HILIC Analysis of Bio Applications

HILIC Analysis for Polar Modifications of mAbs

Barry Boyes, Ph.D.Chief Technology OfficerAdvanced Materials Technology, Inc.Wilmington, Delaware, [email protected]

Outline

• Monoclonal antibody structure• Post-translational modifications

– Polar PTMs• Glycosylation• Deamidation• Oxidation

• Conclusions

2

IgG1 Monoclonal Antibody Structure

• ~ 150 kDa• Two heavy chains, 50 kDa each• Two light chains, 25 kDa each• Disulfide bridges connect the heavy

chains to the light chains and the heavy chains to each other

• Glycans attached at Asn297 of each heavy chain

3

Heavy Chain 1

Heavy Chain 2

Light Chain 1

Light Chain 2

Binding Region

Glycan Glycan

Binding Region

Disulfide Bridges

Polar PTMs discussed today will include:

• N- and O-linked glycosylation, adding one or many carbohydrate residues to the protein

• Deamidation and isomerization at Asn and Gln• Oxidation

Protein Polar Post Translational Modifications (PTMs)

4

GLYCOSYLATION

5

Protein Glycosylation

• Carbohydrates/sugars linked to a protein• Why is glycosylation important for mAbs?

– Impacts the safety/immunogenicity– Influences the efficacy and clearance– Impacts the stability and solubility– As a critical quality attribute, must be characterized

6

LC-MS

m/z 1536.8

m/z 1683.0m/z 1004.0

10 15 20 25 30 35 40 45 50 55 60 min0.20.30.40.50.60.70.80.91.01.11.21.31.41.5

(x10,000)

FA2G

0-Gl

cNAc

A2G0

(G0)

FA2G

0-(G

0F)

Man

5

A2G1

-(G1

)

A2BG

FA2G

1-(G

1F)

A2G2

-(G2

)

FA2G

2-(G

2F)

FA2G

1S1

FA2G

2S1

Glycosylation Variants of Trastuzumab

Galactose (Gal)

Mannose (Man)

N-Acetyl-D-glucosamine (GlcNAc)

Fucose (Fuc)

Sialic acid (Sia)

• HILIC with PNGase F released N-linked glycans reductively aminated by procainamide• Great peak shape and great resolution of compositional isomers• Covers the entire range of N-linked glycovariants that are present in mAbs

7

Column – HALO 90 Å Glycan, 2.7 µm, 2.1 x 150 mm; 0.35 mL/min; 40°C. Gradient Eluents: A –50 mM ammonium formate (pH 4.5); B – ACN Gradient: 75% to 68% ACN in 75 min

O

O-

O P O-

• O-GlcNAc reversibly modifies protein Ser and Thr residues.• O-GlcNAc is a modifier of biological activity, in some cases, with competition for phosphorylation.• Multiple independent sites on a particular protein can be modified by –P or –GlcNAc, near-by or far apart.

Hundreds to thousands of proteins central to biological process and diseases such as cancer, diabetes and neurodegeneration are O-GlcNAcylated.

OHO

HOOH

O

AcHNOHOGT

O-GlcNAcase

Kinase

Phosphorylase

β-O-(N-acetylglucosamine) Modifications of Proteins (O-GlcNAc)

8

LC/MS of O-GlcNAcylated Peptides: RP vs HILICAPP695-14GPep VPTT(OGlcNAc)AASTPDAVDKAPP695-14Pep VPTTAASTPDAVDK

Conditions: Columns - 2.1 x 100 mm HALO 90 Å Penta-HILIC or HALO 160 Å ES-C18; 0.4 mL/min; 60°C. Gradient Eluents: A – 0.1% formic acid/10 mM ammonium formate; B – 90% AcN in A. RP Gradient - 4% to 34% AcN/30 min (1%/min); HILIC Gradient - 90% to 60% AcN/30 min (-1%/min).

RP HILIC

A210

MS

APP695-GP APP695-PepAPP695-Pep

APP695-GP

0.0 2.5 5.0 7.5 10.0 12.5 min

0

50

100

150

200

250

300

350

400

450

500mV

0.48

60.

608

0.68

80.

854

5.86

66.

108

15.0 17.5 20.0 22.5 25.0 27.5 min

0

50

100

150

200

250

300

350

400

450

500mV

19.4

85

21.5

46

22.9

88

0.0 2.5 5.0 7.5 10.0 12.5 min0

2500000

5000000

7500000

10000000

12500000

15000000

17500000

20000000

22500000

25000000

27500000

30000000 TIC(+)

5.90

7/78

9

6.15

6/68

7

15.0 17.5 20.0 22.5 25.0 27.5 min0

2500000

5000000

7500000

10000000

12500000

15000000

17500000

20000000

22500000

25000000

27500000

30000000 TIC(+)

19.5

10/6

87

20.3

10/7

23

21.5

88/7

89

9

HILIC conditions using HALO Penta-HILICRP conditions using HALO 160 Å ES-C18

This sample of 12 peptides and 14 glycopeptides reveals:

HILIC shows higher resolution (3X), at lower variance (2X), compared to RP

Driven by better separation selectivityof this polar modification

Note results for O-GalNAc in modified peptides, so these averages actually refer to O-HexNAc effects.

Peptide DescriptionSequenceMass

(neutral) Rt RP (min)∆ Rt RP (GP-P) Rs RP Rt HILIC (min)

∆Rt HILIC (GP-P) Rs HILIC

APP695-14GPep VPTT(OGlcNAc)AASTPDAVDK 1574.8 5.87 21.55APP695-14Pep VPTTAASTPDAVDK 1371.7 6.11 -0.24 1.90 19.49 2.07 9.41MUC5AC GTTPSPVPTTSTTSAP 1501.6 9.28 16.41MUC5AC-3 GTT(OGalNAc)PSPVPTTSTTSAP 1704.6 8.45 -0.83 6.88 18.68 2.27 13.40MUC5AC-13 GTTPSPVPTTSTT(OGalNAc)SAP 1704.6 8.53 -0.75 5.82 18.51 2.10 10.72MUC5AC3/13 GTT(OGalNAc)PSPVPTTSTT(OGalNAc)SAP 1908.1 7.76 -1.52/2 11.84 20.48 4.07/2 23.35GP-41 Ac-CSTFRPRT(OGlcNAc)SSNAST 1758.8 7.09 18.59P-42 Ac-CSTFRPRTSSNAST 1555.7 7.03 0.06 0.44 17.03 1.56 11.58GP-78 Ac-CQHPPVT(OGlcNAc)NGDTVK 1639.8 6.47 20.32P-84 Ac-CQHPPVTNGDTVK 1436.7 6.56 -0.10 0.66 18.72 1.61 11.23GP-79 Ac-CKIADFGLS(OGlcNAc)KIVEHQ 1932.0 19.36 19.15P-85 Ac-CKIADFGLSKIVEHQ 1728.9 20.80 -1.44 8.16 17.21 1.94 14.76GP-17s CTLHTKAS(OGlcNAc)GMALLHQ 1854.9 13.62 17.29P-20s CTLHTKASGMALLHQ 1651.8 14.23 -0.61 3.06 15.15 2.14 15.38GP-15 Ac-CFELLPT(O-GlcNAc)PPLSP 1557.8 25.16 5.64P-18 Ac-CFELLPTPPLSP 1354.7 27.16 -2.00 8.88 2.71 2.93 20.11GP-46 Ac-CRSSHYGGS(OGlcNAc)LPNVNQI 1975.9 12.48 17.32P-47 Ac-CRSSHYGGSLPNVNQI 1772.8 12.96 -0.48 3.83 15.43 1.89 13.91GP-51 Ac-CSALNRTS(OGlcNAc)SDSALHT 1806.8 9.08 17.23P-52 Ac-CSALNRTSSDSALHT 1603.7 9.55 -0.47 3.85 15.55 1.69 12.42GP-16 Ac-CKIPGVS(OGlcNAc)TPQTL 1487.7 16.41 13.27P-19 Ac-CKIPGVSTPQTL 1284.6 16.98 -0.58 3.74 10.59 2.68 21.63GP-2-p53 Ac-CQLWVDS(OGlcNAc)TPPPG 1543.7 16.43 12.72P-3-p53 Ac-CQLWVDSTPPPG 1340.6 17.66 -1.23 7.23 10.41 2.31 10.28GP-17r Ac-CLHTKAS(OGlcNAc)GMALL 1488.7 16.21 10.59P-20r Ac-CLHTKASGMALL 1285.6 16.98 -0.77 2.79 7.45 3.14 24.73

Average 13.01 -0.73 4.93 15.29 2.17 15.21

Standard Deviation 5.95 0.54 3.32 4.74 0.47 5.13

% RSD 45.7 74.3 67.3 31.0 21.8 33.7

LC/MS of O-GlcNAcylated Peptides: RP vs HILICHILIC ResultsRP Results

10

2.1 mm ID x 100 mm, 0.35 mL/min, 40 C, MS: SQ TIC (+ 300-2000 m/z) @ 0.35/s20 µg Bovine Ribonuclease B tryptic digest (CAM)

RP: 2-58.2%B in B in 45 min (1% AcN/min)B-80%AcN/0.1% FA/10 mM AmmFormA-0.1% FA/10 mM AmmForm

0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5 25.0 27.5 30.0 32.5 min0

250000

500000

750000

1000000

1250000

1500000

1170.40(+)(10.00)1089.40(+)(10.00)1008.40(+)(10.00)927.40(+)(10.00)846.45(+)(10.00)TIC(+)

0.92

5/60

8

1.57

1/59

11.

957/

475

2.13

0/92

8

2.32

8/84

72.

622/

451

2.89

0/39

3 5.00

9/75

1 6.83

2/45

8

8.10

9/80

9

9.23

2/67

69.

783/

536 9.

988/

552

10.2

19/6

10

11.1

75/1

184 11

.331

/104

611

.513

/778

11.6

42/7

7812

.510

/117

5 12.7

04/8

4312

.983

/836

13.8

49/5

4314

.330

/651

14.7

57/1

144

14.9

30/1

145

16.9

76/9

61

19.6

63/8

41

19.8

59/8

4020

.421

/744

20.8

30/8

4121

.217

/841

21.3

89/8

3521

.919

/927

22.6

52/1

080

22.8

19/8

35

23.6

31/8

3523

.926

/835

24.8

20/8

35

HALO 160 Å ES-C18No resolution of high

mannose variant glycopeptide

kml

HILIC: 100-10%B in 45 min (-1% AcN/min)B-90%AcN/0.1% FA/10 mM AmmFormA-40%AcN/0.1% FA/10 mM AmmForm

0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5 25.0 27.5 30.0 32.5 min0

250000

500000

750000

1000000

1250000

1500000

1170.40(+)(20.00)1089.40(+)(20.00)1008.40(+)(20.00)927.40(+)(20.00)846.45(+)(20.00)TIC(+)

0.62

7/66

41.

030/

484

5.81

9/12

14

6.49

8/13

91

9.88

1/52

4

11.1

19/7

12

13.2

45/6

16

13.8

99/5

1514

.047

/543

15.2

83/4

5115

.647

/458

14.7

43/4

75

16.5

19/5

9117

.047

/107

117

.453

/746

18.1

85/1

080

18.5

39/6

32

19.2

37/6

6219

.593

/454

20.0

33/6

63

20.3

36/6

08

20.9

80/8

3521

.468

/835

21.7

68/8

3622

.057

/360

22.9

69/8

4023

.645

/118

4 23.8

60/8

4124

.253

/118

4 24.5

10/1

184

24.9

47/7

95

25.6

23/8

47

26.5

92/9

28

27.4

90/1

009

28.3

09/1

090

31.7

77/1

456

33.9

42/5

74

28.9

43/1

171

HALO Penta-HILIC

Good resolution of high mannose variant

glycopeptide

HILIC Strongly Retains and Resolves N-linked Glycopeptides

11

HILIC SRM analysis of human serum IgGs demonstrating the ability to resolve isomeric glycopeptide glycoforms.

Comparison released N-glycans and glycopeptides of Fetuin. (A) Procainamide labeled released N-glycans. (B) Glycopeptides with the same peptide backbone.

IgG 2/3 (EEQFNSTFR) IgG 4 (EEQFNSTYR) IgG 1 (EEQYNSTYR)

Conditions: Column – 2.1 x 150 mm HALO Penta-HILIC; 0.4 mL/min; 60°C. Gradient Eluents: A –50 mM ammonium formate (pH 4.4)/5% AcN; B – AcNGradient – RG: 80% to 62% AcN/60 min (-1%/min); GP 85% to 48%B 75 min

Huang, Y., Nie, Y., Boyes, B., and Orlando, R. (2016) J. Biomol. Technol., 27, 98-104. Tao, S.J., Huang, Y., Boyes, B.E., and Orlando, R. (2014) Anal. Chem., 86, 10584-10590.

LC-MS/MS of N-Linked Glycans ON and OFF Peptide by HILIC/SRM

12

Separation of Glycopeptides of Hemopexin2.1 mm x 150 mm, A: 0.1% formic acid in water; B: 0.1% formic acid in ACN

Gradient conditions same for Penta-HILIC and BEH Amide; shallower for ZIC-HILIC, 0.3 mL/min, 40 °C

Molecules 2020, 25, 4655

• Increased retention for Penta-HILIC for the di-sialylated glycopeptides

• Reduced retention of sialylatedglycans on ZIC-HILIC may be due to ion repulsion between the negatively charged sulfobetaine functional group and the negatively charged sialic groups of the glycans

• Reduced elution range observed on the BEH amide (< 2 min)

13

Separation of Fucosylated Glycopeptide Isomers

Molecules 2020, 25, 4655

2.1 mm x 150 mm, A: 0.1% formic acid in water; B: 0.1% formic acid in ACNGradient conditions same for Penta-HILIC and BEH Amide; shallower for ZIC-HILIC, 0.3 mL/min, 40 °C

• Normalized EIC chromatograms of A2G2F1 glycoforms of SWPAVGN187CSSALR (PEP1) (A–C) and ALPQPQN453VTSLLGCTH (PEP2) (D–F) on 3 different HILIC columns

• HALO Penta-HILIC resolves the glycoforms of both PEP1 and PEP2

• BEH Amide resolves the glycoformsof PEP1, but not PEP2 while ZIC-HILIC shows complete coelution of the glycoforms of PEP1 and PEP2

14

Separation of Sialylated Glycopeptide Isomers

Molecules 2020, 25, 4655


• Normalized EIC chromatograms of A2G2F1 glycoforms of SWPAVGN187CSSALR (PEP1) (A–C) and ALPQPQN453VTSLLGCTH (PEP2) (D–F) on 3 different HILIC columns

• HALO Penta-HILIC shows baseline resolution of the glycoforms of both PEP1 and PEP2

• BEH Amide is starting to resolve the glycoforms of PEP1 and PEP2 while ZIC-HILIC shows coelutions of the glycoforms of PEP1 and PEP2

15

Separation of Glycopeptides of IgG

Molecules 2020, 25, 4655


• Different selectivity observed across the 3 HILIC phases studied

• Increased resolution for Penta-HILIC compared to the other 2 HILIC phases

16

DEAMIDATION

17

Deamidation/Isomerization of Asparagine• Deamidation of asparagine (N) and glutamine (Q) residues occurs in both peptides and in intact proteins• Rate of these deamidation reactions is strongly condition (pH, T) and sequence dependent (C-terminal G)• Mechanism of ammonium loss is understood to occur as an irreversible reaction through a 5 membered (N) or 6

membered (Q) cyclic intermediate – shown below for Asn with intermediate cyclic succinimidyl structure (sN)• Symmetrical intermediate will hydrolyze to Asp (D) or, via polypeptide backbone rearrangement, to iAsp (iD)• Asp/iAsp dehydration to the cyclic intermediate is reversible, resulting in an equilibrium distribution of Asp/iAsp

containing peptides or polypeptides• Formation of Asp or iAsp results in a more polar modified peptide or polypeptide, amenable to HILIC resolution.

At pH >3 ionization of the carboxylate occurs.

N D

iD

sN

Badgett, M.J., Boyes, B.E., Orlando, R.C. Am. Soc. Mass Spectrom. 28: 818 (2017)

18

Why is Deamidation of mAbs Important?

• Could alter the mAb structure and function– Reduced bioactivity– Change to pharmacokinetics– Change to antigenicity

• Could change the stability, leading to degradation

19

Selected Tryptic Peptides from TrastuzumabIgG Peptides Studied

IgG Peptide with Multiple Asn as a Model System:

GFYPSDIAVEWESN388GQPEN393N394YK

Trastuzumab Light Chain25 ASQDVNTAVAWYQQKPGK 42 N30T

Trastuzumab Heavy Chain76 NTAYLQMNSLR 86 N83S99 WGGDGFYAMDYWGQGTLVTVSSASTK 124 D102G

279 FNWYVDGVEVHNAK 292 D284G306 VVSVLTVLHQDWLNGK 321 N319G375 GFYPSDIAVEWESNGQPENNYK 396 N388G--NN393N394Y421 WQQGNVFSCSVMHEALHNHYTQK 443 N425V--N438H

20

Comparison of HILIC and RP for ResolvingDeamidated and Isomerized Asn Peptides

GFYPSDIAVEWESNGQPENNYKGFYPSDIAVEWESDGQPENNYKGFYPSDIAVEWESiDGQPENNYK

Three peptides from IgG1 – variant at position 14 (IgG N388)

Columns: 2.1 x 150 mm HALO 160 Å ES-C18, 2.7 µm or HALO 90 Å Penta-HILIC, 2.7 µm; Flow rate: 0.4 mL/min; Temp: 60 oC; Mobile Phase A: water/50 mM Ammonium Formate, pH 4.4; Mobile Phase B: acetonitrile/0.1% Formic acidGradient: HILIC – 80%-46.2% in 60 min.; RP - 10-70% B in 60 min; Injection Volume: 4 µL (0.1 µg)

M+2 peptide ions M+2 peptide ions

• The same mobile phase conditions and temperature were employed, reversing the direction of the acetonitrile gradient to effect elution on RP and HILIC columns.

• Note the greater selectivity difference for resolving these peptides in HILIC, compared to RP.• Unlike RP, HILIC reliably resolves Asn/Asp/iAsp, with the retention order shown above

21

Deamidation/Isomerization of Asparagine at pH 9, 37 °C

22

Column: 0.5 x 150 mm Halo Penta-HILIC; Flow rate: 12 µL/min; Temp: 60 oC; Detection: Abs (220 nm) or Orbitrap Velos Pro MS Mobile Phase A: water/50 mM Ammonium Formate, pH4.4; Mobile Phase B: acetonitrile/0.1% Formic acid; Gradient: Hold 80%B for 4 min.; 80%-48% in 64 min.

N388G

D388G iD388GUV

A

B

C

N388G

N388G

D388G

iD388G

N388GD388G

iD388G

iD388G—iD393N

UV

MS

• Mass analysis and MS/MS fragmentation identifies the N388G—(D/iD)393N peptide at 29.8 min.• Degradation of iD388G—N393N rapidly formed two peptides with 1 Da shift at 31.69 and 32.80 min, confirmed as

deamidations to D or iD at position 393, and eventually all peptides degraded to predominantly 34.53 minutes, the iD388G—iD393N peptide. Supported by mass analysis, using CID and ETD fragmentation.

• No evidence of the formation of a triple deamidation was obtained (N394).

Deamidation/Isomerization of Asparagine at pH 9, 37 °C

+1 Da

+2 Da

23

N388G

iD388GD388GN388G

N388G—D/iD393N

Column: 0.5 x 150 mm Halo Penta-HILIC; Flow rate: 12 µL/min; Temp: 60 oC; Detection: Abs (220 nm) or Orbitrap Velos Pro MS Mobile Phase A: water/50 mM Ammonium Formate, pH4.4; Mobile Phase B: acetonitrile/0.1% Formic acid; Gradient: Hold 80%B for 4 min.; 80%-48% in 64 min.

C

D

E

F

• Trastuzumab– Tryptic digest of both native and stressed by incubation in Tris-HCl

pH 9.0 for 7 days at 4 mg/mL• Reduced and alkylated proteins digested 4 hrs in 50 mM Tris-HCl (pH 7.8)/1.5 M

Guanidine-HCl

– Analysis via HILIC capillary LC/MS using the Orbitrap/IT• Extracted ions at the monoisotopic masses of the target sequences were integrated• Reported sequences were confirmed by CID MS/MS fragmentation

Method Conditions for IgG Tryptic Digests

24

Deamidation and Isomerization in IgG Tryptic DigestsGFYPSDIAVEWESNGQPENNYK in Trastuzumab Digest

Native Stressed• 9 sites of potential modification were

analyzed• 3 Asn sites with significant

deamidation, and subsequent isomerization.

25

*JASMS, 28 (2017) 818-826 J Chromatogr A, 1537 (2018) 58-65

Deamidation and Isomerization in IgG Tryptic DigestsGFYPSDIAVEWESNGQPENNYK in Trastuzumab Digest

Native Stressed

WGGDGFYAMDYWGQGTLVTVSSASTK in Trastuzumab Digest

Native Stressed

• 9 sites of potential modification were analyzed

• 3 Asn sites with significant deamidation, and subsequent isomerization.

• 2 sites showed presence of D/iDpairs, neither were strongly affected by “stress”

• All susceptible sites exhibited the cyclic intermediate sN, whether formed from D or N in the native sequence.

26JASMS, 28 (2017) 818-826 J Chromatogr A, 1537 (2018) 58-65

OXIDATION

27

Oxidation of Methionine Mechanism

• Methionine (Met-S) oxidizes to form methionine sulfoxide (Met-SO)• Methionine sulfoxide (Met-SO) can further oxidize to form methionine sulfone (Met-SO2)• Met-SO can be reduced back to methionine using methionine sulfoxide reductase A (MsrA)

28

Why is Oxidation of mAbs Important?

• Met oxidation in mAbs linked to– Function loss– Folding stability decrease– Increase in propensity to aggregate

• Could happen at various development stages– Production– Formulation– Storage

29

Reversed-Phase Separations of Oxidized Peptides

• Tryptic peptides from BSA• Example in A is not resolved well

enough for quantitation

30J. Am. Soc. Mass Spectrom. (2017) 28:818-826

Reversed-Phase Separations of Oxidized Peptides

• Tryptic peptides from BSA• Example in A is not resolved well

enough for quantitation• Example in B is resolved well

enough for quantitation• No consistent resolution with

separation using reversed-phase so cannot be easily predicted

Column: 0.2 x 150 mm HALO 160 Å ES-C18, 5 µm; Flow rate: 2 µL/min; Temp: room temperature; Mobile Phase A: water/0.1% formic acid and 10 mM ammonium formate; Mobile Phase B: acetonitrile/0.1% formic acid and 10 mM ammonium formate; Gradient: 5%-75% B in 120 min.


HILIC Separation of IgG Oxidized Peptide

• IgG oxidized peptide is well resolved from the unmodified form and can easily be quantified

Column: 0.2 x 150 mm HALO 90 Å Penta-HILIC; Flow rate: 2 µL/min; Temp: room temperature; Mobile Phase A: 50 mM Ammonium Formate/0.1% formic acid; Mobile Phase B: acetonitrile/0.1% formic acidGradient: 95%-30% B in 90 min.


(-) DIQMTQSPSSLSASVGDRVTITC(Carbamidomethyl)R(A)-1305

HILIC Separation of IgG Oxidized Peptide from NISTmAb digest

Column: 0.5 x 150 mm HALO 90 Å Penta-HILIC; Flow rate: 50 µL/min; Temp: 60 oC; Mobile Phase A: 50 mM Ammonium Formate, pH 4.4; Mobile Phase B: acetonitrile/0.1% formic acidGradient: 80%-48% B in 55 min. 33

Conclusions

• Detailed characterization of PTMs continues to grow, particularly for therapeutic mAbs

• HILIC separations show increased resolution of polar PTMs of mAbs compared to reversed-phase separations

• HALO Penta-HILIC is well suited for separations of glycovariants and analysis of site occupancy

• HALO Penta-HILIC can yield quantitative details on protein deamidations and other chemical modifications

34

Acknowledgements

• Advanced Materials Technology, Inc.– Barry Boyes, Ph.D.– Andrew Harron, Ph.D.– Arianne Soliven, Ph.D.– Ben Libert– Conner McHale– Stephanie Rosenberg

• Professor Ron Orlando and students at the CCRC, UGA

35

Questions

36

Documents

HILIC Analysis of Bio Applications