Induced Fit—The Key for Understanding LSD Activity? A 4D-QSAR Study on the 5-HT2A Receptor System

Abstract

Using a 4D-QSAR approach (software Quasar) allowing for multiple-conformation, orientation and protonation-state ligand representation as well as for the simulation of induced-®t phenomena, we have validated a family of receptor surrogates for the 5-HT 2A receptor system. The evolution was based on a population of 200 receptor models and simulated during 6,000 cross-over steps, corresponding to 30 generations. It yielded a cross-validated r 2 of 0.951 for the 23 ligands of the training set and a predictive r 2 of 0.859 for the 7 ligands of the test set. In this simulation, all ligand molecules were represented by four different conformers, obtained from a Monte-Carlo search in implicit aqueous solution. A series of six scramble tests (with an average predictive r 2 of À1.05) indicate a high sensitivity of the surrogate family towards the biological data.The quantitative analysis of the contribution of the individual functional groups to the free energy of ligand binding, DG , reveals that the key factors for strong bindingÐand hence activityÐare the ligand desolvation energy and the costs associated with induced ®t, the adaptation of the receptor-binding site to the ligand topology. While the ammonium functionality is essential for recognition, its contribution to DG is not favorable due to a high desolvation energy; important groups are rather methoxy and halide substituents. For most ligand molecules, the evolution does not select the lowest-energy conformer, contrary to previous assumptions in a corresponding 3D-QSAR study.

Two Decades of 4D-QSAR: A Dying Art or Staging a Comeback?

“…The manifestation/magnitude of the local induced fit (the ligand-mediated adaptation of binding pocket to molecular topology/topography) or simulation of the H-bond flip-flop particles (Ser, Thr, Tyr, Cys, His, Asn or Gln amino-acids acting simultaneously as HB donors/acceptors due to a conformationally flexible H-bonding functions) are still challenging issues in the rational drug design. In consequence, the multiple ligand conformation, orientation and protonation representation has been enhanced by an additional level of model abstraction (degree of freedom)—the topology of the quasi-atomistic receptor surrogate [ 98 , 99 ]. The conceptual cascade of the Quasar 4D-QSAR is presented in Figure 9 .…”

Novel Ligands for the Chemokine Receptor-3 (CCR3): A Receptor-Modeling Study Based on 5D-QSAR

“…This alignment problem has long been recognized, [9][10][11] and two principle ways to circumvent it have been explored: in situ generation of the active conformer 16 and multiple ligand representation (4D-QSAR). [17][18][19][20][21][22] The 4D approaches represent the ligand molecules (of both training and test sets) as an ensemble of conformations, orientations, and protonation states. The most likely bioactive representation may then be genetically evolved from this reservoir, for example, using a Boltzmann-weighted selection criterion.…”

“…[19][20][21][22] An adequate treatment of conformationally flexible H-bond donor or acceptor moieties at the true biological receptor, able to engage in differently directed hydrogen bonds with dissimilar ligand molecules, may be simulated through the definition of "H-bond flip-flop" particles (properties) when using quasi-atomistic receptor models. [19][20][21][22] Even with 4D-QSAR, a major unknown persists: manifestation and magnitude of the induced fit, the ligand-induced adaptation of the binding site to the topology of the small molecule. In the absence of the true biological receptor, these quantities cannot unambiguously be determined; the most recently developed 5D-QSAR tools offer a possible solution to the problem; here, several induced-fit scenarios (cf.…”

5D-QSAR: The Key for Simulating Induced Fit?

“…If the underlying pharmacophore hypothesis is based on incorrect assumptions, the resulting surrogate is hardly of any use for predictive purposes. While the alignment problem has long been recognized, − , only the more recently developed 4D-QSAR technologies would seem to provide decent solutions. − In explicit 4D-QSAR approaches, the ligands of both training and test set are provided as an ensemble of conformations, orientations, and protonation states. The most likely bioactive representation is then genetically evolved from this reservoir using a Boltzmann-weighted selection criterion. − An adequate simulation of conformationally flexible H- bond donor or acceptor moieties at the true biological receptor, able to engage in differently directed H bonds with dissimilar ligand molecules (i.e., Ser, Thr, Tyr, Cys, His, Asn, and Gln residues), has become possible with the introduction of H-bond flip-flop particles in quasi-atomistic receptor models. , Inhibitor-dependent H- bond flip-flop has been observed, for example, in purine nucleoside phosphorylase …”

“…Ligand-dependent induced fit has been experimentally detected, for example, in the family of the serine proteases . In the context of quasi-atomistic binding site models, − it would seem necessary to emphasize that the simulation of a receptor-to-ligand adaptation is limited to rather small shifts (rms < 4.0 Å), which implies that more pronounced movements cannot presently be accounted forparticularly, if the large differences are a consequence of the ligand-induced triggering mechanism. Fortunately, the genetic algorithm typically fails under such boundary conditions, i.e., the evolution comes forth only very slowly, if at all.…”