Protein-protein interactions (PPIs) via NMR

Paola Turano

turano@cerm.unifi.it

The chemical shift The magnetic field at the nucleus (the effective field) is generally less than the applied field by a fraction s: B = B0 (1-s)

Simple shielding effects:

•eletronegativity

• “shielding cones”

N

H

more electron

withdrawing--

less shielded

shielding

deshielding

shielding

shield

ing

deshielding

1H (ppm) MORE SHIELDED

Internal protein dynamics

Considering the rotational correlation time of the protein in solution taken as a rigid body as a pivotal point of the timescale, we can define as internal fast motions those faster than the tumbling correlation time and as collective conformational equilibria those involving processes slower than the tumbling correlation time.

tumbling

Exchange process: a dynamic process that exposes a nucleus to at

least two distinct chemical environments

Two-site exchange

obs = fAA + fBB = 0.75 A + 0.25 B

Ix Px

The exchange regime is determined by the chemical shift separation (in Hz). Can be modulated by T (affecting kex)and B0

(affecting ) .

Special case: ligand binding to a protein

Reaction scheme:

The exchanging sites are the free (P) and the complexed form (PL) of the protein.

The exchange rate is given by:

Line shapes simulated for the one-step binding mechanism for increasing populations of the complex (from blue to red). = 250 Hz kex = 2000 Hz.

Thermodynamics of ligand binding

P + L PL Ka = [PL]/[P][C] PL P + L Kd = [P][L]/[PL] Protein-protein interactions (PPIs) span an affinity range that is extremely broad, with dissociation constants Kd from 10-2 to 10-16 M.

Affinity constant and dissociation constant

Kd = [P][L]/[PL] Strong PPIs: dissociation constant Kd < 1 M Weak PPIs: 1 M < Kd < 100 M Ultra weak PPIs: Kd > 100 M

Kinetics

In the assumption of one-step reaction: PL P+ L Kd = koff ⁄ kon Different combinations of koff and kon can give rise to similar affinities.

koff

kon

kex, kon, koff

The NMR exchange rate is given by:

What are the factors affecting kon and koff in PPIs?

Association rate constant kon

Wide spectrum of measured association rate constants. The red vertical line marks the start of the diffusion-controlled regime. The shaded range marks the absence of long-range forces.

Association rate constant kon

The rate of association of a protein complex is limited by diffusion and geometric constraints of the binding sites and may be further reduced by subsequent chemical processes.

The association of two proteins is bounded by the rate at which they find each other through diffusion. To form a stereospecific complex, the two molecules must have appropriate relative orientations when they come together.

A diffusion-controlled rate constant falls on the high end of the spectrum of observed values. A reaction-controlled rate constant falls on the low end A diffusion-controlled association typically involves only local conformational changes between the unbound proteins and the native complex. A reaction-controlled association typically involves gross changes such as loop reorganization or domain movement.

Association rate constant kon

Association rate constant kon: electrostatic contribution

To go beyond the basal rate constant kD0 and reach values > 107 M-1 s-1, as observed for many protein complexes, intermolecular forces must be present. For protein-protein association the dominant long-range force is provided by electrostatic interactions.

For protein-protein association the dominant long-range force is provided by electrostatic interactions, as manifested by complementary charge distributions on the binding partners, which are illustrated in the figure for four protein pairs.

Association rate constant kon

Kinetics vs. thermodynamics

When several proteins compete for the same receptor or when one protein is faced with alternative pathways, kinetic control, not thermodynamic control, dominates in many cases; this is especially true when dissociation is slow. Differences in binding rate between related proteins may serve as an additional mechanism for specificity.

Kinetics vs. thermodynamics

When several proteins compete for the same receptor or when one protein is faced with alternative pathways, kinetic control, not thermodynamic control, dominates in many cases; this is especially true when dissociation is slow. Differences in binding rate between related proteins may serve as an additional mechanism for specificity.

Dissociation rate constant koff

Dissociation is a first order reaction whose rate is dictated by the strength of short range interactions between the interacting proteins (van der Waals interactions, hydrogen bonds, hydrophobic interactions and salt bridges) Typical koff are in the 104-10-7 s-1 range.

Dissociation rate constant koff

koff is directly related to the lifetime of the complex: mean lifetime = 1/koff half-life = ln2/koff

A higher value of koff means a shorter lifetime for the complex and vice versa.

Transient complexes

Transient complexes form when a high turnover is a functional requirement and their components associate and dissociate rapidly, namely with koff 103 s-1 and kon in the range of 107-109 M-1s-1. This results in dissociation constants typical for weak and ultra weak complexes and lifetimes ms. Revealing the presence of such interactions is experimentally challenging because they do not result in a sufficient amount of complexes that can be directly detected.

koff in electrostatic interactions

The association rate constant decreases significantly with increasing ionic strength, whereas the dissociation rate constant is only modestly affected by ionic strenght. As the transient complex lies at the outer boundary of the interaction energy well and hence is close to the native complex, ionic strength is expected to screen electrostatic interactions in the two types of complexes to nearly the same extent. Hence, the association constant and association rate constant are expected to have nearly the same dependence on ionic strength and the dissociation rate would be little affected by ionic strength. Ka = kon ⁄ koff

Association rate constant kon

Electrostatic interactions. Rates beyond the 105–106 M-1 s-1 range implicate electrostatic enhancement. A hallmark of electrostatically enhanced diffusion-limited association is manifested by disparate ionic-strength effects on the association rate kon and the dissociation rate koff. Specifically, kon decreases significantly with increasing ionic strength, but koff is affected by ionic strength only marginally

Affinity constant and dissociation constant

Kd = [A][B]/[C] Strong PPIs: dissociation constant Kd < 1 M Weak PPIs: 1 M < Kd < 100 M Ultra weak PPIs: Kd > 100 M

NMR = unique tool to monitor these interactions

NMR can be used to study protein–ligand interactions, by acquiring multiple NMR spectra along a [P]/[L] titration

coordinate.

As a rule of thumb, in the diffusion controlled association regime:

Kd mM Kd nM Kd M

1H-15N HSQC = gold standard for monitoring intermolecular interactions

The 1H-15N HSQC spectrum is a signature of the protein structure

The 1H-15N HSQC experiment folded

unfolded

Partially folded

1H chemical shift as a signature of the degree of protein folding

1H-15N HSQC in interaction studies

Different strategies for fast, slow and intermediate exchange

THE CHEMICAL SHIFT OR AMIDES OF RESIDUES IN THE INTERACTION AREA IS EXPECTED TO CHANGE UPON COMPLEX FORMATION DUE TO THE CHANGE IN THE CHEMICAL ENVIRONMENT OF THE NUCLEUS. 1H NUCLEI ARE VERY SENSITIVE REPORTES OF CHANGES IN THEIR ENVIRONMENT

1H-15N HSQC: fast chemical exchange

1H-15N HSQC: fast chemical exchange

obs = ffree free+ fbound bound All the other NMR observables are also population-weighted, including relaxation rates. Measurable effects on obs even in the case of

ultra weak interactions (up to mM Kd’s)

Chemical shift perturbation (CSP) mapping via HSQC

1H-15N HSQC = simple and fast experiment, with cryoprobe technology can be acquired in a few minutes even for very low protein concentrations (down to tens of M). In interaction experiments, the 1H-15N HSQC spectrum of one protein is monitored when the unlabeled interaction partner is titrated in and the chemical shift perturbations are recorded for each amino acid. Provided the assignment of the 1H-15N HSQC spectrum is available, shift perturbation measurements identify residues at the interface. If a structural model of the protein exists residues undergoing meaningful chemical shift perturbations are mapped on the protein structure, thus providing the location of the interface. In the case of protein-protein interactions, the procedure is repeated for the second partner.

Δav=[(Δδ2NH+(ΔδN/5) 2)/2]1/2

Garrett plot

Identification of residues undergoing chemical shift perturbations follwoing them in a titration

Mapping of the chemical shift perturbations on the protein surface. Data driven docking to define a structural model

Contact surface areas are identified if assignment is available for both partner macromolecules

1H-15N HSQC CSP mapping: fast exchange

Bertini, Chevance, Del Conte, Lalli, Turano PLoS One. 2011 6(4):e18329.

Data Driven Docking

Chemical shift perturbation mapping yields the interaction area on the individual binding partners, but does not allow defining the relative orientation of the two molecules nor the atom-to-atom interactions at the basis of the recognition process. Nevertheless, residues undergoing chemical shift perturbations can be used as selection filters in data driven computational soft-docking programs

Kd measurements

Simultaneous fit of the chemical shift changes

Under fast exchange conditions:

Kd = [A][B]/[C]

Kd values were obtained by plotting the weighted average chemical shift variations of perturbed residues on the 15N-enriched Bcl-xL as a function of the concentration of the unlabeled partner cytochrome c and were found to be in the 1-3 mM range.

Optimal Kd /[B] when titrating with A

Protein in excess of 10-50*Kd results in very large errors in measurement Small amounts of protein with weak binding require a large excess of ligand to achieve saturation

[B] >> Kd

Lack of curvature makes the fitting difficult/impossible

[B] << Kd

Fitting is good but too much ligand is needed

Curve is good when [B] Kd

Chemical shift mapping: limitations of the approach

In the presence of binding-induced conformational rearrangements, the chemical shift perturbation may extend to residues that are not at the interface, and the chemical shift perturbation fails as a mapping tool, although it still represents an excellent indicator of allosteric processes. Such a situation is associated to slow kon.

1H-15N HSQC: intermediate exchange regime

During the titration signals may be broadened beyond detection due to exchange broadening and can be retrieved at the end (if we can get the final adduct!).

Chemical shift mapping under slow exchange conditions

The problem of signal assignment

Structural info can be derived only if we know the assignment

EXSY experiments: a possible tool for the (straightforward) assignment

under slow exchange regime

Homonuclear 1H-1H or 13C-13C-like experiments

Observing signals of the bound and free forms in the same NMR spectrum is a prerequisite.

EXchange SpectroscopY (EXSY), also known as the zz-exchange experiment. τex≈10–5000 ms; k ex ≈0.2−100 s-1

Slow exchange

If the exchange rate is << R2, then the exchange event has little or no effect on the linewidth. But, if the exchange rate is > R1 it may be possible to measure the rate constants by detecting the exchange of magnetization between the nuclei in the two environments.

EXchange SpectroscopY (EXSY), also known as the zz-exchange experiment. τex≈10–5000 ms; k ex ≈0.2−100 s-1

EXSY experiments (a)

(b)

(c)

Here m is the mixing time!!!

heteronulcear

Slow exchange with respect to the chemical shift time scale, but fast with respect to T1

Other NMR observables for PPIs

•Changes in protein dynamics. •Changes in the overall tumbling time. •Distance and orientational restraints: NOE RDC Paramagnetic effects

Changes in the overall tumbling time: establishment of intermolecular interactions gives rise to changes in the size of the system under study. For protein-protein interactions an overall increase in 15N transverse relaxation rate values is observed, which is consistent with an increase in the overall tumbling correlation time upon complex formation. In the fast exchange regime, the measured transverse relaxation rate is an average of the transverse relaxation rates of the free and bound protein forms, weighted by their molar fraction. This phenomenon can be used to establish complex formation, but also causes signal broadening in the spectra of the bound form of a protein, thus often requiring ad hoc experimental procedures for the obtainment of 1H-15N correlations at the basis of the chemical shift perturbation mapping approach, as detailed below. Under slow exchange signals may become undetectable in HSQC TROSY, CRIPT/CRINEPT TROSY: “monolateral” chemical shift mapping for complexes up to 1 MDa,

NMR for the characterization of PPIs: other tools

Changes in protein dynamics

Reduced solvent accessibility of the contact area with exclusion of bulk water. Some solvent molecules may remain at the interface: bridging; non-bridging; simply trapped.

Changes in internal dynamics. The side chain (exemplified by an arrow) of a residue of the blue protein in its free form exists in a number of energetically similar conformations, that can be represented by a dynamic ensemble of interconverting states. Upon binding the side chain assumes a single conformation

Interaction between two molecules not only implies changes in chemical environment due to local/extended conformational variations but also affects the solvent accessibility and internal protein mobility.

Changes in backbone dynamics

Changes in protein dynamics

J. Mol. Biol. (2003) 327, 719–734

MMP

TIMP-1

Rigified: blue Unaffected : green Faster: red

Establishment of intermolecular interactions gives rise to changes in the size of the system under study. For protein-protein interactions an overall increase in 15N transverse relaxation rate values is observed, which is consistent with an increase in the overall tumbling correlation time upon complex formation.

Changes in the overall tumbling time

Molecular tumbling: r

r is larger for larger molecules Stokes–Einstein–Debye equation, where the molecule is considered as a rigid sphere: r = 4phr3/3kT = hMW/dNAkT h=viscosity r=radius of the molecule (assumed to be spherical) d = density of the molecule (≈103 kg m-3)

T2

-1 is larger for larger molecules NMR linewidth is larger for larger molecules

T2-1 = p1/2

log

(T1

,2)

-log (r)

Under the fast exchange regime, the measured transverse relaxation rate is an average of the transverse relaxation rates of the free and bound protein forms, weighted by their molar fraction. This phenomenon can be used to establish complex formation, but also causes signal broadening in the spectra of the bound form of a protein, thus often requiring ad hoc experimental procedures for the obtainment of 1H-15N correlations at the basis of the chemical shift perturbation mapping approach Under slow exchange signals may become undetectable in HSQC TROSY, Cript/CRINEPT TROSY): “monolateral” chemical shift mapping for complexes up to 1 MDa,

Changes in the overall tumbling time

Optimal MW (kDa)

30-100

>100

CRIPT transfer efficiency

400 KDa

800 KDa

Line narrowing strategies: 2H-enrichment + ad hoc experiments

Chemical shift mapping in slow-exchange

apoHasA in HasA-HasR

3 classes of signals: • Not affected ( < 0.25 ppm) • Disappearing from their original well-resolved position • Behavior not safely defined

HSQC vs. CRINEPT-TROSY

19 kDa

HasR 98 kDa in DPC micelles

Caillet-Saguy, Piccioli, Turano, Izadi-Pruneyre, Delepierre, Bertini, and Lecroisey, JACS 2009

HasA-HasR: interaction surface MONOLATERL INFORMATION

HasR model structure

HasA in the apo “open” conformation

Caillet-Saguy, Piccioli, Turano, Izadi-Pruneyre, Delepierre, Bertini, and Lecroisey, JACS 2009

Intermolecular NOEs

The gold standard for the obtainment of structural information in protein NMR is the Nuclear Overhauser Effect SpectroscopY (NOESY) experiment, in either its 2D homonuclear version or in the 15N- or 13C edited 3D variants, which provides 1H-1H cross peaks for all pairs of protons that fall at short distance (within about 5 Å) one from the other. In a rigid system, translation of NOE intensities into distances is possible thanks to the linear correlation between peak volume and 1/r6 (where r is the proton-proton distance). Dipolar interactions across a binding interface can, in principle, give rise to NOEs, which can thus provide intermolecular distance restraints. However, it is often difficult to determine such NOEs. In the case of a complex, the correlation time for the dipolar interaction can be dominated by the exchange time between the free and bound form. Dissociation rates faster than the tumbling are effective in reducing the dipolar interaction even down to values where the NOE effect is no more measurable. In practice intermolecular NOEs to define the three-dimensional structure of a protein-protein complex are essentially restricted to tight complexes (Kd<10 M). Under these conditions, isotope-edited and -filtered NMR pulse sequences are used to distinguish between inter- and intramolecular NOEs

Intermolecular NOEs may be quenched by fast chemical exchange

h c/r6

c-1=r

-1+ex-1

Now we are interested in measuring intermolecular distances, r. The correlation tome for the intermolecular interaction is determined by : where the former is the correlation time for mulecular tumbling and the second the chemical exchange correlation time. The latter becomes dominant when ex

-1 < r-1

Residual Dipolar Couplings (RDCs)

RDCs provide long range structural restraints for NMR structure determination of macromolecules that are not accessible by most other NMR observables, which are dependent on close spatial proximity of nuclei The measured effect is small and specific experiments have been developed for the measurement of H-N or H-C RDC’s. RDC’s are effective in defining the relative orientation of pairs of nuclei (most commonly 1H-15N of backbone amides) within a molecular frame. RDCs can also be used for the purpose of deriving the relative orientation of two proteins in a complex. An orienting agent is needed. Under fast exchange conditions, in order to overcome the difficulties encountered in obtaining RDCs that emanate from the complex alone, a titration approach can be employed where RDCs are measured in different equilibrium mixtures of the free and bound form. While the RDCs of the free states can be measured directly, the RDCs originating from the bound state will be obtained indirectly by extrapolation of the RDCs in the different equilibrium mixtures.

NMR for the characterization of PPIs: other tools

Paramagnetic derived constraints.

Paramagnetic NMR observables as structural restraints

Observable Dependence upon structural

parameters

Iron-metal distance Angles

Contac shift no no

Pseudocontact shift rMH-3 MH & MH

T1,2-1 contact No no

T1,2-1 dipolar rMH

-6 no

T2-1 Curie rMH

-6 no

Paramagnetic residual dipolar coupling no MH & MH

Cross correlation btw Curie & dipolar relaxation rMH-3 CCR

Suggested readings

Schreiber G, Haran G, Zhou H-X (2009) Fundamental Aspects of Protein-Protein Association Kinetics. Chem. Rev. 109:839-859. Zuiderweg ER (2002) Mapping protein-protein interactions in solution by NMR spectroscopy. Biochemistry 41:1-7. Kleckner IR, Foster MP (2011) An introduction to NMR-based approaches for measuring protein dynamics. Biochim Biophys Acta. 4:942-68. Del Conte R, Lalli D, Turano P (2013) NMR as a tool to target protein-protein interaction. In: Disruption of Protein-Protein Interfaces, Ed. Mangani S. - Springer Heidelberg New York Dordrecht London. pp.: 83-111.