Abstract Antibodies elicited early in an immune response

Abstract Antibodies elicited early in an immune response exhibit lower binding affinity for hapten as compared with mature antibodies which have undergone hypermutation. We have observed that increasing binding affinity correlates with structural rigidity in the family of antibodies raised against the same diketone hapten. This family of antibodies includes the mature aldolase antibody 38 C 2, which has been shown to catalyze an aldol condensation by means of a protein bound enamine intermediate (Science (1995) 270, 1797 -1800). In addition to the catalytic antibody, a family of related noncatalytic antibodies from a primary, secondary, and tertiary response have been characterized according to their binding affinity as measured by fluorescence quenching. The flexibility of the binding site before and after binding has also been characterized from the distribution of fluorescent lifetimes derived from a GLOBALS fit to the frequency data. The catalytic antibody elicited by reactive immunization resembles most closely the primary antibody, suggesting that once covalent bond formation was used to bind the hapten, binding site evolution stopped. By contrast, secondary and tertiary antibodies reveal much narrower lifetime distributions.

Antibody Structure • Antibodies, or immunoglobulin (Ig) molecules, are composed of four chains, two light and two heavy • The structure of Ig molecules can be divided into a constant region, in which the amino acid sequence is largely conserved, and a variable region, where the amino acid sequence for different Ig molecules has considerably more variation • Within the variable region, there are three hypervariable regions, referred to as complementarity determining regions (CDR 1 -3), as they are located at the binding sites of the antibody molecule Ras. Mol image of Ig. G molecule, PDB code 1 IGT C. Branden & J. Tooze, Introduction to Protein Structure, 2 nd ed. , Garland Publishing, Inc. (1999) 1

Affinity Maturation • Upon exposure to an unknown antigen, the immune system begins a process called affinity maturation • The germline antibody that binds the antigen most tightly undergoes somatic hypermutation in the variable region to form antibodies that each have a new active site and, therefore, modified affinity for the antigen • The immune system selects for the modified antibody that most tightly binds the antigen and somatic hypermutation occurs again to produce a more mature antibody with an even more specialized binding site • Therefore, more mature antibodies should show both more rigid binding sites and also smaller dissociation constants 2

Aldolase Family of Antibodies • 38 C 2 is a catalytic antibody, first developed at the Scripps Research Institute - Wagner, J. , Lerner, R. A. , Barbas, C. F. , Science 270, 1797 (1995) • This antibody was raised against a diketone hapten by the process of reactive immunization, and catalyzes the aldol reaction • Reactive immunization involves an actual chemical reaction between the hapten and the antibody, rather than the typical non-covalent interaction between the hapten and the antibody • The evidence for catalytic activity is the observation of an absorption band at 316 nm due to the vinylogous amide formed in the reaction of the diketone hapten with a lysine residue in the binding site of the antibody • The antibodies in this study do not show an absorption peak at 316 nm, indicating that the interaction between the hapten and these antibodies is not covalent 3

Aldolase Family of Antibodies • The antibodies used in this study were raised against the same diketone hapten used to generate the mature 38 C 2 antibody • Previous studies of these antibodies examined the rigidity of the binding site both before and after hapten binding, and demonstrated that the behavior of the primary antibody most closely resembles that of the mature 38 C 2 antibody (Chiu, P. , O’Hara, P. B. , unpublished data) • This indicates that the affinity maturation process for antibodies generated by reactive immunization does not follow the same trend as antibodies raised by interaction with a neutral hapten Antibodies Examined: Primary: collected 12 days after initial exposure to the hapten Secondary: collected 5 days after the first boost Tertiary: collected 5 days after the second boost Mature 38 C 2: collected 14 days after the second boost 4

Ligand Molecules • 1, 3 -Diketone • Hapten all antibodies studied were raised against • Prodan, or 6 -propionyl-2(dimethylamino)naphthalene • Fluorescent molecule • Structural similarities to the 1, 3 -diketone hapten indicated that prodan could bind to the antibodies studied 5

Questions Addressed • Can Trp fluorescence quenching be used to accurately measure the binding of the hapten to the antibodies? • Do the Kd values derived from the Trp fluorescence quenching correlate with the expected binding affinities? • Can prodan, a highly fluorescent and environmentally sensitive small molecule, mimic hapten binding? • Can changes in prodan fluorescence be used to measure the relative binding affinity of the members of this family of antibodies? • How does the binding behavior of the antibodies to the diketone hapten compare with that observed for prodan? 6

Experimental Techniques Antibody dissociation constants were determined using techniques in which the experimental observable is proportional to either the fraction of antibody binding sites filled or the fraction of ligand bound – Fraction of Binding Sites Filled • Diketone hapten was titrated into a solution of antibody in PBS buffer • Tryptophan fluorescence quenching monitored using an ISS K 2 – Fraction of Ligand Bound • Antibody was titrated into a solution of prodan in PBS buffer • Prodan fluorescence experiments done using a Perkin Elmer LS 50 B fluorimeter • Prodan fluorescence anisotropy measured with a Panvera Beacon 2000 using an excitation filter centered at 360 nm and a transmission filter centered at 510 nm 7

Determination of Antibody Dissociation Constants • Quenching of the intrinsic tryptophan fluorescence of the antibodies is related to the fraction of antibody binding sites filled (Fb) according to the following relationship: Fb: Fraction of binding sites filled F 0: Trp fluorescence in the absence of ligand F: Trp fluorescence in the presence of ligand Fmq: Trp fluorescence with maximum quenching • A plot of F 0 -F vs. the total ligand concentration (Lt) can be used to determine the dissociation constants (Kd) of the antibodies according to the following relationship: St: Total concentration of antibody binding sites (all other parameters defined above) 8

Quenching of Intrinsic Tryptophan Fluorescence on Hapten Binding 38 C 2 Increasing hapten Ig. G Increasing hapten • Excitation at 295 nm for 38 C 2, 284 nm for Ig. G • The fluorescence intensity due to intrinsic tryptophan residues in the antibodies decreases with increasing hapten concentration • The fluorescence quenching observed for the mature 38 C 2 antibody (specific binding) is greater than that observed for Ig. G (non-specific binding) 9

Quenching of Intrinsic Tryptophan Fluorescence on Hapten Binding Tertiary lexc=295 nm in all cases Secondary (A 2 c 26. 1) Increasing hapten Primary Secondary (A 2 c 22. 1) Increasing hapten 10

Determination of Antibody Dissociation Constants From Tryptophan Fluorescence Quenching • Ab concentrations 3 m. M • Primary antibody shows non-specific binding similar to Ig. G (data not shown) • Other antibodies show binding behavior more similar to 38 C 2 • Because 38 C 2 forms a covalent bond with the hapten, the value reported is an apparent Kd, Kdq • Non-specific binding occurs at high hapten concentrations • Use of fluorescent ligand would allow for lower ligand concentrations 11

Prodan Emission and Anisotropy • The emission spectrum of prodan blue-shifts with decreasing solvent polarity • The emission spectrum of prodan in acetone resembles that of prodan bound to the antibodies studied • Both free (low anisotropy) and bound (high anisotropy) prodan fluoresce in the range of the emission filter used in the anisotropy experiments, therefore, the anisotropy in this wavelength range can be used to determine the fraction of prodan bound 12

Emission Spectra of Prodan Bound to the Mature 38 C 2 Antibody • Excitation is at 361 nm • Prodan emission shifts to 445 nm upon binding to the 38 C 2 antibody • The emission intensity at 445 nm increases with increasing antibody concentration • At 510 nm, both free and bound prodan contribute to the fluorescence signal 13

Emission Spectra of Prodan Bound to the Tertiary Antibody • Excitation is at 361 nm • As seen with the 38 C 2 antibody, the prodan emission shifts to 445 nm upon binding to the tertiary antibody, and the intensity at 445 nm increases with increasing antibody concentration • Again, both bound and free prodan contribute to the fluorescence signal at 510 nm 14

Determination of Antibody Dissociation Constants • The fluorescence anisotropy of prodan (for the wavelength range 506 - 516 nm) and the fluorescence intensity of prodan at 445 nm are related to the fraction of prodan bound to antibody • A plot of either of these two experimental observables vs. the concentration of free antibody binding sites, [S], allows the dissociation constant, Kd, of the antibodies to be determined according to the following relationship: • The concentration of free antibody was approximated by the total antibody concentration, which is valid as the total prodan concentration used (20 n. M) was at least ten-fold lower than the observed Kd values 15

Determination of Antibody Dissociation Constants From Fluorescence Enhancement 16

Determination of Antibody Dissociation Constants From Anisotropy Data 17

Summary of Antibody Binding Affinity Data Antibody Heavy Chain CDR 3 Sequence % Trp Kd’ /Kdq(38 C 2) Kd /Kd(38 C 2) residues in Trp quenching Prodan anisotropy binding site by hapten (Fluorescence) CKIYKYSFSYW 42% (10/24) 1 1 (1) CIRGGTAYNRYDGAYW 38% (10/26) 3 1. 6 (1) Secondary (A 2 c 26. 1) CATAHYVNPGRFTKTLDYW 38% (10/26) 54 3. 5 Secondary (A 2 c 22. 1) CTRGNYGYVGAYW 38% (10/26) 180* N/A CTRWGYAYW 43% (12/28) Non-specific binding 2. 1 38 C 2 Tertiary Primary *Ligand used was acetyl acetone, not the hapten 18

Conclusions • Kd’ values derived from Trp quenching by hapten set an upper limit on Kd. Both nonspecific binding and low sensitivity of Trp fluorescence preclude more precise measurements using this technique. • Primary antibodies exhibit low affinity binding for hapten as measured by Trp quenching. Quenching of Trp fluorescence from secondary and tertiary antibodies yields Kd’ values that correlate with predicted relative binding affinities and are in the range of 10 to 0. 1 m. M. • Both the blue-shifted emission and increased anisotropy of prodan can be used as a more sensitive probe of binding affinity of antibodies. Though raised against a diketone hapten, these antibodies exhibit Kd values which range from 2 m. M (secondary) to 0. 8 m. M (tertiary) • The maximum anisotropy values observed for prodan with various antibodies are not identical, suggesting that this parameter may be a good indicator of binding site rigidification upon affinity maturation. • The general trend exhibited by the antibodies for binding to hapten and prodan are the same, but the hapten binding exhibits predictably wider range of Kd’ values. 19

Acknowledgements • NSF CRUI collaborators – Prof. Richard Goldsby, creation of antibody family – Prof. David Hansen, synthesis of hapten – Prof. David Ratner and Nalini Sha-Mahoney, genetic characterization of antibodies – Phil Chiu ’ 02, Trp quenching data • Camille and Henry Dreyfus Scholar/Fellow Program for Undergraduate Institutions 20

Substrate Molecules • Hapten • Prodan

Determination of Antibody Dissociation Constants From Fluorescence Enhancement 16

Quenching of Intrinsic Tryptophan Fluorescence on Hapten Binding

Emission Spectra of Prodan Bound to the Mature 38 C 2 Antibody • Excitation is at 361 nm • Emission intensity at 445 nm is due to bound prodan, and is observed to increase with increasing antibody concentration

Summary of Antibody Binding Affinity Data Antibody Heavy Chain CDR 3 Sequence Kd /Kd 38 C 2 % Trp residues in binding site quenching by hapten Kd from prodan anisotropy (fluorescence) CKIYKYSFSYW 42% (10/24) 1 0. 51 ± 0. 06 m. M (0. 75 ± 0. 06 m. M) CIRGGTAYNRYDGAYW 38% (10/26) 3 0. 8 ± 0. 3 m. M (0. 73 ± 0. 04 m. M) Secondary (A 2 c 26. 1) CATAHYVNPGRFTKTLDYW 38% (10/26) 54 1. 8 ± 0. 7 m. M Secondary (A 2 c 22. 1) CTRGNYGYVGAYW 38% (10/26) 180* N/A CTRWGYAYW 43% (12/28) Non-specific binding 1. 1 ± 1. 0 m. M 38 C 2 Tertiary Primary *Ligand used was acetyl acetone, not the hapten 18

Determination of Antibody Dissociation Constants From Anisotropy Data 17

Determination of Antibody Binding Constants From Tryptophan Fluorescence Quenching 11

Summary of Antibody Binding Affinity Data • Prodan Binding – Tertiary: Kd equivalent to 38 C 2 – Secondary: Kd that of 38 C 2 – Primary: Kd that of 38 C 2 • Hapten Binding – Tertiary: Kd equivalent to 38 C 2 – Secondary: Kd that of 38 C 2 – Primary: Kd that of 38 C 2 18