Tải bản đầy đủ - 0 (trang)


Tải bản đầy đủ - 0trang

which the alkyl substituents are attached is asymmetric, both enantiomers of WIN52084 as well as a homologous series of compounds were

evaluated against HRV-14 [22] (Fig. 5). In each case, the S isomer was

considerably more inhibitory than the R, which suggested an enantiomeric effect. Examination of WIN52084 in the pocket clearly showed that

the S-methyl group was in close proximity to a hydrophobic pocket

formed by Leu106 and Ser107 (Fig. 6).

To further analyze the interactions of the methyl group of the two

comformers in the binding site, an energy profiling study was performed.

With the x-ray crystal structure of the S isomer of WIN52084 in the virus

pocket serving as a starting point, a window consisting of all residues

within 8 A˚ of any atom was excised from the starting structure. After

charges had been set on the atoms of the resulting pocket and drug

according to a method in Chem-X [23], and after the hydrogen atoms

had been removed, the intermolecular van der Waals energy was calculated via a 6 –12 function for conformations resulting from the rotation of the oxazoline ring about the bond connected to the phenyl ring,

in increments of 10j. A plot of this function versus the rotation angle

Figure 6 WIN52084 bound to HRV-14. The methyl group on the oxazoline ring

is pointing toward a hydrophobic pocket formed by Leu106 and Ser107.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

showed mainly two peaks at – 90j and 100j. A repeat of this calculation

with the R conformer resulted in a flat valley between 30j and approximately 120j. Similar results were obtained with the R- and S-ethyl

compounds, which showed an even more dramatic pattern that was

significant because the ethyl homologue was more potent (Fig. 7). These

results suggested that the twist angle about the two rings could be an

important factor in determining biological activity. It is possible that

the conformation with the appropriate twist angle may be imposed by

the nature of the binding pocket and that maximum interaction with the

hydrophobic pocket formed by Leu and Ser may also be of importance

[20 – 22].

B. Aliphatic Bridge

The x-ray studies on several analogues in this series of compounds showed

that the chain connecting the isoxazole and phenyl rings adopts a bowed

Figure 7 Plot of energy vs torsion angle from an energy profiling study resulting

from rotating the oxazoline ring of the S isomer of WIN52084 about the phenyl


Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

conformation when bound to HRV-14. It had been assumed that flexibility

of the chain was critical for binding and biological activity. Dynamic

studies by Dr. Andrew McCammon with WIN52084 in HRV-14 revealed

considerable motion of the aliphatic chain during an observation lasting for 10 ps (Fig. 8) (Dr. Andrew McCammon, University of Houston,

personal communication). This result posed several questions regarding

the importance of flexibility vs rigidity of the chain. Would a conformationally rigid chain offer enhanced hydrophobic interactions and consequently improved binding, or are there other factors in the binding

process that would require a flexible chain? To address these issues, several

compounds with rigidity incorporated into the chain were synthesized;

their activity against HRV-14 and HRV-1A examined (Fig. 9) and the

compounds modeled in the respective binding site [24]. WIN54954, which

Figure 8 Molecular dynamics of WIN52084 in HRV-14 during a 10 ps run,

illustrating the movement of the chain. (Courtesy of Andrew McCammon, University of Houston.)

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Figure 9 Table comparing the activity of the E and Z olefin and butyne analogues of WIN54954.

has been clinically evaluated, was used as a comparator. The Z olefin

demonstrated a two- to threefold reduction in activity in comparison to

WIN54954, while the E isomer showed a threefold enhancement in

activity. The potency of the butyne analogue was more than fourfold

greater than that of WIN54954 against HRV-14 and was comparable to

that of the E isomer.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

C. Modeling of Conformationally Restricted


The structures shown in Figure 9 were constructed using WIN54954 as a

template, since its x-ray conformation in HRV-14 had been determined.

The resulting structures were subjected to the Tripos force field (Maximin

2), using Sybyl version 5.41, with default settings. Rotatable bonds in the

alkyl ether chain were defined, and the structures were flexibly fitted to

WIN54954, in virus-bound conformation, for insertion into the HRV-14

binding site (Fig. 10). The optimized fitted structures were inserted into

each serotype by replacement of virus-bound WIN54954. Since the drugbound conformation of the virus binding site with several of the compounds had been determined, revealing only minor variations in compound structure, insertion of the modeled compounds into the binding site

configuration, derived from WIN54954, appeared reasonable.

Two interesting observations emerged from this study. The acetylene

analogue, which was more than fourfold more potent than WIN54954

Figure 10 Overlay of energy-minimized structures of the E and Z isomers and


Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

against HRV-14, was inactive against HRV-1A, and there was a dramatic

difference between the activities of the E and Z olefins against HRV-14.

Because x-ray studies had shown that the HRV-14 binding site was longer

than the 1A site, the results of this study supported the premise that the

activity is dependent on the length of the molecule. The butyne was

modeled in HRV-14, causing no serious steric interactions. However this

was not the case in HRV-1A, where the chlorine atom appeared to interact

with Ile125.

The difference in activity of the E and Z olefins against HRV-14 was

explained by examining the relatively low energy virus-bound conformations. The result of an overlay of WIN-54954 (based on x-ray crystallography data), minimize E- and Z-olefinic structures and the butyne

analogue, suggested that the E isomer showed a reasonable fit while the

Z isomer did not. Furthermore, when the Z isomer was inserted into the

HRV-14 pocket, unfavorable interactions occurred.

The very high minimal inhibitory concentration (MIC) values for the

Z isomer against HRV-14 and HRV-1A may reflect a slow kon in both

cases. The conformational space accessible to the isoxazole of the E and Z

olefins, the butyne, and the three-carbon chained homologue of WIN

54954 by conformational sweep graph and for the Z olefin disclosed a

significant inaccessible region of space, while the butyne, E olefin, and

alkane do not show this deficit. Consequently, binding to this site may be

dependent on conformational permissibility in this region that is required

for entry into the pocket. These results suggested that the activity of these

compounds against the two serotypes is strongly dependent on the

flexibility of the hydrocarbon chain and the ability of the molecule to fit

into the conformational space of both pockets.


Thus far all the compounds that were examined bound to HRV-14, with

the exceptions noted, are oriented with the phenyl ring in a stacking mode

with Tyr128 and Tyr152. Aromatic –aromatic interactions have been

shown to be quite common in protein – protein interactions [25 – 31],

and in many cases have displayed [32,33] an electrostatic component.

Furthermore, such interactions would be expected to contribute extensively to the binding energy [34]. To determine the nature of the aromatic

stacking interactions, an energy profiling study was performed by twist-

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

Tài liệu bạn tìm kiếm đã sẵn sàng tải về


Tải bản đầy đủ ngay(0 tr)