Analysis Using a Gel-Stretching Device NMR Structure...
1
Accurate Measurement of RDC and RCSA for Small Molecule Configuration Analysis Using a Gel-Stretching Device Yizhou Liu, Gary E. Martin, and R. Thomas Williamson NMR Structure Elucidation, Process and Analytical Chemistry (Rahway), Merck & Co. Inc., Kenilworth, NJ, USA liuyiz_189821-0001_ENC_poster_v1.03 04/07/2016 ENC Meeting, Output Size: 44” x 44” Scale: 200% Introduction Stereochemical determination of small drug-like compounds represents a frequent challenge in the pharma industry. Conventionally, we have relied on NOE/ROE and J coupling data in combination with DFT (Density Functional Theory) modeling for such work. However, the limit of this approach manifests when the chiral centers are separated beyond the scope of these short-ranged data. This limitation prompted us to develop techniques that provide long- range structural information. While both RDC and RCSA are known to serve this purpose, RCSA is particularly attractive for proton-deficient molecules in which measurable RDCs are scarce. However, the application of RCSA for small molecules is rare. The main hurdle for RCSA is the lack of a convenient and reliable approach to eliminate isotropic chemical shift changes upon molecular alignment. Here we describe a simple gel-stretching device that provides accurate and clean measurement of RCSA data and allows RDC measurement as well. The utility is demonstrated for several natural products including strychnine, estrone, and retrorsine. Design of the Gel-Stretching NMR Tube The stretching device is an NMR tube with different inner diameters (ID) on two open ends. This design is adopted from a previous report that used a two-stage NMR tube for RCSA measurement in proteins 1 but with improvement for better durability. Isotropic data are collected with the gel in the wide segment (4.2mm ID) while anisotropic data are collected with the gel stretched in the narrow segment (3.0-3.4 mm ID) of the tube. Shimming is easy and good field homogeneity can be obtained. We used the chloroform-compatible PMMA gels 2 for all studies. Isotropic Stretched Spinner here Piston ID: 4.2 mm OD: 5.0 mm ID: <4.2 mm OD: 5.0 mm Wrap with parafilm Rubber plug A. ID: <4.2 mm OD: 5.0 mm Wrap with parafilm B. Figure 1: The design of the 2-stage NMR tube for gel stretching and measurement of RDC and RCSA. A. The setup for the isotropic measurement. B. The setup for the anisotropic measurement. The gel is artificially colored for visualization. References: 1. Liu and Prestegard, J Biomol NMR. (2010), 47, 249-258. 2. Gayathri, Tsaresvsky, and Gil, Chemistry. (2010), 16, 3622-6. Representative Spectral Data and Configuration Analysis Results for Strychnine N N O O H H H H H 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 4 3 1 2 22 12 23 16 8 20 18 14 11 18 20 11 15 17 15 13 A. C. B. 169.4 169.2 (ppm) 64.6 64.4 (ppm) 140.6 140.4 (ppm) 133.0 132.8 (ppm) 122.4 122.2 (ppm) C10 C21 C6 C1 C23 C16 C8 C17 C14 60.2 60.0 (ppm) 42.6 42.4 (ppm) 31.6 31.4 (ppm) Figure 2: Experimental RDC and RCSA data A. Strychnine structure. B. Overlay of J-resolved BIRD-HSQC experiments collected under isotropic (red) and anisotropic (blue) conditions. PMMA signals are enclosed in rectangles. C. Overlays of 13 C spectra under isotropic (red) and anisotropic (blue) conditions, showing RCSA effects at representative carbons; carbonyl and aromatic carbons (top row) tend to have larger RCSA values than aliphatic carbons (bottom row). 20 10 0 -10 -20 -20 -10 0 Exp Value (Hz) 10 20 -5 0 5 Exp Value (Hz) 10 15 Back-calc Value (Hz) 15 10 5 0 -5 Back-calc Value (Hz) -20 -10 0 Exp Value (Hz) 10 20 20 10 0 -10 -20 Back-calc Value (Hz) RDC Q=0.07 S ZZ =-4.9e-4 η=0.51 α=-144° β=-169° γ=30° RCSA Q=0.05 S ZZ =-5.4e-4 η=0.43 α=-158° β=-170° γ=23° S ZZ =-5.1e-4 η=0.48 α=-147° β=-169° γ=27° RDC+RCSA Q=0.08 Figure 3: Evaluation of data quality, showing correlations between experimental values and SVD back-calculated values, showing results for RDC data only (left), RCSA data only (center), and RDC-RCSA data combined (right). The asymmetry parameter η is defined as (|S yy |-|S xx |)/|S zz |. A. B. Q-factors 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0.3 0.25 0.2 0.15 0.1 0.05 0 1a 1b 2 3 4 5 6 7 8 9 10 11 12 13 RDC RCSA RDC+RCSA *Labels the site of inversion. Figure 4: Using RDC and RCSA data for stereochemistry determination. A. Structures of strychnine (1) and its 12 diastereoisomers with inverted stereochemistry. B. Q-factors of RDC alone, RCSA alone, and RDC-RCSA combined are calculated for lowest-energy DFT structures. Enhancing Stereochemistry Discrimination by “Maximally Superimposable Differentiation” (MSD) Maximally superimpose two candidates and only use RDC/RCSA data of the superimposable parts for alignment tensor determination. Then back-calculate RDC/RCSA values for the non- superimposed parts and obtain Q free factors and correlation plots. The subscript “free” means these data were not used in alignment tensor determination. The difference in Q free is expected to be more significant than the difference in Q. More importantly, the numeric differences in Q free factors can be more readily rationalized from a structure perspective by identifying outliers in the correlation plot. Calc Value (Hz) 20 10 0 -10 -20 Pred (Hz) 20 15 10 5 0 -5 -10 -15 20 10 0 -10 -20 20 15 10 5 0 -5 -10 -15 Calc Value (Hz) Pred (Hz) -20 -10 0 Exp Value (Hz) 10 20 -15 -10 -5 0 Exp (Hz) 5 10 15 20 -20 -10 0 10 20 -15 -10 -5 0 5 10 15 20 “Superimposable” (SI) “Non-SI” * Q=0.07 Q=0.13 Q free =0.11 Q free =0.23 A. B. Struct 1 Struct 1 Struct 4 Struct 4 MSD MSD Outlier from the site of inversion Correlation of “non-SI” part Figure 5: MSD analysis for strychnine diastereomers 1 (correct) and 4 (incorrect). Red dots: RDC; green/blue dots: RCSA. A. Maximal superposition of 1 and 4. B. Correlation plots from SVD fail to reveal logical outliers in 4. MSD clearly reveals the RDC at the center of chiral inversion as an outlier. 2 “Superimposable” “Non-SI” O H H H HO Estrone O H H H HO Epi-estrone A. B. Correlation of “SI” part Correlation of “non-SI” part Estrone Q=0.09 Epi-estrone Q=0.13 Q=0.09 Q free =0.14 Q=0.09 Outlier: the D-ring ketone Q free =0.44 10 8 6 4 Back-calc Value (Hz) 2 0 -2 -4 -4-2 0 2 4 6 8 10 10 8 6 4 Back-calc Value (Hz) 2 0 -2 -4 -4-2 0 2 4 6 8 10 10 8 6 4 Back-calc Value (Hz) 2 0 -2 -4 -4-2 0 2 4 6 8 10 10 8 6 4 Back-calc Value (Hz) 2 0 -2 -4 -4 -2 0 Exp Value (Hz) 10 8 6 4 Back-calc Value (Hz) 0 -2 -4 -4-2 0 2 4 6 8 10 10 8 6 4 Back-calc Value (Hz) 2 0 -2 -4 -4-2 0 2 4 6 8 10 Exp Value (Hz) Exp Value (Hz) 246810 MSD MSD Figure 6: MSD analysis for estrone and epi-estrone using RCSA data. A. RCSA only slightly favors estrone over epi-estrone. B. By MSD, the correct isomer (Q free =0.14) has a much better correlation than the incorrect isomer (Q free =0.44). The outlier from D-ring ketone is easily rationalized from the structure differences. N O O O O OH HO H N O O O O OH HO H Retrorsine Diastereomer 2 “Superimposed (SI)" “Non-SI” * A. B. MSD MSD Retrorsine Diastereomer 2 Q=0.15 Q=0.20 Q free =0.18 Q free =0.68 - 15 10 5 Back-calc Value (Hz) 0 5 -10 -15 -15 -10 -5 0 5 10 15 4 2 0 -2 Back-calc Value (Hz) -4 -6 -9 -10 -10 -8 -6 Exp Value (Hz) Exp Value (Hz) -4 -2 0 2 4 15 10 5 Back-calc Value (Hz) 0 5 -10 -15 -15 -10 -5 0 5 10 15 4 2 0 -2 Back-calc Value (Hz) -4 -6 -9 -10 -10 -8 -6 Exp Value (Hz) Exp Value (Hz) -4 -2 0 2 4 Correlation of “non-SI” part Figure 7: MSD analysis for retrorsine using RCSA data. A. RCSA only slightly favors the correct retrorsine structure over the incorrerct one (diastereomer 2). B. By MSD, the correct isomer (Q free =0.18) has a much better correlation than the incorrect isomer (Q free =0.68). Copyright © 2016 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc. All rights reserved.
Transcript of Analysis Using a Gel-Stretching Device NMR Structure...
Accurate Measurement of RDC and RCSA for Small Molecule Configuration Analysis Using a Gel-Stretching Device
Yizhou Liu, Gary E. Martin, and R. Thomas WilliamsonNMR Structure Elucidation, Process and Analytical Chemistry (Rahway), Merck & Co. Inc., Kenilworth, NJ, USA
liuyiz_189821-0001_ENC_poster_v1.03 04/07/2016ENC Meeting, Output Size: 44” x 44” Scale: 200%
Introduction
Stereochemical determination of small drug-like compounds represents a frequent challenge in the pharma industry. Conventionally, we have relied on NOE/ROE and J coupling data in combination with DFT (Density Functional Theory) modeling for such work. However, the limit of this approach manifests when the chiral centers are separated beyond the scope of these short-ranged data. This limitation prompted us to develop techniques that provide long-range structural information. While both RDC and RCSA are known to serve this purpose, RCSA is particularly attractive for proton-deficient molecules in which measurable RDCs are scarce. However, the application of RCSA for small molecules is rare. The main hurdle for RCSA is the lack of a convenient and reliable approach to eliminate isotropic chemical shift changes upon molecular alignment. Here we describe a simple gel-stretching device that provides accurate and clean measurement of RCSA data and allows RDC measurement as well. The utility is demonstrated for several natural products including strychnine, estrone, and retrorsine.
Design of the Gel-Stretching NMR Tube
The stretching device is an NMR tube with different inner diameters (ID) on two open ends. This design is adopted from a previous report that used a two-stage NMR tube for RCSA measurement in proteins1 but with improvement for better durability. Isotropic data are collected with the gel in the wide segment (4.2mm ID) while anisotropic data are collected with the gel stretched in the narrow segment (3.0-3.4 mm ID) of the tube. Shimming is easy and good field homogeneity can be obtained. We used the chloroform-compatible PMMA gels2 for all studies.
Isotropic Stretched
Spinner here
Piston
ID: 4.2 mmOD: 5.0 mm
ID: <4.2 mmOD: 5.0 mm
Wrap withparafilm
Rubber plug
A.
ID: <4.2 mmOD: 5.0 mm
Wrap withparafilm
B.
Figure 1: The design of the 2-stage NMR tube for gel stretching and measurement of RDC and RCSA.A. The setup for the isotropic measurement. B. The setup for the anisotropic measurement. The gel is artificially colored for visualization.
References:1. Liu and Prestegard, J Biomol NMR. (2010), 47, 249-258. 2. Gayathri, Tsaresvsky, and Gil, Chemistry. (2010), 16, 3622-6.
Representative Spectral Data and ConfigurationAnalysis Results for Strychnine
N
N
O
OH
H
HH
H1
2
3 4
5
6 78
910
11
1213
1415
1617
18 1920
2122
23
24
43
12
22 1223
16
8
20 18
14
11182011 15
17
15 13A.
C.
B.
169.4 169.2 (ppm)
64.6 64.4 (ppm)
140.6 140.4 (ppm) 133.0 132.8 (ppm) 122.4 122.2 (ppm)
C10 C21 C6 C1
C23 C16 C8 C17 C14
60.2 60.0 (ppm) 42.6 42.4 (ppm) 31.6 31.4 (ppm)
Figure 2: Experimental RDC and RCSA dataA. Strychnine structure. B. Overlay of J-resolved BIRD-HSQC experiments collected under isotropic (red) and anisotropic (blue) conditions. PMMA signals are enclosed in rectangles. C. Overlays of 13C spectra under isotropic (red) and anisotropic (blue) conditions, showing RCSA effects at representative carbons; carbonyl and aromatic carbons (top row) tend to have larger RCSA values than aliphatic carbons (bottom row).
20
10
0
-10
-20
-20 -10 0Exp Value (Hz)
10 20 -5 0 5Exp Value (Hz)
10 15
Bac
k-ca
lc V
alue
(Hz)
15
10
5
0
-5
Bac
k-ca
lc V
alue
(Hz)
-20 -10 0Exp Value (Hz)
10 20
20
10
0
-10
-20
Bac
k-ca
lc V
alue
(Hz)RDC
Q=0.07
SZZ=-4.9e-4η=0.51α=-144°β=-169°γ=30°
RCSAQ=0.05
SZZ=-5.4e-4η=0.43α=-158°β=-170°γ=23°
SZZ=-5.1e-4η=0.48α=-147°β=-169°γ=27°
RDC+RCSAQ=0.08
Figure 3: Evaluation of data quality, showing correlations between experimental values and SVD back-calculated values, showing results for RDC data only (left), RCSA data only (center), and RDC-RCSA data combined (right). The asymmetry parameter η is defined as (|Syy|-|Sxx|)/|Szz|.
A. B.
Q-fa
ctor
s
0.80.70.60.50.40.30.20.1
0
0.80.70.60.50.40.30.20.1
0
0.3
0.25
0.2
0.15
0.1
0.05
0
1a 1b 2 3 4 5 6 7 8 9 10 11 12 13
RDC
RCSA
RDC+RCSA
*Labels the site of inversion.
Figure 4: Using RDC and RCSA data for stereochemistry determination.A. Structures of strychnine (1) and its 12 diastereoisomers with inverted stereochemistry. B. Q-factors of RDC alone, RCSA alone, and RDC-RCSA combined are calculated for lowest-energy DFT structures.
Enhancing Stereochemistry Discrimination by “MaximallySuperimposable Differentiation” (MSD)
Maximally superimpose two candidates and only use RDC/RCSA data of the superimposable parts for alignment tensor determination. Then back-calculate RDC/RCSA values for the non-superimposed parts and obtain Qfree factors and correlation plots. The subscript “free” means these data were not used in alignment tensor determination. The difference in Qfree is expected to be more significant than the difference in Q. More importantly, the numeric differences in Qfree factors can be more readily rationalized from a structure perspective by identifying outliers in the correlation plot.
Cal
c V
alue
(Hz)
20
10
0
-10
-20
Pre
d (H
z)
201510
50
-5-10-15
20
10
0
-10
-20
201510
50
-5-10-15
Cal
c V
alue
(Hz)
Pre
d (H
z)
-20 -10 0Exp Value (Hz)
10 20 -15 -10 -5 0Exp (Hz)
5 10 15 20
-20 -10 0 10 20 -15 -10 -5 0 5 10 15 20
“Superimposable” (SI) “Non-SI”
*
Q=0.07
Q=0.13
Qfree=0.11
Qfree=0.23
A.
B.
Struct 1 Struct 1
Struct 4 Struct 4
MSD
MSD
Outlier from the site of inversion
Correlation of “non-SI” part
Figure 5: MSD analysis for strychnine diastereomers 1 (correct) and 4 (incorrect).Red dots: RDC; green/blue dots: RCSA. A. Maximal superposition of 1 and 4. B. Correlation plots from SVD fail to reveal logical outliers in 4. MSD clearly reveals the RDC at the center of chiral inversion as an outlier.
2
“Superimposable” “Non-SI”O
H
H HHO
Estrone
O
H
H HHO
Epi-estrone
A. B.
Correlation of“SI” part
Correlation of“non-SI” part
EstroneQ=0.09
Epi-estroneQ=0.13
Q=0.09 Qfree=0.14
Q=0.09
Outlier: theD-ring ketone
Qfree=0.44
10864
Bac
k-ca
lc V
alue
(Hz)
20
-2-4
-4 -2 0 2 4 6 8 10
10864
Bac
k-ca
lc V
alue
(Hz)
20
-2-4
-4 -2 0 2 4 6 8 10
10864
Bac
k-ca
lc V
alue
(Hz)
20
-2-4
-4 -2 0 2 4 6 8 10
10864
Bac
k-ca
lc V
alue
(Hz)
20
-2-4
-4 -2 0Exp Value (Hz)
10864
Bac
k-ca
lc V
alue
(Hz)
0-2-4
-4 -2 0 2 4 6 8 10
10864
Bac
k-ca
lc V
alue
(Hz)
20
-2-4
-4 -2 0 2 4 6 8 10Exp Value (Hz) Exp Value (Hz)
2 4 6 8 10
MSD
MSD
Figure 6: MSD analysis for estrone and epi-estrone using RCSA data.A. RCSA only slightly favors estrone over epi-estrone. B. By MSD, the correct isomer (Qfree=0.14) has a much better correlation than the incorrect isomer (Qfree=0.44). The outlier from D-ring ketone is easily rationalized from the structure differences.
N
O
O
O
OOHHO
H
N
O
O
O
OOHHO
H
Retrorsine Diastereomer 2
“Superimposed (SI)"
“Non-SI”
*
A. B.
MSD
MSD
Retrorsine
Diastereomer 2
Q=0.15
Q=0.20
Qfree=0.18
Qfree=0.68
-
1510
5
Bac
k-ca
lc V
alue
(Hz)
05
-10-15
-15 -10 -5 0 5 10 15
420
-2
Bac
k-ca
lc V
alue
(Hz)
-4-6-9
-10-10 -8 -6
Exp Value (Hz) Exp Value (Hz)-4 -2 0 2 4
1510
5B
ack-
calc
Val
ue (H
z)05
-10-15
-15 -10 -5 0 5 10 15
420
-2
Bac
k-ca
lc V
alue
(Hz)
-4-6-9
-10-10 -8 -6
Exp Value (Hz) Exp Value (Hz)-4 -2 0 2 4
Correlation of“non-SI” part
Figure 7: MSD analysis for retrorsine using RCSA data.A. RCSA only slightly favors the correct retrorsine structure over the incorrerct one (diastereomer 2). B. By MSD, the correct isomer (Qfree=0.18) has a much better correlation than the incorrect isomer (Qfree=0.68).
Copyright © 2016 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc. All rights reserved.
