Protein Clipping Variant Analysis of Vedolizumab using dedicated workflows in BioPharma Compass

Protein  Clipping Variant Analysis  of Vedolizumab 
using dedicated workflows in BioPharma Compass

Proteolytic degradation of biopharmaceuticals during their life cycle can pose health 
risks and require dedicated analysis.

Authors:  Christian Albers 1, Detlev Suckau 1, Tomasz Gozdziewicz 2; 

1

 Bruker Daltonics, Bremen, Germany; 2 Polpharma Biologics S.A., ul. Trzy Lipy 3, 80-034 Gdansk, Poland.

Keywords: 

Proteoform analysis,  

protein degradation, 

enzyme specificity, 

Middle-Up, Middle-Down, 

sequence analysis, 

Elute_UPLC, maXis II 

ETD, BioPharma  

Compass 2021

Abstract

In this work we used enzy-
matically introduced antibody 

truncation variants as a case 
study. These clipping variants 
were detected based on accurate 
mass measurements using a 
maXis II ETD and confirmed by 
LC-ETD and LC-CID Middle- 
Down protein sequencing. 
Typically, ETD yielded higher 
sequence coverages of the 

 
 
proposed clipping candidates. 
However, CID Middle-Down 
spectra provided sufficiently 

long terminal sequence tags to 
confirm candidates as well.

Dedicated clipping variant ana- 
lysis workflows in BioPharma 
Compass 

2021 enabled the 

detection and confirmation of 
these variants rapidly and easily.

Introduction

For Biopharmaceuticals, degra- 
dation of the active drug 

substance during production, 
formulation or storage poses a 
threat to drug efficacy and safety 
and is a critical quality attribute  
(CQA). The detection of such trun-
cated  protein  species  –  clipping  
variants  –  can  be  difficult  to 
achieve through classic peptide 
mapping with trypsin. 

Vedolizumab is a recombinant, 
humanized IgG1 monoclonal antibody 
directed against the human lympho-
cyte  α4β7 integrin, a key mediator 
of gastrointestinal inflammation. It is 
used in the treatment of moderate 
to severe active ulcerative colitis and 
Crohn's disease. 

In this case study, a Vedolizumab 
biosimilar was characterized after 
SpeB treatment, which cleaves IgG1 
predominantly in the hinge region at 
...KTHT | CPPCPAPE... corresponding  
to T229|C230 on Vedolizumab’s 

heavy chain. Minor cleavage products 
with uncharacterized specificity have 
also been observed.

We analyzed the reduced SpeB 
digest products with LC-MS using 
an ultrahigh resolution maXis II ETD 
QTOF instrument to characterize 
the enzyme specificity and evaluate 
the performance of the BioPharma  
Compass 

2021 clipping variants 

detection workflow. Middle-Down 
ETD and CID sequence validation 
measurements were subsequently 
performed to confirm the findings.

Experimental

Samples

A Vedolizumab biosimilar candidate  
drug substance (Polpharma Biologics) 
was measured after treatment with  
FabULOUS and IgGZERO enzyme 
kits (Genovis) followed by reduction  
using TCEP. Concentration after di- 
lution was approximately 0.5 mg/mL. 
The resulting major subunits are 
named following the usual IgG sub-
unit nomenclature: Fd, Fc/2 and LC.

Figure 1: Experimental design for the Vedolizumab protein clipping analysis. SpeB digestion produced Fab and Fc fragments. After reduction Fd, Fc, and LC 
subunits were analyzed by LC-MS, yielding the major expected subunits plus some low abundant fragments. Protein clipping analysis in BioPharma Compass 
provided candidates based on accurate mass and isotope pattern. They were confirmed by direct sequence analysis of the clipping product candidates by 
ETD and CID.

F

ab

F

d

... PKSCDKTHT

1  FabULOUS (SpeB) digest 

in the hinge region

2  Reduction of disulfide 

cross links

3 LC-MS 

Intact mass based 
clipping variant screening

4  CID + ETD Middle-Down 

verification of clipping 
variant candidates

CPPCPAPELAGAPSV ...

F

c

F

c/2

Figure 2: Top: Total Ion Chromatogram of the SpeB-digested, reduced and deglycosylated Vedolizumab 
with the annotated subunits peaks  5 -  7 . Peaks  1 -  4 were further characterized as putative protein 
clipping variants, which elute earlier in the chromatogram.

1 2

3

4

5

6

7

S

ig

na

l in

te

nsi

ty

x107

13.5

14.5

15.5

18.5

16.5

19.5

17.5

13

14

15

18

16

19

17

20

Time [min]

2.0

1.0

3.0

4.0

5.0

6.0

0

Fc/2 NF_15.24

LC_16.85

Fd pQ_18.78

LC-MS and MS/MS

A Bruker Elute UPLC with Waters 
Protein BEH C4, 300Å 1.7 µm, 

2.1 x 100 mm column was used for 
the protein separation.

Solvents: A: 0.1% Formic Acid in Water; 
  

B: 0.1% Formic Acid in ACN

Gradient:  0 min 10% B
 

1.5 min 10% B

2.5 min 15% B

34 min 45% B

35 min 95% B

40 min 95% B

40.2 min 10% B

45 min 10% B

A Bruker maXis II ETD Ultrahigh 

Resolution QTOF was used for 
MS spectra acquisition in the 

150-2800 m/z range. The samples 
were also subjected to ETD and CID  
fragmentation for Middle-Down 

protein sequence analysis using a 
mass range from 100-2400 m/z.

The precursor ion intensity for ETD 
experiments was controlled in all 
experiments using these parameters:

ICC Target Intensity Controlling: 
1000 Mio
ICC Time Accumulation Analyte 
min: 1 ms
ICC Time Accumulation Analyte 
max: 600 ms
Injection Time: 16 - 22 ms
Extended Reaction time: 4 - 6 ms

For CID, a 50/50 Stepping with 25 eV 
and 30 

eV collision energy was 

applied to all precursor ions.

Data Analysis

All datasets were transferred to 
BioPharma Compass 2021 for 
data processing and analysis. The 

Vedolizumab LC and HC sequences 
were imported from [1] as FASTA 
file. N-terminal pyro Q was defined 
as variable modification together 
with G0F, G1F and G2F as major 

glycoforms. The MAM Protein 
Screening workflow was used 
for intact mass analysis of the 
SpeB proteolysis products and the  
Top-Down Sequencing ESI work-
flow for the sequence analysis of the 
CID and ETD spectra. Both work-

flows allowed to specifically analyze  
protein clipping variants. All molecular 
weights calculated were based 
on the monoisotopic masses 

determined using the SNAP 

algorithm in BioPharma Compass.

Results

The regular SpeB cleavage products

The cleavage products resulting 

from Vedolizumab IgGZERO 

deglycosylation, SpeB treatment and 
reduction predominantly yielded the 
expected Fd [1-229], Fc/2 [230-450]  
and LC [1-219] subunits (peaks 5-7). 
In addition 4 additional peaks (1-4)  
were observed by LC-MS analysis 
(Figure 2).

Chromatographic peaks 5-7 contained 
predominantly – but not exclusively,  
see Figure 9 – the expected subunits 
as confirmed by the maximum 
entropy-deconvoluted mass spectra 
(Figure 3).

Figure 3: MaxEnt deconvoluted MS spectra of chromatogram peaks  5 -  7 , containing the IgGZERO deglycosylated Fc/2, the LC and pyro-glutamylated Fd 
subunit. All spectra were isotopically resolved and the calculated isotope patterns (red, LC in peak 6) show perfect alignment with the mass spectra. Minor 
additional peaks are indicative of, e.g., potential protein clipping products of the drug sample (see Figure 9). Monoisotopic molecular weights are annotated 
to the peaks.

5

6

7

In

te

nsi

ty

In

te

nsi

ty

x106

x105

23,900

25,000

23,910

25,100

23,920

25,200 25,300 25,400

Da

Da

0.2

0.2

0.1

0.1

0.3

0.3

0.4

0.4

0.5

0.5

0.6

0.6

0.8

0.8

0.7

0.7

0.9

0.9

1.0

1.0

0

0

In

te

nsi

ty

x105

24,300 24,400 24,500 24,600 24,700

Da

1.00

0.50

0.75

1.00

1.25

1.50

1.75

0

25142.528

Fc/2 NF

23891.777

LC

24576.039

Fd pQ

Figure 4: Top: ETD Middle-Down sequencing spectrum of the Fd subunit in peak 7; c-, y- and z+1 fragment ions were annotated. Center: Sequence coverage  
map of the ETD fragments (blue); top row of bricks represents observed c-ions, bottom row represents matching y- and z+1-ions. Bottom: Sequence 
coverage map of the CID fragments (red) from the CID spectrum (not shown); top row of bricks represents observed a- and b-ions, bottom row represents 
matching y-ions. The yellow bricks indicate tolerated gaps in the observed fragment ion ladders for the calculation of the SVP [2].

In

te

nsi

ty

x103

1000

G

D

y

z+1

z+1 (13)

z+1 (21)

z+1 (34)

z+1 (45)

z+1 (49)

z+1 (2)

D

D

D

T

T

Y

L

N

G

S

S

S S

E V

V

V

V

K

K

K

K

K

K

K

c (27)

c (31)

c

A S V K V S

S G Y

T

S

Y

R

L

I G

W

2000

3000

4000

5000

500

1500

2500

3500

4500

5500

m/z

0.50

0.25

0.75

1.00

1.25

0

1

2

3

Figure 5: MaxEnt deconvoluted MS spectra of chromatogram peaks  1 -  3 . Some clipping candidates are annotated, some aren’t even visible here such as 
HC [246-450] in peak 3 at 23605.831 Da (∆Mr -0.86 ppm).

In

te

nsi

ty

x105

12,800

12,400

12,600

Da

Da

1.00

0.50

0.75

1.00

1.25

1.50

1.75

0

In

te

nsi

ty

x104

12,200

12,600

13,000

12,400

12,800

Da

3.0

1.0

2.0

4.0

5.0

6.0

7.0

0

In

te

nsi

ty

x105

11,000

13,000

12,000

14,000

3.0

1.0

2.0

4.0

5.0

0

12397.121

LC [68-181]

12397.127

LC [68-181]

12698.287

HC [338-450]

12698.292

HC [338-450]

12269.029

HC [342-450]

Subsequent LC-ETD and LC-CID ana- 
lyses of each of the chromatographic 
peaks provided sequence information 
of the fragmented protein subunits, 
which were automatically calculated  
based on the Vedolizumab protein 
sequence. Sequence Coverage (SC) 
and Sequence Validation Percentage 
(SVP) [2] were determined; the SVP 
calculation tolerated a terminal gap 
size of 15 and internal gap size of 3 
missing fragments in the ion series. 
The observed SVP and SC values 
obtained for ETD were Fc/2:41/27%, 

LC: 60/43% and Fd:48/36%, and for 
CID were Fc/2:36/29%, LC:47/38% 
and Fd:28/21%, respectively. In 

average, 50% SVP were obtained from 
LC-ETD analysis for the 3 subunits  
and 37% SVP from LC-CID. The 
results confirmed the Vedolizumab 
subunit sequence even in case of the 
weakest spectra of the Fd (Figure 4).

Clipping variant analysis

The initial list of proposed clipping 
candidates based on intact mass 

can be quite extensive if somewhat 
relaxed mass tolerances, e.g., 10 ppm  
in this case, are applied to match 
sequence candidates (Figure 6). 
However, the applied calibration 

provided for a mass accuracy better 
than 2 ppm, which coincided with 
the only sequence identified by ETD 
analysis in peak 2 (Figure 7). LC 

[68-181] in Figure 5, amongst other 
candidates, was ruled out based on 
poor mass accuracy (Figure 6) and 
the absence of a match with the 
respective ETD spectrum.

Figure 6: Proposed clipping variants in peak 2 after application of a 10 ppm mass tolerance. The 2 
entries at 12698.3 Da. with <2 ppm errors are the best candidates and were further analyzed by ETD 
and CID.

Figure 7: ETD spectrum of the 12698.3 Da protein in peak 2 matched the sequence HC [338-450]. None of the other candidate sequences at 12698.3 Da 
were identified by Middle-Down sequencing.

In

te

nsi

ty

x103

250

750

1250

1750

2250

2750

3250

3750

500

1000

1500

2000

2500

3000

3500

m/z

0.50
0.25

0.75

1.00

1.25

0

K

K

A

G

c (8)

c (11)

c (17)

c (22)

z+2 (32)

z+1 (30)

z+1 (27)

z+1 (25)

z+1
z+2

c (15)

R

V

T

R

Q

Y

S

D

E

G

S

N

Q

L

y

K

H

Q

Q

L

T

K

N

V

Q

S

I

c

The HC [342-450] clipping product  
at 12269.029 Da in peak 3 was 

confirmed by ETD and CID, the CID 
spectrum is shown in Figure 8.

The Mr 24348.231 base peak in the 
MaxEnt spectrum of Peak 4 was 

confirmed by ETD as HC [238-450] 
with a 29% SVP (data not shown).

In addition, peaks were proposed as 
clipping candidates, which weren’t 
directly confirmed by Middle-Down 
sequencing but represented the 
N-terminal and C-terminal fragments  
of the same cleavage sites: 

HC [1-227]|HC [228-450] and 

HC [1-228]|HC [229-450], adjacent 
to the theoretical SpeB cleavage site 

HC [1-229]|HC [230-450] (Figure 10).  
All 4 peaks were matched with an 
average mass accuracy of 0.8 ppm 
and well matching isotope patterns.

Figure 8: CID spectrum of the 12269.029 Da protein in peak 3 matching the sequence HC [342-450]. None of the other candidate sequences at 12698.3 Da 
were identified by Middle-Down sequencing.

In

te

nsi

ty

m/z

1.0

500

1500

3000

4000

5000

6000

7000

8000

1000

2500

3500

4500

5500

6500

7500

2000

2.0

x104

0

Figure 9: Intact mass spectra from chromatographic peaks 5 (top) and 7 (bottom). The annotated peaks show cleavage behind positions HC 227 and HC 228. 
All peaks show a perfect overlap with the calculated isotope pattern (red) with a mass error between 0.4 and 1.5 ppm. The HC [228-450] was matched by 
the SNAP algorithm even against a significant background of chemical noise.

In

te

nsi

ty

x105

24,800

24,900

25,000

25,100

25,200

25,300

25,400

25,500

Da

Da

1.0

2.0

0

15.24

25243.572

HC [229-450]

25380.648

HC [228-450]

In

te

nsi

ty

x105

24,100

24,200

24,300

24,400

24,500

24,600

1.0

2.0

0

18.78

24337.957

[1-227] Gln->pyro-Glu

24475.019

HC [1-228] Gln->pyro-Glu

G

V Y

A V

a
b

G Q

Q V

V

Y

S

A

F

P

D

W

E

E

I

V

Y T

T

L

b (31)

L

L

C

P

S

b (11)

P

H

F

D

Y L

L

F F

D

P

E

E

Q G N

V

S

P

D

G

S

S

y (47)

y (52)

y (67)

H

Y

T

N

y

y (3)

y (18)

y (23)

S

Q

Summary

Three LC-MS runs of SpeB-digested 
and reduced Vedolizumab (LC-MS, 
LC-ETD, LC-CID) confirmed the 
expected cleavage site in the hinge 
region. In addition, the combined 
approach to qualify candidates by 
intact mass measurements for 

Middle-Down sequencing yielded 
6 more, minor, cleavage sites, 3 of 
which were confirmed by direct ETD 
and CID analysis (Figures 7, 8), 2 were 

validated by their complementary 
nature and mass accuracy (Figure 9,)  
and one by mass accuracy and 

isotopic fidelity alone, in the absence of 
alternative possible matches (Figure 
caption 5). The results are summarized 
in Figure 10. In case of Vedolizumab, 
ragged ends of the Fc/2 and the Fd 
were obtained in the hinge region;  
and an overall of 6 unspecific 

cleavage sites were observed with 
the implemented new workflows in 
BioPharma Compass.

The identity and amount of the 
observed cleavage products 
were comparable between the 

Vedolizumab originator and bio-
similar drug substances (data not 
shown), which indicates that clipping  
analysis might become a useful 

protein-based MAM approach 
in comparability, stability and aging 
studies.

Conclusions

•  A workflow was described to identify antibody 

clipping variants based on subunit mass 
measurements and Middle-Down sequencing.

•  A mass accuracy of well below 2 ppm was 

important to provide sufficient specificity to define 
candidate sequences. This was achieved with 
monoisotopic mass assignments using the SNAP 
algorithm. 

•  ETD and CID on the maXis II ETD UHR-QTOF 

were suitable to verify all clipping variants 
subjected to MS/MS analysis, though ETD typically 
provided 20-35% more sequence information.

•  BioPharma Compass 2021 Clipping workflows, 

both for the intact mass-based prediction as well 
as for the Middle-Down sequencing facilitated the 
prediction and confirmation of clipping variants 
significantly to speed up potential drug substance 
clipping analysis.

Figure 10: Vedolizumab HC sequence from [1] with the confirmed main cleavage site (green), CID/ETD-confirmed sites (orange) and sites where the  
identification was based on accurate mass and isotopic pattern matching (red).

B

ru

ke

r D

al

to

ni

cs i

s c

on

tin

ua

ly i

m

pr

ov

in

g i

ts p

ro

du

ct

s a

nd r

es

er

ve

s t

he r

ig

ht 

to

 c

ha

ng

e sp

ec

ifi

ca

tion

s w

ith

out

 n

ot

ic

e.

 © B

ru

ke

r D

al

to

ni

cs 1

1-

20

20

, L

C

M

S

-1

76

, R

ev

. 0

1, 1

8

8

38

59

For Research Use Only. Not for Use in Clinical Diagnostic Procedures.

[email protected] – www.bruker.com

Learn More

You are looking for further Information? 
Check out the link or scan the QR code.

www.bruker.com/biopharma-compass

Bruker Daltonics GmbH & Co. KG

Bremen · Germany
Phone +49 (0)421-2205-0

Bruker Scientific LLC

Billerica, MA · USA 
Phone +1 (978) 663-3660

References

[1]  www.drugbank.ca/drugs/DB09033

[2]  Resemann A, et al. (2016). Full validation of therapeutic antibody sequences by middle-up mass 

measurements and middle-down protein sequencing. MAbs. 8(2):318-30. www.ncbi.nlm.nih.gov/

pmc/articles/PMC4966597/