High throughput screening of pKa by capillary electrophoresis and mass spectrometry (CE/MS)
and long-term validation
Hong Wan
Lead Generation, DMPK and Physical ChemistryAstraZeneca R&D Mölndal
SE-431 83 Mölndal. SWEDENe-mail: [email protected]
5th Physchem Forum, Stevenage, UK, June 19th
• Why screening pKa• Advantages of CE/MS over established methods• High throughput pKa screening method by CE/MS
• Effects of buffer type and ionic strength on pKa• Comparison of CE/UV and CE/MS• Effects of pressure on migration times/sensitivity/mobility
• Comparisons of measured pKa and ref. and predicted values• Reproducibility and accuracy and long-term validation• Summary
Outline
Screening for pKa, why?
• Info about change in charge of molecule……
pKa(dissociation
constant)
Solubility
Lipophilicity
Protein binding
QSAR descriptor Formulation
Bioavailability
AbsorptionPermeabilityHydrogen
bondingpotency
StabilityMetabolism
Is CE/MS better technique for pKa than others?
• Sirius GLpKa – well established and widely used method !
• pION's Gemini for routine pKa ($500/cdp, $760/cpd/co-solvent)
• CE/UV (CombiSep), commercially available
• CE/MS (new technology), AZ, Mölndal, 2003 published.
H. Wan, A. Holmén, M. Någård, W. Lindberg, J. Chromatogr. A. 979 (2002) 369.
H. Wan, A. Holmén, Y. Wang, W. Lindberg, M. Englund, M. Någård, R. Thompson,
Rapid Commun. Mass Spectrom. 17 (2003) 2639.
Advantages of CE/MS over established methods
Specification CE/MS Titration/UV (D-PAS)
Amount of cpd 1 μg, or 1 μL of 10 mM solution
1-2 mg (titration)3 μL of 10 mM DMSO stock (Sirius 3T)
Concentration 1-10 μM (100 μL)
No
> 20 μM (UV)
Purity required
Co-solvent No Yes, for poorly soluble cpds
Accuracy < ± 0.2 ± 0.02-0.2?
Throughputcapacity
>150 cpds/seq.(6h) 4 min/cpd (Sirius T3)
Limitation Poor ionization (ESI) Impurity? poor solubility (precipitation),Ionization group close to chromophore
Yes
Principle of pKa determination by CE
pKa ↔ 50% ionization
}HA{}A}{H{K th
a
−+
=
I0.3281I0.5085z
mmmlog pHpK
2
effm
efftha α+
+⎥⎦
⎤⎢⎣
⎡−
−=
Equations used for pKa calculation
Graphs and table from Miller, Electrophoresis 2002, 23, 2833
High-throughput pKa screening by CE/MS
10-14 different pH buffers Constant ionic strengths
0.2 0.4 0.6 0.8 1 (min)
mAU
0
5
10
15
20
25
30
35
12
3
N
⎟⎟⎠
⎞⎜⎜⎝
⎛−⋅
⋅=−=
eofobs
efftoteofobseff ttV
LLmmm 11
(fit-pKa)
CE separation
Mobility calculation
pKa evaluation
Effective mobility
0 0.5
1 1.5
2 2.5
3 3.5
4 4.5
2.00 4.00 6.00 8.00 10.00 pH
UV non-volatile buffers (Single compound)
20-100 μM
MS Volatile buffers
(Pooled compounds)
pKa2
pKa1
Pressure
Sample pooling and data analysis flow scheme
Reference stand.
2003 manual data analysis
H. Wan, A. Holmén, Y. Wang, W. Lindberg, M. Englund, M. Någård, R. Thompson,
Rapid Commun. Mass Spectrom. 17 (2003) 2639.
Current pKa screening - integrated & automated assay
Automated sample pooling (mass)
CE/MS
Automated migration time conversion
pKa evaluation
pKa reported to database
Generate sample run list (Abase)
AZ10181832-001 NA
pH meff res Fitting Option:2.477 287.909 x 287.806016 -0.102984 287.909 2.4773.012 243.266 x -243.266 243.266 3.0124.046 178.427 x -178.427 178.427 4.046 Value: Ionization: Ion. Func. Gr.4.565 163.527 x -163.527 163.527 4.565 pKa1:5.140 163.972 x -163.972 163.972 5.14 pKa2:5.633 163.325 x -163.325 163.325 5.633 pKa3:6.060 164.855 x -164.855 164.855 6.06 r2:7.086 151.090 x -151.09 151.09 7.0868.020 115.490 x -115.49 115.49 8.028.380 72.070 x -72.07 72.07 8.389.080 26.064 x -26.064 26.064 9.089.470 9.778 x -9.778 9.778 9.47
10.010 5.132 x -5.132 5.132 10.0110.530 2.184 x -2.184 2.184 10.53
0 00 00 00 00 00 0
Comment:
8.33
0.999
3.01
-400
-200
0
0
50
100
150
200
250
300
350
1 2 3 4 5 6 7 8 9 10 11 12 13 14pH
mef
f
B1B1
B2B2
Di-BaseDi-Base
Approved
NVNV
N
N
N
N
B1
B2
CE/MS instrumentation
HPCE3D , 1100 series LC/MSD trap SL (Agilent)
• Untreated fused silica capillary (50-60 cm/50 SEK) can be used as long as it works. • Buffers without filtration (stock solutions can be used up to >3 years).• Sheath liquid (1 L) can be used for more than 6 months (99% recycling).
• Higher initial investment
• Long-term savings
On-line CE/UV/MS experimental setup
UV @ 21.6 cmMS @ > 75 cmMS w/o UV @ 55 cm
Air coolingUV Detector interface
Capillary cassette
Capillary inlet
To MS
(Agilent 1100 series LC/MSD Trap)
Pressure (40-50 mbar)
Courtesy from Agilent (with small modifications)
Sensitivity from volatile vs non-volatile buffers in CE/MS
• Non-volatile buffer decreased signals at higher concentration (ion suppression)
• Ion source contamination by non-volatile buffer
020406080
100120140160
25 m
M NH4A
C50
mM N
H4AC
75 m
M NH4A
C10
0 mM N
H4AC
25 m
M sodiu
m/phos
phate
50 m
M sodiu
m/phos
phate
75 m
M sodiu
m/phos
phate
100 m
M sodiu
m/phos
phate
LidocainePapaverine
S/N
Not detectable
10 μM
Effects of buffer type and ionic strength on pKa
• Titration method uses physiological buffer with ionic strength at 0.15 M.• Ionic strengths (0.025 - 0.15 M) have a small effect on pKa (Δ pKa=0.064)
0
2
4
6
8
10
12
0 2 4 6 8 10 12
CE/UV (non-volatile buffer, I=0.05)CE/MS (volatile buffer, I=0.025)
pKa by CE/UV
pKa by CE/MS
Nicotine
PapaverinePilocarpine
NicotineLidocaine
Codeine
PropranololAtenolol
I0.3281I0.5085z
mmmlog pHpK
2
effm
efftha α+
+⎥⎦
⎤⎢⎣
⎡−
−=
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
10.00
11.00
12.00
pKa1(B) pKa2(HA)
I=0.01I=0.1I=0.15I=0.50I=1.0I=3.0
NH
ONH2
OHTryptophan (ACD)9.55/ 9.34/ 9.01/ 9.32/ 9.3/ 9.92
9.51/2.34 (CE/MS, I=0.025)
9.43/2.30 (CE/UV, I=0.01)
2.37 /2.37/ 2.23/ 2.39/ 2.75
Comparison of UV and MS (Ion trap) sensitivity
UV working concentration: 20-100 μM
0 5 10 15 20 25 30 35 40 Time [min]
0.0
0.2
0.4
0.6
0.8
1.07x10
Intens.[mAU]
Propranolol
AtenololCodeine
LidocainePapverine
Pilocarpine
UV
MS(EIC)
EOF
Sample: 0.5 µMAmmonium acetate,(pH=7) I=0.025Pressure: 40 mbar
Sensitivity and reproducibility of pressure-assisted-CE/MS
.
Total ion chromatogram (TIC)
SENS-C10.D: EIC 260 ±All MS
0
2 4 x10
Intens.
1 2 3 4 5 6 7 8 Time [min]
1 SENS-C20.D: EIC 260 ±All MS
0.0 0.5 1.0
6 x10
1 2 3 4 5 6 7 8 Time [min]
1SENS-C30.D: EIC 260 ±All MS
0.0 0.5
7 x10
1 2 3 4 5 6 7 8 Time [min]
0.1 μM S/N=1
1 μM (propranolol) S/N=235
10 μM S/N=864
1
SENS-C10.D: TIC ±All
0
1 6 x10
Intens.
1 2 3 4 5 6 7 8 Time [min]
0 2
6 x10
1 2 3 4 5 6 7 8 Time [min]
SENS-C30.D: TIC ±All
0
2 7 x10
1 2 3 4 5 6 7 8 Time [min]
0.1 μM
1 μM (11 mixed compounds)
10 μM
SENS-C20.D: TIC ±All
1
1
1
(EIC) and migration times after 30 runs
Ammonium acetate(pH=7) I=0.025;
pressure, 40 mbar
PKA-C036.D: EIC 79; 248 ±All,
2
4
66x10
MS
0 2 4 6 8 10 12 14 16 18Time [min]
0
5
UV-PKAC036.D: UV Chromatogram, 210.4 nmAR-H073379
AR-H073379
DMSO (2.5%)
pH=3
DMSO
?25 μΜ
On-line tandem UV and MS detection
Sensitive and selective MS detection over UV
79.3 125.0 167.0199.0
248.0 (M+1)
317.9
391.0
+MS, 7.3min (#974)
0.00
0.25
0.50
0.75
1.00
1.25
6x10Intens.
100 150 200 250 300 350 400 450 m/z
AR-H073379
Effects of pressure on migration times/sensitivity
+++++++++
+++++
++++
+++
++
+
++ +++++ +
+
+
++
++
+
+
+
++ ++ +
+
+ ++
+
+ ++
+
+ ++
+
+++
+
+
+
+
++++
CE capillary
sheath liquid
nebulizing gas
+++ ++ ++
CE sprayer tip
• Pressure reduced migration times
• Negligible signal suppression
• Improved RSD of effective mobility
• High throughput capacity
160
170
180
190
200
210
220
230
0 10 20 30 40 50 mbar
AtenololCodeineLidocaineQuinidinePilocarpinePropranololNicotine Papaverine
Pilocarpine+Papaverine
Codeine+Propranolol
0
500000
1000000
1500000
2000000
Intensity
0 10 20 30 40 50 mbar
AtenololCodeineLidocaineQuinidinePilocarpinePropranololNicotine Papaverine
Mass scan 70-660 m/zTime (s)
• Effective mobility shift around pKa (an interesting observation)• Mobility shift caused by pressure doesn’t affect pKa values.
H. Wan, A. Holmén, M. Någård, W. Lindberg, J. Chromatogr. A. 979 (2002) 369.
2 3 4 5 6 7 8 9 10 11
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5 Ibuprofen
50 mbar 25 mbar 0 mbar
Mef
f(10-4
cm2 V-1
s-1)
pH
pKa=4.5
Effects of pressure on effective mobility
min0 5 10 15 20 25 30 35
mAU
0
20
40
60
80
100
(c) P=50 mbar
(b) P=25 mbar
(a) P=0 mbar
1 2+3+N
1N
3
1
2
OH
O
OH
O
3 Benzoic acid
2 Ibuprofen
1 Lidocaine
(A) pH=3
40 min
NH
O
N
CH3
CH3
CH3
CH3
How many compounds can be pooled by CE/MS?
• DMSO < 5% in sample
• Constant mobility
• High MS resolution
• More than 150 cpds/sequence
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
0 1 2 3 4 5
PilocarpineLidocainePropranololAtenololPapaverine
012345678
0 1 2 3 4 5
DMSO (Peak height)
DMSO (Peak area)
DMSO %
DMSO %
μeff
0
1
2
3
4
5
6
7
8
9
10
11
0 1 2 3 4 5 6 7 8 9 10 11
N
O
O
OH
F
CH3
NNO
O
O
O
OH
CH3
NN
OO
CH3
pKa (Cal.) ACD6
pKa (CE/MS)
0
1
2
3
4
5
6
7
8
9
10
11
0 1 2 3 4 5 6 7 8 9 10 11
R2=0.992
pKa (literature)
pKa (CE/MS) (a) (b)
Cinoxacin
Flumequine
Phenylbutazone
R2=0.880
Comparison of measured pKa and lit./predicted values
H. Wan, A. Holmén, Y. Wang, W. Lindberg, M. Englund, M. Någård, R. Thompson,
Rapid Commun. Mass Spectrom. 17 (2003) 2639.
Poor prediction for conjugated structures
Long-term validation of CE/MS pKa (2005-2008)
Scatter Plot
pKa(2005)
2
3
4
5
6
7
8
9
10
11
2 3 4 5 6 7 8 9 10 11
Antipyrine
Benzocaine
Diazepam
Furosemide
Ethionamide
Phenylbutazone
WarfarinMethotrexate
Glyburide
Enalapril
Sulfacetamide
Meclizine
Papaverine HCL
Amiodarone
Flumequine
Cimetidine
Tamoxifen
Clozapine
lidocaine
clonidine
Phenytoin
Quinine
Terbutaline
Haloperidol
ClofazimineTerfenadine
propranolol
Imipramine
Paracetmol
O
N
R2 =0.996
pKa: 8.71 (2005)
7.88/7.66 (2008)
Tamoxifen
10 buffers
14 buffers
An example of poorly soluble compound
-0.2
0
0.2
0.4
0.6
0.8
1
1.2
2 3 4 5 6 7 8 9 10 11
1. Ionization not close to chromophore2. Low solubility 0.45 μM (pH 7.4), LogD=6.5pKa (DPAS): nopKa (GlpKa): co-solvent, precipitation ?pKa (CE/MS): 8.71 (2005), 7.88/7.76 (2008)
(ref: 7.6)
Ref. (Karl Box et al., Anal. Chem., 2003, 75, 883-892).
HN+O
CH3
CH3
CH3
NO
CH3
CH3
CH3
H+
pKa = 8.7
Tamoxifen
pKa=8.71
AZ10009015-095 NA
pH meff res Fitting Option:2.362 87.540 x 85.8015775 -1.738423 87.54 2.3622.810 87.797 x 85.8010667 -1.995933 87.797 2.814.120 84.083 x 85.7856568 1.7026568 84.083 4.12 Value: Ionization: Ion. Func. Gr.4.720 89.378 x 85.7373889 -3.640611 89.378 4.72 pKa1:5.331 89.326 x 85.5392192 -3.786781 89.326 5.331 pKa2:5.714 80.971 x 85.1701977 4.1991977 80.971 5.714 pKa3:6.281 78.424 x 83.5164221 5.0924221 78.424 6.281 r2:7.127 71.436 x 71.9841923 0.5481923 71.436 7.1278.454 12.957 x 16.9045823 3.9475823 12.957 8.4548.826 14.763 x 8.09575778 -6.667242 14.763 8.8269.470 11.515 x 1.98220358 -9.532796 11.515 9.479.848 5.279 x 0.84143209 -4.437568 5.279 9.848
10.260 4.675 x 0.32782088 -4.347179 4.675 10.2610.776 2.467 x 0.10018246 -2.366818 2.467 10.776
0 00 00 00 00 00 0
7.76
0.991
Comment:
-20-10
010
0
10
20
30
40
50
60
70
80
90
100
1 2 3 4 5 6 7 8 9 10 11 12 13 14
pH
mef
f
NVNV
B1B1
Mono-BaseMono-Base
Approved
NVNV
O
N
pKa7.8
• Seven compounds as on-line QC for pKa screening (single measurement)
• Reproducibility <0.2 pKa units
Long-term reproducibility of QC (2005-2008)
QC AZ_1 AZ_2 AZ_3 AZ_4 AZ_5 AZ_6 AZ_7
Compound Atenolol Lidocaine Papaverine Nicotine Propranolol Quinine Codeine
Structure
pKa(B1/B2) Mean (n=50) 9.73 8.16 6.05 8.45(3.12) 9.67 8.88(4.00) 8.45
SD 0.09 0.09 0.11 0.10(0.27) 0.08 0.10(0.19) 0.08
Ref. 9.58 7.90 6.39/5.95 8.12(3.12) 9.508.50(4.1)
8.90 8.54
8.218.30
CE/UV 9.61 7.92 6.38 8.02(3.10) 9.49 8.39(4.14) 7.97
NH
O
N
CH3
CH3
CH3
CH3
N
NC
ONH
OH
CH3
CH3
N
N
CH2
OH
O
CH3H
O
N
OH
O
CH3
CH3
H
NH2
OONHCH3
CH3
OH
NH2
OONH
OH
Examples of determination of poorly soluble compounds
N
NH
S
N
O
O
CH3O
CH3 CH3
CH3
(c) Hexetidine
8.9
4.14
• Possible to measure pKa with solubility < 1 µM
• pKa =50% ionization ↔ (max. mobility)/2
• Mobility ≈ charge/mass
N
O
CH3
CH3
CH3
(b) Tamoxifen
(f) Omeprazole
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
(a) Imipramine(b) Tamoxifen(c) Hexetidine(d) Ketoconazole(e) Mebendazole(f) Omeprazole (g) Indomethacin(h) Sulfacetamide
μeff
2 4 6 8 10 (pH)
N N
NH2CH3
CH3 CH3
CH3CH3
pKa fitting example pH 2.5-10.5/10 buffers(2.3-10.8/14 buffers)
-1.5
-1
-0.5
0
0.5
1
1.5
2
2.5
2 4 6 8 10 12
Omeprazole
Clozapine
Chlorpromazine
Phenylbutazone
Ampicillin
Mebendazole
-3
-2
-1
0
1
2
3
2 4 6 8 10 12
SulfacetamideParacetmolClonidineCimetidineDebrisoquinEthionamideAntipyrine
N
NH
NH2
13 (ref)?
N
NH
S
N
O
O
CH3O
CH3 CH3
CH3
Omeprazole
Debrisoquin
S NH
O
O
O
NH2
CH3
5.270.08?
Sulfacetamide
8.9
4.14 mono-base
di-base
mono-base/acid
mono-acid
Summary
• Pressure-assisted CE/MS for pKa:• High throughput screening pKa from 2.3 to 10.8
• Reproducibility and accuracy of pKa ± 0.2 units
• Insensitive to compound purity, requiring minute sample
• Concn: (1-10 µM) beneficial for poorly soluble cpds
• Independent on type of buffer, capillary length, DMSO concn.
• Provide charge and structure/stability information (MS)
Acknowledgements• Anders Holmén, Mats Någård, Walter Lindberg, Yudong Wang, Marie
Englund, Richard Thompson, Johan Ulander
H. Wan, A. Holmén, M. Någård, W. Lindberg, J. Chromatogr. A. 979 (2002) 369.H. Wan, A. Holmén, Y. Wang, W. Lindberg, M. Friberg, M. Någård, R. Thompson, Rapid Commun. Mass Spectrom. 17 (2003) 2639.H. Wan, R. Thompson; Drug Discovery Today, Technologies, 2(2005), 171.H. Wan, J. Ulander, Expert Opinion on Drug Metabolism & Toxicology, 2 (2006), 1389.
• Eva Emanuelsson (validation), Fredrik Bergström (new pKa-fit template test)
• Eva Thorin (Activitybase/Dixy group), Michael Wirth Färdigh (automation)
• Zhiping You (new pKa-fit program), AZ Boston
• Lead generation, Physical Chemistry group, AZ Mölndal