Transcript Determination of Antibody Dissociation Constants From Anisotropy

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 38C2,
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 (CDR1-3), as
they are located at the binding sites
of the antibody molecule
RasMol image of IgG molecule, PDB code 1IGT
C. Branden & J. Tooze, Introduction to Protein
Structure, 2nd 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
• 38C2 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 38C2 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 38C2 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 38C2: collected 14 days after the second boost
4
Ligand Molecules
• 1,3-Diketone
• Hapten all
antibodies studied
were raised against
O
• 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
O
O
O
OH
N
H
O
N
C H3
C H3
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 K2
– Fraction of Ligand Bound
• Antibody was titrated into a solution of prodan in PBS buffer
• Prodan fluorescence experiments done using a Perkin Elmer LS50B
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 
F0  F
F0  F m q
Fb: Fraction of binding sites filled
F0: Trp fluorescence in the absence of ligand
F: Trp fluorescence in the presence of ligand
Fmq: Trp fluorescence with maximum quenching
• A plot of F0-F vs. the total ligand concentration (Lt) can be used to
determine the dissociation constants (Kd) of the antibodies according
to the following relationship:
F0  F  ( F0  F m q )
( K d  S t  Lt ) 
K d  2 K d S t  2 K d Lt  S t  2 S t Lt  Lt
2
2
2
2St
St: Total concentration of antibody binding sites (all other parameters defined above)
8
Quenching of Intrinsic Tryptophan
Fluorescence on Hapten Binding
38C2
Increasing
hapten
Intensity (a.u.)
120x10
3
80
40
320
IgG
400
Increasing
hapten
Intensity (a. u.)
160x10
340
360
380
Wavelength (nm)
3
120
80
40
320
340
360
380
Wavelength (nm)
400
• Excitation at 295 nm for
38C2, 284 nm for IgG
• The fluorescence
intensity due to intrinsic
tryptophan residues in
the antibodies decreases
with increasing hapten
concentration
• The fluorescence
quenching observed for
the mature 38C2
antibody (specific
binding) is greater than
that observed for IgG
(non-specific binding)
9
Quenching of Intrinsic Tryptophan
Fluorescence on Hapten Binding
Tertiary
Secondary (A2c26.1)
Increasing
hapten
Intensity (a. u.)
Increasing 120x103
hapten
100
Intensity (a.u.)
0x10
lexc=295 nm
in all cases
3
80
60
40
20
100
50
340
360
380
Wavelength (nm)
Intensity (a.u.)
Secondary (A2c22.1)
600
500
400
300
200
100
320
320
400
340
360
380
Wavelength (nm)
400
Primary
Increasing
hapten
Increasing
3
120x10
hapten
340
360
380
Wavelength (nm)
Intensity (a.u.)
320
80
40
400
320
340
360
380
Wavelength (nm)
400
10
Determination of Antibody Dissociation Constants
From Tryptophan Fluorescence Quenching
10
Mature 38C2
Kdq= 0.05 ± 0.10 mM
Tertiary
Kd= 0.15 ± 0.09 mM
Secondary (A2c26.1) Kd= 2.7 ± 1.6 mM
Secondary (A2c22.1) Kd= 9.0 ± 0.9 mM
3
F0-F (Arbitrary Units)
100
80
60
40
20
0
0
1
2
3
4
Hapten Concentration (M)
5
6x10
-6
• Ab concentrations  3mM
• Primary antibody shows
non-specific binding
similar to IgG (data not
shown)
• Other antibodies show
binding behavior more
similar to 38C2
• Because 38C2 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
O
Prodan Emission and Anisotropy
N
Filter Transmission
Methanol
Phosphate
Buffer
Intensity (Arbitrary Units)
Acetone
450
500
Wavelength (nm)
550
C H3
C H3
• 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
600 fraction of prodan bound
12
Emission Spectra of Prodan Bound
to the Mature 38C2 Antibody
Intensity (Arbitrary Units)
Bound Prodan
Free Prodan
400
450
500
Wavelength (nm)
550
• Excitation is at 361 nm
• Prodan emission shifts to
445 nm upon binding to
the 38C2 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
600
13
Emission Spectra of Prodan Bound
to the Tertiary Antibody
Free Prodan
Intensity (Arbitrary Units)
Bound Prodan
400
450
500
Wavelength (nm)
550
• Excitation is at 361 nm
• As seen with the 38C2
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
600
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:
F 
( F max  F min )[ S ]
K d  [S ]
 F min
• The concentration of free antibody was approximated by the
total antibody concentration, which is valid as the total prodan
concentration used (20 nM) was at least ten-fold lower than the
observed Kd values
15
Determination of Antibody Dissociation
Constants From Fluorescence Enhancement
Intensity at 445 nm (a. u.)
Mature 38C2 Kd = 0.75 ± 0.06 mM
Tertiary
Kd = 0.73 ± 0.04 mM
0
200
400
600
800
1000
1200
Concentration of Antibody Binding Sites (nM)
1400
16
Determination of Antibody Dissociation
Constants From Anisotropy Data
Mature 38C2
Tertiary
Secondary
Primary
Anisotropy at 518 nm (mP)
300
250
Kd = 0.51 ± 0.06 m
Kd = 0.8 ± 0.3 mM
Kd = 1.8 ± 0.7 mM
Kd = 1.1 ± 1.0 mM
200
150
100
50
0
200
400
600
800
1000
1200
Concentration of Antibody Binding Sites (nM)
1400
17
Summary of Antibody Binding Affinity Data
Antibody
Heavy Chain CDR3
Sequence
Kd’/Kdq(38C2) Kd /Kd(38C2)
% Trp
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
(A2c26.1)
CATAHYVNPGRFTKTLDYW
38%
(10/26)
54
3.5
Secondary
(A2c22.1)
CTRGNYGYVGAYW
38%
(10/26)
180*
N/A
CTRWGYAYW
43%
(12/28)
Non-specific
binding
2.1
38C2
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 mM.
• 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 mM
(secondary) to 0.8 mM (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
Intensity at 445 nm (a.u.)
Mature 38C2 Kd = 360 ± 20 nM
Tertiary
Kd = 370 ± 20 nM
0
200
400
Antibody Concentration (nM)
600
16
Determination of Antibody Dissociation
Constants From Fluorescence Enhancement
Intensity at 445 nm (Arbitrary Units)
Mature 38C2 Kd = 380 ± 30 nM
Tertiary
Kd = 370 ± 20 nM
0
200
400
Antibody Concentration (nM)
600
16
Quenching of Intrinsic Tryptophan
Fluorescence on Hapten Binding
3
3
80
40
340
360
380
Wavelength (nm)
320
340
360
380
Wavelength (nm)
400
340
360
380
Wavelength (nm)
400
3
Intensity (a.u.)
Intensity (a.u.)
320
150x10
120x10
100
80
60
40
20
400
Intensity (a. u.)
Intensity (a.u.)
120x10
120x10
3
80
100
50
40
320
340
360
380
Wavelength (nm)
400
320
Emission Spectra of Prodan Bound to
the Mature 38C2 Antibody
Free Prodan
Intensity (Arbitrary Units)
60
50
40
Bound Prodan
30
20
• Excitation is at 361
nm
• Emission intensity
at 445 nm is due to
bound prodan, and
is observed to
increase with
increasing antibody
concentration
10
0
400
450
500
Wavelength (nm)
550
600
Emission Spectra of Prodan Bound to
the Mature 38C2 Antibody
Bound Prodan
Intensity in Arbitrary Units
120
100
Free Prodan
80
60
40
• Excitation is at 361
nm
• Emission intensity
at 445 nm is due to
bound prodan, and
is observed to
increase with
increasing antibody
concentration
20
400
450
500
Wavelength (nm)
550
600
Summary of Antibody Binding Affinity Data
Antibody
Heavy Chain CDR3
Sequence
Kd /Kd38C2
% Trp
Trp
residues in
binding site quenching
by hapten
Kd from prodan
anisotropy
(fluorescence)
CKIYKYSFSYW
42%
(10/24)
1
0.51 ± 0.06 mM
(0.75 ± 0.06 mM)
CIRGGTAYNRYDGAYW
38%
(10/26)
3
0.8 ± 0.3 mM
(0.73 ± 0.04 mM)
Secondary
(A2c26.1)
CATAHYVNPGRFTKTLDYW
38%
(10/26)
54
1.8 ± 0.7 mM
Secondary
(A2c22.1)
CTRGNYGYVGAYW
38%
(10/26)
180*
N/A
CTRWGYAYW
43%
(12/28)
Non-specific
binding
1.1 ± 1.0 mM
38C2
Tertiary
Primary
*Ligand
used was acetyl acetone, not the hapten
18
Determination of Antibody Dissociation
Constants From Anisotropy Data
Primary
Kd = 2.5 ± 0.4 mM
Secondary Kd = 1.3 ± 0.8 mM
Tertiary
Kd = 410 ± 20 nM
Mature 38C2 Kd = 250 ± 40 nM
Anisotropy at 518 nm (mP)
300
250
200
150
100
50
0
200
400
600
Total Antibody Concentration (nM)
17
Determination of Antibody Binding Constants
From Tryptophan Fluorescence Quenching
F0-F (Arbitrary Units)
100x10
Non-Specific Antibody
Primary Antibody
Secondary Antibody
Tertiary Antibody
Mature 38C2 Antibody
3
80
60
40
20
0
0
1
2
3
4
Hapten Concentration (M)
5x10
-6
11
Summary of Antibody Binding Affinity Data
• Prodan Binding
– Tertiary: Kd equivalent to 38C2
– Secondary: Kd that of 38C2
– Primary: Kd that of 38C2
• Hapten Binding
– Tertiary: Kd equivalent to 38C2
– Secondary: Kd that of 38C2
– Primary: Kd that of 38C2
18