Relative configuration and conformation of key intermediates for matrix metalloproteinase inhibitors as determined by NMR-based restrained simulated annealing and validated by X-ray analysis [2]

Full text of DOI:10.1021/ja982415e

11512
J. Am. Chem. Soc. 1998, 120, 11512-11513
Relative Configuration and Conformation of Key In-
termediates for Matrix Metalloproteinase Inhibitors
as Determined by NMR-Based Restrained Simulated
Annealing and Validated by X-ray Analysis
Hans Matter,* Martin Knauf, Wilfried Schwab, and
Erich F. Paulus
Hoechst Marion Roussel
Core Research Functions & Chemical Research
D-65926 Frankfurt am Main, Germany
ReceiVed July 9, 1998
ReVised Manuscript ReceiVed September 8, 1998
The information about relative and absolute configuration of
stereogenic centers in natural products and synthetic key inter-
mediates is indispensable for modern medicinal chemistry. To
achieve spatial complementarity between ligands and specific
target enzymes, an understanding of stereoselective transforma-
tions is required. Here we describe the application of a strategy
to simultaneously determine configuration and conformation of
key intermediates in the synthesis of enzyme inhibitors using 2D
NMR spectroscopy and simulated annealing1 including a floating
chirality approach.2
Since matrix metalloproteinases are of growing interest in
pharmaceutical research,3 we have investigated key intermediates
for stromelysin (MMP-3) inhibitors with S1′ directed biphenyl-
sulfonyl side chains.4 Their relative configurations were unam-
biguously obtained using this approach and could be validated
using X-ray structure analysis. To our knowledge, this is the
first example of this strategy combined with an independent
validation showing its potential to rapidly obtain relevant stere-
ochemical information.5
The synthetic route (cf. Figure 1) started from (R)-tetrahy-
droisoquinolin-3-carboxylate 1, which was N-sulfonated.6 2 was
converted to the aldehyde 4 by reduction to 3 using BMS7 and
oxidation using o-iodoxybenzoic acid IBX.8 The loss of optical
rotation from 3 to 4 indicates racemization despite the extremely
mild conditions. Due to lack of stability, 4 was directly converted
to the configurationally stable N-methylnitrone 5. Heating a
solution of 5 resulted in 1,3-dipolar cycloaddition to give the
desired isoxazolidines 6 in a 48.6:51.4 ratio of 6a (RSS/SRR, trans)
and 6b (RSR/SRS, cis). The low cis/trans stereoselectivity has
been overcome in special cases.9 The stereodirecting effect of
R-chiral nitrones during olefin cycloaddition was described using
different transition state models.10 Our data suggest that cycload-
dition of tetrahydroisoquinolin nitrones preferably results in 4R/
3S (4S/3R) stereochemistry.
Figure 1. Synthesis of both diastereomeric MMP-3 inhibitor key
intermediates 6a and 6b.
1
After the assignment of H NMR resonances11 for 6a and 6b
(cf. Supporting Information), distance constraints were extracted
from 2D NOESY and ROESY spectra by conversion of cross-
peak volumes into interproton distances using the isolated spin
pair approximation (ISPA). Thirty-one and 26 nontrivial distance
constraints were obtained for 6a and 6b, respectively.12 In
addition, homonuclear J-coupling constants (cf. Supporting
Information) were converted into dihedral angle information.13
For both compounds the NMR spectra indicate the existence of
a single preferred conformation.14
The NOE restrained simulated annealing calculations1 were
performed using SYBYL.15,16 Since it is not possible to determine
absolute configurations in distance space, carbon C4 was set to
R-chirality. Thus, four different starting configurations for
carbons C4, C3, and C1 (RSS, RSR, RRS, and RRR) were
considered for each intermediate. The NMR-derived distance
constraints were applied as a biharmonical function. An energetic
force field was used where experimentally derived distance
constraints have higher force constants than the terms maintaining
individual atom chiralities. Thus, chiral centers are allowed to
invert during the simulation. For each starting configurations,
50 structures were calculated, which were all combined for
analysis yielding 200 individual structures for 6a and 6b,17
respectively.
Acceptable structures were selected on the basis of the
maximum pairwise rmsd violations18 as objective criteria. For
each structure, the maximum rmsd to all structures with lower
(10) (a) Vasella, A.; Voeffray, R. HelV. Chim. Acta 1982, 65, 1134-1144.
(b) DeShong, P.; Li, W.; Kennington, J. W.; Ammon, H. L. J. Org. Chem.
1991, 56, 1364-1373. (c) Baggiolino, E. G.; Iacobelli, J. A.; Hennessy, B.
M.; Batcho, A. D.; Sereno, J. F.; Uskokovic, M. R. J. Org. Chem. 1986, 51,
3098-3108.
(11) NMR samples were prepared by dissolving 6a and 6b in 0.6 mL of
CDCl3. NMR spectra were obtained in the phase sensitive mode on a BRUKER
DRX 600 spectrometer at 300 K using TPPI for quadrature detection in F1.
For ROESY experiments, a mixing time of 150 ms was used for both epimers.
The mixing times for the NOESY experiments of 6a and 6b were set to 600
ms and 1 s, respectively.
(12) For nondiastereotopically assigned CH2 protons and methyl groups,
90 and 100 pm were added to the upper bounds as pseudoatom corrections
(Wu¨thrich, K. NMR of Proteins and Nucleic Acids; Wiley: New York, 1986).
Upper and lower limits have been set to (10% of the calculated distances.
(13) A Karplus-type of equation (Karplus, M. J. Chem. Phys. 1959, 30,
11-15) with the following parameters was used: A ) 9.5, B ) -1.6, C )
1.8 (DeMarco, A.; Llinas, M.; Wu¨thrich, K. Biopolymers 1978, 17, 617-
636).
(14) The diastereotopic protons show significant chemical shift differences
and different vicinal coupling constants. The line shapes of the resonance
signals do not indicate any conformational heterogeneity (Kessler, H. Angew.
Chem., Int. Ed. Engl. 1982, 21, 512-523).
(15) SYBYL Molecular Modelling Package, Versions 6.3, Tripos, St. Louis,
MO, 1996.
(16) All energy calculations were based on the TRIPOS 6.0 force field
(Clark, M.; Cramer, R. D., III; Van Opdenbosch, N. J. Comput. Chem. 1989,
10, 982-1912) including Gasteiger-Marsili charges (Gasteiger, J.; Marsili,
M. Tetrahedron 1980, 36, 3219-3228).
* To whom the correspondence should be addressed. Phone: ++49-69-
305-84329. Fax: ++49-69-331399. E-mail: hans.matter@hmrag.com.
(1) Nilges, M.; Clore, G. M.; Gronenborn, A. M. FEBS Lett. 1988, 229,
317-324.
(2) Weber, P. L.; Morrison, R.; Hare, D. J. Mol. Biol. 1988, 3, 483-487.
(3) (a) Zask, A.; Levin, J. I.; Killar, L. M.; Skotnicki, J. S. Curr. Pharm.
Des. 1996, 2, 624-661. (b) Hagmann, W. K.; Lark, M. W.; Becker, J. W.
Ann. Rep. Med. Chem. 1996, 231-240. (c) Ye, Q. Z.; Hupe, D.; Johnson, L.
Curr. Med. Chem. 1996, 3, 407-418.
(4) Esser, C. K.; Bugianesi, R. L.; Caldwell, C. G.; Chapman, K. T.; Durette,
P. L.; Girotra, N. N.; Kopka, I. E.; Lanza, T. J.; Levorse, D. A.; MacCoss,
M.; Owens, K. A.; Ponpipom, M. M.; Simeone, J. P.; Harrison, R. K.;
Niedzwiecki, L.; Becker, J. W.; Marcy, A. I.; Axel, M. G.; Christen, A. J.;
McDonnell, J.; Moore, V. L.; Olszewski, J. M.; Saphos, C.; Visco, D. M.;
Shen, F.; Colletti, A.; Krieter, P. A.; Hagmann, W. K. J. Med. Chem. 1997,
40, 1026-1040.
(5) For a pioneering approach, see: Reggelin, M.; Hoffmann, H.; Ko¨ck,
M.; Mierke, D. F. J. Am. Chem. Soc. 1992, 114, 3272-3277.
(6) Experimental conditions, see Supporting Information.
(7) Lane, C. F. Aldrichim. Acta 1977, 10, 41-51.
(8) (a) Frigerio, M.; Santagostino, M.; Sputore, S.; Palmisano, G. J. Org.
Chem. 1995, 60, 7271-7276. (b) Plumb, J. B.; Harper, D. J. Chem. Eng.
News 1990, 3.
(9) Mukai, C.; Kim, I. J.; Cho, W. J.; Kido, M.; Hanaoka, M. J. Chem.
Soc., Perkin Trans. 1 1993, 2494-2503.
10.1021/ja982415e CCC: $15.00 © 1998 American Chemical Society
Published on Web 10/27/1998

Communications to the Editor
J. Am. Chem. Soc., Vol. 120, No. 44, 1998 11513
Figure 3. Comparison between representative solution structures (white
carbon atoms) and solid-state structures (grey carbon atoms) for diaster-
eomers 6a (left) and 6b (right).
for epimer 6a is slightly violated by 0.11 Å (averaged over 20
conformers). A similar agreement is observed for the second
epimer 6b, although the resonance overlap of protons H2, H3,
and H5 led to a higher uncertainty in the NOE distance restraints,
causing a higher NOE violation for all 28 RSR conformers.
Figure 2. Results of the NMR-based analysis of both epimers. In the
upper left panel, the conformational ensemble of 20 acceptable structures
for 6a (RSS chirality) is given. In the lower left panel, the maximum of
the rms deviation is shown between pairs of conformers with a rms NOE
violation smaller than a cutoff value as objective criterium for ensemble
selection. The sample of structures for rms deviation gets progressively
larger with increasing cutoff values for the NOE violation, as plotted on
the x-axis. On the y-axis, the maximum rms deviation values are given.
In the upper right part, the ensemble of 28 acceptable structures for
diastereomer 6b (RSR chirality) is shown in combination with the
corresponding maximum rmsd graph (lower right panel).
For 6a, strong NOE between H1 and H4 and a weaker
correlation for H1 and H3 count for a H1-H3 trans orientation.
However, the full differentiation between all possible stereoiso-
mers at C4, C3, and C1 could only be achieved in a straightfor-
ward manner using this combined experimental and computational
approach. The situation was even more difficult for 6b since no
unambiguous NOE effects between protons H1-H4 and H1-
H3 were observed, whereas the calculation unambiguously led
to the RSR diastereomer with a cis orientation of H1 and H3.
constraint violations was plotted on the y-axis as a function of
the constraint violation (Figure 2, lower panel). For a conformer
family, the maximum rmsd raises continuously as a function of
the violation of the experimental data. However the scatter plots
in Figure 2 reveal significant steps, indicating that new confor-
mational and/or configuration states become accessible, which
differ significantly from all structures with lower violations. These
plateaus with higher violations and rmsd values correspond to
alternative configurations at carbons C1 and/or C3.
For 6a, 69 structures with low NOE violations are acceptable.
All reveal RSS chirality at C4, C3, and C1 (Figure 2, box in lower
panel). Interestingly, 40% of those structures were obtained from
different starting configurations by inversion of chiral centers.
In fact, the structure with the lowest NOE violation results from
an RSR starting diastereomer. In contrast, all diastereomers
starting with RSS maintain their chiralities. For analysis, the first
well-defined plateaus in Figure 2 (left) with rms NOE violations
between 0.03 and 0.04 Å were considered, leading to 20
representative structures for 6a (Figure 2, upper left panel). The
second plateau for 6a with 49 conformers also corresponds to
RSS configurations but now represents another conformational
state characterized by a different orientation of the biphenyl
moiety.
Subsequently an X-ray structure analysis19 for validation
confirmed the relative configurations for both epimers obtained
by NMR. The comparison between solid state and solution
structures for 6a and 6b is displayed in Figure 3. The only
remarkable difference between solution and solid-state structures
is the orientation of the biphenyl moiety in both epimers, which
might result from crystal packing effects. Concerning the
heterocyclic system, conformations and distances are very similar.
This validated approach for the simultaneous determination of
conformation and configuration rapidly and efficiently deduces
the correct relative stereochemistry. As earlier proposed,5 force
field modifications allowing chiral centers to convert produce
configurations consistent with all experimental data. The unbiased
structure analysis based on NOE rms violation of individual
conformers unambiguously allows the selection of acceptable
structures in accord with experimental data. For further testing
during implementation of this strategy, compound 1 of reference
5 was used, leading to identical results as described therein. Thus,
our results clearly demonstrate the potential of this powerful
approach to obtain relevant information in a timely and straight-
forward manner.
The same analysis for 6b revealed the RSR configuration at
C4, C3, and C1. Again, no chirality inversion was observed for
the RSR starting structures. The most acceptable 28 conformers
and the maximum rmsd plots are displayed in Figure 2 on the
right.
All NOE-derived constraints are fulfilled for both epimers (cf.
Supporting Information). Only one NOE in the biphenyl residue
Acknowledgment. The work of Holger Thiel during the synthesis of
the herein described molecules is gratefully acknowledged.
(17) A time step of 1.0 fs was used for the integration of Newton’s equation
of motion for a duration of 200 ps each. The kinetic energy was included by
coupling the system to a thermal bath (Berendsen, H. J. C.; Postma, J. P. M.;
van Gunsteren, W. F.; Di Nola, A.; Haak, J. R. J. Chem. Phys. 1984, 81,
3684-3690). The atomic velocities were applied following a Boltzmann
distribution about the center of mass to obtain a starting temperature of 1000
K. The force constants for upper and lower boundaries were initially set to
50 kcal/mol‚Å2. After simulating for 1.5 ps at 1000 K, the system temperature
was stepwise reduced using an exponential function over a 2.5 ps period to
reach a final temperature of 200 K. Resulting structures were sampled every
4 ps and minimized resulting a protocol of 50 cycles with 4 ps simulation
time for each of the 4 starting isomers. The total CPU time per diastereomer
(200 structures) was approximately 3 h. per diastereomer (200 structures) was
approximately 3 h.
Supporting Information Available: Experimental conditions and
four additional tables with 1H chemical shifts, homonuclear experimental
and backcalculated 3J(H,H) coupling constants, and experimentally derived
distance constraints in comparison to computed averaged distances for
6a and 6b (5 pages, print/PDF). See any current masthead page for
ordering information and Web access instructions.
JA982415E
(18) Widmer, H.; Widmer, A.; Braun, W. J. Biomol. NMR 1993, 3, 307-
(19) Experimental conditions, see Supporting Information. Paulus, E. F.;
Schwab, W.; Knauf, M.; Matter, H. 1998, in preparation.
324.