THE JOWNAL OF BIOLOGICAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 269, No. 18, Issue of May 6 , pp. 13458-13464, 1994 Printed in U S A .

Molecular Characterization of Protein Sir, a Streptococcal Cell Surface Protein That Binds Both Immunoglobulin A and Immunoglobulin G*

(Received for publication, December 28, 1993, and in revised form, February 18, 1994)

Lars Stenbere, Paul W. OTooleS§, Jiri Mesteckm and Gunnar LindahlSII From the Wepartment of Medical Microbiology, Lund University, S-223 62 Lund, Sweden and the Wepartment of Microbiology, University of Alabama at Birmingham, Birmingham, Alabama 35294

Cell surface proteins that bind to the Fc part of immu- noglobulin (Ig) A andor IgG are expressed by many strains of the group A Streptococcus, an important hu- man pathogen. Two extensively characterized proteins in this group of molecules are protein Arp that prefer- entially binds IgA and protein H that binds IgG. In ad- dition, recent work has shown that many group A strep- tococcal strains express a novel type of Fc-binding protein, designated protein Sir, that binds both IgA and IgG. Protein Sir22, the molecule expressed by a strain of serotype M22, has now been purified and characterized after expression of the cloned gene in Escherichia coli. Dot-blot analysis with a large number of purified mono- clonal Igs showed that protein Sir22 reacted with 19 out of 20 IgA proteins and with 19 out of 24 IgG proteins. The affinity constants for the reactions between protein Sir22 and Ig were determined to be 7.0 x 108 M-’ for serum IgA, 2.4 x 10’ M - ~ for secretory IgA, and 7.8 x 10’ M - ~ for IgG. Inhibition experiments showed that the bindings of IgA and IgG to protein Sir22 were mutually exclusive, indicating shared or contiguous binding sites. Analysis of the sequence of the sir22 gene indicated a gene prod- uct with 365 amino acid residues, including a 41-residue signal peptide. The processed form of the protein, 324 residues, has a calculated M, of 37,186. Deletion analysis of the sir22 gene showed that a 156-residue NH,-terminal fragment of protein Sir22 retained the ability to bind both IgA and IgG. The overall organization of protein Sir22 is similar to that of the IgA-binding protein Arp and the IgG-binding protein H. All three of these pro- teins are members of the M protein family and have a central repeat region of the C type.

Several species of Gram-positive bacteria express cell wall proteins that bind to the Fc part of IgG or IgA,’ the two major

* This work was supported by the Swedish Medical Research Council (Project 9490). the Medical Faculty of Lund University, The Royal Physiographic Society in Lund, The Swedish Society for Medical Re-

Bergvall, Crafoord, Kock, and Osterlund, and United States Public search, King Gustaf V’s 80-year Foundation, the Foundations of

Health Service (Bethesda, MD) Grant DK-28537. The costs of publica- tion of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

0 Present address: Dept. of Biochemistry and Microbiology, Univer- sity of Victoria, Victoria B. C., V8W 3P6 Canada.

11 To whom correspondence should be addressed: Dept. of Medical Microbiology, Lund University, Solvegatan 23, S-223 62 Lund, Sweden.

The abbreviations used are: Ig, immunoglobulin; IgA, IgD, IgE, I&, and IgM, immunoglobulin A, D, E, G, and M, respectively; PAGE, poly- acrylamide gel electrophoresis; BSA, bovine serum albumin; SPRIA, solid phase radioimmunoassay; PVDF, polyvinylidene difluoride; ORF, open reading frame.

Tel.: 46-46-173244; Fax: 46-46-189117.

Ig classes in man. Well known examples of such proteins in- clude protein A from Staphylococcus aureus and protein G from the group G Streptococcus, both of which preferentially bind IgG (1-3). Two different IgA-binding proteins have also been studied in detail: protein Arp from the group A Streptococcus and protein Bac from the group B Streptococcus (4-8). These various Ig-binding proteins are valuable as immunochemical tools and as model systems for studies of the interactions be- tween Ig and specific receptors. In addition, the study of bac- terial Ig-binding proteins can contribute to our understanding of the host-parasite relationship during infection, since it is likely that these proteins help the bacteria to evade the defense mechanisms of the infected host (3).

Studies of the Ig-binding proteins expressed by different strains of the group A Streptococcus, an important human pathogen, have demonstrated that these proteins are remark- ably heterogeneous with regard to structure and binding prop- erties (9-15). However, all Ig-binding proteins of the group A Streptococcus are members of the same family of proteins, the M protein family, and have similar overall organization (12). nYo of these Ig-binding proteins in the M protein family have been purified and extensively characterized after expression in Escherichia coli: protein Arp that preferentially binds IgA and protein H that binds IgG (4, 5, 11, 16-18).

Further characterization of the Ig-binding proteins ex- pressed by the group A Streptococcus has recently shown that some strains express a protein, designated protein Sir, that binds human IgA and IgG of all subclasses and therefore has broader reactivity than any bacterial Fc-binding protein de- scribed previously (14). Although the IgA-binding protein Arp also binds some monoclonal IgG proteins, most IgG molecules are not bound to protein Arp and the affinity constant for the binding of protein Arp to polyclonal IgG was too low to be measurable (4). Taken together, these data indicated that pro- tein Sir represents a novel type of Ig-binding protein, and fur- ther molecular characterization of the molecule was therefore warranted.

For the preliminary study of protein Sir (14), we used a streptococcal strain of serotype M22 and the protein, desig- nated Sir22, was purified after extraction from the streptococ- cal cell wall. However, it is notoriously dificult to isolate cell wall proteins directly from the group A Streptococcus, and only minute amounts of protein Sir22 could be obtained by this method. A new procedure has therefore been established for purification of protein Sir22 after expression of the cloned gene in E. coli, and the protein has been characterized immuno- chemically. Furthermore, the amino acid sequence of protein Sir22 has been deduced from the sequence of the gene and a comparison with the known sequences of the IgA-binding pro- tein Arp and the IgG-binding protein H is presented. Finally, the binding activity of protein Sir22 for Ig was localized to the NH,-terminal half of the molecule.

13458

&A- and IgG-binding Protein Sir 13459

MATERIALS AND METHODS Bacterial Strains, Bacteriophage, and Plasmids-The sir22 gene, the

structural gene for protein Sir22, was cloned from Streptococcus pyo- genes strain AL168 (serotype M22), using the vector A EMBL3 as de- scribed previously (14). Plasmid pSIR2202 is a pK19 (19) derivative with a 2.2-kilobase Hind111 insert from one of the A clones. pSIR2202 carries the entire gene sir22. Plasmid pARP401 carrying the gene arp4 has been described (9). Plasmid vectors pUC18/19 (20) and pK18/19 were used in subcloning experiments. The E. coli strains TB1 and JM83 (20) were used as hosts for chimeric plasmids.

DNA Preparations, Cloning Techniques, and Sequence Determina- tion-Bacteriophage A DNA was prepared from phage particles purified on CsCl step gradients, essentially as described by Sambrook et al. (21). Plasmid DNA was purified by the Magic DNA purification systems (Promega, Madison, WI) following the recommendations of the manu- facturer. Standard procedures were used for restriction enzyme diges- tions and cloning experiments (21). A colony blot technique (21) and SDS-PAGE followed by Western blots were used to follow expression of cloned genes encoding Ig-binding proteins. Double-stranded chain ter- mination DNA sequencing was performed with Sequenase (U. S. Bio- chemical Corp.) and T7 DNA polymerase (Pharmacia, Uppsala, Swe- den), and thermo-cycle sequencing was performed using the Femto- mole kit from Promega. Template plasmid DNA was obtained by subcloning of restriction enzyme fragments and by the method of Misra (22), using exonuclease Bal-31 to create ordered sets of deletions in cloned fragments. Custom primers were employed to complete the sec- ond strand in certain positions. The nucleotide sequence was deter- mined for both strands of the plasmid DNA.

Purification of Proteins Sir22 and Arp4 Expressed in E. coli-E. coli strain JM83 harboring plasmid pSIR2202 or pARP401 was grown in 4 liters of LB broth supplied with the appropriate antibiotic (21) to A, = 0.3. The cells (about 15 g, wet weight) were subjected to osmotic shock to release periplasmic proteins (23). The osmotic shock lysate (-1200 ml), containing the periplasmic proteins in 0.5 m~ MgCI,, was subjected to anion-exchange chromatography on DEAE Bio-Gel A (Bio-Rad). Fif- teen ml of gel equilibrated in 10 m~ Tris, pH = 8.0, was added to the lysate, which was stirred gently a t 4 "C for 4 h. The gel was transferred to a glass column and adsorbed proteins were eluted with a 500-ml linear salt gradient (0-0.2 M NaCl in 10 m~ Tris, pH = 8.0) followed by 1 M NaCl(25 ml). Fractions (10 ml) were collected and analyzed for the presence of Sir22 or Arp4 by SDS-PAGE and Western blot analysis. Appropriate fractions were pooled, concentrated, and subjected to mo- lecular sieve chromatography in a column (90 x 2.5 cm) of Sephacryl S-300 HR (Pharmacia) in PBSA (0.03 M phosphate, 0.12 M NaCl, 0.02% NaN,, pH = 7.2). Fractions (5 ml) containing protein Sir22 o r b 4 were pooled and frozen. The yield was 0.2-0.3 mg of Sir22 or Arp4/g of bacteria.

Zmmunoglobulins and Other Proteins-All proteins used were of hu- man origin, unless otherwise stated. Polyclonal serum IgA and secre- tory IgA were purchased from Cappel-Organon Teknika (Turnhout, Bel- gium) and polyclonal IgG from Kabi (Uppsala, Sweden). Monoclonal IgM was from The Binding Site (Birmingham, Great Britain). The monoclonal IgG proteins were kindly provided by Dr. F. Skvaril (World Health Organizatiodnternational Union of Immunologic Societies Im- munoglobulin Subcommittee, Bern, Switzerland). Monoclonal IgA pro- teins of the IgAl and IgA2 subclasses were isolated from serum or plasma of patients with IgA multiple myeloma. Purification steps (24) included precipitation with ammonium sulfate, starch-block electro- phoresis, andlor DEAE-cellulose ion-exchange chromatography and mo- lecular sieve chromatography on Sephadex G-200 and Sepharose 6B (Pharmacia) columns. Purity of these isolated proteins was tested a t a concentration of approximately 10 mg/ml by immunoprecipitation with polyvalent reagents to human serum proteins.

The designations of the monoclonal IgA and IgG preparations are given in the legend to Fig. 2. Monoclonal IgD was the gift of Dr. Anders Grubb (Lund University), and monoclonal IgE was the gift of Dr. Inge Olsson (Lund University). Rabbit and mouse IgG, human and bovine serum albumin (BSA) were from Sigma, and human fibrinogen was from Kabi. Protein H (16) was the kind gift of Dr. Lars Bjorck (Lund University). Polyclonal Fab fragments of IgG were from Cappel-Or- ganon Teknika, and polyclonal Fc fragments of IgG were from Calbio- chem. The preparation of Fab and Fc fragments derived from an IgAl myeloma protein has been described (4).

Inhibition Experiments and Determination of Binding Constants-A solid phase radioimmunoassay (SPRIA) was used to analyze the bind- ing of IgA and IgG to protein Sir22. Microtiter plates (Falcon 3912, Becton Dickinson and Co., Oxnard, CA) were coated with protein Sir22

I . I .

. .I

21.5-

1 4 . 6

STAIN BLOT e ,

Pmk: IgA-Fc Pmk: IgG-Fc

FIG. 1. Western blot analysis of three Ig-binding proteins of the group A Streptococcus: protein Sir22, protein Arp4, and protein H. Similar amounts of the three proteins (all purified after expression in E. coli) were separated by SDS-PAGE and stained with Coomassie Brilliant Blue. Two identical gels were electroblotted to PVDF mem- branes and probed with radiolabeled IgA-Fc or IgG-Fc. Molecular mass markers, indicated on the left, are in kilodaltons.

by incubation overnight with 100 pl of a solution (1-5 pg/ml) of the protein in PBS (0.03 M phosphate, 0.12 M NaCl, pH = 7.2). The wells were blocked by washing with 0.15 M NaCl + 0.05% Tween. For inhibi- tion experiments, 100 p1 of 1251-labeled polyclonal IgAor IgG (about 1-2 ng, 20,00040,000 cpm) in PBSAT (PBS with 0.02% NaN, and 0.05% Tween 20) and various amounts (85 pg to 5 pg) of unlabeled IgA or IgG in 100 pl of PBSAT were added to the wells. After 2 h of incubation the wells were washed three times with 0.15 M NaCl + 0.05% Tween, and the remaining radioactivity counted in a y-counter. Binding constants were determined by SPRIA essentially as described by Lindahl and kers t rom (4) and plotted according to Scatchard (25). Nonspecific binding (less than 1%) was determined in wells coated with buffer (PBS) alone. Addition of 10 m~ BSA did not inhibit binding of IZ5I-Ig. All incubations were done at room temperature (20-22 " 0 .

Other MethodsSDS-PAGE, Western blotting, and dot-blot analysis were performed as described (14). Proteins were radiolabeled with '''1 using a modified lactoperoxidase method (26) or with the Bolton-Hunter reagent (Amersham International, Amersham, Great Britain). Specific activities were between 0.17 and 0.5 MBq/pg. Total protein concentra- tions were determined with the Micro BCA protein assay reagent (Pierce Chemical Co.). Automated amino acid sequence analysis of pro- tein bands transferred to a polyvinylidene difluoride (PVDF) membrane (Immobilon, Millipore, Bedford, MA) was performed directly on the membranes, using an Applied Biosystems 470A gas-liquid solid phase sequenator as described (27).

RESULTS

Purification of Protein Sir22 Expressed in E. co l iS t r a in JM83 harboring pSIR2202 was used to purify protein Sir22 expressed in E. coli. Protein Sir22 was secreted into the periplasmic space and could be purified from osmotic shock lysates, using a combination of anion-exchange chromatogra- phy and molecular sieve chromatography. In a typical experi- ment, 4 liters of bacterial culture yielded 3.5 mg of pure protein Sir22. The protein could also be purified by affinity chromatog- raphy on immobilized Ig, but this method was not used rou- tinely due to poor recoveries.

Purified protein Sir22 was analyzed by Western blotting, using Fc fragments of IgA and IgG as probes (Fig. 1). In the same experiment, protein Sir22 was compared with purified preparations of the other two group A streptococcal proteins mentioned above: protein Arp4 that preferentially binds IgA and protein H that binds IgG. In SDS-PAGE, protein Sir22 migrated as a doublet band, corresponding to 37.5- and 38.5- kDa peptides. These two polypeptides had the same NH,-ter- minal sequence: Glu-Ser-Ser-Asn-Asn-Ala. The size heteroge- neity of protein Sir22 was not unexpected, since doublet or

1.5 0.75 0.375 0.188

e

Arp4

FIG. 2. Dot-blot analysis of the binding specificities of protein Sir22 and protein Arp4. The indicated amounts of Ig (or other serum proteins) diluted in PBS were applied in 100-pl aliquots to a PVDF membrane. After blocking, the membranes were probed with the radiolabeled streptococcal protein indicated to the right, washed, and subjected to autoradiography. The monoclonal IgG and IgApreparations are, from left to right: IgGl(A) LOB, I ~ G U K ) Schw, 1gG1(~) Ho, I ~ G U K ) Hoch, I ~ G U K ) Bau, 1gG1(~) Hu; IgG2(A) Pa, IgG2(A) Fi, 1gG2(~) Sa, I~GGB(K) G236, I @ ~ ( K ) Bi, 1gG2(~) Os; IgG3 (A) Fi, IgG3 (A) Ruf, IgG3 (A) Hii, IgG3 ( K ) b y , IgG3 (K) Stu, IgG3 ( K ) Gee; IgG4 (A) Ste, IgG4 (A) We, IgG4 ( K ) Spa, IgG4 ( K )

Hru, IgG4 ( K ) Bru, IgG4 ( K ) Rei; IgAl (A) JM, IgAl (A) JM, IgAl (A) SapII, IgAl ( K ) Ber, IgAl ( K ) JM, IgAl ( K ) JM; IgA2 (A) Fel, IgA2 (A) Sch, IgA2 ( K ) Gir, IgA2 ( K ) Kes, IgA2 ( K ) Mau, IgA2 (K) Bel. The designations of the different monoclonal proteins refer to the patient of origin, except for the abbreviation JM, which indicates that the protein was from the collection of Jiri Mestecky. All proteins, except rabbit IgG and mouse IgG, were of human origin.

0.1251A Serum IgA: hz7.0 x 108 W1

OSsraory IgA: K&4 x l@ M1

0 1 t 0 0.05 0.10 0.15 020

0"251B

X

0 4 < 0 0.1 0 2 0.3 0.4

BOUND $A (nM) BOUND $0 (nM)

FIG. 3. Scatchard plots for the reactions between protein Sir22 and human polyclonal serum IgA, eecretory IgA and IgG. The affinity constants for the reactions between protein Sir22, immobilized in the wells of microtiter plates, and IgA (A) or IgG ( B ) were determined in a solid-phase radioimmunoassay. 0.1 ml of radiolabeled Ig was mixed with various amounts (85 pg to 5 pg) of unlabeled Ig in 0.1 ml of phosphate-buffered saline, pH = 7.2, containing 0.02% NaN, and 0.05% Tween 20. After 2 h a t 25 "C the wells were washed three times and the remaining radioactivity measured in a y-counter. Amounts of bound and free Ig were calculated and plotted according to Scatchard (25). Each point represents the mean of the values from triplicate experiments. The best fit line through the points was estimated. The affinity constants of the reactions then eaual the absolute values of the sloDes of the curves. and the intersection of the line with the x axis gives the maximal amount of bound Ig.

triplet bands are commonly seen when streptococcal cell sur- face proteins are expressed in E. coli (9,28). The blotting anal- ysis demonstrated that each of the two polypeptides in protein Sir22 reacted with both IgA and IgG, but in the blot shown in Fig. 1 these two bands merge into one broader band. In agree- ment with previous results, protein Arp bound IgA-Fc frag- ments and also reacted very weakly with IgG-Fc fragments (too weakly to be seen in Fig. l), whereas protein H bound only IgG-Fc fragments (4,161. For protein H, weak blotting bands of high molecular mass were also seen, possibly due to polymer- ization of the molecule (18). Protein H migrated more slowly than expected in the SDS-PAGE, as described also for other streptococcal cell surface proteins (29).

A dot-blot analysis with Fc and Fab fragments of IgA and IgG showed that purified protein Sir22 bound only Fc fragments of these Igs (data not shown).

Protein Sir22 Binds Monoclonal IgG and IgA Proteins of All Subclasses-The binding properties of protein Sir22 were ana- lyzed in a dot-blot experiment, using a large number of purified monoclonal Igs representing the four subclasses of human IgG and the two subclasses of human IgA (Fig. 2). For comparison, the same Ig preparations were also tested for ability to bind protein Arp4, which binds IgA and also binds to some molecules of IgG (14). In interpreting the data in Fig. 2, we concluded that protein Sir22 or Arp4 showed binding to a protein when reac-

tivity was apparent also at the lowest dilution tested. Protein Sir22 bound to all monoclonal IgGl and IgG4 pro-

teins tested (six of each subclass), to five of the six IgG2 pro- teins tested, and it also bound to two of the six IgG3 proteins analyzed. In contrast, only 4 of the 24 IgG proteins reacted with protein Arp4. In particular, protein Arp bound only 1 out of 12 molecules representing the IgGl and IgG2 subclasses, the two major human IgG subclasses. The dot-blot analysis with mono- clonal IgA preparations showed that protein Sir22 bound mono- clonal IgA molecules of both subclasses, reacting with 11 out of 12 IgA proteins. Binding was also observed with eight other monoclonal IgA proteins (data not shown). Thus, 19 out of 20 monoclonal IgA proteins (9 out of 9 IgAl proteins and 10 out of 11 IgA2 proteins) were shown to bind protein Sir22. The IgA- binding properties of protein Arp4 were similar to those of protein Sir22, and binding was demonstrated for 17/20 IgA proteins tested. In agreement with previous results (14, 301, both protein Sir22 and protein Arp4 lacked reactivity with IgM, IgD, IgE, fibrinogen, and albumin, but bound to rabbit IgG.

Affinity Constants-The equilibrium constants (affinity con- stants) for the reactions between protein Sir22 and serum IgA, secretory IgA, or IgG were determined by SPRIA (Fig. 3). The Scatchard plot for the interaction between protein Sir22 and serum IgAis linear, indicating a single binding reaction with an affinity constant of 7.0 x lo* d . The affinity constant for the

IgA- and IgG-binding Protein Sir 13461

ImlA

1 1 10 100 lo00 UNLABELED lg ADDED (nM)

0.1 1 10 lo0 UNLABELED lg ADDED (nM)

F I G . 4. Inhibition of binding of IgA or IgG to protein Sir22. Protein Sir22 was immobilized in the wells of microtiter plates and incubated with 0.1 ml of radiolabeled IgA(A) or 0.1 ml radiolabeled IgG ( B ) mixed with the indicated amounts of unlabeled Ig in 0.1 ml of phosphate-buffered saline, pH = 7.2, containing 0.02% NaN, and 0.05% Tween 20. ARer 2 h at 25 "C the wells were washed three times and remaining radioactivity measured in a y-counter. The inhibition is ex- pressed as percent of maximal inhibition, where 100% inhibition rep- resents no detectable binding and 0% represents the binding of radio- labeled Ig to protein Sir22 when no unlabeled Ig was present. Each

Addition of 10 rn BSA did not inhibit binding. point represent the mean of the values from triplicate experiments.

reaction between protein Sir22 and secretory IgA was %fold lower, 2.4 x 10' M-'. With regard to I&, analysis of the binding data indicates an affinity constant of 7.8 x 10' M-'. However, when large amounts of IgG were added, a reaction with lower affinity was observed. The results also indicate that protein Sir22 carries the same number of binding sites for serum IgA, secretory IgA and IgG, disregarding the lower affinity binding

The Bindings of IgA and IgG to Sir22 Are Mutually Exclusive-Inhibition experiments showed that IgA and IgG were equally effective at inhibiting the binding of IgG to protein Sir (Fig. 4B ). Both IgA-and IgG also inhibited the binding of IgA to protein Sir, but about~~10 times more IgG than IgA was required to obtain a 50% inhibition (Fig. 4A).

Sequencing of the sir22 Gene and Comparison of Proteins Sir22, H, and Arp4-Plasmid pSIR2202 carries the complete sir22 gene, as shown by Southern blot analysis (data not shown). Sequencing of the entire insert in pSIR2202 (Fig. 5) revealed the presence of an open reading frame (ORF) of 1095 nucleotides from which one can deduce a protein with 365 amino acid residues. Determination of the NH,-terminal se- quence of the mature protein (see above) showed that protein Sir22 starts with a 41-residues-long signal peptide with 66-

to IgG.

98% residue identity to signal peptides of other Ig-binding pro- teins from the group A Streptococcus (9-12, 15, 31-33). The processed form of the protein, 324 amino acid residues, has a calculated M, of 37,186. The amino acid composition of protein Sir22 is similar to that of protein Arp4 and protein H and other members of the streptococcal M protein family, with a high content of Glu, Lys, Ala, Leu, and absence of Cys. Analysis of the secondary structure of protein Sir22 predicts that it is almost totally a-helical, like proteins Arp and H and other members of the M protein family (34).

Protein Sir22 is a member of the streptococcal M protein family. Like other proteins in this family, protein Sir22 has a conserved signal peptide, a central repeat region, and a con- served COOH-terminal region (12,34,35). The central repeats are of the C type, i.e. protein Sir22 is a member of class C in the M protein family (12). The number of C repeats varies between two and four in different members of class C in the M protein family. Based on the definition of C repeats in protein Arp4 (91, two such sequences, C1 and C2, are found in protein Sir22. The C2 sequence is 42 amino acid residues long, whereas the C1 sequence lacks the first 7 residues found in C2.

In Fig. 6, protein Sir22 is compared with protein Arp4 and protein H, both of which are members of class C in the M protein family (9, 11, 12). The NH,-terminal parts of these molecules are most likely responsible for the ability of the pro- teins to bind Ig (see below). In protein Sir22 this region can be divided into several subregions, based on the comparison with protein Arp and protein H. The most NH,-terminal part of the processed form of the protein (41 amino acid residues) is unique, but following this sequence, there are regions which show extensive residue identity to protein Arp4 andor protein H (Fig. 6). It seems likely that at least one of these regions corresponds to the Ig-binding p a d s ) of protein Sir22. The non- repeated NH,-terminal region is longer in protein Sir22, that binds both IgA and IgG, than in protein Arp or protein H. Nevertheless, protein Sir22 is the shortest bacterial Ig-binding protein sequenced so far, since it has only two C repeats.

Analysis of the complete nucleotide sequence of the insert in plasmid pSIR2202 (Fig. 5) indicated that the sir22 gene is surrounded by two other genes encoding proteins in the M protein family, as described also in other strains of S. pyogenes (35). These genes probably encode Mrp22 and Enn22 proteins (12, 36-38).

Deletion Analysis of the sir22 Gene-In an attempt to define the Ig-binding regiods) in protein Sir22, a number of Bal-31 deletions extending from the 3' end of the sir22 gene were analyzed for expression of Ig-binding protein fragments. The fragment expressed by one of the deletion clones, designated E14, comprises the first 156 amino acids of the mature protein Sir22 and corresponds to the nonrepeated NH,-terminal region and part of the C1 sequence (Fig. 5). The exact end point of the deletion in the E14 clone was determined by nucleotide se- quencing. The E14 protein fragment retains the ability to bind IgA and IgG, which shows that both Ig binding activities of protein Sir22 are located in the NH,-terminal half of the pro- tein (Fig. 7). The migration of the E14 fragment in SDS-PAGE corresponds to a molecular weight considerably larger than that expected from the deduced amino acid sequence of the protein, which can be explained by read-through into the vec- tor. Extensive attempts to obtain clones expressing fragments of protein Sir22 shorter than the E14 protein were unsuccess- ful.

DISCUSSION

The streptococcal M protein family comprises a number of cell surface molecules that interact with the immune system of the infected host. Since the group A Streptococcus is a major

13462 IgA- and IgG-binding Protein Sir

1

121

241

-41 3.6 1

481 -30

601 11

721 51

841 91

961 131

108 1 171

1201 211

132 1 251

1441 291

1561 1681

1801

1921

2041

2161

AAGCmAGGAOCTC~CCRGATACTAAACCTGGCAATAAAGAOGTTCC~MGACCATCACAAACMGMCAAACACTMTAAAOCTCCTATOOCGCAAA~GAGACMTTAC 120 A S G A P K P D T K P G N K E V P T R P S Q T R T N T N K A P M A Q T K R Q L P

C G T C A A C A G L O O C G I M C C M C C C A ~ T T C A C T G C A ~ A G ~ T T G A C A G T G A ~ G C A T C T ~ A G O C G T ~ T T ~ C C T ~ G C A A A G M G R A A A C T M G T C C ~ C C ~ A C T A T 240 S T G E E T T N P F F T A A A L T V I A S A G V L A L K R K E E N * * *

360

-10 -10 161 M A R K D T N K Q Y S -31

GCfiAC.AAAATTAAAAACAGGTACATCAOTAGCGOTC~TAOCGGTCGCTGTOOCTGTTTTAGGAOCAOOCTTTGCAAACCAAACMC~TTMGGCGGAGTCATC~TMTGCGGAGTCA~AAA 600 L R K L K T G T A S V A V A V A V L G A G F A N Q T T V K A E S S N N A E S S N 10

CATTTCTCAAGAAAGCAAACTMTAAATACRTTGACTGATG-TGAGAAACTCAGAOMGAGCTCC~AGTATTATGCATTAAGTGATGCT~GMGMG~CTAGGTAT~C 720 I S Q E S K L I N T L T D E N E K L R E E L Q Q Y Y A L S D A K E E E P R Y K A 50

A T T G A G A G G C G R A A A T C M G A T C T T C G G G ~ A A A G ~ T A C C A G G A T R A A A T ~ T T A G M G ~ G ~ ~ C C T A G ~ T C A G M G A T G T A G M C G T C A C T A 840 L R G E N Q D L R E X E R K Y Q D K I K K L E E X E K N L E K K S E D V E R H Y 90

T C T T ~ T A G A T C A A G ~ A T A A A G A A C ~ M G A A C G T C ~ T C T A G A ~ M C T C G M C G T C A A A G T C A A C G ~ A A A T A G ~ A A ~ G T T A T C M G ~ M C T C C ~ A 960 L K K L D Q E E K E Q Q E R Q K N L E E L E R Q S Q R E I D K R Y Q E Q L Q K Q 130

A C M C A A T T AO-AGAAAATCTCAGAAGCT~TC~MGAGCCTMGTCG~ACCTTGMGCGTCTCGTGCAGCTMG~GTAOMOCAG~CT~T~TCTTMTGC 1080

-41

t +1

Q Q ~ ~ ~ K Q I S E A S R K S L S R D L E A S R A + A K K K V E A D L A A L N E14 1 % ca 170 T G A G C A C C ~ C T C A A O O ~ C ~ T C ~ A G A C G C A A G C C G T C M G O C C T A A G C C G T G A C C ~ M O C G T C T C G C G M O C T ~ ~ G T A G M G C A O A C T T A G C C G ~ C 1200 E H Q K L K E E K Q I S D A S R Q G L S R D L E A S R E A K K K V E A D

AAATAOCAARCTTCAAGCCC~~CTAAACAAAGAGCTTG~MGGTAAG-TTATCAG~AAAAAG~GAGTTACMGCMG~TA~GCTGAAGC-GCTCTTM N S K L Q A L E K L N K E L E E G K K L S E K E K A E L Q A R L E A E A K A L K

AGAGCMTTOOCTAAACMGCTGM~CTTGC-CT-GGCMCCAAACACCAAACOCT~GTAOCCCCACAAGCTMCCGTTCMGATCAOCMTGAC~MC~AGMC E Q L A K Q A E E L A K L K G N Q T P N A K V A P Q A N R S R S A M T Q Q K R T

G T T A C C G T C M C A O O C G A C A O C T M C C ~ T T T A L P S T G E A A N P F F T A A A A T V M V S A G M L A L K R K E E N * * *

-10 RD. M A R Q Q T K K N Y S L R K L

ARRACCGGTACGGCTTCAOTA~CGTTGC~ACCGTTTTOOGCOCAGGTT~C~CCAAACOOMGTMGAGC~AT~GCAGTTTCTOO-GT~MGT-GAAAGT~ K T G T A S V A V A L T V L G A G F A N Q T E V R A D E A V S G K V E V K E S E

AAAGAGACTMGTATMGACGTTGGCCTTMGAOOTGRAAATGCTGACC~AGAAACGTAAATGC-TATTTAGAG-TT~GCAOMGMG~T~TTAOAAAAA K E T K Y K T L A L R G E N A D L R N V N A K Y L E K I N A E E E K N K K L E K

G A A A A A C M G A G T T A G - C C M G C C C T T M C ~ C M G A T A E K Q E L E N Q A L N F Q D V I E T Q E K E K E D L K T T L A K A T K E N E I S

GAAGCTAGCCGTAAAGGGTTAAGCCGAGACTTAGMGCTT E A S R K G L S R D L E A

2 10

1320 250

1440 290

1560 324

1680 1800

1920

2040

2160

2200

insert of streptococcal DNA in plasmid pSIR2202 and the derived amino acid sequences of the three longest open reading frames. Possible -35 and FIG. 5. Nucleotide sequence of gene sir22 and partial nucleotide sequences of two surrounding genes. The figure shows the entire

-10 regions are underlined, possible ribosome-binding sites (RBS) are boxed, and potential transcription termination signals are underlined with arrows. The first (truncated) ORF represents the putative Mrp22 protein, the second ORF represents protein Sir22, and the third (truncated) ORF represents the putative Enn22 protein (12, 14,38). The arrows labeled C l and C2 indicate the position of the C repeats in protein Sir22. The arrow

protein Sir22 begins with the first amino acid of the processed protein (indicated with a uerticul arrow and +1) as determined by NH,-terminal labeled E14 indicates the end of the protein Sir22-fragment E14, as determined by nucleotide sequencing. Numbering of amino acid residues in

sequencing of the purified protein. This sequence has been submitted to the EMBUGenBank data base under accession number X75750.

human pathogen, much work has been devoted to the charac- terization of molecules in this family of proteins. In addition to the classical M proteins, which interfere with phagocytosis, the M protein family also includes several Ig-binding proteins with different specificities. Among these Ig-binding cell surface mol- ecules, protein Sir is unique in its ability to bind both IgA and IgG with high affinity.'

Protein Sir reacts with most, but not all, IgA and IgG mol- ecules. The dot-blot analysis with purified monoclonal Igs (Fig. 2) indicated that protein Sir reacted with 19 out of 20 IgA proteins and with 19 out of 24 IgG proteins. The Ig molecules that bound represented all subclasses of IgA and IgG, but it is noteworthy that only two out of six IgG3 proteins reacted with protein Sir. This finding was not surprising, since previous work has shown that two other bacterial IgG-binding proteins, staphylococcal protein A and streptococcal protein Mrp, show limited reactivity with molecules of the IgG3 subclass (12, 39). This limited ability to bind IgG3 molecules may be related to the unique structure of the heavy chain in this IgG subclass (40).

The affinity constants for the binding of protein Sir to poly- clonal serum IgA and polyclonal IgG are of similar magnitude, 7-8 x lo* M-'. These affinity constants are also similar to those

* The protein described in this report was isolated from a strain of serotype M22 and was therefore designated protein Sir22. However, the available data indicate that a protein with similar properties is ex- pressed also by strains of other serotypes (14,42), and the molecule will therefore be referred to as protein Sir in this discussion.

reported for the two other group A streptococcal Ig-binding proteins that have been compared to protein Sir in this report. Thus, the affinity constants for the binding of serum IgA to protein Arp and for the binding of IgG to protein H were de- termined to be 5.6 x lo8 and 1.6 x lo9 "', respectively (4, 16). The affinity constant for the binding of protein Sir22 to secretory IgA was about 3-fold lower than that for serum IgA. This finding is in agreement with previous results, showing that streptococcal IgA-binding proteins have lower affinity for secretory IgA than for serum IgA (6, 17). This difference in affinity can be explained by steric interference due to the pres- ence of secretory component in secretory IgA (41).

The Ig binding properties of protein Sir indicate that this molecule combines the IgA binding ability of protein Arp with the IgG binding ability of protein H. Since these three strepto- coccal proteins are structurally similar, one could speculate that protein Sir has arisen through a genetic recombination event between two genes encoding protein Arp and protein H, respectively. However, the available data indicate that expres- sion of protein Arp or protein Sir is a common property among clinical isolates of the group A Streptococcus (14, 421, whereas protein H has so far been found only in one strain of serotype Ml(11, 16). These data suggest that evolution has favored the appearance of streptococcal cell surface molecules like protein Arp and protein Sir, whereas protein H may have arisen through a unique genetic event in the strain expressing that protein.

Previous work has demonstrated that protein Arp, which

IgA- and IgG-binding Protein Sir M I N O ACID RESIDUE NUMBER:

- 4 1 +1 5 0 100 150 200 250 300 350

13463

Protein Arp4 N c1 cz e3 . . . . . .

Protein Arp4 N -C . . . . . .

Protein Sir22 N -C . . . . .

FIG. 6. Comparison of three Ig-binding proteins from the group AStreptococcus. Protein Arp4 is expressed by a strain of serotype M4, protein Sir22 by a strain of serotype M22, and protein H by a strain of serotype M1. The figures between the aligned proteins represent amino acid sequence identity. As described in the text, the ability of the three proteins to bind Ig most likely resides in the nonrepeated NH,-terminal parts of the molecules. Regions of extensive amino acid residue identity in these NH,-terminal parts of the three proteins are indicated, and the corresponding amino acid sequences are given in the lower part of the figure. Alternative alignments, yielding slightly different results, are possible. The data for protein Arp4 are from Frithz et al. (9), and the data for protein H are from Gomi et al. (11). SI‘, signal peptide; C1, C2, C3, repeated regions; M, putative “membrane anchor” region. The arrow labeled E14 indicates the end of an N€&-terminal fragment of protein sir22 (see Figs. 5 and 7).

A B C i A B C I . I 0

66 - 45 -

I I

31 - 1 ) : - 1 - I

21.5- I I

14.4-

Probe: IgA-Fc Probe: IgG-Fc FIG. 7. Western blot analysis of the N&terminal E14 fragment

of protein Sir22. The figure shows the Ig-binding proteins detected in three derivatives of E. coli carrying different plasmids: A, plasmid pE14, expressing the NH,-terminal E14 fragment of protein Sir22 (see Figs. 5 and 6); B, plasmid pSIR2202, expressing the entire protein Sir22; C, plasmid pK18, lacking insert of streptococcal DNA. 200-pl overnight cultures of the E. coli strains were concentrated to 50 pl, boiled for 3 min in a solution containing 2% SDS and 5% 2-mercapto- ethanol, separated by SDS-PAGE in duplicate, electroblotted to PVDF membranes, and probed with either radiolabeled IgA-Fc or IgG-Fc. Molecular mass markers, indicated on the left, are in kilodaltons.

binds IgA with high affinity, also binds some IgG molecules (4, 14). Protein Arp can therefore be regarded as a molecule that binds both IgA and I&, like protein Sir. However, protein Arp reacted poorly or not at all with most monoclonal IgG proteins studied here (Fig. 2). In particular, protein Arp reacted with only 1 out of 12 molecules representing the IgGl and IgG2 subclasses, which are the two major IgG subclasses in man and comprise 90% of human serum IgG. We previously suggested that protein Arp preferentially binds to IgG molecules of the IgG3 subclass (141, but the more extensive data now available show that certain IgG molecules of other subclasses are also able to bind to protein Arp. It should be noted that the inability of protein Arp to bind to the majority of IgG molecules can explain the finding that the affinity of this protein for poly- clonal IgG was too low to be measurable (4).

The Ig binding activities of protein Sir are located in the NH,-terminal half of the molecule, as shown by characteriza-

tion of the E14 fragment. This fragment includes the nonre- peated NH,-terminal region of the molecule and part of the C1 repeat (Figs. 5 and 6). For protein H, it has been shown that the IgG-binding region is located in the nonrepeated NH,-terminal part of the molecule (43), and there is evidence that the same is true for protein Arp (17). It therefore seems likely that the ability of the E14 fragment to bind Ig is due to the presence of the nonrepeated NH,-terminal part of protein Sir. The simplest hypothesis is clearly that the Ig-binding regions of proteins Sir, H, and Arp are located in the NH,-terminal regions where these three proteins show extensive residue identity (Fig. 6). The data reported in this paper provide the basis for future analysis of this question. Future characterization of the binding re- gion(s) in protein Sir will also reveal whether this protein has a single binding site that recognizes the Fc region of both IgA and IgG or if protein Sir has two separate binding sites, one for IgA and one for IgG. Although inhibition experiments indicate that IgAand IgG may bind to the same site (Fig. 41, the data are also compatible with the existence of two separate, but adja- cent, binding sites where the binding of IgA sterically interferes with the binding of IgG and vice versa.

In summary, we have reported the gene sequence, purifica- tion, and some immunochemical properties of protein Sir, a streptococcal cell surface protein that binds both IgA and IgG with high affinity. The biological function of protein Sir is not known, but it seems possible that the bacteria use this protein to cover themselves with IgA or with IgG, depending on the class of Ig that predominates at the site where the bacteria are multiplying in the infected host. Further characterization of protein Sir is therefore of interest not only from an immuno- chemical point of view, but also for the understanding of the host-parasite relationship during streptococcal infection. Fi- nally, it should be noted that cell surface proteins binding ei- ther IgA or IgG have previously been described both in bacteria and in mammals (44,45). Since a protein binding both IgA and IgG has now been found in bacteria, it seems possible that proteins with similar properties may be found also in higher organisms.

13464 &A- and IgG-binding Protein Sir Acknowledgments-We are indebted to Dr. Lars Bjorck for the gi f t of

purified protein H, to Dr. F. Skvaril (World Health Organization/ International Union of Immunological Societies Immunoglobulin Sub- committee, Bern, Switzerland) for the gift of monoclonal human IgG proteins and to Dr. B. herstrijm for helpful discussions. Annika hdersson and Rose Kulhavy provided excellent technical assistance.

REFERENCES 1. Forsgren, A., and Sj6quist. J. (1966) J. Immunol. 97,822-827 2. Bjorck, L., and Kronvall, G. (1984) J. Immunol. 133, 969-974 3. Boyle, M. D. P. (ed) (1990) Bacterial Immunoglobulin-binding Proteins, Vol. I,

4. Lindahl, G., and Akerstrom, B. (1989) Mol. Microbiol. 3, 239-247 5. Lindahl, G., herstram, B., Stenberg, L., Frithz, E., and HedBn, L.-0. (1991) in

Genetics and Molecular Biology of Streptococci, Lactococci and Enterococci (Dunny, G. M., Cleary, P. P., and McCay, L. L., eds) pp. 155-159, American

6. Lindahl, G., Akerstmm, B., Vaerman, J.-P., and Stenberg, L. (1990) EUI: J. Society for Microbiology, Washington, D. C.

Immunol. 20,2241-2247 7. HedBn, L.-O., Frithz, E., and Lindahl, G. (1991) Eur. J. Zmmunol. 21, 1481-

8. Jerlstrom, P. G., Chhatwal, G. S. , and Timmis, K N. (1991) Mol. Microbiol. 6, 1490

843-849

10. Heath, D. G., and Cleary, P. P. (1989) Proc. Natl. Acad. Sci. U. S. A. 86, 9. Frithz, E., HedBn, L.-O., and Lindahl, G. (1989) Mol. Microbiol. 3,1111-1119

11. Gomi, H., Hozumi, T., Hattori, S., Tagawa, C., Kishimoto, F., and Bjorck, L.

12. Omole, P., Stenberg, L., Rissler, M., and Lindahl, G. (1992) Proc. Natl. h a d .

13. Raeder, R., Otten, R., Chamberlin, L., and Boyle, M. D. P. (1992) J. Clin.

14. Stenberg, L., OTbole, P., and Lindahl, G. (1992) Mol. Microbiol. 6, 1185-1194 15. Retnoningnun, D. S., Podbielski, A., and Cleary, P. P. (1993) J. Immunol. 160,

16. Akesson, P., Cooney, J., Kishimoto, F., and Bjorck, L. (1990) Mol. Immunol. 27,

17. Akerstrom, B., Lindqvist, A,, and Lindahl, G. (1991) Mol. Immunol. 28, 349-

18. Akerstmm, B., Lindahl, G., Bjorck, L., and Lindqvist, A. (1992) J. Immunol.

Academic Press, New York

47414745

(1990) J. Immunol. 144,4046-4052

Sei. U. S. A. 89, 86614665

Microbiol. 30, 3074-3081

2332-2340

523-531

357

19. Pridmore, R. D. (1987) Gene (Amst.) 66, 309-312 20. Yanisch-Perron, C., Vieira, J., and Messing, J. (1985) Gene (Amst.) 33,103-119 21. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: A

Laboratory Manual, 2nd Ed., Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

148,32364243

22. Misra, T. K. (1985) Gene (Amst.) 34,263-268 23. Anraku, Y., and Heppel, L. A. (1967) J. Biol. Chem. 242,2561-2569 24. Mestecky, J., and Kilian, M. (1985) Methods Enzymol. 116, 37-75 25. Scatchard, G. (1949) Ann. N. Y. A d . Sci. 61, 660-672 26. Myhre, E. B., and Kuusela, P. (1983) Infect. Immun. 40, 29-34 27. Matsudaira, P. (1987) J. Biol. Chem. 282, 10035-10038 28. Fischetti, V. A., Jones, K F., Manjula, B. N., and Scott, J. R. (1984) J. Exp. Med.

29. Hollingshead, S . K., Fischetti, V. A., and Scott, J. R. (1986) J. Biol. Chem. 261,

30. Lindahl, G. (1989) Mol. & Gen. Genet. 216,372-379 31. Bessen, D., and Fischetti, V. (1992) Infect. Immun. 60, 124-135 32. Hedh, L.-O., and Lindahl, G. (1993) J. Gen. Microbiol. 139,2067-2074 33. Akesson, F!, Schmidt, K-H., Cooney, J., and Bjorck, L. (1994) Biochem. J . , in

34. Fischetti, V. (1989) Clin. Microbiol. Rev. 2, 285-314 35. Kehoe, M. A. (1994) New Compr. Biochem. 27,217-261 36. Haanes, E. J., Heath, D. G., and Cleary, P. P. (1992) J. Bacteriol. 174, 4967-

37. Jeppson, H., Frithz, E., and HedBn, L.-0. (1992) FEMS Microbiol. Lett. 92,

38. Whatmore, A. M., and Kehoe, M. A. (1994) Mol. Microbiol. 11,363-374 39. van Loghem, E., Frangione, B., Recht, B., and Franklin, E. C. (1982) Scad. J.

40. Michaelsen, T. E., Frangione, B., and Franklin, E. C. (1977) J. Biol. Chem. 262,

41. Akerstmm, B., Lindqvist, A., Vander Maelen, C., Grubb, A,, Lindahl, G., and

42. Lindahl, G., and Stenberg, L. (1990) Epidemiol. Infect. 106,87-93 43. Frick, I."., Akesson, P., Cooney, J., Sjobring, U., Schmidt, K.-H, Gomi, H.,

Hattori, S. , Tagawa, C., Kishimoto, F., and Bjork, L. (1994) Mol. Microbiol, in press

44. Maliszewski, C. R., March, C. J., Schoenborn, M. A., Gimpel, S . , and Shen, L. (1990) J. Exp. Med. 172,1665-1672

45. van de Winkel, J. G. J., and Capel, P. J. A. (1993) Immunol. 'Ibday 14,215-221

169, 1083-1095

1677-1686

press

4976

139-146

Immunol. 16.275-278

883-889

Vaerman, J.-P. (1994) Mol. Immunol., in press