An absolute structure template for a unique voltage-gated sodium channel  binding site

Rex A. Palmer; Brian S. Potter; Michael J. Leach; Terence C. Jenkins; Babur Z. Chowdhry

doi:10.1039/C0MD00043D

View PDF VersionPrevious ArticleNext Article

DOI: 10.1039/C0MD00043D (Concise Article) Med. Chem. Commun., 2010, 1, 45-49

Show CompoundsShow Chemical TermsShow Biomedical Terms

An absolute structure template for a unique voltage-gated sodium channel binding site†

Rex A. Palmer *^a, Brian S. Potter ^a, Michael J. Leach ^b, Terence C. Jenkins ^c and Babur Z. Chowdhry *^b
^aSchool of Crystallography, Birkbeck College, University of London, Malet Street, London, UK WC1E 7HX. E-mail: rex.palmer@btinternet.com; Tel: +44 208 449 1049
^bSchool of Science, University of Greenwich at Medway, Chatham Maritime, Kent, UK ME4 4TB. E-mail: b.z.chowdhry@greenwich.ac.uk; Fax: +44 208 331 9805; Tel: +44 208 331 8208
^cMorvus Technology Limited, Tŷ Myddfai, Llanarthne, Carmarthen, UK SA32 8HZ

Received 12th April 2010 , Accepted 18th May 2010

First published on 9th June 2010

Abstract

The X-ray crystallographic structures of the mesylate salts of a novel voltage-gated sodium channel-binding ligand R-(−)-BW202W92 and its much less active S-(+)-enantiomer (BW203W92) have been determined to establish their absolute configurations. Each enantiomer exists as two distinct atropisomeric forms in the solid state and the crystal structures for each enantiomer are stabilized by quite distinct patterns of intermolecular hydrogen bonding. Such structural differences will influence the pharmacological properties of the enantiomers and hence dictate their contrasting receptor binding properties.

Introduction

R-(−)-2,4-Diamino-6-fluoromethyl-5-(2,3,5-trichlorophenyl)pyrimidine (BW202W92, (I) in Fig. 1) is derived from chemical classes of compounds (triazines and pyrimidines) typified by agents such as lamotrigine (Lamictal™; a clinical antiepileptic agent), and sipatrigine (619C89; a neuroprotective agent).^1,2 Electrophysiological studies^3,4 have been used to demonstrate that BW202W92 is a potent/selective sodium-ion channel blocker with no effect on N, P/Q and T type calcium channels and a weak effect on R type calcium channels. These different kinds of structurally homologous channels are distinguished by their physiological roles, inhibition by specific toxins, high- or low-voltage activation as well as activation/inactivation kinetics.⁵ BW202W92 is also, in an in vivo model of cerebral ischaemic/stroke damage, a potent neuroprotective agent.³ Recent binding and ligand binding displacement studies, using a [³H]-labelled form of BW202W92 as a ligand, have revealed a novel stereoselective drug binding site for sodium channels in rat forebrain synaptosomes.⁶ Although the level of specific [³H]-BW202W92 binding in a standard incubation medium is relatively poor, the addition of low concentrations of tetrodotoxin (EC₅₀ = 2–3 nM) greatly enhances the binding, apparently by increasing the affinity of the implicated binding sites. Tetrodotoxin-dependent binding is markedly stereoselective such that under these experimental conditions the S-(+)-enantiomer (BW203W92, (II) in Fig. 1) is up to 30-fold less potent, and extremely sensitive to inhibition by elevated K⁺ ion concentrations (IC₅₀ = 5.9 mM). The latter effect has been ascribed to changes in membrane potential.⁶ This discovery has extremely important and wide-ranging implications: (i) the identification of a novel sodium channel binding site, and (ii) more importantly, that this site may represent a new target for drug intervention.⁶ A detailed structural study for BW202W92 is now warranted and will be of paramount importance to a broad spectrum of scientific disciplines, including medicinal chemistry, molecular physiology, pharmacology, and biophysics. To this end, we have determined the X-ray crystal structures for the mesylate salts of BW202W92 (R-form) and BW203W92 (S-form) to establish their absolute chiral configurations and critical spatial properties. The mesylate acid addition salts of the two enantiomers crystallized in different space groups, each comprising two crystallographically independent A and B molecules per asymmetric unit. Further, atropisomerism is exhibited by the molecules present in each structural arrangement. Herein we report the crystal structures of both the biochemically active R-form and the relatively inactive S-form and contrast their molecular presentations.


	Fig. 1 (Top) Schematic chemical structure for the sterically hindered rotamers of I and II, respectively. (Bottom) General numbering scheme used for the enantiomeric compounds (R-form = BW202W92, S-form = BW203W92). Note that the pyrimidine ring is N1′-protonated in each of the mesylate acid addition salts.

Results and discussion

Experimental details are given in the ESI.† Crystal and structure refinement data for I and II are summarized in Table 1. Atom coordinates and equivalent isotropic thermal parameters are given in Tables S1A (I) and S1B (II), ESI,† respectively. Molecular geometry data are shown in Tables S2A (I) and S2B (II), ESI† (bond lengths and angles), and Tables S3A (I) and S4B (II), ESI† (torsion angles). Hydrogen bond geometries are tabulated in Tables S4A (I) and S4B (II), ESI,† respectively.

Table 1 Crystal data and structure refinement for the R-form (BW202W92; I) and S-form (BW203W92; II)

Identification code	BW202W92 (R) (I)	BW203W92 (S) (II)
a The CAD4 diffractometer was set to discard extremely weak reflections in order to save collection time and as a precaution against excessive deterioration of the crystal.
Empirical formula	C₁₂H₁₂Cl₃FN₄O₃S	C₁₂H₁₂Cl₃FN₄O₃S
Chemical formula	C₁₁H₉Cl₃FN₄CH₃SO₃	C₁₁H₉Cl₃FN₄CH₃SO₃
Formula weight	417.67	417.67
T/K	293(2)	123(2)
Wavelength/Å	1.54180	0.71073
Crystal system	Monoclinic	Triclinic
Space group	P2₁	P1
a/Å	8.384(2)	7.716(2)
b/Å	16.984(3)	8.120(2)
c/Å	12.480(3)	13.719(3)
α (°)	90	74.91(3)
β (°)	104.14(2)	87.69(3)
γ (°)	90	89.83(3)
Volume/Å³	1723.4(7)	829.2(3)
Z	4	2
d _calcd (mg m⁻³)	1.609	1.673
μ/mm⁻¹	6.238	0.709
F(000)	844	422
Crystal size/mm	0.30 × 0.16 × 0.16	0.30 × 0.20 × 0.20
θ range for
data collection (°)	4.49 to 74.42	2.60 to 27.47
Index ranges	−10 ≤ h ≤ 10	−9 ≤ h ≤ 9
	−8 ≤ k ≤ 21	−10 ≤ k ≤ 10
	−6 ≤ l ≤ 15	−17 ≤ l ≤ 15
Reflections collected	3990	6456
Independent reflections	3192 [R(int) = 0.0496] 81.9%	5302 [R(int) = 0.0733] 93.9%
Completeness to θ = 74.42°^a	81.9%	93.9%
Refinement method	Full-matrix least-squares on F²	Full-matrix least-squares on F²
Data/restraints/parameters	3192/1/450	5302/3/452
Goodness-of-fit on F²	1.042	0.938
Final R indices [I > 2σ(I)]	R ₁ = 0.0490	R ₁ = 0.0614
Final R indices [I > 2σ(I)]	wR₂ = 0.1254	wR₂ = 0.1204
R indices (all data)	R ₁ = 0.0644	R ₁ = 0.1248
R indices (all data)	wR₂ = 0.1324	wR₂ = 0.1468
Absolute structure parameter	0.05(3)	−0.10(9)
Extinction coefficient	0.00150(16)	0.0071(12)
Largest diff. peak and hole e.Å⁻³	0.339 and −0.510	0.384 and −0.351

For the R-form the absolute structure parameter⁷ is 0.05(3), indicating that the absolute configuration has been correctly assigned. Similarly for the S-form the absolute structure parameter⁷ is −0.10(9), which again confirms that the correct absolute configuration has been assigned. Furthermore, independent crystallization of the two enantiomeric molecules in different, and, more importantly, non-centrosymmetric space groups is a strong indication, although not a definite proof, that the R and S forms of this molecule are not easily interconvertible simply through rotation about the central bond.

Fig. 1 shows the formula and atom numbering scheme used for the H[BW202W92]⁺ cation and associated counterion in the mesylate salt. ORTEP/RASTER-3D^8,9 views of the molecules A and B in the R-form and S-form are shown in Fig. 2a and 2b, respectively. To aid comparison, the view direction is shown perpendicular to the pyrimidine ring.


	Fig. 2 ORTEP/RASTER-3D views of the paired A and B molecules in the crystal structures for the (a) R-enantiomer and (b) S-enantiomer, showing the 50% probability thermal motion ellipsoids. A common perpendicular orientation for the pyrimidine rings is used for comparison. For clarity, the mesylate counteranions are not shown.

Molecular geometry

Bond lengths in the R-form structure are determined to approximately within ±0.005 Å and bond angles to ±0.4°; corresponding values for the S-form were ±0.006 Å and to ±0.4°, respectively. Bond lengths and angles in each structure compare well with those found in organic compounds,¹⁰ exhibit no unusual values, and reveal no significant differences between corresponding regions of the different molecules. All ring atoms in each ring of both isomeric forms are coplanar (within 0.004 Å) and all bonded atoms for each ring are also positioned in the corresponding plane (i.e., within 0.004 Å).

Axial chirality and atropisomerism

The presence of two symmetry-independent molecules in the asymmetric unit of a crystal structure provides an ideal opportunity to assess variations in the molecular geometry, which here corresponds to differences in relative orientation for the rings in molecules A and B in both the R- and S-enantiomeric forms. The presence of three bulky flanking groups about the pivot bond C(1)-C(5′), Cl(2) on the benzene ring, and N(4′)H₂ and C(7′)H₂F(7′) on the pyrimidine ring, provides a very high barrier to free rotation such that two stable rotamers can exist without dynamic interchange at room temperature (Fig. 1). Interestingly this behaviour is observed for both the R- and S-enantiomers of this compound, with molecules A and B evident in each crystal structure. Fig. 3(a) shows views of molecules A and B in the R-enantiomer, respectively, viewed along the C(5′)–C(1) bond from the pyrimidine ring side, clearly showing that they are distinct atropisomers of this type. The distinction is characterized by the torsion angle τ = C(2)–C(1)–C(5′)–C(4′), which is −111.9(6)° in molecule A and −69.0(7)° in molecule B. Similarly, Fig. 3(b) shows molecules A and B in the S-enantiomer for which τ = 104.8(6)° in molecule A and 71.7(6)° in molecule B. Thus, both the R- and S-forms exhibit clear atropisomerism within their respective crystal forms. It is evident from Fig. 2 that the presence of three distinct chemical groups provides the basis for axial chirality in these compounds and, importantly, that the correct absolute configurations are assigned to each enantiomer (see crystallographic discussion, below). This is confirmed by the rotation direction of the priority sequence 1[N(4′)]–2[C(6′)]–3[Cl(2)], which is clockwise and anti-clockwise for the R- and S-forms, respectively. Formation of the two atropisomers in each compound presumably occurs during the synthesis procedure. Tetrodotoxin-dependent binding of this agent to rat forebrain synaptosomes has been shown to be stereoselective, as the less active S-(+)-enantiomer BW203W92 is up to 30-fold less potent on a dose basis compared to the active R-(−)-form.⁶ It is also likely that, in addition to this enantioselectivity, one atropisomeric form of the R-isomer (i.e., molecule A or B) may exhibit superior binding to the implicated bioreceptor.


	Fig. 3 Views for A and B molecules respectively in the (a) R-isomer and (b) S-isomer, viewed along the C(5′)–C(1) chiral axis and showing the absolute configurations. The ring planes are viewed edge-on and the interplane torsion is determined by C(2)–C(1)–C(5′)–C(4′).

Hydrogen bonding in the R- and S-forms

Both crystal structures are characterized by distinct patterns of hydrogen bonding. Fig. 4 shows the local H-bonding arrangement between molecule A or B and their associated mesylate counteranions. Fig. 5 reveals characteristically different extended networks of H-bonds for each of the enantiomeric forms. In the R-form, the BW202W92 molecules are H-bonded to the O atoms of the mesylate through N(1′), N(2′) and N(4′). There is also a short contact between F(7′) and N(2′) of a neighbouring molecule. In contrast, for the S-enantiomer, the BW203W92 molecules form complementary hydrogen bonds between their N(4′)H₂ groups in symmetry-related molecules, and also to the O atoms of the mesylate anion through N(1′), N(2′) and N(4′). Knowledge of these interactions will inform an understanding of why the R-form selectively binds to sodium channel receptors.


	Fig. 4 Local H-bonded contacts between cationic molecules A and B and their corresponding mesylate (CH₃SO₃⁻) anions in the crystal structures for (a) the R-form and (b) the S-form.


	Fig. 5 Intermolecular hydrogen-bonded network patterns for the component molecules in the crystal structures for (a) the R-enantiomer and (b) the S-enantiomer.

Molecular modelling

Cronin et al.¹¹ modelled the binding of lamotrigine using the known structure of the central helices on the potassium channel template and a model has also been determined for lamotrigine binding to the sodium channel on the basis of mutation analysis.¹² Currently, however, there is no suitable published structural template available exclusively for the sodium channel and particularly for brain sodium channel subtypes. In the absence of such a template this necessarily precludes any meaningful modelling studies for BW202W92 and BW203W92 receptor interactions.

It is anticipated that molecular modelling using each of the BW202W92 and BW203W92 ligand components would be more informative than previous attempted studies, as all other known members of the lamotrigine family crystallize in centrosymmetric space groups and thus contain equal contributions from their R and S enantiomers. On this basis, models derived from other structures, including lamotrigine itself, will be unable to distinguish possible differences for the R and S modes of binding indicated from the marked contrast in drug behaviour evident for (I) and (II).

Conclusions

It is established from previous work that BW202W92 binds to sodium channels at a novel stereoselective site such that the R-form is biochemically potent whereas its S-enantiomer (BW203W92) is relatively inactive. In addition, other compounds of this lamotrigine family of similar structure also display markedly different therapeutic preferences, e.g. for epilepsy, pain, or neurodegeneration. A major challenge for medicinal chemists is to account for such differences in drug activity at the molecular level. This structural study has highlighted that there are large and significant differences in both local structure and H-bonding for the R- and S-enantiomers of BW202W92. These structural differences are likely to reflect the different pharmacological behaviors of the two agents in vivo and their mechanism(s) for drug interaction at the receptor site. The detailed structural information obtained for the active/inactive forms of the two enantiomers of this drug provide a firm basis to discern subtle differences between agents of the important lamotrigine class and to inform drug design for next-generation agents. In particular, the atropisomerism behaviour can now be exploited in dynamic molecular modelling studies of drug interaction with target receptors to gauge critical stereochemical factors that will be important for QSAR studies.

Acknowledgements

We thank Dr P. Barraclough (University of Greenwich) for the synthesis and provision of samples of BW202W92 and BW203W92 compounds. Low-temperature X-ray intensity data were collected using the EPSRC single-crystal X-ray data facility at the University of Southampton, UK.

References

M. J. Leach, A. D. Randall, A. Stefani and A. H. Hainsworth, in Antiepileptic Drugs; R. H. Levy, R. H. Mattson, B. S. Meldrum, E. Perucca, eds.; Lippincott, Williams & Wilkins: USA 2002, 363–369, 5th edition Search PubMed.
A. H. Hainsworth, A. Stefani, P. Calabresi, T. W. Smith and M. J. Leach, CNS Drug Rev., 2000, 6, 111–134 CAS.
L. Caputi, A. H. Hainsworth, F. Lavaroni, M. J. Leach, N. C. McNaughton, N. B. Mercuri, A. D. Randall, F. Spadoni, J. H. Swan and A. Stefani, Brain Res., 2001, 919, 259–268 CrossRef CAS.
A. H. Hainsworth, N. C. McNaughton, A. Pereverzev, T. Schneider and A. D. Randall, Eur. J. Pharmacol., 2003, 467, 77–80 CrossRef CAS.
S. P. H. Alexander, A. Mathie and J. A. Peters, Brit. J. Pharm. (Guide to Receptors & Channels 4th Ed.) 2009, 158, Supplement 1, S127–S128; S146–S147 Search PubMed.
D. R. Riddall, M. J. Leach and J. Garthwaite, Mol. Pharmacol., 2006, 69, 278–287 CAS.
H. D. Flack, Acta Crystallogr., Sect. A: Found. Crystallogr., 1983, 39, 876–881 CrossRef.
C.L.Barnes ORTEP-3 for Windows-a version of ORTEP-III with a Graphical User Interface (GUI). J. Appl. Cryst., 30 ( 1997), 30, pp. 568–589 [Based on ORTEP-III (v 1.0.3) by C.K. Johnson and M.N. Burnett.] Search PubMed.
E. A. Merrit and D. J. Bacon, Raster 3D Graphics version 2.7c., Methods Enzymol., 1997, 277, 505–524 CAS [Implemented in WinGX (qv) and generated by Ortep-3 for Windows.].
M. F. C. Ladd, R. A. Palmer, Structure Determination by X-ray Crystallography, Klewer-Plenum: N.Y., USA ( 2003) p503, 4th Edition Search PubMed.
N. B. Cronin, A. O'Reilly, H. Duclohier and B. A. Wallace, J. Biol. Chem., 2003, 278, 10675–10682 CrossRef CAS.
V. Yarov-Yarovoy, J. Brown, E. M. Sharp, J. J. Clare, T. Scheuer and W. A. Catterall, J. Biol. Chem., 2000, 276, 20–27 CrossRef.

Footnote

† Electronic supplementary information (ESI) available: Experimental details and Supplementary tables. CCDC reference numbers 299482, 299483. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c0md00043d