Isothermal titration calorimetry

The binding affinity for the inhibitors of penicillin acylase was measured using kinetic data. Isothermal titration calorimetry (ITC) is a more general method for determining binding constants which does not depend on catalytic activity. It also has the considerable advantage of providing a measure of the affinity and enthalpy ofbinding in a single experiment.

A schematic diagram of an ITC experiment is shown in Figure 2. A solution of one reactant (usually the protein) is maintained at a constant temperature in a sample cell. This is kept warmer than the surrounding isothermal jacket so that there is a constant thermal gradient and thus a constant input of energy is required to the sample heating coil. A concentrated solution

Figure 1. Ligand binding to penicillin acylase [2]. (a) Structure of the enzyme determined with phenylacetic acid (yellow) bound (pdb code: lpnl). Red is the A chain, blue the B chain of the enzyme, with the active site serine and oxyanion hole coloured yellow. (b) Detail of the active site of the unliganded structure (pdb code: lpnk), showing the active site serine. The pink grid surface shows the solvent accessible surface calculated with a probe of 1A. (c) Overlay of the structures of phenylacetic acid derivatives in complex with the enzyme. Blue = m-nitrophenylacetic acid (pdb code: 1ai5) andp-hydroxyphenylacetic acid (pdb code: 1ai6).

Figure 1. Ligand binding to penicillin acylase [2]. (a) Structure of the enzyme determined with phenylacetic acid (yellow) bound (pdb code: lpnl). Red is the A chain, blue the B chain of the enzyme, with the active site serine and oxyanion hole coloured yellow. (b) Detail of the active site of the unliganded structure (pdb code: lpnk), showing the active site serine. The pink grid surface shows the solvent accessible surface calculated with a probe of 1A. (c) Overlay of the structures of phenylacetic acid derivatives in complex with the enzyme. Blue = m-nitrophenylacetic acid (pdb code: 1ai5) andp-hydroxyphenylacetic acid (pdb code: 1ai6).

Table 1. Chemical structures of the phenylacetic acid derivatives whose structure has been determined complexed to penicillin acylase, together with measured inhibition constants [2]

Compound

Structure

Ki (mM)

thiopheneacetic acid

rrv s^i 0

0.062

phenylacetic acid

cnr

0.072

phenol

a"

0.12

m -11 itro phenyl acetic acid

\yr

0.19

p-hydroxyphenylacetic acid

jOnr

0.23

3,4-di hydroxy phenylacetic acid

rcnr

1.06

2,5-dihydroxyphenylacetic acid

OX

1.22

U 0

3.84

of the ligand is then injected. If the binding of the ligand to the protein is exothermic, then heat is liberated; if it is an endothermic process then heat is absorbed. The change in the amount of energy required to keep the sample cell at the constant temperature can then be measured. A series of injections is made and the sample allowed to return to thermal equilibrium after each. The total heat released or absorbed is then determined. A typical trace is shown in

1,0 1.5 20 Molar Ratio

Figure 3. Typical output from an ITC experiment (in this case the binding of the ligand KNapK to OppA, measured at pH 7 and 22 °C).

Figure 3 - as the ligand saturates the protein binding site, so the calorimeter trace decreases and the final spikes in the trace show the heat of dilution of the ligand in the sample buffer. From the shape of the curve a binding affinity can be estimated and from the amount of heat evolved the enthalpy of the interaction can be calculated. Knowing both AG and AH, the entropy of the interaction can also be calculated.

The measurement is limited by the sensitivity of the thermocouples and the amount of heat generated or taken up during the reaction. This gives two other requirements:

6. The binding of ligand to protein must involve a measurable enthalpy change.

7. The dissociation constant for the interaction should be in the 10-8 to 10-3 M range.

The next exploratory system was thrombin. Interestingly, Weber and co-workers [3] have studied the binding of a series oftripeptides with varying primary amine side chains in the P1 position. They determined structures and used kinetics to measure Ki, but were unable to rationalise the full range of observed affinities. Our collaborators, Glaxo-Wellcome, had available supplies of the enzyme and also a series of ligands associated with a drug discovery programme. For one of these ligands, argatroban, we are able to measure binding with calorimetry. However, this is a ligand with a K approaching nanomolar, beyond the sensitivity of calorimetry. Unfortunately, most of the other available ligands were not very soluble in aqueous solution. They could be taken up using DMSO but the heat of dilution would then be very large and mask the enthalpy of binding of the ligand. This gives another requirement for this type of experiment:

8. Both protein and ligand need to be highly soluble in aqueous buffers.

We also explored measuring the binding ofthe neu-5-Ac-2-en series ofinhib-itors to the influenza virus enzyme, sialidase [4]. Although it was possible to obtain good quality calorimetry traces for this system, it proved too expensive to produce the large quantities of the enzyme that were required for a systematic study. In addition, it was not possible to develop a satisfactory protocol for recovery of the enzyme after calorimetry. This adds a more focussed requirement on protein availability as:

9. Large quantities ofprotein available, recoverable post-calorimetry.

Figure 4. Secondary structure schematic diagram of the structure of the OppA protein complexed with KnapK [6] as pink van der Waals spheres (pdb code: 1b0h).
0 0

Post a comment