LC-NMR in Drug Discovery

Brian L Marquez, Ph.D. Principal Scientist Structure Elucidation Group Pharmaceutical Sciences NMR Laboratory Pfizer Inc. 445 Eastern Point Road, MS 8118A-2011 Groton, CT 06340

Tel: Fax: Email:

(860) 686-1360 (860) 686-6227 [email protected]

Personal Background

• •

Undergraduate in Biochemistry and Biophysics in 1997 Ph.D. in Medicinal Chemistry in 2001 – Pharmacognosy -- Marine Natural Products

• • •

Wyeth Pharmaceuticals (PA) -- NMR Spectroscopist in Discovery Amgen Inc. (CA) -- NMR Spectroscopist in Discovery/PKDM/Process Chemistry Pfizer Inc. (CT) -- NMR Group Leader in Pharmaceutical Sciences (Development)

What we do in the Pharm. Sci. NMR group

• NMR need throughout all of drug development. This includes but is not limited to: – Filing characterizations – Lot confirmations – Solution confirmations – New technology development and integration – STRUCTURE ELUCIDATION • Impurities (e.g. process related) • Degradants (e.g. process related or forced degradation)

Intent of this Lecture

– What is LC-NMR – No NMR theory, just application oriented. – Limitations – Configurations/Options – Practical Considerations – Overlying Theme • When/When not to use it

LC-NMR: Simplistic Concept

• HPLC is plumbed in line with a “flow” NMR system • Sample components are physically separated by HPLC • Each component flows, in turn, from the LC column and UV detector to the NMR sample cell – Multiple different configurations for this

• NMR is performed on each desired fraction / peak – Always the rate limiting step

• Continuous or stopped flow mode – Additional methods are also available – Peak “parking” and “trapping”.

Modes of Operation

• Continuous Flow – Eluent sampled in “real-time” as flowing through NMR Detection Coil

• Time Slices – Regions, or “time-slices” of interest are analyzed

• Stopped Flow – Pump is stopped at desired location and data acquired

• Peak Parking – Peaks of interest are “parked” in off-line sample loops

• Peak Trapping – Solid Phase Extraction cartridges are used to “re-concentrate” samples.

General Schematic for an LC-NMR

NMR

LC Control System

Control System

NMR Spectrometer

LC Pump

Injector

X

36 Loop Cassette

Column/ Column Heater

Waste Or Fraction Collector OR

UV

General Cartoon of Loop Collection Loop Collection ….

NMR Spectrometer

…. LC Pump

X

DAD Detector

36 Loop Cassette

Direct Stop Flow

LC-NMR Hardware Configuration

Binary or Quaternary LC Pump

Degasser

UV Detector

Manual Injection

RF Gradients

Temperature Control Unit

LC Column Temperature Control

Loop Collection

600 MHz NMR Console

LC-NMR Probe Schematic NMR detection coil built directly onto flow cell (4mm OD)

240µL

120µL

IN

From LC

OUT

To Waste

Traditional Probe Configurations

• Most common configuration – inverse probes – best proton sensitivity • Flow Cells – Active Volumes – – – –

3mm – 60µL Active Volume 4mm – 120µL 5mm – 240 µL Others “non-traditional” will be covered later

• Typically probes are outfitted with z-gradients – For gradient experiments and shimming

General Application Strategy 1. Structural determination of metabolites:

M2 P M3 M1

~50 µg of metabolites From microsomal

4.6 mm or 6mm Reverse-phase HPLC Mobile phase: CH3CN/D 2O+(0.1%TFA or DCOOD), solvent gradient < 2%/min

P

incubations

2. Structural determination of impurities:

i1

i2

4.6 mm or 6 mm reverse-phase HPLC Large volume injections in high % of aqueous phase More then 1 mg

Mobile phase: CH3CN/D 2O+(0.1%TFA or DCOOD), solvent gradient < 2%/min

Min~1 µg

Structure Elucidation Using NMR

Homonuclear Coupling Heteronuclear Coupling

1H-1H

COSY

1H-1H

NOESY

1H-13C

HMBC

1H-13C

HSQC

O

H H

H

CH3

• 1D 1H and homonuclear NMR experiments are the most sensitive and accessible experiments for LC-NMR • 1D 13C and heteronuclear NMR experiments are very insensitive and are typically inaccessible to LC-NMR applications (in most cases).

When to use LC -NMR

• • • •

Fairly resolved peaks. Relative abundance of entities similar. Known stability issues. A significant amount of information is known about the compound

Impurity Analysis:

Low Volume (20µ (20µL) Injection of Sample

Intens. [mAU]

7

800

20µL injected t 0 48 50 50.5 51

%B 10 100 100 10 10

1.5mL/min 4.6x150mm Luna 5u C8(2) A: 0.1% TFA/D 20 B: 0.1% TFA/MeCN

600 ~0.3-0.5% Impurities (215nm) 400

200 4

0

5

10

5

6

15

8

20

25

30

35

40

45Time [min]

In order to achieve sufficient NMR sensitivity it was necessary to overload the HPLC column without sacrificing peak resolution. This goal was achieved by maximizing sample concentration in the highest content of aqueous phase followed by large volume column injections.

Impurity Analysis:

Large Volume (500µL) Injection of Sample

7

Intens. [mAU]

500µL injected t 0 48 50 50.5 51

%B 10 100 100 10 10

1mL/min 4.6x150mm Luna 5u C8(2) A: 0.1% TFA/D 20 B: 0.1% TFA/MeCN

800

600

6 5 8

400 4 200

Pressure UV (215.0nm)

0 0

5

10

15

20

25

30

35

40

45 Time [min]

Impurity Analysis: Solved Structures 10

8

7

B

1

A #8

11, 12 87 9

6

11

3-5

3’-5’

12

C

3 4

10

#7

3-5 11,12 8 7 9

8

7

5

A

B

11

1 2

12

6,6’

3 7

8

4 5

B

10

A 879

#6

11

1

11-13,3-5

13 12

2 6,6’

3 4

7

5

10

A 1

3-5 7 8 9

#5

6,6’

3 7

9 #4

A

1

3-5 7,8

6,6’

* 3 4

8.0

7.5

7.0

6.5

6.0

5.5

5.0

4.5

4.0

3.5

3.0

2.5

2.0

1.5

1.0

ppm

5

8

4

B

5

8

B D

C

Metabolite Analysis:

100µL of ~4mg/mL injected

Intens. [mAU] 2000

t 0 5 30 37 38 40

%B 35 35 95 95 35 35

1500

4

5 1000

1.0mL/min 4.6x250mm YMC-AQ 5u 120A A: 0.1% d-FA/D20 500 B: 0.1% d-FA/MeCN

3

1 2

Analog only DAD (254)

0 0

5

10

15

20

25

30

35 Time [min]

Metabolite Analysis: Solved Structures

13

M_678351_1, LC/MS run 396

F

A

10

F

7

11 7 10 2 5

4 9 6

3

O

F

F

7 11

3

3

6

A 7

O

6

O

11

N

5

8

N

N

5

N

12

10,4

42

4

10 9

95 6 3

2

428/429

F

4

1

4

4

OH

3

13

13

444/446

12 F

1

4

6

3

3

3 O

6

A

N

1

10

F

7 11 5 10

9

F

1 OH

O

8

1

1

5

8

13

13

N

5

N

O

12

6

O

N

12

5

8 1

N

9 11

1 8

6

MS(H)/MS(D)

F

12

O

9 7

2

O

N

12

6

O

11

N

5

2

N

41

3

2

5

8

4

3

O 13

13

444/445

F

F

12 9

2

1

11 7 10

9

5

7

6

O

N

N

12

6

O

11

13

N

8

9

4

OH

4

460/462 F

4

5 6

3

9

12

11 8.5

8.0

7.5

7.0

6.5

6.0

5.5

5.0

4.5

4.0

3.5 ppm

13

N

N

6

O N

8

O

9.0

7

11

2

6

A

10

O

1

OH

OH

F

7 10

1

5

8

13

F

OH

5N

O

12

1

10

F

4

8

6

A

5

478/482

5 4

3

1

O

N

2

4

2 3

Metabolite Analysis:

500µL injected t 0 2 40 45 46 50

%B 20 20 90 90 20 20

1.0mL/min 6x150mm YMC-AQ 3u 12n A: 0.1% d-AcOH/D20 B: 0.1% d-AcOH/MeCN

Intens. [mAU]

5000 5 4000

3000

2000 6 1000 3 1 2

4 DAD (254)

0 0

5

10

15

20

25

30

35

40

Time [min]

Metabolite Analysis: Solved Structures

2, 3, 4, 5

A

S

5

1

1

4

6

O

2

3

O

4 1 6

2 5

A

S

5

1

3, 4

2

3

4 6 1

5

4

2

A

S

5

1

3

3

HO

2 OH

9

A

S

5 7, 8 1

2

6

5

HO

3

1 3

2, 5

3, 4

7, 8

2

A

S

5

1 6

4

1 3

8.5

8.0

7.5

7.0

6.5

6.0

5.5

5.0

4.5

4.0

3.5

3.0

2.5

2.0

1.5

ppm

2

0 10 20 30 43.497 44.583 45.288

40 50 60 70 82.475

78.252

89.427

65.991

580

582

50.444

46.903

582

80.553

51.044

582

525

47.221

24 45

80

86.873 87.267

78.883 79.723

76.648

73.743 74.227

71.301

65.099

28

72.122

62.646 63.083 64.261

58.518

54.584

52.958

27

57.740

52.481

45.589

60

48.372

120

47.831 48.828 49.137 49.509

598

140

46.437

40

44.950

42.550

20 33.190 34.340

29.261

23.354

12.429

18.085

mAU

21.642

15.382

11.740

When not to use LC-NMR

VWD1 A, Wavelength=254 nm (DC110403\1EAR4633.D)

17

100

80

23

0 min

Regiochemistry

• Sample submitted for determination of the regiochemistry of the primary amine moiety

NH2 N O

Structure A

N

NH2 O

Structure B

1H

Assignment

Very little help using Chemical Shift Arguments

E

NH2 N O A D

N

B

C

NH2 O

E

D C

A

C

D

B

1.93

1.06

1.05

7.9 7.8 7.7 7.6 7.5 7.4 7.3 7.2 7.1 7.0 6.9 6.8 6.7 6.6 6.5 6.4 6.3 6.2 6.1 6.0 5.9 ppm 1.00

B

1.00

A

E

1H-15N

HMBC Data

Background -- Simplified

• Pulse sequence allows one to detect 1H’s longrange coupled (~8 Hz) to 15N – Depending on the experiment used one can choose to omit or retain the 1JNH

O N

nJ

NH

(n = 2,3) ? Will result in a correlation centered at the 1H and 15N chemical shift

N H 1J

NH

? Will result in a correlation centered at the 15N chemical shift and a split signal centered on 1H

1H-15N

HMBC Data Allows Assignment

1H-15N HMBC OF CP-325,366 LOT52203-059-01 ~ 22, MG IN 0.8 ML DMSO, BY AJJ TEMP=25C, Liquids 600 MHz NMR Instrument, GNMR06

A

E

C

D

NH2

B

N

A ppm

B

70 Hz

O D

C

80

Primary Amine

E

n

90

1 100

JNH (n = 2,3) JNH

110 120

E

130 140 150 160

A

C

D

B

170 180

Tertiary Amide 6.8

6.6

6.4

6.2

6.0 ppm

7.9 7.8 7.7 7.6 7.5 7.4 7.3 7.2 7.1 7.0 6.9 6.8 6.7 6.6 6.5 6.4 6.3 6.2 6.1 6.0 5.9 ppm 1.93

7.0

1.06

7.2

1.00

7.4

1.05

7.6

1.00

7.8

190

Other Options Available

• LC-NMR/MS – Allows on-line MS and NMR evaluation of samples

• Peak Trapping (Column Trapping) – Potentially allows multiple LC peaks to be “trapped” and concentrated prior to NMR data acquisition.

• Microcoil Probes – Has potential to allow microscale separation mechanisms (e.g. CapLC).

• CryoProbe Technology – Significantly lowers “noise floor” through cryo-cooling RF electronics in the probe.

LC-NMR/MS

LC & MS Control System

NMR Control System

600MHz NMR Spectrometer

LC Pump

Injector

X

Loop Cassette

Column/ Column Heater

Waste Or Fraction Collector

OR Splitter

95% 5% UV

Mass Spectrometer

MS Component

Actual System Hardware Setup for LC-NMR-MS ( without magnet ) HPLC

LC-NMR-MS Interface, BNMI

Esquire Ion Trap Mass-Spectrometer

NMR Spectrometer electronics

UV detector BPSU-36/ BSFU-O

MS-Rough pump Graphic Borrowed from Bruker-Biospin

Actual System

Column Heater/Injector

36 Loop Cassette

Agilent 1100

BNMI

Esquire 3000+ Ion Trap MS

3 Port Switching Valves DAD Detector

MS Rough Pump

NMR Electronics

Other Options

LC-SPE-MS-NMR Peak Trapping

Chromatographic System

Spark Prospekt 2

Drying

Spark Prospekt 2

NMR Spectrometer

Closer Look at the SPE

Spark Holland Prospekt SPE

Robot gripper for SPE cartridges

2 flow lines where trap cartridges are inserted

Trap cartridge size 10mm * 2mm (ID) 10mm * 1mm (ID) 10mm * 3mm (ID) ~ 2 $ per cartridge

SPARK HOLLAND SPE-UNIT

Various commercial packings available Bruker provides a set of 4 different solid phase types to start

Graphic Borrowed from Bruker-Biospin

Major New Developments

• Miniaturization – Microcoil Probes • Integration of new (to commercial NMR) MS • Cryogenic NMR Flow Probes

MicroCoil Probe • Horizontal copper RF solenoid Coil • Vertical (Z) pulse field gradient (PFG) coil • Flow cell is surrounded by CF-43 fluorocarbon for susceptibility matched to copper coil • 1.5 µL active volume with a 5 µL total volume • 7 µL total volume from inlet to outlet (3 µL transfer from injection assembly) • Lock power > 45 db to prevent saturation • π/2 pulse width of 8.4 µs at 18 db power level • Low power needed for 90% H2O/ 10% D2O (75 db) saturation

Advantages vs. Disadvantages

• Advantages • Extremely mass sensitive • Capillary-scale fluidics allow transport of µL volume samples over distances of 5-10 meters with virtually no degradation in analyte peak volume. • Diffusion and mixing effects at the capillary scale are very limited so that peaks can be parked overnight with negligible loss of S/N. • Residual protonated solvents are significantly reduced – need for multiple solvent suppression avoided in most cases. • Acquisition of data in fully protonated solvents is reasonable.

• Disadvantages • Manipulation of 5µL aliquots can tedious. • Availability of HT platform poor • Samples of poor solubility

Capillary Probe Data O

One Hour acquisition time per sample

Sample in 5µL cell (conc. mg/mL) 100µg (20 mg/mL)

F

S/N

440

25µg (5 mg/mL)

229

HOD Saturation

12.5µg (2.5 mg/mL)

116

6.3µg (1.3 mg/mL)

62

3.1µg (0.63 mg/mL) 1.5µg (0.31 mg/mL) 780ng (0.16 mg/mL) 390ng (0.08 mg/mL) 5.5 5.0

N N N H

845

50µg (10 mg/mL)

7.0 6.5 6.0

N

4.5 4.0 3.5 3.0 2.5 2.0 ppm

36 25 18 16

Bruker’s New MicroTOF -- Metabolites

•

•

•

The microTOF-LC can provide the exact mass of the analyzed sample and therefore access to the sum formula. microTOF-LC allows HyStar™ to trigger collection of chromatographic peaks into loops (LC-NMR), or SPE cartridges (LC-SPE™ NMR) based on mass chromatograms However, no additional structural information is provided (e.g. fragmentation). This information is sometimes more relevant than the molecular formula.

Picture taken from www.bruker-biospin.con

CryoProbes Installation of a 600MHz triple resonance 1H{13C,15N} CryoProbeTM system.

Picture taken from presentation made by Dr. Kim Colson at Bruker Biospin

Overall Conclusions

• LC-NMR is an extremely useful tool in very specific instances. • Additional “hyphenation”, in some cases, provides an enormous amount of pertinent structural information. • Decreasing the noise floor (cryoprobes) is allowing NMR to routinely analyze samples that were previously impossible by NMR

Acknowledgements

• David Chow • Kim Colson • Linda Lohr and Andy Jensen