Classifying spa Types in Complexes Improves Interpretation of Typing Results for Methicillin-Resistant Staphylococcus aureus

Abstract

A total of 382 isolates of methicillin-resistant Staphylococcus aureus originating from three Austrian regions and one adjacent Italian region (Vienna, Lower Austria, North Tyrol, and South Tyrol) were typed by DNA sequence analysis of the variable repeat region of the protein A gene (spa typing). The strain collection consisted of arbitrarily chosen isolates originating from clinical specimens taken in the years 2003 to 2005 at 17 hospitals. The most common spa types found were t001 (28.8% of all isolates), t190 (27.0%), t008 (14.1%), and t041 (11.3%). The 42 remaining spa types accounted for ≤2.4% each. The dominating spa types varied between the different regions. As short sequence DNA repeat units are unstable entities, the 46 spa types were classified into seven spa complexes with respect to short sequence repeat unit composition and organization. Such classification into complexes can provide additional information for the hospital epidemiologist, empowering one to differentiate the introduction of a new strain from mere variation of endemic spa types.

Staphylococcus aureus is one of the most significant health care-associated pathogens and is responsible for a wide range of hospital infections (7, 27, 44). Since the first identification of methicillin-resistant Staphylococcus aureus (MRSA) in the early 1960s (18), the incidence of S. aureus bacteremia has been rising and has multiplied in some European countries in the past 25 years (11, 25). This increase coincides with an increased rate of community-acquired MRSA infections (5, 17, 26, 36). The emergence of community-associated MRSA strains has further increased public health concerns (26). Therefore, S. aureus typing has become an important tool in the study of strain origin, surveillance of health care-associated infections, and epidemiological outbreak investigation. A large number of molecular methods have been developed for typing of MRSA strains. The most widely used molecular typing method for the study of MRSA epidemiology is pulsed-field gel electrophoresis (PFGE) (1, 29). However, the use of PFGE remains limited by problematic interlaboratory comparison, problems of interpretation, low throughput, and high costs (6, 42). PFGE is a suitable method for the determination of clonal relationships but not for long-term epidemiological investigations (2, 3). Typing techniques based on DNA sequencing have an obvious advantage in speed, unambiguous data interpretation, simplicity of database creation, and standardization among diverse laboratories (24). Multilocus sequence typing (MLST) is a highly discriminatory method for strain typing and characterizes bacterial isolates on the basis of the sequence fragments of seven housekeeping genes. An isolate is defined by the allelic profile, or sequence type (9). Similar sequence types are grouped into clonal complexes. The low mutation rate of the seven housekeeping genes makes MLST more suitable for long-term and global epidemiological studies (10, 24). However, MLST of S. aureus is an expensive and laborious method that requires the sequencing of approximately 3,500 nucleotides and does not have the resolving power of PFGE. Recently, DNA sequencing of short sequence repeats (SSR) of the polymorphic X region of the protein A gene (spa) was proposed as an accurate method for typing S. aureus (13, 14, 34, 37). Although spa typing does not have the fine resolution of PFGE, it has advantages in terms of speed, interpretation, and interlaboratory comparison (34, 44). The polymorphic X region consists of a variable number of 21-bp to 27-bp repeats and is located upstream of the region encoding the cell wall attachment sequence (38). The determination of spa types was simplified by developing appropriate software (16, 34), and to date 111 diverse repeat units and more than 1,400 spa types have been described (http://spaserver.ridom.de). Although the function of the octapeptide of the polymorphic X region is not known, SSR variations are clearly related to bacterial pathogenesis and virulence (22, 32, 41, 43).

Despite the variability of the DNA sequence of the variable X region, the amino acid sequences of the repeats remain relatively constant. The peptides of diverse spa types show similar amino acid hydropathicities and similar secondary structures (20). Thus, mutations mainly affect the wobble base and all repeats are in frame, so that deletions or additions of repeat units will not alter the reading frame.

In most cases, rearrangements of repetitive sequences are caused by replication slippage or the repair of double-strand breaks during DNA replication (4, 12, 43). Conflicting results on the variability of the X region of the spa gene have been reported (13, 19, 34, 37, 39). Without selective pressure (i.e., under laboratory conditions), the frequency of genetic variation of SSR is low (13, 31, 37, 39). However, under selective pressure (i.e., in vivo), up to 10% of the analyzed MRSA strains showed mutations within the polymorphic X region of the spa gene (19, 39). It is probably a limitation to use the X region as a sound target for epidemiological investigations (40). The hypervariability of the SSR region and the huge number of spa types identified demand further classification into spa complexes with respect to SSR unit similarity (10, 30). The spa complexes so determined showed a good correlation to the recently determined clonal complexes (10, 21, 30), amplified fragment length polymorphism complexes (25), and DNA microarray data (20). Thus, analysis of spa complexes may increase the usefulness of spa typing for short-term as well as long-term epidemiological investigations.

For the first time, MRSA isolates collected from three Austrian regions and one adjacent Italian region were analyzed by spa typing using this bifunctional approach to view the MRSA population. The diversity of the MRSA isolates was determined by the number of distinct spa types. The clonal structure of the MRSA population was determined by the number of spa complexes, which were defined as a group of spa types with similar SSR profiles. These spa complexes were correlated to clonal complexes or to an MLST if the specific sequence type had not been assigned to a clonal complex.

MATERIALS AND METHODS

Bacterial strains.

The 382 MRSA isolates spa typed originated from the region of South Tyrol, Italy (isolates provided by Sanitary Service Bozen, n = 56), and the Austrian regions of North Tyrol (University Hospital Innsbruck, n = 76), Lower Austria (Hospital Gmuend, Danube-Clinic Tulln, Hospital St. Poelten, Hospital Korneuburg/Stockerau, Danube-Clinic Gugging, and Waldviertel-Clinic Horn, n = 80), and Vienna (Hospital Rudolfstiftung, Hospital Meidling, Lorenz-Boehler Hospital, Hospital Speising, Hospital Lainz, Hospital Kaiser-Franz-Josef, Vienna Medical Center East, Hospital Wilheminen, and Hospital St. Elisabeth, n = 170).

spa sequencing.

A loopful of material from an overnight culture (grown on blood agar plates) was suspended in 0.5 ml sterile H₂O. The suspension was heated to 95°C for 10 min and was immediately frozen at −20°C for at least 30 min. After thawing, 1.0 μl of the suspension was used for PCR.

The variable X region of the spa gene was amplified in a 50-μl reaction volume using RedTaq Ready mix (Sigma, St. Louis, Mo.). The oligonucleotides used for amplification correspond to the 5′ end (1113F, 5′-TGTAAAACGACGGCCAGTTAAAGACGATCCTTCGGTGAGC) and the 3′ end (1514R, 5′-CAGGAAACAGCTATGACCCAGCAGTAGTGCCGTTTGCTT) (16, 28) containing an M13 primer sequence (31). PCR conditions were 95°C for 5 min; 35 cycles each of 95°C for 15 s, 58°C for 30 s, and 72°C for 45 s; and a final step at 72°C for 10 min. Prior to sequencing, 10 μl of the amplified products was analyzed on 1.5% agarose gels. The remaining 40 μl was purified using a GenElute PCR clean-up kit (Sigma) according to the manufacturer's instructions.

Sequence analysis was performed by cycle sequencing (SequiTerm Excel II cycle sequencing kit; Epicenter, Madison, Wis.) with fluorescent-labeled primers M13 univ. (5′-TGTAAAACGACGGCCAGT) and M13 rev. (5′-CAGGAAACAGCTATGACC) (MWG-Biotech, Ebersberg, Germany), using a LI-COR 4200S automated DNA sequencer (LI-COR Bioscience, Lincoln, Nebr.) according to the manufacturer's instructions.

Sequence data analysis.

The standard chromatogram files of the forward and reverse sequences obtained from each sample were assembled and edited and spa types were determined using Ridom StaphType software (16).

Classification of spa types.

spa types with similar repeat profiles were grouped into a spa complex. Since sequence-based alignment using algorithms is not useful for analysis of repeat regions, visual analysis of repeat organization is a more reliable and easier method for comparing related spa types. A spa complex was characterized by spa types that shared identical SSR units arranged in an identical order. Within a complex, spa types were differentiated due to deletions or duplications of SSR units, which is the major cause of genetic rearrangement within an SSR region (4, 41). Exceptions were due to less frequent mutation events such as deletion, duplication, and exchange of triplets or single nucleotide polymorphisms, resulting in a change of single SSR units within a repeat region. The general structure of repeat units of a repeat region may not be changed by single nucleotide polymorphisms.

RESULTS

Typing of the 382 clinical MRSA isolates from the four regions, South Tyrol, North Tyrol, Lower Austria, and Vienna, yielded 46 spa types (Tables 1 to 7). The most common spa types identified were t001 (28.8% of all isolates), t190 (27.0%), t008 (14.1%), and t041 (11.3%). The 42 remaining spa types found accounted for ≤2.4% each. Considering SSR unit similarity of the SSR region, the 46 spa types identified were classified into seven spa complexes (Tables 1 to 7). spa complexes were defined by the following repeat profiles: complex I, 26-(30/23)-(17-34-17-20/2)_n-(17-12)_n-17-16; complex II, 11-(19-21-12-21)-17-(34-24-34)-22-25; complex III, 15/1/8-(12)-(16-2)_n-(25)-(17)-(24/25)-24_n; complex IV, 8/9-(16)-(2-16-34-13-17-34)-16-34; complex V, 26-(23)_n-13-23-(31-29/5-17)_n-25-17-25-16-28; complex VI, which might consist of two subcomplexes, 4-20/21-(12-17)_n-20-17-12-12/17-17 and 7-23-12-21-21-(17)-20-17-(12)_n-(17); and complex VII, which might also consist of subcomplexes, 7-23-(21)-12/17/16-(34)_n-(33)-34/12/13-(12-23)-(2-12-23) (Tables 1 to 7).

TABLE 1.

spa complex I

spa typea	SSR profile																No. of isolatesb
																	ST	NT	LA	VIE	Total
t661	26	30	17	34	17	20	17	34	17	20	17	12	17	12	17	16				1	1
t838	26	30	17	34	17	20	17	34	17	20	17	12	17		17	16			1	1	2
t896	26	30	17	34	17	20	17	34	17		12	12		12	17	16		4			4
t041	26	30	17	34	17	20	17	34	17	20	17	12			17	16	11	5	13	14	43
t001	26	30	17	34	17	20	17					12			17	16	5	56	38	11	110
t002	26	23	17	34	17	20	17					12			17	16		1	2		3
t1062	26	23	17	34	17	2	17					12			17	16				2	2
t1003	26				17	20	17	34	17	20	17	12			17	16		1			1
t003	26				17	20	17					12	17		17	16				2	2
t811	26	30	17	34	17	20	17								17	16			1	1	2
t551	26	30	17	34	17	20									17	16			1		1
t288	26	30													17	16			2		2
Total																	16	67	58	32	173

^at001, t002, t003, and t041 were assigned to CC5.

^bThe frequency of each spa type within a region is shown. ST, South Tyrol; NT, North Tyrol; LA, Lower Austria; VIE, Vienna.

TABLE 7.

spa complex VII

spa typea	SSR profile													No. of isolatesb
														ST	NT	LA	VIE	Total
t084	7	23		12	34	34		12	12	23	2	12	23	1				1
t091	7	23	21	17		34			12	23	2	12	23	1				1
t593	7	23		12	34						2	12	23				1	1
t127	7	23	21	16		34	33	13								1		1
t044	7	23		12	34	34	33	34						2		1	6	9
t416	7	23		12	34												1	1
Total														4		2	8	14

^aspa complex VII probably consists of at least two subcomplexes, which is reflected by the assignment of the spa types to diverse clonal complexes and sequence types (CC1, CC15, ST7, and ST80).

^bThe frequency of each spa type within a region is shown. ST, South Tyrol; NT, North Tyrol; LA, Lower Austria; VIE, Vienna.

One hundred seventy-three isolates, or 45.4%, belonged to spa complex I (Table 1). The spa types t001, t002, t003, and t041 of complex I were assigned to clonal complex 5 (CC5). The South German MRSA type t001 and the Rhine Hesse MRSA type t002 represent the prototypes within this complex. In North Tyrol and Lower Austria, spa types of complex I prevailed, with frequencies of 88.2% and 72.5%, respectively (Table 1). In both regions, t001 was the most prevalent spa type, with a frequency of 73.7% in North Tyrol and with a frequency of 47.5% in Lower Austria (Table 1).

One hundred sixty-two isolates, or 42.5%, belonged to spa complex II (Table 2). The spa types t008, t190, t051, and t024 were assigned to CC8. spa types of complex II were most frequent for Vienna and South Tyrol, with frequencies of 60.5% and 62.5%, respectively (Table 2). With a frequency of 52.3%, spa type t190 was the predominant spa type in Vienna. spa type t190 was not found in South Tyrol. In this region, spa type t008 was the dominant spa type, with a frequency of 58.9%. spa type t008 was the exclusive representative of complex II in North Tyrol (Table 2).

TABLE 2.

spa complex II

spa typea	SSR profile											No. of isolatesb
												ST	NT	LA	VIE	Total
t051	11	19	21	12	21	17	34	24	34	22	25	1				1
t008	11	19		12	21	17	34	24	34	22	25	33	6	2	13	54
t024	11			12	21	17	34	24	34	22	25			1		1
t746	11	19		12	21	17				22	25	1				1
t190	11					17	34	24	34	22	25			13	90	103
t662	11					17			34	22	25				1	1
t810	11					17				22	25			1		1
Total												35	6	17	104	162

^at051, a North German MRSA, Archaic/Iberian type (ST247:MRSA:SCCmec I), may represent the prototype of spa complex II. t051, t008, t024, and t190 were assigned to CC8.

^bThe frequency of each spa type within a region is shown. ST, South Tyrol; NT, North Tyrol; LA, Lower Austria; VIE, Vienna.

Fifteen isolates, or 3.9%, belonged to spa complex III (Table 3). The spa types t012, t018, and t019 of complex III were assigned to CC30/39, and spa types t030 and t037 of this complex were assigned to CC239. The spa type t018 and the Vienna MRSA type t037 might represent the prototypes of this complex. Thirteen of 15 isolates of complex III were found in the Vienna region (Vienna and bordering districts of Lower Austria).

TABLE 3.

spa complex III

spa typea	SSR profile											No. of isolatesb
												ST	NT	LA	VIE	Total
t018	15	12	16	2	16	2	25	17	24	24	24			1		1
t012	15	12	16	2	16	2	25	17		24	24				1	1
t582	1	12	16	2	16	2	25			24	24				1	1
t019	8		16	2	16	2	25	17			24				1	1
t037	15	12	16	2			25	17		24	24		1		1	2
t030	15	12	16	2						24	24		1		4	5
t425	15	12	16	2			25	17	25						1	1
t930	15		16	2						24	24				2	2
t633	8										24				1	1
Total													2	1	12	15

^at018 may represent the prototype of spa complex III. t012, t018, and t019 were assigned to CC30, whereas t030 and t037 were assigned to CC239.

^bThe frequency of each spa type within a region is shown. ST, South Tyrol; NT, North Tyrol; LA, Lower Austria; VIE, Vienna.

The spa complex IV consisted of four different spa types, derived from five isolates (Table 4). The spa types t015, t065, and t390 were assigned to CC45. The Berlin MRSA type t004 (ST45:MRSA:SCCmec IV) (where ST45 is sequence type 45 and SCCmec IV is staphylococcal chromosome cassette mec type IV) might represent the prototype of this complex. All isolates were from the adjacent geographical areas Vienna and Lower Austria.

TABLE 4.

spa complex IV

spa typea	SSR profile										No. of isolatesb
											ST	NT	LA	VIE	Total
t015	8	16	2	16	34	13	17	34	16	34				2	2
t065	9		2	16	34	13	17	34	16	34				1	1
t550	8						17	34	16	34				1	1
t390	8	16							16	34			1		1
Total													1	4	5

^at015, t065, and t390 were assigned to CC45. t004 (ST45:MRSA:SCCmec IV) may represent the prototype of spa complex IV.

^bThe frequency of each spa type within a region is shown. ST, South Tyrol; NT, North Tyrol; LA, Lower Austria; VIE, Vienna.

The spa complex V consisted of nine isolates (Table 5). With the exception of spa type t005, which was a single isolate from South Tyrol, all other isolates were from Vienna. The spa types t005, t032, and t223 were assigned to CC22.

TABLE 5.

spa complex V

spa typea	SSR profile																No. of isolatesb
																	ST	NT	LA	VIE	Total
t032	26	23	23	13	23	31	29	17	31	29	17	25	17	25	16	28				5	5
t379	26	23	23	13	23	31	29	17				25	17	25	16	28				2	2
t005	26	23		13	23	31	5	17				25	17	25	16	28	1				1
t223	26	23		13	23		5	17				25	17	25	16	28				1	1
Total																	1			8	9

^aThe Barnim MRSA type t032 is the prototype of spa complex V. t032, t005, and t223 were assigned to CC22.

^bThe frequency of each spa type within a region is shown. ST, South Tyrol; NT, North Tyrol; LA, Lower Austria; VIE, Vienna.

The spa complex VI consisted of four isolates that belong to four different spa types (Table 6). The spa types of this complex might consist of two different subcomplexes due to the first SSR units of the SSR region and show, at least to some extent, similarity to the spa types of complex I and complex VII. Similar spa types of complex VI were assigned to CC25.

TABLE 6.

spa complex VI

spa typea	SSR profile													No. of isolatesb
														ST	NT	LA	VIE	Total
t897	4			20	12	17	12	17	20	17	12	17	17		1			1
t349	4			21	12	17			20	17	12	12	17			1		1
t585	7	23	12	21	12				20	17	12	12					1	1
t148	7	23	12	21	12	17			20	17	12	12	17				1	1
Total															1	1	2	4

^aMLST profiles of these spa types were not available. Similar spa types were assigned to CC25.

^bThe frequency of each spa type within a region is shown. ST, South Tyrol; NT, North Tyrol; LA, Lower Austria; VIE, Vienna.

Complex VII was a heterogeneous group (Table 7), and might consist of two or even more subcomplexes. However, because of common repeat units and due to the limited number of similar spa types these spa types were grouped into a common complex. Nine of 14 isolates (64%) of the six spa types within this complex belonged to spa type t044. The spa types of this complex were assigned to CC1 (t127), CC15 (t084), ST7 (t091), and ST80 (t044 and t416). spa types within this complex were found in South Tyrol and the Vienna region.

DISCUSSION

The Antibiotic Resistance, Prevention and Control (ARPAC) project, a European Commission DG Research-funded concerted action, identified S. aureus resistant to methicillin/oxacillin as an alert organism of concern in Europe (23). Typing alert organisms by use of molecular techniques is considered a high-priority recommendation for national health authorities.

To determine the diversity of the MRSA population in Austria, isolates from diverse regions and hospitals were typed. Because of the high level of polymorphism within the X region of the protein A gene and the fact that spa typing is associated with speed, low costs, and comparability of data, spa typing was used in this study for typing of MRSA isolates. Forty-six different spa types were found, of which the majority, 82%, belong to four spa types only. However, as SSR are generally described as unstable entities that undergo frequent DNA sequence variation (19, 33, 39, 40, 41), these 46 spa types were grouped into spa complexes with respect to SSR unit composition and organization (34). This system of classification of related types had already been used to group MLSTs into clonal complexes (8). The grouping of related types (spa, MLST, or PFGE) is a valuable tool for strain analysis and supports a hospital epidemiologist's work by simplifying the identification of epidemiologically related strains. However, molecular subtyping on its own is not able to elucidate whether distinct subtypes have evolved from one another or whether they represent single introductions from outside. Without additional epidemiological information, it is, for instance, impossible to say whether spa type t008 in Vienna is the precursor of the most frequent spa type, t190, or whether these spa types occurred independently. Moreover, without epidemiological data it is even impossible to postulate a correlation between rare spa types, for example, t037 in Vienna and in North Tyrol. Nevertheless, these rare spa types are those of special interest for epidemiological investigations, because the chance of elucidating chains of infection is increased compared to doing so with widely distributed types. Thus, dominant spa types may be too frequent to provide useful hints in investigating the chains of transmission. Additional markers are necessary to differentiate these frequent strains.

The allocation of similar spa types into spa complexes facilitates the differentiation of new, emerging spa types into descendants of the local spa complexes and new spa types and unrelated types. The grouping of similar spa types considers that various spa types are encountered among strains with similar overall genotypes, which is an indication that the speed of SSR evolution does not reflect the speed of overall genome evolution (25, 39). The spa complexes generated with respect to these features showed a good correlation to recently determined clonal complexes, coagulase groups, and amplified fragment length polymorphism clusters (8, 10, 25, 35). Moreover, recently reported microarray data correlate with SSR composition and organization (20). Thus, this bifunctional application combines a rapidly evolving marker useful for outbreak investigations with the more stable core structure of a complex useful for long-term epidemiology by sequencing a single DNA fragment.

However, one discrepancy in comparison to the MLST complexes could be observed. Although spa types of spa complex III have a uniform SSR structure, spa types of this complex were assigned to the different clonal complexes CC30/39 and CC239. Strains of the clonal complex CC239 are related to strains of the clonal complex CC8 (8). CC8 is equivalent to spa complex II. Thus, we might have a situation where similar spa types have evolved from different precursors. The opposite, the evolution of different MLSTs from a common spa type, cannot be ruled out. Further work has to be done to elucidate this question.

Due to the lack of a sufficient number of spa types and isolates, the spa complex VI may contain spa types that belong to two related but different spa complexes. Although the MLST of our isolates of spa complex VI is not known, similar spa types were assigned to CC25. The SSR structure of spa complex VI shows a distant relation to spa complex I.

As with complex VI, there is a high probability that the spa types of spa complex VII belong to different groups. This is supported by the fact that spa types of spa complex VII have been assigned to different MLSTs (ST1, ST15, ST7, and ST80). Nevertheless, spa types of spa complex VII have similar SSR structures. Moreover, it was previously shown that isolates of ST1 and ST80 both carry the Panton-Valentine leucocidin gene and SCCmec IV (15). Thus, there might be a phylogenetic correlation of these spa types within complex VII that is reflected by a related SSR structure, although the MLST is different.

With respect to the classification of spa types into spa complexes, each region has its distinct spa population. Interestingly, the neighboring regions of South Tyrol and North Tyrol and those of Vienna and Lower Austria have quite different spa populations. There is a higher similarity between the spa populations of the geographically distant regions North Tyrol and Lower Austria on the one hand and Vienna and South Tyrol on the other hand. In our opinion, this finding underlines the importance of local efforts to cope with local MRSA problems. Whereas South Tyrol, North Tyrol, and Lower Austria have quite uniform spa populations, consisting of mainly two spa complexes each (with a few spa types belonging to other complexes), due to a homogenous population, the occurrence of all spa complexes in Vienna probably reflects the multinational population structure of this large city.

It is of interest that the prevalent spa type within spa complex II in South Tyrol is t008, whereas it is t190 in Vienna. This prevalence of different spa types, both of one complex, may be the result of adaptation to different ecological conditions (39).

In conclusion, we consider spa typing a highly effective and rapid typing tool for S. aureus that has significant advantages over other typing techniques, with the potential to benefit largely the surveillance of antimicrobial resistance and infection control. The system presented here of reporting complexes in addition to the mere spa type should foster the use of spa typing as a tool for long- and short-term epidemiological investigations. Although some spa types cannot be grouped unambiguously into a distinct spa complex, the grouping of spa types into complexes, in addition to the simple reporting of spa types, seems to be a promising tool to simplify the use of spa typing for epidemiological investigations. We recommend a reclassification of the >1,400 already-determined spa types that would consider the relatedness of certain spa types, for example, I-001 (I for the spa complex and 001 for the spa type). The grouping of the large number of spa types of the Ridom database and the determination of MLSTs of so far-unclassified isolates will show a more complete and accurate situation as to the numbers and types of spa complexes. Even difficult-to-group types (i.e., spa types consisting of one, two, or three SSR units) should be amenable to grouping into a complex with the additional information of the MLST.