- Email: [email protected]
Characterisation of mouse monoclonal antibodies for pneumolysin: fine epitope mapping and V gene usage
Immunology Letters 88 (2003) 227 /239 www.elsevier.com/locate/
Characterisation of mouse monoclonal antibodies for pneumolysin: fine epitope mapping and V gene usage ´ lvarez a,1, Marı´a del Mar Garcı´a-Sua´rez b,1, Francisco J. Me´ndez a, Beatriz Sua´rez-A Juan R. de los Toyos b,* a
´ rea de Microbiologı´a, Facultad de Medicina, Universidad de Oviedo, c/Julia´n Claverı´a s/n, 33006 Oviedo, Spain A ´ rea de Inmunologı´a, Facultad de Medicina, Universidad de Oviedo, c/Julia´n Claverı´a s/n, 33006 Oviedo, Spain A
b
Received 23 January 2003; received in revised form 9 April 2003; accepted 14 April 2003
Abstract Pneumolysin (PLY) is a cholesterol-dependent cytolysin (CDC) produced by Streptococcus pneumoniae , the main cause of community-acquired pneumonia. We have applied a set of diverse molecular methodologies (PCR-derived PLY peptides, biopanning of a library of phage-displayed random nonapeptides, indirect ELISA and competition tests with soluble peptides) to achieve concordant complementary observations in order to obtain a fine epitope mapping of three mouse monoclonal antibodies (PLY-4, PLY-7 and PLY-8) for PLY. PLY-4 seems to recognise a conformation-dependent epitope with a core reactivity involving R232. The epitopes recognised by PLY-7 and PLY-8 are within the sequences (401)GQDLTAH(407) and (450)KRTISIWGT(458), respectively. PLY-7 also recognises suilysin (SLY), in which the homologous reactive amino acid stretch is (429)GVNLTSH(435). In a homology model of PLY with the crystal structure of perfringolysin O (PFO), R232 is part of a well-exposed contorted loop on the edge of the concave and convex faces of domain 1. The sequences reactive with PLY-7 and PLY-8 would conform one of the loops at the bottom of domain 4 and a b strand of one of the two b sheets of this domain, respectively. Western blot analyses carried out with anti-PLY rabbit IgG and polyclonal mouse serum identified stretches comprising residues 40 /98, 199 /248, 352 /414 and 415 / 471 of PLY as immunogenic and antigenic; altogether with their recognition by the monoclonal antibodies herein considered, these results stress the immunological significance of domains 1 and 4 of the PLY molecule. PLY-4, PLY-7 and PLY-8 share the same Vk chain; this chain and that of the PLY-5 monoclonal antibody are essentially in germline configuration, whereas the VH regions of these monoclonals come from diverse gene segments and are mutated. # 2003 Elsevier Science B.V. All rights reserved. Keywords: Monoclonal antibody; Pneumolysin; Epitope mapping; V gene
1. Introduction Cholesterol-dependent cytolysins (CDCs) are produced by certain species of five different genera of Gram-positive bacteria (Archanobacterium , Bacillus , Clostridium , Listeria and Streptococcus ). They make up a group of pore-forming toxins on eukaryotic cell membranes. Once they have interacted with cholesterol, they oligomerise and insert into the membrane giving
* Corresponding author. Tel.: /34-985-10-3660; fax: /34-985-103658. E-mail address: [email protected] (J.R. de los Toyos). 1 These two authors contributed equally to this work.
rise to the opening of holes and leading to its permeabilisation [1]. CDCs share a high degree of sequence similarity (between 40 and 70%). CDCs are supposed to have evolved from a single progenitor gene and to adopt similar 3D foldings. The crystal structure of a soluble monomeric form of perfringolysin O (PFO) has been the first of a CDC to be solved [2]. The monomer is structured in four discontinuous domains which conform an elongated mushroom-shaped molecule. Domains 1, 2 and 4 make up the cylindrical stem, domain 3 being packed laterally against domain 2. Domain 1 is supposed to be involved in oligomerisation. Its convex face has been predicted to form extensive contacts with amino acid residues of the concave face of the neigh-
0165-2478/03/$ - see front matter # 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0165-2478(03)00081-6
228
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A
bouring monomer. Domain 4, which comprises the last 109 C-terminal residues including the well conserved Trp-rich stretch or cysteine motif, adopts a compact bsandwich folding, with three tip loops predominantly hydrophobic; it would be the membrane binding domain [3]. A model of pore assembly for the CDCs has been proposed according to which bound monomers form a large prepore complex on the membrane before insertion [4]. Membrane binding and insertion is accompanied by major structural changes. In addition to domain 4, domain 3 has also been shown to penetrate the membrane; it could be involved in oligomer formation as well. We generated [5] a panel of mouse monoclonal antibodies (MAbs) to PLY, the CDC produced by Streptococcus pneumoniae . Three of these (PLY-4, PLY-5 and PLY-7) have anti-hemolytic properties. PLY-5 and PLY-7 have been shown to prevent the binding of PLY to the erythrocyte membrane, whereas PLY-4 seems to interfere with the oligomerisation process; PLY-8 is a non-neutralising MAb. A rough epitope mapping of these MAbs was made from the recognition of truncated versions of PLY. The epitope recognised by PLY-5 was identified in more detail later [6]. We have now applied a set of diverse mapping methodologies to provide a more accurate description of the epitopes identified by the PLY-4, PLY-7 and PLY-8 MAbs; the present findings are discussed in accordance with the structural features assigned to domains of the PLY molecule based on a homology model with the crystal structure of PFO. The V gene usage of these MAbs has also been analysed.
2. Material and methods 2.1. Bacterial strains S. pneumoniae GB05 strain was grown on agar blood plates at 37 8C. Streptococcus suis NCTC 10234 was grown at 37 8C in Todd /Hewitt broth (DIFCO), additioned with 0.1% b-mercaptoethanol [7]. A filtered 48 h-supernatant was subjected to Western-blot and hemolysis analyses. 2.2. Polyclonal sera and MAbs The generation of the PLY-specific mouse hybridomas herein considered has been previously described [5]. The three secrete IgG1, kappa MAbs. The antibody 1.3C9.91 is also a mouse IgG1, kappa monoclonal generated against human pepsinogen C [8]. MAbs were purified by thiophilic adsorption chromatography (AFFI-T gel columns, Kem-En-Tec, Denmark) from serum- and protein-free culture supernatants (HybriMax, Sigma).
Anti-PLY polyclonal mouse serum refers to a pool made from the sera of the mice which were used to generate the above mentioned hybridomas. The generation of the anti-PLY rabbit IgG preparation has already been described [9 /11]. 2.3. Immunoblot analyses of SLY-containing supernatants Twenty microlitres of culture supernatant were assayed. Primary antibodies were probed at 1 mg; peroxidase conjugates to F(ab?)2 fragments of sheep IgG antimouse IgG (Jackson ImmunoResearch Laboratories) were used at 1:1000. Specific protein bands were detected applying the BM chemiluminescence blotting substrate (POD) (Roche Diagnostics). 2.4. Construction and Western blot analyses of recombinant full-length PLY and PCR-derived PLY fragments Table 1 shows the primers used to generate diverse recombinant products. Fig. 1 shows a graphical representation of their relationships to the putative domains of the PLY molecule. Recombinant PLY was PCR amplified from a colony of S. pneumoniae GB05. Amplification products were directly cloned into pGEM-T vector (Promega) and sequenced. Recombinant PLY (R) and recombinant fragments (A /S) were amplified from this gene, and cloned into Nde I and BamH I endonuclease restriction sites of pET-11a (Novagen). This expression plasmid was used to transform Escherichia coli DH10B (GibcoBRL). Plasmids with correct insert were used to transform E. coli BL21 (DE3) pLysS (Stratagene) for producing the corresponding recombinant proteins. Overnight cultures were diluted into 2 /TY broth supplemented with ampicillin (100 mg/ml) and chloramphenicol (25 mg/ml). Bacteria were grown at 37 8C until an optical density at 600 nm (OD600)/0.8 and then induced with 1 mM isopropyl-b-thiogalactopyranoside (IPTG). After incubation for 30 min, rifampicin (240 mg/ ml) was added to the culture and incubated for an additional 90 min at 37 8C. Bacteria were harvested by centrifugation, and boiled for 5 min in electrophoresis sample buffer. Whole-cell lysates were subjected to 16.5%T/3%C Tricine- or 12%- SDS-polyacrylamide gel electrophoresis. The proteins were transferred to Immun-Blot PVDF Membrane (0.4 or 0.2 mm) (Bio Rad) with a Mini Trans-Blot Electrophoretic Transfer Cell (Bio Rad) and the membranes were then blocked overnight at 4 8C in 1% blocking reagent (Roche Diagnostics). Membranes were incubated with antiPLY mouse polyclonal serum (1:500), anti-PLY rabbit IgG (1:50) or MAbs (10 mg). Mouse and rabbit polyclonal sera were preincubated with lyophilised
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A
229
Table 1 Synthetic oligonucleotide primers used for recombinant-PLY fragments Fragment Amino acid residues
5? Primer
3? Primer
Ra A B C D E F G H I J K N O P Q S
5? GAGGTAG/CATATG/GCAAATAAAGC 5? GAGGTAG/CATATG/GCAAATAAAGC 5? GAGGGTAAT/CATATG/CCCGATGAG 5? GATCGTGCT/CATATG/ACTTATAG 5? ATGTCCCAGCT/CATATG/CAGTATG 5? TTTTAAG/CATATG/TATTATACAGTCAG 5? GGGCG/CATATG/TATCTCAAGTTG 5? GGCGACCCA/CATATG/GGTGCCCGAG 5? GTACAGAC/CATATG/GAGACTAAG 5?CACTAGTATT/CATATG/AAAGGGAATG 5? GAGGTAG/CATATG/GCAAATAAAGC? 5? GATCGTGCT/CATATG/ACTTATAG 5? GGGCG/CATATG/TATCTCAAGTTG 5? ATGTCCCAGCT/CATATG/CAGTATG 5? GAGGTAG/CATATG/GCAAATAAAGC 5? ATGTCCCAGCT/CATATG/CAGTATG 5? GAGGTAG/CATATG/GCAAATAAAGC
5? 5? 5? 5? 5? 5? 5? 5? 5? 5? 5? 5? 5? 5? 5? 5? 5?
1 /471 1 /49 40 /98 97 /151 148 /207 199 /248 247 /305 299 /353 352 /414 415 /471 1 /151 97 /248 247 /414 148 /305 1 /305 148 /471 1 /359
TCTC/GGATCC/TAGTCATTTTCTACCTTA GTCGACA/GGATCC/GCTACTTTCTTTC CCAGGC/GATCC/ATACTCTAAGTCATC CATGCTG/GGATCC/GTTATTTATTCATAC ATCTCCT/GGATCC/TAAACAGCGTC CGT/GGATCC/CAACTAGAGATAGAC CTCTACC/GGATCC/ACCTAGCCTG CTGTAAGCT/GGATCC/TAAGTCTCAAC GATTACGAA/GGATCC/CTTATAAAGG TCTC/GGATCC/TAGTCATTTTCTACCTTA CATGCTG/GGATCC/GTTATTTATTCATAC CGT/GGATCC/CAACTAGAGATAGAC GATTACGAA/GGATCC/CTTATAAAGG CTCTACC/GGATCC/ACCTAGCCTG CTCTACC/GGATCC/ACCTAGCCTG TCTC/GGATCC/TAGTCATTTTCTACCTTA ACGC/GGATCC/CTATCTGTAAGCTGTAACCTTAG
Restriction sites for Nde I and BamH I are underlined. Stop codons are in bold. a Full-length PLY.
whole-cell lysates of non-transformed E. coli , at 2.5 mg/ ml, for 2 h at 37 8C and overnight at 4 8C. Similar preparations of normal mouse IgG (Sigma), preimmune normal rabbit serum or the 1.3C3.91 MAb were
parallely assayed as primary antibody negative controls. After several washing steps, blots were incubated with a 1:1000 dilution of anti-mouse IgG-HRP-conjugated (Sigma) or with a 1:5000 dilution of anti-rabbit IgG-
Fig. 1. Schematic representation of S. pneumoniae PLY and recombinant-PLY fragments. The first and last amino acid residues of PLY included in each recombinant protein are indicated.
230
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A
HRP conjugated (Sigma). Reactive proteins were detected with BM Chemiluminescence blotting substrate (POD) (Roche Diagnostics). 2.5. Phage display library The pVIII-9aa library was kindly provided by Dr Alessandra Luzzago, IRBM, Italy. This library is made up of random sequences of nine amino acids inserted into the amino terminus of the coat protein VIII (pVIII) of bacteriophage f1 [12], and contains 9.4 /107 independent clones. 2.6. Growth and purification of phages Amplification of the library was essentially done as described previously [13]. Briefly, over 1011 CFU of the pVIII-9aa library in E. coli strain XL1-blue (Stratagene) were inoculated in 500 ml of 2 /TY broth containing 15 mg/ml of tetracycline (TC, Sigma), 100 mg/ml of ampicillin (AMP, Sigma), and 1% glucose (2 /TY /TC / AMP /GLU) and incubated at 37 8C until the OD600 was approximately 0.5. This culture was then infected with M13-K07 helper phage (Amersham Pharmacia Biotech) (multiplicity of infection /50) by standing incubation for 30 min at 37 8C and afterwards with shaking for another 30 min period. Subsequently, 0.1 mM (final concentration) IPTG was added and incubated with strong shaking for 5 h at 37 8C. Bacterial cells were removed by centrifugation (twice for 30 min at 8600 /g ), and the phages present in the supernatant were precipitated twice with polyethylene glycol /NaCl (20% PEG-8000, 2.5 M NaCl). Finally, the phages were resuspended in Tris /buffered saline (TBS), 0.02% NaN3 to a final concentration of about 1.5 /1014 CFU/ml. 2.7. Biopanning Three rounds of biopanning were applied for each MAb. Briefly, Maxisorp immunotubes (5 ml; Nunc, Denmark) were incubated with 4 ml of specific-MAb (100 mg/ml in 10 mM phosphate-buffered saline [PBS], pH 7.3) at 4 8C overnight. Subsequently, the immunotubes were blocked with blocking buffer (3% bovine serum albumin [BSA] in PBS, 0.05% Tween-20) for 2 h at 37 8C, and washed three-times with PBS. The phage library (between 1 and 3 /1012 CFU in blocking buffer) was added, and the tubes were incubated for 30 min at room temperature (RT), rotated continuously on a turntable, followed by a standing incubation for 90 min at RT. The tubes were then washed ten-times with PBS, 0.05% Tween-20 and ten-times with PBS to remove the unbound phages. Bound phages were eluted with 1 ml of 0.1 M Glycine /HCl, pH 2.2, rotating for 10 min at RT. Eluted phages were neutralised with 0.5 ml of 1 M Tris /HCl, pH 7.4, and used to infect E. coli strain XL1-
blue. To determine the number of eluted phages, infected E. coli cells were plated by serial dilution on 2 /TY /AMP /IPTG-X /Gal plates. After the first round of specific biopanning, there was a ‘negative’ selection step in which phages non-specifically bound to the MAb used for ‘positive’ selection were eliminated using immunotubes coated with an irrelevant MAb, 1.3C3.91. The remaining phages were amplified and further enriched by two additional rounds of ‘positive’ selection performed as described above. For the last round of biopanning, the specific MAb was used at 50 mg/ml. After this, phages were individually isolated from infected E. coli cells to test their reactivity in ELISA. 2.8. Phage ELISA The reactivity of MAbs with monophages was assayed by means of indirect ELISA tests. Flat-bottom 96-well polystyrene enzyme immunoassay plates (Costar) were coated with serial dilutions of PEGprecipitated monophages (100 ml/well in PBS, pH 7.3) overnight at 4 8C. The plates were then blocked with blocking buffer for 2 h at 37 8C. After being washed, the wells were successively incubated with 100 ml of antipneumolysin (PLY) MAb (10 mg/ml in 1% BSA /PBS / Tween-20); anti-mouse IgG-HRP conjugated (Sigma) (dilution 1:1000); and 50 mM phosphate-citrate buffer, pH 5.2, containing 10 mg of O-phenylenediamine dihydrochloride (Sigma) and 0.05% H2O2. Plates were finally incubated at 37 8C for 30 min in the dark. Colour development was stopped by adding 100 ml/well of 2 M H2SO4, and read at 492 nm. Plates were washed threetimes with PBS, 0.05% Tween-20 between incubations. All determinations were made in triplicate. The 1.3C3.91 MAb was used as an irrelevant negative-control antibody. 2.9. Sequencing of phage insert DNA The monophages which developed the highest absorbances were selected for sequencing. Single stranded DNA was prepared by phenol extraction following the methods already described [14] and sequenced after applying the sequencing primer M13 (/40). 2.10. Peptide ELISA tests Synthetic peptides, /95% pure, were supplied by Neosystem, France. These tests were carried out as the phage ELISA ones, the antigens being peptides, suspended in 10 mM PBS, pH 7.3, at 1 mg/100 ml per well. Blocking buffer was 1% BSA in PBS, 0.05% Tween-20.
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A
Anti-PLY MAbs were assayed at 10 mg/ml; anti-PLY mouse polyclonal serum at 1:500; and normal mouse IgG (Sigma) at 5 mg/ml, in blocking buffer. 2.11. ELISA competition tests with soluble peptides For this, ELISA plates were coated with 50 ng of recombinant PLY in 100 ml of 10 mM PBS, pH 7.3. The reactivity of each MAb with PLY was titered in ELISA tests carried out as described above for the phage ELISA. A minimum amount of each Mab, which rendered a reliable absorbance signal was chosen as a positive control. For competition tests, MAbs were mixed with different amounts of the synthetic peptide in blocking buffer and preincubated for 2 h at 37 8C and overnight at 4 8C. The mixtures were then added to the PLY-coated wells and incubated for 20 min at 37 8C. The rest of ELISA continued as described above. Anti-mouse IgG-HRP conjugated (Sigma) was used at a dilution 1:5000. 2.12. Hemolysis neutralisation assays These were performed as previously described [5,6]. Titration of toxin preparations was done immediately before neutralisation tests. Both PLY and SLY were assayed at two hemolytic units (HU). The neutralisation capacity of each MAb was also titered. For neutralisation tests, MAbs were assayed at their minimum neutralising amount. For neutralisation-blocking assays, MAbs were preincubated with peptides as above. 2.13. V gene sequencing 1/3/107 MAb-producing hybridoma cells were used for each preparation of mRNA using the QuickPrep Micro mRNA Purification kit (Amersham Pharmacia Biotech). cDNA was PCR amplified with V regionspecific primers [15]. PCR products were then cloned into the pGEM-T vector (Promega) and submitted for sequencing using a Thermo Sequenase Cycle Sequencing kit (Amersham Pharmacia Biotech). The sequences were analysed by means of BLASTN searches of GenBank databases (http://www.ncbi.nih.gov). Germlines were those defined and recorded in the IgBLAST database of GenBank. Amino acid residue assignments within V regions and definition of subgroups were done applying ad hoc entries in the Kabat Database of Proteins of Immunological Interest (http://www.immuno.bme.nwu.edu). 2.14. Alignment analyses and structural models CDCs sequences were obtained from the Swiss-prot database. Sequence alignments were done using the gap
231
and PILEUP programs of the Genetics Computer Group sequencing package [16]. Structural models of PLY were generated with the SWISS-MODEL program (http://www.expasy.ch/swissmod/SWISS-MODEL.html). Images were produced using the SWISS-PDBVIEWER (http://www.expasy.ch/ spdbv/mainpage.html) and WEBLAB VIEWERLITE (http://www.accelrys.com) programs. 2.15. Sequence accession numbers The complete sequences of the V regions herein characterised are deposited in GenBank (accession numbers AY194240, AY194241, AF509572, AF509573, AF509574, AF509575, AF509576 and AF509577).
3. Results 3.1. Reactivity of MAbs with SLY in immunoblot assays and analyses of their neutralising properties After a first description of the anti-PLY monoclonal antibodies developed by us, suilysin (SLY), secreted by S. suis , was characterised as the most similar CDC to PLY [7]. We were, therefore, interested in analysing the recognition of SLY by our anti-PLY MAbs. The PLY-5 antibody has already been reported to recognise all the CDCs tested and to block their binding to eukaryotic membranes [6]. The PLY-7 MAb was also probed on SLY [10]. By applying Western-blot chemiluminescence enhancement methodology to a supernatant of the SLYproducing S. suis NCTC 10234 strain, we have also confirmed that the PLY-7 antibody does recognise SLY and neutralise its hemolytic activity whereas PLY-4 and PLY-8 do not. 3.2. Western blot analysis of the PCR-derived PLY fragments The results of this analysis are shown in Tables 2 and 3. Out of the nine different, approximately contiguous and 50 amino acid residues long, peptides examined (Table 2), the anti-PLY mouse polyclonal serum tested was reactive with five of such (B, E, G, H and I) whereas the rabbit IgG preparation recognised only four (B, E, H and I). To be reactive with fragment I, the corresponding whole-cell lysate had to be boiled for at least 15 min, suggesting that this peptide needed to be sufficiently unfolded to deploy its antigenic sites. PLY7 and PLY-8 unambiguously identified peptides H and I, respectively; surprisingly, PLY-4 and PLY-5 did not recognise any of the fragments tested.
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A
232
Table 2 Reactivity of polyclonal sera and MAbs with recombinant-PLY fragments demonstrated by enhanced chemiluminescence immunoblotting analyses Fragment
a
R (1 /471) A (1 /49) B (40 /98) C (97 /151) D (148 /207) E (199 /248) F (247 /305) G (299 /353) H (352 /414) I (415 /471) a b
Anti-PLY rabbit IgG
b
/ / / / / / / / / /
Anti-PLY polyclonal mouse serum
/ / / / / / / / / /
MAbs PLY-4
PLY-5
PLY-7
PLY-8
/ / / / / / / / / /
/ / / / / / / / / /
/ / / / / / / / / /
/ / / / / / / / / /
Full-length PLY. Immunoreactive, /; non-reactive: /.
Table 3 Further enhanced-chemiluminescence-immunoblotting analyses of the reactivity of PLY-4 with longer recombinant-PLY fragments
Table 4 Deduced amino acid sequence of the inserts of monophages rescued after biopanning with the PLY-4, PLY-7 and PLY-8 MAbs
Fragment
Reactivitya
Insert sequence
Frequencya
ELISAb
J (1 /151) K (97 /248) N (247 /414) O (148 /305) P (1 /305) Q (148 /471) S (1 /359)
/ / / // /// /// ////
PLY-4 RPPRYLANQ RPPRYLSQLc RLPRYMQEV NRPRWVKEA RPPRFQLSS RPPRYMVQS RPPKWVSAQc RKPRYLTQS
3 3 2 1 1 1 1 1
1.35f 0.95f 1.4f 1.7 1.6 1.5 0.9 0.4
PLY-7 GATLGPYPP GGQTFTDREd KGIYPLSYE
16 1 1
1.2f 1.4 0.7
PLY-8 KRAVQKWAPe
12
1.3f
a
Non-reactive, /; very weak, /; weak, //; strong, ///; very strong, ////.
To further investigate the reactivity of PLY-4, some longer PCR fragments were studied. Table 3 summarises the results obtained. Fragment S corresponds to the union of domains 1, 2 and 3. From the data shown in Table 3, it seems that PLY-4 requires fragments at least 150 amino acids long to be reactive with and that it recognises a core structure located somewhere between positions 151(E) and 247(Y) of PLY.
a
3.3. Biopanning of a random nonapeptide library, and rescue and identification of monophages
Number of times each monophage was independently recovered. Reactivity on monophage-coated indirect ELISA tests. c Residues present in the (232)RPLVYISSV(240) amino acid stretches of PLY are in bold. d Residues present in the (400)NGQDLTAHF(408) amino acid stretches of PLY are in bold. e Residues present in the (450)KRTISIWGT(458) amino acid stretches of PLY are in bold. f Mean of the absorbance developed by the monophage isolates.
After each round of selection, the number of eluted phages was determined and the percentage of phage that bound to the selecting MAb was defined as the ratio between this number and the corresponding input. This percentage increased significantly after the second round, indicating that the biopanning had been successful and that antibody-specific phages were selected. Monophages were isolated from each round. Some of them were randomly tested in ELISA, exhibiting diverse reactivities. Monophages which reacted more strongly with the specific MAb than with the irrelevant MAb,
1.3C3.91, and which developed an OD above that of the wild phage (with no insert), were taken as specific. Table 4 shows the deduced amino acid sequence of the inserts from rescued specific monophages, the number of times they were independently recovered and their ELISA reactivity. Twenty-seven monophages, isolated after the third round of biopanning with PLY-4, were randomly tested in ELISA. Of these, 22 were considered as positive; from these, 13, which developed the highest ELISA readings,
b
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A
were selected for sequencing. All the insert sequences recovered contained at least a pair of contiguous RP residues. When these amino acid sequences were aligned to the PLY sequence, RPPRFQLSS aligned to the 144(V)/ 152(K) amino acid stretch. It seems unlikely that this stretch encompasses the PLY-4 epitope, as it only overlaps in two residues with the 151(E) /247(Y) amino acid stretch and because, if this were the case, fragment J (1 /151) could have been recognised by this MAb as well. However, two other sequences, RPPRYLSQL and RPPKWVSAQ, coincided in their alignment to (232)RPLVYISSV(240) within the 151(E) /247(Y) amino acid stretch whereas the rest of the sequences aligned outside of it. Therefore, from these observations, the PLY-4 epitope could be located within the (232)RPLVYISSV(240) amino acid stretch of PLY. In the case of PLY-7, 29 monophages out of 32 from the third round and 11 out of 12 from the second round were considered as specific. From these, 12 from the third round and six from the second were selected for sequencing. Sixteen monophage clones presented the same amino acid sequence whereas the remaining two were unrelated. These sequences did not reveal a common motif, and strikingly the most reactive monophage was isolated only once after the second round. The recovered inserts were then aligned to the PLY sequence. The KGIYPLSYE and GATLGPYPP inserts aligned to amino acid stretches included as part of the PCR-derived peptide A and F, respectively, whereas GGQTFTDRE aligned to the (400)NGQDLTAHF(408) stretch, which is within the PCR-derived peptide H; the homologous SLY sequence is (428)NGVNLTSHW(436). With respect to the PLY-8 MAb, 20 individual phages from the third round were isolated. Twelve of the 16 monophages considered to be specific were randomly selected for sequencing. All presented the same amino acid sequence insert, KRAVQKWAP; this insert aligned to the (450)KRTISIWGT(458) stretch of PLY, included in fragment I. 3.4. ELISA and neutralisation-blocking tests These were performed to confirm the nature of the sequences recognised by PLY-4, PLY-7 and PLY-8; the PLY-5 MAb was used as control. By homology with the resolved crystallographic structure of PFO, Glu231 and Arg232 of PLY are part of a well surface-exposed contorted loop on the edge of the concave and convex faces of domain 1; their side chains bulge prominently on the protein (Figs. 4 and 5). Pro233 is covered by the (13)MNYD(16) stretch. The (234)LVYISSV(240) amino acid stretch is mostly buried under another loop, at the centre of which is D338, on
233
the convex face of domain 1. Arg232 has been predicted to form a salt bridge with Asp338 [17]. Therefore, in characterising the PLY-4 epitope, we also directed our attention to the amino acid residues preceding Arg232/Pro233. The following peptides were assayed: (227)GISAERPL(234) and (226)RGISAERPLVYISSV(240); in order to give a loop conformation to the (227)GISAERPL(234) peptide, another peptide was synthesised including a Cys at each end of the (227)GISAERPL(234) sequence. None of these peptides was reactive with PLY-4 in indirect ELISA tests or showed any interfering/blocking activity even when assayed at 100 mg. We attributed these results to the inability of the assayed peptides to properly recreate the antigenic features of the PLY-4 epitope. We then considered whether a longer peptide could recreate these antigenic features. Therefore, another peptide, 24 amino acid long, half of fragment E, from (216)D to (239)S positions of PLY, was probed as antigen in indirect ELISA tests. The anti-PLY mouse polyclonal serum and PLY-4 did recognise this peptide but not PLY-5; otherwise, 50 mg of this peptide did not block the neutralising capacity of PLY-4. A synthetic peptide, named as p21PLY, comprising residues from (393)T to (413)P of PLY and, therefore, encompassing the (400)NGQDLTAHF(408) stretch, was similarly probed. This peptide was recognised by PLY-7 and the anti-PLY mouse polyclonal serum but not by PLY-5. Amino acid residues (401)GQDLTAH(407) of PLY would conform a wellexposed b strand of one of the loops at the bottom of domain 4, between its two b-sheets, whereas amino acid residues N(400) and F(408) seem to be drawn up and already part of the neighbouring b strands. Thus, we also decided to test a synthetic peptide, called p7PLY, which corresponds to amino acid residues (401)GQDLTAH(407) of PLY, and a p7SLY, corresponding to the (429)GVNLTSH(435) amino acid stretch of SLY. The p7PLY and the p7SLY peptides were both recognised in indirect ELISA tests and blocked the neutralising ability of the PLY-7 MAb (Table 5). A peptide p5, corresponding to amino acid residues (425)IRECTGLAWEWWRTV(439), which includes the cysteine motif and, therefore, the PLY-5 epitope [6], behaved similarly in relation to the PLY-5 MAb. In the opposite way, neither the p7PLY nor the p7SLY peptide was recognised by or affected the neutralising characteristics of the PLY-5 MAb. Upon titration and for ELISA-competition tests, PLY-5 was assayed at 1 mg/well and PLY-8 at 20 ng/ well. A p8 peptide, corresponding to amino acid residues (450)KRTISIWGT(458) of PLY, was able to compete with the recognition of native PLY by the PLY-8 MAb in a dose-dependent manner (Fig. 2), as did the p5 peptide in relation to PLY-5. As a control, the p7PLY
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A
234
Table 5 Neutralisation-blocking assays of PLY-7 and PLY-5 PLY (2 HU)
PLY-7 (1 mg)
PLY-5 (2 mg)
p7PLY (mg)
p7SLY (mg)
p5 (mg)
Lysis
/a / / / / / / / /
/ / / / / / / / /
/ / / / / / / / /
/ / / 25 / / 25 / /
/ / / / / 25 / 25 /
/ / / / 31 / / / 31
YES NO NO YES NO YES NO NO YES
a
/, added; /, not added.
The VH regions are made of diverse VH and JH gene segment combinations; the two VL regions identified are also made of two different VL and Jk gene segment combinations. The VH regions of the PLY-8 and PLY-5 MAbs show a number of amino acid replacements threetimes higher than the number found in their VL partners; in relation to the PLY-4 and PLY-7 MAbs, their VH and VL regions show a similar number of amino acid substitutions.
Fig. 2. ELISA competition tests with the p8 and p5 peptides.
peptide, assayed at 100 mg, did not block such recognition by the PLY-8 or the PLY-5 MAbs. 3.5. V gene usage The four MAbs have used unique VH genes as their closest germlines are not repeated (Fig. 3, panel a). These VH genes belong to three different subgroups (Table 6). PLY-4 is made of a 7183 gene of the VH5 family. The PLY-7 and PLY-8 MAbs have utilised J558 genes of the VH1 gene family; the PLY-5 MAb has utilised a 36 /60 VH gene of the VH3 family. Overall, they display between 83 and 95% of nucleotide identities with the most homologous VH germline gene found. Nucleotide substitutions have led to amino acid replacements in all four CDRH1s and CDRH2s. The CDRH3s of PLY-4 and PLY-7 are of the same length (11 amino acid residues); that of PLY-8 is eight amino acid residues long. The CDRH3 of PLY-5 comprises only three amino acids (Glu/Ala /Ser). With respect to the k chains, it is surprising that PLY4, PLY-7 and PLY-8 share the V region (Fig. 3, panel b), with the same pattern of amino acid replacements. This V region shows a 96% nucleotide identity (Table 6) with its closest germline 19/20. The PLY-5 Vk region has been made from a different Vk gene (cw9) although also within the V subgroup; in this case, there is a nucleotide identity of 97% with this cw9 germline.
4. Discussion Our Western-blot observations from different contiguous PCR-derived PLY peptides, 50 amino acid residues long, and from mouse and rabbit PLY-specific polyclonal sera and the herein studied MAbs have allowed us to identify stretches of amino acid residues 40 /98, 199 /248, 299/353, 352/414 and 415/471 as immunogenic and antigenic. According to the structural data and folding models of PLY which parallel those described for PFO [2,17], all these stretches include well accessible and surface exposed sites. Most of fragment B (positions 40 /98) folds as an ample loop on the concave face of domain 1 (Fig. 4). Fragment E (amino acid residues 199/248) is mostly within domain 1. Fragment G encompasses two distinct stretches: amino acid residues 299 /318 are part of domain 3, whereas those running from positions 319 to 353 are part of a contorted loop on the convex face of domain 1. Fragment H (positions 352/414) corresponds to the first four b strands of the b-sandwich of domain 4; and fragment I (positions 415/471), to the last four. Altogether, these observations highlight the immunological significance of domains 1 and 4. A former screening with human and rabbit hyperimmune sera and partially overlapping synthetic peptides [18] had identified two common immunoreactive sites at the aminoterminal end of the PLY molecule, which comprise residues from (29)E to (48)R and from (131)L to (152)K positions. Our fragment B partially overlaps with the
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A Fig. 3. Amino acid sequences of heavy (a) and light-chain (b) variable regions of four anti-PLY MAbs, PLY-4, PLY-7, PLY-8 and PLY-5. The sequences are aligned to their closest available germline sequence in the IgBLAST database of GenBank. Dashes indicate identities. Sequence stretches used in primers are underlined. The FR/CDR assignments were according to the Kabat database of proteins of immunological interest.
235
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A
236
Table 6 Characteristics of the V regions of the four anti-PLY MAbs PLY-4
PLY-7
PLY-8
PLY-5
Gene usage VH familya Subgroupb VH germ-line segmentc D germ-line segment JH germ-line segment VL family Subgroup VL germ-line segment JL germ-line segment
7183 (VH5) IIID VH 69.1 (95%) DSP 2.7/8 (100%) JH4 (91%) Vk 19/28 V Vk19 /20 (96%) Jk2 (100%)
J-558 (VH1) IIA VMU-3.2 (91%) / JH3 (97%) Vk 19/28 V Vk19 /20 (96%) Jk2 (97%)
J-558 (VH1) IIA V23 (83%) DFL16.1/2 (100%) JH2 (97%) Vk 19/28 V Vk19 /20 (96%) Jk2 (100%)
36 /60 (VH3) IA VH 36 /60 (86%) / JH1 (91%) Vk 9/10 V cw9 (97%) Jk1 (100%)
CDR3 length (amino acids ) Heavy chain Light chain
11 8
11 8
7 8
3 9
Number of mutations (nucleotides ) VH segment 13 VL segment 10
24 10
48 11
40 8
Number of mutations (amino acids ) VH segment 9 VL segment 8
12 8
22 8
18 6
a b c
Family and germlines were assigned according to the IgBLAST database of GenBank. Subgroup is according to the Kabat database of proteins of immunological interest. Nucleotide homology with the closest germline.
Fig. 4. A homology model of the PLY molecule which shows the spatial distribution of the amino acid stretches, corresponding to PCRderived fragments, identified as immunogenic/antigenic in the present study. Fragments B, E and G are in grey. Domain 4 comprises fragments H and I. The location of the E231, R232, and D338 residues is indicated.
first of these immunoreactive sites, but otherwise no similar occurrences are seen. These observational discrepancies could be attributable to technical limitations inherent to the different epitope mapping methologies applied. PLY and SLY share 50% of amino acid residue identities. We have confirmed that PLY-7 and PLY-5
recognise linear epitopes on SLY and also neutralise its hemolytic activity, while PLY-4 and PLY-8 do not have these properties. From a pepscan analysis covering the whole amino acid sequence of PLY, the epitope recognised by PLY-5 was unambiguously assigned to the well-conserved undecapeptide motif present in all CDCs [6]; this motif would conform a well-exposed elongated loop at the bottom of domain 4. The PLY-5 epitope would be within the sequence WEWWRT (Fig. 6). We have found that PLY-5 does not recognise any of the herein 50-residue-long PCR-derived PLY peptides. This suggests that epitope recognition by this MAb is greatly dependent on the characteristics of the peptide, which encompasses it and on the context in which this peptide is expressed and/or tested. Proper analysis of the antigenicity of peptides and epitope identification requires concordant results from diverse immunoassay formats and epitope mapping methodologies [19]. In this context, we have to stress that this has been done for the present epitope mapping work and that especially, from the Western-blot analyses, we have obtained unambiguous results as no misleading cross-reactions among PCR-derived fragments were observed. Because fragments derived from proteinase K nicked PLY were not reactive with PLY-4, it was suggested that the cleavage site would be in the centre of the PLY-4 epitope, located around residue 142 [5]. Based on our present observations from full-length recombinant PLY and PCR-derived PLY fragments, it seems that the
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A
PLY-4 epitope, located in domain 1, is greatly conformation-dependent, and that encompasses a core structure between the 151(E) and the 247(Y) amino acid residues. According to the biopanning and indirect ELISA tests, this core reactivity could reside around the Arg at position 232. Based on PFO, domain 1 of PLY would be made of intermingled non-contiguous amino acid stretches, conforming a highly interdependent ordered structure [1,17]. The 199(Y) to the 248(L) amino acid stretch, which corresponds to the whole length of fragment E, is sandwiched between fragments B and G (Figs. 4 and 5) and the first N-terminal 20 amino acid residues. The PLY-4 epitope could be made of and/or determined by residues belonging to these three fragments. Residual conformational features of the PLY-4 epitope could explain its recognition in Western blot assays; conversely, any significant structural alteration of the PLY molecule would be greatly detrimental for its antigenicity. Proper reconstitution of the epitope recognised by the PLY-4 antibody seems to require rather long amino acid stretches. The elucidation of the ultimate characteristics of this epitope requires other experimental approaches. In PFO, it has been shown that insertion into the membrane is preceded by oligomerisation which implies the formation of a large prepore [4]. In the oligomer and pore of PLY, the Arg232 would be positioned at the outer surface of the ring. As the PLY-4 epitope seems to include Arg232, a residue which is predicted to be directly involved in the oligomerisation step [17], the hemolysis-neutralising capacity of the PLY-4 MAb would be exerted through its interference with the oligomerisation/formation of prepore events.
Fig. 5. A CPK representation of a lateral view of the packing of fragments A, B, E and G in which the E231, R232 and D338 residues are highlighted.
237
A number of observations from diverse truncates and mutants of CDCs [20 /23] suggests that some residues within domains 1 and 3 are key factors for oligomerisation and hemolysis by intervening in their correct folding and three dimensional structure; this would apply in particular to positions of PLY from 199(Y) to 248(L) (fragment E) which could be pivotal for the conformation of the PLY-4 epitope. Cross-inhibition ELISA tests between pairs of MAbs had shown that the binding of PLY-7 on PLY had positive modulation effects upon the subsequent binding of PLY-5; the reverse phenomenum was not observed. Analyses by means of flow cytometry of epitope recognition led to the conclusion that the neutralising capacity of PLY-5 and PLY-7 is due to the blocking of the binding of the toxin to the eukaryotic membrane and that their epitopes become embedded in the membrane upon binding of the toxin. Similar findings were made later on in relation to SLY [10]. The experimental findings herein described lead us to conclude that PLY-7 recognises an epitope in the (401)GQDLTAH(407) sequence of PLY and in the (429)GVNLTSH(435) of SLY. By sequence alignment and structural homology with PFO, the (401)GQDLTAH(407) stretch would conform one of the three parallel loops at the bottom of domain 4 (Fig. 6). The WEWWRT motif conforms one of the external loops, recognised by PLY-5, the other being the one conformed by this amino acid stretch recognised by the PLY-7 MAb. It is worth stressing that, while this WEWWRT loop and the intermediate one correspond to well conserved sequence stretches in the CDCs, the external loop recognised by the PLY-7 MAb shows a high sequence variability among the CDCs; this would explain the restricted reactivity of this MAb to PLY and SLY. The close spatial vicinity of the PLY-7 epitope to the PLY-5 within domain 4 could explain the above mentioned immunomodulatory and epitope recognition observations on membranes. A wealth of observations suggests that the cysteine motif and in particular the Trp-rich stretch is primarily involved in the interaction with cholesterol and in membrane insertion [1]; these events are accompanied by several conformational changes in the CDCs [20,21,24 /26]. Together with cholesterol, other membrane molecules could mediate toxin binding as well. To our knowledge, the only mutation so far generated within the PLY-7 epitope-containing homologous stretch of any CDC is that recently reported [27] in which leucine 461 of native LLO was substituted by threonine present in PFO as well as in PLY (position 405), SLY (position 433) and other CDCs. This conversion resulted in an increased hemolytic activity of LLO at a neutral pH but the mechanism(s) by which this effect is exerted remain(s) to be precisely explained.
238
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A
Fig. 6. (A) A homology model of domain 4 of PLY; the location of the sequences comprising the PLY-7, PLY-8 and PLY-5 epitopes is highlighted. (B, C and D) Structural details of the PLY-8, PLY-7 and PLY-5 sequences, respectively.
The blocking of the PLY-7 MAb on toxin binding could have, at least, two different explanations. First and based on our ELISA-derived modulation observations, the binding of PLY-7 to domain 4 could induce a conformational change that would lead the PLY-5 epitope-containing Trp-rich loop to lose affinity for cholesterol. Second, that binding could interpose a steric hindrance to the interaction between this toxin site and cholesterol and other membrane components. PLY-8 is a PLY non-neutralising MAb. From its reactivity with truncated forms of PLY, the epitope recognised by this MAb appeared to be in the last six Cterminal amino acid residues of PLY [5]. Nevertheless, our present findings indicate that its epitope lies between amino acid residues (450)KRTISIWGT(458) of PLY which would conform a b strand in the b sheet opposite to the WEWWRT loop (Fig. 6). No substitution mutants within the PLY-8 epitope-containing amino acid stretch have been generated, but, apart from its contribution to the folding of domain 4, this toxin site does not seem to be primarily involved with membrane interaction and/or oligomerisation events; therefore, the binding of the PLY-8 MAb to its epitope would not interfere with these processes and this would explain why this MAb has no neutralising capacity. The herein assigned location is in agreement with our previous
observations according to which the PLY-8 epitope would become embedded in the membrane after toxin binding. It is surprising that PLY-4, PLY-7 and PLY-8 share the same Vk region; in relation to its closest germline, this Vk region shows a nucleotide identity of 96%. The Vk region of the PLY-5 MAb has a nucleotide homology of 97% with its closest germline. All this suggests that these Vk chains are essentially germline encoded and that accumulation/selection of amino acid replacement within these Vk domains is not important for PLY binding. As the epitopes recognised by these four MAbs are different, it is obvious that the epitope specificity of these MAbs is dictated by the VH counterparts; a preponderant role of the heavy chain in the determination of the specificity of the secondary antibody response to protein antigens has already been reported [28,29]. We rescued two hybridomas from two different mice that happened to synthesise the same anti-hemolytic MAb (PLY-5/PLY-6). PLY-6 is another anti-PLY IgG1, kappa MAb from which it was inferred that it recognised the same epitope as PLY-5 [5]. The sequences of the VH and Vk genes of PLY-6 have been previously reported [30]. In agreement with those former observations, we have now found that PLY-5 and PLY-6 have
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A
identical VH and Vk regions. Therefore, it seems that the PLY-5 epitope strongly selects for a particular VH/VL pairing. The four MAbs herein considered express VH genes very likely deriving from different germlines and belonging to different subgroups; furthermore, these VH genes are recombined with different JH segments, adding support to the assumption that the four MAbs are independently derived. This diverse use of VH genes is a common feature of the humoral immune response to complex protein antigens [31]. The VH regions of the PLY-7, PLY-8 and PLY-5 MAbs have a nucleotide homology of 91, 83 and 86%, respectively, with its closest germline; since these MAbs recognise epitopes within domain 4 which is involved in toxin binding, it seems that the efficient antibody recognition of this toxin domain is associated with mutated VH regions. This may not be the case of the PLY-4 epitope, mediating oligomerisation events within domain 1, as the VH region of the PLY-4 MAb has a 95% of nucleotide homology and a low number of amino acid replacement in relation to its closest germline.
Acknowledgements The authors are indebted to Dr Alessandra Luzzago, Istituto di Richerche di Biologia Molecolare P Angeletti (IRBM), Pomezia, Italy, for providing the pVIII nonapeptide library. We also wish to thank David H. Wallace for the English language correction of the ´ lvarez manuscript. This work and Beatriz Sua´rez-A were supported by Proyecto PB-MED01-08, FICYT, Principado de Asturias, Spain; Marı´a del Mar Garcı´aSua´rez was financed in part by Proyecto 1FD97-1185, Ministerio de Ciencia y Tecnologı´a, Spain.
References [1] R.K. Tweten, M.W. Parker, A.E. Johnson, Curr. Top. Microbiol. Immunol. 257 (2001) 15 /33. [2] J. Rossjohn, S.C. Feil, W.J. McKinstry, R.K. Tweten, M.W. Parker, Cell 89 (1997) 685 /692. [3] R. Ramachandran, A.P. Heuck, R.K. Tweten, A.E. Johnson, Nat. Struct. Biol. 9 (2002) 823 /827. [4] L.A. Shepard, O. Shatursky, A.E. Johnson, R.K. Tweten, Biochemistry 39 (2000) 10284 /10293. [5] J.R. de los Toyos, F.J. Me´ndez, J.F. Aparicio, F. Va´zquez, M.M. Garcı´a Sua´rez, A. Fleites, C. Hardisson, P.J. Morgan, P.W. Andrew, T.J. Mitchell, Infect. Immun. 64 (1996) 480 /484.
239
[6] T. Jacobs, M.D. Cima-Cabal, A. Darji, F.J. Me´ndez, F. Va´zquez, A.A.C. Jacobs, Y. Shimada, Y. Ohno-Iwashita, S. Weiss, J.R. de los Toyos, FEBS Lett. 459 (1999) 463 /466. [7] R.P.A.M. Segers, T. Kenter, L.A.M. de Haan, A.A.C. Jacobs, FEMS Microbiol. Lett. 167 (1998) 255 /261. [8] E.P. Diamandis, S. Nadkarni, B. Bhaumik, A. Abdelrahman, D.N. Melegos, G. Borchert, M.H. Black, M. Alonso, A. Salas, J.R. de los Toyos, A. Sampedro, C. Lo´pez-Otı´n, Clin. Chem. 43 (1997) 1365 /1371. [9] M.D. Cima-Cabal, F. Va´zquez, J.R. de los Toyos, F.J. Me´ndez, J. Clin. Microbiol. 37 (1999) 1964 /1966. [10] M.D. Cima-Cabal, Ph.D. thesis, Universidad de Oviedo, Spain, 1999. [11] M.D. Cima-Cabal, F.J. Me´ndez, F. Va´zquez, M.M. Garcı´aSua´rez, J.R. de los Toyos, J. Immunoassay Immunochem. 22 (2001) 99 /112. [12] F. Felici, L. Castagnoli, A. Musacchio, R. Jappelli, G. Cesareni, J. Mol. Biol. 222 (1991) 301 /310. [13] A. Luzzago, F. Felici, in: S. Cabilly (Ed.), Methods in Molecular Biology: Combinatorial Peptide Library Protocols: Combinatorial Peptide Library Protocols, vol. 87, Humana Press, 1998, pp. 155 /164. [14] J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, 1989. [15] R. Orlandi, D.H. Gussow, P.T. Jones, G. Winter, Proc. Natl. Acad. Sci. USA 86 (1989) 3833 /3837. [16] J. Devereux, P. Haeberli, O. Smithies, Nucleic Acids Res. 12 (1984) 387 /395. [17] J. Rossjohn, R.J.C. Gilbert, D. Crane, P.J. Morgan, T.J. Mitchell, A.J. Rowe, P.W. Andrew, J.C. Paton, R.K. Tweten, M.W. Parker, J. Mol. Biol. 284 (1998) 449 /461. [18] P. Salo, A. Na¨rva¨nen, M. Leinonen, Infect. Immun. 61 (1993) 2822 /2826. [19] M.H.V. Van Regenmortel, Biologicals 29 (2001) 209 /213. [20] M. Palmer, P. Saweljew, I. Vulicevic, A. Valeva, M. Kehoe, S. Bhakdi, J. Biol. Chem. 271 (1996) 26664 /26667. [21] E.M.A. Ghani, S. Weis, I. Walev, M. Kehoe, S. Bhakdi, M. Palmer, Biochemistry 38 (1999) 15204 /15211. [22] I. Yamamoto, H. Kimoto, Y. Taketo, A. Taketo, Biosci. Biotechnol. Biochem. 65 (2001) 2682 /2689. [23] H. Baba, I. Kawamura, C. Kohda, T. Nomura, Y. Ito, T. Kimoto, I. Watanabe, S. Ichiyama, M. Mitsuyama, Biochem. Biophys. Res. Commun. 281 (2001) 37 /44. [24] L.A. Shepard, A.P. Heuck, B.D. Hamman, J. Rossjohn, M.W. Parker, K.R. Ryan, A.E. Johnson, R.K. Tweten, Biochemistry 37 (1998) 14563 /14574. [25] M. Palmer, I. Vulicevic, P. Saweljew, A. Valeva, M. Kehoe, S. Bhakdi, Biochemistry 37 (1998) 2378 /2383. [26] O. Shatursky, A.P. Heuck, L.A. Shepard, J. Rossjohn, M.W. Parker, A.E. Johnson, R.K. Tweten, Cell 99 (1999) 293 /299. [27] I.J. Glomski, M.M. Gedde, A.W. Tsang, J.A. Swanson, D.A. Portnoy, J. Cell. Biol. 156 (2002) 1029 /1038. [28] A.J. Caton, S.E. Stark, J. Kavaler, L.M. Staudt, D. Schwartz, W. Gerhard, J. Immunol. 147 (1991) 1675 /1686. [29] R. Tuteja, Immunology 97 (1999) 1 /8. [30] M. Garcia, G. Strachan, A.J.R. Porter, W.J. Harris, FEMS Immunol. Med. Microbiol. 22 (1998) 225 /231. [31] S. Clarke, R. Rickert, M.K. Wloch, L. Staudt, W. Gerhard, M. Weigert, J. Immunol. 145 (1990) 2286 /2296.
Characterisation of mouse monoclonal antibodies for pneumolysin: fine epitope mapping and V gene usage ´ lvarez a,1, Marı´a del Mar Garcı´a-Sua´rez b,1, Francisco J. Me´ndez a, Beatriz Sua´rez-A Juan R. de los Toyos b,* a
´ rea de Microbiologı´a, Facultad de Medicina, Universidad de Oviedo, c/Julia´n Claverı´a s/n, 33006 Oviedo, Spain A ´ rea de Inmunologı´a, Facultad de Medicina, Universidad de Oviedo, c/Julia´n Claverı´a s/n, 33006 Oviedo, Spain A
b
Received 23 January 2003; received in revised form 9 April 2003; accepted 14 April 2003
Abstract Pneumolysin (PLY) is a cholesterol-dependent cytolysin (CDC) produced by Streptococcus pneumoniae , the main cause of community-acquired pneumonia. We have applied a set of diverse molecular methodologies (PCR-derived PLY peptides, biopanning of a library of phage-displayed random nonapeptides, indirect ELISA and competition tests with soluble peptides) to achieve concordant complementary observations in order to obtain a fine epitope mapping of three mouse monoclonal antibodies (PLY-4, PLY-7 and PLY-8) for PLY. PLY-4 seems to recognise a conformation-dependent epitope with a core reactivity involving R232. The epitopes recognised by PLY-7 and PLY-8 are within the sequences (401)GQDLTAH(407) and (450)KRTISIWGT(458), respectively. PLY-7 also recognises suilysin (SLY), in which the homologous reactive amino acid stretch is (429)GVNLTSH(435). In a homology model of PLY with the crystal structure of perfringolysin O (PFO), R232 is part of a well-exposed contorted loop on the edge of the concave and convex faces of domain 1. The sequences reactive with PLY-7 and PLY-8 would conform one of the loops at the bottom of domain 4 and a b strand of one of the two b sheets of this domain, respectively. Western blot analyses carried out with anti-PLY rabbit IgG and polyclonal mouse serum identified stretches comprising residues 40 /98, 199 /248, 352 /414 and 415 / 471 of PLY as immunogenic and antigenic; altogether with their recognition by the monoclonal antibodies herein considered, these results stress the immunological significance of domains 1 and 4 of the PLY molecule. PLY-4, PLY-7 and PLY-8 share the same Vk chain; this chain and that of the PLY-5 monoclonal antibody are essentially in germline configuration, whereas the VH regions of these monoclonals come from diverse gene segments and are mutated. # 2003 Elsevier Science B.V. All rights reserved. Keywords: Monoclonal antibody; Pneumolysin; Epitope mapping; V gene
1. Introduction Cholesterol-dependent cytolysins (CDCs) are produced by certain species of five different genera of Gram-positive bacteria (Archanobacterium , Bacillus , Clostridium , Listeria and Streptococcus ). They make up a group of pore-forming toxins on eukaryotic cell membranes. Once they have interacted with cholesterol, they oligomerise and insert into the membrane giving
* Corresponding author. Tel.: /34-985-10-3660; fax: /34-985-103658. E-mail address: [email protected] (J.R. de los Toyos). 1 These two authors contributed equally to this work.
rise to the opening of holes and leading to its permeabilisation [1]. CDCs share a high degree of sequence similarity (between 40 and 70%). CDCs are supposed to have evolved from a single progenitor gene and to adopt similar 3D foldings. The crystal structure of a soluble monomeric form of perfringolysin O (PFO) has been the first of a CDC to be solved [2]. The monomer is structured in four discontinuous domains which conform an elongated mushroom-shaped molecule. Domains 1, 2 and 4 make up the cylindrical stem, domain 3 being packed laterally against domain 2. Domain 1 is supposed to be involved in oligomerisation. Its convex face has been predicted to form extensive contacts with amino acid residues of the concave face of the neigh-
0165-2478/03/$ - see front matter # 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0165-2478(03)00081-6
228
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A
bouring monomer. Domain 4, which comprises the last 109 C-terminal residues including the well conserved Trp-rich stretch or cysteine motif, adopts a compact bsandwich folding, with three tip loops predominantly hydrophobic; it would be the membrane binding domain [3]. A model of pore assembly for the CDCs has been proposed according to which bound monomers form a large prepore complex on the membrane before insertion [4]. Membrane binding and insertion is accompanied by major structural changes. In addition to domain 4, domain 3 has also been shown to penetrate the membrane; it could be involved in oligomer formation as well. We generated [5] a panel of mouse monoclonal antibodies (MAbs) to PLY, the CDC produced by Streptococcus pneumoniae . Three of these (PLY-4, PLY-5 and PLY-7) have anti-hemolytic properties. PLY-5 and PLY-7 have been shown to prevent the binding of PLY to the erythrocyte membrane, whereas PLY-4 seems to interfere with the oligomerisation process; PLY-8 is a non-neutralising MAb. A rough epitope mapping of these MAbs was made from the recognition of truncated versions of PLY. The epitope recognised by PLY-5 was identified in more detail later [6]. We have now applied a set of diverse mapping methodologies to provide a more accurate description of the epitopes identified by the PLY-4, PLY-7 and PLY-8 MAbs; the present findings are discussed in accordance with the structural features assigned to domains of the PLY molecule based on a homology model with the crystal structure of PFO. The V gene usage of these MAbs has also been analysed.
2. Material and methods 2.1. Bacterial strains S. pneumoniae GB05 strain was grown on agar blood plates at 37 8C. Streptococcus suis NCTC 10234 was grown at 37 8C in Todd /Hewitt broth (DIFCO), additioned with 0.1% b-mercaptoethanol [7]. A filtered 48 h-supernatant was subjected to Western-blot and hemolysis analyses. 2.2. Polyclonal sera and MAbs The generation of the PLY-specific mouse hybridomas herein considered has been previously described [5]. The three secrete IgG1, kappa MAbs. The antibody 1.3C9.91 is also a mouse IgG1, kappa monoclonal generated against human pepsinogen C [8]. MAbs were purified by thiophilic adsorption chromatography (AFFI-T gel columns, Kem-En-Tec, Denmark) from serum- and protein-free culture supernatants (HybriMax, Sigma).
Anti-PLY polyclonal mouse serum refers to a pool made from the sera of the mice which were used to generate the above mentioned hybridomas. The generation of the anti-PLY rabbit IgG preparation has already been described [9 /11]. 2.3. Immunoblot analyses of SLY-containing supernatants Twenty microlitres of culture supernatant were assayed. Primary antibodies were probed at 1 mg; peroxidase conjugates to F(ab?)2 fragments of sheep IgG antimouse IgG (Jackson ImmunoResearch Laboratories) were used at 1:1000. Specific protein bands were detected applying the BM chemiluminescence blotting substrate (POD) (Roche Diagnostics). 2.4. Construction and Western blot analyses of recombinant full-length PLY and PCR-derived PLY fragments Table 1 shows the primers used to generate diverse recombinant products. Fig. 1 shows a graphical representation of their relationships to the putative domains of the PLY molecule. Recombinant PLY was PCR amplified from a colony of S. pneumoniae GB05. Amplification products were directly cloned into pGEM-T vector (Promega) and sequenced. Recombinant PLY (R) and recombinant fragments (A /S) were amplified from this gene, and cloned into Nde I and BamH I endonuclease restriction sites of pET-11a (Novagen). This expression plasmid was used to transform Escherichia coli DH10B (GibcoBRL). Plasmids with correct insert were used to transform E. coli BL21 (DE3) pLysS (Stratagene) for producing the corresponding recombinant proteins. Overnight cultures were diluted into 2 /TY broth supplemented with ampicillin (100 mg/ml) and chloramphenicol (25 mg/ml). Bacteria were grown at 37 8C until an optical density at 600 nm (OD600)/0.8 and then induced with 1 mM isopropyl-b-thiogalactopyranoside (IPTG). After incubation for 30 min, rifampicin (240 mg/ ml) was added to the culture and incubated for an additional 90 min at 37 8C. Bacteria were harvested by centrifugation, and boiled for 5 min in electrophoresis sample buffer. Whole-cell lysates were subjected to 16.5%T/3%C Tricine- or 12%- SDS-polyacrylamide gel electrophoresis. The proteins were transferred to Immun-Blot PVDF Membrane (0.4 or 0.2 mm) (Bio Rad) with a Mini Trans-Blot Electrophoretic Transfer Cell (Bio Rad) and the membranes were then blocked overnight at 4 8C in 1% blocking reagent (Roche Diagnostics). Membranes were incubated with antiPLY mouse polyclonal serum (1:500), anti-PLY rabbit IgG (1:50) or MAbs (10 mg). Mouse and rabbit polyclonal sera were preincubated with lyophilised
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A
229
Table 1 Synthetic oligonucleotide primers used for recombinant-PLY fragments Fragment Amino acid residues
5? Primer
3? Primer
Ra A B C D E F G H I J K N O P Q S
5? GAGGTAG/CATATG/GCAAATAAAGC 5? GAGGTAG/CATATG/GCAAATAAAGC 5? GAGGGTAAT/CATATG/CCCGATGAG 5? GATCGTGCT/CATATG/ACTTATAG 5? ATGTCCCAGCT/CATATG/CAGTATG 5? TTTTAAG/CATATG/TATTATACAGTCAG 5? GGGCG/CATATG/TATCTCAAGTTG 5? GGCGACCCA/CATATG/GGTGCCCGAG 5? GTACAGAC/CATATG/GAGACTAAG 5?CACTAGTATT/CATATG/AAAGGGAATG 5? GAGGTAG/CATATG/GCAAATAAAGC? 5? GATCGTGCT/CATATG/ACTTATAG 5? GGGCG/CATATG/TATCTCAAGTTG 5? ATGTCCCAGCT/CATATG/CAGTATG 5? GAGGTAG/CATATG/GCAAATAAAGC 5? ATGTCCCAGCT/CATATG/CAGTATG 5? GAGGTAG/CATATG/GCAAATAAAGC
5? 5? 5? 5? 5? 5? 5? 5? 5? 5? 5? 5? 5? 5? 5? 5? 5?
1 /471 1 /49 40 /98 97 /151 148 /207 199 /248 247 /305 299 /353 352 /414 415 /471 1 /151 97 /248 247 /414 148 /305 1 /305 148 /471 1 /359
TCTC/GGATCC/TAGTCATTTTCTACCTTA GTCGACA/GGATCC/GCTACTTTCTTTC CCAGGC/GATCC/ATACTCTAAGTCATC CATGCTG/GGATCC/GTTATTTATTCATAC ATCTCCT/GGATCC/TAAACAGCGTC CGT/GGATCC/CAACTAGAGATAGAC CTCTACC/GGATCC/ACCTAGCCTG CTGTAAGCT/GGATCC/TAAGTCTCAAC GATTACGAA/GGATCC/CTTATAAAGG TCTC/GGATCC/TAGTCATTTTCTACCTTA CATGCTG/GGATCC/GTTATTTATTCATAC CGT/GGATCC/CAACTAGAGATAGAC GATTACGAA/GGATCC/CTTATAAAGG CTCTACC/GGATCC/ACCTAGCCTG CTCTACC/GGATCC/ACCTAGCCTG TCTC/GGATCC/TAGTCATTTTCTACCTTA ACGC/GGATCC/CTATCTGTAAGCTGTAACCTTAG
Restriction sites for Nde I and BamH I are underlined. Stop codons are in bold. a Full-length PLY.
whole-cell lysates of non-transformed E. coli , at 2.5 mg/ ml, for 2 h at 37 8C and overnight at 4 8C. Similar preparations of normal mouse IgG (Sigma), preimmune normal rabbit serum or the 1.3C3.91 MAb were
parallely assayed as primary antibody negative controls. After several washing steps, blots were incubated with a 1:1000 dilution of anti-mouse IgG-HRP-conjugated (Sigma) or with a 1:5000 dilution of anti-rabbit IgG-
Fig. 1. Schematic representation of S. pneumoniae PLY and recombinant-PLY fragments. The first and last amino acid residues of PLY included in each recombinant protein are indicated.
230
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A
HRP conjugated (Sigma). Reactive proteins were detected with BM Chemiluminescence blotting substrate (POD) (Roche Diagnostics). 2.5. Phage display library The pVIII-9aa library was kindly provided by Dr Alessandra Luzzago, IRBM, Italy. This library is made up of random sequences of nine amino acids inserted into the amino terminus of the coat protein VIII (pVIII) of bacteriophage f1 [12], and contains 9.4 /107 independent clones. 2.6. Growth and purification of phages Amplification of the library was essentially done as described previously [13]. Briefly, over 1011 CFU of the pVIII-9aa library in E. coli strain XL1-blue (Stratagene) were inoculated in 500 ml of 2 /TY broth containing 15 mg/ml of tetracycline (TC, Sigma), 100 mg/ml of ampicillin (AMP, Sigma), and 1% glucose (2 /TY /TC / AMP /GLU) and incubated at 37 8C until the OD600 was approximately 0.5. This culture was then infected with M13-K07 helper phage (Amersham Pharmacia Biotech) (multiplicity of infection /50) by standing incubation for 30 min at 37 8C and afterwards with shaking for another 30 min period. Subsequently, 0.1 mM (final concentration) IPTG was added and incubated with strong shaking for 5 h at 37 8C. Bacterial cells were removed by centrifugation (twice for 30 min at 8600 /g ), and the phages present in the supernatant were precipitated twice with polyethylene glycol /NaCl (20% PEG-8000, 2.5 M NaCl). Finally, the phages were resuspended in Tris /buffered saline (TBS), 0.02% NaN3 to a final concentration of about 1.5 /1014 CFU/ml. 2.7. Biopanning Three rounds of biopanning were applied for each MAb. Briefly, Maxisorp immunotubes (5 ml; Nunc, Denmark) were incubated with 4 ml of specific-MAb (100 mg/ml in 10 mM phosphate-buffered saline [PBS], pH 7.3) at 4 8C overnight. Subsequently, the immunotubes were blocked with blocking buffer (3% bovine serum albumin [BSA] in PBS, 0.05% Tween-20) for 2 h at 37 8C, and washed three-times with PBS. The phage library (between 1 and 3 /1012 CFU in blocking buffer) was added, and the tubes were incubated for 30 min at room temperature (RT), rotated continuously on a turntable, followed by a standing incubation for 90 min at RT. The tubes were then washed ten-times with PBS, 0.05% Tween-20 and ten-times with PBS to remove the unbound phages. Bound phages were eluted with 1 ml of 0.1 M Glycine /HCl, pH 2.2, rotating for 10 min at RT. Eluted phages were neutralised with 0.5 ml of 1 M Tris /HCl, pH 7.4, and used to infect E. coli strain XL1-
blue. To determine the number of eluted phages, infected E. coli cells were plated by serial dilution on 2 /TY /AMP /IPTG-X /Gal plates. After the first round of specific biopanning, there was a ‘negative’ selection step in which phages non-specifically bound to the MAb used for ‘positive’ selection were eliminated using immunotubes coated with an irrelevant MAb, 1.3C3.91. The remaining phages were amplified and further enriched by two additional rounds of ‘positive’ selection performed as described above. For the last round of biopanning, the specific MAb was used at 50 mg/ml. After this, phages were individually isolated from infected E. coli cells to test their reactivity in ELISA. 2.8. Phage ELISA The reactivity of MAbs with monophages was assayed by means of indirect ELISA tests. Flat-bottom 96-well polystyrene enzyme immunoassay plates (Costar) were coated with serial dilutions of PEGprecipitated monophages (100 ml/well in PBS, pH 7.3) overnight at 4 8C. The plates were then blocked with blocking buffer for 2 h at 37 8C. After being washed, the wells were successively incubated with 100 ml of antipneumolysin (PLY) MAb (10 mg/ml in 1% BSA /PBS / Tween-20); anti-mouse IgG-HRP conjugated (Sigma) (dilution 1:1000); and 50 mM phosphate-citrate buffer, pH 5.2, containing 10 mg of O-phenylenediamine dihydrochloride (Sigma) and 0.05% H2O2. Plates were finally incubated at 37 8C for 30 min in the dark. Colour development was stopped by adding 100 ml/well of 2 M H2SO4, and read at 492 nm. Plates were washed threetimes with PBS, 0.05% Tween-20 between incubations. All determinations were made in triplicate. The 1.3C3.91 MAb was used as an irrelevant negative-control antibody. 2.9. Sequencing of phage insert DNA The monophages which developed the highest absorbances were selected for sequencing. Single stranded DNA was prepared by phenol extraction following the methods already described [14] and sequenced after applying the sequencing primer M13 (/40). 2.10. Peptide ELISA tests Synthetic peptides, /95% pure, were supplied by Neosystem, France. These tests were carried out as the phage ELISA ones, the antigens being peptides, suspended in 10 mM PBS, pH 7.3, at 1 mg/100 ml per well. Blocking buffer was 1% BSA in PBS, 0.05% Tween-20.
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A
Anti-PLY MAbs were assayed at 10 mg/ml; anti-PLY mouse polyclonal serum at 1:500; and normal mouse IgG (Sigma) at 5 mg/ml, in blocking buffer. 2.11. ELISA competition tests with soluble peptides For this, ELISA plates were coated with 50 ng of recombinant PLY in 100 ml of 10 mM PBS, pH 7.3. The reactivity of each MAb with PLY was titered in ELISA tests carried out as described above for the phage ELISA. A minimum amount of each Mab, which rendered a reliable absorbance signal was chosen as a positive control. For competition tests, MAbs were mixed with different amounts of the synthetic peptide in blocking buffer and preincubated for 2 h at 37 8C and overnight at 4 8C. The mixtures were then added to the PLY-coated wells and incubated for 20 min at 37 8C. The rest of ELISA continued as described above. Anti-mouse IgG-HRP conjugated (Sigma) was used at a dilution 1:5000. 2.12. Hemolysis neutralisation assays These were performed as previously described [5,6]. Titration of toxin preparations was done immediately before neutralisation tests. Both PLY and SLY were assayed at two hemolytic units (HU). The neutralisation capacity of each MAb was also titered. For neutralisation tests, MAbs were assayed at their minimum neutralising amount. For neutralisation-blocking assays, MAbs were preincubated with peptides as above. 2.13. V gene sequencing 1/3/107 MAb-producing hybridoma cells were used for each preparation of mRNA using the QuickPrep Micro mRNA Purification kit (Amersham Pharmacia Biotech). cDNA was PCR amplified with V regionspecific primers [15]. PCR products were then cloned into the pGEM-T vector (Promega) and submitted for sequencing using a Thermo Sequenase Cycle Sequencing kit (Amersham Pharmacia Biotech). The sequences were analysed by means of BLASTN searches of GenBank databases (http://www.ncbi.nih.gov). Germlines were those defined and recorded in the IgBLAST database of GenBank. Amino acid residue assignments within V regions and definition of subgroups were done applying ad hoc entries in the Kabat Database of Proteins of Immunological Interest (http://www.immuno.bme.nwu.edu). 2.14. Alignment analyses and structural models CDCs sequences were obtained from the Swiss-prot database. Sequence alignments were done using the gap
231
and PILEUP programs of the Genetics Computer Group sequencing package [16]. Structural models of PLY were generated with the SWISS-MODEL program (http://www.expasy.ch/swissmod/SWISS-MODEL.html). Images were produced using the SWISS-PDBVIEWER (http://www.expasy.ch/ spdbv/mainpage.html) and WEBLAB VIEWERLITE (http://www.accelrys.com) programs. 2.15. Sequence accession numbers The complete sequences of the V regions herein characterised are deposited in GenBank (accession numbers AY194240, AY194241, AF509572, AF509573, AF509574, AF509575, AF509576 and AF509577).
3. Results 3.1. Reactivity of MAbs with SLY in immunoblot assays and analyses of their neutralising properties After a first description of the anti-PLY monoclonal antibodies developed by us, suilysin (SLY), secreted by S. suis , was characterised as the most similar CDC to PLY [7]. We were, therefore, interested in analysing the recognition of SLY by our anti-PLY MAbs. The PLY-5 antibody has already been reported to recognise all the CDCs tested and to block their binding to eukaryotic membranes [6]. The PLY-7 MAb was also probed on SLY [10]. By applying Western-blot chemiluminescence enhancement methodology to a supernatant of the SLYproducing S. suis NCTC 10234 strain, we have also confirmed that the PLY-7 antibody does recognise SLY and neutralise its hemolytic activity whereas PLY-4 and PLY-8 do not. 3.2. Western blot analysis of the PCR-derived PLY fragments The results of this analysis are shown in Tables 2 and 3. Out of the nine different, approximately contiguous and 50 amino acid residues long, peptides examined (Table 2), the anti-PLY mouse polyclonal serum tested was reactive with five of such (B, E, G, H and I) whereas the rabbit IgG preparation recognised only four (B, E, H and I). To be reactive with fragment I, the corresponding whole-cell lysate had to be boiled for at least 15 min, suggesting that this peptide needed to be sufficiently unfolded to deploy its antigenic sites. PLY7 and PLY-8 unambiguously identified peptides H and I, respectively; surprisingly, PLY-4 and PLY-5 did not recognise any of the fragments tested.
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A
232
Table 2 Reactivity of polyclonal sera and MAbs with recombinant-PLY fragments demonstrated by enhanced chemiluminescence immunoblotting analyses Fragment
a
R (1 /471) A (1 /49) B (40 /98) C (97 /151) D (148 /207) E (199 /248) F (247 /305) G (299 /353) H (352 /414) I (415 /471) a b
Anti-PLY rabbit IgG
b
/ / / / / / / / / /
Anti-PLY polyclonal mouse serum
/ / / / / / / / / /
MAbs PLY-4
PLY-5
PLY-7
PLY-8
/ / / / / / / / / /
/ / / / / / / / / /
/ / / / / / / / / /
/ / / / / / / / / /
Full-length PLY. Immunoreactive, /; non-reactive: /.
Table 3 Further enhanced-chemiluminescence-immunoblotting analyses of the reactivity of PLY-4 with longer recombinant-PLY fragments
Table 4 Deduced amino acid sequence of the inserts of monophages rescued after biopanning with the PLY-4, PLY-7 and PLY-8 MAbs
Fragment
Reactivitya
Insert sequence
Frequencya
ELISAb
J (1 /151) K (97 /248) N (247 /414) O (148 /305) P (1 /305) Q (148 /471) S (1 /359)
/ / / // /// /// ////
PLY-4 RPPRYLANQ RPPRYLSQLc RLPRYMQEV NRPRWVKEA RPPRFQLSS RPPRYMVQS RPPKWVSAQc RKPRYLTQS
3 3 2 1 1 1 1 1
1.35f 0.95f 1.4f 1.7 1.6 1.5 0.9 0.4
PLY-7 GATLGPYPP GGQTFTDREd KGIYPLSYE
16 1 1
1.2f 1.4 0.7
PLY-8 KRAVQKWAPe
12
1.3f
a
Non-reactive, /; very weak, /; weak, //; strong, ///; very strong, ////.
To further investigate the reactivity of PLY-4, some longer PCR fragments were studied. Table 3 summarises the results obtained. Fragment S corresponds to the union of domains 1, 2 and 3. From the data shown in Table 3, it seems that PLY-4 requires fragments at least 150 amino acids long to be reactive with and that it recognises a core structure located somewhere between positions 151(E) and 247(Y) of PLY.
a
3.3. Biopanning of a random nonapeptide library, and rescue and identification of monophages
Number of times each monophage was independently recovered. Reactivity on monophage-coated indirect ELISA tests. c Residues present in the (232)RPLVYISSV(240) amino acid stretches of PLY are in bold. d Residues present in the (400)NGQDLTAHF(408) amino acid stretches of PLY are in bold. e Residues present in the (450)KRTISIWGT(458) amino acid stretches of PLY are in bold. f Mean of the absorbance developed by the monophage isolates.
After each round of selection, the number of eluted phages was determined and the percentage of phage that bound to the selecting MAb was defined as the ratio between this number and the corresponding input. This percentage increased significantly after the second round, indicating that the biopanning had been successful and that antibody-specific phages were selected. Monophages were isolated from each round. Some of them were randomly tested in ELISA, exhibiting diverse reactivities. Monophages which reacted more strongly with the specific MAb than with the irrelevant MAb,
1.3C3.91, and which developed an OD above that of the wild phage (with no insert), were taken as specific. Table 4 shows the deduced amino acid sequence of the inserts from rescued specific monophages, the number of times they were independently recovered and their ELISA reactivity. Twenty-seven monophages, isolated after the third round of biopanning with PLY-4, were randomly tested in ELISA. Of these, 22 were considered as positive; from these, 13, which developed the highest ELISA readings,
b
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A
were selected for sequencing. All the insert sequences recovered contained at least a pair of contiguous RP residues. When these amino acid sequences were aligned to the PLY sequence, RPPRFQLSS aligned to the 144(V)/ 152(K) amino acid stretch. It seems unlikely that this stretch encompasses the PLY-4 epitope, as it only overlaps in two residues with the 151(E) /247(Y) amino acid stretch and because, if this were the case, fragment J (1 /151) could have been recognised by this MAb as well. However, two other sequences, RPPRYLSQL and RPPKWVSAQ, coincided in their alignment to (232)RPLVYISSV(240) within the 151(E) /247(Y) amino acid stretch whereas the rest of the sequences aligned outside of it. Therefore, from these observations, the PLY-4 epitope could be located within the (232)RPLVYISSV(240) amino acid stretch of PLY. In the case of PLY-7, 29 monophages out of 32 from the third round and 11 out of 12 from the second round were considered as specific. From these, 12 from the third round and six from the second were selected for sequencing. Sixteen monophage clones presented the same amino acid sequence whereas the remaining two were unrelated. These sequences did not reveal a common motif, and strikingly the most reactive monophage was isolated only once after the second round. The recovered inserts were then aligned to the PLY sequence. The KGIYPLSYE and GATLGPYPP inserts aligned to amino acid stretches included as part of the PCR-derived peptide A and F, respectively, whereas GGQTFTDRE aligned to the (400)NGQDLTAHF(408) stretch, which is within the PCR-derived peptide H; the homologous SLY sequence is (428)NGVNLTSHW(436). With respect to the PLY-8 MAb, 20 individual phages from the third round were isolated. Twelve of the 16 monophages considered to be specific were randomly selected for sequencing. All presented the same amino acid sequence insert, KRAVQKWAP; this insert aligned to the (450)KRTISIWGT(458) stretch of PLY, included in fragment I. 3.4. ELISA and neutralisation-blocking tests These were performed to confirm the nature of the sequences recognised by PLY-4, PLY-7 and PLY-8; the PLY-5 MAb was used as control. By homology with the resolved crystallographic structure of PFO, Glu231 and Arg232 of PLY are part of a well surface-exposed contorted loop on the edge of the concave and convex faces of domain 1; their side chains bulge prominently on the protein (Figs. 4 and 5). Pro233 is covered by the (13)MNYD(16) stretch. The (234)LVYISSV(240) amino acid stretch is mostly buried under another loop, at the centre of which is D338, on
233
the convex face of domain 1. Arg232 has been predicted to form a salt bridge with Asp338 [17]. Therefore, in characterising the PLY-4 epitope, we also directed our attention to the amino acid residues preceding Arg232/Pro233. The following peptides were assayed: (227)GISAERPL(234) and (226)RGISAERPLVYISSV(240); in order to give a loop conformation to the (227)GISAERPL(234) peptide, another peptide was synthesised including a Cys at each end of the (227)GISAERPL(234) sequence. None of these peptides was reactive with PLY-4 in indirect ELISA tests or showed any interfering/blocking activity even when assayed at 100 mg. We attributed these results to the inability of the assayed peptides to properly recreate the antigenic features of the PLY-4 epitope. We then considered whether a longer peptide could recreate these antigenic features. Therefore, another peptide, 24 amino acid long, half of fragment E, from (216)D to (239)S positions of PLY, was probed as antigen in indirect ELISA tests. The anti-PLY mouse polyclonal serum and PLY-4 did recognise this peptide but not PLY-5; otherwise, 50 mg of this peptide did not block the neutralising capacity of PLY-4. A synthetic peptide, named as p21PLY, comprising residues from (393)T to (413)P of PLY and, therefore, encompassing the (400)NGQDLTAHF(408) stretch, was similarly probed. This peptide was recognised by PLY-7 and the anti-PLY mouse polyclonal serum but not by PLY-5. Amino acid residues (401)GQDLTAH(407) of PLY would conform a wellexposed b strand of one of the loops at the bottom of domain 4, between its two b-sheets, whereas amino acid residues N(400) and F(408) seem to be drawn up and already part of the neighbouring b strands. Thus, we also decided to test a synthetic peptide, called p7PLY, which corresponds to amino acid residues (401)GQDLTAH(407) of PLY, and a p7SLY, corresponding to the (429)GVNLTSH(435) amino acid stretch of SLY. The p7PLY and the p7SLY peptides were both recognised in indirect ELISA tests and blocked the neutralising ability of the PLY-7 MAb (Table 5). A peptide p5, corresponding to amino acid residues (425)IRECTGLAWEWWRTV(439), which includes the cysteine motif and, therefore, the PLY-5 epitope [6], behaved similarly in relation to the PLY-5 MAb. In the opposite way, neither the p7PLY nor the p7SLY peptide was recognised by or affected the neutralising characteristics of the PLY-5 MAb. Upon titration and for ELISA-competition tests, PLY-5 was assayed at 1 mg/well and PLY-8 at 20 ng/ well. A p8 peptide, corresponding to amino acid residues (450)KRTISIWGT(458) of PLY, was able to compete with the recognition of native PLY by the PLY-8 MAb in a dose-dependent manner (Fig. 2), as did the p5 peptide in relation to PLY-5. As a control, the p7PLY
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A
234
Table 5 Neutralisation-blocking assays of PLY-7 and PLY-5 PLY (2 HU)
PLY-7 (1 mg)
PLY-5 (2 mg)
p7PLY (mg)
p7SLY (mg)
p5 (mg)
Lysis
/a / / / / / / / /
/ / / / / / / / /
/ / / / / / / / /
/ / / 25 / / 25 / /
/ / / / / 25 / 25 /
/ / / / 31 / / / 31
YES NO NO YES NO YES NO NO YES
a
/, added; /, not added.
The VH regions are made of diverse VH and JH gene segment combinations; the two VL regions identified are also made of two different VL and Jk gene segment combinations. The VH regions of the PLY-8 and PLY-5 MAbs show a number of amino acid replacements threetimes higher than the number found in their VL partners; in relation to the PLY-4 and PLY-7 MAbs, their VH and VL regions show a similar number of amino acid substitutions.
Fig. 2. ELISA competition tests with the p8 and p5 peptides.
peptide, assayed at 100 mg, did not block such recognition by the PLY-8 or the PLY-5 MAbs. 3.5. V gene usage The four MAbs have used unique VH genes as their closest germlines are not repeated (Fig. 3, panel a). These VH genes belong to three different subgroups (Table 6). PLY-4 is made of a 7183 gene of the VH5 family. The PLY-7 and PLY-8 MAbs have utilised J558 genes of the VH1 gene family; the PLY-5 MAb has utilised a 36 /60 VH gene of the VH3 family. Overall, they display between 83 and 95% of nucleotide identities with the most homologous VH germline gene found. Nucleotide substitutions have led to amino acid replacements in all four CDRH1s and CDRH2s. The CDRH3s of PLY-4 and PLY-7 are of the same length (11 amino acid residues); that of PLY-8 is eight amino acid residues long. The CDRH3 of PLY-5 comprises only three amino acids (Glu/Ala /Ser). With respect to the k chains, it is surprising that PLY4, PLY-7 and PLY-8 share the V region (Fig. 3, panel b), with the same pattern of amino acid replacements. This V region shows a 96% nucleotide identity (Table 6) with its closest germline 19/20. The PLY-5 Vk region has been made from a different Vk gene (cw9) although also within the V subgroup; in this case, there is a nucleotide identity of 97% with this cw9 germline.
4. Discussion Our Western-blot observations from different contiguous PCR-derived PLY peptides, 50 amino acid residues long, and from mouse and rabbit PLY-specific polyclonal sera and the herein studied MAbs have allowed us to identify stretches of amino acid residues 40 /98, 199 /248, 299/353, 352/414 and 415/471 as immunogenic and antigenic. According to the structural data and folding models of PLY which parallel those described for PFO [2,17], all these stretches include well accessible and surface exposed sites. Most of fragment B (positions 40 /98) folds as an ample loop on the concave face of domain 1 (Fig. 4). Fragment E (amino acid residues 199/248) is mostly within domain 1. Fragment G encompasses two distinct stretches: amino acid residues 299 /318 are part of domain 3, whereas those running from positions 319 to 353 are part of a contorted loop on the convex face of domain 1. Fragment H (positions 352/414) corresponds to the first four b strands of the b-sandwich of domain 4; and fragment I (positions 415/471), to the last four. Altogether, these observations highlight the immunological significance of domains 1 and 4. A former screening with human and rabbit hyperimmune sera and partially overlapping synthetic peptides [18] had identified two common immunoreactive sites at the aminoterminal end of the PLY molecule, which comprise residues from (29)E to (48)R and from (131)L to (152)K positions. Our fragment B partially overlaps with the
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A Fig. 3. Amino acid sequences of heavy (a) and light-chain (b) variable regions of four anti-PLY MAbs, PLY-4, PLY-7, PLY-8 and PLY-5. The sequences are aligned to their closest available germline sequence in the IgBLAST database of GenBank. Dashes indicate identities. Sequence stretches used in primers are underlined. The FR/CDR assignments were according to the Kabat database of proteins of immunological interest.
235
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A
236
Table 6 Characteristics of the V regions of the four anti-PLY MAbs PLY-4
PLY-7
PLY-8
PLY-5
Gene usage VH familya Subgroupb VH germ-line segmentc D germ-line segment JH germ-line segment VL family Subgroup VL germ-line segment JL germ-line segment
7183 (VH5) IIID VH 69.1 (95%) DSP 2.7/8 (100%) JH4 (91%) Vk 19/28 V Vk19 /20 (96%) Jk2 (100%)
J-558 (VH1) IIA VMU-3.2 (91%) / JH3 (97%) Vk 19/28 V Vk19 /20 (96%) Jk2 (97%)
J-558 (VH1) IIA V23 (83%) DFL16.1/2 (100%) JH2 (97%) Vk 19/28 V Vk19 /20 (96%) Jk2 (100%)
36 /60 (VH3) IA VH 36 /60 (86%) / JH1 (91%) Vk 9/10 V cw9 (97%) Jk1 (100%)
CDR3 length (amino acids ) Heavy chain Light chain
11 8
11 8
7 8
3 9
Number of mutations (nucleotides ) VH segment 13 VL segment 10
24 10
48 11
40 8
Number of mutations (amino acids ) VH segment 9 VL segment 8
12 8
22 8
18 6
a b c
Family and germlines were assigned according to the IgBLAST database of GenBank. Subgroup is according to the Kabat database of proteins of immunological interest. Nucleotide homology with the closest germline.
Fig. 4. A homology model of the PLY molecule which shows the spatial distribution of the amino acid stretches, corresponding to PCRderived fragments, identified as immunogenic/antigenic in the present study. Fragments B, E and G are in grey. Domain 4 comprises fragments H and I. The location of the E231, R232, and D338 residues is indicated.
first of these immunoreactive sites, but otherwise no similar occurrences are seen. These observational discrepancies could be attributable to technical limitations inherent to the different epitope mapping methologies applied. PLY and SLY share 50% of amino acid residue identities. We have confirmed that PLY-7 and PLY-5
recognise linear epitopes on SLY and also neutralise its hemolytic activity, while PLY-4 and PLY-8 do not have these properties. From a pepscan analysis covering the whole amino acid sequence of PLY, the epitope recognised by PLY-5 was unambiguously assigned to the well-conserved undecapeptide motif present in all CDCs [6]; this motif would conform a well-exposed elongated loop at the bottom of domain 4. The PLY-5 epitope would be within the sequence WEWWRT (Fig. 6). We have found that PLY-5 does not recognise any of the herein 50-residue-long PCR-derived PLY peptides. This suggests that epitope recognition by this MAb is greatly dependent on the characteristics of the peptide, which encompasses it and on the context in which this peptide is expressed and/or tested. Proper analysis of the antigenicity of peptides and epitope identification requires concordant results from diverse immunoassay formats and epitope mapping methodologies [19]. In this context, we have to stress that this has been done for the present epitope mapping work and that especially, from the Western-blot analyses, we have obtained unambiguous results as no misleading cross-reactions among PCR-derived fragments were observed. Because fragments derived from proteinase K nicked PLY were not reactive with PLY-4, it was suggested that the cleavage site would be in the centre of the PLY-4 epitope, located around residue 142 [5]. Based on our present observations from full-length recombinant PLY and PCR-derived PLY fragments, it seems that the
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A
PLY-4 epitope, located in domain 1, is greatly conformation-dependent, and that encompasses a core structure between the 151(E) and the 247(Y) amino acid residues. According to the biopanning and indirect ELISA tests, this core reactivity could reside around the Arg at position 232. Based on PFO, domain 1 of PLY would be made of intermingled non-contiguous amino acid stretches, conforming a highly interdependent ordered structure [1,17]. The 199(Y) to the 248(L) amino acid stretch, which corresponds to the whole length of fragment E, is sandwiched between fragments B and G (Figs. 4 and 5) and the first N-terminal 20 amino acid residues. The PLY-4 epitope could be made of and/or determined by residues belonging to these three fragments. Residual conformational features of the PLY-4 epitope could explain its recognition in Western blot assays; conversely, any significant structural alteration of the PLY molecule would be greatly detrimental for its antigenicity. Proper reconstitution of the epitope recognised by the PLY-4 antibody seems to require rather long amino acid stretches. The elucidation of the ultimate characteristics of this epitope requires other experimental approaches. In PFO, it has been shown that insertion into the membrane is preceded by oligomerisation which implies the formation of a large prepore [4]. In the oligomer and pore of PLY, the Arg232 would be positioned at the outer surface of the ring. As the PLY-4 epitope seems to include Arg232, a residue which is predicted to be directly involved in the oligomerisation step [17], the hemolysis-neutralising capacity of the PLY-4 MAb would be exerted through its interference with the oligomerisation/formation of prepore events.
Fig. 5. A CPK representation of a lateral view of the packing of fragments A, B, E and G in which the E231, R232 and D338 residues are highlighted.
237
A number of observations from diverse truncates and mutants of CDCs [20 /23] suggests that some residues within domains 1 and 3 are key factors for oligomerisation and hemolysis by intervening in their correct folding and three dimensional structure; this would apply in particular to positions of PLY from 199(Y) to 248(L) (fragment E) which could be pivotal for the conformation of the PLY-4 epitope. Cross-inhibition ELISA tests between pairs of MAbs had shown that the binding of PLY-7 on PLY had positive modulation effects upon the subsequent binding of PLY-5; the reverse phenomenum was not observed. Analyses by means of flow cytometry of epitope recognition led to the conclusion that the neutralising capacity of PLY-5 and PLY-7 is due to the blocking of the binding of the toxin to the eukaryotic membrane and that their epitopes become embedded in the membrane upon binding of the toxin. Similar findings were made later on in relation to SLY [10]. The experimental findings herein described lead us to conclude that PLY-7 recognises an epitope in the (401)GQDLTAH(407) sequence of PLY and in the (429)GVNLTSH(435) of SLY. By sequence alignment and structural homology with PFO, the (401)GQDLTAH(407) stretch would conform one of the three parallel loops at the bottom of domain 4 (Fig. 6). The WEWWRT motif conforms one of the external loops, recognised by PLY-5, the other being the one conformed by this amino acid stretch recognised by the PLY-7 MAb. It is worth stressing that, while this WEWWRT loop and the intermediate one correspond to well conserved sequence stretches in the CDCs, the external loop recognised by the PLY-7 MAb shows a high sequence variability among the CDCs; this would explain the restricted reactivity of this MAb to PLY and SLY. The close spatial vicinity of the PLY-7 epitope to the PLY-5 within domain 4 could explain the above mentioned immunomodulatory and epitope recognition observations on membranes. A wealth of observations suggests that the cysteine motif and in particular the Trp-rich stretch is primarily involved in the interaction with cholesterol and in membrane insertion [1]; these events are accompanied by several conformational changes in the CDCs [20,21,24 /26]. Together with cholesterol, other membrane molecules could mediate toxin binding as well. To our knowledge, the only mutation so far generated within the PLY-7 epitope-containing homologous stretch of any CDC is that recently reported [27] in which leucine 461 of native LLO was substituted by threonine present in PFO as well as in PLY (position 405), SLY (position 433) and other CDCs. This conversion resulted in an increased hemolytic activity of LLO at a neutral pH but the mechanism(s) by which this effect is exerted remain(s) to be precisely explained.
238
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A
Fig. 6. (A) A homology model of domain 4 of PLY; the location of the sequences comprising the PLY-7, PLY-8 and PLY-5 epitopes is highlighted. (B, C and D) Structural details of the PLY-8, PLY-7 and PLY-5 sequences, respectively.
The blocking of the PLY-7 MAb on toxin binding could have, at least, two different explanations. First and based on our ELISA-derived modulation observations, the binding of PLY-7 to domain 4 could induce a conformational change that would lead the PLY-5 epitope-containing Trp-rich loop to lose affinity for cholesterol. Second, that binding could interpose a steric hindrance to the interaction between this toxin site and cholesterol and other membrane components. PLY-8 is a PLY non-neutralising MAb. From its reactivity with truncated forms of PLY, the epitope recognised by this MAb appeared to be in the last six Cterminal amino acid residues of PLY [5]. Nevertheless, our present findings indicate that its epitope lies between amino acid residues (450)KRTISIWGT(458) of PLY which would conform a b strand in the b sheet opposite to the WEWWRT loop (Fig. 6). No substitution mutants within the PLY-8 epitope-containing amino acid stretch have been generated, but, apart from its contribution to the folding of domain 4, this toxin site does not seem to be primarily involved with membrane interaction and/or oligomerisation events; therefore, the binding of the PLY-8 MAb to its epitope would not interfere with these processes and this would explain why this MAb has no neutralising capacity. The herein assigned location is in agreement with our previous
observations according to which the PLY-8 epitope would become embedded in the membrane after toxin binding. It is surprising that PLY-4, PLY-7 and PLY-8 share the same Vk region; in relation to its closest germline, this Vk region shows a nucleotide identity of 96%. The Vk region of the PLY-5 MAb has a nucleotide homology of 97% with its closest germline. All this suggests that these Vk chains are essentially germline encoded and that accumulation/selection of amino acid replacement within these Vk domains is not important for PLY binding. As the epitopes recognised by these four MAbs are different, it is obvious that the epitope specificity of these MAbs is dictated by the VH counterparts; a preponderant role of the heavy chain in the determination of the specificity of the secondary antibody response to protein antigens has already been reported [28,29]. We rescued two hybridomas from two different mice that happened to synthesise the same anti-hemolytic MAb (PLY-5/PLY-6). PLY-6 is another anti-PLY IgG1, kappa MAb from which it was inferred that it recognised the same epitope as PLY-5 [5]. The sequences of the VH and Vk genes of PLY-6 have been previously reported [30]. In agreement with those former observations, we have now found that PLY-5 and PLY-6 have
´ lvarez et al. / Immunology Letters 88 (2003) 227 /239 B. Sua´rez-A
identical VH and Vk regions. Therefore, it seems that the PLY-5 epitope strongly selects for a particular VH/VL pairing. The four MAbs herein considered express VH genes very likely deriving from different germlines and belonging to different subgroups; furthermore, these VH genes are recombined with different JH segments, adding support to the assumption that the four MAbs are independently derived. This diverse use of VH genes is a common feature of the humoral immune response to complex protein antigens [31]. The VH regions of the PLY-7, PLY-8 and PLY-5 MAbs have a nucleotide homology of 91, 83 and 86%, respectively, with its closest germline; since these MAbs recognise epitopes within domain 4 which is involved in toxin binding, it seems that the efficient antibody recognition of this toxin domain is associated with mutated VH regions. This may not be the case of the PLY-4 epitope, mediating oligomerisation events within domain 1, as the VH region of the PLY-4 MAb has a 95% of nucleotide homology and a low number of amino acid replacement in relation to its closest germline.
Acknowledgements The authors are indebted to Dr Alessandra Luzzago, Istituto di Richerche di Biologia Molecolare P Angeletti (IRBM), Pomezia, Italy, for providing the pVIII nonapeptide library. We also wish to thank David H. Wallace for the English language correction of the ´ lvarez manuscript. This work and Beatriz Sua´rez-A were supported by Proyecto PB-MED01-08, FICYT, Principado de Asturias, Spain; Marı´a del Mar Garcı´aSua´rez was financed in part by Proyecto 1FD97-1185, Ministerio de Ciencia y Tecnologı´a, Spain.
References [1] R.K. Tweten, M.W. Parker, A.E. Johnson, Curr. Top. Microbiol. Immunol. 257 (2001) 15 /33. [2] J. Rossjohn, S.C. Feil, W.J. McKinstry, R.K. Tweten, M.W. Parker, Cell 89 (1997) 685 /692. [3] R. Ramachandran, A.P. Heuck, R.K. Tweten, A.E. Johnson, Nat. Struct. Biol. 9 (2002) 823 /827. [4] L.A. Shepard, O. Shatursky, A.E. Johnson, R.K. Tweten, Biochemistry 39 (2000) 10284 /10293. [5] J.R. de los Toyos, F.J. Me´ndez, J.F. Aparicio, F. Va´zquez, M.M. Garcı´a Sua´rez, A. Fleites, C. Hardisson, P.J. Morgan, P.W. Andrew, T.J. Mitchell, Infect. Immun. 64 (1996) 480 /484.
239
[6] T. Jacobs, M.D. Cima-Cabal, A. Darji, F.J. Me´ndez, F. Va´zquez, A.A.C. Jacobs, Y. Shimada, Y. Ohno-Iwashita, S. Weiss, J.R. de los Toyos, FEBS Lett. 459 (1999) 463 /466. [7] R.P.A.M. Segers, T. Kenter, L.A.M. de Haan, A.A.C. Jacobs, FEMS Microbiol. Lett. 167 (1998) 255 /261. [8] E.P. Diamandis, S. Nadkarni, B. Bhaumik, A. Abdelrahman, D.N. Melegos, G. Borchert, M.H. Black, M. Alonso, A. Salas, J.R. de los Toyos, A. Sampedro, C. Lo´pez-Otı´n, Clin. Chem. 43 (1997) 1365 /1371. [9] M.D. Cima-Cabal, F. Va´zquez, J.R. de los Toyos, F.J. Me´ndez, J. Clin. Microbiol. 37 (1999) 1964 /1966. [10] M.D. Cima-Cabal, Ph.D. thesis, Universidad de Oviedo, Spain, 1999. [11] M.D. Cima-Cabal, F.J. Me´ndez, F. Va´zquez, M.M. Garcı´aSua´rez, J.R. de los Toyos, J. Immunoassay Immunochem. 22 (2001) 99 /112. [12] F. Felici, L. Castagnoli, A. Musacchio, R. Jappelli, G. Cesareni, J. Mol. Biol. 222 (1991) 301 /310. [13] A. Luzzago, F. Felici, in: S. Cabilly (Ed.), Methods in Molecular Biology: Combinatorial Peptide Library Protocols: Combinatorial Peptide Library Protocols, vol. 87, Humana Press, 1998, pp. 155 /164. [14] J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, 1989. [15] R. Orlandi, D.H. Gussow, P.T. Jones, G. Winter, Proc. Natl. Acad. Sci. USA 86 (1989) 3833 /3837. [16] J. Devereux, P. Haeberli, O. Smithies, Nucleic Acids Res. 12 (1984) 387 /395. [17] J. Rossjohn, R.J.C. Gilbert, D. Crane, P.J. Morgan, T.J. Mitchell, A.J. Rowe, P.W. Andrew, J.C. Paton, R.K. Tweten, M.W. Parker, J. Mol. Biol. 284 (1998) 449 /461. [18] P. Salo, A. Na¨rva¨nen, M. Leinonen, Infect. Immun. 61 (1993) 2822 /2826. [19] M.H.V. Van Regenmortel, Biologicals 29 (2001) 209 /213. [20] M. Palmer, P. Saweljew, I. Vulicevic, A. Valeva, M. Kehoe, S. Bhakdi, J. Biol. Chem. 271 (1996) 26664 /26667. [21] E.M.A. Ghani, S. Weis, I. Walev, M. Kehoe, S. Bhakdi, M. Palmer, Biochemistry 38 (1999) 15204 /15211. [22] I. Yamamoto, H. Kimoto, Y. Taketo, A. Taketo, Biosci. Biotechnol. Biochem. 65 (2001) 2682 /2689. [23] H. Baba, I. Kawamura, C. Kohda, T. Nomura, Y. Ito, T. Kimoto, I. Watanabe, S. Ichiyama, M. Mitsuyama, Biochem. Biophys. Res. Commun. 281 (2001) 37 /44. [24] L.A. Shepard, A.P. Heuck, B.D. Hamman, J. Rossjohn, M.W. Parker, K.R. Ryan, A.E. Johnson, R.K. Tweten, Biochemistry 37 (1998) 14563 /14574. [25] M. Palmer, I. Vulicevic, P. Saweljew, A. Valeva, M. Kehoe, S. Bhakdi, Biochemistry 37 (1998) 2378 /2383. [26] O. Shatursky, A.P. Heuck, L.A. Shepard, J. Rossjohn, M.W. Parker, A.E. Johnson, R.K. Tweten, Cell 99 (1999) 293 /299. [27] I.J. Glomski, M.M. Gedde, A.W. Tsang, J.A. Swanson, D.A. Portnoy, J. Cell. Biol. 156 (2002) 1029 /1038. [28] A.J. Caton, S.E. Stark, J. Kavaler, L.M. Staudt, D. Schwartz, W. Gerhard, J. Immunol. 147 (1991) 1675 /1686. [29] R. Tuteja, Immunology 97 (1999) 1 /8. [30] M. Garcia, G. Strachan, A.J.R. Porter, W.J. Harris, FEMS Immunol. Med. Microbiol. 22 (1998) 225 /231. [31] S. Clarke, R. Rickert, M.K. Wloch, L. Staudt, W. Gerhard, M. Weigert, J. Immunol. 145 (1990) 2286 /2296.