Article pubs.acs.org/ac
Higher Order Inclusion Complexes and Secondary Interactions Studied by Global Analysis of Calorimetric Titrations Christian Schönbeck,†,‡ René Holm,*,‡ and Peter Westh† †
NSM, Research Unit for Functional Biomaterials, Roskilde University, Universitetsvej 1, DK-4000 Roskilde, Denmark Preformulation, H.Lundbeck A/S, Ottiliavej 9, DK-2500 Valby, Denmark
‡
S Supporting Information *
ABSTRACT: This paper investigates the use of isothermal titration calorimetry (ITC) as a tool for studying molecular systems in which weaker secondary interactions are present in addition to a dominant primary interaction. Such systems are challenging since the signal pertaining to the stronger primary interaction tends to overshadow the signal from the secondary interaction. The methodology presented here enables a complete and precise thermodynamic characterization of both the primary and the weaker secondary interaction, exemplified by the binding of β-cyclodextrin to the primary and secondary binding sites of the bile salt glycodeoxycholate. Global regression analysis of calorimetric experiments at various concentrations and temperatures provide a precise determination of ΔH, ΔG°, and ΔCp for both binding sites in glycodeoxycholate (K1 = 5.67 ± 0.05 × 103 M−1, K2 = 0.31 ± 0.02 × 103 M−1). The results are validated by a 13C NMR titration and negative controls with a bile salt with no secondary binding site (glycocholate) (K = 2.96 ± 0.01 × 103 M−1). The method proved useful for detailed analysis of ITC data and may strengthen its use as a tool for studying molecular systems by advanced binding models.
A
complexes in addition to 1:1 complexes. If the secondary interaction is much weaker than the primary, the two binding events occur in different concentration ranges and may be studied separately, but if the magnitudes of the two binding constants are not sufficiently different, the binding isotherms for the two binding events will overlap and may be hard to distinguish. The latter situation has been treated for various experimental techniques, including liquid chromatography,2 affinity capillary electrophoresis,3 NMR titration,4 selective electrodes,5 and isothermal titration calorimetry (ITC).6 ITC has the fundamental advantage over the other techniques that it directly measures the change in site occupancy resulting from a small change in concentration. Other techniques measure the total amount of bound or free species. Conversely, the interpretation of ITC is limited by the detection of the cumulated heat from different binding events in contrast to, e.g., NMR, where the binding to different sites may be observed separately. In cases where two sets of reasonably strong sites occur, they can be readily resolved by ITC.7,8 However, weaker effects are often discarded as “heat of dilution” or experimental noise, so the challenge is to separate this from the useful information. The extraction of reliable binding parameters may be facilitated by a global analysis in which experimental data
large number of physical techniques are available for the determination of binding constants in molecular complexes. In all of these techniques, the basic principle is to vary the relative concentrations of the binding species and measure the resulting change in one or more observables that are sensitive to complex formation, e.g., absorbance, fluorescence, electrophoretic mobility, heat content, or NMR chemical shifts.1 The resulting binding isotherm is then fitted with a mathematical model to yield the binding constant. The most common model is based on the assumption that only one type of equal and independent binding sites is present, which is probably a reasonable assumption in many cases. However, there are situations where the presence of multiple equilibria is significant, and it is no longer reasonable to assume that only one type of sites is present. More complex models have been developed to describe these situations1 but as the complexity of the model increases so does the number of fitting parameters, and as a consequence, these models are capable of fitting almost any binding isotherm. A good fit of the model to the experimental data is not a proof that the chosen model is a reasonable description of the binding process. Especially in the case where a primary binding interaction dominates, weaker secondary binding interactions may drown in the signals from the primary interaction. In such cases, both the simple 1 site model and more complex models are capable of fitting the binding isotherm and the detection and quantification of weaker secondary interactions may be extremely difficult. A typical example is the formation of weaker higher-order © 2012 American Chemical Society
Received: November 10, 2011 Accepted: January 31, 2012 Published: January 31, 2012 2305
dx.doi.org/10.1021/ac202842s | Anal. Chem. 2012, 84, 2305−2312
Analytical Chemistry
Article
55 °C. BSs are known to form micelles in aqueous solution, and the critical micelle concentrations of GC and GDC are 4 and 2 mM, respectively.20 Thus, self-association of BSs should not occur in the investigated concentration range, and this is confirmed by blank titrations of BS into buffer and by NMR (see below). A total of 36 titrations were conducted. Each titration consisted of one 2 μL injection and 27 × 10 μL injections. The peak resulting from the first injection was subsequently ignored. Heats of dilution were measured in the same way by titrating 10 mM βCD into buffer and subtracted from the enthalpograms. Examples of raw data for the titrations of βCD into BS and buffer are shown in Figures S-1 and S-2, Supporting Information. NMR Spectroscopy. Stock solutions were made by weighing dried βCD and GDC and dissolving in 50 mM phosphate buffer in D2O. Mixtures at various molar ratios were made by mixing the two stock solutions in the NMR tubes. Preliminary investigations indicated the formation of BS micelles at concentrations >1.5 mM. Therefore, the solutions were designed such that the concentration of noncomplexed GDC was