0161-5890/92$5.00 + 0.00 Pergamon Press Ltd

Molecular Immunology, Vol. 29, No. 5, pp. 677-687, 1992 Printed in Great Britain.

LOCALIZATION OF SEQUENCE-DETERMINED NEOEPITOPES AND NEUTROPHIL DIGESTION FRAGMENTS OF C-REACTIVE PROTEIN UTILIZING MONOCLONAL ANTIBODIES AND SYNTHETIC PEPTIDES* S.-C. YING,~ E. SHEPHARD,$F. C. DE BEER,§ J. N. SIEGEL,? D. HARRIS,~ B. E. GE\KLTRZ,~ M. FRIDKIN~ and H. GEWURZ~~J**

tRush Medical College, Chicago, Illinois, U.S.A.; SUniversity of Capetown Observatory, South Africa; $University of Kentucky, Lexington, Kentucky, U.S.A. and IThe Weizmann Institute of Science, Rehovot, Israel (First received 21 May 1991; accepted in revised form 30 September

1991)

Abstract-We recently described 17 anti-CRP mAb, seven to native- (or conformational) and 10 to neo- (or sequence-determined) epitopes, including several anti-neo-CRP mAb specific for CRP peptide 199-206. In the present study, four new anti-native- and four new anti-neo-CRP mAb were generated and characterized by ELISA reactivity with native and modified human and rabbit CRP, as well as binding to pronase fragments of human CRP in Western blots. Assays with 17 synthetic CRP peptides identified anti-neo-CRP mAb specific for peptides 1-16, 1424 and 137-152, respectively. The anti-neo-CRP mAb were reacted with fragments obtained by digesting CRP with multiple additional enzymes, including Staphylococcal V8 protease, trypsin, elastase, plasmin, thrombin and alpha-chymotrypsin. Native CRP was remarkably resistant to enzymic digestion, particularly in the presence of calcium, but was readily cleavable upon denaturation. Twenty-three informative fragments served to further distinguish mAb reactivity with at least four additional neo-CRP epitopes, which presumptively included residues in the regions of amino acids 22-45,41-61, 114-121 and 13&138, respectively. The eight epitopes identified corresponded well with predicted regions of CRP antigenicity. In addition, at least six distinct native or conformation-determined epitopes were delineated. Reactivity of the anti-neo-CRP mAb with fragments of CRP generated by PMN enzymes indicated that regions sensitive to cleavage by neutrophil enzymes are located at approximately 3, 10 and 16 kD from the amino terminus of the CRP subunit. We expect that the anti-CRP mAb described and mapped herein will be useful tools for the elucidation of CRP structure and function.

INTRODUCTION Human C-reactive protein (CRP) is a trace protein in normal serum (Claus et al., 1976), which elevates up to lOOO-fold in association with tissue necrosis and inflammation, and hence is known as an acute phase reactant. The native CRP molecule is a pentamer comprised of five identical, non-covalently linked subunits, and is a

member of the “pentraxin” family of proteins (Osmand et al., 1977). Modified forms of CBP h&e been described (Potempa et al., 1983, 1987) which show antigenic as well as electrophoretic and binding characteristics distinct from those of native CRP

The reported functions of CRP include the ability to activate the complement system (Kaplan and Volanakis, 1974; Siegel et al., 1974), and bind to and modulate the responsiveness of neutrophils (Zeller et al., 1986a; Shephard et al., 1988, 1989; Kew et al., 1990), monocytes *Presented in part to the Annual Meeting of the American Association of Immunologists, New Orleans, June 1990 (Zeller et al., 1986b; Zahedi et al., 1989) and platelets (Potempa et al., 1988). The molecular conformations (Ying et al., 1990). ((Author to whom correspondence should be addressed at: and sites of CRP involved in these functions are beginDepartment of Immunology/Microbiology, Rush Medical ning to be identified (Buchta and Fridkin, 1986; Robey College, 1753 W. Congress Parkway, Chicago, IL 60612, et al., 1987; Shephard et al., 1990; Jiang et al., 1991); U.S.A. present evidence indicates that the native conformation **Holder of the Thomas J. Coogan, Senior Chairman in is necessary for complement activation while modified Immunology established by Marjorie Lindheimer Everett. CRP displaying neo-CRP (i.e. sequence-determined) epiAbbreviations: CRP, C-reactive protein; DTT, dithiothreitol; topes preferentially activates leukocytes and platelets HRP, horse radish peroxidase; KLH-PC, phosphoryl(Potempa et al., 1988; Jiang et al., 1991). choline-derivatized keyhole limpet hemocyanin; mAb, In a previous study involving 17 mAb, at least five monoclonal antibody (-ies); PMN, polymorphonuclear epitopes on native CRP, and at least three epitopes on leukocytes; raCRP, rabbit C-reactive protein; RT, room temp; PC, phosphorylcholine; TBS, Tris-buffered saline modified CRP, were identified (Ying et al., 1989). (pH 7.4). In order to help elucidate the structure-function 677

S.-C. YING et al.

678

relationships of CRP, an additional eight anti-CRP mAb were produced. The reactivities of the 16 previous and eight new mAb were tested with multiple enzyme-generated fragments of CRP and synthetic CRP peptides. At least six distinct conformational and eight distinct linear epitopes were identified, and the linear epitopes were localized to include regions between residues l-16, 14-24, 2245, 41-61, 114121, 130-138, 147-152 and 199-206, respectively, from the N-terminus. The mAb characterized in this way were used to identify fragments obtained upon digestion of CRP with neutrophil enzymes. MATERIALS

AND METHODS

Reagents

KLH-PC was prepared as previously described (Potempa et al., 1983). Trypsin, alpha-chymotrypsin, plasmin, thrombin and Staphylococcus aureus V8 protease (V8 protease) were purchased from Sigma Chemicals (St Louis, MO). Elastase was obtained from Calbiochem (San Diego, CA). The protein standard used in analyses of the SDS gels was obtained from Bio-Rad Laboratories (Richmond, CA). Preparation of CRP

Human CRP was purified by affinity chromatography on phosphorylcholine-substituted Biogel 0.5 A with citrate elution, followed by DE-52, gel filtration and second PC affinity chromatography, and biotinylated as previously described (Ying et al., 1989). Rabbit CRP was purified by similar affinity chromatography followed by gel filtration on Sephacryl S-200 in the absence of added calcium and Protein A Sepharose column chromatography as previously described (Cabana et al., 1982). CRP peptides

Seventeen synthetic CRP peptides were generously prepared and provided by Drs F. A. Robey (National Institute of Dental Research, Bethesda, MD) (Robey et al., 1987), R. Houghton (Scripps Clinic and Research Foundation, La Jolla, CA) (Ying et al., 1989), synthesized by Research Genetics (Atlanta, GA), or prepared in our laboratories as previously described (Buchta and Fridkin, 1986). Polyclonal antibodies

Polyclonal antibodies used included unconjugated and HRP-conjugated goat anti-mouse IgG + IgM (H + L chain) obtained from Pel-Freez Biologicals (Rogers, AK) and HRP-conjugated avidin obtained from Boehringer Mannheim Biochemicals (Indianapolis, IN). Monoclonal antibodies (mAb)

In addition to the 17 anti-CRP mAb previously described (Ying et al., 1989), which were raised by a regimen involving booster immunizations with both rabbit and human CRP, eight new mAb were prepared by identical procedures, except all immunizations involved only native- or urea-modified human CRP. The

mAb were further purified by protein A Sepharose affinity chromatography by using the Affi-Gel protein A MAPS kit obtained from Bio-Rad Laboratories, using individual columns for each mAb and eluting with 2 M KSCN. The affinity-isolated mAb were dialyzed into Tris-buffered saline, sterile-filtered and stored at - 70°C. For certain experiments mAb IgG was prepared by precipitation with 50% ammonium sulfate as previously described (Ying et al., 1989). Enzyme-linked

immunosorbent assay (ELISA)

ELISA assays were performed to quantify reactivities with both native- and modified- (neo-) CRP determinants as described previously (Ying et al., 1989). In brief, CRP was either captured on immobilized KLH-PC to express native CRP epitopes selectively or directly immobilized on polystyrene plates to express neo-CRP epitopes selectively; serial dilutions of mAb were added, and detected with peroxidase conjugated-goat anti-mouse IgG. In some assays, CRP was reduced in 0.1 M DTT for 2 hr at 37°C and alkylated with 0.06 M iodoacetamide for 15 min at RT before immobilization on polystyrene plates. Dot blot assays

Dot blots were carried out as previously described (Ying et al., 1989). SDS-PAGE

and immunoblots

SDS-PAGE was carried out on 15% polyacrylamide gels (Bio-Rad Laboratories) using the buffer system described by Laemmli (1970). Immunoblots were performed as previously reported (Ying et al., 1989). In some experiments enzyme-digested CRP was reduced with 0.1 M DTT before loading on the gels. Limited proteolysis of CRP

Purified human CRP at 1 mg/ml was digested (37°C 16 hr) with multiple enzymes in 10mM Tris-HCl0.15 M NaCl (TBS), pH 7.4, containing 2 mM calcium. CRP:enzyme ratios (w/w) were 1O:l for alphachymotrypsin (Kinoshita et al., 1989); 5:l for trypsin, plasmin, thrombin and V8 protease; and 2: 1 for elastase. For digestions in the absence of calcium, CRP (1 ml) was dialyzed overnight against 4 1 TBS, and reacted in TBS when trypsin, alpha-chymotrypsin or plasmin were used; TBS adjusted to pH 8.7 when elastase and thrombin were used; and 0.05 M NH,HCOj when V8 protease was used. Denatured CRP was prepared by boiling in the presence or absence of calcium in “sample buffer” (0.0625 M Tri-HCl, pH 6.8, containing 1% SDS and 10% glycerol) for 5 min, and digested by incubation with the enzymes at 37°C for 20 min. All digestion reactions were stopped by boiling 5 min in 1% SDS sample buffer, in certain experiments in the presence of 0.1 M DTT, and the products were separated by SDS-PAGE. The apparent molecular masses of the intact subunit and the fragments were estimated with reference to protein standards on SDS-PAGE in the presence of DTT; residue numbers were estimated based on size, reactivity with

679

Localization of CRP neoepitopes and PMN digestion fragments

anti-CRP peptide mAb and closest enzymesensitive cleavage sites.

confirmed

degradation of CRP with PMN lyso~omalenzymes Digestion of CRP by neutrophil lysosomal enzymes was performed as previously described (Shephard et al., 1988). Briefly, purified human peripheral blood neutrophils were washed and resuspended in 0.15 M PBS. Lysosomal enzymes were released by stimulation (3 min, 37°C) with FMLP in the presence of cytochalasin B. CRP (100 pg) was digested by the cell-free lysosomal enzymes (500~1) (pH 7.4, 3 hr) in the absence of

calcium. Digestion was terminated by the addition of sample buffer containing 1% SDS, 1 M urea and 5% 2-mercaptoethanol; the final CRP concn was 100 pg/1.2 ml. RIEWLTS

Generation and characterizationof new anti-CRP mAb In order to extend the mapping of both native and modified (neo-) CRP epitopes, eight new anti-CRP mAb (indicated in bold letters in Table 1) were generated and characterized. Four of the new mAb were selective for

Tabfe 1. Reactivity of Anti-CRP n&b with native and modified CRP, CRP fragments and CRP peptides by ELISA and Western blot assays” Antibody

Isotype

Calcium

Epitope site

Group I: React with native CRP only 1. I-4H2 2a dep 32-1BI’ 1 dep 2. 15-lD6 2a enh 22-3612’ 1 enh 3. 26-8D8 4. 32-5All 29”6BQ’

1 1 1

ind ind ind

PC

C

Ra

+ + -

-

-

-

+ -

+ -

Group II: React with native CRP strongly and modified CRP weakly 5. 15-2ClO 2a ind 26-lA8 1 ind 32-933’ 2b ind 6. 21-3Al 1 E-enh Group III: React with modified CRP only 7. 29-4810 1 ind 8. 31-6A12 1 ind 9. 26-8ClO 1 ind 10. 29-7AlO 3 ind 11. 26-6B7 1 ind 12. 26-7A8 2b ind 13. 32-9H8 1 ind 14. 33-3812 1 ind 26-2H5’ 1 ind 26-9C9’ 1 ind 15-368 1 ind 26-7C6 1 ind 13-12D7’ 2a ind

1-16 14-24 22-4.5” 41-61” 114-121’ 130-138’ 147-152 199-206 199-206 199-206 199-206 199-206 199-206

-

-

+ + + + $ + + + +

ELISA analyses Native CRP Modified CRP 0.16’ 0.35b 0.07 0.21 0.11 0.05 0.18 0.05 0.06 0.07 24.50 > 20.0 > 20.0 > 20.0 > 20.0 > 20.0 >20.0 > 20.0 > 20.0 > 20.0 > 20.0 > 20.0 > 20.0 > 20.0

Western blot

>40.0 > 40.0

-

>20.0 > 20.0 > 20.0 >40.0 > 40.0

-

3.04d 1.48d 3.57d >40.0

+d +d

0.04 0.004

1.59 2.18 1.69 0.68 0.002 0.16 0.16 0.21 0.26 0.41 0.90

+d + ++++ ++++ ++++ ++++ ++++ ++++ -t+++ ++++ I-+++ ++++ ++++ f+++ ++++

“mAb were assayed in 0.01 M CaCl, and 0.01 M EDTA, respectively, and the reactivities were scored as calcium-dependent (dep), calcium-enhanced (enh), calcium-independent (ind), or enhanced in the presence of EDTA (E-enh). Numerals at the far left refer to demonstrably distinct epitopes of CRP. The mAb identification numbers (in the first column) preceding the dash indicate the fusion number; mAb not previously reported or described (from fusions 29, 31 and 32) are indicated in bold letters. The presumptive sites of the epitopes reactive with the anti-neo-CRP mAb are indicated as amino acid residues numbered from the ~-te~inal of the CRP subunit. Inhibitability by phospho~lcholine (PC), inhibition of ability to react with Clq and activate the complement system (C), and cross-reactivity with rabbit CRP (Ra) are indicated by the “+” sign; “ -” denotes absence of these reactivities; titers in ELISA assays were quantified as the amount (pg/ml mAb Ig) of mAb yielding 50%-maximal reactivity, and Western blot assays of CRP in SDS-PAGE were scored from negative (- ) to maximal (+ + + +), using the affinity purified mAb at concns at least 5-fold greater than those required for maximal reactivity in the ELISA assays. bAssayed using anti-CRP mAb instead of PC to “capture” biotinylat~ CRP on plate; titers of 1Bl and 4H2 in PC capture assays were 0.59 and 1.75 pg/ml, respectively. ‘mAb with binding reactivities not distinguishable from those of mAb designated by number immediately above. dmAb 2C10, lA8 and 9E3 react in ELISA assays with modified CRP subjected to reduction-alkylation at levels of > 20 pg/ml, and with weaker positive strength in Western blots. ePresumptive epitopes not yet confirmed by reactivity with synthetic peptides.

680

S.-c.

YING

native CRP epitopes, with one ( 1B 1) expressing calciumdependent, PC-inhibitable binding to CRP, and four were directed against neo-CRP epitopes. In contrast to the anti-CRP mAb previously raised in our laboratories using booster immunizations with rabbit as well as human CRP (e.g. Table 1, fusion 26) immunizations were performed with human CRP only, and mAb were selected for reactivity with human CRP in the absence of cross-reactivity with rabbit CRP (Table 1, fusions 29, 31 and 32). The eight new mAb, along with the 16 previously described anti-CRP mAb (Ying et al., 1989), were divided into three groups based upon their reactivities with native and modified CRP. Group I was comprised of seven mAb (4H2, lB1, lD6, 3Gl2, 8D8, 5Al1, 6B4) which reacted only with native CRP. Among these, the binding of two mAb (4H2 and 1Bl) was calcium-dependent and PC-inhibitable, indicating reactivity with the PC-combining region; the reactivity of two mAb (lD6 and 3Gl2) with CRP was enhanced in the presence of 2 mM calcium; and the binding of the other three mAb (8D8, 5All and 6B4) was unchanged by either calcium or EDTA. Group II was comprised of four mAb (2Cl0, lA8,9E3 and 3Al) which reacted strongly with native CRP, but also reacted weakly (to an approximately 500- to lOOO-fold lesser degree when reduced and alkylated modified CRP was used) with modified CRP; in contrast to the Group I mAb, these mAb also reacted weakly with CRP in Western blot assays performed in SDS gels. The binding of one of the Group II mAb (3Al) was enhanced by 10 mM EDTA. Group III mAb included 13 antibodies (4Bl0, 6Al2, 8Cl0, 7Al0, 6B7. 7A8, 9H8, 3Hl2, 2H5, 9C9, 3G8, 7C6 and 12D7) each of which reacted only with modified CRP. As indicated in Table 1, the latter six of these mAb may be directed against the same epitope. Only one (of the seven) Group I mAb (8D8), but nine of the 13 group III mAb (8C10, 6B7, 7A8, 3H 12, 2H5, 9C9, 3G8, 7C6 and 12D7), strongly cross-reacted with raCRP in appropriate ELISA assays. Reactivity

of anti-CRP

mAb with synthetic

CRP peptides

We next tested the ability of each of the distinctive anti-neo-CRP mAb to bind to 17 CRP peptides (summarized in Table 2) by dot blot assays. Four mAb with anti-peptide activities were identified. mAb 4BlO bound to peptides l-l 5 and l-16; mAb 6Al2 bound to peptide 14-24; mAb 9H8 bound to peptide 137-152 (as well as to pronase digestion fragment 1477206); and mAb 3Hl2 bound to peptides 199-206 and 201-106 (Table 2). These binding specificities shown in Tables 1 and 2 were confirmed by ELISA inhibition assays. The binding to immobilized CRP of mAb 4BlO was inhibited by peptide l-16, 6A12 was inhibited by peptide 14-24, 9H8 was inhibited by peptide 137-152 and 3Hl2 was inhibited by peptide 199-206, with 0.1, 8.5, 2.4 and 0.18 pgg/ml of these peptides (0.54, 7.20, 1.20 and 0.20 PM), respectively, required for 50% inhibition (Fig. 1). The small amounts of peptide required for inhibition, as shown in

et

ai.

Table 2. Reactivity of anti-CRP mAb with synthetic CRP peptides by dot blot assays Residues l-15 1-16 14-24 23-30 3746 37-58 7&75 77-82 83-90 90-102 109-123 124-136 137-152 160-165 193-200 199-206 201-206

4BlO 6A12 8ClO 7AlO 6B7

7A8

_ _ _ _ _ -

-

_

-

_ -

+ + -

+ _

-

_ _

-

_ _ -

-

_ _ -

Fig. 1, supported the specificities indicated binding studies presented in Table 1. Proteolysis of native and modljied and absence of calcium

CRP

9H8 3H12

+ -

+ +

by the direct

in the presence

Since reactivity with the available CRP peptides served to map only four of the eight anti-neo-CRP mAb, we attempted further mapping of neo-CRP epitopes by use of enzyme-generated fragments. Purified CRP first was digested in the presence of calcium with multiple enzymes including trypsin, alpha-chymotrypsin, elastase, plasmin, thrombin and V8 protease, and analyzed by SDS-PAGE. As previously described (Kinoshita et al., 1989), native CRP proved to be strikingly resistant to digestion at 37°C in the presence of calcium, migrating with an apparent molecular mass of 28.2 kD for the intact subunit in reduced SDS-PAGE (Fig. 2A). Fragmentation was observed only with elastase, which generated an identifiable fragment of approximately 25.0 kD (Fig. 2A). Further, as previously reported when CRP was treated with pronase in the absence of calcium (Kinoshita et al., 1989), fragmentation was observed only after enzyme-treated CRP was exposed to denaturing conditions. By contrast, when native CRP was digested at 37°C in the absence of calcium and analyzed by SDS-PAGE, significant cleavages regularly were observed with three of the enzymes tested (Fig. 2B). Alpha-chymotrypsin generated a major fragment of 18.2 kD and minor fragments (not visible in the figure) of 19.2 and 7.3 kD; elastase generated a major fragment of 25.0 kD and minor fragment of 19.0 kD; and V8 protease generated major fragments of approximately 26.2, 19.7, 16.0, 12.2, 10.7 and 7.6 kD, respectively. Plasmin, thrombin and trypsin did not produce any detectable fragments. Again, the fragments were separated only after enzymetreated CRP was (Table 3) exposed to denaturing conditions.

Localization

681

of CRP neoepitopes and PMN digestion fragments

‘(JO- A. pepttde l-16

4610

‘001

B. poptide 14.24

‘w-

0.

80 -

3H12 loo-

ap

C.

poptldo

137-152

poptldo

199-206

so-

3A12 0

peptide

‘005

.05

.50

5.0

50

(pa/ml)

Fig, 1. ELISA inhibition assays. In (A) mAb 4BlO and 3Hl2 were preincubated with synthetic CRP peptide l-16 at 37°C for 30 min prior to exposure to immobilized CRP; in (B) mAb 6Al2 and 3Hl2 were preincubated with synthetic CRP peptide 14-24 at 37°C for 30min prior to reactivity with immobilized CRP; in (C) mAb 9H8 and 3Hl2 were preincubated with synthetic CRP peptide 137-152 and in (D) mAb 3Hl2 and 6Al2 were preincubated with synthetic CRP peptide 199-206.

Fig. 2. SDS-PAGE analyses of native CRP exposed to various enzymes (37°C 16 hr) in the presence (A) and absence (B) of calcium, and similar SDS-PAGE analyses of CRP denatured by boiling in SDS exposed to the same enzymes (37°C 20 min) in the presence (C) and absence (D) of calcium. The samples were reduced with 0.1 M DTT. In each of the panels, lane 1 = alpha-chymotrypsindigested CRP; lane 2 = alpha-chymotrypsin alone; lane 3 = elastase-digested CRP; lane 4 = elastase alone; lane 5 = trypsin-digested CRP; lane 6 = trypsin alone; lane 7 = protein standards; lane 8 = Staph V8 protease-digested CRP; lane 9 = Staph V8 protease alone; 10 = pIasmin-digested CRP; lane 11 = plasmin alone; lane 12 = thrombin-digested CRP; lane 13 = thrombin alone; and lane 14 = CRP alone.

S.-C. YING ef

682

al.

Table 3. Informative

Enzyme --___ _.._ StphV8 StphVX Elastase StphV8 Elastase Elastase StphV8 Trypsin Trypsin Trypsin StphV8 Trypsin StphV8 Trypsin StphVS Trypsin Trypsin StphV8 Trypsin Trypsin Elastase StphV8 StphV8

CRP cleavage fragments identified by Western blot assays with anti-CRP mAb -CRP Fragment Reactivity with mAb _____-.-.--..~ -.--.- ___~___.~___ .__ State Approx. Est. kD of CRP” residues 4B10 6A12 8CIO 7AIO 6B7 7A8 9H8 3H12 __.-_..__ 28.2 l-206 + + + if + + + 26.4 1-197 + A-SDS + + f + + + 25.8 I-193 + A-SDS + + i+ + + 25.0 I-190 + + + + + + + 19.7 I-147 + A-SDS + + + + f 19.0 1-136 f + + + + + 17.4 1-132 + A-SDS + + + + 15.8 l-130 + A-SDS + + + -t14.5 I-119 + A-SDS + + + + 13.2 l-116 + A-SDS + + + 12.0 8-116 A-SDS + + -I11.8 I-108 + A-SDS + + -t 9.5 l-69 + A-SDS + -t + 7.5 l-60 + A-SDS + + -t 27.5 7-206 A-SDS + + + -t+ + + 21.5 43-206 A-SDS + + + + + 20.2 59-206 A-SDS -t + + + 18.3 70-206 A-SDS + + + + 12.4 109-206 A-SDS + + + + 10.0 123-206 A-SDS + f i9.5 124-206 A-SDS + + + 8.7 142-206 A-SDS -Ii8.2 1477206 A-SDS + + 5.4 164-206 A-SDS +

“Native CRP (--) and CRP denatured by boiling (5 min in 1% SDS; A-SDS) were incubated for 16 hr at 37’C and 20min at 37°C. respectively, with the indicated enzymes; StphV8 = StaphylococcaI V8 protease.

When denatured (boiled in SDS) CRP was reacted with this group of enzymes, more extensive fragmentation regularly was observed both in the presence (Fig. 2C) and absence (Fig. 2D) of calcium; degradation was greater when calcium was present. Alpha-chymotrypsin generated major fragments or‘ 15.7, 14.2, 13.2, 10.7 and 9.7 kD and a minor fragment of 26.4 kD in the absence of calcium. Elastase produced a major fragment of 17.4 kD and a minor fragment of 8.7 kD. Trypsin produced major fragments of 27.5$ 20.2, 18.3, 14.5, 13.2, 12.0 and 9.5 kD, and a minor fragment of 10.0 kD, in the presence of calcium. V8 protease generated major fragments of26.4,25.8, 19.7, 15.8, 12.4, 11.8 and 8.2 kD, and minor fragments of 21.5, 7.5 and 5.4 kD. Plasmin and thrombin did not generate visible fragments at all.

When analyzed by Western blot assays, 23 of the fragments produced by enzymic digestion proved to be informative, i.e. identified distinct fragments which could be localized. These included several minor fragments which were not visible in Fig. 2 by Coomassie protein staining (Fig. 3 and Table 3). The positions of the cleavages were calculated based on molecular size and binding with mAb shown to be reactive with CRP peptides l-16, 137-152 and 199-206, and fragments

l-146 and 147-206 (Table 2), and related to the known amino acid sequence and theoretic cleavage sites of CRP. Representative data are shown in Fig. 3 and the results are summarized in Table 3. Additional neo-epitopes were identified in four distinct regions of the CRP subunit by this approach. For example, 8CIO reacted with a 7.5 kD fragment of V8 protease-digested boiled CRP which also reacted with two mAb (4BlO and 6A12) known to react at the N-terminus (Fig. 3A) and hence was comprised of approximately residues I-60. By contrast, 8ClO failed to react with a 21.5 kD fragment in this digest which reacted well with 3H12 and 9H8, two mAb which react with C-terminal peptides, and hence this fragment was judged to consist of residues 43-206. Since it failed to react with peptides l--16 and 14-24, 8ClO therefore was provisionally judged to react with an epitope which included amino acids in the region of residues 22-45 (granting a three amino acid margin of error) from the N-terminus. Similarly, 7AlO reacted with a 21.5 kD fragment (residues 43-206) of V8 protease-digested boiled CRP (Fig. 3A) but not with a 20.5 kD fragment of trypsin-digested boiled CRP (Fig. 3B) which, because it reacted with mAb 3H12 and 9H8, was judged to include the C-terminus of the subunit (residues 59-206). Thus, 7AlO was provisionally deemed to react with an epitope which included amino acids in the region of

Localization

of CRP neoepitopes and PMN digestion fragments

683

Fig. 3. Western blots of denatured (boiled) CRP digested with V8 protease (panel A), trypsin (panel B) and elastase (panel C), and native CRP digested with elastase (panel D). All digestions were performed in the absence of calcium, and the samples were reduced with 0.1 M DTT before separation on SDS gels. Lane 1, CRP alone; lane 2, enzyme alone; lane 3, protein standard; and lanes 413, enzyme-digested CRP. Lanes 14 were stained with Coomassie Blue, while lanes 5-12 were blotted with mAb 4B10, 6A12, 8C10, 7A10, 6B7, 7A8, 9H8 and 3H12, respectively; lane 13 was blotted with an isotype control mAb. residues 41-61. In a similar manner, mAb 6B7 was provisionally judged to react with an epitope which included residues 114-121 (Fig. 3B) and mAb 7A8 with an epitope including residues 130-138 from the N-terminus (Fig. 3C and D; Table 3). In all, four distinct linear regions with CRP epitopes were confirmed by reactivity with synthetic CRP peptides to include residues in the regions of amino acids 1-16, 1424, 147-152 and 199-206 from the N-terminus, respectively. In addition, four epitopes were provisionally (designated by the “ - ” symbol in Table 1) localized primarily by reactivity with CRP digestion fragments to include amino acids in the regions of residues 2245, 41-61, 114121 and 130-138.

Identljication of CRP fragments produced by PMN enzymes by use of anti-CRP mAb The generation of CRP fragments by digestion with PMN membranes and granule contents has been reported (Shephard et al., 1988, 1989, 1990). We utilized the mAb described above to map the CRP fragments arising through degradation by PMN granule enzymes. Fragments of 21, 16, 14 and 8 kD, as well as smaller fragments (~6 kD), were observed during periods of limited proteolysis (Fig. 4). These were blotted with mAb 4B10, 8C10, 9H8 and 9C9, which recognize epitopes located between residues l-16, 2245, 147-152 and 199-206, respectively (Table 1; Fig. 5). The 21 kD fragment was recognized by mAb 9H8 and 9C9 but not mAb 4BlO or 8C10, indicating that it originated from the C-terminus. The 16 kD fragment reacted with mAb

4B10 and 8ClO but not 9H8 or 9C9, indicating that it originated from the N-terminus. The 8 kD fragment reacted with 9H8 and 9C9 but not 4B10 or 8C10, indicating that it originated from the C-terminus and probably derived from the same cleavage which gave rise to the 16 kD fragment. The 14 kD fragment reacted with 9H8 and 9C9 but not 4B10 or 8C10, indicating that it is a C-terminus fragment. No 10 kD fragment reactive with the N-terminus mAb was observed, perhaps because of further cleavages into smaller peptides which did not transfer in the Western blot procedure. Thus, cleavages seemed to occur at 3, 10 and 16 kD from the N-terminus, corresponding approximately to residues 26, 90 and 137, respectively. When CRP degradation by lysosomal enzymes was allowed to proceed to end-point and the peptides solubilized with TCA to a final concn of lo%, substantial accumulation of soluble CRP peptides of < 14 kD along with loss of recognizable larger fragments of 21, 16 and 14 kD identical to that previously described (Shephard et al., 1988), was observed. Dot blot analysis of this mixture indicated that small peptides reactive with each mAb 4B10, 8C10, 9H8 and 9CP, probably deriving from the larger fragments, were present (data not shown). DISCUSSION In a previous study involving 17 anti-CRP mAb, at least eight epitopes were reported on human CRP, with at least five present on native CRP and at least three present on modified CRP (Ying et al., 1989). These mAb

Fig. 4. Wcstcrn blots of CRP incubated with neutrophil lysosomal cnrymcs 3 hr at 37 C‘. Lanes a (after transfer) and f (before transfer) are autoradiographs of cleaved CRP. while lanes b--e are immunoblots stained with mAb 4BlO. 8ClO. 9H8 and 3Hl2. rcspectivcly.

were prepared by immunizations with both human and rabbit CRP, in order to favor generation of mAb to epitopes expressed in conserved regions of the molecule. All of the anti-neo-CRP mAb generated reacted with rabbit as well as human CRP, as did one of the anti-native CRP mAb. In order to further characterize the epitopes of human CRP, eight new mAb were raised immunizing only with human CRP and selecting for

III

PMNl

V8

I

I

I

I

I

1

I

I

I

PM;JP

+&

I I I

1

I PM;3 --

I 1 I

V’S

Fig. 5. Schematic representation of the presumptive localization of epitopes identified by the anti-neo-CRP mAb on the CRP subunit; the dotted arrows and breaks in the subunit refer to selective informative cleavages induced by trypsin (tryp), Staphylococcal V8 protease (V8) and PMN enzymes (PMN I-PMN3). and the darkened areas indicate peptides with which certain mAb react.

specificity for human CRP; these were found to involve four anti-native and four anti-neo-CRP specificities. The 24 mAb were divided into three groups, based upon their reactivity with native and modified CRP. Group I included mAb which reacted with calcium-dependent, PC-inhibitable, exclusively native mAb epitopes or with calcium-enhanced or calcium-independent native CRP epitopes not inhibited by PC and not retained on modified CRP. Group II reacted with native CRP epitopes which were retained (although with greatly decreased reactivity) on modified CRP. Group III mAb reacted exclusively with sequence-determined neo-CRP epitopes not expressed on the native CRP molecule. This division into three groups, rather than four groups as previously presented (Ying et d.. 1989) reflects the merging of the previous groups III and IV into a single group of relatively defined linear neo-CRP epitopes on modified CRP. The CRP sequence-determined neo-CRP epitopes reactive with Group III mAb were mapped by defining the binding of these mAb to synthetic CRP peptides and enzyme-generated fragments. Dot blot and ELISA analyses using synthetic peptides showed that mAb 4BlO binds to peptides 1 15 and l-16; mAb 6A12 binds to peptide 14-24; mAb 9HX binds to peptide 137-l 52; and mAb 3Hl2, 2H5, 9C9, 3G8, 7C6 and 12D7 bind to peptides 1999206 and 2Oll206. CRP fragments next were generated by digestion with multiple enzymes. Native CRP proved to be markedly resistant to enzymic digestion in the presence of calcium, but somewhat more

Localization

of CRP neoepitopes and PMN digestion fragments

enzyme-sensitive in the absence of calcium (Kinoshita et al., 1989). However, modified or denatured CRP (e.g. CRP boiled in SDS) proved to be markedly enzymesensitive both in the presence and absence of calcium; this preferential sensitivity of modified CRP to digestion might be a significant factor in the catabolism of this pentraxin. Twenty-three informative fragments were identified by immunoblotting the products of the various digestion mixtures, and served to define eight distinct antigenic regions dispersed along the CRP subunit (Fig. 5). These were localized to include amino acids between residues 1-16, to which mAb 4BlO binds; residues 14-24, to which 6A12 binds; residues 22-45, to which 8ClO binds; residues 4161, to which 7AlO binds; residues 114-121, to which mAb 6B7 binds; residues 130-138, to which mAb 7A8 binds; residues 147-152, to which mAb 9H8 binds; and residues 199-206, to which mAb 3H12 and several additional mAb bind. These eight mapped and presumptive linear epitopes correspond well with regions of predicted CRP antigenicity based on CRP structure, as predicted by the Protein Analysis Module of the Sequence Analysis Software Program of the Genetic Computer Group (Devereux et al., 1984). However, we found that some of mAb, for reasons not yet clear, did not bind to synthetic CRP peptides involving these presumptive binding sites. For example, 6B7 did not bind to peptide 109-123 and 7A8 did not bind to peptide 124-136. Perhaps these peptides lacked critical secondary structure or key adjacent residues required for binding of the appropriate mAb to occur. Neutrophil and macrophage enzymes are known to digest CRP, with generation of biologically active peptides (Buchta and Fridkin, 1986; Robey et al., 1987; Shephard et al., 1988, 1989, 1990). The mAb defined in the present study permitted identification of the fragments obtained when CRP is digested by PMN enzymes. Thus, C-terminal fragments of 21, 14 and 8 kD, and an N-terminal fragment of 16 kD, were identified, resulting from cleavages approximately 3, 10 and 16 kD (at approximately residues 26, 90 and 137, respectively) from the N-terminus of the molecule (Fig. 4). A cleavage 16 kD from the N-terminus was previously appreciated upon digestion of native CRP with Pronase or Nagarse protease (Kinoshita et al., 1989) as well as by several of the isolated enzymes utilized in the present study. Low mol. wt peptides which migrate to the front of gel were evident, and were inferred to arise through cleavages at protease-sensitive sites 3 kD from the N-terminus and from continued degradation of the larger fragments. Fragments from the N-terminus seemed particularly prone to continued degradation, since the 10 kD Nterminal fragment from the cleavage which gave rise to the C-terminus 14 kD fragment was not recognized by Western blots of SDS-PAGE. Certain of these fragments have been reported to modulate neutrophil function (Shephard et al., 1988, 1990) including peptides homologous with the amino acid sequence between residues 70-103 and 200-206. Since protease-sensitive cleavage sites occur annroximatelv d at residues 90 and ._

685

137, we would expect functional peptides containing residues 70-90 and 201-206 at their C-termini, which could account for the bioactivity of the CRP peptides c 14 kD, to be generated. Comparison of the sequences of human CRP (Oliveira et al., 1979; Lei et al., 1985; Woo et al., 1985) and rabbit CRP (Wang et al., 1982; Hu et al., 1986), together with the data presented in Tables l-3 and Fig. 3, suggest a further localization of the binding sites of mAb 4B10, 8C10, 6A12, 7A10, 6B7 and 7A8 on the human CRP subunit (Fig. 5). Since 4BlO bound to peptides 1-15 and 1-16 but not to raCRP, which is identical to human CRP at residues 7-16, residues l-6 probably are involved in its binding; this is consistent with the result that 4BlO does not bind to fragment 7-206 (Fig. 3B, Table 3). Similarly, 8C10, which was presumptively localized to include residues 2245, cross-reacts with raCRP; since rabbit and human CRP have identical sequences between residues 28-37, it seems likely that this area is involved in the 8ClO-reactive site. Since 7AlO presumptively reacted with an epitope which includes residues in the region of amino acids 41-61 but did not bind to raCRP or react with peptide 37-58, it seems likely that it reacts with an epitope which includes residues 57-61. Lastly, 6B7 and 7A8, which were presumptively localized to the regions of residues 114-121 and 130-138 of human CRP, respectively, cross-react with reCRP; identical sequences in human and rabbit CRP at residues 110-115, 117-125 and 130-137 are consistent with this provisional localization. In addition to the eight neo-CRP epitopes described above, at least six epitopes have been identified to date on native CRP utilizing the mAb described herein, including: (a) an epitope whose expression is inhibited in the presence of PC; (b) an epitope which shows enhanced expression in the presence of calcium; (c) an epitope which is not influenced by the presence of calcium; (d) an epitope which also is found on rabbit CRP and in that species is observed only in the presence of calcium; (e) an epitope which shows enhanced expression in the presence of EDTA; and (fl an epitope which is retained on modified and denatured CRP. In addition, it recently was found that the native epitope also found on rabbit CRP (which is reactive with mAb 8D8) is involved in binding of human CRP to Clq and CRP-induced complement consumption (Jiang et al., 1991). None of the mAb described herein, including anti-native-CRP mAb 1Bl and 4H2 which are directed to the PC-binding site, reacted with the PC-binding peptide recently described by Swanson and Mortensen (1990); these mAb also did not show strong reactivity with the PC-binding CRP of the rabbit and horseshoe crab (unpublished observations). It is of interest that different biological responses seem to be induced by native CRP displaying only native CRP epitopes and modified CRP preferentially expressing neo-CRP epitopes. For instance, C consumption but not leukocyte activation has been induced by CRP complexes which display native conformational and not neo-CRP linear epitopes (Jiang et al., 1991), while

686

S-C. YtNC el al.

modified forms of CRP expressing neo-CRP epitopes have been capable of activating leukocytes and platelets but not the C system (Potempa et a/., 1988). We expect that the anti-CRP mAb described and mapped herein will continue to serve as useful tools for the identification of both the linear and the conformational regions of CRP responsible for its ability to initiate these various biological responses, and in addition, help to elucidate the chemical mechanisms involved. Acknowledgements-We

appreciate the generous gifts of CRP peptides 23-30, 109-123 and 199-206 from Dr Frank A. Robey of the National Institute of Dental Research; 137-152 from Dr Richard A. Houghton of the Scripps Clinic and Research Institute; and 47-63 from Dr Richard F. Mortensen of Ohio State University. We also appreciate the capable typing assistance of MS Mary Rolfe-Shaw.

REFERENCES Buchta R. and Fridkin M. (1986) Synthetic peptides from C-reactive protein containing tuftsin-related sequences. Peptides 7, 961-968. Cabana V. G., Gewurz H. and Siegel J. N. (1982) Interaction of very low density lipoproteins (VLDL) with rabbit Creactive protein. J. Immun. 128, 2342-2348. Claus D. R., Osmand A. P. and Gewurz H. (1976) Radioimmunoassay of human C-reactive protein and levels in normal sera. J. Lab. r&z. Med. 87, 120-128. Devereux J., Haeberli P. and Smithies 0. (1984) A comprehensive set of sequence analysis programs for the Vax. Nucleic Acids Res. 12, 387-395.

Hu S.-I., Miller S. M. and Samols D. (1986) Cloning and characterization of the gene for rabbit C-reactive protein. biochemistry 25, 7834-7839.

Jiang H., Siegel J. N. and Gewurz I-I. (1991) Binding and complement activation by C-reactive protein via the collagen-like region of Clq and inhibition of this reaction by monoclonal antibodies to CRP and Clq. J. Immun. 146, 23242330. Kaplan M. H. and Volanakis J. E. (1974) Interaction of C-reactive protein with the complement system. I. Consumption of human complement associated with the reaction of C-reactive protein with C-polysaccharides and with the choline phosphatides, lecithin and sphingomyelin. J. Immun. 112, 2135-2147. Kew R. R., Hyers T. M. and Webster R. 0. (1990) Human C-reactive protein inhibits neutrophil chemotaxis in vitro: Possible implications for the adult respiratory distress syndrome. J. Lab. c/in. Med. 115, 339-345. Kinoshita C. M., Ying S-C., Hugli T. E., Siegel J. N., Potempa L. A., Jiang H., Houghten R. A. and Gewurz H. (1989) elucidation of a protease-sensitive site involved in the binding of calcium to C-reactive protein. Biochemistry 28, 9840-9848.

Laemmli U. K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680-685. Lei K. J., Liu T., Zen G., Soravia E., Liu T.-Y. and Goldman

N. D. (1985) Genomic DNA sequence for human C-reactive protein. J. biol. Chem. 260, 13,377-13,383. Oliveira E. B., Gotschlich E. C. and Liu T.-Y. (1979) Primary structure of human C-reactive protein. J. biol. Chem. 2% 489-502.

Osmand A. P., Friedenson B., Gewurz H., Painter R. H., Hoffman T. and Shelton E. (1977) Characterization of C-reactive protein and the complement subcomponent Clt as homologous proteins displaying cyclic pentameric symmetry (pentraxins). Proc. natn. Acad. Sci. U.S.A. 74, 739-743. Potempa L. A., Maldonado B. A., Laurent P., Zemel E. S. and Gewurz H. (1983) Antigenic, electrophoretic and binding alteration of human C-reactive protein modified selectively in the absence of calcium. Molec. fmmun. 20, 1165-l 175.

Potempa L. A., Siegel J. N., Fiedel B. A., Potempa R. T. and Gewurz H. (1987) Expression, detection and assay of a neoantigen (neo-CRP) associated with a free, human Creactive protein subunit. Mofec. Immun. 24, 531-541. Potempa L. A., Zelier J. M., Fiedel B. A., Kinoshita C. M. and Gewurz H. (1988) Stimulation of human neutrophils, monocytes, and platelets by modified C-reactive protein (CRP) expressing a neoantigenic specificity. Inf?ummation 12, 391-40s.

Robey F. A., Ohura K., Futaki S., Fuji N., Yajima H., Goldman N., Jones D. and Wahl S. (1987) Proteolysis of human C-reactive protein produces peptides with potent immunomodulating activity. J. bid. Chem. 262, 7053-7057. Shephard E. G., Anderson R., Beer S. M., Jansen C. E., Rensburg V. and de Beer F. C. (1988) Neutrophil lysosomal degradation of human CRP: CRP-derived peptides modulate neutrophil function. Clin. exp. Immun. 73, 139-145. Shepard E. G., Anderson R., Rosen O., Myer M. S., Fridkin M., Strachan A. F. and de Beer F. C. (1990) Peptides generated from C-reactive protein by a neutrophil membrane protease. Amino acid sequence and effects of peptides on neutrophil oxidative metabolism and chemotaxis. J, Immun. 145, 1469-1476.

Shephard E. G., Beer S. M., Anderson R., Strachan A. F., Nel A. E. and de Beer F. C. (1989) Generation of biologically active C-reactive protein peptides by neutral protease on the membrane of phorbol myristate acetate-stimulated neutrophils. J. Immun. 143, 2974-2981. Siegel J. N., Rent R. and Gewurz H. (1974) Interaction of C-reactive protein with the complement system. I. Protamine-induced consumption of complement in acute phase sera. J. exp. Med. 140, 631-647. Swanson S. J. and Mortensen R. F. (1990) Binding and immunological properties of a synthetic peptide corresponding to the phosphorylcholine-binding region of C-reactive protein. Molec. Immun. 27, 679-687. Wang C. M., Nguyen N. Y., Yonaha K., Robey F. A. and Liu T.-Y, (1982) Primary structure of rabbit C-reactive protein. J. bid. Chem. 257, 13,610-13,615. Woo P., Korenberg J. R. and Whitehead A. S. (1985) Characterization of genomic and complementary DNA sequence of human C-reactive protein, and comparison with the complementary DNA sequence of serum amyloid P component. J. biol. Chem. 260, 13,38413,388. Ying S-C., Gewurz H., Kinoshita C. M., Potempa L. A. and Siegel J. N. (1989) Identification and partial characterization of multiple native and neoantigenic epitopes of human C-reactive protein by using monoclonal antibodies. J. fmmun. 143, 221-228.

Ying S.-C., Shephard E., deBeer F. C., Fridkin M., Siegel J. N., Harris D. and Gewurz H. (1990) Identification of trypsin- and neutrophil (PMN) enzyme-generated fragments of CRP utilizing new mAb and peptides. FASEB J. 4, A2226

Localization

of CRP neoepitopes and PMN digestion fragments

Zahedi K., Tebo J. M., Siripont J., Klimo G. F. and Mortensen R. F. (1989) Binding of human C-reactive protein to mouse macrophages is mediated by distinct receptors. J. Immun. 142, 23842392. Zelfer J. M., Landay A. L., Lint T. F. and Gewurz H. (1986~) Aggregated C-reactive protein binds to human polymor-

687

phonuclear leukocytes and potentiates Fc receptor-mediated chemiluminescence. .l. Lab. clin. Med. 108,567-575. Zeller J. M., Landay A. L., Lint T, F. and Gewurz H. (1988) Enhan~ment of human peripheral blood rnon~~e respiratory burst activity by aggregated C-reactive protein. J. Lark. B&l. 40, 769-783,