IgA receptor from group B streptococci
Eur. J. Immunol. 1990. 20: 2241-2247
2241
Gunnar Lindahl+, Bo AkerstromO, Jean-Pierre Vaerman. and Lars Stenberg+
Characterization of an IgA receptor from group B streptococci: specificity for serum IgA*
Department of Medical Microbiology+ Department of Physiological Chemistryo, University of Lund, Lund and Experimental Medicine Unit, International Institute of Cellular and Molecular Pathology, Catholic University of Louvain., Brussels
Some strains of group B streptococci express a cell surface protein which binds 1gA.This report describes some properties of such an IgA receptor and compares it with a previously described IgA receptor from group A streptococci.The group B receptor was released in an almost pure form from bacteria incubated at elevated pH, and could be isolated by IgA-Sepharose affinity chromatography. The sequence of the N-terminal19 amino acid residues was unique.The receptor preferentially binds IgA of human origin, as shown in immunoblotting experiments with purified IgA from nine different species. The affinity constant of the purified receptor for serum IgA was determined to be 3.5 X lo8 M-l, but for secretory IgA it was too low to allow determination. This result indicates that secretory component and/or J chain interferes with the binding of IgA to this type of bacterial receptor, which may be one of the physiological functions of these polypeptides. A reduction in affinity was also observed for another complexed form of IgA, al-microglobulin-IgA. The group B receptor is antigenically unrelated to the IgA receptor from group A streptococci (protein Arp), but competitive inhibition experiments indicate that they bind to the same region in IgA. The implications of these findings, and the biological role of bacterial IgA receptors, are discussed.
1 Introduction Cell surface receptors that bind to the Fc region of IgA have been found in certain strains of both group A streptococci and group B streptococci, two common human pathogens [l,21. For both of these bacterial species, the IgA receptor (IgAR) has been purified and partially characterized. In group B streptococci, where the IgAR was first defined as an antigen called p [3,4], the purified receptor was shown to have a molecular weight of about 130000 [l]. There is evidence that this IgA-binding protein acts as a virulence factor [ 5 ] , but it is not yet known whether this is due to the capacity to bind IgA or to some other property of the protein. The IgAR in group A streptococci, called protein Arp, has been purified from two different streptococcal strains after expression of the cloned genes in Escherichia coli, and was shown t o have an apparent molecular weight of about 41 000 [6, 71. The biological function of protein Arp is not known, but the fact that this receptor shows extensive sequence similarity to streptococcal M proteins [8], which are well-known virulence factors [9], suggests that this receptor may also contribute to bacterial virulence. In this report,we describe a simple method for isolating the group B receptor, which has been immunochemically
characterized and compared with the previously isolated receptor from group A streptococci, protein Arp.The most remarkable property of the group B receptor is that the affinity of this receptor for secretory IgA is much lower than for serum IgA. This suggests that one function of the secretory component and/or J chain is to interfere with the binding of secretory IgA to bacterial IgAR.
2 Materials and methods 2.1 Bacterial strains
Two different series of group B streptococcal strains were used. These strains were isolated from clinical specimens sent to the Laboratory of Clinical Microbiology, Lund University Hospital. One series of 48 strains, the SB strains, obtained from Drs. G. Kronvall and C. Schonbeck, originate from a variety of specimens, including urine, vaginal secretions and pus. A second series, the 27 BS strains, represent all septicemia strains of group B streptococci isolated in our laboratory during a period of 7 years. The group B streptococcal strain A909 [lo] was obtained from Dr. C. Schalen. The group A streptococcal strains AW43 and AP4, both of which bind IgA, have been described [6, 71. The streptococcal strains were grown in Todd-Hewitt broth (Oxoid), Basingstone, GB).
[I 85981
*
This work was supported by grants from the Swedish Medical Research Council, the Medical Faculty of the University of Lund, King Gustaf V’s 80-year Foundation, Fysiografiska Sallskapet in Lund, the Foundations of Osterlund, Johansson, and Kock and by HighTech Receptor, Inc., Malmo, Sweden.The partial support of the Belgian State-Prime Minister’s Office Science Policy Programming is also acknowledged.
Correspondence:Gunnar Lindahl, Department of Medical Microbiology, University of Lund, Solvegatan 23, s-223 62 Lund, Sweden 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1990
2.2 Selection of IgA-binding strains The ability of bacterial strains to bind radiolabeled serum IgA was tested in a standard binding assay with 2 x 108 bacteria in 200 p1 of buffer, as described [6]. With a few exceptions, the strains could be unequivocally classified as binders or non-binders, showing > 40% or < 5% binding, respectively. Five out of 48 strains in the SB series and 5 out of 27 strains in the BS series were able to bind IgA. None of these strains could bind radiolabeled IgG or IgM. One of 0014-2980/90/1010-2241$3.SO + .25/0
2242
G. Lindahl, B. kerstrom, J.-PVaerman and L. Stenberg
the strains, SB35, was chosen for the biochemical work described in this report.This strain,which was isolated from urine, is of serotype Ic (typing kindly performed by Ms. I. Wilson, Department of Medical Microbiology, University of Lund) .
2.3 Proteins Ig were of human origin and polyclonal, unless otherwise stated. The origins of some preparations have been described [6]. Purified preparations of human monoclonal IgG proteins of all four subclasses were provided by Dr. F. Skvaril, WHOAUIS Immunoglobulin Subcommittee, Berne, Switzerland. The preparations of purified secretory IgA from the rabbit [ l l ] , rat [12], guinea pig [13], dog [14], pig [151, cow [161and horse [171, as well as sheep mesenteric lymph IgA [18] have been described. The preparation of al-microglobulin-IgA, purified as described [19], was the gift of Dr. Anders Grubb, Department of Clinical Chemistry, University of Lund. Two mouse IgA myeloma proteins,TEPC 15 and MOPC 315,were purchased from Litton Bionetics, Kensington, MD. Proteins Arp4 and Arp60 were purified as described [6, 71. 2.4 Purification of group B receptor The method used is based on the observation that the receptor is released from bacteria suspended in buffer of high pH (see Sect. 3.1). Strain SB35 was grown overnight in 20 1 Todd-Hewitt broth, collected by centrifugation, and washed twice with 50mM Tris, pH 7.0.The washed bacteria (about 30 g) were then suspended to 10% (v/v) in 50 mM Tris, pH 11.0, giving a final pH of 9.7. The suspension was incubated at 37°C for 4 h, centrifuged, and the SN was immediately dialyzed against PBS. The dialyzed preparation was frozen in portions, from which the receptor was later purified by affinity chromatography on IgA-Sepharose, as described [6]. Using this procedure, about 20 mg of receptor can be recovered from 30 g of washed bacteria.
Eur. J. Immunol. 1990. 20: 2241-2247
2.6 Amino acid sequence analysis Automated amino-terminal amino acid sequencing was performed with an Applied Biosystems (Foster City, CA) 470A gas-liquid solid-phase sequenator as described [19]. Amino acid sequence homology analysis was done with the program WORDSEARCH, which is included in the “Wisconsin” software [20]. NBRF Protein and Sequence Data Library and NBRF New Sequence Data Library were used as sources for published amino acid sequences. 2.7 Competitive inhibition experiments
The wells of microtiter plates (highly activated flat-bottom Titertek PVC Immunoassay microplates; Flow Laboratories, Aberdeen, GB) were coated overnight at 4°C with 100 p1 of polyclonal IgA, 10 pg/ml in PBS. Remaining binding sites were blocked by the addition of 100 pl of VBS (10 mM veronal buffer, 0.15 M NaCl, pH 7.4) containing 0.25% gelatin and 0.25% Tween 20, and incubation was continued at room temperature for 4 h and then at 4°C overnight. After washings with 0.12 M NaCl, 0.03 M sodium phosphate, 0.02% NaN3, 0.05% Tween 20; pH7.2 (PBSAT), approximately 5 ng (about 1.5 x 104 cpm) of radiolabeled IgAR and various amounts (10 ng-4 pg) of unlabeled IgAR receptor were added in 50 p1 PBSAT and the plates were incubated at room temperature for 4 h.The wells were then washed with PBSAT, separated by cutting, and counted in a gamma counter. 2.8 Determination of equilibrium constants
The equilibrium constants for the binding between group B receptor and various molecular forms of IgA were determined by a solid-phase RIA [6]. Briefly, the receptor, coated to microtiter plates, was incubated overnight with approximately 1ng 1251-labeledIgA and various amounts of unlabeled IgA of the same molecular form. The amount of bound radiolabeled IgA was determined, and the equilibrium constant was obtained after plotting the data and calculating according to Scatchard [21]. All incubations were done at room temperature.
2.5 Electrophoresis, immunoblotting and dot blots The electrophoretic methods used have been described [6]. Proteins separated by PAGE were transferred to membranes of polyvinylidene difluoride (Immobilon, Millipore Cop,Molsheim, France). The blocking, washing, incubations with 12%labeled protein probes and autoradiography were then performed as described [6]. Dot blots were also performed as described [6], except that Immobilon membranes were used. In the dot blot for which antisera were used for the analysis (see Fig. 8), the Immobilon membranes were first blocked in a buffer containing gelatin [6] and then incubated for 1 h at room temperature in rabbit antiserum, or preimmune serum, diluted 1000-foldin 15 ml of “VBS-gel” (10 mM veronal buffer, 0.15 M NaCl, 0.1% gelatin, pH 7.4). After washings (4 X 20 min) in 250 ml of 1M NaCl, 10 mM EDTA, 0.25% gelatin and 0.25% Tween 20, the filters were incubated for 3 h in 15 ml of “VBS-gel” containing radiolabeled protein G (2 x lo5 c p d m l) , followed by four washes as before [7]. All these procedures were carried out at room temperature and were followed b y autoradiography at - 70 “C.
2.9 Other methods
Radiolabeling was performed with carrier-free 1251(Amersham International, Amersham, GB), using a modified lactoperoxidase method [22] or chloramin T [23]. Binding assays with whole bacteria were performed as described [6]. Antibodies to purified group B receptor were obtained by S.C. immunization of a rabbit with 100 pg of the protein emulsified in CFA. Two immunizations were made with an interval of 5 weeks. The rabbit was bled before immunization and 1week after the last immunization.The antiserum to protein Arp4 has been described [7].
3 Results 3.1 Release of IgAR from strains of group B streptococci incubated at elevated pH
Previous work with strains of group B streptococci that bind IgA has shown that the IgAR is present in the growth
IgA receptor from group B streptococci
Eur. J. Immunol. 1990. 20: 2241-2247
medium, but this protein is size heterogenous [24]. Such a size-heterogeneous IgAR is also found in the growth medium of strain SB35, the strain studied here (Fig. 1). A method for isolating more size-homogeneous receptor was described by Russell-Jones et al. [l], who purified the protein from an extract prepared by boiling bacteria in 2% SDS. However, during work with strain SB35 we noted that the size-homogeneous receptor can also be easily obtained by incubating washed bacteria at elevated pH. The protein released by this treatment appears to be almost pure receptor. Minor amounts of contaminating proteins can be removed by affinity chromatography on IgA-Sepharose (Fig. 1). The molecular weight of the purified receptor is about 130O00, similar to the protein described by RussellJones et al. [l]. The release of receptor at elevated pH has not yet been studied systematically, but preliminary experiments indicated that optimal yields of receptor were obtained by suspension of washed bacteria in glycine-buffer, pH 11.0, and incubation at 37°C for 4 h. The final pH of such a suspension was 9.7. Suspension in buffers of lower pH (2-5) also caused release of receptor but this material was more size heterogeneous (data not shown). There was little release of receptor at 0 "C, but addition of sodium azide (0.05%) did not affect the yield at 37°C (data not shown). BLOT
STAIN 200
-
116-
E
20 10
5 2.5 1.25 0.63 0.3 1
Figure 3. Dot-blot assay on Immobilon, showing binding of radiolabeled group B receptor to different proteins. The indicated amounts of protein, diluted in PBS, were applied in 10O-pl aliquots. After blocking, the membrane was probed with '251-labeled receptor,washed and autoradiographed. All Ig preparations except IgM were monoclonal.
The release of IgAR at elevated pH is not a unique property of strain SB35, since similar results were obtained with 6 other IgA-binding strains of group B streptococci, including the commonly used type Ic strain A909. For all of these strains, the receptor was found to have a molecular weight of about 130000 (data not shown). On the other hand, no receptor was recovered from two different IgA-binding strains of group A streptococci treated in the same way.
3.2 N-terminal amino acid sequence of the group B receptor
97-
The N-terminal amino acid sequence of the group B receptor is shown in Fig. 2. Positions nos. 1,12and 16could not be identified, and are represented by the letter X in the figure. The sequence was compared with published sequences and was found to be unique.
66-
43-
1
2
3
4
5
6
7
8 9 1 0
Figure 1 . Release of IgAR from strain SB35 incubated at elevated pH. Bacteria were grown overnight in Todd-Hewitt broth, washed twice in 50 m M Tris, pH 7.0, and then resuspended to 10% (v/v) in buffers of different pH. After incubation for 4 h at 37"C, the suspensions were centrifuged and samples of the SN were separated on SDS-PAGE (T = 8%, C = 3.3%) under reducing conditions. Lanes 1 and 6: SN from the culture in Todd-Hewitt broth, concentrated &fold, lanes 2 and 7: bacteria suspended in 50 m M Tris, pH 7.0, lanes 3 and 8: bacteria suspended in 50 mM Tris, pH 9.0, lanes 4 and 9: bacteria suspended in 50 mM glycine, pH 11.0, lanes 5 and 10: group B receptor further purified by affinity chromatography. Lanes 1-5 were stained with Coomassie brilliant blue; lanes 6-10 were blotted onto Immobilon membranes and probed with radiolabeled IgA. Molecular mass (kDa) markers are on the left.
1
2243
5
10
15
Figure 2. N-terminal sequence of the group B receptor. Residues nos. 1, 12 and 16 could not be identified and are denoted by "X.
3.3 Binding specificity of the group B receptor Radiolabeled receptor was used in a dot-blot assay to study the specificity of the binding (Fig. 3). Membranes of polyvinylidene difluoride (Immobilon) were used for this analysis, since preliminary experiments unexpectedly showed very poor binding of the receptor to IgA immobilized on nitrocellulose membranes. The data in Fig. 3 show that the receptor binds to monoclonal IgA proteins of both subclasses. As expected, the receptor binds t o Fc fragments, but not to Fab fragments, of monoclonal 1gA.There was no binding to IgG, IgM, IgD or IgE, nor to albumin or fibrinogen. These binding properties of the purified receptor are similar to those of whole SB35 bacteria, which bind 19
Eur. J. Immunol. 1990. 20: 2241-2247
G. Lindahl, B. kerstrom, J.-PVaennan and L. Stenberg
2244
1 2 3 4
IgA, but not IgG, IgM, albumin or fibrinogen (data not shown). Our results also agree with the data of RusselJones et al. [l],who used an inhibition test to study the specificity of their receptor.
5 6 7 8 9101112
-
92
- 67
The purified group B receptor binds to the a H chain of IgA separated by SDS-PAGE (Fig. 4). Surprisingly, there was also a weaker reactivity with IgG y chains, but no binding to IgM p chains. Prolonged exposure of the X-ray film also showed a very weak binding to the L chains of these Ig (not shown). When whole plasma was analyzed in the same way, two bands appeared in the blotting experiment, corresponding t o a and y chains. These two bands were of similar strength, although the serum concentration of IgG is about sixfold higher than that of 1gA.These data indicate that the denaturing conditions of the SDS-PAGE cause exposure of y chain determinants that have affinity for the group B receptor. O n the other hand, the SDS-PAGE analysis did not show reactivity of the receptor with any other plasma protein, except for the a and y chains.
- 46 - 30
- 14
- 92
- 67 - 46
To study the species specificity of the receptor, preparations of purified IgA from nine different species were analyzed in a Western blot experiment, using radiolabeled receptor as the probe. Binding was observed only for a chains of human origin (Fig. 5). 3.4 Equilibrium constant of the binding between the
group B receptor and different molecular forms of IgA In man, IgA occurs in at least three different forms. Most of the IgA is secreted as secretory IgA which is usually a dimer containing a J chain and a chain of secretory component in addition to the two IgA monomers [25]. In contrast, most of the IgA present in serum is monomeric, but about 5% of this IgA occurs in a complex, where it is covalently linked to the serum protein al-microglobulin in a molar ratio of 1: 1 [19]. The affinity constant (equilibrium constant) of the group B receptor for these three forms of IgA was
- 30 - 14
Figure.5. Analysis of the species specificity of the group B receptor. Two micrograms of animal IgA (lanes 1-10), human secretory IgA (lane 11) or human serum IgA (lane 12) were separated by SDS-PAGEand stained with Coomassie brilliant blue (A). An identical gel was transferred to an Immobilon membrane and incubated with 1251-labeled group B receptor (B). The animal IgA were from rabbit (lanes 1-2), cow (lanes 3-4). sheep (lane 5), rat (lane 6), guinea pig (lane 7), horse (lane 8), pig (lane 9),and dog (lane 10). The animal IgA were secretory IgA, except that from sheep, which originates from lymph. The molecular masses of the standard proteins are given in kDa.
K,(M-1*10-8)
BLOT
STAIN
0
Serum IgA 3.5
0 al-m-lgA
66
-
.
45
-
m
29 24
--
20
-
0.14
C
a 0.1
i
1
2
3
4
5
6
7
%
Figure 4. Binding of group B receptor to separated a chains of IgA and y chains of IgG.Ten micrograms of serum IgA (lanes 1 and 5), 10 pg of IgG (lanes 2 and 6), 10 pg of IgM (lanes 3 and 7) and 50 p1 of human plasma diluted 100-fold were separated on SDS-PAGE (T = 10%;C = 3.3%).The gel wasstained with Coomasie brilliant blue (lanes 1-4), or transferred to an Immobilon membrane, which was incubated with 1251-labeledreceptor (lanes 5-8). Molecular mass markers, indicated on the left, are in kDa.
2
3 Bound lgA (nM)
4
5
Figure 6. Scatchard plot of the binding between group B receptor and two different molecular forms of IgA. The wells of microtiter plates were coated with group B receptor overnight, and after washing they were incubated for 5 h with approximately 1 ng '251-labeled serum IgA or a,-microglobulin-IgA, and varying amounts of unlabeled IgA of the same type. For each IgA concentration the ratio boundfree IgA was calculated and plotted against bound IgA. The equilibrium constant was then calculated from the slope of the linear diagram. Each point represents the
mean of triplicate experiments.
Eur. J. Immunol. 1990. 20: 2241-2247
IgA receptor from group B streptococci
2245
determined in a solid-phase RIA system with the receptor immobilized to the plastic walls of microtiter plates. The resulting Scatchard plots are shown in Fig. 6. The affinity constant for serum IgA was found to be 3.5 X lo8 M-l, but for the al-microglobulin-IgA complex it was 25-fold lower, 0.14 x lo8 M - ~ .The affinity constant for the binding to secretory IgA was too weak to permit Scatchard analysis.
interest to study the ability of these various forms of IgA to bind to whole bacteria, i.e. to receptor on the bacterial cell surface. The data in Fig. 7 show that the three forms of IgA differ both in the rate with which they bind to the bacteria and in the maximal amount of binding achieved. These results are in good agreement with the equilibrium constants described above, since serum IgA and secretory IgA show the highest and lowest rate of binding, respectively.
3.5 Binding of different molecular forms of IgA to whole bacteria
3.6 The group B receptor and protein Arp are immunologically unrelated but bind to the same region in IgA
The finding that purified group B receptor has different affinity for the three molecular forms of IgA made it of
60
t
0
serum IgA
0 Orl-m-IgA A secretory IgA
,.
0
The IgAR isolated from group A and group B streptococci were compared in two ways. First, the two types of IgAR were compared immunologically, using rabbit antisera raised against purified preparations of the receptors. As shown in Fig. 8, there was no cross-reactivity between the two types of receptor. O n the other hand, two variants of protein Arp showed strong cross-reactivity. As expected, no binding was observed when preimmune rabbit serum was used in the test (data not shown).
.
a 1
2
-A-
3
,
loot
Time (h)
Figure 7. Binding of different molecular forms of IgA t o strain SB35 as a function of time.Washed bacteria (2 X 108 cells in 200 p1 of buffer) were mixed with 2-5 ng (about 104cpm) of radiolabeled protein and incubated at room temperature for the time indicated. Ice-cold buffer (2 ml) was then added, the samples were centrifuged and radioactivity in the pellet was measured.
/ /
/group
A' /
'
B rec.
p' 8
2
100
1000
Unlabeled IgA receptor added (ng)
1
antiArp4
0.5 0.25 0.13
Figure 9. Inhibition of binding of lzI-labeled protein Arp4 to immobilized human serum IgA by unlabeled protein Arp4 or group B receptor. See Sect 2.7 for details.
0.06
2 1
antigroup B roc.
0.5 0.25 0.13
I
0.06
Figure 8. Dot-blot analysis of cross-reactivity between streptococcal IgAR. The indicated amounts of protein, diluted in PBS, were applied to Immobilon membranes in 100-pl aliquots. After blocking, the membranes were incubated with rabbit antiserum of the type indicated. The filter was then washed and bound rabbit IgG detected with 1251-labeledprotein G.
A second type of comparison between the two streptococcal IgAR was designed to study whether they bind to the same region in the Fc part of IgA. This was studied in a competitive inhibition experiment, as previously used in comparisons between different bacterial IgG receptors [26].The data in Fig. 9 show that the binding of radiolabeled protein Arp4 to IgA was inhibited by both receptors. Similarly, the binding of radiolabeled group B receptor to IgA was also inhibited by both receptors (data not shown). These data suggest that the two IgAR bind to the same region in the Fc part of IgA. This inhibition was not unspecific, since addition of a control protein (human haptoglobin) was without effect.
2246
G. Lindahl, B. Akerstrom, J.-RVaennan and L. Stenberg
4 Discussion Two different streptococcal IgAR have now been characterized in considerable detail: protein Arp from group A streptococci and the receptor from group B streptococci studied here. These two bacterial IgAR are structurally different, as shown by the fact that they are antigenically unrelated and are of very different size. The mode of cell wall attachment also appears to be different for the two receptors, since the group B receptor, but not protein Arp, is released in almost pure form from bacteria incubated at elevated pH. The mechanism of this release is not known, but it forms the basis of a simple method for purifying the group B receptor. Similar release of receptor has also been observed at low pH, and it seems possible that it could favor the establishment of an infection in tissues where the pH is not neutral. A remarkable property of the purified group B receptor is that the affinity constant for the binding to secretory IgA is much lower than for serum IgA, and could not be determined. In agreement with this result, binding experiments with whole bacteria showed very poor binding of secretory IgA (Fig. 7). Poor binding of secretory IgA to whole bacteria has also been observed by Brady and Boyle [24], who suggested that the solubilized receptor has different properties and is able to bind secretory IgA. This conclusion was based on the observation that radiolabeled secretory IgA can be used to detect the solubilized receptor in a Western blot. However, this apparent binding of secretory IgA can be readily explained by antibody activity against the receptor in the Ig preparation used [27]. Such antibody activity has indeed been observed by us in polyclonal Ig preparations (data not shown). The presence of such antibody activity in polyclonal preparations of human IgA is not surprising, since group B streptococci are part of the normal flora in man [28]. Since secretory IgA is the predominant antibody on mucous membranes, it was surprising that the group B receptor binds poorly to this molecular form of IgA. However, seen from the point of view of the host, this finding suggests that one function of secretory component and/or J chain is to interfere with the binding of secretory IgA to specific receptors. On the other hand, the function and mechanism of action of the group B receptor in bacterial infection remain unclear. It even seems possible that the in vivo function of this bacterial virulence factor is to bind to some protein other than IgA, e.g. a cell adhesion molecule, some of which are known to be structurally related to Ig [29]. An indication that the IgAR might also be able to bind to other members of the Ig superfamily comes from a Western blot experiment (Fig. 4),which shows that under denaturing conditions the receptor binds not only to a chains of IgA but also to y chains of IgG. In addition to secretory IgA, human IgA also occurs in another complexed form, al-microglobulin-IgA [19]. In this complex, which represents about 5% of the IgA found in serum, one molecule of IgA is covalently linked to one molecule of al-microglobulin, an immunomodulatory serum protein with a molecular weight of 26000 [30, 311. The equilibrium constant for the binding of the group B receptor to this complex was found to be 25-fold lower than for serum IgA. The group B receptor, therefore, has
Eur. J. Immunol. 1990. 20: 2241-2247
reduced affinity also for this complexed form of IgA, which shows that the bound al-microglobulin molecule can drastically affect the ability of IgA to bind to a specific receptor.This finding may be of significance with regard to the physiological function of al-microglobulin. Several lines of evidence now indicate that the IgAR isolated from group A and group B streptococci have similar binding specificity. Both receptors bind in the Fc region and they recognize both subclasses of IgA [ l , 6, 7, 321. The competitive binding experiments reported here (Fig. 9) also indicate that the two receptors bind to the same or overlapping site(s) within the Fc region of IgA. Similar results have been obtained with three structurally different bacterial IgGR, all of which appear to bind in the same region of the Fcy fragment [26, 33, 341. Another similarity between the two streptococcal IgAR is that both receptors apparently are specific for IgA of human origin (Fig. 5) [1]*. Finally, both receptors show higher affinity for serum IgA than for the two complexed forms of IgA, although the differences in affinity are less dramatic for protein Arp than for the group B receptor*. A difference between the two IgA receptors is that protein Arp, but not the group B receptor, also binds weakly to the Fcy region [6, 71. However, this binding to IgG apparently takes place at a separate site on the protein Arp molecule [6]*. In conclusion, the results reported here show that the IgAR in group A and group B streptococci are structurally unrelated but have similar binding properties. This implies that the two receptors are the result of convergent evolution, which has favored the appearance of these different IgA receptors in streptococci. Our results also suggest that one of the functions of secretory component and/or J chain is to interfere with the binding of secretory IgA to specific receptors. Regardless of the biological functions of the different IgAR, these proteins are potentially useful as immunological tools and may also be used as model systems for the analysis of Ig binding to specific receptors. We are grateful to Dr. Anders Crubb for kindly donating the preparation of a!-microglobulin-lgA and to Drs. J. E. Butler, W C. Hanley, A . J. Husband and F? C. Montgomery for generous gifts of purified animal IgA. The technical assistance of Ms. Gunilla Lovhult and Ms. Ann-Charlotte Johansson is gratefully acknowledged. The amino acid sequence analysis was kindly performed by Ms. Ingrid Dahlqvist. Monoclonal human IgG proteins of different subclasses were provided by Dr. E Skvaril, WHOIIUIS Immunoglobulin Subcommittee, Berne, Switzerland.
Received June 5. 1990.
5 References 1 Russell-Jones, G. J., Gotschlich, E. C. and Blake, M. S., J. Exp. Med. 1984. 160: 1467. 2 Christensen, P. and Oxelius,V. A., Acta Pathol. Microbiol. Scand. [C] 1975. 83: 184. 3 Wilkinson, H. W. and Eagon, R. G., Infect. Immun. 1971. 4: 596. 4 Bevanger, L. and Maeland, J. A., Acta Pathol. Microbiol. Scand. [B] 1979. 87: 51.
* Akerstrom, B., Lindqvist, A. and Lindahl, G., Mol. Immunol. 1990, in press.
Eur. J. Immunol. 1990. 20: 2241-2247 5 Bevanger, L. and Naess, A. I., Acta Pathol. Microbiol. Scand. [B] 1985. 93: 121. 6 Lindahl, G. and h e r s t r o m , B., Mol. Microbiol. 1989. 3: 239. 7 Lindahl, G., Mol. Gen. Genet. 1989. 216: 372. 8 Frithz, E., HedCn, L. 0. and Lindahl, G., Mol. Microbiol. 1989. 3: 1111. 9 Fischetti,V. A., Clin. Microbiol. Rev. 1989. 2: 285. 10 Lancefield, R. C., McCarty, M. and Everly,W. N., J. Exp. Med. 1975. 142: 165. 11 Cebra, J. and Robbins, J. B., J. Immunol. 1966. 97: 12. 12 Acosta Altamirano, G., Barranco-Acosta, C.,Van Roost, E. and Vaerman, J. P., Mol. Immunol. 1980. 17: 1525. 13 Vaerman, J. P. and Heremans, J. F., J. Immunol. 1972. 108: 637. 14 Vaerman, J. P., Studies on IgA Immunoglobulins in Man and Animals, Thesis, Sintal Press, Louvain 1970. 15 Vaerman, J. P., Arbuckle, J. B. and Heremans, J. F., lnt. Arch. Allergy Appl. lmmunol. 1970. 39: 323. 16 Butler, J. E., Peterson, L. and McGivern, P. L., Mol. Immunol. 1980. 17: 757. 17 Vaerman, J. P., Querinjean, P. and Heremans, J. F., Immunology 1971. 21: 443. 18 Sheldrake, R. F., Husband, A. J. ,Watson, D. L. and Cripps, A. W., J. Immunol. 1984. 132: 363. 19 Grubb, A., Mendez, E., Fernandez-Luna, J. L., Lopez, C., Mihaesco, E. and Vaerman, J.-P., J. Biol. Chem. 1986. 261: 14313.
IgA receptor from group B streptococci
2247
20 Devereux, J., Haeberli, I? and Smithies, D., Nucleic Acids Res. 1984. 12: 387. 21 Scatchard, G., Ann. NYAcad. Sci. 1949. 51: 660. 22 Myhre, E. B. and Kuusela, P., Infect. Immun. 1983. 40: 29. 23 Greenwood, F. C., Hunter,W. M. and Glover, J. S., Biochem. J. 1963. 89: 114. 24 Brady, L. J. and Boyle, M. D. P., Infect. Immun. 1989. 57: 1573. 25 Mestecky, J. and McGhee, J., Adv. Immunol. 1987. 40: 153. 26 Reis, K. J., Ayoub, E. M. and Boyle, M. D. I?, J. Immunol. 1984. 132: 3098. 27 Thangavelu, C. P. and Koshi, G., J. Clin. Microbiol. 1980.12: 1. 28 Person, K. M. S., Bjerre, B., Elfstrom, L., Polberger, S. and Forsgren, A., Eur. J. Clin. Microbiol. 1986. 5: 156. 29 Williams, A. F., Immunol. Toda 1987. 8: 298. 30 Bergglrd, B., Ekstrom, B. and kerstrom, B., Scand. J. Clin. Lab. Invest. 1980. 40 [Suppl. 1541: 63. 31 Logdberg, L. and Akerstrom, B., Scand. J. Immunol. 1981.13: 383. 32 SchalCn, C., Acta Pathol. Microbiol. Scand. [C] 1980. 88: 271. 33 Nardella, F. A., Schroder, A. K., Svensson, M., Sjoquist, J., Barber, C. V. and Christensen, P., J. Immunol. 1987. 138: 922. 34 Stone, G. C., Sjobring, U.,Bjorck,L., Sjoquist, J., Barber, C.V. and Nardella, F. A., J. Immunol. 1989. 143: 565.
L
Eur. J. Immunol. 1990. 20: 2241-2247
2241
Gunnar Lindahl+, Bo AkerstromO, Jean-Pierre Vaerman. and Lars Stenberg+
Characterization of an IgA receptor from group B streptococci: specificity for serum IgA*
Department of Medical Microbiology+ Department of Physiological Chemistryo, University of Lund, Lund and Experimental Medicine Unit, International Institute of Cellular and Molecular Pathology, Catholic University of Louvain., Brussels
Some strains of group B streptococci express a cell surface protein which binds 1gA.This report describes some properties of such an IgA receptor and compares it with a previously described IgA receptor from group A streptococci.The group B receptor was released in an almost pure form from bacteria incubated at elevated pH, and could be isolated by IgA-Sepharose affinity chromatography. The sequence of the N-terminal19 amino acid residues was unique.The receptor preferentially binds IgA of human origin, as shown in immunoblotting experiments with purified IgA from nine different species. The affinity constant of the purified receptor for serum IgA was determined to be 3.5 X lo8 M-l, but for secretory IgA it was too low to allow determination. This result indicates that secretory component and/or J chain interferes with the binding of IgA to this type of bacterial receptor, which may be one of the physiological functions of these polypeptides. A reduction in affinity was also observed for another complexed form of IgA, al-microglobulin-IgA. The group B receptor is antigenically unrelated to the IgA receptor from group A streptococci (protein Arp), but competitive inhibition experiments indicate that they bind to the same region in IgA. The implications of these findings, and the biological role of bacterial IgA receptors, are discussed.
1 Introduction Cell surface receptors that bind to the Fc region of IgA have been found in certain strains of both group A streptococci and group B streptococci, two common human pathogens [l,21. For both of these bacterial species, the IgA receptor (IgAR) has been purified and partially characterized. In group B streptococci, where the IgAR was first defined as an antigen called p [3,4], the purified receptor was shown to have a molecular weight of about 130000 [l]. There is evidence that this IgA-binding protein acts as a virulence factor [ 5 ] , but it is not yet known whether this is due to the capacity to bind IgA or to some other property of the protein. The IgAR in group A streptococci, called protein Arp, has been purified from two different streptococcal strains after expression of the cloned genes in Escherichia coli, and was shown t o have an apparent molecular weight of about 41 000 [6, 71. The biological function of protein Arp is not known, but the fact that this receptor shows extensive sequence similarity to streptococcal M proteins [8], which are well-known virulence factors [9], suggests that this receptor may also contribute to bacterial virulence. In this report,we describe a simple method for isolating the group B receptor, which has been immunochemically
characterized and compared with the previously isolated receptor from group A streptococci, protein Arp.The most remarkable property of the group B receptor is that the affinity of this receptor for secretory IgA is much lower than for serum IgA. This suggests that one function of the secretory component and/or J chain is to interfere with the binding of secretory IgA to bacterial IgAR.
2 Materials and methods 2.1 Bacterial strains
Two different series of group B streptococcal strains were used. These strains were isolated from clinical specimens sent to the Laboratory of Clinical Microbiology, Lund University Hospital. One series of 48 strains, the SB strains, obtained from Drs. G. Kronvall and C. Schonbeck, originate from a variety of specimens, including urine, vaginal secretions and pus. A second series, the 27 BS strains, represent all septicemia strains of group B streptococci isolated in our laboratory during a period of 7 years. The group B streptococcal strain A909 [lo] was obtained from Dr. C. Schalen. The group A streptococcal strains AW43 and AP4, both of which bind IgA, have been described [6, 71. The streptococcal strains were grown in Todd-Hewitt broth (Oxoid), Basingstone, GB).
[I 85981
*
This work was supported by grants from the Swedish Medical Research Council, the Medical Faculty of the University of Lund, King Gustaf V’s 80-year Foundation, Fysiografiska Sallskapet in Lund, the Foundations of Osterlund, Johansson, and Kock and by HighTech Receptor, Inc., Malmo, Sweden.The partial support of the Belgian State-Prime Minister’s Office Science Policy Programming is also acknowledged.
Correspondence:Gunnar Lindahl, Department of Medical Microbiology, University of Lund, Solvegatan 23, s-223 62 Lund, Sweden 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1990
2.2 Selection of IgA-binding strains The ability of bacterial strains to bind radiolabeled serum IgA was tested in a standard binding assay with 2 x 108 bacteria in 200 p1 of buffer, as described [6]. With a few exceptions, the strains could be unequivocally classified as binders or non-binders, showing > 40% or < 5% binding, respectively. Five out of 48 strains in the SB series and 5 out of 27 strains in the BS series were able to bind IgA. None of these strains could bind radiolabeled IgG or IgM. One of 0014-2980/90/1010-2241$3.SO + .25/0
2242
G. Lindahl, B. kerstrom, J.-PVaerman and L. Stenberg
the strains, SB35, was chosen for the biochemical work described in this report.This strain,which was isolated from urine, is of serotype Ic (typing kindly performed by Ms. I. Wilson, Department of Medical Microbiology, University of Lund) .
2.3 Proteins Ig were of human origin and polyclonal, unless otherwise stated. The origins of some preparations have been described [6]. Purified preparations of human monoclonal IgG proteins of all four subclasses were provided by Dr. F. Skvaril, WHOAUIS Immunoglobulin Subcommittee, Berne, Switzerland. The preparations of purified secretory IgA from the rabbit [ l l ] , rat [12], guinea pig [13], dog [14], pig [151, cow [161and horse [171, as well as sheep mesenteric lymph IgA [18] have been described. The preparation of al-microglobulin-IgA, purified as described [19], was the gift of Dr. Anders Grubb, Department of Clinical Chemistry, University of Lund. Two mouse IgA myeloma proteins,TEPC 15 and MOPC 315,were purchased from Litton Bionetics, Kensington, MD. Proteins Arp4 and Arp60 were purified as described [6, 71. 2.4 Purification of group B receptor The method used is based on the observation that the receptor is released from bacteria suspended in buffer of high pH (see Sect. 3.1). Strain SB35 was grown overnight in 20 1 Todd-Hewitt broth, collected by centrifugation, and washed twice with 50mM Tris, pH 7.0.The washed bacteria (about 30 g) were then suspended to 10% (v/v) in 50 mM Tris, pH 11.0, giving a final pH of 9.7. The suspension was incubated at 37°C for 4 h, centrifuged, and the SN was immediately dialyzed against PBS. The dialyzed preparation was frozen in portions, from which the receptor was later purified by affinity chromatography on IgA-Sepharose, as described [6]. Using this procedure, about 20 mg of receptor can be recovered from 30 g of washed bacteria.
Eur. J. Immunol. 1990. 20: 2241-2247
2.6 Amino acid sequence analysis Automated amino-terminal amino acid sequencing was performed with an Applied Biosystems (Foster City, CA) 470A gas-liquid solid-phase sequenator as described [19]. Amino acid sequence homology analysis was done with the program WORDSEARCH, which is included in the “Wisconsin” software [20]. NBRF Protein and Sequence Data Library and NBRF New Sequence Data Library were used as sources for published amino acid sequences. 2.7 Competitive inhibition experiments
The wells of microtiter plates (highly activated flat-bottom Titertek PVC Immunoassay microplates; Flow Laboratories, Aberdeen, GB) were coated overnight at 4°C with 100 p1 of polyclonal IgA, 10 pg/ml in PBS. Remaining binding sites were blocked by the addition of 100 pl of VBS (10 mM veronal buffer, 0.15 M NaCl, pH 7.4) containing 0.25% gelatin and 0.25% Tween 20, and incubation was continued at room temperature for 4 h and then at 4°C overnight. After washings with 0.12 M NaCl, 0.03 M sodium phosphate, 0.02% NaN3, 0.05% Tween 20; pH7.2 (PBSAT), approximately 5 ng (about 1.5 x 104 cpm) of radiolabeled IgAR and various amounts (10 ng-4 pg) of unlabeled IgAR receptor were added in 50 p1 PBSAT and the plates were incubated at room temperature for 4 h.The wells were then washed with PBSAT, separated by cutting, and counted in a gamma counter. 2.8 Determination of equilibrium constants
The equilibrium constants for the binding between group B receptor and various molecular forms of IgA were determined by a solid-phase RIA [6]. Briefly, the receptor, coated to microtiter plates, was incubated overnight with approximately 1ng 1251-labeledIgA and various amounts of unlabeled IgA of the same molecular form. The amount of bound radiolabeled IgA was determined, and the equilibrium constant was obtained after plotting the data and calculating according to Scatchard [21]. All incubations were done at room temperature.
2.5 Electrophoresis, immunoblotting and dot blots The electrophoretic methods used have been described [6]. Proteins separated by PAGE were transferred to membranes of polyvinylidene difluoride (Immobilon, Millipore Cop,Molsheim, France). The blocking, washing, incubations with 12%labeled protein probes and autoradiography were then performed as described [6]. Dot blots were also performed as described [6], except that Immobilon membranes were used. In the dot blot for which antisera were used for the analysis (see Fig. 8), the Immobilon membranes were first blocked in a buffer containing gelatin [6] and then incubated for 1 h at room temperature in rabbit antiserum, or preimmune serum, diluted 1000-foldin 15 ml of “VBS-gel” (10 mM veronal buffer, 0.15 M NaCl, 0.1% gelatin, pH 7.4). After washings (4 X 20 min) in 250 ml of 1M NaCl, 10 mM EDTA, 0.25% gelatin and 0.25% Tween 20, the filters were incubated for 3 h in 15 ml of “VBS-gel” containing radiolabeled protein G (2 x lo5 c p d m l) , followed by four washes as before [7]. All these procedures were carried out at room temperature and were followed b y autoradiography at - 70 “C.
2.9 Other methods
Radiolabeling was performed with carrier-free 1251(Amersham International, Amersham, GB), using a modified lactoperoxidase method [22] or chloramin T [23]. Binding assays with whole bacteria were performed as described [6]. Antibodies to purified group B receptor were obtained by S.C. immunization of a rabbit with 100 pg of the protein emulsified in CFA. Two immunizations were made with an interval of 5 weeks. The rabbit was bled before immunization and 1week after the last immunization.The antiserum to protein Arp4 has been described [7].
3 Results 3.1 Release of IgAR from strains of group B streptococci incubated at elevated pH
Previous work with strains of group B streptococci that bind IgA has shown that the IgAR is present in the growth
IgA receptor from group B streptococci
Eur. J. Immunol. 1990. 20: 2241-2247
medium, but this protein is size heterogenous [24]. Such a size-heterogeneous IgAR is also found in the growth medium of strain SB35, the strain studied here (Fig. 1). A method for isolating more size-homogeneous receptor was described by Russell-Jones et al. [l], who purified the protein from an extract prepared by boiling bacteria in 2% SDS. However, during work with strain SB35 we noted that the size-homogeneous receptor can also be easily obtained by incubating washed bacteria at elevated pH. The protein released by this treatment appears to be almost pure receptor. Minor amounts of contaminating proteins can be removed by affinity chromatography on IgA-Sepharose (Fig. 1). The molecular weight of the purified receptor is about 130O00, similar to the protein described by RussellJones et al. [l]. The release of receptor at elevated pH has not yet been studied systematically, but preliminary experiments indicated that optimal yields of receptor were obtained by suspension of washed bacteria in glycine-buffer, pH 11.0, and incubation at 37°C for 4 h. The final pH of such a suspension was 9.7. Suspension in buffers of lower pH (2-5) also caused release of receptor but this material was more size heterogeneous (data not shown). There was little release of receptor at 0 "C, but addition of sodium azide (0.05%) did not affect the yield at 37°C (data not shown). BLOT
STAIN 200
-
116-
E
20 10
5 2.5 1.25 0.63 0.3 1
Figure 3. Dot-blot assay on Immobilon, showing binding of radiolabeled group B receptor to different proteins. The indicated amounts of protein, diluted in PBS, were applied in 10O-pl aliquots. After blocking, the membrane was probed with '251-labeled receptor,washed and autoradiographed. All Ig preparations except IgM were monoclonal.
The release of IgAR at elevated pH is not a unique property of strain SB35, since similar results were obtained with 6 other IgA-binding strains of group B streptococci, including the commonly used type Ic strain A909. For all of these strains, the receptor was found to have a molecular weight of about 130000 (data not shown). On the other hand, no receptor was recovered from two different IgA-binding strains of group A streptococci treated in the same way.
3.2 N-terminal amino acid sequence of the group B receptor
97-
The N-terminal amino acid sequence of the group B receptor is shown in Fig. 2. Positions nos. 1,12and 16could not be identified, and are represented by the letter X in the figure. The sequence was compared with published sequences and was found to be unique.
66-
43-
1
2
3
4
5
6
7
8 9 1 0
Figure 1 . Release of IgAR from strain SB35 incubated at elevated pH. Bacteria were grown overnight in Todd-Hewitt broth, washed twice in 50 m M Tris, pH 7.0, and then resuspended to 10% (v/v) in buffers of different pH. After incubation for 4 h at 37"C, the suspensions were centrifuged and samples of the SN were separated on SDS-PAGE (T = 8%, C = 3.3%) under reducing conditions. Lanes 1 and 6: SN from the culture in Todd-Hewitt broth, concentrated &fold, lanes 2 and 7: bacteria suspended in 50 m M Tris, pH 7.0, lanes 3 and 8: bacteria suspended in 50 mM Tris, pH 9.0, lanes 4 and 9: bacteria suspended in 50 mM glycine, pH 11.0, lanes 5 and 10: group B receptor further purified by affinity chromatography. Lanes 1-5 were stained with Coomassie brilliant blue; lanes 6-10 were blotted onto Immobilon membranes and probed with radiolabeled IgA. Molecular mass (kDa) markers are on the left.
1
2243
5
10
15
Figure 2. N-terminal sequence of the group B receptor. Residues nos. 1, 12 and 16 could not be identified and are denoted by "X.
3.3 Binding specificity of the group B receptor Radiolabeled receptor was used in a dot-blot assay to study the specificity of the binding (Fig. 3). Membranes of polyvinylidene difluoride (Immobilon) were used for this analysis, since preliminary experiments unexpectedly showed very poor binding of the receptor to IgA immobilized on nitrocellulose membranes. The data in Fig. 3 show that the receptor binds to monoclonal IgA proteins of both subclasses. As expected, the receptor binds t o Fc fragments, but not to Fab fragments, of monoclonal 1gA.There was no binding to IgG, IgM, IgD or IgE, nor to albumin or fibrinogen. These binding properties of the purified receptor are similar to those of whole SB35 bacteria, which bind 19
Eur. J. Immunol. 1990. 20: 2241-2247
G. Lindahl, B. kerstrom, J.-PVaennan and L. Stenberg
2244
1 2 3 4
IgA, but not IgG, IgM, albumin or fibrinogen (data not shown). Our results also agree with the data of RusselJones et al. [l],who used an inhibition test to study the specificity of their receptor.
5 6 7 8 9101112
-
92
- 67
The purified group B receptor binds to the a H chain of IgA separated by SDS-PAGE (Fig. 4). Surprisingly, there was also a weaker reactivity with IgG y chains, but no binding to IgM p chains. Prolonged exposure of the X-ray film also showed a very weak binding to the L chains of these Ig (not shown). When whole plasma was analyzed in the same way, two bands appeared in the blotting experiment, corresponding t o a and y chains. These two bands were of similar strength, although the serum concentration of IgG is about sixfold higher than that of 1gA.These data indicate that the denaturing conditions of the SDS-PAGE cause exposure of y chain determinants that have affinity for the group B receptor. O n the other hand, the SDS-PAGE analysis did not show reactivity of the receptor with any other plasma protein, except for the a and y chains.
- 46 - 30
- 14
- 92
- 67 - 46
To study the species specificity of the receptor, preparations of purified IgA from nine different species were analyzed in a Western blot experiment, using radiolabeled receptor as the probe. Binding was observed only for a chains of human origin (Fig. 5). 3.4 Equilibrium constant of the binding between the
group B receptor and different molecular forms of IgA In man, IgA occurs in at least three different forms. Most of the IgA is secreted as secretory IgA which is usually a dimer containing a J chain and a chain of secretory component in addition to the two IgA monomers [25]. In contrast, most of the IgA present in serum is monomeric, but about 5% of this IgA occurs in a complex, where it is covalently linked to the serum protein al-microglobulin in a molar ratio of 1: 1 [19]. The affinity constant (equilibrium constant) of the group B receptor for these three forms of IgA was
- 30 - 14
Figure.5. Analysis of the species specificity of the group B receptor. Two micrograms of animal IgA (lanes 1-10), human secretory IgA (lane 11) or human serum IgA (lane 12) were separated by SDS-PAGEand stained with Coomassie brilliant blue (A). An identical gel was transferred to an Immobilon membrane and incubated with 1251-labeled group B receptor (B). The animal IgA were from rabbit (lanes 1-2), cow (lanes 3-4). sheep (lane 5), rat (lane 6), guinea pig (lane 7), horse (lane 8), pig (lane 9),and dog (lane 10). The animal IgA were secretory IgA, except that from sheep, which originates from lymph. The molecular masses of the standard proteins are given in kDa.
K,(M-1*10-8)
BLOT
STAIN
0
Serum IgA 3.5
0 al-m-lgA
66
-
.
45
-
m
29 24
--
20
-
0.14
C
a 0.1
i
1
2
3
4
5
6
7
%
Figure 4. Binding of group B receptor to separated a chains of IgA and y chains of IgG.Ten micrograms of serum IgA (lanes 1 and 5), 10 pg of IgG (lanes 2 and 6), 10 pg of IgM (lanes 3 and 7) and 50 p1 of human plasma diluted 100-fold were separated on SDS-PAGE (T = 10%;C = 3.3%).The gel wasstained with Coomasie brilliant blue (lanes 1-4), or transferred to an Immobilon membrane, which was incubated with 1251-labeledreceptor (lanes 5-8). Molecular mass markers, indicated on the left, are in kDa.
2
3 Bound lgA (nM)
4
5
Figure 6. Scatchard plot of the binding between group B receptor and two different molecular forms of IgA. The wells of microtiter plates were coated with group B receptor overnight, and after washing they were incubated for 5 h with approximately 1 ng '251-labeled serum IgA or a,-microglobulin-IgA, and varying amounts of unlabeled IgA of the same type. For each IgA concentration the ratio boundfree IgA was calculated and plotted against bound IgA. The equilibrium constant was then calculated from the slope of the linear diagram. Each point represents the
mean of triplicate experiments.
Eur. J. Immunol. 1990. 20: 2241-2247
IgA receptor from group B streptococci
2245
determined in a solid-phase RIA system with the receptor immobilized to the plastic walls of microtiter plates. The resulting Scatchard plots are shown in Fig. 6. The affinity constant for serum IgA was found to be 3.5 X lo8 M-l, but for the al-microglobulin-IgA complex it was 25-fold lower, 0.14 x lo8 M - ~ .The affinity constant for the binding to secretory IgA was too weak to permit Scatchard analysis.
interest to study the ability of these various forms of IgA to bind to whole bacteria, i.e. to receptor on the bacterial cell surface. The data in Fig. 7 show that the three forms of IgA differ both in the rate with which they bind to the bacteria and in the maximal amount of binding achieved. These results are in good agreement with the equilibrium constants described above, since serum IgA and secretory IgA show the highest and lowest rate of binding, respectively.
3.5 Binding of different molecular forms of IgA to whole bacteria
3.6 The group B receptor and protein Arp are immunologically unrelated but bind to the same region in IgA
The finding that purified group B receptor has different affinity for the three molecular forms of IgA made it of
60
t
0
serum IgA
0 Orl-m-IgA A secretory IgA
,.
0
The IgAR isolated from group A and group B streptococci were compared in two ways. First, the two types of IgAR were compared immunologically, using rabbit antisera raised against purified preparations of the receptors. As shown in Fig. 8, there was no cross-reactivity between the two types of receptor. O n the other hand, two variants of protein Arp showed strong cross-reactivity. As expected, no binding was observed when preimmune rabbit serum was used in the test (data not shown).
.
a 1
2
-A-
3
,
loot
Time (h)
Figure 7. Binding of different molecular forms of IgA t o strain SB35 as a function of time.Washed bacteria (2 X 108 cells in 200 p1 of buffer) were mixed with 2-5 ng (about 104cpm) of radiolabeled protein and incubated at room temperature for the time indicated. Ice-cold buffer (2 ml) was then added, the samples were centrifuged and radioactivity in the pellet was measured.
/ /
/group
A' /
'
B rec.
p' 8
2
100
1000
Unlabeled IgA receptor added (ng)
1
antiArp4
0.5 0.25 0.13
Figure 9. Inhibition of binding of lzI-labeled protein Arp4 to immobilized human serum IgA by unlabeled protein Arp4 or group B receptor. See Sect 2.7 for details.
0.06
2 1
antigroup B roc.
0.5 0.25 0.13
I
0.06
Figure 8. Dot-blot analysis of cross-reactivity between streptococcal IgAR. The indicated amounts of protein, diluted in PBS, were applied to Immobilon membranes in 100-pl aliquots. After blocking, the membranes were incubated with rabbit antiserum of the type indicated. The filter was then washed and bound rabbit IgG detected with 1251-labeledprotein G.
A second type of comparison between the two streptococcal IgAR was designed to study whether they bind to the same region in the Fc part of IgA. This was studied in a competitive inhibition experiment, as previously used in comparisons between different bacterial IgG receptors [26].The data in Fig. 9 show that the binding of radiolabeled protein Arp4 to IgA was inhibited by both receptors. Similarly, the binding of radiolabeled group B receptor to IgA was also inhibited by both receptors (data not shown). These data suggest that the two IgAR bind to the same region in the Fc part of IgA. This inhibition was not unspecific, since addition of a control protein (human haptoglobin) was without effect.
2246
G. Lindahl, B. Akerstrom, J.-RVaennan and L. Stenberg
4 Discussion Two different streptococcal IgAR have now been characterized in considerable detail: protein Arp from group A streptococci and the receptor from group B streptococci studied here. These two bacterial IgAR are structurally different, as shown by the fact that they are antigenically unrelated and are of very different size. The mode of cell wall attachment also appears to be different for the two receptors, since the group B receptor, but not protein Arp, is released in almost pure form from bacteria incubated at elevated pH. The mechanism of this release is not known, but it forms the basis of a simple method for purifying the group B receptor. Similar release of receptor has also been observed at low pH, and it seems possible that it could favor the establishment of an infection in tissues where the pH is not neutral. A remarkable property of the purified group B receptor is that the affinity constant for the binding to secretory IgA is much lower than for serum IgA, and could not be determined. In agreement with this result, binding experiments with whole bacteria showed very poor binding of secretory IgA (Fig. 7). Poor binding of secretory IgA to whole bacteria has also been observed by Brady and Boyle [24], who suggested that the solubilized receptor has different properties and is able to bind secretory IgA. This conclusion was based on the observation that radiolabeled secretory IgA can be used to detect the solubilized receptor in a Western blot. However, this apparent binding of secretory IgA can be readily explained by antibody activity against the receptor in the Ig preparation used [27]. Such antibody activity has indeed been observed by us in polyclonal Ig preparations (data not shown). The presence of such antibody activity in polyclonal preparations of human IgA is not surprising, since group B streptococci are part of the normal flora in man [28]. Since secretory IgA is the predominant antibody on mucous membranes, it was surprising that the group B receptor binds poorly to this molecular form of IgA. However, seen from the point of view of the host, this finding suggests that one function of secretory component and/or J chain is to interfere with the binding of secretory IgA to specific receptors. On the other hand, the function and mechanism of action of the group B receptor in bacterial infection remain unclear. It even seems possible that the in vivo function of this bacterial virulence factor is to bind to some protein other than IgA, e.g. a cell adhesion molecule, some of which are known to be structurally related to Ig [29]. An indication that the IgAR might also be able to bind to other members of the Ig superfamily comes from a Western blot experiment (Fig. 4),which shows that under denaturing conditions the receptor binds not only to a chains of IgA but also to y chains of IgG. In addition to secretory IgA, human IgA also occurs in another complexed form, al-microglobulin-IgA [19]. In this complex, which represents about 5% of the IgA found in serum, one molecule of IgA is covalently linked to one molecule of al-microglobulin, an immunomodulatory serum protein with a molecular weight of 26000 [30, 311. The equilibrium constant for the binding of the group B receptor to this complex was found to be 25-fold lower than for serum IgA. The group B receptor, therefore, has
Eur. J. Immunol. 1990. 20: 2241-2247
reduced affinity also for this complexed form of IgA, which shows that the bound al-microglobulin molecule can drastically affect the ability of IgA to bind to a specific receptor.This finding may be of significance with regard to the physiological function of al-microglobulin. Several lines of evidence now indicate that the IgAR isolated from group A and group B streptococci have similar binding specificity. Both receptors bind in the Fc region and they recognize both subclasses of IgA [ l , 6, 7, 321. The competitive binding experiments reported here (Fig. 9) also indicate that the two receptors bind to the same or overlapping site(s) within the Fc region of IgA. Similar results have been obtained with three structurally different bacterial IgGR, all of which appear to bind in the same region of the Fcy fragment [26, 33, 341. Another similarity between the two streptococcal IgAR is that both receptors apparently are specific for IgA of human origin (Fig. 5) [1]*. Finally, both receptors show higher affinity for serum IgA than for the two complexed forms of IgA, although the differences in affinity are less dramatic for protein Arp than for the group B receptor*. A difference between the two IgA receptors is that protein Arp, but not the group B receptor, also binds weakly to the Fcy region [6, 71. However, this binding to IgG apparently takes place at a separate site on the protein Arp molecule [6]*. In conclusion, the results reported here show that the IgAR in group A and group B streptococci are structurally unrelated but have similar binding properties. This implies that the two receptors are the result of convergent evolution, which has favored the appearance of these different IgA receptors in streptococci. Our results also suggest that one of the functions of secretory component and/or J chain is to interfere with the binding of secretory IgA to specific receptors. Regardless of the biological functions of the different IgAR, these proteins are potentially useful as immunological tools and may also be used as model systems for the analysis of Ig binding to specific receptors. We are grateful to Dr. Anders Crubb for kindly donating the preparation of a!-microglobulin-lgA and to Drs. J. E. Butler, W C. Hanley, A . J. Husband and F? C. Montgomery for generous gifts of purified animal IgA. The technical assistance of Ms. Gunilla Lovhult and Ms. Ann-Charlotte Johansson is gratefully acknowledged. The amino acid sequence analysis was kindly performed by Ms. Ingrid Dahlqvist. Monoclonal human IgG proteins of different subclasses were provided by Dr. E Skvaril, WHOIIUIS Immunoglobulin Subcommittee, Berne, Switzerland.
Received June 5. 1990.
5 References 1 Russell-Jones, G. J., Gotschlich, E. C. and Blake, M. S., J. Exp. Med. 1984. 160: 1467. 2 Christensen, P. and Oxelius,V. A., Acta Pathol. Microbiol. Scand. [C] 1975. 83: 184. 3 Wilkinson, H. W. and Eagon, R. G., Infect. Immun. 1971. 4: 596. 4 Bevanger, L. and Maeland, J. A., Acta Pathol. Microbiol. Scand. [B] 1979. 87: 51.
* Akerstrom, B., Lindqvist, A. and Lindahl, G., Mol. Immunol. 1990, in press.
Eur. J. Immunol. 1990. 20: 2241-2247 5 Bevanger, L. and Naess, A. I., Acta Pathol. Microbiol. Scand. [B] 1985. 93: 121. 6 Lindahl, G. and h e r s t r o m , B., Mol. Microbiol. 1989. 3: 239. 7 Lindahl, G., Mol. Gen. Genet. 1989. 216: 372. 8 Frithz, E., HedCn, L. 0. and Lindahl, G., Mol. Microbiol. 1989. 3: 1111. 9 Fischetti,V. A., Clin. Microbiol. Rev. 1989. 2: 285. 10 Lancefield, R. C., McCarty, M. and Everly,W. N., J. Exp. Med. 1975. 142: 165. 11 Cebra, J. and Robbins, J. B., J. Immunol. 1966. 97: 12. 12 Acosta Altamirano, G., Barranco-Acosta, C.,Van Roost, E. and Vaerman, J. P., Mol. Immunol. 1980. 17: 1525. 13 Vaerman, J. P. and Heremans, J. F., J. Immunol. 1972. 108: 637. 14 Vaerman, J. P., Studies on IgA Immunoglobulins in Man and Animals, Thesis, Sintal Press, Louvain 1970. 15 Vaerman, J. P., Arbuckle, J. B. and Heremans, J. F., lnt. Arch. Allergy Appl. lmmunol. 1970. 39: 323. 16 Butler, J. E., Peterson, L. and McGivern, P. L., Mol. Immunol. 1980. 17: 757. 17 Vaerman, J. P., Querinjean, P. and Heremans, J. F., Immunology 1971. 21: 443. 18 Sheldrake, R. F., Husband, A. J. ,Watson, D. L. and Cripps, A. W., J. Immunol. 1984. 132: 363. 19 Grubb, A., Mendez, E., Fernandez-Luna, J. L., Lopez, C., Mihaesco, E. and Vaerman, J.-P., J. Biol. Chem. 1986. 261: 14313.
IgA receptor from group B streptococci
2247
20 Devereux, J., Haeberli, I? and Smithies, D., Nucleic Acids Res. 1984. 12: 387. 21 Scatchard, G., Ann. NYAcad. Sci. 1949. 51: 660. 22 Myhre, E. B. and Kuusela, P., Infect. Immun. 1983. 40: 29. 23 Greenwood, F. C., Hunter,W. M. and Glover, J. S., Biochem. J. 1963. 89: 114. 24 Brady, L. J. and Boyle, M. D. P., Infect. Immun. 1989. 57: 1573. 25 Mestecky, J. and McGhee, J., Adv. Immunol. 1987. 40: 153. 26 Reis, K. J., Ayoub, E. M. and Boyle, M. D. I?, J. Immunol. 1984. 132: 3098. 27 Thangavelu, C. P. and Koshi, G., J. Clin. Microbiol. 1980.12: 1. 28 Person, K. M. S., Bjerre, B., Elfstrom, L., Polberger, S. and Forsgren, A., Eur. J. Clin. Microbiol. 1986. 5: 156. 29 Williams, A. F., Immunol. Toda 1987. 8: 298. 30 Bergglrd, B., Ekstrom, B. and kerstrom, B., Scand. J. Clin. Lab. Invest. 1980. 40 [Suppl. 1541: 63. 31 Logdberg, L. and Akerstrom, B., Scand. J. Immunol. 1981.13: 383. 32 SchalCn, C., Acta Pathol. Microbiol. Scand. [C] 1980. 88: 271. 33 Nardella, F. A., Schroder, A. K., Svensson, M., Sjoquist, J., Barber, C. V. and Christensen, P., J. Immunol. 1987. 138: 922. 34 Stone, G. C., Sjobring, U.,Bjorck,L., Sjoquist, J., Barber, C.V. and Nardella, F. A., J. Immunol. 1989. 143: 565.
L
