Molecular immunology, Vol. 29, NO. 3, pp. 303-312, 1992 Printed in Great Britain.

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

CHARACTERIZATION OF INTERACTIONS INVOLVING ANTI-~ETATY~~ ANTIBODIES AND IMMUNE COMPLEXES* KARLA

Department

M. WEIDNER

of Microbiology,

and EDWARD W. Voss,

University

of Illinois,

Urbana,

JR?

IL 61801, U.S.A.

(First received 25 March 1991; accepted in revised form 8 July 199 I) Abstract-Immunizations of high affinity anti-fluorescein monoclonal antibody 4-4-20 affinity labeled with fluorescein 5-isothiocyanate into a rabbit elicited antibodies specific for the liganded conformation of 4-4-20 (termed “anti-metatype” antibodies). Reaction of liganded 4-4-20 with anti-metatypc antibodies caused significant delay (up to 23-fold) in the rate of dissociation of fluorescein ligand from the active site. In this study, structural analogues of Buorescein, including fluorescein S-isothiocyanate, fluorescein 4-isothiocyanate, 5-djchlorotriazinyl am~no~uorescein and %carboxyfluorescein, were bound by monoclonal antibody 4-4-20 and anti-metatype antibody reactivity was observed through delay in the dissociation rate of ligand from Mab 4-4-20. Significant delays (ranging from 5- to 242-fold) were observed for all structural analogues examined indicating that 4-4-20 maintained similar but not necessarily identical conformations upon binding fluorescein structural analogues. Additionally, fluorescein 5-isothiocyanate and fluorescein 6-isothiocyanate were conjugated to carrier molecules of increasing mol. wt (ranging from 225 to 14,600 D) in an attempt to sterically interfere with “metatopes” at the mouth of the active site and localize regions of anti-metatype antibody binding. These fluorescein-conjugated compounds were reacted with 4-4-20, and binding of anti-metatype antibodies delayed dissocation rates from 24- to > l500-fold. These results indicated that the mechanism whereby anti-metatype antibodies delay the release of fluorescyl ligands from the active site probably does not solely involve steric hindrance of the ligand due to binding of anti-metatype antibodies at the mouth of the active site. Studies with 4-4-20 Fab fragments and a single chain derivative of 4-4-20 (consisting of the variable regions tethered by a 14 amino acid linker) indicated that anti-metatype reactivity was specific for the immunoglobulin variable region.

INTRODUCTION Immunologically

distinct

conformational

states

have

been demonstrated in antibody variable regions upon ligand binding (Voss et al., 1988, 1989; Weidner and Voss, 1991). The unique liganded conformational state has been molecule.

defined Newly

as the “metatype” of formed epitopes exposed

an antibody after ligand

*This work was supported by a grant from The Biotechnology Research Development Corporation, Peoria, Illinois. tTo whom correspondence should be addressed: Department of Microbiology, University of Illinois, 131 Burrill Hall, 407 South Goodwin Avenue, Urbana, IL 61801, U.S.A. Abbreviations: BPB, bradykinin potentiator B; 5-CFL, 5carboxyfluorescein; %DCTAF, 5dichlorotriazinyl aminofluorescein; DMSO, dimethyl sulfoxide; EDTA, ethylenediamene tetraacetic acid; ELISA, en~me-linked immunosorbent assay; Fab, 50 kd antigen binding fragment derived by papain digestion of an immunoglobulin molecule; Fc, 50 kd constant region fragment derived by papain digestion of an immunoglobulin molecule; FDS, fluorescein disodium salt; FITC I, fluorescein 5isothiocyanate; FITC II, ffuorescein 6-isothi~yanate; FL, fluorescein; Ig, immunoglobulin; Mab, monoclonal antibody; MEK, methyl ethyl ketone; Met, metatype; PCMB, p-chloromercuribenzoic acid; Q,,,, , maximum fluorescence quenching expressed as a percent; RNAse, ribonuclease; SCA, single chain antibody.

binding have been termed “metatopes,” and antibodies specific for the metatypic state have been labeled “antimetatype” antibodies (Voss et al., 1988). Anti-metatype antibodies react with liganded sites in a manner distinguishable from anti-idiotype interactions. Additionally, the reaction of polyclonal anti-metatype antibodies with liganded antibody results in significant delay in the rate of ligand dissociation from the active site. This delay results in an apparent enhancement of antibody affinity through effects on the liganded complex (Voss et al., 1988; Weidner and Voss, 1991). Recently, other investigators have observed a similar phenomenon in various systems. In such studies, the affinity of a monoclonal antibody for the homologous antigen was enhanced through binding of another monoclonal antibody recognizing a different epitope on the same antigen (Erlich et al., 1982; Holmes and Parham, 1983; Howard et al., 1979; Laibeuf et al., 1981). However, the metatype system is unique in that it appears to be sensitive to conformational changes that occur within the antibody variable domains upon ligand binding. Mechanisms of protein--protein interactions between polyclonal anti-metatype antibodies and liganded antibody molecules have not been determined. To better understand the basis of anti-metatype binding and the mechanisms whereby delayed dissociation rates result, the fluorescein hapten system was employed. Murine monoclonal anti-fluorescein antibody 4-4-20 (IgG,, , K.

K. M. WEIDNER and E. W. Voss. JR

304

Ka = 1.7 x 10” M-‘) was selected because the ligand binding properties have been studied extensively (Kranz et al., 1982; Bates et al., 1985), the ligand-antibody complex crystallized (Gibson et al., 19X8), and the atomic structure resolved to 2.7 Angstroms (Herron et al., 1989). Previous studies suggested the mechanism of the delayed dissociation rate of fluorescyl ligand from the antibody active site was due to steric hindrance by anti-metatype antibodies binding directly at the “mouth” of the antibody active site (Weidner and Voss, 1991). A refined Mab-4-4-20 crystal structure indicates that the 5-position of the phenyl carboxyl ring of fluorescein is not directly involved in bonding interactions with antibody active site amino acid residues. Thus, substitutions in the 5-position of the fluorescein Iigand do not directly influence specificity. In fact, substituents at the 5-position become exposed to the aqueous phase surrounding the ligand-antibody complex. Therefore, increasing molecular weight substitutions at this position were employed in an attempt to mask metatopes at the mouth of the antibody active site and to distinguish between anti-metatype antibodies reacting directly with the mouth of the antibody active site and those reacting with determinants peripheral or distal to the active site. Additionally, ligand analogues and isomers of fluorescein bound by monoclonal antibody 4-4-20 were studied, and effects of analogue binding on the conformation of the antibody variable region were observed through reactivity with anti-metatype antibodies. Structural analogues of fluorescein and Auorescein compounds substituted at the 5-position used in these studies were selected on the basis of intrinsic fluorescent properties and monoclonal antibody 4-4-20 binding affinity. These compounds of defined structural and chemical properties were employed in an attempt to describe interactions involved in the formation of the anti-metatype-liganded antibody ternary complex. MATERIALS

AND METHODS

Monclonal anti-fluorescein antibody Hybridoma 4-4-20 mediated fusion of splenocytes from a with fluorescein-keyhole in Freund’s complete

was derived through a chemically Sp2jO-Ag14 myeloma cells with BALB/e mouse hyperimmunized limpet hemocyanin emulsified adjuvant (Kranz et NI., 1982).

into 0.1 M P04, pH 8.5. Affinity labeling of the active site was achieved with the use of Ruorescein isothiocyanate isomer I (FTTC I, Molecular Probes, Inc.). FITC I was dissolved and added in a 1: 1 stoichiometric molar ratio (FITC: antibody active site), incubated at room temp. for 2-3 hr, then dialyzed extensively into 0.1 M PO,, pH 8.0 to remove non-covalently bound FITC I.

~rnrnun~~~t~on~ with ignite

i~beled 4-4-20

Affinity labeled 4-4-20 was “weakly” emulsified in Freund’s Complete Adjuvant (Difco) and immunized into a New Zealand white rabbit. Five separate injections approx. 4 weeks apart led to the production of polyclonal rabbit anti-metatype antibodies (Weidner and Voss, 1991).

Purljication of xenogenic anti-metatype antibodies Polyclonal rabbit anti-Met was purified as described (Weidner and Voss, 1991). Briefly, a purified gamma globulin fraction was adsorbed with 4-4-20-Sepharose to remove potential anti-constant region and anti-idiotype activity, followed by adsorption with fluorescein-Sepharose to remove any anti-metatype activity due to unexpected conjugation of FITC to lysine residues external to the antibody active site.

FITS corrugation Fluorescein isothiocyanate (isomers I and II) was conjugated to the following molecules (see Fig. 1): N-acetyl-L-lysine (Sigma Chemical Co.), bradykinin potentiator B (Peninsula Laboratories; Kato and Suzuki, t971), lactalbumin (Sigma Chemical Co.), and ribonuclease (Sigma Chemical Co.). FITC-N-acetyi-L-lysine and FITC-bradykinin potentiator B were prepared by incubation each with excess FITC in carbonate buffer for 2-3 hr. Fluorescein-labeled compounds were purified by preparative thin-layer chromatography on silica gel using methyl ethyl ketone (MEK) saturated with 0.1 M acetate buffer (pH 5.0) as the solvent. Visible bands containing the labeled reagents were harvested from the plate and solubilized in 0.1 M PO,, pH 8.0. FITC-lactalbumin and FITC-RNAse were prepared using a similar procedure. FITC was dissolved and added (2-3 mol FITC: mole protein) to 3.0 ml of both lactalbumin and ribonuclease to which 50 mg of KzCO1 had been added. These conjugates were dialyzed in 0.1 M PO, to remove unreactive FITC.

Preparation of imm~nogen

Synthesis qf p-chi,romercuribe~lzoure deritzztiue qf F1TC

Ascites fluid containing monoclonal antibody 4-4-20 was delipified with dextran sulfate followed by precipitation of the gamma globulin fraction with 50% saturated ammonium sulfate. The gamma globulin fraction was dissolved in 0.1 M PO,, pH 8.0 and dialyzed in the same buffer to remove residual ammonium sulfate. The gamma globulin fraction was incubated with fluoresceinSepharose (prepared as described by Sundberg and Porath, 1974) and Mab 4-4-20 was eluted with 0.23 M glycine-HCI, pH 2.6. Purified Mab 4-4-20 was dialyzed

Potassium carbonate (3.8 mg) and FITC I (3.9 mg) were added to 7.5 mg of glycine dissolved in 1.Oml dH,O and reacted for I hr at room temp. Reaction products were analyzed by thin-layer chromatography using 0.1 N NaOH saturated methyl ethyl ketone. The FITC-glycyl derivative was harvested from the silica gel plate and the pH adjusted to 5.5 with the addition of 3.0 M sodium acetate. p-Chloromercuribenzoic acid (5.0 mg) was added and the reaction incubated for I hr at room temp with stirring. Reaction products were analyzed and

Anti-Met

interactions

305 FL H

7

YL H-y-$-/+(CH,),-Y-COO HsH

YL NH-C-(CH&

~JH-~-~-HcJ-~-COO~ IAH

YH F=O

-CH3

A

H;H

CH3

N-tetradecanoyl-

doo-

FITC-N-acetyl-L-lysine*

FITC-PCMB

MW = 577

aminofluorescein

MW = 787

MW = 557

FL

FL

I FL

H-N-;-N-(74).,

H-ljl-C-N-(FY)4

I pE-G-L-P-P-R-P-K-I-P-P

FITC-bradykinin MW =

s

potentiator

B*

lactalbumin

:

FITC-lactalbumin*

FITC-ribonuclease*

MW r* 15,200

1571

ribonuclease

MW =

14,400

Fig. 1. Structures of fluorescein-conjugated compounds with corresponding mol. wts. The asterisk (*) denotes those compounds to which both fluorescein 5-isothiocyanate and fluorescein 6-isothiocyanate were conjugated. (FL = fluorescein isothiocyanate.)

obtained using preparative thin-layer chromatography on silica gel with a NaOH-MEK solvent system. A compound distinguishable from the FITC-glycyl derivative remained at the origin (Rr = 0) and was harvested and chromatographed twice (successively) in the NaOHMEK solvent, and finally eluted with dH>O. Under the conditions of thin-layer chromatography employed (2 parts isoamyl alcohol: 1 part DMSO), the compound migrated with an R, of 0.14, while the unreactive p-chloromercuribenzoic acid remained at the origin. The p-chloromercuribenzoate derivative of FITC (FITC IPCMB) was harvested from the silica gel and eluted with dH,O.

affinity of 5 x 10’ Mm’, about 2-3-fold less than 4-4-20 whole immunoglobulin molecule (Bedzyk et al., 1990) or Fab fragments. 4-4-20 Fah ,fragments

Other analogues (see Fig. 2) used in these studies included: fluorescein isothiocyanate isomer I, fluorescein isothiocyanate isomer II, 5-(4,6-dichlorotriazinyl)aminofluorescein, 5-carboxyfluorescein, and N-tetradecanoyl aminofluorescein (all obtained from Molecular Probes Inc.). Compounds were dissolved in 50 mM PO,, pH 8.0, when possible, or in 100% ethanol followed by dilution into 50 mM PO,, pH 8.0.

Fab fragments of Mab 4-4-20 were obtained by digestion with papain using the following procedure. Affinity purified Mab 4-4-20 was dialyzed into 50 mM Tris, 150 mM NaCl, 2 mM EDTA at pH 7.5 (Oi and Herzenberg, 1979). Mercuripapain (Worthington Biochemical Corp.) in the same buffer was activated at 37°C for 15 min with the addition of dithioerythritol (Sigma Chemical Co.) to a final concn of 1 mM. Activated papain was was added to Mab 4-4-20 at an enzyme: protein ratio of 1: 100 (w/w) and incubated at 37°C for 15 min. The reaction was terminated with the addition of 1 mM dehydroascorbic acid (Aldrich Chemical Co.) and dialyzed into 50 mM PO, buffer, pH 8.0, Fab fragments of 4-4-20 were purified with protein A-Sepharose (Zymed Laboratories) and Fc contamination was determined by reactivity of HRP-protein G (Zymed Laboratories) with adsorbed 4-4-20 Fab fragments in solid phase ELISA.

Spectral

Fluorescence

Fluorescein

structural

analysis

analogues

of Juorescein

and structural

analogues

Visible absorption spectra of fluorescein and all structural analogues were obtained using a Beckman DU-64 spectrophotometer. Single chain antibody

of 4-4-20

Single chain antibody of 4-4-20 was engineered by joining the variable heavy chain to the variable light chain through a 14 amino acid linker designated “212” (G-S-T-S-G-S-G-K-S-S-E-G-K-G). The protein was expressed in E. coli and purified as described (Bird et al., 1988; Denzin et al., 199 1). SCA 4-4-20/212 possessed an

quenching

assay

Fluorescence quenching measurements were performed using an Aminco Bowman spectrophoto(maximum fluorescence fluorometer, and Q,,, quenching) was determined as described by Watt and Voss (1978). Dissociation

rate assay

Dissociation rates defining the interaction of fluorescein ligand and structural analogues (FITC I, FITC II, 5-carboxyfluorescein, 5-dichlorotriazinyl aminofluorescein, and FITC-labeled molecules) bound to Mab 4-4-20

K. M. WEIDNERand E. W. Voss. JR

306

H”c3fjlcgo

H”yjfLgo

coo Fluorescein MW =

5-Carboxyfluorescein

331

MW =

376

N=C=S

Fluorescein

Fluorescein

5-isothiocyanate

MW =

6-isothiocyanate

MW =

389

CIyN)-~~

389

0: I coo-

Cl 5-DCTAF MW =

Fig. 2. Fluorescein

structural

analogues

were determined at 2°C and 20°C as described (Herron, 1984). Additionally, dissociation rates were performed in the presence of a saturating concentration of the rabbit anti-metatype reagent which was determined by titration of anti-Met whole immunoglobulin and Fab fragments until maximum delay of the dissociation rate was observed (Fig. 3). This concn of anti-Met antibodies (whole Ig) was used throughout the kinetic studies. Dissociation rate measurements in the absence and presence of anti-metatype were based on the equilibria:

495

with corresponding

mol. wts.

,O;

k, Ab + ligand + anti-Met

Ab . ligand + anti-Met

1 k,

It

k: Ab . anti-Met

+ ligand

Ab

1

ligand . anti-Met

k;

0

molar

where k, and kf are the second order association constants (M-’ set’) for fluorescyl ligand, k, and k! are the first order dissociation constants (set-‘), and k, > kf . The assay was performed by the addition of 1.0 ml of Saminofluorescein (1 x 10Ph M), a non-fluorescent

Fig. 3. Titration

excess

of

active

sites

of anti-Met and the effect on the dissociation lifetime of the 4-4-20-fluorescylkligand complex at 2°C. Mab 44-20 was 90% saturated with fluorescyl ligand and titrated with either anti-Met Ig (closed triangles) or anti-Met Fab fragments (closed circles).

Anti-Met interactions analog of fluorescein, to 100~1 of antibody (5.6 x 10m7 M) with 90% active sites filled. Increase in fluorescence was monitored as a function of time and data were plotted by first order dissociation kinetics. Dissociation rates were expressed as z, the lifetime of the Ab ligand complex, which is equal to l/k1 .

antibodies typically produce such shifts in the ligand’s absorption maximum upon binding (Watt and Voss, 1977; Bates et al., 1985). Collectively, these results confirmed the ability of 4-4-20 to bind each of the fluorescyl analogues tested. Dissociation

RESULTS

Analysis

of xenogenic

anti-metatype

antibodies

Five separate immunizations of affinity labeled Mab4-4-20 over a 5 month period elicited antibodies specific for the Ig liganded state. Solid phase ELISA confirmed reactivity with the liganded conformation and demonstrated little or no reactivity with the non-liganded state (Weidner and Voss, 1991). Results of a competitive inhibition ELISA demonstrated that only soluble liganded Mab 4-4-20 inhibited binding of antimetatype antibodies to absorbed liganded 4-4-20 Fab fragments (Weidner and Voss, 1991). FITC

conjugation

Spectrophotometric analysis indicated that each fluorisothiocyanate derivative (N-acetyl-L-lysine, escein bradykinin potentiator B, lactalbumin and ribonuclease) had an absorption maxima of 491-492 nm, indicative of fluorescein. Analysis

of p-chloromercuribenzoate

derivative

of FITC

A compound possessing the characteristics of a pchloromercuribenzoate derivative of FITC I (FITC IPCMB) was harvested by thin-layer chromatography and the mercury content analyzed by atomic adsorption. Quantitative results indicated the presence of 1 Hg atom per mole of fluorescein. Fluorescein content was analyzed spectrophotometrically and verified by an absorption maxima of 493 nm at a basic pH (A,,,,, of fluorescein is 492 nm). Normalized to equal concns of fluorescein the FITC I-PCMB compound showed a fluorescence quantum yield (0) of 0.85 relative to fluorescein. FITC I-PCMB and fluorescein were titrated with Mab4-4-20 in a fluorescence quenching assay and demonstrated superimposable quenching curves. Fluorescence quenching indicated specific binding of the FITC I-mercure-compound by Mab 4-4-20, and superimposition of the quenching curves indicated the FITC I conjugated to PCMB was as accessible for binding as unconjugated fluorescein. Mah 4-4-20 binding to fluorescein structural and fluorescein -conjugated analogues

analogues

Binding of Mab 4-4-20 to various structural analogues (Figs 1 and 2) was determined by fluorescence analyquenching in Q,,, assays and spectrophotometric ses. Significant fluorescence quenching was observed for each of fluorescyl-compounds upon addition of 4-4-20 (data not shown). A substantial bathochromic shift (lo-20 nm) in the absorption maximum of each of the compounds was also observed upon binding 4-4-20. Both monoclonal and polyclonal anti-fluorescein

307

rate assays

Addition of saturating concentrations of anti-Met antibodies caused significant delay (up to 23-fold) in the dissociation rate of fluorescyl ligand from Mab 4-4-20 (Voss et al., 1988; Weidner and Voss, 1991). A similar effect was observed (to varying degrees) with liganded anti-fluorescein antibodies that were idiotypically related to Mab 4-4-20 when reacted with anti-metatype antibodies (Weidner and Voss, 1991). “Saturation” was determined by a sequential titration of anti-Met antibodies coupled to an observed increase in the lifetime of the 4-4-20-fluorescyl ligand complex until maximum delay in the dissociation rate was achieved (Fig. 3). Titrations with both anti-Met gamma globulin and anti-Met Fab fragments showed only slight differences in their ability to delay the dissociation rate. It is, therefore, not likely that the delayed dissociation rate is due to antibody bivalency or polyvalent complex formation (Holmes and Parham, 1983; Posner et al., 1991). Due to difficulty in obtaining adsorbed rabbit antimetatype fractions concentrated enough to saturate active sites and delay the dissociation rate, purified unadsorbed gamma globulin fractions of the anti-Met reagent were used routinely. Before use in kinetic studies, the gamma globulin fraction was tested for anti-fluorescein activity in a Q,,, assay. The gamma globulin fraction did not quench fluorescein fluorescence, even at high concentrations. Another concern was the presence of anti-idiotype antibodies in the gamma globulin fraction, which may interfere with dissociation rate measurements. Previous studies utilizing anti-idiotype reagents generated by immunizations with non-liganded 4-4-20, however, indicated that anti-idiotype antibodies did not retard the dissociation rate (Weidner and Voss, 1991). Therefore, the use of anti-metatype gamma globulin fractions was considered preferable and valid. The following ligands were bound to Mab 4-4-20 in the dissociation rate assay: fluorescein, fluorescein isothiocyanate (isomers I and II), 5-dichlorotriazinyl 5-carboxyfluorescein. Disaminofluorescein, and sociation rates were performed at 2°C and the lifetime (7) of the antibody-ligand complex was determined for fluorescein and each structural analogue. The same assay was performed for each of the analogues in the presence of anti-metatype antibodies (Table 1). Kinetic studies of fluorescein structural analogues were best performed at 2°C due to rapid dissociation rates observed at 20°C. Under these conditions, Mab-4-4-20 and fluorescein displayed an intrinsic t of 1840.0 set while the addition of anti-metatype antibodies led to a 23-fold increase in r to 41,600 sec. FITC I and FITC II displayed lifetimes of 1413.4 and 0.3 set with Mab-4-4-20, respectively, and bound anti-metatype antibodies increased the lifetimes to approx. 32,000 and 86 set, respectively.

K. M. WEIDNERand E. W. Voss. JR

308

Table 1. Effect of anti-metaty~ antibodies on dissociation rate kinetics of Mab4-4-20 with fluorescein and various analogues” at 2°C Ligand” __--_______~ FDS FITC I 5-DCTAF 5-CFL FITC II

z (set)

Met zc (set)

1840.1 1413.4 7220.0 0.7 0.3

41600.0 3 1746.0 33333.3 169.4 62. I

At” -~23 x 23 x 5X 242 x 207 x

%’ 100 76.3 80.1 0.4 0.2

“Values based on the average of duplicate trials. *FDS (fluorescein disodium salt), FITC I (fluorescein isothiocyanate isomer I), 5-DCTAF (5-djchlorotriazinyl aminofluores~ein), 5-CFL (S-carboxyfluorescein), FITC II (fluorescein isothiocyanate isomer II). ‘At saturating concns of anti-Met. dChange in lifetime due to presence of anti-Met. ‘Percent of q,,,, (the lifetime of 4-4-20 and FDS in the

presence of anti-Met).

S-Dichlorotriazinyl aminofluorescein exhibited a lifetime of 7220.0 set with Mab4-4-20 (greater than that of Mab 4-4-20 and fluorescein). This enhanced lifetime might have been due to interaction of the chloride groups with residues at the mouth of the active site. However, the addition of anti-metatype antibodies increased 1:to approx. 33,000 set, below that of Mab 4-420 and fluorescein. Mab 4-4-20 and 5-carboxyfluorescein demonstrated a 0.7 set lifetime, and the addition of anti-metatype antibodies increased z to 169.4 sec. Reactivity with the anti-metatype gamma globulin fraction caused delays in the dissociation rates of all structural analogues tested. However, it is important to note that while the anti-metatype reagent increased 7 in a range of 5-242 x , the value oft was never increased beyond that of Mab 4-4-20 and ~uorescein in the presence of antimetatype antibodies (z,,,).

The dissociation rate assay was also performed using Mab 4-4-20 and fluorescein with increasing molecular weight carrier substitutions (Table 2) at either the 5- or &position. Dissociation rates were performed at 20°C to circumvent the slow reaction rates observed at 2°C. A comparison of the increase in lifetime (5) of the antibody-ligand complex in the presence of anti-metatype antibodies at 2 and 20°C revealed that reaction rates were proportional (within 2-fold) and, therefore, the relative rates of increase in z in the presence of antimetatype antibodies would be independent of temperature. At 20°C the Mab 4-4-20-fluorescein complex had a lifetime of approx. 231 set and the addition of antimetatype antibodies increased that value to 10,700 set or a 46fold increase in r. FITC I-acetyl-lysine and FITC I-bradykinin potentiator B behaved similarly yielding lifetimes of 131 and 117 set, respectively, which in the presence of anti-metatype antibodies increased to approx. 5900 and 7000 set (a 45 and 59-fold increase in t). FITC I conjugated to lactalbumin displayed a lifetime of 5.5 set when bound to Mab 4-4-20. Binding of anti-metatype antibodies led to a 1460-fold increase in z (Fig. 4). FITC I-ribonuclease dissociated from Mab 4-420 with a lifetime of 390 set (greater than Mab 4-4-20 and fluorescein), however, this effect may have been due to interactions of ribonuclease with non-active site side chains of the immunoglobulin molecule. Bound antimetatype antibodies increased r to approx. 9400 set, a value close to that of Mab 4-4-20 and fluorescein. yet not exceeding that value. Both the FITC I-PCMB and Ntetradecanoyl-aminofluorescein exhibited an increase in r in the presence of anti-metatype antibodies from 74.5 and 48.4 set, respectively to 4950.0 and 4048.0 sec. FITC II-acetyl-lysine and FITC I -bradykinin potentiator B showed 1500- and 960-fold increases in r upon binding with anti-metatype antibodies. Even though the increased z values were substantial, anti-metatype

Table 2. Effect of anti-metatype on dissociation rate kinetics of Mab 4-4-20 and FITC-labeled compounds” at 20°C Compound”

f (set)

Met T’ (set)

AZ”

FDS FITC I-RNAse FITC I-la~talbumin FITC I-BPB (peptide) FITC I-a~etyI-~ysine FITC I-PCMB 5-N-tetradecanoylamino-FL FITC II-BPB (peptide) FITC II-acetyl-lysine FITC II-lactalbumin FITC II-RNAse

230.6 390.7 5.5 117.0 131.0 74.5 48.4 4.2 1.9 NB’ NB’

10695.2 9389.7 8032.1 6944.4 5882.3 4950.0 4048.6 4024.1 2916.2 -

46x 24X 1460 x 59x 45 x 104x 84x 958 x 1566x -. -

%’

100 87.5 75.1 64.9 55.0 46.3 37.8 37.6 27.8 --

“Values based on the average of duplicate trials. ‘FITC 1 (fluorescein .5-isothiocyanate) and FITC 11 (fluorescein h-isothiocyanate). ‘At saturating concentrations of anti-Met. ‘Change in lifetime due to presence of anti-Met. ‘Percent of Tmax(the lifetime of 4-4-20 and FDS in the presence of anti-Met}. ’ NB = No binding,

309

Anti-Met interactions

0

40

20

60 TIE

80

100

bed

Fig. 4. Effect of anti-metatype antibodies on the dissociation rate of monoclonal antibody 4-4-20 and FITC I-lactalbumin at 20°C (C corresponds to the concn of ligand bound at time t and Co is the initial concn of bound ligand). The dissociation rate of FITC I-lactalbumin and 4-4-20 is represented by closed triangles and the delayed dissociation rate observed by the addition of anti-metatype antibodies is represented by closed circles. enhancement was lower than the values achieved for FITC I-conjugated derivatives. It has previously been suggested that binding of unconjugated FITC II leads to

a more open configuration of the 4-4-20 active site as a result of steric constraints (Swindlehurst and Voss, 1991), which may contribute to the lower values obtained for enhanced lifetimes. However, Mab 4-4-20 showed no binding to FITC II-lactalbumin and FITC IIribonuclease in the dissociation rate assay (as determined by lack of fluorescence quenching). These results were supported by non-reactivity of Mab4-4-20 with FITC II-lactalbumin and FITC II-ribonuclease in solid phase ELISA. Lack of binding by 4-4-20 may simply be due to the inaccessibility of FITC II due to steric hindrance in these low substituted compounds. To confirm that the delayed dissociation rate was due to binding of anti-metatype antibodies to 4-4-20 immunoglobulin variable regions, 4-4-20 Fab fragments and SCA 4-4-20/2 12 were examined. FITC I-bradykinin potentiator B was selected as the ligand because of an intermediate molecular weight and sensitive fluorescent properties. Dissociation rates of FITC I-bradykinin potentiator B with 4-4-20 Fab fragments and SCA 4-420/212 in the presence and absence of anti-Met were determined at 20°C (Table 3). Fab fragments of 4-4-20 displayed a lifetime of 70.8 set with the FTTC I-peptide, which increased to 4830.9 set when anti-metatype antiTable

3. Dissociation kinetics” of FITC Ibradykinin potentiator B with 4-4-20 Fab fragments and SCA 4-4-20/212 at 20°C

Molecule -4-4-20 Fab SCA 4-4-2012 12

r (set)

Met r (set)’

AT’

70.8 26.8

4830.9 4878.0

68 x 182x

“Values based on the average of duplicate trials. bAt saturating concentrations of anti-Met. ‘Change in lifetime due to the presence of anti-

Met. MIMM

29,3-B

body was added, a value close to that of 4-4-20 Ig molecule. SCA 4-4-201212 demonstrated a lifetime of 26.8 set with the FITC I-peptide and addition of antimetatype antibodies led to a 182 x increase in z to 4878 sec. Although SCA 4-4-20/212 exhibited a 3-fold shorter lifetime in the absence of anti-metatype antibodies, addition of the anti-Met reagent increased the lifetime of SCA---ligand complex to the same values as that for the Fab-ligand complex. DWXJSSION

Anti-metatype (anti-FITC I-affinity labeled 4-4-20) antibodies are specific for the Iiganded conformation of Mab 4-4-20, and it has been assumed that this liganded state represents induced conformational changes within the variable domains. Because some Ig proteins, including the Meg light chain (Bence-Jones) dimer, exhibit extensive conformational changes in the binding cavity to accommodate diverse ligands (Edmundson et al., 1984), it was assumed that the binding of fluorescyl structural analogues to Mab4-4-20 may also produce this effect. Induced conformational changes were monitored by reactivity of anti-metatype antibodies with liganded Mab 4-4-20. Anti-metatype antibodies reacted to some degree with Mab 4-4-20 bound to most of the fluorescein structural analogues tested, suggesting that upon binding these compounds 4-4-20 assumed similar liganded conformations. Values for T increased from 5to 242-fold for kinetic studies performed at 2°C and from 24- to 1566-fold for kinetic studies at 20°C. It is not yet clear, however, if values of enhanced lifetimes closest to T,,, (the lifetime of Mab 4-4-20 and fluorescein in the presence of anti-Met) indicate analogous liganded conformations or if the degree to which anti-metatype antibodies can enhance the lifetime indicates more closely related liganded st~ctures. Nonetheless, the value of r for each compound (at 2 and 20°C) in the presence of anti-Met antibodies never exceeded zmaX.

310

K. M. WEIDNER and E. W. Voss, JR

The value of T,~,..~must, therefore, be related to the average affinity of the polyclonal anti-metatype reagent for the fluorescyl-ligand complex. Previous studies with anti-metatype reagents suggested that two residues (Tyr H.53 and Asn L28) located at the mouth of the liganded active site may important in the formation of a metatope (Weidner and Voss, 1991). To test this hypothesis, FITC 1 was conjugated to carrier molecules of increasing molecular weight in an attempt to sterically hinder the binding of antimetatype antibodies at the mouth of the 4-4-20 active site. Additionally, FITC II was conjugated to most of the same molecules to observe the effects of orientation of the fluorescein group relative _to the carrier. Conjugation of FITC II to a carrier may provide protection of slightly different regions of the 4-4-20 “mouth.” However, all of the molecules tested, substituted at the 5- or 6-position and bound by 4-4-20, demonstrated a substantial delay in the dissociation rate (Table 2) upon reactivity with anti-metatype antibodies. The activity did not appear to be due exclusively to binding of anti-Met antibodies at the mouth of the active site which suggests that anti-Met antibodies may react with regions peripheral or distal to the active site. The results are consistent with the concept that rather than steric hindrance of the release of ligand, binding of anti-Met may further stabilize the liganded conformer (or the variable domains) thereby slowing the rate of ligand dissociation of (Fig. 5). A general trend exists (both Tables 1 and 2) that compounds exhibiting relatively short lifetimes in the absence of anti-Met antibodies tend to have enhanced lifetimes substantially below T,,,~~.This may suggest that there is a “critical” lifetime which must exist between the 4-4-20-ligand complex before anti-metatype antibodies can bind. If it is assumed that cooperative interactions exist between two or three anti-Met antibodies to delay hgand dissociation, the lifetime of the antibody-ligand complex must be greater than the time it takes for all anti-Met antibodies to bind. Perhaps binding of one or two anti-Met antibodies may enhance z, but all metatopes must be bound to reach values near t,,,. Interpretation of the final value and amount of increase in z in the presence of anti-Met antibodies for each FITC-conjugated analogue proved to be compiex (especially for molecules as large as FITC I-lactalbumin). Because other interactions are probably occurring

upon Mab 4-4-20 binding of such large molecules, attention was not focused on the actual value of z or the amount of increase in the lifetime in the presence of anti-Met antibodies. However, some general trends were observed. As the size of substitutions increased at the 5-position, the lifetimes of the complexes when bound by anti-Met antibodies were closer to T,,, Increasing molecular weight substitutions on fluorescein appeared to have a stabilizing effect on the antibody-ligand--antiMet ternary complex. The degree of stabilization upon binding of anti-metatype antibodies seemed to be influenced by: ( I ) interface factors at the mouth of the 4-4-20 site; (2) spacer distance between the 5-carbon position and occurrence of side chain branching evident in the peptide or protein derivatives; (3) preference of sulfur atoms relative to oxygen atoms near the interface (e.g. 5-CFL). and (4) pI of the protein or peptide linked to antibody-bound Buorescein. Finally, the significant difference between the effect of bound isomers I and II on stabilization by anti-Met reagents indicated that the difficulty of the 4-4-20 site to accommodate FITC II compounds (Swindlehurst and Voss, 1991) is direcly correlated with anti-Met stabilization. Regardless of the interactions of Mab 4-4-20 with FITC-derivatives (Table 2) the most striking observation is that while there is a wide range of z for all compounds in the absence of anti-Met, addition of anti-metatype antibodies increased T to near the same level (within 2-3-fold of r,,,) for almost of all FITC-conjugated compounds tested. This suggests that Mab 4-4-20 assumes nearly the same hganded conformation upon binding these analogues, and anti-metatype binds of each of these complexes with a similar affinity regardless of the lifetime of the antibody-ligand complex in the absence of anti-Met (assuming T is greater than the “critical” lifetime). Because there is precedence for conformational changes in the Fc portion of antibody molecules upon ligand binding (Nemazee and Sato, 1982) it was important to prove that the dissociation rate delay was due to anti-Met binding to antibody variable regions. This was verified using 4-4-20 Fab fragments and SCA 4-4-20/2 12 (Table 3). SCA 4-4-20/212 possessed an affinity 2-3-fold lower than the 4-4-20 Ig molecule, and its dissociation lifetime with FITC I-bradykinin potentiator B was approx. 223-fold lower. However, binding of anti-Met to the SCA 4-4-20!212-FITC I-BPB complex increased 5 to

Fig. 5. Hypothetical representation of conformational changes in the antibody active site upon formation of an immune complex. A hapten linked to a carrier molecule is bound by antibody inducing conformat~onal changes in the antibody active site. The asterisks (*) denote newly exposed or accessible regions (metato~s) upon binding by the antibody molecule. Anti-metatype antibodies may bind these regions stabilizing the immune complex delaying the dissociation rate of the complex.

311

Anti-Met interactions

approx. 4900 set, almost identical to that of 4-4-20 Fab fragments and close to the value obtained for 4-4-20 Ig molecule. This suggests that the delayed dissociation rate of the FITC-conjugated compounds is due to binding of anti-Met antibodies to the Ig variable regions, and that SCA 4-4-20/212 and 4-4-20 Fab fragments assume similar conformations upon binding such compounds. Even though the original studies with anti-Met used the hapten fluorescein as a model (Voss et al., 1988; Voss et al., 1989; Weidner and Voss, 1991), these studies have been expanded to include ligands of high mol. wt. The anti-metatype-metatype concept may now apply to biologically significant antigens of much larger size. Most intriguing, however, are the implications in z+o of antibodies (anti-metatype) which can delay the dissociation rate of large mol. wt ligands from antibody active sites. If these antibodies exist in biological systems. the consequences of such interactions in the immune system may prove important. For example, stimulation of B-cells could be enhanced by antimetatype antibodies’ by increasing the lifetime of the antigen---surface immunoglobulin complex. This time delay may also be sufficient to enhance T-cell (T, or T,) activation. Metatype-anti-metatype interactions could serve as a model for specific interactions in the immune system not necessarily involving only antibody-antibody complexes. Voss (1990) proposed metatype as a model for the T-cell receptor, and Lenhert et al. (1990) showed that T-cell glycoprotein CD4 binds the Fab portion of many human immunoglublins yet binds antibody coupled with antigen 100 times greater. Clearly, the idiotype-based Network Hypothesis (Jerne, 1975) could be influenced by metatype-anti-metatype interactions, yet how anti-metatype might affect this intricate regulatory system is not yet understood. A logical progression of these studies would be to generate monoclonal anti-metatype antibodies. However, it is becoming increasingly evident that production of a monoclonal antibody which can delay the dissociation rate may be difficult. If anti-Met antibodies act as stabilizers of the liganded conformation, such stabilization may require cooperative interactions of two or more antibodies binding to different “metatopes” on the surface. Therefore, the generation of monoclonals will become important to further define the interactions of the metatype-anti-nletatype complex.

REFERENCES

Bates R. M., Ballarcl D. W. and Voss E. W., Jr (1985) Comparative properties of monoclonal antibodies comprising a high affinity anti-fluorescyl idiotype family. Molec. Immun. 22, 871-877. Bedzyk W. D., Weidner K. M., Denzin I,. K., Johnson L. S., Hardman K. D., Pantoliano M. W., Asel E. K. and Voss E. W., Jr (1990) Immunological and structural characterization of a high affinity anti-fluorescein single-chain antibody. J. biof. Chem. 265, I8,615-18.620. Bird R. E., Hardman K. D., Jacobson J. W., Johnson S., Kaufman B. M., Lee S. M., Lee T., Pope S. H., Riordan

G. S. and Whitlow M. (1988) Single chain antigen binding proteins. Science 242, 423-426. Cheong H. S., Chang J. S., Park J. M. and Byun S. M. (1990) Affinity enhancement of bispecific antibody against two different epitopes in the same antigen. B&h. biophys. Res. Commun.

173, 795-800.

Denzin L. K., Whitlow M. and Voss E. W., Jr (1991) Singlechain site specific mutations of fluorescein-amino acid contact residues in high affinity monoclonal antibody 4-4-20. J. biol. Chem. 266, 14,095%14,103. Edmundson A. B., Ely K. R. and Herron J. N. (1984) A search for site-filling ligands in the Meg Bence-Jones dimer: Crystal binding studies of fluorescent compounds. .Uolec. rmmun. 21, 561-576. Erlich P. H., Moyle W. R., Moustafa Z. A. and Canfield R. E. (1982) Mixing two monoclonal antibodies yields enhanced affinity for antigen. J. Zmmun. 128, 2709-2713. Gibson A. L., Herron J. N., He X. E., Patrick V. A., Mason M. L., Lin J. N., Kranz D. M., Voss E. W., Jr and Edmundson A. B. (1988) Differences in crystal properties and ligand affinities of an anti-fluorescyl Fab (4-4-20) in two solvent systems. Proteins 3, 155-160. Herron J. N. and Voss E. W., Jr (1984) Analysis of heterogeneous dissociation kinetics in polyconal populations of anti-fluorescyl antibodies. In Fluorescein Hapten: An rmmunoio~i~al Probe (Edited by E. W. Voss, Jr.), pp. 77-96. CRC Press, Boca Raton, FL. Herron J. N., He X., Mason M. L., Voss E. W., Jr and Edmundson A. B. (1989) Three-dimensional structure of a fluorescein-Fab complex crystallized in 2-methyl-2,4-pentanediol. Proteins 5, 27 l-280. Holmes N. J. and Parham P. (1983) Enhancement of monoclonal antibodies against HLA-A2 is due to antibody bivalency. J. biol. Chem. 258, 1580-1586. Howard J. C., Butcher G. W., Galfre G., Milstein C. and Milstein C. P. (1979) Monoclonal antibodies as tools to analyze the serological and genetic complexities of major transplantation antigens. ln~mun. Rer. 47, 139-174. Jerne N. K. (197.5) The immune system: A web of V-domains. Hatwey Lect. 70, 93-l 10. Kato H. and Suzuki T. (1971) Bradykinin-potentiating peptides from the venom of Agkislrodon ha!ys blotnho_fJi. Isolation of five bradykinin potentiators and the amino acid sequences of two of them, potentiators B and C. Biochemistry 10,972-980. Kranz D. M., Herron J. N. and Voss E. W.. Jr (1982) Mechanisms of ligand binding by monoclonal anti-fluorescyl antibodies. .I. biol. Chem. 257, 6987-6995. Lenert P., Kroon D., Spiegelberg H., Golub E. S. and Zanetti M. (1990) Human CD4 binds immunoglobulins. Science 248, 1639-1643. Liabeuf A., Le Borgne de Kaouel C., Kourilsky F. M., Malissen B., Manuel Y. and Sanderson A. R. (1981) An antigenic determinant of human /?2-microglobulin masked by the association with HLA heavy chains at the cell surface: Analysis using monoclonal antibodies. J. immun. 127, 1542-1548. Nemazee D. A. and Sato V. L. (1982) Enhancing antibody: a novel component of the immune response. Pror. natn. Acad. Sci. U.S.A.

79, 3828-3832.

Oi V. T. and Herzenberg L. A. (1979) Localization of murine Ig-I b and Ig-la (IgG,,) altotypic determinants detected with monoclonal antibodies. Mrllec. Immun. 16, 1005-1017. Posner R. G., Erickson J. W., Holowka D., Baird B. and Goldstein B. (1991) Dissociation kinetics of bivalent

312

ligand-immunoglobulin

K. M. WEIDNERand E. W. Voss, JR E aggregates in solution. Biochem-

istry 30, 2348-2356.

Sundberg L. and Porath J. (1974) Preparation of adsorbents for biospecific affinity chromatography-I. Attachment of group containing ligands to insoluble polymers by means of befunctional oxiranes. J. Chromnt. 90, 87-98. Swindlehurst C. A. and Voss E. W., Jr (1991) Fluorescence measurements of immune complexes of Mab4-4-20 with isomeric haptens. Biophys. J. 59, 619-628. Voss E. W., Jr (1990) Anti-metatype reactivity: A model for T-cell receptor recognition. Imtnun. Today 11, 355-360. Voss E. W., Jr, Dombrink-Kurtzman M. A. and Ballard D. W. (1989) Inter-relationship between immunoglobulin idiotype and metatype. M&c. Itnmun. 26, 971-977.

Voss E. W., Jr, Miklasz S., Petrossian A. and DombrinkKurtzman M. A. (1988) Polyclonal antibodies specific for liganded active site (metatype) of a high affinity antimonoclonal antibody. Molec. rmrn~n. 25, hapten 751-759.

Watt R. M. and Voss E. W., Jr (1977) Mechanism of quenching of fluorescein by anti-fluorescein IgG antibodies. Immunochemistry

14, 533-541.

Watt R. M. and Voss E. W., Jr (1978) Characterization of a~nity-labeled fluorescyl ligand to specifically purified rabbit IgG antibodies. ~~zm~nochemi.~try 15, 875-882. Weidner K. M. and Voss E. W., Jr (1991) Immunological characterization of xenogenic anti-metatype antibodies. 1. hiol. Chem. 266, 2513-2519.