn
	BMB77UR	S. epidermidis	9	axilla	L	88/h	14	Pn, Gn
	BMB78UR	S. epidermidis	10	axilla	M	877/a/r	15	Pn, Gn, Sxt
	BMB82UR	S. epidermidis	10	groin	N	77/l	9	Pn, Gn, Sxt
	BMB86UR	S. epidermidis	11	axilla	P	62/o	17	Pn, Sxt
	BMB87UR	S. epidermidis	11	groin	P	62/o	17	Pn, Sxt
	BMB88UR	S. epidermidis	11	nose	P	62/o	17	Pn
	BMB12UR	S. haemolyticus	1	catheter	A	193/e	1	Pn, Tc, Gn, Sxt, Co
	BMB13UR	S. haemolyticus	1	axilla	A	193/e	1	Pn, Tc
	BMB30UR	S. haemolyticus	2	groin	C	10/e	4	Pn, Er, Gn, Sxt
	BMB60UR	S. haemolyticus	7	nose	I	170/e	11	Pn, Gn, Tc, Er, Sxt, Co, Cp, Cl
	BMB61UR	S. haemolyticus	7	groin	I1	181/e	11	Pn, Tc, Er, Co, Cp, Cl
	BMB83UR	S. haemolyticus	11	nose	O	18/j	16	Pn, Tc, Sxt
	BMB84UR	S. haemolyticus	11	nose	O	18/j	16	Pn, Tc, Gn, Sxt
	BMB67UR	S. hominis	8	ND	K	225/d	12	Pn, Tc

a Antibiotype means antimicrobial resistant pattern for Pn, penicillin G (10 U); SxT, trimethoprim (1.25 μg) plus sulphamethoxazole (23.75 μg); Gn, gentamicin (10 μg); Tc, tetracycline (30 μg); Er, erytromycin (15 μg); Cl, clindamycin (2 μg); Co, chloramphenicol (30 μg).

3.2. Antibiotic susceptibility 

Besides being resistant to β-lactam antibiotics, 15 of the 22 MRSE isolates tested were also resistant to trimethoprim-sulfamethoxazole, 13 to gentamicin, 4 to tetracycline, 4 to erytromycin and 1 to chloramphenicol. Among the 7 MRSHae obtained, 5 were resistant to trimethoprim-sulfamethoxazole, 4 to gentamicin, 6 to tetracycline, 3 to erythromycin, 3 to chloramphenicol and 2 to clindamycin. The unique MRSHom isolate, in addition to β-lactam resistance, was resistant only to tetracycline (Table 1).

3.3. Methicillin resistant 

It was observed that no borderline or susceptible isolate grew on plates containing methicillin at concentration of 25 μg/ml. Any growth was taken as an indication of methicillin resistance and examined further by using a mecA-specific DNA probe. There was a 100% agreement between growth on screen agar plates and the mecA probe.

3.4. PFGE analysis 

After restriction of the genomic DNA with SmaI, PFGE resolved the 22 MRSE isolates into 11 PFGE types (B, D, E, F, G, H, J, L, M, N, and P; Table 1). The pattern B had one subtype, assigned B1, that had 2 band differences from B. Another pattern designed as E also had 2 subtypes (E1 and E2), differing in 4 bands each from pattern E. Finally, the pattern J had um subtype (J1) that differed in 2 bands (Fig. 1A,B). Among the 7 MRSHae, 4 major patterns were observed (A, C, I and O). The pattern assigned I had one subtypes (I1) differing by 2 PFGE bands. The only strain of MRSHom displayed a pattern called K. All variants of a specific type (subtypes) were isolated from a same patient. Strains of MRSE or MRSHae, showing the same or related genetic background, have not been detected colonizing different patients.

View Large Image Download to PowerPoint

Fig. 1. A and B. PFGE patterns of MRSCN isolates obtained from CAPD patients. B, B1, D, E, E1, E2, F, G, H, J, J1, L, M, N, P: PFGE patterns of isolates of MRSE, A, C, I, I1, O: PFGE patterns of isolates of MRSHae and K: PFGE pattern of the isolate of MRSHom. λ-ladder and LMW (low-molecular weight) are DNA standard markers.

3.5. mecA location 

When the SmaI fragments were transferred to a nylon membrane and hybridized with the mecA DNA probe, different patterns of hibridization were observed. Strains of MRSE or MRSHae displaying exactly the same PFGE patterns had the mecA gene located at the same PFGE fragment size and PFGE band (Table 1). In addition, isolates displaying different PFGE patterns had different SmaI/mecA polymorphisms, except for the MRSE isolate displaying subtype E2, that has the same mecA location observed for type E strains. It is interesting that all MRSE subtypes (except for subtype E2) and MRSHae subtypes I and I1 had mutations in the mecA gene or in its vicinity affecting the SmaI restriction site. Thus, the MRSE subtype B1 had the mecA gene inserted in a fragment size of 260 kb in the PFGE band “c” while pattern B had the mecA gene inserted in a fragment size of 253 kb in the band “c.” In the same way, MRSE subtype E1 had the mecA gene located in a fragment of 233 kb in the band “e” while pattern E had the mecA gene in a fragment of 231 kb in the band “e.” The same was observed for the MRSE subtype J1, in which the mecA gene was inserted in a fragment of 368 kb in the band “b” while the pattern J had the mecA in the band “b” of 358 kb. Finally, for the MRSHae subtype I1, the mecA was in a PFGE fragment of 181 kb in the band “e” while the mecA gene in the PFGE pattern I was located also in the band “e” but in a fragment of 170kb (Table 1).

3.6. ClaI/mecA polymorphism 

The chromosomal vicinity of the mecA gene in the isolates was also investigated with a DNA probe internal to the mecA gene, after ClaI restriction and conventional DNA gel electrophoresis (Table 1). Twelve ClaI/mecA polymorphisms were obtained among the 22 MRSE strains displaying 15 different PFGE patterns. However, strains belonged to PFGE pattern E displayed two different ClaI/mecA polymorphisms 6 and 7. The contrary was also observed, the same ClaI/mecA polymorphism type 9 was verified among MRSE strains belonging to three different PFGE patterns (G, J and N). Concerned to the PFGE subtypes, a different ClaI/mecA polymorphism was verified for B and B1 variants (ClaI/mecA types 2 and 3; respectively), for J and J1 (ClaI/mecA types 9 and 13, respectively) and finally for E1 and E2 (ClaI/mecA types 6 and 7).

The 7 MRSHae studied were classified in 4 ClaI/mecA polymorphisms (Table 1). However, the MRSHae isolates displaying subtype I1 had the same ClaI/mecA polymorphism of the isolate displaying PFGE pattern I (type 11; Table 1).

3.7. AP-PCR 

All ten BEC isolates used as control of the PCR reproducibility, displaying the same genotype, had exactly the same and reproducible AP-PCR pattern (Fig. 2).

View Large Image Download to PowerPoint

Fig. 2. AP-PCR pattern of different MRSA isolates belonging to the BEC type, using primer 1, showing the reproducibility of the technique to correctly identify staphylococcal strains displaying the same genetic background. λ-HindIII and 123 kb ladder are DNA standard markers.

The DNA of 20 MRSE isolates used for AP-PCR corresponded to 14 PFGE different patterns (11 PFGE types and 3 subtypes). Seventeen different amplification patterns were verified using the primer 1 (Fig. 3). Each PFGE pattern corresponded to only one AP-PCR type, except for 3 strains from pattern E, that displayed different AP-PCR types (strain BMB41UR, AP-PCR type III; BMB42UR, type IV and BMB63UR, type V) and 2 strains from PFGE pattern P, that displayed AP-PCR pattern XVI (BMB86UR) and pattern XVII (BMB87UR; Table 2). In relation to the subtypes analyzed, the pattern B isolate could be discriminated from the subtype B1, the pattern E from the subtype E2 and so as the pattern J from the subtype J1 (Table 2).

View Large Image Download to PowerPoint

Fig. 3. A and B. AP-PCR of different MRSE isolates, using primer 1. The different AP-PCR types displayed by the isolates were assigned I to XII. The difference in the run of strain 76UR and 77UR observed in this illustration was due to the amount of DNA applied in the gel. λ-HindIII and 123 kb ladder DNA are DNA standard markers.

Table 2. Classification of the isolates in genotypes by the association of PFGE, mecA location and ClaI/mecA polymorphism and comparison with AP-PCR types

	Genotype	PFGE/mecAlocation/ClaI/mecA	AP-PCR (Isolate)
	1	B/253c/2	I (BMB15UR)
	2	B1/260c/3	II (BMB435 UR)
	3	D/73e/5	VII (BMB31 UR)
	4	E/231e/6	III (BMB41 UR)
	5	E/231e/7	IV (BMB42 UR), V (BMB63 UR)a
	6	E2/231e/7	VI (BMB73 UR)
	7	F/301e/8	VIII (BMB46 UR), (BMB55UR)b
	8	G/348b/9	IX (BMB59 UR)
	9	H/159e/10	X (BMB62 UR)
	10	J/358b/9	XI (BMB65 UR), (BMB68 UR)b
	11	J1/368a/13	XII (BMB69 UR)
	12	L/88h/14	XIII (BMB76 UR), (BMB77 UR)b
	13	M/877a, 62r/15	XIV (BMB78 UR)
	14	N/771/9	XV (BMB82 UR)
	15	P/62o/17	XVI (BMB86 UR), XVII (BMB87 UR)a

a Strains of genotypes 5 and 15 had 2 different AP-PCR types each. b The same genotype was displayed by two different isolates.

The 20 MRSE analyzed was discriminated in 15 genotypes when 3 of the molecular techniques used in this study were associated (PFGE, mecA location and ClaI/mecA polymorphism). In only two cases (for genotypes 5 and 15), the AP-PCR discriminated isolates of the same genotype in different amplification patterns (Table 2). However, if the antibiotype (antibiotic resistance pattern) was also associated to the genotype, the two isolates (BMB42UR and BMB63UR) displaying genotype 5 (E/231e/7) could be further discriminated in antibiotypes Pn/Gn/Sxt and Pn/Tc/Gn/Sxt. However, we did not find any difference that justified the discrimination by the AP-PCR of the strains BMB86UR and MBM87UR; presenting both genotype 15 (P/62o/17) and antibiotype Pn/Sxt.