Adjustable Locks and Flexible Keys: Plasticity of Epitope−Paratope Interactions in Germline Antibodies

Supplementary Material References Subscriptions Permissions Email Alerts

Tarique Khan and Dinakar M. Salunke J Immunol 2014; 192:5398-5405; Prepublished online 30 April 2014; doi: 10.4049/jimmunol.1302143 http://www.jimmunol.org/content/192/11/5398

http://www.jimmunol.org/content/suppl/2014/04/30/jimmunol.130214 3.DCSupplemental.html This article cites 41 articles, 15 of which you can access for free at: http://www.jimmunol.org/content/192/11/5398.full#ref-list-1 Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscriptions Submit copyright permission requests at: http://www.aai.org/ji/copyright.html Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/cgi/alerts/etoc

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 9650 Rockville Pike, Bethesda, MD 20814-3994. Copyright © 2014 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606.

Downloaded from http://www.jimmunol.org/ at University of California San Francisco on January 27, 2015

This information is current as of January 27, 2015.

The Journal of Immunology

Adjustable Locks and Flexible Keys: Plasticity of Epitope–Paratope Interactions in Germline Antibodies Tarique Khan*,1 and Dinakar M. Salunke*,†

T

he clonal selection theory (1) predicts a correlation between antigenic determinants and the corresponding Abs, implying that each B lymphocyte expresses a unique BCR (Ab) that can respond to an incoming binding Ag without having been previously exposed to it. Degenerate reactivity of individual germline Abs has been suggested to be a physiological requirement in response to the potentially infinite antigenic repertoire to be encountered by the humoral immune system (2–5), although immune evasion by pathogens through escape mutations provides evidence for the limitations of the germline repertoire. In any case, the population of B cells available at any point does not present the entire potential repertoire. Therefore, to be able to respond each time exposure occurs, the humoral immune system has to be able to use its resources economically, including being able to use the same BCR to recognize different Ags as well as to use different BCRs to recognize the same Ag. Naive germline BCRs are likely to need rapid identification and selection upon exposure to initiate an immune response. In contrast, affinity-matured Abs develop through selection by prolonged exposure to the dominant conformation of the Ag. It is only after the selection of high-affinity B cells in germinal centers that affinity maturation and clonal expansion follow (6, 7). Therefore, it is interesting to address the structural mechanisms adopted by germline encoded BCRs for Ag *National Institute of Immunology, New Delhi 110067, India; and †Regional Centre for Biotechnology, Gurgaon 122016, India 1 Current address: Department of Biochemistry, University of Texas Southwestern Medical Center, Dallas, TX.

Received for publication August 16, 2013. Accepted for publication March 27, 2014. This work was supported by the Department of Biotechnology, Government of India. The coordinates and structure factors presented in this article have been submitted to the Research Collaboratory for Structural Bioinformatics Protein Data Bank (http:// www.pdb.org) under accession codes 4bh7 and 4bh8. Address correspondence and reprint requests to Dr. Dinakar M. Salunke, Regional Centre for Biotechnology, 180, Udyog Vihar Phase-I, Gurgaon, Haryana 122016, India. E-mail address: [email protected] The online version of this article contains supplemental material. Abbreviations used in this article: CNS, Crystallography and NMR System; Gdp, GDPRPSYISHLL; PDB, Protein Data Bank; PEG, polyethylene glycol; Ppy, PPYPAWHAPGNI; RMSD, root mean square deviation. Copyright Ó 2014 by The American Association of Immunologists, Inc. 0022-1767/14/$16.00 www.jimmunol.org/cgi/doi/10.4049/jimmunol.1302143

recognition. Previous structural studies have provided insights on how germline Ab pluripotency may enhance the BCR repertoire diversity (4, 5, 8–10). Comparative analysis of degeneracy in interactions of germline versus affinity-matured Abs with antigenic targets would contribute further to the understanding of the structural mechanisms operating in the configuration of B cell immune responses. In the current study, we have addressed the versatility of Ag recognition at the initial encounter by analyzing the binding of Ags with overlapping epitopes by genetically independent germline Abs. The crystal structures of mAb 36-65 in complex with the peptide epitopes GDPRPSYISHLL (Gdp) and PPYPAWHAPGNI (Ppy) (2) were investigated by x-ray crystallography. These structures were compared with complexes of the same peptides with another independent mAb BBE6.12H3 to understand the structural bases of their recognition. The H chain V region of mAb 36-65 has been shown to be constructed from VH J558, DH Fl16.1, and JH2 gene segments, whereas mAb BBE6.12H3 consists of VH 186.2, DH Fl16.1, and JH2 (11–13). Furthermore, mAbs 36-65 and BBE6.12H3 use L chains of the k and l isotypes, respectively (12, 13). L chain CDRs adopt distinct canonical conformations as defined by Chothia et al. (14). Thus, mAbs 36-65 and BBE6.12H3 represent Abs of independent origin having distinct CDR sequences (Fig. 1). Their structures and binding properties have already been extensively investigated (2, 4, 5). We have previously explored mechanistic details of mAb BBE6.12H3 binding to these peptides (4). Further analyses of the same peptides bound to the germline mAb 36-65 have allowed us to compare the recognition of a flexible Ag by independent germline Abs. Comparative analyses of these structures show that entirely different CDR sequences are involved in the interaction with the peptide Ags. These data provide structural understanding of the ways in which specific recognition of a given Ag can be achieved by many different VDJ/VJ combinations in the remarkably adaptable naive immune receptor repertoire. Our studies also reveal how unrelated Abs structurally adjust to recognize a flexible Ag and indicate that the primary B cell response is likely composed of BCRs having a high degree of structural adaptability. In contrast, our comparison of flexible Ag recognition by germline versus affinity-matured Abs shows that they adopt distinct structural mechanisms.

Downloaded from http://www.jimmunol.org/ at University of California San Francisco on January 27, 2015

Ag recognition by independent primary Abs against a small flexible Ag with overlapping epitopes was analyzed to address the determinants of Ag specificity during the initial encounter. Crystal structures of two distinct dodecapeptide Ags, GDPRPSYISHLL and PPYPAWHAPGNI, in complex with the germline mAb 36-65 were determined and compared with the structures of the same Ags bound to another independent germline mAb, BBE6.12H3. For each peptide Ag, the two germline mAbs recognized overlapping epitopes, but in different topologies. The peptide structures differed, and the two paratopes attained discrete conformations, leading to different surface topologies, in a mode that can be described as adjustable locks and flexible keys. This is in contrast to mature mAbs, in which conformational convergence of different paratopes while binding to a common epitope in a similar conformation has been reported. These results suggest that the primary immune receptor repertoire is highly versatile as compared with its mature counterpart. Germline and mature mAbs adopt distinct mechanisms for recognizing a flexible epitope. Whereas conservation of conformational repertoire is a key characteristic of mature mAbs achieved through affinity maturation, the germline mAbs, at the initial stages of Ag encounter, maintain substantial plasticity, accommodating a broad specificity repertoire. The Journal of Immunology, 2014, 192: 5398–5405.

The Journal of Immunology

5399 Refinement and model building

Peptides

The Crystallography and NMR System (CNS) suite was used for structure refinement (19). Both Rwork (Crystallographic R-factor) and Rfree (Free Rfactor) (19) values were monitored during the refinement. We used 10% of the total reflections in each case for calculation of Rfree values. Rigid body refinement was carried out for the complete F(ab) molecule, and VH, VL, CH, and CL domains were treated as discrete units. The models were further refined by using positional and temperature factor refinement protocols of CNS. COOT was used for model building and to display electron-density maps (20). Final structures for the F(ab) were obtained after several rounds of manual model building in COOT (20). The peptide molecules were built into the Ag-binding cavity of the refined F(ab) models using the electron density evident in the Fo-Fc maps. Water molecules were added using the water pick program in CNS. The quality of the model was checked with MolProbity (21). Structural models were generated using PyMOL (http:// www.pymol.org), and structural superimpositions were done using program superpose in CCP4 (22). The Ab–Ag interactions and buried surface areas were analyzed using PISA (23). Structure factors and coordinates of the crystal structures of the 36-65 F(ab) bound to peptides Gdp and Ppy were deposited in the Protein Data Bank(http://www.pdb.org) under the accession codes 4BH8 and 4BH7, respectively.

The dodecameric peptides used in this study were identified in a previous study by using a phage display peptide library kit (New England Biolabs, Cambridge, MA) (2). Briefly, the germline mAb BBE6.12H3 was shown to bind to a series of independent peptides from a random phage display library. From this set, two independent clones, 3 (GDPRPSYISHLL) and BA7-09 (PPYPAWHAPGNI), which hereafter would be referred to as Gdp and Ppy, were selected for crystallographic analyses with the mAb 36-65 in the current study. These peptides were synthesized by the solid-phase method in an automated peptide synthesizer (431A; Applied Biosystems, Foster City, CA). Peptides were cleaved from the resin by trifluoroacetic acid (SigmaAldrich). A linear gradient of acetonitrile containing 0.1% trifluoroacetic acid was used to purify the peptides on a Delta Pak C18 column (Waters, Milford, MA). Purified peptides were characterized by mass spectrometry.

Ab Hybridoma cells secreting the IgG mAb 36-65 were cultured in DMEM containing 10% FCS (12). Male BALB/c mice were g-irradiated (4 Gy) and primed with Freund’s incomplete adjuvant 72 h prior to i.p. injection of 5 3 106 hybridoma cells in 500 ml Dulbecco’s PBS per mouse. Ascitic fluid generated in the mouse peritoneal cavity was tapped after ∼4 to 5 d. All animal experiments were approved by the Institutional Animal Ethics Committee.

F(ab) preparation A three-step purification protocol was followed to purify IgG from the ascitic fluid. The 40% ammonium sulfate–precipitated fraction containing IgG was resuspended in 10 mM Tris buffer at pH 8.5. Further purification was carried out by affinity chromatography followed by ion-exchange chromatography using Protein-G Sepharose and DEAE 5PW anion exchange column on an HPLC system (Waters Delta 600; Waters), respectively. F(ab) fragments were prepared by papain digestion of the purified IgG at pH 7.1. The digestion mixture was dialyzed in 10 mM Tris buffer (pH 8) and loaded onto an anion-exchange column (DEAE 5PW) to purify the F(ab) fragments. F(ab) purity was tested by SDS-PAGE, and concentration was estimated by the BCA protein assay kit (Pierce, Rockford, IL).

Crystallization and data collection Crystals were grown by hanging-drop vapor-diffusion method (15) at 28˚C by using a starting F(ab) concentration of 10 mg/ml. Crystals of the 36-65– peptide complex were obtained from a solution in which Gdp or Ppy peptides were preincubated with F(ab) for 24 h before initiating crystallization. A 20fold molar excess of the peptide was used for cocrystallization experiments. The binary complex crystallized in the presence of 15–24% polyethylene glycol (PEG) of different molecular weights in 50 mM sodium cacodylate buffer in the pH range of 6.5–7 containing 0.0–0.1 M zinc chloride. Diffraction data for the mAb 36-65 F(ab) in complex with peptides Gdp and Ppy were collected from the crystals grown in 21% PEG 6000, 50 mM sodium cacodylate (pH 6.7), 50 mM zinc chloride, and 17% PEG 3350 in 50 mM sodium cacodylate (pH 6.5), respectively. Diffraction data for both complexes were collected on the home source, RU300 (Rigaku, Tokyo, Japan). Images were registered on Mar345dtb image plate with an oscillation of 1˚ per image. The crystals were cryoprotected by soaking in mother liquor containing 25% glycerol and flash frozen. Data were integrated with MOSFLM (16) and scaled with SCALA (17).

Structure determination The reported structure of the Ag-free mAb 36-65 F(ab) (Protein Data Bank [PDB] code 2A6J) was used as an initial search model in MOLREP (18) for determining the structure of 36-65 F(ab)–peptide complexes. This model produced good solutions for data corresponding to both 36-65–Gdp and 36-65–Ppy structures with correlation coefficient values of 72.2 and 67.7, respectively. Subsequent refinements were conducted using the structures from these molecular replacement solutions.

FIGURE 1. Alignment of the CDR sequences of mAbs BBE6.12H3 and 36-65. Starting and ending residue numbers of each CDR are shown; identical amino acids are marked with asterisks. Deletions in CDRs are marked as a dash.

Results Peptide Ag Gdp and Ppy bound 36-65 F(ab) structures Sequence alignment comparing the H and L chain CDRs of mAb 36-65 and BBE6.12H3 is shown in Fig. 1. The crystal structure of the mAb 36-65 F(ab) in complex with the independent phage ˚ resdisplay-derived dodecapeptide Gdp was determined at 2.4 A olution (Fig. 2A). This complex was named 36-65–Gdp. The L and H chains were named A and B and contained 211 and 220 residues, respectively. The structure has 98% residues in the allowed region of the Ramachandran map. The crystal structure of the mAb 36-65 F(ab) in complex with another independent phage ˚ display–derived dodecapeptide, Ppy, was determined at 2.9 A resolution (Fig. 2B). This complex was named 36-65–Ppy, and the L and H chains were named A and B, respectively. In this complex, 97.5% of the residues were in the allowed region of the Ramachandran map. In both of these complexes, the structure of the entire F(ab) molecule could be built into the electron density with the exception of a loop extending from residues Ala138B to Thr140B. This loop region has also been shown to be disordered in many other previously reported F(ab) structures. Crystal data and refinement statistics for both the structures are shown in Table I. Because the N-terminal of the peptide Gdp moves into the solvent, the first three residues from the N terminus, Gly, Asp, and Pro, could not be mapped. The C-terminal residue of the Gdp peptide was also not visible in the electron density (Fig. 2A, Supplemental Fig. 1A). The average temperature factors for the Gdp peptide were slightly higher than those for the F(ab) molecule as the N-terminal residues were solvent exposed and did not interact with the CDR (Tables I, II). Among the bound residues, the N-terminal of the Gdp peptide is seen to interact with the paratope generated by the 36-65 H chain CDRs. In the case of the 36-65–Ppy complex, the first and the last three residues of the Ppy peptide were not visible in the electron density, presumably because they were solvent exposed (Fig. 2B, Supplemental Fig. 1B). The Ab–Ag interactions were analyzed in terms of the total buried surface areas upon binding. In the case of 36-65–Gdp

Downloaded from http://www.jimmunol.org/ at University of California San Francisco on January 27, 2015

Materials and Methods

5400

GERMLINE Abs SHOW PLASTICITY OF SPECIFICITY Table II. Interactions of Gdp peptide with mAbs 36-65 and BBE6.12H3 Gdp Residues

Arg4 Pro5 Ser6 Tyr7 Ile8 Ser9 His10

Tyr102B (1) Tyr102B (6), Gly103B (3) Tyr101B (7), Tyr102B (8), Gly103B (10) Val100B (1), Tyr101B (8), Tyr102B (1) Tyr32B (3), Tyr101B (1) Ser31B (4), Tyr32B (12), Gly33B (5), Ser99B (13), Val100B (13),Tyr101B (10), Tyr106B (5) 50B Tyr (3), Asn52B (5), Tyr101B (1), Tyr106B (5)

BBE6.12H3 Residuesa

Tyr32L (4) Tyr32L (6), Trp91L (2) Tyr32L (8), Trp91L (2) Trp33H (2), Tyr97H (1)

Trp33H (4), Arg50H (1) Arg50H (3), Asp52H (5), Ser54H (1), Gly56H (10)

Epitope–paratope interactions were calculated using van der Waals radii in contact program of CCP4 (16). The 12-mer peptide (Gdp) residues, which were common in both complexes, are shown in boldfaced text. The number of interactions shown by each Ab residue is shown in parentheses. a Data taken from Khan and Salunke (4).

Cross-reacting peptide Ag: Gdp in different germline Ab environments

FIGURE 2. Two independent crystal structures of the germline mAb 36-65 F(ab) bound to peptide Ags Gdp (A) and Ppy (B). The two peptidebound states are shown as ribbons highlighting secondary structural features. Peptide molecules are shown as blue sticks. Surface representations of mAb 36-65 in complex with Gdp and Ppy peptides, respectively, are shown in the bottom panel. The interacting regions of the paratope surfaces are highlighted in red along with the interacting residues (sticks).

complex, the total buried surface areas of F(ab) and peptide were ˚ 2, respectively. In the case of 36-65–Ppy complex, 2180 and 504 A ˚ 2, the corresponding total buried surface areas were 2290 and 426 A respectively. Table I. Data collection and refinement statistics

Data-processing statistics Space group Cell dimensions ˚) a, b, c (A b (˚) ˚) Maximum resolution (A Rmerge (%) Average (I)/(Sig I) Completeness (%) No. of unique reflections Multiplicity Refinement statistics ˚) Resolution range (A No. of reflections Rwork/Rfree (%) No. of atoms Protein Peptide Water B-factor Protein Peptide Water RMSDs ˚) Bond length (A Bond angle (˚)

36-65–Gdp

36-65–Ppy

P21

P21

54.4,77.0, 59.2 101.7 58–2.4 10.0 (40.0) 12.0 (3.2) 100.0 (100.0) 19,076 4.5 (4.5)

54.5, 76.8, 59.1 102.3 57–2.9 12.0 (32.0) 9.4 (3.9) 97.8 (99.7) 10,548 4.7 (4.7)

50–2.4 18,818 23.0/25.1

57.8–2.89 9,490 22.9/25.4

3,310 69 190

3,313 68 77

24.5 84.4 32.5

22.5 64.5 23.0

0.007 1.7

0.01 1.8

Data in parentheses are for highest-resolution shell.

Comparison of the structure of Gdp determined in this study in complex with mAb 36-65 with that in the previously reported complex with mAb BBE6.12H3 (PDB code 2Y06) (4), provided interesting insights into the structural behavior of Gdp in the context of BCR recognition. As shown in Fig. 3A, the paratope surfaces of mAbs 36-65 and BBE6.12H3 bound to Gdp were very different. Superimpositions of these complexes show that the CDR conformations of these two germline mAbs differ substantially. Although CDRs H1 and H2 seem to be similar (root mean square ˚ ), as they adopt the same canonical deviation [RMSD] ,0.5 A conformations. Although the sequences of CDRH1 and CDRH2 were also similar (Fig. 1), the individual interacting residues involved were very different. The elbow angles and the buried ˚ 2, surface areas of the two F(ab) molecules differ by 10˚ and 670 A respectively. Eight residues from Arg4P to Leu11P could be unambiguously traced into the electron density in the 2Fo-Fc map for the 36-65– Gdp complex (Supplemental Fig. 1A). In contrast, nine residues from Asp2P to His10P were evident in the BBE6.12H3–Gdp complex. Structural comparison of Gdp in the two Ab complexes revealed that the gross topology of the peptide is conserved. They both exhibit approximately a right-angled arrangement of the backbone with a change in direction at the proline residue (Fig. 3B). However, neither the backbone nor the side chain atoms of any residue were superimposable. Comparison of RMSD values for each ˚ ) as well as the peptide residue in terms of the Ca positions (2.5 A ˚ side-chain atoms (4.6 A) also indicated that the conformations of the individual residues of the Gdp peptide were different when bound to mAbs 36-65 and BBE6.12H3. Interestingly, in both complexes, the peptide bound in the same orientation, albeit in different regions of the paratope (Fig. 4A, 4B). The backbone and the side chains of the peptide residues oriented differently to interact with distinct paratope residues in the 36-65–Gdp and BBE6.12H3–Gdp complexes (Fig. 4A, 4B). Some peptide residues interacted with similar amino acids but their spatial locations on the paratope differed substantially, as outlined in Tables II and III and Fig. 3C and 3D. The middle region of the peptide (i.e., the residues PSYISH) was conformationally similar when bound to either mAb 36-65 or mAb BBE6.12H3. The backbone conformations of this region were similar, although side-chain orientations differed. This is particularly

Downloaded from http://www.jimmunol.org/ at University of California San Francisco on January 27, 2015

Leu11

36-65 Residues

The Journal of Immunology

5401

evident for Tyr7P and His10P (Fig. 3B). In addition, the CDR residues of mAbs BBE6.12H3 and 36-65, which interact with

these residues, were entirely different (Table II). One residue of the Gdp peptide, Ile8P, was solvent exposed in the BBE6.12H3– Gdp complex, but the same residue had numerous interactions in 36-65–Gdp (Tables II, III). Table III. Comparison of atomwise polar interactions of 36-65 and BBE6.12H3 with Gdp peptide Gdp [Atoms]

36-65 [Atoms]

4

FIGURE 4. Comparison of Gdp peptide-binding mode on independent Ab paratope surfaces. Peptide is displayed as sticks. In the surface view, mAbs 36-65 and BBE6.12H3 are shown in blue and magenta, respectively. The interacting paratope surfaces in both complexes are colored orange. Close-up of the interacting Connolly surfaces of the Abs are decorated according to their hydropathy feature (red, hydrophobic; blue, charged). (A) 36-65–Gdp. (B) BBE6.12H3–Gdp.

Arg [N] Pro5 [N] Ser6 [N] Ser6 [O] Ser6 [OG] Tyr7 [O] Ile8 [N] Ile8 [O] Ser9 [O] Ser9 [N] Ser9 [OG] His10 [N] His10 [O] His10 [ND1] His10 [NE2]

B:Tyr102 [OH] B:Gly103 [N]

BBE6.12H3 [Atoms]a

L:Tyr32 [OH] L:Tyr32 [OH] L: Trp91 [NE1] L:Trp91 [NE1] H:Tyr95 [OH]

B:Tyr102 [N], Gly103 [N] B: Tyr101 [O] B: Tyr101 [N] H:Arg50 [NH2] 101

B:Tyr [O] B:Tyr32 [OH] B:Tyr101 [N] B: Ser31 [O] B:Ser31 [O], Ser99 [O] B:Gly33 [N], Ser99 [O]

Polar contacts were calculated using PISA (23). a Data taken from Khan and Salunke (4).

H: Arg50 [NE], Arg50 [NH2]

Downloaded from http://www.jimmunol.org/ at University of California San Francisco on January 27, 2015

FIGURE 3. Differences in the CDR conformations and the interactional network of mAbs, 36-65, and BBE6.12H3 to recognize the peptide Ag Gdp. (A) Stereoscopic view of the structural alignment of the CDRs of the Gdp peptide bound mAbs 36-65 (blue) and BBE6.12H3 (orange). (B) Structure alignment to compare the conformations of the peptide Gdp when bound to mAbs 36-65 (red) and BBE6.12H3 (green). Stereoscopic diagrams displaying ˚ distance) of the bound interacting residues (within a 4-A peptide are shown as thin sticks, and the stretch of the Gdp peptide evident in the electron-density map of its complexes with mAbs 36-65 and BBE6.12H3 is shown as thick sticks. (C) 36-65–Gdp. (D) BBE6.12H3–Gdp.

5402

FIGURE 5. The differences in the CDR conformations and the interactional network of mAbs 36-65 and BBE6.12H3 to recognize the peptide Ag Ppy. (A) Stereoscopic view of the structural alignment of the CDRs of the Ppy peptide-bound mAbs 36-65 (blue) and BBE6.12H3 (orange). (B) Structure alignment to compare the conformations of the peptide Ppy when bound to mAbs 36-65 (green) and BBE6.12H3 (red). Residues involved in the b-turn formation in Ppy peptide are marked with dotted square (green) and dotted circle (red) in 36-65–Ppy and BBE6.12H3–Ppy, respectively. Stereoscopic views display˚ distance) of the bound ing interacting residues (within a 4-A peptide are shown in thin-stick and the stretch of the Gdp peptide evident in the electron density map of its complexes with mAbs 36-65 and BBE6.12H3 is shown in thick-stick representation. (C) 36-65–Ppy. (D) BBE6.12H3–Ppy.

the case of BBE6.12H3-Gdp, both L and H chain CDRs contributed equally to the polar interactions. Detailed comparisons of atomwise polar contacts of each residue are provided in Table III. Cross-reacting peptide Ag: Ppy in different germline Ab environments The structure of Ppy bound to the mAb BBE6.12H3 (PDB code 2Y07) has been reported previously (4). A comparative analysis of the 36-65–Ppy and BBE6.12H3-Ppy paratopes revealed distinct topologies and charge distributions (Figs. 5A, 6A, 6B). Superimposition of the secondary structural elements also showed different CDR conformations except for CDRs H1 and H2 (Fig. 5A). Analyses at the level of individual interacting residues indicated entirely different side-chain orientations in all of the CDRs. The elbow angles and the buried surface areas differed by 10˚ and ˚ 2, respectively. 220 A The structures of the Ppy peptide bound to the germline mAbs 36-65 and BBE6.12H3 were superimposed to compare peptide

Downloaded from http://www.jimmunol.org/ at University of California San Francisco on January 27, 2015

The Gdp peptide bound at different sites in the Ag-combining grooves of mAbs BBE6.12H3 and 36-65, involving entirely different residues (Fig. 3C, 3D). Gdp did not interact with the L chain CDRs in the 36-65–Gdp complex. Even in the H chain CDRs, entirely different sets of residues were involved in the interaction with the two mAbs (Fig. 3C, 3D, Table II). In BBE6.12H3–Gdp, several residues from the L and H chains made van der Waals contacts with the peptide residues (Table II). In the 36-65–Gdp complex, the peptide Ag bound over the CDRH3 loop in a zigzag pattern. Thus, the majority of the van der Waals contacts were made by CDRH3 residues in 36-65–Gdp (Fig. 3C, Table II). In 36-65– Gdp, Tyr7P and Ile8P were oriented toward the CDRH3 and hence had additional interactions when compared with those in mAb BBE6.12H3. Similarly, His10P in the 36-65–Gdp complex made more contacts as it snugly fit into a groove between CDRH3 and CDRH1 (Fig. 4A, 4B) and also contributed maximally toward the buried surface area. The majority of the polar interactions in 36-65– Gdp were formed by the H chain CDRs H1 and H3. In contrast, in

GERMLINE Abs SHOW PLASTICITY OF SPECIFICITY

The Journal of Immunology

5403 Table IV. Interactions of Ppy peptide with mAbs 36-65 and BBE6.12H3 Ppy Residuesa

36-65 Residues

BBE6.12H3 Residuesa

Tyr3

Lys59B (5), Lys65B (1)

Pro4 Ala5 Trp6

Tyr57B (4), Lys59B (3) Tyr57B (3) Tyr50B (1), Tyr106B (3)

Trp91L (4), Ser93L (2), Asn94L (2) Trp91L (6), Trp33H (7), Arg50H (8), Ala58H (2) Trp33H (2), Arg50H (1)

His7

Tyr32A (8), Gly91A (4), Asn92A (5), Arg96A (3), Gly104B (7), Ser105B (8) Asn92A (3), Thr93A (4), Leu94A (6), Arg96A (1) Thr93A (4), Leu94A (7) Leu94A (2)

Pro2

Pro9 Gly10 FIGURE 6. Comparison of Ppy peptide-binding mode on independent Ab paratope surfaces. Peptide is displayed as sticks. In the surface view, mAbs 36-65 and BBE6.12H3 are shown in blue and magenta, respectively. The interacting paratope surfaces in both complexes are colored orange. Close-up of the interacting Connolly surfaces of the Abs are decorated according to their hydropathy feature (red, hydrophobic; blue, charged). (A) 36-65–Ppy. (B) BBE6.12H3–Ppy.

conformation in the two complexes (Fig. 5B). Interestingly, Ppy adopted a b-turn conformation in both of the structures. However, this turn consisted of distinct sets of residues: AWHAP and YPAW in 36-65–Ppy and BBE6.12H3–Ppy, respectively (Fig. 5B). Superimposition of the Ppy peptide from the two complexes revealed different conformations. The RMSD values in the positions of Ca ˚ , respectively. The peptide and side chains were 2.1 and 4.6 A residues had considerable differences in their side-chain orientations, perhaps matching the surface charge distribution on the paratope surfaces of mAbs 36-65 and BBE6.12H3 (Fig. 5A, 5B). Changes in the side chains of Tyr3P and Trp6P were particularly clearly evident. Although eight residues from Pro2P to Pro9P could be traced into the electron density in the 2Fo-Fc map for the 36-65–Ppy complex, only six residues from Pro1P to Trp6P were discernible in the BBE6.12H3–Ppy complex (Fig. 5C, 5D). The electron-density map for the Ppy peptide is shown in Supplemental Fig. 1B. Although detailed conformational features of the peptide in the two structures were distinct, the overall fold of the peptide remained similar (Fig. 5B). The residues of the two independent F(ab) molecules involved in interactions with Ppy peptide are shown in Table IV and Fig. 5C and 5D. As is evident in Table V, polar interactions were observed in both complexes but with different residues of the mAbs 36-65 and BBE6.12H3. In the 36-65–Ppy complex, whereas the majority of the van der Waals contacts were formed by CDRL3, some residues from CDRL1, CDRH2, and CDRH3 also contributed to it (Fig. 5C, 5D, Table IV). The peptide Ppy bound at different sites in the mAbs 36-65 and BBE6.12H3. The middle region of the peptide made van der Waals contacts with Tyr57B. The C terminus of the peptide Ag showed a major interaction with CDRL3 (Fig. 6A, 6B). The majority of the van der Waals contacts in this case were through polar residues. In contrast, van der Waals contacts in BBE6.12H3-Ppy complex were formed with all types of amino acids. The N terminus of the peptide in the complexes with mAbs 36-65 and BBE6.12H3 bound at distinct paratope surfaces because it had major interactions with CDRH2 and CDRL3, respectively (Fig. 5C, 5D, Table IV). Out of the 12 peptide residues, 5 comprising the sequence PYPAW were seen bound to both of the mAbs. However, the

Epitope–paratope interactions were calculated using van der Waals radii in contact program of CCP4 suite (16). The 12-mer peptide (Ppy) residues, which were common in both complexes, are shown in boldfaced text. The number of interactions shown by each Ab residue is shown in parentheses. a Data taken from Khan and Salunke (4).

conformation and interactions of these residues were different in the two Ab environments (Fig. 5B, Table IV). In the 36-65–Ppy complex, Pro2P did not interact with the paratope, as it was solvent exposed. In contrast, the same residue made several contacts with the paratope in the BBE6.12H3–Ppy complex (Table IV). These observations demonstrate that these two germline mAbs use different kinds of interactions and with different sets of residues for recognizing the same peptide epitope.

Discussion The phenomenon of degeneracy in Ag binding in germline mAbs is likely to be significant for enabling responsiveness to a diversity of pathogens (2, 5, 8, 10). The high specificity of recognition in the immune system is countered by evasion strategies in pathogens which, among other modalities (24), use mechanisms to change their antigenic surfaces (25–27) and/or mimic host moieties (28, 29). Structurally ill-defined antigenic determinants could well add further difficulties for the host during the initial encounter when germline Abs would be involved. It was evident that the two germline mAbs used in this study, BBE6.12H3 and 36-65, adopt different structural strategies while recognizing a common epitope, even though their affinities are comparable. The Ag-combining sites in the two mAbs show overlap, with similar footprints for the common epitope. However, they adopt entirely different paratope topologies with substantial differences in

Table V. Comparison of atomwise polar interactions of 36-65 and BBE6.12H3 with Ppy peptide Ppy [Atoms]

36-65 [Atoms]

BBE6.12H3 [Atoms]a

Pro2 [N] Tyr3 [OH]

B:Lys59 [NZ], Lys65 [NZ]

L:Tyr32 [OH] H:Arg50 [NE], 50Arg50 [NH1]

Pro4 [O] Ala5 [N] His7 [N] His7 [ND1] His7 [NE2] Pro9 [O]

B:Tyr57 [OH], Lys59 [NZ] B:Tyr57 [OH] B:Arg96 [NH2] B:Gly104 [O], Ser105 [N] A:Asn92 [O] A:Leu94 [N]

Polar contacts were calculated using PISA (23). a Data taken from Khan and Salunke (4).

Downloaded from http://www.jimmunol.org/ at University of California San Francisco on January 27, 2015

Ala8

Trp33H (8), Tyr95H (3), Tyr97H (2)

5404

to further delineate the structural mechanisms involved and their physiological significance.

Acknowledgments We thank Drs. Tim Manser and K.V.S. Rao for the gift of the hybridomas, Drs. Deepak T. Nair, Jasmita Gill, and Satyajit Rath for critically reading the manuscript, and H.S. Sarna for technical assistance.

Disclosures The authors have no financial conflicts of interest.

References 1. Burnet, F. M. 1957. A modification of Jerne’s theory of antibody production using the concept of clonal selection. Aust. J. Sci. 20: 67–69. 2. Manivel, V., F. Bayiroglu, Z. Siddiqui, D. M. Salunke, and K. V. Rao. 2002. The primary antibody repertoire represents a linked network of degenerate antigen specificities. J. Immunol. 169: 888–897. 3. Nguyen, H. P., N. O. Seto, C. R. MacKenzie, L. Brade, P. Kosma, H. Brade, and S. V. Evans. 2003. Germline antibody recognition of distinct carbohydrate epitopes. Nat. Struct. Biol. 10: 1019–1025. 4. Khan, T., and D. M. Salunke. 2012. Structural elucidation of the mechanistic basis of degeneracy in the primary humoral response. J. Immunol. 188: 1819– 1827. 5. Sethi, D. K., A. Agarwal, V. Manivel, K. V. Rao, and D. M. Salunke. 2006. Differential epitope positioning within the germline antibody paratope enhances promiscuity in the primary immune response. Immunity 24: 429–438. 6. Ochiai, K., M. Maienschein-Cline, M. Mandal, J. R. Triggs, E. Bertolino, R. Sciammas, A. R. Dinner, M. R. Clark, and H. Singh. 2012. A self-reinforcing regulatory network triggered by limiting IL-7 activates pre-BCR signaling and differentiation. Nat. Immunol. 13: 300–307. 7. Berek, C., and C. Milstein. 1987. Mutation drift and repertoire shift in the maturation of the immune response. Immunol. Rev. 96: 23–41. 8. James, L. C., P. Roversi, and D. S. Tawfik. 2003. Antibody multispecificity mediated by conformational diversity. Science 299: 1362–1367. 9. Notkins, A. L. 2004. Polyreactivity of antibody molecules. Trends Immunol. 25: 174–179. 10. Wedemayer, G. J., P. A. Patten, L. H. Wang, P. G. Schultz, and R. C. Stevens. 1997. Structural insights into the evolution of an antibody combining site. Science 276: 1665–1669. 11. Wong, Y. W., D. S. Gill, B. Parhami-Seren, M. K. Short, S. R. Sompuram, and M. N. Margolies. 1998. Structural requirements for a specificity switch and for maintenance of affinity using mutational analysis of a phage-displayed antiarsonate antibody of Fab heavy chain first complementarity-determining region. J. Immunol. 160: 5990–5997. 12. Vora, K. A., J. V. Ravetch, and T. Manser. 1997. Amplified follicular immune complex deposition in mice lacking the Fc receptor gamma-chain does not alter maturation of the B cell response. J. Immunol. 159: 2116–2124. 13. Wong, Y. W., P. H. Kussie, B. Parhami-Seren, and M. N. Margolies. 1995. Modulation of antibody affinity by an engineered amino acid substitution. J. Immunol. 154: 3351–3358. 14. Chothia, C., A. M. Lesk, A. Tramontano, M. Levitt, S. J. Smith-Gill, G. Air, S. Sheriff, E. A. Padlan, D. Davies, W. R. Tulip, et al. 1989. Conformations of immunoglobulin hypervariable regions. Nature 342: 877–883. 15. Wlodawer, A., and K. O. Hodgson. 1975. Crystallization and crystal data of monellin. Proc. Natl. Acad. Sci. USA 72: 398–399. 16. Leslie, A. G. W. 1992. Joint CCP4 and ESFEACMB newsletter on protein. Protein Crystallogr. 26: 27–33. 17. Collaborative Computational Project, Number 4. 1994. The CCP4 suite: programs for protein crystallography. Acta Crystallogr. D Biol. Crystallogr. 50: 760–763. 18. Navaza, J. 1994. AMoRe: an automated package for molecular replacement. Acta Crystallogr. A 50: 157–163. 19. Br€unger, A. T. 1992. Free R value: a novel statistical quantity for assessing the accuracy of crystal structures. Nature 355: 472–475. 20. Emsley, P., and K. Cowtan. 2004. Coot: model-building tools for molecular graphics. Acta Crystallogr. D Biol. Crystallogr. 60: 2126–2132. 21. Chen, V. B., W. B. Arendall, III, J. J. Headd, D. A. Keedy, R. M. Immormino, G. J. Kapral, L. W. Murray, J. S. Richardson, and D. C. Richardson. 2010. MolProbity: all-atom structure validation for macromolecular crystallography. Acta Crystallogr. D Biol. Crystallogr. 66: 12–21. 22. Kabsch, W. 1976. A solution for the best rotation to relate two sets of vectors. Acta Crystallogr. A 32: 922–923. 23. Krissinel, E., and K. Henrick. 2005. Detection of protein assemblies in crystals. Comput. Life Sci. 3695: 163–174. 24. Mulhern, O., B. Harrington, and A. G. Bowie. 2009. Modulation of innate immune signalling pathways by viral proteins. Adv. Exp. Med. Biol. 666: 49–63. 25. Manel, N., and D. R. Littman. 2011. Hiding in plain sight: how HIV evades innate immune responses. Cell 147: 271–274. 26. Sandbulte, M. R., K. B. Westgeest, J. Gao, X. Xu, A. I. Klimov, C. A. Russell, D. F. Burke, D. J. Smith, R. A. Fouchier, and M. C. Eichelberger. 2011. Discordant antigenic drift of neuraminidase and hemagglutinin in H1N1 and H3N2 influenza viruses. Proc. Natl. Acad. Sci. USA 108: 20748–20753.

Downloaded from http://www.jimmunol.org/ at University of California San Francisco on January 27, 2015

the conformations of the CDR loops. The common epitopes also adopt distinct conformations when bound to different mAbs. However, greater conformational rearrangements were observed in the germline mAb paratopes than in the corresponding epitopes. These findings indicate that structural plasticity mechanisms to recognize the Ag are inherent to the germline mAbs. Interestingly, in previous work, independent mature Ab clones exhibited structural convergence against an immunodominant epitope, PS1, known to be flexible in solution (30). The PS1 epitope was driven to a common conformational state and the CDRs of each of the independent mature mAbs against this epitope adopted identical conformations while binding to it (31). Thus, contrary to the broad conformational repertoire of the primary Abs observed in the current study, that of the mature Abs was conserved during Ag binding (31–34). In this context, the mode of physical contacts between an Ag and two distinct germline mAbs as deciphered in the current study might have implications for rapid recognition of flexible epitopes by a limited Ab repertoire. It is plausible that the enhanced flexibility in CDRs compared with that in the peptide Ags might enable recruitment of multiple independent naive B cell clones. It has been shown that both the framework region and the CDRs have a considerable amount of inherent conformational plasticity (35–37). Therefore, it is not surprising that distinct germline Abs recognize the same epitope by rearranging the CDR conformations. This may well have implications of Ag specificity beyond the naive BCR repertoire, because Kaji et al. (38) have shown in a recent report that the B cell memory can contain both germline-encoded and somatically mutated BCRs. Polyclonal Abs generated against an Ag are generally expected to bind diverse epitopes with distinct structures. If the Ag is small, such as a dodecapeptide as in the present case, the polyclonal response might become effectively monospecific due to the inevitable overlap of reacting surfaces. If this small Ag is flexible, different Ab clones binding to it may possibly recognize different topological states of the molecule. Indeed, the peptide Ags Gdp and Ppy binding to two independent mAbs BBE6.12H3 and 36-65 in distinct topological states reflects this possibility. The inherent conformational potential of a flexible peptide is of considerable structural interest. Functionally relevant conformational preferences of flexible molecules with known physiological receptors can indeed be deciphered in the form of their complexes. However, the conformational preferences of linear peptides, which do not embody a well-defined structural motif and do not have known physiological receptors, are harder to dissect. Abs as receptors, particularly when they have flexible binding sites, as is the case in germline Abs, can provide snapshots of possible conformational preferences for such flexible peptides. This is particularly attractive when structural complexes of more than one such independent flexible receptor with a common ligand are available. The crystal structures of the two linear peptides bound to two different germline Abs determined in the current study provided this scenario. Our data illustrate the structural versatility of germline Abs, which is relevant during initial antigenic encounter, whereas mature Abs of the secondary immune response do not appear to retain such a wide conformational diversity. Indeed, the many instances of independent mature Abs showing convergence of their paratope topologies while binding a common epitope support this possibility (33, 39–41). Such use of distinct structural strategies by BCRs at germline versus mature stages may have physiological significance. Because the current analysis is based on 12-mer peptides and their interactions with germline mAbs, additional studies with protein Ags and their cognate Abs, both germline and mutated, would help

GERMLINE Abs SHOW PLASTICITY OF SPECIFICITY

The Journal of Immunology

36.

37.

38.

39.

40.

41.

17/9 and three different Fab-peptide complexes. J. Mol. Biol. 234: 1098– 1118. Stanfield, R. L., M. Takimoto-Kamimura, J. M. Rini, A. T. Profy, and I. A. Wilson. 1993. Major antigen-induced domain rearrangements in an antibody. Structure 1: 83–93. Arevalo, J. H., C. A. Hassig, E. A. Stura, M. J. Sims, M. J. Taussig, and I. A. Wilson. 1994. Structural analysis of antibody specificity. Detailed comparison of five Fab’-steroid complexes. J. Mol. Biol. 241: 663–690. Kaji, T., A. Ishige, M. Hikida, J. Taka, A. Hijikata, M. Kubo, T. Nagashima, Y. Takahashi, T. Kurosaki, M. Okada, et al. 2012. Distinct cellular pathways select germline-encoded and somatically mutated antibodies into immunological memory. J. Exp. Med. 209: 2079–2097. Xu, Y., P. A. Ramsland, J. M. Davies, M. Scealy, K. S. Nandakumar, R. Holmdahl, and M. J. Rowley. 2004. Two monoclonal antibodies to precisely the same epitope of type II collagen select non-crossreactive phage clones by phage display: implications for autoimmunity and molecular mimicry. Mol. Immunol. 41: 411–419. Ramsland, P. A., W. Farrugia, T. M. Bradford, P. Mark Hogarth, and A. M. Scott. 2004. Structural convergence of antibody binding of carbohydrate determinants in Lewis Y tumor antigens. J. Mol. Biol. 340: 809– 818. Scheid, J. F., H. Mouquet, B. Ueberheide, R. Diskin, F. Klein, T. Y. Oliveira, J. Pietzsch, D. Fenyo, A. Abadir, K. Velinzon, et al. 2011. Sequence and structural convergence of broad and potent HIV antibodies that mimic CD4 binding. Science 333: 1633–1637.

Downloaded from http://www.jimmunol.org/ at University of California San Francisco on January 27, 2015

27. Hilleman, M. R. 2004. Strategies and mechanisms for host and pathogen survival in acute and persistent viral infections. Proc. Natl. Acad. Sci. USA 101(Suppl. 2): 14560–14566. 28. Moran, A. P., M. M. Prendergast, and B. J. Appelmelk. 1996. Molecular mimicry of host structures by bacterial lipopolysaccharides and its contribution to disease. FEMS Immunol. Med. Microbiol. 16: 105–115. 29. Stebbins, C. E., and J. E. Gala´n. 2001. Structural mimicry in bacterial virulence. Nature 412: 701–705. 30. Nayak, B. P., R. Tuteja, V. Manivel, R. P. Roy, R. A. Vishwakarma, and K. V. Rao. 1998. B cell responses to a peptide epitope. V. Kinetic regulation of repertoire discrimination and antibody optimization for epitope. J. Immunol. 161: 3510–3519. 31. Nair, D. T., K. Singh, Z. Siddiqui, B. P. Nayak, K. V. Rao, and D. M. Salunke. 2002. Epitope recognition by diverse antibodies suggests conformational convergence in an antibody response. J. Immunol. 168: 2371–2382. 32. Goldman, E. R., W. Dall’Acqua, B. C. Braden, and R. A. Mariuzza. 1997. Analysis of binding interactions in an idiotope-antiidiotope protein-protein complex by double mutant cycles. Biochemistry 36: 49–56. 33. Berger, C., S. Weber-Bornhauser, J. Eggenberger, J. Hanes, A. Pl€uckthun, and H. R. Bosshard. 1999. Antigen recognition by conformational selection. FEBS Lett. 450: 149–153. 34. Fields, B. A., F. A. Goldbaum, X. Ysern, R. J. Poljak, and R. A. Mariuzza. 1995. Molecular basis of antigen mimicry by an anti-idiotope. Nature 374: 739–742. 35. Schulze-Gahmen, U., J. M. Rini, and I. A. Wilson. 1993. Detailed analysis of the free and bound conformations of an antibody. X-ray structures of Fab

5405