RPRP

High Efficiency LC/MS Analysis of O GlcNAc Modified Synthetic PeptidesHigh Efficiency LC/MS Analysis of O‐GlcNAc Modified Synthetic Peptides g y / y y p1 2 3 l 3 h i h 1 l d 2 3Barry Boyes1, 2, 3; Alex Harvey3; Stephanie Schuster1; Ron Orlando2, 3y y ; y ; p ;

1Advanced Materials Technology Inc Wilmington DE; 2Complex Carbohydrate Research Center University of Georgia Athens GA; 3GlycoScientific Inc Athens GA1Advanced Materials Technology Inc., Wilmington, DE; 2Complex Carbohydrate Research Center, University of Georgia, Athens, GA; 3GlycoScientific, Inc., Athens, GA

f f /GC/ S C f C f /G C /GObjectives Prediction of Resolution of Peptide/Glycopeptide Pairs LC/MS Conditions for RP and HILIC of Peptides/Glycopeptides HILIC Better Resolves Matched Peptide/Glycopeptide PairsObjectivesE i di i f f i hi h l i LC/MS i f

• A mix of Ammonium Formate/Formic Acid is an attractive mobile phase compared to Formic Acid alone showing narrow• Examine conditions for performing high resolution LC/MS separations of h d f d d l l h ff f To uncover new polypeptide glycosylation sites prediction of RP and HILIC retention effects of site occupancy

500mV500mV

A mix of Ammonium Formate/Formic Acid is an attractive mobile phase, compared to Formic Acid alone, showing narrow and symmetrical peaks, improved load tolerance (linear isotherm), and good compatibility with online ESI‐MS detectionhexosamine modified peptides, particularly the effects of separation  RP HILIC

To uncover new polypeptide glycosylation sites, prediction of RP and HILIC retention effects of site occupancy by an O‐linked carbohydrate would be useful. As shown below, the correlation of resolution of GP/P pairs400

450

.866

8 400

450

and symmetrical peaks, improved load tolerance (linear isotherm), and good compatibility with online ESI MS detection (McCalley, J. Chromatogr. A, 1038 (2004), p. 77; Schuster, Boyes, Wagner, Kirkland, J. Chromatogr. A, 1228 (2012), p. 232, mode (RPC and HILIC) 

APP695‐GP APP695 GP

by an O linked carbohydrate would be useful. As shown below, the correlation of resolution of GP/P pairs using RP and HILIC modes of separation is poor, implying that the mechanisms are very different. 350

400 56.

108

350

400

( y g ( ) p y g g ( ) pJohnson, Boyes, Fields, Kopkin, Orlando, Journal of Biomolecular Techniques: JBT (2013).)• Evaluate the utility of LC with online MS for identification and purity 

AAPP695 GP APP695‐Pep

APP695‐PepAPP695‐GP g p p , p y g y

Resolution Correlation300

0.60

8

300 21.5

46

• RP and HILIC conditions can employ the same mobile phases, with reversal of the gradient elution composition: RPl d l ( h l )

analysis of hexosamine modified peptides.  A210APP695‐Pep Resolution Correlation 

200

250

200

250

9.48

5

increases acetonitrile, HILIC decreases acetonitrile (water is the strong solvent).• Figure 1 presents the TIC chromatograms for a standard peptides separated in RP using the Halo Peptide ES C18 column

• Define the retention and selectivity differences between RPC and HILIC y = 0 2402x + 1 2808

15.00150

200

150

200 1

• Figure 1 presents the TIC chromatograms for a standard peptides separated in RP using the Halo Peptide ES‐C18 column and in HILIC mode using the Halo Penta HILIC column Column efficiencies are very comparable and resolution of thismode of separating hexosamine modified and unmodified peptides y = 0.2402x + 1.2808

R² = 0 1379100

0.68

8 100

988

and in HILIC  mode using the Halo Penta HILIC column. Column efficiencies are very comparable, and resolution of this mixture is readily achieved by either mode of separation.

R  = 0.137910.00P0

50

0.48

60

0.85

4

0

50 22.

mixture is readily achieved by either mode of separation. • Retention (Rt), peak widths (W1/2), resolution (Rs) and signal intensities in UV absorbance and in ESI‐MS total ion currents • Introduction Rs

 R0.0 2.5 5.0 7.5 10.0 12.5 min

0

15.0 17.5 20.0 22.5 25.0 27.5 min

0

( ), p ( 1/2), ( ) gwere determined for the 26 peptides shown in Table 1, using conditions similar to those shown in Figure 1. The 

IntroductionRecent developments HPLC instruments and column packing materials are permitting 5.00R

1.15

Inten. (x10,000,000)

687 3

9.0

Inten. (x1,000,000)

788.827500000

30000000 TIC(+)

156/

687

1.15

Inten. (x10,000,000)

687.3

8.5Inten. (x1,000,000)

788.827500000

30000000 TIC(+)

glycopeptides and peptides were injected at known quantities, to allow assessment of ESI‐MS signals, normalized to UV‐Recent developments HPLC instruments and column packing materials are permitting faster separations, for reversed‐phase (RP) and Hydrophilic Interaction Liquid  0.95

1.00

1.05

1.10 687.3

7.5

8.0

8.5

25000000

89

6.1

0 95

1.00

1.05

1.10

7.0

7.5

8.0

25000000

0/68

7

derived concentration data.p , p ( ) y p q

Chromatography (HILIC) analyses of peptides, protein fragments and intact proteins.  The  Figure 1. RP and HILIC Peptide Separations 0.000.80

0.85

0.90

6 0

6.5

7.0

20000000

22500000

5.90

7/78

0.80

0.85

0.90

0.95

6.0

6.5

20000000

22500000

19.5

10

89

recent popularity of sub‐2 μm diameter particles and the new development of small 

9

Figure 1. RP and HILIC Peptide Separations2.1 mm ID x 100 mm, 0.35 mL/min, 40°C, MS: SQ TIC (+ 300‐2000 m/z) @ 0.35/s MS 0.00 5.00 10.00 15.00 20.00 25.00 30.00

0 60

0.65

0.70

0.75

5.0

5.5

6.0

17500000

20000000

0.65

0.70

0.75

4.5

5.0

5.5

17500000

20000000

21.5

88/7

diameter superficially porous particles (SPP) designed for biomolecules is permitting f t d hi h l ti ti f tid d t i f t

3000000TIC(+)

1.562/499

/478

7/464

7S5 S4S3 S1HILIC: 100‐10%B in 45 min (‐1% AcN/min)

MSRs HILIC

0.45

0.50

0.55

0.60

3.5

4.0

4.5

150000000.45

0.50

0.55

0.60

3.5

4.0

15000000

faster and higher resolution separations of peptides and protein fragments.  2500000 11

12.717/

13.65

4.287/45

549/443S5 S4

S2HILIC: 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 Given that the peak widths are the same for HILIC and RP, improved HILIC resolution of GP/P pairs must be0.30

0.35

0.40

2.0

2.5

3.0

10000000

12500000

0.30

0.35

0.40

2.0

2.5

3.0

10000000

12500000

Modification of serine or threonine residues of proteins by β‐D‐N‐acetylglucosamine2000000

14

17.5A‐40%AcN/0.1% FA/10 mM AmmForm

Halo Penta‐HILICGiven that the peak widths are the same for HILIC and RP, improved HILIC resolution of GP/P pairs must be driven by selectivity differences. Comparing the plot of the differences in retention time (ΔRt, representing 

0.10

0.15

0.20

0.25

0 5

1.0

1.5

7500000

10000000

0 10

0.15

0.20

0.25

1.0

1.5

7500000

10000000

Modification of serine or threonine residues of proteins by β D N acetylglucosamine(GlcNAc) has emerged as a significant biological signaling mechanism. O‐GlcNAcylation of  1500000

y y p g p p gAcN difference at elution) for each mode, for each peptide pair, also yields a poor correlation. 500 750 1000 1250 1500 1750 m/z

0.00

0.05

1372.9

500 750 1000 1250 1500 1750 m/z0.0

0.5

1051.5 1577.1

5000000500 750 1000 1250 1500 1750 m/z

0.00

0.05

0.10 1372.9

636.6

500 750 1000 1250 1500 1750 m/z0.0

0.5

687.3 1577.11051.65000000

310/

723

(G c c) as e e ged as a s g ca t b o og ca s g a g ec a s . O G c cy at o orelevant sites can involve meaningful cross‐talk with phosphorylation targets, both  1000000

h l i f l i b h i bili f i f i d0

2500000

0

2500000 20.3

nearby and at distant sequences. Similarly modification of polypeptides by α‐GalNAc500000 MUC5AC3

The poor correlation of ΔRt results, in part, by the greater variability of separation performance in RP mode, and the smaller effect of the HexNAc modification on retention for RP the mean value of ΔRt (Table 1) is700mV800mV

52

0.0 2.5 5.0 7.5 10.0 12.5 min0

15.0 17.5 20.0 22.5 25.0 27.5 min0

(mucin antigen Tn) is of significant interest. For a variety of purposes, we have prepared 8 30 id h i id i h d i h O Gl NA difi i i d

0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5 minTIC(+)3000000 RP HILICMUC5AC3/13 MUC5AC13

MUC5AC3 and the smaller effect of the HexNAc modification on retention; for RP the mean value of ΔRt (Table 1) is‐0 73 (74% rsd) compared with 2 17 (22% rsd) for HILIC HILIC is better in both absolute terms for resolving600 9 20

.465

700

7.75

8.44

4

.304

8‐30 residue synthetic peptides with and without O‐GlcNAc modifications at serine and threonine residues LC/MS methods to qualify the purity and identities of such peptides 2500000

S5S4S3RP: 2‐58.2%B in B in 45 min (1% AcN/min)B 80%A N/0 1% FA/10 M A F

RP HILICMUC5AC

MUC5AC3/13

MUC5ACMUC5AC3/13

‐0.73 (74% rsd), compared with 2.17 (22% rsd) for HILIC. HILIC is better in both absolute terms for resolving GP/P pairs, as well as less variable in the retention effect of a HexNAc addition to the peptide sequences.500

600

16.4

06

496 18

.669

2

600 8.52

3

9.

threonine residues. LC/MS methods to qualify the purity and identities of such peptides using both Hydrophilic‐interaction chromatography (HILIC) and reversed‐phase 2000000 91

/464

411/478

333/499

S2B‐80%AcN/0.1% FA/10 mM AmmFormA‐0.1% FA/10 mM AmmForm

l id S 8 AMUC5AC

MUC5AC3MUC5AC

/ GP/P pairs, as well as less variable in the retention effect of a HexNAc addition to the peptide sequences. Δ Rt Correlation400

500

18.4

500using both Hydrophilic‐interaction chromatography (HILIC) and reversed‐phase chromatography (RPC) are compared. With many peptide/glycopeptide pairs we observe 1500000

2000000

35/443

805/457

10.79

13.4

16.3

S1Halo Peptide ES‐C18 A210

MUC5AC3MUC5AC13

Δ Rt Correlation 400

400chromatography (RPC) are compared.  With many peptide/glycopeptide pairs we observe much improved separations and LC/MS features using HILIC, compared to RPC. LC/MS 

1500000

8.03 9.

S1

0 6387 0 64844300

8

300 0.61

0

p p / g , p /using high performance HILIC permits rapid and sensitive analysis of O‐GlcNAcylated

1000000 y = 0.6387x ‐ 0.6484R² = 0 32663

200

88 88 7016

.878

8 6 039

0

200

862

.275

8 8.84

5

peptides. A collection of sequence‐matched glycopeptide/peptide pairs are compared for  0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5 min500000

R  = 0.32663

RP

100 15.5

815

.88

16.7

7

17.3

1

17.7

5

19.

19.4

1

22.4

30

100

96 0.69

0 0.8

34 7.52

0

8.18

38.

8.70

89.

056

9.56

5

various separation features in RP and HILIC mode, as well as for detection by absorbance t l l th (210 ) d li l t i i ti t t (ESI Table 1 Paired comparison of RP and HILIC for Glycopeptide/Peptide LC/MS Th 26 tid h l d b LC/MS i

2

ΔRt 

10.0 12.5 15.0 17.5 20.0 22.5 min0

0.0 2.5 5.0 7.5 10.0 12.5 min0 0.

49

1.13 7 9

at low wavelength (210 nm) and on‐line electrospray ionization mass spectrometry (ESI‐MS) using the single quadrupole analyzer

Table 1. Paired comparison of RP and HILIC for Glycopeptide/Peptide LC/MS. The 26 peptides shown were analyzed by LC/MS using 2.1 x 100 mm columns, with a flow rate of 0.4 mL/min at 60°C. Gradient conditions: A – 0.1% formic acid/10 mM ammonium formate; B – 90% 1

Δ

3 75

4.00

Inten. (x1,000,000)

954.9

5.25

5.50

Inten. (x1,000,000)

751.8

14000000

15000000 TIC(+)

14000000

15000000 TIC(+)

55MS) using the single quadrupole analyzer.  2.1 x 100 mm columns, with a flow rate of 0.4 mL/min at 60 C. Gradient conditions: A  0.1% formic acid/10 mM ammonium formate; B  90% Acetonitrile in A.  500 pmol was injected, with detection at 210 nm, MS operated +4.5 kV, with 0.45 s scan from 500 to 2000 m/z. RP Gradient ‐ 4% to 

/ ( / ) / ( / ) f

10 95

1.00

1.05

Inten. (x1,000,000)

751.7

3.25

3.50

3.75

4.50

4.75

5.00

13000000

14000000

92/9

5513000000

14000000

7.79

2/95

339/

752

Materials and Methods34% AcN/30 min (1%/min); HILIC Gradient ‐ 90% to 60% AcN/30 min (‐1%/min). The sequences were selected based on known O‐HexNAc modified sites in a variety of proteins Peak widths; means were 0 092 min for RP and 0 090 min for HILIC Average and range of Rt for the modes were similar

00.80

0.85

0.90

0.95

2.75

3.00

3.50

3.75

4.00

4.25

11000000

12000000

20.4

9

11000000

12000000 9.3

Materials and MethodsC l f HALO P tid ES C18 d H l P t HILIC d d t Ad d

in a variety of proteins. Peak widths; means were 0.092 min. for RP and 0.090 min. for HILIC. Average and range of Rt for the modes were similar.

M Δ R RP Δ R HILIC

0.00 1.00 2.00 3.00 4.000.65

0.70

0.75

1502.12.00

2.25

2.50

2.75

3.00

3.25

3 50

9000000

10000000

752

9000000

10000000

5/85

37Columns of HALO Peptide ES‐C18 and Halo Penta HILIC were produced at Advanced 

Materials Technology Inc (Wilmington DE) Both materials employ Fused Core® silicaPeptide Description Sequence

Mass (neutral) Rt RP (min)

Δ Rt RP (GP‐P) Rs RP Rt HILIC (min)

Δ Rt HILIC (GP‐P) Rs HILIC

APP695‐14GPep VPTT(OGlcNAc)AASTPDAVDK 1574 8 5 87 21 55MS Δ Rt HILIC

0.50

0.55

0.60

1.50

1.75

2.00

2.25

2.50

8000000

9000000

16.4

65/7

6 /8538000000

9000000

8.47

555

0/15

77

Conclusions and Future DirectionsMaterials Technology Inc. (Wilmington, DE). Both materials employ Fused‐Core  silica particles of 2.7 μm diameter, a solid core of 1.7 μm diameter, and a shell thickness of 0.5

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 16.41 0.35

0.40

0.45

0 75

1.00

1.25

1 00

1.25

1.50

1.75

6000000

7000000

541/

1706

18.7

00/

6000000

7000000 8.5

Conclusions and Future Directionsparticles of 2.7 μm diameter, a solid core of 1.7 μm diameter, and a shell thickness of 0.5 μm. The Halo Peptide shell has 160 Å pores, and the Penta HILIC column 90 Å pores. 

U 5 S S S 50 6 9 8 6MUC5AC‐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 10.72 0.20

0.25

0.30

0.25

0.50

0.75

0.25

0.50

0.75

1.00

1502.2

4000000

5000000 18.5

88/9

047 74000000

5000000

μ p p , pSynthetic peptides were synthesized at the CCRC (Lab of Prof. Geert‐Jan Boons), or  •High performance HILIC and RP HPLC separations were conducted to define preferred mode of

MUC5AC3/13 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 18.59 0.05

0.10

0.15

694 0

500 750 1000 1250 1500 1750 m/z0.00

974.0853.4 1273.1 1909.4

500 750 1000 1250 1500 1750 m/z0.00

694.2 1007.9

3000000

4000000

19.0

89.

467/

897

486/

1107

3000000

4000000

obtained from AnaSpec (Freemont, CA), or from Colin Mant and Prof. R.S. Hodges (U. High performance HILIC and RP HPLC separations were conducted to define preferred mode of practice to resolve O‐GlcNAc and O‐GalNAc modified peptides.

P‐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 20.32

500 750 1000 1250 1500 1750 m/z0.00

694.0

1000000

2000000 19

22.4

1000000

2000000

Colorado, Aurora, CO). The Retention Standard Mix uses the S1‐S5 sequences:S1 RGAGGLGLGK A

p p p•Formic acid/Ammonium formate mixtures with acetonitrile gradient elution was effective for LC/MS 

P‐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 19.15 10.0 12.5 15.0 17.5 20.0 22.5 min

00.0 2.5 5.0 7.5 10.0 12.5 min

0

S1     RGAGGLGLGK‐Am S2 Ac RGGGGLGLGK Am

/ g /analysis of the glycopeptide and peptide mixtures using both RP and HILIC operation. Comparing RP versus HILIC for LC/MS analyses of the collection of paired glycopeptides and peptides (GP and P) reveals:

P‐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 17.29

S2 Ac‐RGGGGLGLGK‐AmS3 Ac‐RGAGGLGLGK‐Am

y g y p p p p g p•HILIC better resolved glycopeptide/peptide mixtures.

Comparing RP versus HILIC for LC/MS analyses of the collection of paired glycopeptides and peptides (GP and P) reveals:• As previous observed for unmodified peptides and tryptic digests, RP and HILIC are orthogonal (retention in each mode does not

P‐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 A CFELLPTPPLSP 1354 7 27 16 2 00 8 88 2 71 2 93 20 11S3 Ac RGAGGLGLGK Am 

S4 Ac‐RGVGGLGLGK‐Am •HILIC exhibited some potential to resolve O‐GalNAc modifications  at variant sites (3 and 13) on the As previous observed for unmodified peptides and tryptic digests, RP and HILIC are orthogonal (retention in each mode does not significantly correlate to the other). 

P‐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 91S c G GG G G

S5 Ac‐RGVVGLGLGK‐Am MUC5AC peptide sequence.g y )

• In all cases, glycopeptides were retained less in RP, eluting at or before the unmodified peptide with the same sequence. P‐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 42

The instrument was the Shimadzu Nexera LC‐30 components (40 µL mixer), with the  •Additional glycopeptide/peptide pairs will be examined to improve the predictability of retention • In HILIC operation, glycopeptides were retained more, eluting after the unmodified peptide.P‐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 21.63

SPD20 Absorbance detector, fitted with a high sensitivity semi‐micro flow cell, and the MS 2020 i l d l MS d 4 5 kV ill i l i

prediction for O‐HexNAc peptide modifications.• The difference in retention time (%AcN) between paired GP and P were much larger in HILIC than RP. F ll i f l tid / tid l ti i hi h i HILIC th RP

P 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 10.28

MS‐2020 single quadrupole MS, operated at +4.5 kV capillary potential, scanning between 300 2000 m/z at 0 3 or 0 45 sec/scan Chromatographic parameter measures d b

• For all pairs of glycopeptides/peptides, resolution is higher in HILIC than RP. • TIC signal is slightly lower ( 35%) in RP than HILIC for either GP or P but variable with sequence This was the case comparing either

pGP‐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 24.73between 300‐2000 m/z at 0.3 or 0.45 sec/scan. Chromatographic parameter measures 

used the absorbance data at 210 nm generated with LabSolutions v 5 54 software•Supported by NIH NIGMS GM099355, GM093747.• TIC signal is slightly lower (‐35%) in RP than HILIC, for either GP or P, but variable with sequence. This was the case comparing either 

all peptides and glycopeptides as groups or by pair‐wise GP/P comparisons HILIC conditions appear less noisyAverage 13.01 ‐0.73 4.93 15.29 2.17 15.21used the absorbance data at 210 nm, generated with LabSolutions v.5.54 software.. all peptides and glycopeptides as groups, or by pair wise GP/P comparisons. HILIC conditions appear less noisy.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