Characterization of SCCmec Instability in Methicillin-Resistant Staphylococcus aureus Affecting Adjacent Chromosomal Regions, Including the Gene for Staphylococcal Protein A (spa)

ABSTRACT

Staphylococcal cassette chromosome mec (SCCmec) represents a sequence of clear clinical and diagnostic importance in staphylococci. At a minimum the chromosomal cassette contains the mecA gene encoding PBP2a but frequently also includes additional antibiotic resistance genes (e.g., ermA and aadC; macrolide and aminoglycoside resistance, respectively). Certain regions within SCCmec elements are hot spots for sequence instability due to cassette-specific recombinases and a variety of internal mobile elements. SCCmec changes may affect not only cassette stability but the integrity of adjacent chromosomal sequences (e.g., the staphylococcal protein A gene; spa). We investigated SCCmec stability in methicillin-resistant Staphylococcus aureus (MRSA) strains carrying one of four SCCmec types cultured in the absence of antimicrobial selection over a 3-month period. SCCmec rearrangements were first detected in cefoxitin-susceptible variants after 2 months of passage, and most commonly showed precise excision of the SCCmec element. Sequence analysis after 3 months revealed both precise SCCmec excision and a variety of SCCmec internal deletions, some including extensive adjacent chromosomal loss, including spa. No empty cassettes (i.e., loss of just mecA from SCCmec) were observed among the variants. SCCmec stability was influenced both by internal mobile elements (IS431) as well as the host cell environment. Genotypically similar clinical isolates with deletions in the spa gene were also included for purposes of comparison. The results indicate a role for host-cell influence and the IS431 element on SCCmec stability.

INTRODUCTION

Methicillin-resistant Staphylococcus aureus (MRSA) continues to be a pathogen of global importance (1). Genes associated with methicillin resistance, primarily mecA, are encoded in a family of mobile genetic elements called staphylococcal cassette chromosome mec (SCCmec) (2). Some SCCmec elements also encode other antimicrobial resistance genes such as ermA or aadC (macrolide and aminoglycoside resistance, respectively). The stable chromosomal location of SCCmec, due to specific insertion into the orfX locus, has allowed development of a variety of PCR-based diagnostic tests, including rapid detection of MRSA in nasal swabs, blood culture bottles, and wound specimens (3–6). These tests have been used successfully for almost a decade. However, SCCmec is a known hot spot for DNA rearrangements, with 14 currently recognized types and numerous subtypes (7). This instability can lead to deletions of a portion or all of the SCCmec element, sometimes encompassing adjacent chromosomal sequences (8–14), such as the gene for staphylococcal protein A (spa) (15–17) positioned relatively close (e.g., ca. 40 kb) to SCCmec. SCCmec-associated rearrangements have been primarily described in clinical isolates, presenting challenges to molecular detection. However, little is known regarding the extent to which such events might occur in nonclinical settings, serving as a reservoir of potentially problematic strains upon migration to a clinical environment. We sought to investigate this issue by assessing the frequency and extent to which deletions associated with, but not limited to, SCCmec might spontaneously occur over time in different MRSA populations growing in the absence of selection. For purposes of comparison, S. aureus clinical isolates identified as mecA-negative and spa-nontypeable were also included in the analysis.

RESULTS

Longitudinal SCCmec stability.

The stability of the four different SCCmec elements during 3 months of serial subculturing at room temperature is shown in Table 1. While strain growth differences in brain heart infusion (BHI) broth may have had a minor influence on the frequency of deletion mutants, resulting in methicillin susceptibility over time (e.g., BK11515; Table 1), this did not appear to be widespread since significant differences between susceptible and resistant organisms were not observed in spectrophotometric growth curve experiments (data not shown). SCCmec type I in S. aureus COL was stable (i.e., no variants detected) over the 3-month period. However, when transduced from strain COL to strain RN4220 (i.e., strain CRG2358), all susceptible isolates exhibited extensive SCCmec I instability after 3 months of subculture involving deletions within SCCmec extending to adjacent chromosomal regions, including spa.

TABLE 1.

Frequency of methicillin-susceptible mutants recovered during serial subculture

MRSA strain	SCCmec type	Mutants recovered (%) during:
		Mo 1	Mo 2	Mo 3
COL	I	0.00	0.00	0.00
CRG2358	I	0.00	0.00	0.71a
HFH-30820	II	0.00	8.33	12.12b
BK11515	III	0.00	0.25	0.12c
FPR3757	IV	0.00	0.53	0.94c

^a100% SCCmec internal deletions extending to the adjacent chromosome. No empty cassette or site-specific total cassette excision detected.

^b42% site-specific total cassette excision, 58% SCCmec internal and internal plus adjacent chromosomal deletions. No empty cassettes detected.

^c100% site-specific total cassette excision.

Strain HRH-30830 (SCCmec II) exhibited the highest frequency and greatest variety of susceptibility-associated deletions after 3 months of subculture. PCR testing indicated that 42% of deletions were precise SCCmec excision, while the remainder of the isolates showed a mixture of internal SCCmec deletions and deletions of SCCmec plus adjacent chromosomal sequences (Fig. 1).

FIG 1.

Diagrammatic representation of sequenced deletions (indicated by the pattern bars) associated with SCCmec I, II, and IV (A, B, and C, respectively) in S. aureus strains. In each case, SCCmec internal deletions are shown in the upper diagram. Instances of deletions including adjacent chromosomal regions are indicated by exit arrows and shown in larger context in the lower diagram. Clinical isolates are indicated with an asterisk.

PCR analysis confirmed site-specific total SCCmec excision in all cefoxitin-susceptible BK11515 (SCCmec III) and FPR3757 (SCCmec IV) isolates, which was initially detected at 2 months of subculture.

Interestingly, empty cassettes, i.e., the specific loss of only mecA from the SCCmec element and no other sequences (9), were not detected in cefoxitin-susceptible isolates from any of the SCCmec types.

WGS analysis of SCCmec-associated deletion.

Cefoxitin-susceptible colonies derived from CRG2358 and HRH-30830 after 3 months of subculture were shown by PCR analysis to have deletions other than precise SCCmec excision. The genomes of these colonies were analyzed by whole-genome sequencing (WGS) (Fig. 1). CRG2358 (SCCmec I) derivatives exhibited four deletion patterns (labeled CRG2522, 2523, 2524, and 2525) compared to the parent strain (Fig. 1A). In each case, deletions began at the 3′ end of IS431, which is 5′ to mecA, and extended approximately 126 kb to 201 kb, including spa, which is approximately 40 kb downstream from SCCmec I. The deletions observed in HFH-30820 and its derivatives containing SCCmec II were more variable, showing eight different patterns (Fig. 1B); however, the deletions always originated at the 3′ end of one of the two IS431 elements flanking plasmid pUB110 preceding (i.e., 5′ to) mecA. In three instances (CRG2531, 2532, and 2537), deletions ranging from 18 kb to 33 kb were internal to SCCmec II and included loss of mecA. The remaining deletions (CRG2530, 2534, 2535, 2036), which ranged from 56 kb to 78 kb, eliminated SCCmec II and adjacent chromosomal regions, the largest of which (CRG2536) came within 7 kb of (but did not include) spa.

WGS analysis of clinical isolates with suspected SCCmec deletions.

Seven oxacillin-susceptible S. aureus clinical isolates (Table 2), all of which were negative for spa in a commercial assay, had deletions in or around the SCCmec II and IV elements, as determined by PCR amplification. WGS analysis (i.e., whole-genome multilocus sequence typing [wgMLST] and traditional MLST) revealed six of the isolates to most likely have previously carried SCCmec II (i.e., sequence type 5 [ST5] or ST105 and the presence of pUB110 in two of the isolates; Fig. 1B). Similar to the results found with longitudinal analysis, when mapped against SCCmec II strain Mu50, each of these isolates showed deletions originating at the 3′ end of IS431 either 5′ or 3′ to plasmid pUB110 and extending from 88 kb to 141 kb, always including the remainder of SCCmec and spa (Fig. 1B). The SCCmec IV isolate 16320 (ST8) was mapped to SCCmec IV reference strain FPR3757. In comparison to the precise SCCmec excision seen in longitudinal analysis, the isolate exhibited extensive deletion (Fig. 1C) also originating at the 3′ end of IS431-mec and extending to remove the remainder of SCCmec and spa, ca. 63 kb or 94 kb depending on whether an ancestral arginine catabolic metabolic element (ACME) adjacent to SCCmec may have been present.

TABLE 2.

Bacterial strains

Strain	Relevant characteristics	Reference or source
MRSA stability isolates
COL	ST5, SCCmec I	GenBank accession no. CP000046.1
CRG2358	RN4220, ST8, transduced with SCCmec I from strain COL (CP000046.1)	This study
HFH-30820	ST 5, SCCmec II	ATCC BP-BAA-1699
BK11515	ST72, SCCmec III	Kreiswirtha
FPR3757	ST8, SCCmec IV, ACME	30
mecA-spa-variant clinical isolatesb
CRG2256	ST5, SCCmec IIc	Iowa
14625	ST105, SCCmec IIc	Maryland
14636	ST5, SCCmec IIc	Connecticut
14638	ST5, SCCmec IIc	Oregon
14671	ST105, SCCmec IIc	New York
14758	ST105, SCCmec IIc	Kansas
16320	ST8, SCCmec IVc	Smolensk, Russia

^aCourtesy of Barry Kreiswirth, Public Health Research Institute, New York.

^bCepheid strain collection.

^cPresumed ancestral SCCmec type based on ST.

For each of the sequenced isolates (in vitro or clinical) deletions specifically began at the 3′ end of an IS431 element but terminated randomly either within SCCmec or in adjacent chromosomal regions (often including spa) with no apparent similarity to what might have represented a target sequence.

DISCUSSION

Since its discovery in 2000 (2), SCCmec architecture has been crucial to our understanding of the development and spread of the methicillin resistance gene mecA among staphylococcal species, especially S. aureus. Previous studies have reported three different iterations of SCCmec rearrangement—precise SCCmec excision, empty cassettes (loss only of mecA), and partial loss of internal SCCmec sequences, which may or may not extend into the adjacent chromosome. This study generated two of the three deletion types, i.e., precise excision of the entire SCCmec element with no loss of adjacent chromosomal sequences, and the partial loss of internal SCCmec sequence but often including much larger deletions, up to 201 kb. The deletions begin at IS431 in the SCCmec element and extend downstream to include spa and other chromosomal sequences. What was not observed among any of the cultured strains or clinical isolates was the empty cassette phenomenon (16). This may be an artifact of the selection of the clinical strains for the study, since they were all suspected of having spa deletions, but the lack of empty cassettes among the isolates passaged in the absence of selection was surprising. There is some evidence for the role of antibiotic pressure (i.e., vancomycin) in inducing empty cassettes (9).

Key aspects of past studies have been deletions associated with SCCmec II IS431 elements (18–23). However, to date, questions regarding the role of specific SCCmec architecture, other host-cell influence, and the extent to which growth in the absence of selection might relate to such rearrangements have not been examined. As a step toward addressing these issues, we examined the stability of different common SCCmec types (I through IV) during 3 months of growth under nonselective conditions. Included were two different strains carrying isogenic SCCmec I elements, the first with the element found natively in strain COL and the second when the element was transduced from COL to RN4220 (CRG2358). Of the strains tested, SCCmec I in strain COL was the most stable, with no deletions detected after 3 months of passage. However, in the CRG2358 host, deletions in SCCmec I were seen at 3 months, all involving extensive loss (i.e., 126 to 201 kb) of adjacent chromosomal sequence including spa (Fig. 1A). This implicates as yet unknown host factors in SCCmec stability. Host cell influence was also implicated in SCCmec IV carriage. This element in FPR3757 exhibited instability at 2 months, which was due to precise excision of SCCmec. However, clinical isolate 16320, which is highly related to FPR3757 by conventional and wgMLST, exhibited loss of SCCmec and extensive adjacent chromosomal sequence including spa. Taken together, these results also point to uncharacterized host cell factors external to but influencing SCCmec stability that are deserving of further investigation.

WGS analysis clearly confirmed the influence of IS431 on SCCmec stability. In every case, deletions in susceptible isolates (whether of in vitro or clinical origin) began at the 3′ end of the IS431 elements within SCCmec. Deletions of SCCmec II showed more variability than SCCmec I, perhaps owing to the two IS431 elements flanking pUB110. However, multiple IS431 elements alone were not the sole influence, since SCCmec III in BK11515 was stable over 3 month’s longitudinal analysis despite containing four IS431 elements (i.e., flanking internal SCCmec and pT181) (24). Factors affecting the extent of IS431-associated deletions also remain unclear since termination did not occur in recognizable target sequences. Events such as one-ended transposition, producing identical deletion starting points but asymmetrical terminations in plasmids (25), IS3-related “single-ended attack” where a 3′ OH group at a cleaved IS element promotes nucleophilic attack on the opposite end (26), and studies of IS-associated large-scale genomic rearrangements in Gram-negative organisms (27) may hold clues to what was observed here and deserve further investigation. Recombinase genes ccrA and ccrB, known to be involved in insertion and excision within SCCmec, may also play a role here (28, 29).

This study has several limitations. First, this study would have benefited from inclusion of additional SCCmec types and host cell backgrounds. In addition, the inclusion and sequence analysis of isolates with empty cassettes may have indicated other rearrangements not previously considered relevant. Nonetheless, our data serve as a proof of principle and provide a sense of the multifactorial nature of SCCmec rearrangements. While antibiotic selection may influence SCCmec loss and rearrangement (9), study results indicate that cassette-associated changes may frequently occur in nature even in the absence of antimicrobial selection providing a potential reservoir of S. aureus variants. These include precise cassette excision and internal SCCmec deletions often extending to the adjacent chromosome. Rates of change were difficult to quantitate due to the influence of host strain and SCCmec type. The interrelationship between SCCmec architecture and host-cell influence identified here deserves additional investigation to more fully understand the dynamics of SCCmec stability and maintenance.

MATERIALS AND METHODS

Bacterial strains.

The bacterial strains used in the study are listed in Table 2. For in vitro analysis, MRSA strains with SCCmec types I through IV were employed. Seven clinical isolates of oxacillin-susceptible S. aureus (six containing SCCmec type II and one containing SCCmec type IV) from different patients and diverse geographic locations were also studied as examples of naturally occurring deletion mutants.

Longitudinal serial subculturing.

For SCCmec stability studies, MRSA isolates were initially grown on brain heart infusion (BHI) agar with cefoxitin (FOX, 8 μg/mL) to ensure uniform resistance. Isolates were then serially subcultured on BHI agar slants at room temperature every 4 to 5 days and surveyed for resistance stability once a month for a total of 3 months. For each monthly survey, cells were grown overnight in BHI broth (37°C) and serially diluted and plated on BHI agar. After overnight incubation (37°C), the resulting colonies (ca. 300 to 700 per isolate per assay) were replica plated to BHI-FOX plates and again incubated overnight at 37°C. Colonies unable to grow on BHI-FOX (i.e., presumptive methicillin-susceptible) were plated on both BHI and BHI-FOX agar to confirm viability and susceptibility, respectively. Changes in SCCmec and adjacent chromosomal regions in confirmed methicillin-susceptible isolates were characterized at the 3-month point as noted below.

Growth analysis.

MRSA and susceptible derivatives were inoculated into 100 mL of BHI broth in Nephelo culture flasks with a sidearm (Bellco Glass, Vineland, NJ). Bacterial growth was monitored by optical density at 540 nm (OD₅₄₀) absorbance measured every 30 min during 5 h of incubation with shaking at 37°C.

Molecular characterization.

(i) PCR and whole-genome sequencing (WGS). Site-specific excision of the entire SCCmec element in methicillin-susceptible variants of SCCmec I, II, and III was detected by the presence of a PCR product by utilizing orfX-specific forward primers and strain-specific reverse primers targeting chromosomal sequences 3′ to the SCCmec element. For strains containing SCCmec IV and its derivatives, which may also contain the arginine catabolic metabolic element (ACME) in addition to SCCmec, an additional reverse primer in ACME was included to determine the extent of deleted sequences (Table 3). Amplification conditions were as follows: initial denaturation step of 94°C for 2 min, 30 cycles of denaturing at 94°C for 1 min, annealing at 53°C for 1 min, extension at 72°C for 1 min, and a final extension at 72°C for 5 min.

TABLE 3.

PCR primers for empty-cassette detection

Oligonucleotide	5′–3″ sequence	SCCmec type	Reference or source
orfx/SCCmec junction
Xsau 325	GGATCAAACGGCCTGCACA	All	31
SCCmec and chromosome or ACME junctiona
A-rev	GAAACTTCATTGGTATATTTAC	I, II, IV	This study
ACME-R	CCTCCTTCACTTAGCACTG	IV	This study
mecIII-R	ACGGTTAGCTTTGGTGGCTT	III	This study

^aSCCmec IV primers were designed to detect both SCCmec IV and SCCmec IV-ACME cassette loss.

For WGS, genomic DNA was extracted (DNeasy kit; Qiagen, Germantown, MD), Nextera XT libraries were prepared, and the libraries were sequenced on a MiSeq instrument according to the manufacturer’s instructions (Illumina, San Diego, CA). Sequences were mapped to the following reference strains: for SCCmec I, S. aureus strain COL (GenBank accession number CP000046.1), for SCCmec II, strain Mu50 (NC_002758.2), and for SCCmec IV strain FPR3757 (CP000255) using BioNumerics v.8 (Applied Maths, Belgium) using default settings. No SCCmec III-containing isolates were sequenced, because they all were precise excisions of the SCCmec element.