1.
1
Some Unusual Aggregation Phenomena in Recombinant Proteins
John S. Philo, Alliance Protein Laboratories
ABSTRACT: Biopharmaceuticals exhibit a wide variety of aggregate sizes and types. Here we
will describe some interesting and unusual phenomena found in recombinant proteins from 3
different clients.
Protein X is a very sticky glycoprotein that is difficult to measure by SEC and which is subject to
freeze/thaw damage. Sedimentation velocity (SV) on freeze/thaw stressed samples showed high
levels of non-covalent aggregates that could not be detected by the standard SEC method. After
developing an improved SEC method that correlates well with SV data we discovered that
freeze/thaw creates metastable aggregates that will slowly dissociate to monomer over hours to
days, depending on the temperature.
For Protein Y we investigated the appearance of aggregates over time after the lyophilized
formulation was reconstituted. SV analysis did show formation of ~0.5% aggregate over 8 hours.
More interesting however is that the reconstituted sample initially contained several percent of a
partially-unfolded monomer, and this species decreased over time as this species slowly refolded.
Aggregation then was arising not from the native state, but from this partially-unfolded state, in a
reaction competing with refolding to the native structure.
Protein Z is a test antigen for vaccine development that exhibits high levels of an early-eluting
putative dimer peak in stressed samples. However when we used SEC-MALLS to confirm the
dimer assignment we discovered this species is actually an expanded, partially-unfolded monomer
rather than an aggregate. This interpretation was also confirmed by SV analysis.
© copyright 2006 Alliance Protein Laboratories

2.
2
Protein X:
metastable non-covalent aggregates

3.
3
0
1
2
3
4
0.00
0.02
0.04
0.06
0.08
normalizedc(s)
13.2%
sedimentation coefficient (Svedbergs)
80.1%
0.13%
0.30%
0.33%
0.56%
1.38%
1.26%
0.85%
1.85%
X 50
Sedimentation velocity reveals high levels of aggregates in a
Protein X drug substance sample stressed by 4 freeze/thaw cycles
19.9% total aggregate is measured by SV, after 4X freeze/thaw

4.
4
The same sample measured by SV in the standard SEC
elution buffer for Protein X shows only 4% aggregate
(because the buffer dissociates non-covalent aggregates)
0
5
10
0.0
0.1
0.2
96.0%
normalizedc(s)
sedimentation coefficient (Svedbergs)
2.8%
0.36%
0.82%
X 50

5.
5
A new SEC method gives good correlation with SV results
A sample highly stressed by freeze/thaw cycles gives 63.4%
monomer by SEC, 63.2% by SV; 12.2% dimer by SEC, 11.4% by SV
(note that dimer is much better resolved by SV, so the ‘dimer’ peak
in SEC is probably contaminated by trimer)
10 15 20 25
Retention Time (minutes)
Absorbance(215nm)
24.4%large
aggregates
12.2%dimer
63.4% monomer
0
3
6
0.00
0.06
0.12
13.8%
63.2%
normalizedc(s)
sedimentationcoefficient (Svedbergs)
0.62%
0.74%
4.73%
1.51%
1.47%
4.36%
0.56%
0.66%
0.92%
1.30%
3.43%
0.60%
0.96%
1.14%
X50

6.
6
10 15 20 25
Retention Time (minutes)
Absorbance(215nm)
1.8%largeaggregates
11.6%dimer
86.6% monomer
A freshly-thawed* sample of bulk drug substance shows similar
levels of dimer but much lower levels of larger aggregates
This sample is 86.6% monomer by SEC, 83.2% by SV;
11.6% dimer by SEC, 14.1% by SV
*Note however that the SV measurement requires about 6 hr.
0
3
6
0.00
0.06
0.12
14.1%
83.2%
normalizedc(s)
sedimentation coefficient (Svedbergs)
0.05%
0.02%
0.27%
0.22%
0.02%
0.27%
0.08%
0.32%
0.20%
0.07%
1.18%
X 50

7.
7
10 15 20 25
Retention Time (minutes)
Absorbance(215nm)
1.6%largeaggregates
1.6%dimer 96.8% monomer
After 15 hr at 29 °C, the dimer content of a freshly-thawed
sample drops from 13% down to 1.5%
There is a corresponding increase in monomer, and also a
small decrease in the larger oligomers
0
3
6
0.00
0.06
0.12
96.5%
normalizedc(s)
sedimentationcoefficient (Svedbergs)
0.18%
0.02%
0.29%
0.26%
0.21%
0.30%
1.38%
0.15%
0.11%
0.14%
0.14%
0.30%
X50

8.
8
The rate of dissociation of dimer to monomer is strongly
dependent on temperature
10 15 20 25
Retention Time (minutes)
Absorbance(215nm)
— t = 0
— 1 day at 4°C
— 1 day at 29°C
— 1 day at 40°C

9.
9
Take-home lessons from Protein X
1. Sedimentation velocity can serve as a “gold
standard” to help develop better SEC methods
• the goal is high correlation between SEC and
SV, not perfect quantitative agreement
2. Metastable aggregates are not uncommon!
• it may take hours to days for re-equilibration
after any change in concentration, solvent, or
temperature

10.
10
Protein Y:
after reconstitution the lyophilized product
contains transient partially-unfolded
monomers

11.
11
Sedimentation velocity analysis of lyophilized Protein Y
over time after re-hydration
1 10 100
0.00
0.06
0.12
0 hr
2 hr
8 hr
s×c(s)
sedimentation coefficient (Svedbergs)
0.02%
0.07%
0.38%
0.007%
X200
1. new large aggregates appear and increase
2. slowly-sedimenting partially-unfolded monomer decreases
3. native monomer increases

12.
12
Conclusions about Protein Y
1. The reconstituted protein initially contained several
percent of a partially unfolded monomer, which could
slowly refold to the native state on a time scale of a few
hours
2. The ~0.5% total aggregate that formed over 8 hours was
arising from association of the partially unfolded
monomer
• the aggregation reaction competes with the re-folding
reaction

13.
13
Protein Z:
An “aggregate” that isn’t an aggregate

14.
14
volume (ml)
5.0 6.0 7.0 8.0 9.0 10.0
relativescale
0.0
0.2
0.4
0.6
0.8
1.0
monomer
dimer?
large
aggregates
This highly stressed sample of a VaxGen test antigen showed
high levels of an SEC peak eluting near the position
expected for a dimer

15.
15
However SEC-MALLS immediately shows that alleged
aggregate is actually an altered form of monomer!
5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0
Elution Volume (ml)
0.0
Signal(volts)
LS 90°(scaled) RI
ratio of LS to RI peak
heights is the same,
therefore MW is same

16.
16
molar mass vs. volume
volume
5.0 6.0 7.0 8.0 9.0 10.0
molarmass(g/mol)
5
1.0x10
6
1.0x10
7
1.0x10
Here is the actual molecular mass chromatogram calculated from
the light scattering data.
The apparent mass of the alleged aggregate is somewhat higher than the monomer mass
due to co-elution of sticky large aggregates, but clearly this is a monomer.

17.
17
Sedimentation velocity confirms formation of an
expanded monomer that sediments slowly
0 1 2 3 4 5 6 7 8
0
1
2
3.6%,4.04S
30.7%
4.60 S
3.6%,2.43S
45.2%
3.62 S
normalizedc(s)
sedimentation coefficient (Svedbergs)
native
monomer
expanded
monomer

18.
18
Acknowledgements
• Byeong Chang, Integrity Biosolution
• Amer Jaber, Serono
• VaxGen