Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 2 November 2018
1
Article
2
IgG Charge
3
Danlin Yang1, Rachel Kroe-Barrett1, Sanjaya Singh2, and Thomas Laue3,*
4
5
6
7
8
9
10
doi:10.20944/preprints201811.0052.v1
Biotherapeutics Discovery Research, Boehringer Ingelheim Pharmaceuticals, Inc., Ridgefield, Connecticut
06877, USA. Present address: Janssen BioTherapeutics, Janssen Research & Development, LLC, Spring
House, Pennsylvania 19477, USA
2
Janssen BioTherapeutics, Janssen Research & Development, LLC, Spring House, Pennsylvania 19477, USA
3
Department of Molecular, Cellular and Biomedical Sciences, University of New Hampshire, Durham,
New Hampshire, 03861, USA; tom.laue@unh.edu
* Correspondence: tom.laue@unh.edu; Tel.: +01-603-978-5579
1
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Abstract: It has been known since the 1930’s that all immunoglobulins carry a weak negative
charge in physiological solvents. However, there has been no systematic exploration of this
fundamental property. Accurate charge measurements have been made using membrane
confined electrophoresis in two solvents (pH 5.0 and pH 7.4) on a panel of twelve mAb IgGs, as
well as their F(ab’)2 and Fc fragments. The following observations were made at pH 5.0: 1) the
measured charge differs from the calculated charge by ~40 for the intact IgGs, and by ~20 for the
Fcs; 2) the intact IgG charge depends on both Fv and Fc sequences, but does not equal the sum of
the F(ab)’2 and Fc charge; 3) the Fc charge is consistent within a class. In phosphate buffered
saline, pH 7.4: 1) the intact IgG charges ranged from 0 to -13; 2) the F(ab’)2 fragments are nearly
neutral for IgG1s and IgG2s, and about -5 for some of the IgG4s; 3) all Fc fragments are weakly
anionic, with IgG1 < IgG2 < IgG4; 4) the charge on the intact IgGs does not equal the sum of the
F(ab’)2 and Fc charge. In no case is the calculated charge, based on H+ binding, remotely close to
the measured charge. The charge on IgGs in physiological solvent is sufficiently small to
minimize its contribution to thermodynamic nonideality. Some of the mAbs carried a charge in
physiological salt that was outside the range observed for serum-purified human poly IgG. To
best match physiological properties, a therapeutic mAb should have a measured charge that falls
within the range observed for serum-derived human IgGs.
29
30
Keywords: Analytical electrophoresis; IgG subclasses; monoclonal IgG, Protein charge
31
1. Introduction
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
It has been known for over 80 years that all serum proteins, including the immunoglobulins,
carry a net negative charge under physiological conditions [1]. More recently, it was shown that
freshly prepared human polyclonal IgGs have a Debye-Hückel-Henry charge, ZDHH [2], between -3
and -9 [3]. The narrow range of charge is somewhat surprising since isoelectric focusing analysis of
the same sample yielded isoelectric points (pIs) covering the pH range from less than 4 to greater
than 10 [3]. Charge is a system property that depends on temperature and solvent composition, and
it is believed that the narrow range of ZDHH under physiological conditions is a consequence of anion
binding.
It is known that charge and charge distribution are important contributors to protein solubility
and solution viscosity [4–7]. The majority of biotherapeutic mAbs exhibit pIs >= 8, and carry a positive
charge in the pH 5 – 6 range where they are formulated [5–7]. However, there is no published charge
data for these mAbs in physiological solvents, and it is not known whether their charge falls into the
range observed for normal human IgGs. It is apparent that a systematic analysis of the charge on
mAbs would be useful.
Presented here is an analysis of the charge on twelve anti IL-13 IgGs. Using membrane confined
electrophoresis, MCE, charge data have been acquired for three IgGs, mAb1, mAb2 and mAb3, that
© 2018 by the author(s). Distributed under a Creative Commons CC BY license.
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 2 November 2018
doi:10.20944/preprints201811.0052.v1
2 of 14
48
49
50
51
52
53
54
bind to different IL-13 epitopes [3]. For each mAb, ZDHH has been measured for four subclasses, IgG1,
IgG2, IgG4 and IgG4Pro. Furthermore, the charge on the Fc and F(ab’)2 fragments was measured to
determine whether the intact IgG charge is the sum of the Fc and F(ab’)2 fragment charges, and to
assess how the charge is distributed over the IgG structure. Finally, the charge on the IgGs and their
fragments were measured at both pH 5.0 and pH 7.4 to determine how the charge varies between
formulation and physiological conditions. The results illustrate how little is known about protein
charge and demonstrates the power of analytical electrophoresis in assessing this property.
55
1.1 Background
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
Protein charge is significant to a variety of biochemical, biophysical and biological phenomena
[8]. Thermodynamically, charge is a system property that depends on temperature, pressure, salt
concentration, salt type and pH [9]. At present there is no way to calculate protein charge accurately.
However, charge may be measured with both precision and accuracy [2,10,11]. Of the measurement
methods, membrane confined electrophoresis [12,13] is the most accurate and flexible [2,14].
There are a variety of charge descriptions (e.g. ζ potential, Zeffective, ZDHH) [2]. While each
description is useful, here we will use ZDHH, which is the unitless valence resulting from the ratio of
the protein charge (in coulombs) to the proton unit charge (e.g. Ca2+ has a valence of +2, Cl-1 has a
valence of -1). Calculation of ZDHH from the free-boundary electrophoretic mobility removes the
effects of electrophoresis and the solvent ion cloud [2,12,15]. Thus, ZDHH reflects any changes in
protein charge that accompany changes in solvent pH, salt type or salt concentration [2].
Though pH may contribute to protein charge, ZDHH reflects binding by all solvent ions (e.g. Na+,
2PO4 , Cl-) and not just H+. It has been known for over 60 years that proteins bind anions to a greater
extent than cations [16–18]. Two non-exclusive models have emerged for the mechanism of anion
binding. One model focuses on the tendency for anions to accumulate preferentially at hydrophobic
surfaces [17]. Based on NMR data, the other model suggests that anion binding may involve amide
protons [18].
Because ion binding and dissociation occur rapidly, ZDHH values are time averages. The extent
of fluctuation about the mean value is proportional to the change in charge with ion chemical
potential (i.e. the slope of the curve of Z versus log[X]) [19]. If the titration curve is flat (i.e. dZ/dlog[X]
~0), there will be very little charge variation, and the charge distribution about the average value will
be narrow. A steep titration curve, however, indicates large charge variations which, particularly if
they swing around neutrality, result in the inter-molecular attractions that reduce solubility and
cause higher viscosities. Thus, measurement of ZDHH as a function of solvent ion concentration
(including pH) may be helpful in finding solvent conditions that optimize solubility and viscosity.
81
2. Materials and Methods
82
2.1. Monoclonal and human serum IgGs
83
84
85
86
87
88
89
90
91
92
93
94
95
Twelve anti-IL13 IgGs comprising three unique variable regions, each constructed as four
human IgG subclasses, IgG1, IgG2, wild-type IgG4(Ser222), and a hinge mutant IgG4(Pro222), were
made from stable NS0 cell clone at Boehringer Ingelheim. Human serum derived from male AB
plasma was purchased from Sigma (cat# H4522). The IgGs were purified by ÄKTA affinity
chromatography system and MabSelect Sure resin (GE Healthcare) following standard methods
[20]. The quality of the purified mAb IgGs and their fragments generated by subsequent enzymatic
digestion was evaluated by analytical size-exclusion ultra-performance liquid chromatography (SEUPLC) using a BEH200 column on the Waters Acquity UPLC system (Waters Corporation). The
mobile phase buffer consisted of 50 mM sodium phosphate (pH 6.8), 200 mM arginine, and 0.05%
sodium azide. For each sample run, 10 µg of material was injected onto the column with the
running flow rate at 0.5 mL/min for 5 min.
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 2 November 2018
doi:10.20944/preprints201811.0052.v1
3 of 14
96
2.2. IgG fragmentation
97
98
99
100
101
102
103
104
105
A FragIT kit with individual spin columns containing the active IdeS, a cycstein protease
secreted by Streptococcus pyogenes covalently coupled to agarose beads was used (Genovis, cat# A2FR2-025). After the IgG sample was buffer exchanged into the cleavage buffer (10 mM sodium
phosphate, 150 mM NaCl) and the column was equilibrated with the cleavage buffer, the IgGenzyme mixture was incubated at 37 oC for an hour on an orbital shaker. The digested fragments
were separated from the immobilized enzyme, followed by the purification of F(ab’)2 using a
supplied CaptureSelect column containing Fc affinity matrix (Thermo Fisher). Upon the collection
of the F(ab’)2 in the flow-through, the Fc was eluted using the 0.1 M glycine (pH 3.0) elution buffer
and immediately neutralized by adding 10% v/v of 1 M Tris (pH 8.0).
106
2.3. Sample preparation
107
108
109
110
111
112
113
114
115
Each sample was dialyzed into desired buffers at 4-10 oC overnight using Zeba desalting
columns (Thermo Fischer), after which the concentration was determined using appropriate
extinction coefficients in NanoDrop™ 8000 Spectrophotometer (Thermo Fischer). Two solvents
were used: 10 mM sodium acetate, 50 mM NaCl, pH 5.0; and Dulbecco's PBS (pH 7.4) containing 8
mM sodium phosphate dibasic, 1.5 mM potassium phosphate monobasic, 2.7 mM KCl, and 138 mM
NaCl. The acetate buffer was prepared by diluting chemicals purchased from Sigma into distilled
deionized water from a Milli-Q Plus filtration system (Millipore) and titrating to the desired pH 5.0
with 10 N NaOH solution. For all measurements, the sample solutions were used within a week of
preparation and stored at 4 oC between measurements.
116
2.4 Liquid Chromatography Mass Spectrometry (LC-MS)
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
The sequences of the purified mAbs and respective F(ab’)2 and Fc fragments were evaluated by
LC-MS using a PoroShell 300SB-C8 column (5 µm, 75 x1.0 mm) on the Agilent HPLC system
followed by analysis in the Agilent 6210 time-of-flight mass spectrometer (Agilent Technologies).
The composition of the mobile phase A was 99% water, 1% acetonitrile, and 0.1% formic acid, and
that of mobile phase B was 95% acetonitrile, 5% water, and 0.1% formic acid. The gradient started
with 20% B at 0 min and increased to 85% B at 10 min with the constant flow rate of 50 µl/min. Each
sample was subjected to a native run, a reduced run after incubation with TCEP (Sigma), and a
deglycosylated run after incubation with TCEP and PNGase F (New England Biolabs). The
MassHunter Qualitative Analysis program (version B.06.00) was used to deconvolute the raw data.
2.5 Analytical Ultracentrifugation (AUC)
The solution properties of the purified mAbs and cleaved F(ab’)2 and Fc were evaluated by
sedimentation velocity experiments in an Optima XL-I AUC equipped with absorbance optics
(Beckman Coulter). Each sample was prepared in three concentrations with 1:3 serial dilutions
starting from 0.5 mg/mL in the correponding buffer, and 400 µl of the prepared solution was loaded
into the sample chamber, whereas buffer was loaded into the reference chamber of an AUC cell
assembled with standard double-sector centerpieces and quartz windows. The experiments were
conducted at 20 °C using an An60Ti 4-hole rotor spinning at 40,000 rpm. The sedimentation process
was monitored by collecting absorbance data at 280 nm wavelength and 30-µm radial increments.
The collected data was analyzed using the SEDANAL software by which the apparent
sedimentation coefficient distribution g(s*) was derived [21]. The resulting analysis was initially
plotted as g(s*) vs. s* in which the areas under the peaks provided the concentration for the
boundary corresponding to each peak in the distribution. The weight average sedimentation
coefficient (sw) was computed by selecting a range over which to do the average on the plots. The
plots were concentration-normalized to enable the inspection for reversible interactions. The Stokes
radius, Rs, which is used for ZDHH calculation is derived from the Svedberg equation:
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 2 November 2018
doi:10.20944/preprints201811.0052.v1
4 of 14
=
(1 − ῡ )
6
(1)
144
145
where M is the molar mass, ῡ is the partial specific volume, ρ is the solvent density, s is the
sedimentation coefficient, NA is the Avogadro’s number, and η is the viscosity of the solvent.
146
2.6. Imaged capillary isoelectric focusing (icIEF)
147
148
149
150
151
152
153
154
155
156
157
158
The pI and charge heterogeneity of the IgG samples were determined on an iCE3 system
(Protein Simple) [22,23]. Briefly, the pH gradient was created by an ampholyte mixture consisted of
44% (v/v) of 1% methylcellulose, 1.25% (v/v) of pharmalyte 3-10 solution, 3.75% (v/v) µl of
pharmalyte 5-8 solution, 1.25% (v/v) of servalyte 9-11 solution, 0.63% (v/v) of pI marker pH 6.14,
0.63% (v/v) of pI marker pH 8.79, 6.3% (v/v) of 200 mM iminodiacetic acid, and 43% (v/v) of water.
After sample preparation at 1 mg/mL in DI water, 40 µl of the diluted sample was mixed with 160
µl of ampholyte mixture and centrifuged for 5 min. The operating protocol used an initial potential
of 1500 volts for 1 min, followed by a potential of 3000 volts for 20 min. For samples containing
highly basic species, pI markers at pH 7.55 and pH 9.77 (0.63% v/v) and a focus period of 10 min at
3000 volts was used. Separation was monitored at 280 nm, and the data analyzed using the iCE CFR
software to calibrate the pI values and to select the markers. Subsequetly, the data files were
exported to Empower for analysis using the cIEF processing method.
159
2.7. Membrane-confined electrophoresis (MCE) and ZDHH determinations
160
161
162
163
164
165
166
167
168
169
170
Protein valence was measured in the MCE instrument (Spin Analytical), which provides a
direct measurement of the electrophoretic mobility (µ) to derive the Zeff and the ZDHH [12,13]. In each
experiment, 20 µl of sample at 1 mg/mL was loaded into a 2 x 2 x 4 mm quartz cuvette whose ends
were sealed with semipermeable membranes (MWCO 3 kDa, Spectra/Por Biotech grade). An
electric field was applied (4.3 V/cm for IgG, 8.5 V/cm for F(ab’)2 and Fc, and 19.8 V/cm for serum
IgGs) longitudinally across the cell. The applied electric field, E, is a function of the applied current,
i, the buffer conductivity (κ, 5.8 mS for 10 mM acetate, 50 mM NaCl [pH 5.0] and 16.8 mS for PBS
[pH 7.4]), and the cross-sectional area of the cuvette, A, as = . Image scans of the cuvette were
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
acquired with 25 μm resolution at 280 nm every 10-20 seconds. Time difference analysis provided
an apparent electrophoretic mobility distribution, g(μ) versus μ, uncorrected for diffusion. Values of
μ were converted to charge using the Spin Analytical software:
=
=
(2)
1+
(
(3)
)
where µ is the electrophoretic mobility, f is the translational frictional coefficient, e is the elementary
proton charge , ĸD is the inverse Debye length, a is the sum of the Stokes radius of the
macromolecule and its counterion (0.18 nm for Cl-1 and 0.122 nm for Na+), and H(κD a) is Henry’s
function that accounts for electrophoretic effects. For reference, under the experimental conditions
used here, κD a ~ 2 and H(κD a) ~1.1, though exact values are calculated for each experiment.
2.8 Calculated charge, ZCal, and calculated isoelectric pH, pICal
Sednterp was used to calculate pI values, pICal, as well as the H+ titration curve from which ZCal
was determined [24]. These calculations are based on the amino acid composition and use pKa
values from Edsall and Wyman [25]. It was assumed that the N-terminal amino groups were not
blocked.
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 2 November 2018
doi:10.20944/preprints201811.0052.v1
5 of 14
186
2.9 Dynamic light scattering (DLS) and kD determinations
187
188
189
190
191
192
193
194
195
196
197
A DynaPro Plate Reader (Wyatt) running Dynamics (version 7.4.0.72) was used to determine
the diffusion interaction parameter, kD. Each sample was prepared at 5 concentrations ranging from
10 mg/mL to 0.625 mg/mL in 2-fold serial dilutions. 35 µl of each solution was added to a 384-well
UV-Star Clear Microplate (Greiner Bio-One), spun in a centrifuge for 2 mins to remove air bubbles
and then placed into the plate reader. The experiment was started after the temperature inside the
reader reached 20 oC. A total of 10 acquisitions at 20 s per acquisition were obtained for each
sample. A well image was acquired after the last acquisition measurement to look for bubbles or
deposited aggregates. The mutual diffusion coefficient (Dm) was plotted against the sample
concentration = 0(1+ ), with D0 and kD determined by linear regression analysis using
GraphPad Prism (version 7.03). The error for kD was determined by calculating the propagation of
the standard error of the coefficients from the linear regression.
198
3. Results
199
200
201
202
203
204
205
206
All purified IgGs contain > 99% monomer content as assessed by analytical SE-UPLC and are
sequence confirmed by LC-MS. These twelve mAbs also displayed homogeneous solution
properties within each mAb group in both pH 5.0 acetate and pH 7.4 PBS buffer conditions as
illustrated in Figure 1. Overlaps between the IgG subclasses within each mAb group are observed,
in which the weight-average sedimentation coefficients (sw) are 6.37 ± 0.06, 6.37 ± 0.05, and 6.43 ±
0.09 in pH 5.0 acetate, and 6.28 ± 0.04, 6.27 ± 0.07, and 6.31 ± 0.06 in pH 7.4 PBS for mAb 1, mAb 2,
and mAb 3, respectively. These sw values are consistent with the molecular weight of ~150 kDa IgG
antibodies.
207
208
209
210
Figure 1. Sedimentation velocity analysis of IgG subclasses from mAb1, mAb2, and mAb3 in pH 5.0
acetate (red) and pH 7.4 PBS (blue) solutions. Normalized g(s*) sedimentation distributions are obtained from
IgG1 (solid line), IgG2 (dotted line), IgG4 (dashed line), and IgG4Pro (dot-dashed line) in both buffers. The
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 2 November 2018
doi:10.20944/preprints201811.0052.v1
6 of 14
211
212
purity assessed by SE-UPLC and cleavage sites between F(ab’)2 and Fc identified by LC-MS are summarized in
Table 1.
213
Table 1. Quality of IgG fragments from IdeS digestion.
Subclass
V region
Cleaved Site
IgG1
mAb 1
mAb 2
mAb 3
mAb 1
mAb 2
mAb 3
mAb 1
mAb 2
mAb 3
mAb 1
mAb 2
mAb 3
…CPPCPAPELLG
/ GPSVF…
IgG2
IgG4
IgG4Pro
214
215
216
217
218
219
220
221
222
223
…CPPCPAPPVA
/ GPSVF…
…CPSCPAPELLG
/ GPSVF…
…CPPCPAPELLG
/ GPSVF…
F(ab’)2 purity
(%)
95
100
100
100
100
100
95
95
97
97
100
100
Fc purity
(%)
100
98
97
The solution homogeneity of each cleaved fragment was assessed by SV-AUC. All IgG
fragments showed sedimentation distribution profiles like that in Figure 2 for mAb 1, where the
superposition of the three concentrations of F(ab’)2 and Fc samples indicate homogeneity and the
absence of self-association. The weight-average sedimentation coefficients (sw) from the Fc
evaluations are 3.45 ± 0.02, 3.46 ± 0.02, and 3.38 ± 0.18 for IgG1, IgG2, and IgG4, respectively. These
values are consistent with the molecular weight of ~50 kDa, which indicates Fc homodimer in
solution despite cleavage below the hinge region. The sw from the F(ab’)2 evaluations are 4.86 ± 0.01,
5.14 ± 0.06, 4.90 ± 0.02, and 4.95 ± 0.01 for IgG1, IgG2, IgG4, and IgG4Pro, respectively. These values
are consistent with the molecular weight of ~100 kDa, which is expected for a bivalent Fab linked by
hinge.
IgG1
IgG2
IgG4
224
(a)
(b)
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 2 November 2018
doi:10.20944/preprints201811.0052.v1
7 of 14
Figure 2. Sedimentation velocity analysis of IgG1, IgG2, and IgG4 cleaved (a) Fc and (b) F(ab’)2 from
mAb1 in pH 5.0 acetate. Normalized g(s*) sedimentation distributions obtained with the concentration of
0.5 mg/mL (red), 0.167 mg/mL (blue), and 0.056 mg/mL (green).
228
229
230
231
232
All IgGs exhibited pI profiles like that in Figure 3 for mAb1 IgG1. Three-peaks are observed,
acidic, main and basic. The pI values for each IgG are presented in Table 2, along with the calculated
pI. For each mAb, the subclass pIs followed the trend: IgG1 > IgG2 > IgG4, with those of IgG4 and
IgG4Pro being identical. The measured main species pI and the calculated pI are correlated (Figure
4), though the intercept (-1) suggests that pICal corresponds to the more acidic species.
0.28
Main pI
0.26
8.065
225
226
227
0.24
0.22
0.20
0.16
0.12
Low pI marker
High pI marker
0.10
Basic pI
Acidic pI
6.136
7.917
0.08
0.06
8.791
0.14
8.235
Absorbance
0.18
0.04
0.02
0.00
6.00
6.20
6.40
6.60
6.80
7.00
7.20
7.40
7.60
7.80
8.00
8.20
8.40
8.60
8.80
9.00
233
234
235
Figure 3. Electrophoretogram image of mAb1 IgG1. The peaks to the left and to the right of the main peak
indicates acidic and basic charge variant, respectively.
236
Table 2. Measured and calculated pI values of IgG.
ID
mAb1
mAb2
mAb3
237
Subclass
pIcal*
IgG1
IgG2
IgG4
IgG4Pro
IgG1
IgG2
IgG4
IgG4Pro
IgG1
IgG2
IgG4
IgG4Pro
7.7
6.9
6.6
6.6
8.2
7.3
7.0
7.0
8.2
7.4
7.1
7.1
Minutes
pIicIEF
Acidic peak
Main peak
Basic peak
7.9
6.9
6.2
6.2
8.2
7.9
7.4
7.4
8.2
7.2
7.5
7.5
8.1
7.0
6.3
6.3
8.4
8.0
7.6
7.6
8.4
8.0
7.7
7.7
8.2
7.3
6.5
6.5
8.6
8.2
7.7
7.7
8.6
8.1
7.8
7.8
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 2 November 2018
doi:10.20944/preprints201811.0052.v1
8 of 14
238
239
240
241
242
243
244
245
Figure 4. Linear regression analysis and correlation between experimental pI as measured by icIEF and
theoretical pI calculated from the IgG sequence. Dotted lines indicate the 95% confidence interval.
Using MCE, the electrophoretic mobility was determined for each IgG and its cleaved F(ab’)2
and Fc in pH 5.0 acetate and pH 7.4 PBS as illustrated in Figure 5. By applying the Debye-Hückel
approximation to correct for the solvent shielding effects, Henry’s function to correct for
electrophoretic effects, and using the sum of the measured protein Stokes radius and its counterion,
the ZDHH distribution may be calculated from the electrophoretic mobility (Figure 5, right-hand
panels).
IgG
F(ab’)2
Fc
246
247
248
249
250
251
252
253
254
255
256
257
258
(a)
(b)
Figure 5. ZDHH determination of IgG, F(ab’)2, and Fc by MCE in pH 5.0 acetate. (a) Raw MCE scans
over time during electrophoresis. The data (left panel) shows the light intensity (I, vertical axis) as a
function of the distance moved from the membrane (cm, horizontal). Time difference curves (ΔI/Δt)
are calculated from data between the green and red highlighted scans. The electrophoretic mobility
distribution is calculated from distance moved from the membrane, x, divided by the product of the
electric field, E, and average elapsed time for the middle scan ̅, = ̅. (b) The vertical axis shows
∙
the time derivative (ΔI/Δt) of the intensity data in panel (a) as a function of ZDHH (horizontal axis).
ZDHH was calculated from the mobility using T = 20 oC; viscosity = 0.98 cp; conductance = 16.8 mS;
E = -19.8 V/cm, D = 78; counterion radius, 0.18 nm; Stokes radius, 5.5 nm. The peak ZDHH position is
displayed above the curve.
Table 3 and Table 4 summarize the ZDHH measurements, as well as the calculated charge, ZCal,
in pH 5.0 acetate and pH 7.4 PBS, respectively. A 0 charge was assigned if no boundary formed
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 2 November 2018
doi:10.20944/preprints201811.0052.v1
9 of 14
259
260
261
262
263
during electrophoresis regardless of the E field direction or magnitude. In acetate pH 5.0 all IgGs
and their fragments are cationic (Table 3). However, in all cases the measured ZDHH is substantially
lower than Zcal. In PBS pH 7.4 (Table 4), all intact IgGs are neutral (Mab2/IgG1) or anionic, despite
the fact the Zcal is cationic in some cases. For all mAbs, ZDHH decreases with subclass in the rank
order of IgG1 > IgG2 > IgG4.
264
Table 3. Measured and calculated Z values of IgG, F(ab’)2, and Fc in pH 5.0 acetate.
ID
mAb
1
mAb
2
mAb
3
265
266
267
F(ab')2
Fc
ZDHH
Zcal
ZDHH
Zcal
ZDHH
Zcal
IgG1
7.7 ± 0.2
57.3
3.3 ± 0.2
31.2
6.2 ± 0.1a
26.30
IgG2
3.9 ± 0.1
50.0
0
25.9
4.9 ± 0.1b
24.30
IgG4
IgG4Pro
IgG1
IgG2
IgG4
IgG4Pro
IgG1
IgG2
IgG4
IgG4Pro
1.4 ± 0.2
1.4 ± 0.8
10.6 ± 0.1
10.1 ± 0.2
5.6 ± 0.2
5.6 ± 0.2
12.5 ± 0.1
10.3 ± 0.2
7.7 ± 0.2
7.8 ± 0.2
46.7
46.7
61.0
53.7
50.4
50.4
65.8
58.5
55.1
55.1
1.3 ± 0.1
1.5 ± 0.2
8.6 ± 0.2
4.7 ± 0.1
6.2 ± 0.1
6.2 ± 0.1
9.4 ± 0.1
5.3 ± 0.2
7.1 ± 0.1
7.3 ± 0.1
27.9
27.9
34.9
29.6
31.6
31.6
39.6
34.3
36.3
36.3
0.45 ± 0.1c
18.98
c
a
b
c
c
a
b
c
c
pooled IgG1-Fc dialyzed into acetate from mAb1, mAb2, and mAb3 digestions
b
c
pooled IgG2-Fc dialyzed into acetate from mAb1, mAb2, and mAb3 digestions
pooled IgG4-Fc dialyzed into acetate from mAb1, mAb2, and mAb3 digestions
Table 4. Measured and calculated Z values of IgG, F(ab’)2, and Fc in pH 7.4 PBS.
ID
mAb 1
mAb 2
mAb 3
272
273
274
275
276
IgG
a
268
269
270
271
Subclass
F(ab')2
IgG
Subclass
IgG1
IgG2
IgG4
IgG4Pro
IgG1
IgG2
IgG4
IgG4Pro
IgG1
IgG2
IgG4
IgG4Pro
Fc
ZDHH
Zcal
ZDHH
Zcal
ZDHH
Zcal
-5.6 ± 0.1
-7.7 ± 0.6
-10.6 ± 0.5
-13 ± 0.3
0
-3.2 ± 0.2
-7.4 ± 0.2
-9.6 ± 0.4
-5.3 ± 0.5
-6.1 ± 0.3
-6.1 ± 0.2
-10.7 ± 0.4
1.8
-4.4
-6.5
-6.5
5.8
-0.4
-2.5
-2.5
6.0
-0.1
-2.2
-2.2
0
0
-4.3 ± 0.8
-5.05 ± 0.5
0
0
0
0
0
0
0
0
-0.48
-4.59
-2.61
-2.61
3.5
-0.61
1.38
1.38
3.45
-0.36
1.63
1.63
-2.8 ± 0.1
-6.0 ± 0.6e
-10.4 ± 0.3f
d
1.50
-0.48
-4.60
f
d
e
f
f
d
e
f
f
pooled IgG1-Fc dialyzed into PBS from mAb1, mAb2, and mAb3 digestions
d
pooled IgG2-Fc dialyzed into PBS from mAb1, mAb2, and mAb3 digestions
e
f
pooled IgG4-Fc dialyzed into PBS from mAb1, mAb2, and mAb3 digestions
While ZDHH and Zcal are correlated in either solvent (Figure 6), the slope is about ½ - ¾ of what
would be expected if there were a 1:1 correspondence between the expected H+ uptake/release and
ZDHH. These data are consistent with a model in which an anion is bound for every 1.3 – 2 H+ bound.
Similarly, ZDHH for the intact IgGs correlates with the sum of ZDHH from fragments (Figure 7), albeit
with a slope that is about ½ of that expected if the charge on the fragments simply summed. We
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 2 November 2018
doi:10.20944/preprints201811.0052.v1
10 of 14
Measured ZDHH
have no mechanism or explanation for the data in Figure 7 and present them here in the hope that
they will encourage future work.
Measured ZDHH
277
278
(a)
279
280
281
282
(b)
Figure 6. Linear regression analysis and correlation between experimental ZDHH measured by MCE and
theoretical Z calculated from the IgG sequence. (a) pH 5.0 acetate. (b) pH 7.4 PBS. Dotted lines indicate
the 95% confidence interval.
284
285
286
287
288
289
290
291
292
293
294
295
296
IgG ZDHH
IgG ZDHH
283
(a)
(b)
Figure 7. Linear regression analysis and correlation between ZDHH measured from intact IgG and the sum
of ZDHH from the fragments. (a) pH 5.0 acetate. (b) pH 7.4 PBS. Dotted lines indicate the 95% confidence
interval.
4. Discussion
Protein charge directly influences the structure, stability, solubility, and ability to interact with
other macromolecules [26]. Charge-charge repulsion is important for overcoming the attractive
forces that lead to high viscosities in high-concentration protein solutions [27]. Because protein
charge can vary with solvent conditions, it is a system property rather than a property of the protein.
The systematic analysis of twelve mAbs and their F(ab’)2 and Fc fragments provides several insights
into IgG charge and raises several important questions about our understanding of protein charge.
Charge-charge repulsion contributes to thermodynamic nonideality and, consequently, the
colloidal stability of protein solutions [9]. It is clear from the data in Tables 3 and 4 that charge
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 2 November 2018
doi:10.20944/preprints201811.0052.v1
11 of 14
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
calculations based solely on H+ binding lead to highly inaccurate estimates of IgG charge. Thus,
even though there is a correlation between the measured and calculated charge (Figure 6), charge
calculations should not be considered reliable. Given its potential importance to colloidal stability,
it is important to determine the impact of charge on nonideality.
At low to moderate protein concentrations (< ~15 mg/mL), the net sum of all repulsive and
attractive interactions is described by the second virial coefficient, B22 or A2. The diffusion
interaction parameter, kD, is related to and often used as a stand-in for these quantities [28], with
more positive values of kD correlating with more positive values of B22, i.e. greater repulsive
interactions. If charge-charge repulsion contributes significantly to nonideality, there should be a
positive correlation of charge with kD. Figure 8 shows the correlation of ZDHH with the diffusion
interaction parameter, kD. Under formulation conditions (Figure 6, panel a) increasing ZDHH
correlates with increased repulsive interaction (i.e. kD becomes more positive). This suggests that
charge measurements may be a useful parameter for selecting candidate mAbs for development.
It should be noted that it is the effective charge, Zeff, rather than ZDHH, that impacts thermodynamic
nonideality [2]. This distinction is important because Zeff includes the contribution of the solvent
ions, with Zeff decreasing (i.e. repulsive interactions decreasing) as salt concentration is increased
[9]. Because salt diminishes charge-charge interactions, thus reducing colloidal stability, it should
be no surprise that most mAbs are manufactured and formulated in low-salt solvents.
While charge does contribute to nonideality under formulation conditions, there is no
correlation between ZDHH and kD under physiological conditions (Figure 8, panel b). This result
means that it is unfavorable solvent displacement energies that keep mAbs in solution, for all other
protein-protein interactions are attractive [29]. Similarly, it is likely that it is the protein solvation
shell that dominates the solubility of serum IgG.
320
321
322
323
Figure 8. Linear regression analysis and correlation between ZDHH measured for intact IgG and the
concentration-dependence of the diffusion coefficient, kD. (b) pH 7.4 PBS. Dotted lines indicate the 95%
confidence interval.
324
325
326
327
328
One surprising result of our work is that freshly prepared human IgG exhibits a rather narrow
ZDHH distribution in physiological solvent (from approximately -10 to -2, Figure 9), even though
isoelectric focusing shows that the same sample has species ranging from pI < 4 to pI > 10. [30] This
exact same ZDHH range may be calculated from electrophoretic mobility measurements published 80
years ago [1]. Figure 9 shows that most, but not all, of the mAbs in this study exhibit ZDHH that fall
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 2 November 2018
doi:10.20944/preprints201811.0052.v1
12 of 14
329
330
331
in the range for human serum poly IgG. It is not clear whether there are any physiological or
medical consequences associated with a mAb ZDHH that falls outside the normal physiological
range. Thus, these results are presented in the hopes of stimulating further research.
332
333
334
335
336
337
Figure 9. ZDHH distribution for freshly prepared human IgG in DPBS. ZDHH was calculated for T = 20 oC,
viscosity = 0.98 cp, electric field = -14.88 V/cm, ionic strength = 0.167 M, conductivity = 16.6 ms, protein
radius = 5.5 nm, counterion radius = 0.18 nm, D = 78. The ZDHH for the twelve intact IgGs in this study are noted
(inverted triangles) along with bars indicating the measurement uncertainty.
5. Conclusions
338
339
340
341
342
343
344
345
346
347
Protein charge contributes to producing the colloidally stable mAb solutions needed during
development, manufacture and formulation. At this time, protein charge cannot be calculated with
any accuracy by even the most detailed structural information using the most sophisticated
algorithms. Protein charge, however, is readily measured with accuracy and precision. In this first
systematic and comprehensive examination of the charge on IgGs it is clear that: 1) IgGs bind
significant quantities of anions, 2) anion binding will contribute to the desolvation energy, thus
preventing IgG aggregation, 3) mAb charge measurements may be useful in selecting candidate
molecules for development and 4) mAb charge measurements under physiological conditions may
be useful in determining whether a candidate molecule falls within the normal range for human
IgGs.
348
349
350
351
Author Contributions: Conceptualization, D.Y., T.L., S.S., and R.K-B.; Methodology, T.L and D.Y.; Analysis,
D.Y. T.L.; Investigation, D.Y.; Resources, R.K-B.; Writing-Original Draft Preparation, D.Y and T.L.; WritingReview & Editing, R.K-B. and S.S.; Supervision, R.K-B. and T.L.; Project Administration, R.K-B. and T.L.;
Funding Acquisition, R.K-B and S.S..
352
353
354
355
356
357
Funding: Boehringer-Ingelheim
Acknowledgments: The authors wish to thank Boehringer-Ingelheim for supporting the doctorate research of
Danlin Yang, a portion of which is published here. Special thanks to her Ph.D. committee members, David Hayes
and Christopher Roberts, who encouraged this work and offered helpful advice. We also are thankful for the
encouragement and interest expressed by the members of the Biomolecular Interactions Technology Center
(BITC). This paper is dedicated to the memory of Eric and Betty Laue.
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 2 November 2018
doi:10.20944/preprints201811.0052.v1
13 of 14
358
Conflicts of Interest: The authors declare no conflict of interest.
359
References
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
1.
Tiselius, A.; Kabat, E. A. An electrophoretic study of immune sera and purified antibody
preparations. J. Exp. Med. 1939, 69, 119–131.
2.
Filoti, D. I.; Shire, S. J.; Yadav, S.; Laue, T. M. Comparative study of analytical techniques for
determining protein charge. J. Pharm. Sci. 2015, 104, 2123–2131, doi:10.1002/jps.24454.
3.
Yang, D.; Correia, J. J.; Iii, W. F. S.; Roberts, C. J.; Singh, S.; Hayes, D.; Kroe-Barrett, R.; Nixon,
A.; Laue, T. M. Weak IgG self- and hetero-association characterized by fluorescence analytical
ultracentrifugation. Protein Sci. 2018, 27, 1334–1348, doi:10.1002/pro.3422.
4.
Cohn, E. J. Studies in the physical chemistry of the proteins: I. the solubility of certain proteins
at their isoelectric points. J. Gen. Physiol. 1922, 4, 697–722, doi:10.1085/jgp.4.6.697.
5.
Raut, A. S.; Kalonia, D. S. Opalescence in Monoclonal Antibody Solutions and Its Correlation
with Intermolecular Interactions in Dilute and Concentrated Solutions. J. Pharm. Sci. 2015, 104,
1263–1274, doi:10.1002/jps.24326.
6.
Shire, S. J.; Shahrokh, Z.; Liu, J. Challenges in the development of high protein concentration
formulations. J. Pharm. Sci. 2004, 93, 1390–1402, doi:10.1002/jps.20079.
7.
Li, L.; Kumar, S.; Buck, P. M.; Burns, C.; Lavoie, J.; Singh, S. K.; Warne, N. W.; Nichols, P.;
Luksha, N.; Boardman, D. Concentration dependent viscosity of monoclonal antibody
solutions: explaining experimental behavior in terms of molecular properties. Pharm. Res.
2014, 31, 3161–3178, doi:10.1007/s11095-014-1409-0.
8.
Mathews, C. K.; Holde, K. E. van; Appling, D. R.; Anthony-Cahill, S. J. Biochemistry; 4 edition.;
Pearson: Toronto, 2012; ISBN 978-0-13-800464-4.
9.
Tanford, C. Physical Chemistry of Macromolecules; First Edition edition.; John Wiley & Sons Inc:
New York, 1961; ISBN 978-0-471-84447-1.
10.
Moody, T. P.; Kingsbury, J. S.; Durant, J. A.; Wilson, T. J.; Chase, S. F.; Laue, T. M. Valence and
anion binding of bovine ribonuclease A between pH 6 and 8. Anal. Biochem. 2005, 336, 243–
252, doi:10.1016/j.ab.2004.09.009.
11.
Her, C.; Filoti, D. I.; McLean, M. A.; Sligar, S. G.; Alexander Ross, J. B.; Steele, H.; Laue, T. M.
The Charge Properties of Phospholipid Nanodiscs. Biophys. J. 2016, 111, 989–998,
doi:10.1016/j.bpj.2016.06.041.
12.
Ridgeway, T. M.; Hayes, D. B.; Moody, T. P.; Wilson, T. J.; Anderson, A. L.; Levasseur, J. H.;
Demaine, P. D.; Kenty, B. E.; Laue, T. M. An apparatus for membrane-confined analytical
electrophoresis. Electrophoresis 1998, 19, 1611–1619, doi:10.1002/elps.1150191016.
13.
Laue, T. M.; Shepard, H. K.; Ridgeway, T. M.; Moody, T. P.; Wilson, T. J. Membrane-confined
analytical electrophoresis. Methods Enzymol. 1998, 295, 494–518.
14.
Kyne, C.; Jordon, K.; Filoti, D. I.; Laue, T. M.; Crowley, P. B. Protein charge determination and
implications for interactions in cell extracts. Protein Sci. Publ. Protein Soc. 2017, 26, 258–267,
doi:10.1002/pro.3077.
15.
Moody, T. P.; Shepard, H. K. Nonequilibrium thermodynamics of membrane-confined
electrophoresis. Biophys. Chem. 2004, 108, 51–76, doi:10.1016/j.bpc.2003.10.009.
16.
Scatchard George The attractions of proteins for small molecules and ions. Ann. N. Y. Acad.
Sci. 1949, 51, 660–672, doi:10.1111/j.1749-6632.1949.tb27297.x.
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 2 November 2018
doi:10.20944/preprints201811.0052.v1
14 of 14
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
17.
Gokarn, Y. R.; Fesinmeyer, R. M.; Saluja, A.; Razinkov, V.; Chase, S. F.; Laue, T. M.; Brems, D.
N. Effective charge measurements reveal selective and preferential accumulation of anions,
but not cations, at the protein surface in dilute salt solutions. Protein Sci. Publ. Protein Soc.
2011, 20, 580–587, doi:10.1002/pro.591.
18.
Miao, L.; Qin, H.; Koehl, P.; Song, J. Selective and specific ion binding on proteins at
physiologically-relevant concentrations. FEBS Lett. 2011, 585, 3126–3132,
doi:10.1016/j.febslet.2011.08.048.
19.
Edsall, J. T.; Wyman, J. Chapter 9 - Polybasic Acids, Bases, and Ampholytes, Including
Proteins. In Biophysical Chemistry; Edsall, J. T., Wyman, J., Eds.; Academic Press, 1958; pp. 477–
549 ISBN 978-1-4832-2946-1.
20.
Harlow, E.; Lane, D. P. Antibodies: A Laboratory Manual; 1 edition.; Cold Spring Harbor
Laboratory Press, 1988; ISBN 978-0-87969-314-5.
21.
Stafford, W. F.; Braswell, E. H. Sedimentation velocity, multi-speed method for analyzing
polydisperse solutions. Biophys. Chem. 2004, 108, 273–279, doi:10.1016/j.bpc.2003.10.027.
22.
Bjellqvist, B.; Ek, K.; Giorgio Righetti, P.; Gianazza, E.; Görg, A.; Westermeier, R.; Postel, W.
Isoelectric focusing in immobilized pH gradients: Principle, methodology and some
applications. J. Biochem. Biophys. Methods 1982, 6, 317–339, doi:10.1016/0165-022X(82)90013-6.
23.
Mao, Q.; Pawliszyn, J. Capillary isoelectric focusing with whole column imaging detection for
analysis of proteins and peptides. J. Biochem. Biophys. Methods 1999, 39, 93–110,
doi:10.1016/S0165-022X(99)00006-8.
24.
Hayes, D.; Laue, T.; Philo, J. Program Sednterp: Sedimentation Interpretation Program version 1.09.;
Alliance Protein Laboratories: Thousand Oaks, CA, USA, 1995;
25.
Edsall, J. T.; Wyman, J. Biophysical chemistry; Academic Press: New York, 1958; ISBN 978-0-12232201-3.
26.
Laue, T. Charge matters. Biophys. Rev. 2016, 8, 287–289, doi:10.1007/s12551-016-0229-3.
27.
Tomar, D. S.; Kumar, S.; Singh, S. K.; Goswami, S.; Li, L. Molecular basis of high viscosity in
concentrated antibody solutions: Strategies for high concentration drug product development.
mAbs 2016, 8, 216–228, doi:10.1080/19420862.2015.1128606.
28.
Saluja, A.; Fesinmeyer, R. M.; Hogan, S.; Brems, D. N.; Gokarn, Y. R. Diffusion and
sedimentation interaction parameters for measuring the second virial coefficient and their
utility as predictors of protein aggregation. Biophys. J. 2010, 99, 2657–2665,
doi:10.1016/j.bpj.2010.08.020.
29.
Laue, T. Proximity energies: a framework for understanding concentrated solutions. J. Mol.
Recognit. JMR 2012, 25, 165–173, doi:10.1002/jmr.2179.