Characterization of DNA Sequences Required for the CcrAB-Mediated Integration of Staphylococcal Cassette Chromosome mec, a Staphylococcus aureus Genomic Island

Abstract

The mobile element staphylococcal cassette chromosome mec (SCCmec), which carries mecA, the gene responsible for methicillin resistance in staphylococci, inserts into the chromosome at a specific site, attB, mediated by serine recombinases, CcrAB and CcrC, encoded on the element. This study sought to determine the sequence specificity for CcrB DNA binding in vitro and for CcrAB-mediated SCCmec insertion in vivo. CcrB DNA binding, as assessed in vitro by electrophoretic mobility shift assay (EMSA), revealed that a 14-bp sequence (CGTATCATAAGTAA; the terminal sequence of the orfX gene) was the minimal requirement for binding, containing an invariant sequence (TATCATAA) found in all chromosomal (attB) and SCCmec (attS) integration sites. The sequences flanking the minimal attB and attS binding sites required for insertion in vivo were next determined. A plasmid containing only 37 bp of attS and flanking sequences was required for integration into the attB site at 92% efficiency. In contrast, at least 200 bp of sequence within orfX, 5′ to the attB core, and 120 bp of specific sequence 3′ to the orfX stop site and attB core were required for the highest insertion frequency. Finally, an attS-containing plasmid was inserted into wild-type Staphylococcus aureus strains without integrated SCCmec (methicillin susceptible) at various frequencies which were determined both by sequences flanking the att site and by the presence of more than one att site on either the chromosome or the integration plasmid. This sequence specificity may play a role in the epidemiology of SCCmec acquisition.

INTRODUCTION

Since the first clinical isolate was reported in 1961, methicillin-resistant Staphylococcus aureus (MRSA) infections continue to increase in prevalence and have become a severe health problem throughout the world. According to the Centers for Disease Control (CDC) data, in 2004 MRSA infections accounted for 63% of the total number of staphylococcal infections in the United States (4). Recently, MRSA strains have been isolated in the community and appear to be more virulent than those acquired from hospitals (33). MRSA strains produce a novel penicillin binding protein (PBP), PBP2a, which has a low affinity for β-lactam antibiotics and can mediate cell wall cross-linking when susceptible PBPs are β-lactam inactivated (3, 6). PBP2a is encoded by mecA, which is located on a mobile genomic island, SCCmec. SCCmec integrates into and excises from the chromosome in a site-specific manner mediated by a group of serine recombinases, CcrAB and CcrC, encoded within the element (14, 16, 17, 19, 31). Serine recombinases, such as phage integrases, bring together two attachment (att) sites and then catalyze DNA cleavage, strand exchange, and recombination (1, 5, 28, 29, 30). For SCCmec insertion, CcrB binds to specific sites, one on the incoming, circular SCCmec and the other on the staphylococcal chromosome (attS and attB, respectively), and mediates integration with the help of CcrA by protein-protein interaction. Although CcrA itself does not bind DNA, both proteins are required for integration of attS sequences with attB (31). Such recombination leaves directly repeated copies of the att site at each end of the inserted SCCmec, known as attR and attL (Fig. 1). SCCmec excision occurs between attR and attL and regenerates attB on the chromosome and attS on the nonreplicating, circular element. Alternative att sites yielding variant excised products have been described in some MRSA strains carrying type II and type IV SCCmec (9, 18, 31).

Fig 1.

Schematic diagram of the integration and excision of SCCmec. Sequences are from S. aureus strain N315. Black and red double lines designate the chromosome and SCCmec, respectively. The orfX gene is marked as a blue dashed arrow. The att sites are marked as black short arrows. attR2 is an alternative site for excision. The sequences of att sites are shown below. The 8 identical bases are in boldface and underlined. The DNA sequences of att sites are black if they come solely from the chromosomal att site (attB) and red if they are from the circular, incoming SCCmec element (attS) or contained entirely within SCCmec (attR2). The arrows above sequences indicate inverted repeats that are either close together on the incoming circular element (attS) or at each end of the inserted element (attR1 and attL).

The chromosomal insertion site, attB, is the carboxyl terminus of a highly conserved gene of unknown function, orfX, located near the S. aureus origin of replication. Sequences downstream of attB are identical among some MRSA strains following precise excision of SCCmec but differ from those downstream of attB in methicillin-susceptible S. aureus (MSSA) strains not known to carry SCCmec (23). These data suggest that sequences flanking the attB site may determine the strain specificity of SCCmec insertion. In the following study we attempted to determine the sequence specificity for CcrB binding in vitro and CcrAB-mediated SCCmec insertion in vivo, using both standard laboratory strains and wild-type MSSA strains as recipients. Our data suggest that sequences both 3′ and 5′ of the minimal attB sequence determine the frequency and efficiency of SCCmec insertion.

MATERIALS AND METHODS

Strains and media.

All of the strains used in this work are listed in Table 1. S. aureus strains were cultured in tryptic soy broth (TSB). Escherichia coli Top10 (Invitrogen, Carlsbad, CA) was used for gene cloning. E. coli BL21(DE3) (pLys) (Invitrogen, Carlsbad, CA) was used for protein expression and purification. The antibiotics and concentrations used were as follows: 100 μg/ml ampicillin (Ap) for selection of E. coli strains following transformation and 10 μg/ml chloramphenicol (Cml), 5 μg/ml erythromycin (Erm), and 5 μg/ml tetracycline (Tet) for selection of S. aureus strains following electroporation (26) or transduction (24).

Table 1.

Strains and plasmids used in this study

Strain or plasmid	Relevant characteristic	Reference or source
Strains
E. coli
Top10	Host for DNA cloning	Invitrogen
BL21	Host for protein expression and purification	Invitrogen
S. aureus
NCTC8325	MSSA strain	24
RN4220	Restriction-defective derivative of RN450	31
RN450	NCTC8325 cured of ϕ11, ϕ12, and ϕ13	24
NRS199	MSSA strain	23
15575	MSSA strain	23
15580	MSSA strain	23
15585	MSSA strain	23
15604	MSSA strain	23
15666	MSSA strain	23
AW2	RN4220-derived strain with the first 380 bp of orfX deleted	This work
AW5	RN4220-derived strain with gfpmutinserted next to the stop codon of orfX	This work
AW6	RN4220-derived strain with gfpmutinserted 19 bp downstream of orfX	This work
AW7	RN4220-derived strain with gfpmutinserted 120 bp downstream of orfX	This work
AW8	RN4220-derived strain with the orfX and 120-bp downstream sequence deleted	This work
AW14	RN4220-derived strain with the first 280 bp of orfX deleted	This work
AW16	RN4220-derived strain with the 150-bp downstream sequence of orfX replaced by that from 15575	This work
AW17	15575-derived strain with the 200-bp downstream sequence of orfX replaced by that from RN4220	This work
Plasmids
pCN38	E. coli (Apr)-S. aureus(Cmr) shuttle vector to clone attB-containing sequences	7
pCN57	Apr Cmr; template for the amplification of gfpmut	7
pET30a	Apr; vector to clone the CcrB protein expression construct	Invitrogen
pKOR1	Apr; vector for the allelic replacement in S. aureus	2
pSC-B	Apr; vector to clone the PCR products	Stratagene
pSR	Apr Tetr Ts rep; ccrA2B2 driven by its own promoter from N315	19
pSRatt	A 1-kb fragment containing attS and attR2 amplified from circular type II SCCmec in N315 cloned on pSR, used in studies shown in Table 5	19
pWA62	A 274-bp attB-containing fragment cloned from RN450 into pSC-B	This work
pWA63	A 456-bp attS and attR2-containing fragment from pSRatt cloned into pSC-B	This work
pWA86	The first 381-bp-deleted vector of orfX derived from pKOR1 to construct mutant strain AW2	This work
pWA95	ccrB gene cloned on pET30a at the NdeI/XhoI sites to express the CcrB protein with C-terminal His tag in E. coli	This work
pWA117	Insert gfpmut next to the stop codon of orfX in RN4220 to construct mutant strain AW5	This work
pWA118	Insert gfpmut 19 bp downstream of orfX in RN4220 to construct mutant strain AW6	This work
pWA119	Insert gfpmut 120 bp downstream of orfX in RN4220 to construct mutant strain AW7	This work
pWA126	About a 400-bp BamHI-KpnI fragment from pWA62 cloned into pCN38 for cointegration with pWA159 in vivo	This work
pWA136	orfX-deleted vector derived from pKOR1 to construct mutant strain AW14	This work
pWA137	The 120-bp fragment attB-8325-6 from Table 3 cloned into pSC-B	This work
pWA138	The 120-bp fragment attB-8325-9 from Table 3 cloned into pSC-B	This work
pWA139	The 120-bp fragment attB-8325-5 from Table 3 cloned into pSC-B	This work
pWA140	The 120-bp fragment attB-8325-11 from Table 3 cloned into pSC-B	This work
pWA141	About a 300-bp BamHI-KpnI fragment from pWA137 cloned into pCN38 for cointegration with pWA159 in vivo	This work
pWA142	About a 300-bp BamHI-KpnI fragment from pWA138 cloned in pCN38 for cointegration with pWA159 in vivo	This work
pWA143	About a 300-bp BamHI-KpnI fragment from pWA139 cloned into pCN38 for cointegration with pWA159 in vivo	This work
pWA144	About a 300-bp BamHI-KpnI fragment from pWA140 cloned into pCN38 for cointegration with pWA159 in vivo	This work
pWA151	A 40-bp attS-containing fragment from pSRatt cloned into pSC-B; sequences are shown in Table 4	This work
pWA153	A 38-bp attS-containing fragment from pSRatt cloned into pSC-B; sequences are shown in Table 4	This work
pWA157	An 8-kb fragment containing Ts rep, tetL, ccrA2B2 from pSR cloned in the XbaI site of pWA151	This work
pWA158	An 8-kb fragment containing Ts rep, tetL, ccrA2B2from pSR cloned in the XbaI site of pWA153	This work
pWA159	pUC19ori, bla, Ts rep, tetL, ccrA2B2, 57-bp attS; derived from pSR	31
pWA175	A 104-bp attS-containing fragment from pSRatt cloned into pSC-B	This work
pWA178	An 8-kb fragment containing Ts rep, tetL, ccrA2B2 from pSR cloned in the XbaI site of pWA175	This work
pWA184	A 358-bp attS-containing fragment without attR2 amplified from pSRatt cloned into pSC-B	This work
pWA185	An 8-kb fragment containing Ts rep, tetL, ccrA2B2 from pSR in the XbaI site of pWA184	This work
pWA186	An 8-kb fragment containing Ts rep, tetL, ccrA2B2 from pSR in the XbaI site of pWA63 containing attS and attR2, used in studies shown in Fig. 5A and B	This work
pWA200	A 443-bp attS-containing fragment from pSRatt cloned into pSC-B	This work
pWA201	An 8-kb fragment containing Ts rep, tetL, ccrA2B2 from pSR in the XbaI site of pWA200	This work
pWA202	A 40-bp attS-containing fragment from pSRatt cloned into pSC-B; sequences are shown in Table 4	This work
pWA203	An 8-kb fragment containing Ts rep, tetL, ccrA2B2 from pSR in the XbaI site of pWA202	This work
pWA214	The first 281-bp-deleted vector of orfX in RN4220/RN450 derived from pKORI	This work
pWA216	A 41-bp attS-containing fragment from pSRatt cloned into pSC-B; sequences are shown in Table 4	This work
pWA217	A 40-bp attS-containing fragment from pSRatt cloned into pSC-B; sequences are shown in Table 4	This work
pWA218	A 35-bp attS-containing fragment from pSRatt cloned into pSC-B; sequences are shown in Table 4	This work
pWA220	A 33-bp attS-containing fragment from pSRatt cloned into pSC-B; sequences are shown in Table 4	This work
pWA221	An 8-kb fragment containing Ts rep, tetL, ccrA2B2 from pSR in the XbaI site of pWA216	This work
pWA222	An 8-kb fragment containing Ts rep, tetL, ccrA2B2 from pSR in the XbaI site of pWA217	This work
pWA223	An 8-kb fragment containing Ts rep, tetL, ccrA2B2 from pSR in the XbaI site of pWA218	This work
pWA225	An 8-kb fragment containing Ts rep, tetL, ccrA2B2 from pSR in the XbaI site of pWA220	This work
pWA226	A 37-bp attS-containing fragment from pSRatt cloned into pSC-B; sequences are shown in Table 4	This work
pWA227	An 8-kb fragment containing Ts rep, tetL, ccrA2B2 from pSR in the XbaI site of pWA226	This work
pWA229	A 150-bp fragment downstream of orfX replaced by the corresponding part from 15575, derived from RN4220	This work
pWA245	A 2.4-kb fragment containing orfX from RN450 in the BamHI-EcoRI site of pCN38 for cointegration with pWA159 in vivo	This work
pWA250	A 0.7-kb fragment containing orfX from RN450 in the BamHI-EcoRI site of pCN38 for cointegration with pWA159 in vivo	This work
pWA252	A 200-bp fragment downstream of orfX replaced by the corresponding part from RN4220, derived from 15575	This work
pWA257	A 1.8-kb fragment containing orfX from RN450 in the BamHI-EcoRI site of pCN38 for cointegration with pWA159 in vivo	This work
pWA258	A 1.1-kb fragment containing orfX from RN450 in the BamHI-EcoRI site of pCN38 for cointegration with pWA159 in vivo	This work
pWA355	A 55-bp attR2 fragment cloned onto pWA187, also containing Ts rep, tetL, ccrA2B2 from pSR; used in Fig. 6	This work

Purification of CcrB.

The ccrB gene was cloned into pET30a, generating a C-terminal His tag (pWA95), and expressed in BL21(DE3) (pLys) cells (Novagen). The cells were grown at 37°C in LB medium containing 50 μg/ml kanamycin and 34 μg/ml chloramphenicol for 2 h. The temperature was then reduced to 18°C, and isopropyl-β-d-thiogalactopyranoside (IPTG) was added to 40 μM after the optimal density at 600 nm reached 0.7. Incubation was continued overnight, and the next day cells were harvested, pelleted, and kept at −80°C. The cell pellet was resuspended in buffer A [50 mM Tris-HCl (pH 8.0), 0.3 M NaCl, 5 mM imidazole, 5% glycerol, 1 mM phenylmethylsulfonyl fluoride (PMSF), 1 mM Tris (2-carboxyethyl) phosphine (TCEP)] and ruptured using an Emulsiflex C-3 (Avestin, Inc., Ottawa, Canada) operating at >20,000 lb/in². The cell lysates were centrifuged at 25,000 rpm for 30 min, and the pellet was resuspended in 40 ml buffer B (50 mM Tris-HCl [pH 8.0], 0.3 M NaCl, 5 mM imidazole, 5% glycerol, 7 M urea, 1 mM PMSF) to solubilize CcrB. This was then centrifuged again at 25,000 rpm for 30 min, and the supernatant was loaded on an Ni-nitrilotriacetic acid (NTA) agarose column prequilibrated with buffer B. All subsequent purification steps were carried out at 4°C. CcrB bound to the Ni column was washed with buffer B and then eluted with buffer B containing increasing concentrations of 30 to 150 mM imidazole. Fractions containing CcrB were combined and dialyzed against buffer C (50 mM Tris-HCl [pH 8.5], 0.4 M NaCl, 5 M urea, 20% glycerol, 0.5 mM EDTA, 5 mM β-mercaptoethanol [β-ME], 0.2 mM PMSF). CcrB was further dialyzed against buffer C with gradually decreasing urea concentrations from 5 M to 0 M and concentrated to 12 mg/ml. Purified CcrB resulting from this process is shown in Fig. 2A. The two minor bands were shown by N-terminal sequence analysis to be cleavage products of CcrB and could not be removed by additional purification or protease inhibition. However, electrophoretic mobility shift assay (EMSA) showed that neither product bound DNA or interfered with CcrB binding (data not shown) and were, therefore, not felt to interfere with subsequent analyses. Our previous study (31) was performed with glutathione S-transferase (GST)-tagged CcrB. We have performed the EMSA described in the results below with both His-tagged and GST-tagged CcrB and found the results to be identical. However, we chose to report all of the data in this study with the smaller, His-tagged preparation.

Fig 2.

(A) A polyacrylamide gel showing purified, His-tagged CcrB used for electromobility shift analysis (EMSA). Lane 1, size markers (Precision Plus Std Prestained; Bio-Rad, Hercules, CA); lane 2, refolded His-tagged CcrB. (B) EMSA of DNA fragments described in Table 3. Concentrations (μM) of CcrB are indicated above each lane. The numerals to the left of panel 1 designate the number of bands shifted in this panel for comparison to bands in each subsequent panel.

EMSA.

EMSA was performed according to the digoxigenin (DIG) gel shift kit protocol (2nd generation; Roche, Indianapolis, IN). C-terminal His-tagged CcrB was prepared as described above. All DNA fragments used in this assay were PCR amplified using primers shown in Table 2. Fragments of ≤50 bp were synthesized and annealed to form double-stranded DNA molecules. A total of 0.8 to 1.2 ng of the DIG-11-ddUTP-labeled DNA fragments was incubated with 0 to 22 μM purified protein and 0.05 μg/μl poly(dI-dC) and 0.005 μg/μl poly-l-lysine in the binding buffer [100 mM HEPES (pH 7.6), 5 mM EDTA, 50 mM (NH₄)₂SO₄, 5 mM dithiothreitol (DTT), 1% (wt/vol) Tween 20, 150 mM KCl] at room temperature for 20 min. The samples were then separated on a 6% retardation gel (Invitrogen, Carlsbad, CA) and transferred to a positively charged nylon membrane (Roche, Indianapolis, IN). DIG chemiluminescence was detected by exposure to X-ray film.

Table 2.

Primers used in this work

Code	Use of amplicon	Sequence
ccrB-N315e-7	ccrB into the NdeI and XhoI sites of pET30a	GGAATTCCATATGCAACAACTTAAAACAAAACGTG
ccrB-N315e-8	ccrB into the NdeI and XhoI sites of pET30a	CCGCTCGAGAATAGTAAGATATAGTGTTTGGGGC
attB	attB-8325-1 for EMSA in Table 3	CATTTAAGATTATGCGTGGAGAAGC
B-8325-2	attB-8325-1 for EMSA in Table 3	TTATCGTGATATATCTTATATATTG
B-8325-3	attB-8325-2 for EMSA in Table 3	TGCACAAGGACGTCTTACAACGC
orfX-8325-8	attB-8325-2 for EMSA in Table 3	ATTGGTATATTTACTATTTTTTGTCAA
orfX-8325-9	attB-8325-3 with primer B-8325-2 for EMSA in Table 3	AACTAAAAAATTCTGTATGAGGAG
B-8325-4	attB-8325-4 with primer B-8325-3 for EMSA in Table 3	TTACTTATGATACGCCTCTCCTCGC
8325-inv-5	attB-8325-5 with primer B-8325-3 for EMSA in Table 3	GGAGAGGCGTATCATAAGTAATCGCATAATCTTAAATGCTCTGTAC
orfX-8325-7	attB-8325-6 with primer B-8325-2 for EMSA in Table 3	AACTAAAAAATTCTGTATGAGGAGATAA
8325-inv-6	attB-8325-7 with primer B-8325-2 for EMSA in Table 3	TTACTTATGATACGCCTCTCCAACTAAAAAATTCTGTATGAGGAGAT
B-8325-7	attB-8325-8 for EMSA in Table 3	ATATGCGCAATTATCGTGATATATC
B-8325-8	attB-8325-8 for EMSA in Table 3	GGCGTATCATAAGTAAAACTAAAAA
B-8325-14	attB-8325-9 with primer B-8325-7 for EMSA in Table 3	GCGTATCATAAGTAAAACTAAAAAATTCTGT
B-8325-9	attB-8325-10 with primer B-8325-7 for EMSA in Table 3	GTATCATAAGTAAAACTAAAAAATTCTGT
B-8325-11	attB-8325-11 with primer B-8325-3 for EMSA in Table 3	TTTTTAGTTTTACTTATGATACGCC
B-8325-26	attB-8325-23 with primer B-8325-3 for EMSA in Table 3	TTACTTATGATACGCCTCTCCTCGC
B-8325-15	attB-8325-12 with primer B-8325-3 for EMSA in Table 3	TTTTACTTATGATACGCCTCTCCTCGC
B-8325-16	attB-8325-13 with primer B-8325-7 for EMSA in Table 3	GCGCATCATAAGTAAAACTAAAAAATTCTGTATGAGGAGAT
B-8325-17	attB-8325-14 with primer B-8325-3 for EMSA in Table 3	GCGTATCGTAAGTAAAACTAAAAAATTCTGTATGAGGAGAT
orfX-KN-1	An upstream fragment for the orfX deletion in AW8, AW2, and AW14	GGGGACAAGTTTGTACAAAAAAGCAGGCTGGAGAAAGAGATGACGTTAGAACGCGTG
orfX-KN-2	An upstream fragment for the orfX deletion in AW8, AW2, and AW14	CCGGAATTCGTCGTTTGGTCCTCCAAATTGTTA
orfX-KN-3	A downstream fragment for the orfX deletion in AW2	CCGGAATTCAACTAAAAAATTCTGTATGAGGAGA
orfX-KN-4	A downstream fragment for the orfX deletion in AW2	GGGGACCACTTTGTACAAGAAAGCTGGGTTTTCTGTAGTTGAAGAATTCGATAG
orfX-KN-5	A downstream fragment for the orfX deletion in AW8	CGGAATTCTAGCTAACAGTTGAAATTAGGCCCT
orfX-KN-6	A downstream fragment for the orfX deletion in AW8	GGGGACCACTTTGTACAAGAAAGCTGGGTTCAATTCCGAAGTCATAATCAATCA
orfX-KN-8	A downstream fragment with primer orfX-KN-3 for the orfX deletion in AW14	CCGGAATTCCAACGCATGACCCAAGGGC
mR8	A 460-bp attS-containing fragment on pWA186	ATGAAAGACTGCGGAGGCTAACT
mL1	A 460-bp attS-containing fragment on pWA186	GAATCTTCAGCATGTGATTTA
attS-2-19	An upstream fragment with mL1 for the attR2 site deletion	GGGGTACCTAAAAACATAACAGCAATTCACATAA
attS-2-17	A downstream fragment for the attR2 site deletion	GGGGTACCTGAGGTTCATGATTTTTGACATAGT
attS-2-18	A downstream fragment for the attR2 site deletion	ATAGATCACCTTTATCCCAATCAGA
attS1-2	A 358-bp attS-containing fragment on pWA185 with mL1	TGTTCTGATACATTCAAAATCCCTT
attS1-1	A 104-bp attS-containing fragment on pWA178 with attS1-2	GGACAATTAAAAAAACCTCATCATT
attS-N315-1	A 933-bp attS-containing fragment on pWA238	ACTGCTTGGGTAACTTATCATGGA
attS-N315-2	A 933-bp attS-containing fragment on pWA238	ATCACAGTAGTGCAAAGCACGTCGA
orfX-1	A 3.7-kb attB-containing fragment for pWA245	TGCCCGGAATTCCCTTCAACACAAGAAACATG
orfX-4	A 3.7-kb attB-containing fragment for pWA245	CGGCCCAAGCTTGGATCAGTCGCATCAAATGT
orfX-8325-B-1	A 707-bp attB-containing fragment on pWA250	CGCGGATCCAAAAGAGAAATATTGGAAGCAAGCC
orfX-8325-B-2	A 707-bp attB-containing fragment on pWA250	CGGAATTCAGCTACGTCAATGACTGTTGATTATG
cL1	A 274-bp attB-containing fragment on pWA126	ATTTAATGTCCACCATTTAACA
cR1	A 274-bp attB-containing fragment on pWA126	AAGAATTGAACCAACGCATGA
pCN38-1	Verify the insertion in the vector pCN38	CTGTTAACTTACTAACTCTTTCCGCAAA
pCN38-2	Verify the insertion in the pCN38 vector	CCGCCTTTGAGTGAGCTG
pT3	Verify the insertion in the pSC-B vector	AATTAACCCTCACTAAAGGGAAC
pT3-rev	Verify the insertion in the pSC-B vector	CTGCAGCGACCAATGTGG
orfX-g-1	A 1-kb upstream fragment for gfpmut insertion in AW5	GGGGACAAGTTTGTACAAAAAAGCAGGCTGCATAGAGGAACTGTTAGTAGCGG
orfX-gfp-1	A 1-kb upstream fragment for gfpmut insertion in AW5	CGCGGATCCTTACTTATGATACGCCTCTCCTCGC
orfX-gfp-2	A 1-kb upstream fragment for gfpmutinsertion in AW6 with orfX-g-1	CGCGGATCCCATACAGAATTTTTTAGTTTTACTTATGA
orfX-gfp-3	A 1-kb upstream fragment for gfpmutinsertion in AW7 with orfX-g-1	CGCGGATCCATGCGCAATTATCGTGATATATCTT
orfX-gfp-4	A 1-kb downstream fragment for gfpmut insertion in AW5, AW6, and AW7	CGGAATTCTAGCTAACAGTTGAAATTAGGCCCT
orfX-gfp-7	A 1-kb downstream fragment for gfpmutinsertion in AW5, AW6, and AW7	GGGGACCACTTTGTACAAGAAAGCTGGGTTCAATTCCGAAGTCATAATCAATCA
gfpmut-1	An ∼700-bp gfpmut for the insertion in AW5, AW6, and AW7	CGCGGATCCATTTAAGAAGGAGATATACATATGAG
gfpmut-2	An ∼700-bp gfpmut for the insertion in AW5, AW6, and AW7	CGGAATTCTATTTGTATAGTTCATCCATGCCA
orfX-8325-p	A 1-kb upstream fragment of the attB site with orfX-KN-1 for AW16	[p]TTACTTATGATACGCCTCTCCTCG
orfX-15575-1	A 150-bp 3′ flanking sequence of the attB site amplified from 15575 for AW16	[p]TGAGGTTCATGATTTTTGACATAGT
orfX-15575-2	A 150-bp 3′ flanking sequence of the attB site amplified from 15575 for AW16	CGGAATTCTTTTACCGTAATTTTACTATATTTATTTAC
orfX-15575-4	A 1-kb upstream fragment of the attB site amplified from 15575 for AW17	GGGGACAAGTTTGTACAAAAAAGCAGGCTTCCCCAAATACACAAGTTATACACTG
orfX-15575-5	A 1-kb upstream fragment of the attB site amplified from 15575 for AW17	[p]TTACTTATGATACGCCTCTCCACG
orfX-15575-8	A 214-bp 3′ flanking sequence of the attB site amplified from RN4220 for AW17	[p]AACTAAAAAATTCTGTATGAGGAGATAATAA
orfX-15575-9	A 214-bp 3′ flanking sequence of the attB site amplified from RN4220 for AW17	CGGAATTCATTTAATAAGGACAAAAAGAAGCATTC
orfX-15575-6	A 1-kb downstream fragment of the attB site amplified from 15575 for AW17	CGGAATTCGAGCGAATTTCCAATGATAGACAAC
orfX-15575-7	A 1-kb downstream fragment of the attB site amplified from 15575 for AW17	GGGGACCACTTTGTACAAGAAAGCTGGGTTTCAAAGCAAACATTTCATTTAGCG
attR2-55	A 55-bp fragment containing attR2 site, the sequence is from N315	TTATGTGAATTGCTGTTATGTTTTTAAGAAGCTTATCATAAGTAATGAGGTTCAT
attR2-55-r	A 55-bp fragment containing attR2 site, the sequence is from N315	ATGAACCTCATTACTTATGATAAGCTTCTTAAAAACATAACAGCAATTCACATAA
att-15580-1	Assess the insertion of pSRatt at an attsite 7 kb downstream of orfX in 15580	ATTTGTATTATCGGAGAGCTATGTT
att-15580-2	Assess the insertion of pSRatt at an attsite 7 kb downstream of orfX in 15580	TAATAAATTTCAATGTCGCTAAGTG

Construction of attS-containing plasmids.

17 attS-containing fragments were cloned into plasmids that were used as integration vectors and are shown in Table 1. The basic method was as follows. Fragments either were amplified with primers shown in Table 2 or, if less than 40 bp, were synthesized and complementary strands were annealed. PCR amplicons or synthesized oligonucleotides were ligated into commercial vector pSC-B (Stratagene, La Jolla, CA). An 8.1-kb XbaI fragment containing a copy of native ccrA2B2 with its own promoter, a temperature-sensitive origin of replication and a tetracycline resistance gene, all from pSR (19), were then inserted at the XbaI site of pSC-B, yielding the pWA series of plasmids shown in Table 1.

Construction of orfX-deleted mutants.

Chromosomal sequences within and flanking the orfX gene in RN4220 were deleted by allelic replacement mutagenesis, using vector pKOR1 (2). This plasmid allows efficient cloning without the use of restriction enzymes and ligase, using the lambda recombination cassette (Invitrogen, Carlsbad, CA) and positive selection for secondary recombination. As shown in Table 1, a series of strains containing the orfX deletions, designated AW2, AW8, AW14, AW16, and AW17, were created by PCR-amplifying fragments flanking the orfX deletion of interest using primers shown in Table 2. The amplicons were digested by EcoRI and ligated overnight; the ligated product was mixed with the plasmid pKOR1 and the BP clonase II enzyme (Invitrogen, Carlsbad, CA); the mixture was incubated at room temperature for 5 h, and the resulting construct was then transformed into E. coli TOP10 competent cells. The plasmids (pWA136, pWA86, pWA214, pWA229, and pWA252) were recovered from E. coli, introduced into RN4220 by electroporation (29), and selected on TSB agar plates containing 10 μg/ml chloramphenicol. A colony of RN4220 containing the transformed plasmid was grown at 42°C to inhibit replication of the ts plasmid, and colonies growing on chloramphenicol were again selected, presumably containing only chromosomally integrated sequences. Colonies that had undergone secondary recombination between the sequences flanking orfX, expelling the plasmid and the orfX gene, were selected by inducing the culture with 1 μg/ml of anhydrotetracycline. The plasmid contains an antisense construct to the essential gene secY that is driven by a tet-inducible promoter. Any colonies that still contain the integrated plasmid and have not undergone secondary recombination will not grow following tetracycline induction. Chromosomal deletion of orfX was confirmed by PCR amplification and DNA sequencing.

Construction of plasmids for attB-attS cointegrate formation.

The vectors carrying various sizes of attB-containing fragments for cointegration with the attS-containing vector pWA159 were based on the S. aureus-E. coli shuttle vector pCN38 (7). The attB-containing fragments were amplified from the RN450 genome using the primers shown in Table 2. The plasmids and the PCR amplicons used for their construction are shown in Table 1. attB-containing vectors were electroporated into the orfX-deleted strain AW8 containing pWA159.

gfpmut insertion in RN4220.

The insertion of the green fluorescent protein gene (gfpmut) into the chromosome of RN4220 was performed by selection of allelic replacement mutations in the chromosome of S. aureus using pKOR1 as described above. Three insertion points were chosen in relation to orfX: one next to the stop codon, a second 19 bp downstream, and a third 120 bp downstream. The construction of the first insertion will be described in detail (generating insertion plasmid pWA117; Table 1). The other two insertion plasmids were constructed in a similar manner (pWA118 and pWA119). Two 1-kb fragments, one upstream and one downstream of orfX, were amplified (primers orfX-g-1 and orfX-gfp-1 for upstream, orfX-gfp-4 and orfX-gfp-7 for downstream). These sequences were used only to provide homology for integration and did not change any sequence upstream of the gfpmut insertion sites, as determined by DNA sequencing of the final constructs. A gfpmut gene was amplified from pCN57 (7) by using primers gfpmut-1 and gfpmut-2, and the downstream and gfpmut fragments were digested by EcoRI and ligated together overnight. A second round of amplification was used to ligate the products using primers gfpmut-1 and orfX-gfp-7. The PCR product was then digested with BamHI and ligated to the BamHI site of the upstream fragment. The ligated product containing the gfpmut insertion was mixed with pKOR1 and BP clonase II and incubated at room temperature for at least 5 h. Two microliters of the recombination product was then transformed into E. coli TOP10 competent cells, generating pWA117. Each of the three insertion vectors, pWA117, pWA118, and pWA119, was introduced into RN4220, and colonies were selected on TSA plates containing 10 μg/ml chloramphenicol. Allelic replacement with inducible counterselection was performed as described above, creating gfpmut insertion strains AW5, AW6, and AW7 (Table 1). Insertions were confirmed by PCR and DNA sequencing.

Statistical analysis.

One-way comparisons of means with a control were made by Dunnett's method, while single comparisons were performed using the Wilcoxon rank-sum test. Statistical software used was JMP (version 8.0; SAS Institute, Cary, NC).

RESULTS

Determination of the att-associated CcrB DNA binding site.

A comparison of MRSA att sites (attB, attS, attR, and attL) revealed a 21-bp common core consensus sequence (GGAGAGGCGTATCATAAGTAA) (Fig. 3). This 21 bp and flanking sequences found around the strain 8325 (NCTC8325) (Table 1) attB site (orfX carboxyl terminus) were used in EMSA to assess the minimal sequence required for DNA binding (Table 3). We used large fragments with 100 bp of DNA flanking either the 5′ or the 3′ end of the 21 bp to determine specific sequences required for binding. The 21-bp sequence could be shortened to 14 bp at its 5′ end with no flanking DNA, and CcrB binding still occurred (attB-8325-13), but the 3′ end required at least 7 bp of flanking DNA (attB-8325-11), with no sequence specificity (attB-8325-14), for gel shift to take place at the lowest concentration of CcrB protein.

Fig 3.

(A) Sequence alignment of different att sites. Asterisks are used to designate conserved bases in all of the att sequences. The rectangle shows the 14-bp minimal attB-containing sequence required for gel shift by EMSA. Among these, 8 bases (TATCATAA; shaded sequence) are highly conserved in all att sites. (B) The 21 nucleotides that form the center of the attB site around which DNA binding and strand exchange occur in strains 8325 and N315. They are numbered for reference in the text.

Table 3.

Characterization of the DNA element required for CcrB binding by EMSA^a

Fragment	Sequenceb	Shift abilityc
attB-8325-1	10 bp---ATGCGA GGAGAGGCGTATCATAAGTAA AACTAAAAAA-80 bp	+++
attB-8325-2	100 bp-ATGCGA	−
attB-8325-3	AACTAAAAAA-100 bp	−
attB-8325-4	100 bp-ATGCGA GGAGAGGCGTATCATAAGTAA	−
attB-8325-5	100 bp-ATGCGA TTACTTATGATACGCCTCTCC	+++
attB-8325-6	GGAGAGGCGTATCATAAGTAA AACTAAAAAA-100 bp	+++
attB-8325-7	AATGAATACTATGCGGAGAGG AACTAAAAAA-100 bp	−
attB-8325-8	GGCGTATCATAAGTAA AACTAAAAAA-100 bp	+++
attB-8325-9	GCGTATCATAAGTAA AACTAAAAAA-100 bp	+++
attB-8325-13	CGTATCATAAGTAA AACTAAAAAA-100 bp	+++
attB-8325-10	GTATCATAAGTAA AACTAAAAAA-100 bp	−
attB-8325-12	100 bp-ATGCGAGG AGAGGCGTATCATAAGTAA AA	−
attB-8325-23	100 bp-ATGCGAGG AGAGGCGTATCATAAGTAA AACTAA	++
attB-8325-11	100 bp-ATGCGAGG AGAGGCGTATCATAAGTAA AACTAAAAAA	+++
attB-8325-14	100 bp-ATGCGAGG AGAGGCGTATCATAAGTAA GGGGGTTAT	+++

^aDIG-labeled double-stranded DNA was mixed with 0, 0.11, 1.1, 5.5, and 22 μM CcrB-His. After a 30-min incubation at room temperature, the samples were separated on a 6% retardation gel.

^bUnderline, central binding sequence; italics, reverse complement of the central sequence; bold, inverted central sequence.

^c+++, shifted with ≤0.11 μM CcrB-His; ++, shifted with 1.1 or 5.5 μM CcrB-His; +, only shifted with 22 μM CcrB-His; −, no shift.

To further identify the essential bases among the 14 required for binding, single base changes were introduced, mutating each base into the other three, and EMSA was repeated for each of the mutant fragments. Since the sequence TATCATAA is conserved in all of the att sites (Fig. 3), mutations targeted only these bases. The only single base that could not be substituted with any of the others without completely abolishing gel shift was the C at position 13 in the 21-bp central fragment shown in Fig. 3B. Single bases of the three-base palindrome around the C could be changed one at a time, and gel shift was retained, but two base changes to any other base within one of the three-base palindromes or deletion of a single base within one of the palindromes was not tolerated. Thus, the core appears to require a central cytosine around which there is a three-base pair palindrome, two bases of which on both sides need to be retained for CcrB binding to occur. The cytosine appears to be an essential base for DNA binding, but this may not be the crossover point for strand exchange during catalysis. As shown in Fig. 1, strand exchange may occur 10 nucleotides 5′ to this cytosine.

Characterization of the functional attS sequence in vivo.

Circular SCCmec is known to integrate into the chromosome by the specific sites attS on the incoming SCCmec element and attB on the chromosome. As shown in Fig. 3, the attS site shares an 8-bp core sequence with attB, and as detailed above, this core sequence is required for CcrB binding. We next sought to establish the sequences flanking the binding site required for CcrAB-mediated recombination and integration in vivo. We first investigated the integration of sequences flanking attS that are present in S. aureus strain N315. Sequences of various lengths and base compositions were introduced into a plasmid that contained a temperature-sensitive origin of replication, a tetracycline resistance gene (tetL), and the ccrAB operon driven by its native promoter (pWA187). The target attB site was in the chromosome of RN4220, the strain in which all of these experiments were conducted. Integration percentage was assessed by the ratio of the number of colonies growing on tetracycline agar following growth at 42^o (the nonpermissive temperature for plasmid replication; all colonies represented integration of the plasmid into the chromosome) to the number of colonies growing at the permissive temperature for replication (30°C). Results are the means from at least three experiments. Integration was confirmed by PCR amplification and sequence determination of junction fragments from 6 individual colonies for each integration experiment. The results of the integration studies are shown in Table 4 and can be summarized as follows. First, there are a minimum number of bases flanking the core sequence that are required for integration. The fragment could be reduced in size progressively from 456 to 37 bp with retention of >90% integration frequency. However, just as with DNA binding, there needed to be a minimum number of bases both 3′ and 5′ to the core for efficient integration.

Table 4.

Chromosomal integration of plasmids containing attS

Vector	attS sequence	Length (bp)	Integration frequency (%)
pWA186	267 bp-AACCTCATCATTAACTGATACGCAGAAGCATATCATAAATGATGCGGTTT-139 bp	456	8.80 ± 1.52
pWA201	267 bp-AACCTCATCATTAACTGATACGCAGAAGCATATCATAAATGATGCGGTTT-126 bp	443	92.83 ± 3.15
pWA185	267 bp-AACCTCATCATTAACTGATACGCAGAAGCATATCATAAATGATGCGGTTT-41 bp	358	96.90 ± 0.28
pWA178	13 bp-AACCTCATCATTAACTGATACGCAGAAGCATATCATAAATGATGCGGTTT-41 bp	104	98.95 ± 1.48
pWA159	AAAAACCTCATCATTAACTGATACGCAGAAGCATATCATAAATGATGCGGTTTTTTa	56	96.80 ± 4.52
pWA157	CATCATTAACTGATACGCAGAAGCATATCATAAATGATGC	40	96.00 ± 5.66
pWA158	CTGATACGCAGAAGCATATCATAAATGATGCGGTTTTTT	38	93.65 ± 1.63
pWA203	ATTAACTGATACGCAGAAGCATATCATAAATGATGCGGTT	40	91.80 ± 6.56
pWA221	ACCTCATCATTAACTGATACGCAGAAGCATATCATAAATGA	41	86.38 ± 3.01
pWA222	AAAAACCTCATCATTAACTGATACGCAGAAGCATATCATAAA	40	0
pWA227	TTAACTGATACGCAGAAGCATATCATAAATGATGCGG	37	91.86 ± 2.22
pWA225	AACTGATACGCAGAAGCATATCATAAATGATGC	33	0
pWA223	TACGCAGAAGCATATCATAAATGATGCGGTTTTTT	35	0

^aThe inverted repeat sequences present on the circular N315 SCCmec element are underlined.

Second, when a second att site, attR2 (31) (Fig. 1), was included in a 456-bp fragment (pWA186), the integration frequency dropped from >90% to 8.8%. The decrease in integration frequency was confirmed to be due to having both attS and attR2 on the same fragment by replacing the 19-bp attR2 site with a 6-bp KpnI site and observing a restoration in integration frequency to >90% (pWA201). Additional experiments performed using attS and attR2 alone showed that the attR2 site by itself had an integration frequency into attB that was the same as that of attS alone.

Finally, there are inverted repeats of 8 to 26 bp present at each end of the integrated SCCmec sequence just internal to the attR and attL sites in published SCCmec sequences (17, 19). These sequences come together upon excision and flank the attS site as part of the circular element. The importance of these inverted repeats as part of the attS flanking sequences required for insertion was examined using a 56-bp construct that contained the complete imperfect 26-bp inverted repeat found in strain N315 (19) (pWA159 in Table 4; illustrated in Fig. 1). The fragment was shortened at its 5′ end, removing 17 of the 26 nucleotides in the repeat (pWA158), and at its 3′ end, removing 11 nucleotides (pWA221), and the fragments still integrated at 94% and 86% frequencies, respectively. Thus, CcrAB recognizes the core and specific flanking sequences on incoming SCCmec DNA but needs only from 5 to 20 bp of these flanking sequences to mediate highly efficient integration. The inverted repeat can be reduced by two-thirds and integration still occurs at wild-type frequency.

Characterization of the functional attB site in vivo.

The role that sequences flanking the attB core and binding sequences play in SCCmec insertion was investigated in two ways. First, the RN450/RN4220 attB site and various sequences flanking this site were included on a plasmid that was introduced into the orfX- and attB-deleted strain AW8 that also contained pWA159. The latter plasmid contained attS sequences that were found to integrate into the chromosomal attB site at a frequency of 96% (Table 4). Recombination of the two plasmids to form cointegrates was the measure of CcrAB-mediated insertion of attS into attB sequences and was assessed by the ratio of tetracycline-resistant cells at the nonpermissive versus the permissive temperature for plasmid replication as with chromosomal integration. The presence of plasmid cointegrates was confirmed by PCR of cointegration junctions and the detection of a larger-sized plasmid on agarose gels. Using this technique, the maximum frequency of plasmid cointegration was found to be only 5%. We have previously shown that there is a critical ratio between the quantity of CcrA and CcrB and target sites (31). This ratio was likely disrupted by the high plasmid copy numbers. However, there was enough difference between the cointegration frequencies associated with selected sequences that we could detect significant differences using this method. The results are shown in Fig. 4. The highest cointegration frequency of 5.52% ± 0.32% was found when a 2.4-kb fragment, with 1.2 kb flanking each side of attB (pWA245), was used. Either the 5′ (pWA258) or the 3′ (pWA257) fragment flanking attB could be reduced by 1.2 kb with no decrease in cointegration frequency (5.31% ± 0.07% and 5.20% ± 0.52%, respectively), but when both 5′ and 3′ flanking sequences were reduced to a total fragment length of 700 bp, the frequency dropped to 0.96% ± 0.13% (Fig. 4). Similarly, as the total fragment length was further reduced, including fragments that bound to CcrB in gel shift experiments (attB-8325-5, attB-8325-6, attB-8325-8, and attB-8325-11, respectively) (Table 3), cointegration frequencies dropped to undetectable (no colonies at the nonpermissive temperature and 10⁸ at the permissive temperature) (Fig. 4). These data suggested that a far more specific sequence flanking attB was required for SCCmec insertion than was required for attS.

Fig 4.

Characterization of attB-flanking sequences required for attS integration determined by plasmid cointegration formation. The schematic map and sequences illustrate attB-carrying fragments of different sizes used for plasmid cointegration. Plasmids pWA141, pWA142, pWA143, and pWA144 contained fragments with attB-8325 numbers that were shown to bind CcrB protein (Table 3). The orfX gene is marked as a striped arrow. Both vectors were introduced into the orfX-deleted strain AW8. The integration frequency was measured by the ratio of colonies surviving at 42°C over those at 30°C (%) on TSA plates with 5 μg/ml tetracycline. Zeroes indicate no colonies were found out of 10⁸ plated. The integration was verified by PCR and DNA sequencing.

Second, in order to directly assess sequences surrounding attB required for SCCmec insertion, we introduced deletions or insertions around the attB site in RN4220. We used the integration plasmid pWA186, containing both attS and attR2, for these studies because its lower integration frequency (8.8%; Table 4) allowed us to detect differences in integration frequency; the high integration frequency (>95%) with the other plasmids masked these differences. The specificity and nature of CcrAB-mediated integrations of pWA186 in these experiments were assessed by DNA sequence analysis of the insertion sites from 10 colonies of RN4220 and 10 colonies from each of the five insertion site mutants shown in Fig. 5 and discussed below. All insertions were into the attB site. However, in all 60 insertion site sequences, there had been a recombination between the attS and attR2 sites on pWA186, deleting the intervening 84 bp, such that only the attR2 site integrated into attB. There was no integration of control plasmids containing no att sequences, further demonstrating the attB site specificity of integration.

Fig 5.

Integration of the attS-containing vector pWA186 into wild-type and attB-flanking sequence mutant strains. (A) Integration of pWA186 into the attB upstream-sequence-deleted mutants. In AW2, the N-terminal 380 bp of the orfX gene were deleted, leaving only the C-terminal 100 bp of sequence, including the attB site. In AW14, the N-terminal 280 bp of orfX were deleted, leaving 200 bp. (B) Integration of pWA186 into the attB downstream-sequence-inserted mutants. A copy of gfpmut was inserted just after the stop codon of orfX (AW5), 19 bp downstream (AW6), and 120 bp downstream (AW7). *, P < 0.05; **, P < 0.0001 (each AW strain compared to RN4200).

The first mutant targets created were strains AW2 and AW14, which had deletions of orfX leaving just 100 bp or 200 bp 5′ of the attB site, respectively. In AW2, the integration frequency of pWA186 decreased dramatically to 0.69% ± 0.01%, whereas in AW14 it only decreased by 15.7% to 7.52% ± 0.85%, suggesting that 100 to 200 bp of the sequence 5′ to the attB site were required for efficient integration (Fig. 5A). Mutant sequences 3′ to the attB site were created by inserting a gfpmut2 gene at three locations: immediately following the stop codon of orfX (AW5), 19 bp downstream (AW6), and 120 bp downstream (AW7) of orfX. Compared to the RN4220 parent, the AW5 integration frequency was reduced by 85% (1.3% ± 0.5% versus 8.8% ± 1.6%), AW6 by 50% (4.4% ± 0.7% versus 8.8% ± 1.6%), and AW7 by 28% (6.4% ± 0.9% versus 8.8% ± 1.6%) (Fig. 5B). Thus, at least 120 bp of the specific sequence 3′ of attB are required for efficient CcrAB-mediated SCCmec integration. It is not likely that any other genes are contained within this DNA that facilitate integration. There are no other genes within the 200 bp 5′ to the attB site except the orfX gene, and while there is a 150-bp open reading frame (ORF) within the 120-bp 3′ sequences, this ORF is not present in many of the MSSA strains that supported high-frequency integration.

The integration of plasmids containing attSCCmecII in MSSA strains.

Noto et al. (22, 23) have previously shown that when SCCmec was excised, by overproducing CcrAB, from strains carrying type I (COL), type II (N315), and type IV (NRS384 and J39) SCCmec, the sequences for 120 bases 3′ to the orfX stop codon were identical. An analysis of sequences 3′ to orfX in 11 MSSA strains found that these sequences differed from those of strains in which CcrAB-mediated SCCmec excision had been verified (23). It was hypothesized that these strains could only support SCCmec integration and excision at a low frequency. In order to test the integration aspect of this hypothesis, six of these MSSA strains were chosen for study. First, the 14-bp attB site and 100 bp of 3′ DNA were amplified from each strain and examined by EMSA for CcrB binding. Each fragment demonstrated gel shift at the lowest concentration of protein, confirming that the sequence of DNA flanking the core is not a determinant of protein binding. A plasmid containing attS was then introduced into each strain, and insertion frequency was measured, using integration into RN4220 as the standard. A plasmid with an integration frequency into RN4220 of 96% (pWA159) was first chosen for study and was found to integrate into the attB site of each of the 7 MSSA strains at a similar frequency of >90%. Thus, as with integration studies of attB mutants, an integration plasmid was next chosen with a relatively low integration frequency so that significant increases and decreases compared to the standard could be assessed. This integration plasmid, pSRatt, contained a total of 1 kb of sequences, including attS and attR2, and CcrAB expressed from their native promoter. It integrated into RN4220 at a frequency of 1.2% (31). The pSRatt plasmid was able to insert into the attB site of each MSSA strain at a frequency higher than that of the RN4220 control strain (Table 5). In order to assess the role of 3′ sequences in the increased integration frequency, 150 bp of DNA 3′ to orfX from MSSA 15575, which integrated at a frequency four times greater than that of RN4220, was inserted into the chromosome downstream of orfX in RN4220 (strain AW16), and RN4220 sequences were inserted into the chromosome at the same location in MSSA 15575 (AW17). The integration frequency doubled for AW16 (P = 0.01) and decreased 3-fold for AW17 (P = 0.01).

Table 5.

Chromosomal integration of attS-containing plasmids into MSSA strains

Strain	No. of att sites	Character of attB structure	Integration frequency (%)
RN4220	1	8325	1.25 ± 0.16
RN450	1	8325	1.18 ± 0.25
15580	2	MRSA252 after type II SCCmec excision	4.08 ± 0.24
15585	1	Followed by unknown ORF with no known homologues in staphylococci	3.73 ± 0.27
15575	1	MW2 after type IV SCCmec excision	3.89 ± 0.44
15604	1	Followed by a putative restriction-modification system	2.34 ± 0.20
15666	1	Followed by unknown ORF with no known homologues in staphylococci	1.27 ± 0.15
NRS199	1	Followed by a putative restriction-modification system	1.61 ± 0.25
AW16	1	8325; a 150-bp fragment just downstream of orfX came from 15575	2.17 ± 0.24
AW17	1	15575; a 200-bp fragment just downstream of orfX came from RN4220	1.25 ± 0.10

We also assessed the integration of att site-containing plasmids into the chromosome of strain 15580 (23), which has two att sites, one at orfX and another approximately 6.3 kb downstream separated by five ORFs of unknown function. We used the following integration plasmids: pSRatt, containing both attS and attR2; pWA159, containing only attS; and pWA355, containing only attR2. The overall integration frequency of pSRatt into strain 15580 was 4.08%, four times greater than that into RN4220 (Table 5). The overall integration frequency of both pWA159 (attS alone) and pWA355 (attR2 alone) was >90% (data not shown). Forty colonies were picked for pSRatt insertions and 30 each for pWA159 and pWA355, and their insertion sites were assessed by PCR using the primers shown in Fig. 6. The sequence of insertion sites was determined for 10 sites for each of the three inserted plasmids. As seen in Fig. 6, 93.3% of the pWA159 (attS alone) insertions were into the attB orfX site, either with (23.3%; site 1) or without (70%; site 3) the 6.3 kb of DNA separating the two att sites. In contrast, 73% of the pWA355 (attR2 alone) insertions were into the second, non-orfX att site. The pSRatt insertions were also predominantly (65%) into the second, non-orfX att site (site 2), consistent with sequence analysis showing that, as with pWA186, the two att sites (attS and attR2) had recombined into a single attR2 site. Thus, the presence of non-attS integration sequences on SCCmec can direct insertion into sites other than attB.

Fig 6.

Diagram of the insertion sites for plasmids containing attS alone (pWA159), attR2 alone (pWA355), and both attS and attR2 (pSRatt) into the chromosome of MSSA strain 15580. The arrows designate the sites and names of primers used to identify the insertion sites into attB (site 1), alternate att site (site 2), or the hybrid att site when the DNA between sites 1 and 2 had excised (site 3). The table at the bottom compares the frequency and sites of insertion of the relevant plasmids.

Excision of integrated sequences.

Another possible determinant of integration frequency is the rate of excision of integrated sequences. That is, if integration takes place at a high frequency but the excision frequency is just as high, then the overall observed integration frequency will appear to be low. In order to test this hypothesis, colonies were picked from strains that had integrated plasmids at high frequency (pWA159, pWA201) and those that had low integration frequencies (pWA186, pSRatt). Colonies from these strains were grown overnight in broth without antibiotics at 30°C and then plated on agar containing tetracycline and incubated at 30°C and at 42°C. Any excised plasmids would be unable to replicate at 42°C, and thus the ratio of colonies growing on tetracycline-containing agar at 30°C and at 42°C would be an indication of the spontaneous excision frequency. All of the integrated plasmids were stable, with a >95% ratio of colonies growing on tetracycline agar at 42°C versus 30°C. In addition, four colonies were picked from each plate, and PCR was performed using primers that would amplify only circular, excised plasmids. There were no PCR products indicating plasmid excision, while all strains had PCR products from integrated plasmids. Thus, the integration frequencies observed in our experiments were a result of the sequences included on the integration plasmids and not a balance between integration and excision.

DISCUSSION

The family of large serine recombinases, which includes CcrB, has separate domains for DNA binding and catalytic functions. Binding takes place at specific sites on the chromosome (attB) and on the incoming element (SCCmec) (attS), with subsequent joining of the two sequences by DNA cleavage and strand exchange. In this study, we sought first to determine the CcrB DNA binding sequences in vitro using EMSA and then both incoming element and target chromosomal sequences required for integration in vivo. We found the minimal nucleotide-specific chromosomal binding sequence (attB) of strain 8325 to be 14 bp (CGTATCATAAGTAA), of which 8 bp (TATCATAA) are common to attB and attS sites from SCCmec types I to IV that contain ccrAB (Fig. 3). Our gel shift data, coupled with mutational analysis, showed the following. First, the central cytosine residue was essential; it could not be mutated without abolishing gel shift. Second, more nucleotides were required to the right of the cytosine (at least 14) than to the left (only 5), and the last six of the nucleotides to the right could be completely nonspecific. However, the orientation around the cytosine was partially sequence specific, because if the 21-bp central sequence (Fig. 3B) was inverted there was still a requirement for at least 14 bases, now located to the left of the central cytosine (Table 3, attB-8325-7). These observations are better understood with reference to the solved structures of two serine recombinases that have close resemblance to the predicted structure of CcrB: transposon protein γδ and the plasmid resolvase SIN (21, 32). Both SIN and γδ bind DNA as dimers but in an asymmetric fashion, with the helix-turn-helix motif of one monomer binding first to one side of the binding site followed by the binding of the second monomer. Initial binding of the first monomer is required for binding of the second, and more nucleotides are contacted during the interaction of the first monomer with DNA than with the second. For CcrB, the 14 nucleotides to the right of the essential cytosine required for binding would mark the contact point of the first monomer, with protein-nucleotide interactions farthest from the cytosine more nonspecific compared to the specific interactions closer to the cytosine. The second monomer, binding after the first, contacts only nucleotides to the left of the cytosine. This order of binding would occur even if the sequences were inverted relative to flanking DNA. However, these data seem inconsistent with the position of the apparent crossover point for catalysis, as shown in Fig. 1. The points where the attR1 and attL sequences diverge from attB and attS (change from black to red) are 10 nucleotides removed from the cytosine identified as the only nucleotide essential for binding. This discrepancy could be explained if the gel shifts with shorter sequences were the result of monomer rather than dimer binding. This is suggested in Fig. 2 by the presence of more shifted bands at lower CcrB concentrations for the longest oligomer (attB-8325-1) than for the shorter oligomers, a result of the binding of larger protein complexes. Thus, more nucleotides to the left of the cytosine would be included in dimer binding and would involve the crossover site. This would suggest that the first monomer binds most avidly to a sequence that does not include the crossover site and that catalysis occurs only with the binding of the second monomer to form a dimer.

In contrast to the minimal sequences required for binding, longer sequences and more extensive flanking DNA were required for integration in vivo. While only 37 bp of DNA, including the binding core sequence, were required for attS, 100 to 200 bp of DNA in orfX, 5′ to the chromosomal attB binding core, and at least 120 bp of DNA 3′ to attB were needed for maximal integration frequency. Large phage serine recombinases such as ϕC31and Bxb1 bind to both the chromosomal attB and phage attP sites as a dimer, forming a tetramer at the synapse of the two att sites (11, 12). Large protein complexes like these may require extensive DNA contact for proper orientation of the catalytic unit. Both CcrB and CcrA have been shown to be required for integration of attS-containing plasmids, with CcrA binding to CcrB (31), possibly to assist in orientation of the binding complexes and/or catalysis at the synapse (25). It was also interesting that the inverted repeats found at each end of SCCmec elements that contain CcrAB (types I to IV) (15, 17, 19) seemed to play no role in attS-containing plasmid integration. It is possible that inverted repeats have a role in excision rather than integration. There is evidence that both sequence and recombinase requirements for integration and excision differ. For example, certain phage, such as ϕC31, TP901-1, and Bxb1, require the participation of an additional protein, Xis, for excision to occur (11, 12, 28). ϕC31 integrase alone is catalytically inert on att substrates (attL and attR) following phage integration (12).

Specificity of the sequence flanking the 3′ end of the binding site may determine the strains into which SCCmec can disseminate. The acquisition of SCCmec and spread of MRSA clones worldwide has occurred within a fairly narrow range of sequence types and clonal complexes, as determined by multilocus sequence typing (10). In this study, we assessed the ability of an SCCmec construct to insert into wild-type MSSA strains of various multilocus sequence types and with a variety of sequences 3′ to the orfX attB site. We had previously found that the sequences flanking the 3′ end of attB in strains from which SCCmec had been excised were identical (23), leading us to hypothesize that MSSA strains of sequence types not commonly found among MRSA and with differing 3′ flanking sequence would not support insertion of SCCmec. We found that to be only partially true in this study. While exchange of the sequence 3′ to attB between RN4220 and an MSSA strain significantly altered the insertion frequency into each strain, the insertion of attS-containing plasmids into 7 MSSA strains occurred at frequencies at least equal to RN4220. Factors other than insertion site sequences that may also be important in determining successful SCCmec acquisition may include recipient strain fitness following mecA acquisition (20) and the SCCmec excision rate. A balance that favored excision over insertion could yield an insertion-negative phenotype, though none was observed here.

Another factor that affects the insertion frequency is the presence of multiple att sites, either on the incoming element or on the chromosome. A second att site (attR2) is present on the incoming SCCmec DNA of all type II to IV SCCmec elements at the attR end of the element (this is absent in type I elements). We showed in this study that the presence of two att sites on the insertion plasmid markedly lowered the insertion frequency. One explanation for the lower insertion frequency may be found in the sequence of the insertion site. In every insertion site sequenced, the two plasmid att sites had undergone recombination, leaving only the attR2 site. The low insertion frequency may therefore be due to competition between the two sites for integration versus recombination, both potentially mediated by CcrAB. It is not clear why recombination does not occur between the two att sites on incoming SCCmec type II to IV elements but may relate to the high copy number of plasmid constructs as opposed to the single copy of incoming SCCmec elements. The presence of multiple att sites on the chromosome can also affect the number and type of inserted elements. Clinical staphylococcal isolates can acquire heterologous elements, such as ACME, at a duplicate att site flanking SCCmec, and complex SCCmec elements can be constructed that contain more than one set of Ccr genes (8, 9, 13, 14, 27). In this study, we found that the CcrAB-mediated insertion of plasmid constructs containing either or both of two att sites (attS and attR2) directed insertion to a different one of three chromosomal att sites in a clinical MSSA isolate.

Thus, we have shown in this study that while SCCmec and chromosomal integration sequences required for CcrB binding are short, DNA sequences flanking the att sites required for element integration are more extensive and to some degree specific, suggesting a role in strain preference for insertion. In addition, the presence of multiple att sites on either the chromosome or SCCmec can alter site specificity and lead to the accumulation of complex SCCmec elements.