Vaccine, Vol. 13, No. 1, pp.

41--44, 1995
E M A N N Copyright © 1994 Elsevier Science Ltd
Printed in Great Britain. All rights reserved
0264-410X/95 $10.00 + 0.00

Contribution of electrostatic and

hydrophobic interactions to the
adsorption of proteins by
aluminium-containing adjuvants
R a g h e b H. A1-Shakhshir*, F r e d E. Regnier*, J o e L. W h i t e ~;a n d S t a n l e y L. H e m *~
The effect of ionic strength and ethylene glycol on the adsorption of bovine serum albumin
(BSA) or lysozyme by a commercial aluminium hydroxide or aluminium phosphate
adjuvant was studied at pH 7.4 and 25°C. The adsorption of BSA by aluminium hydroxide
adjuvant and lysozyme by aluminium phosphate adjuvant was found to be inversely related
to ionic strength. This indicates that electrostatic attractive forces contribute to adsorption.
The adsorption of lysozyme by aluminium phosphate adjuvant was reduced by the addition
of ethylene glycol. However, no change in the adsorption of BSA by aluminium hydroxide
adjuvant was noted when up to 40% ethylene glycol was present. This behaviour indicates
that hydrophobic forces contribute to the adsorption of lysozyme but not of BSA. However,
virtually no adsorption was observed when the protein and the adjuvant had the same
surface charge. Thus, attractive forces may not be sufficient to produce adsorption of an
antigen by an aluminium-containing adjuvant if electrostatic repulsiveforces are present.

Keywords: Protein adsorption; aluminium-containing adjuvants; electrostatic forces

Antigens are frequently adsorbed by aluminium-
containing adjuvants during the formulation of vaccines
in order to enhance the immune response 1. Unfortunately,
the adsorption mechanism is not fully understood
although recent studies using model proteins 2 as well as
malaria antigens 3 have demonstrated that adsorption is
pH-dependent, with maximum adsorption occurring
under pH conditions where the antigen and adjuvant are
oppositely charged. This behaviour suggests that
electrostatic forces contribute to adsorption. A more
direct test of the presence of electrostatic adsorption
forces is the effect of ionic strength on the adsorptive
capacity. The contribution of electrostatic attractive
forces to antigen adsorption will be studied by
determining the effect of ionic strength on the adsorption
of two model proteins by a commercial aluminium
hydroxide or aluminium phosphate adjuvant 4.
Hydrophobic attractive forces are a major factor in
determining the conformation of proteins. It is possible
that hydrophobic interactions may contribute to
the adsorption of antigens by aluminium-containing
*Department of Industrial and Physical Pharmacy, tDepart-
ment of Chemistry and tDepartment of Agronomy, Purdue
University, West Lafayette, IN 47907, USA. °°To whom
correspondence should be addressed at: 1136 Pharmacy
Building, Purdue University, West Lafayette, IN 47907, USA.
(Received 15 February 1994; revised 2 May 1994; accepted
2 May 1994)
adjuvants. Ethylene glycol is frequently added to mobile
phases in hydrophobic interaction chromatography to
reduce the hydrophobic interaction between protein and
sorbent 5. It functions by stabilizing the hydration layer
of the protein, which renders hydrophobic interactions
thermodynamically unfavourable. Thus, the effect of
ethylene glycol will also be studied to determine if hydro-
phobic attractive forces play a role in the adsorption of
antigens by aluminium-containing adjuvants.
MATERIALS AND METHODS
Two commercially available aluminium-containing adju-
vants were studied: aluminium hydroxide (Alhydrogel
'85', Superfos Biosector a/s, Vedbaek, Denmark) and
aluminium phosphate (Adju-Phos, Superfos Biosector
a/s). X-ray diffraction, infra-red spectroscopy, trans-
mission electron microscopy and energy dispersive
spectrometry were used as previously described4 to
identify Alhydrogel '85' as crystalline aluminium
oxyhydroxide and Adju-Phos as amorphous aluminium
hydroxyphosphate. Crystallized bovine serum albumin
(Sigma, St Louis, MO) and chicken egg white lysozyme
(Sigma) were selected as the model proteins.
Protein adsorption by the adjuvants at pH 7.4 and
25°C was studied in 15ml suspensions containing a
quantity of adjuvant equivalent to 0.85 mg AI/0.5 ml. This
concentration of adjuvant was selected because FDA

Vaccine 1995 Volume 13 Number 1 41

Protein adsorption by aluminium-containing adjuvants: R.H, AI-Shakhshir et al.

regulations allow no more than 0.85 mg of aluminium RESULTS AND DISCUSSION

per dose of vaccine 6.
Aliquots of the aluminium hydroxide adjuvant (3%
equivalent AI(OH)3) containing 25.5 mg A1 were diluted
to 6ml in 0.06, 0.25 or 0.75M NaC1. The pH was
adjusted to 7.4 with either hydrochloric acid or sodium
hydroxide solutions having normalities corresponding to
the diluting solution. The final volume was adjusted to
10ml using the same NaCI solution. Another set of
aliquots also containing 25.5 mg A1 were diluted to 6 ml
in 0, 20 or 40% ethylene glycol and 0.06 M NaCI solution.
The pH of these aliquots was adjusted to 7.4 with either
0.06M HC1 or 0.06M NaOH. The final volume was
adjusted to 10 ml with the initial diluting solution.
Stock solutions containing 30mg bovine serum
albumin (BSA) or lysozyme per ml were prepared by
dissolving the protein in 0.06, 0.25 or 0.75 M NaC1 or
solutions containing 0, 20, or 40% ethylene glycol and
0.06 M NaC1. The pH was adjusted to 7.4 as above.
Aliquots containing 0-105 mg of protein were diluted to
5 ml with the same dilution solution at pH 7.4.
Adsorption studies were performed by adding 5 ml of
the protein stock solution to 10ml of the appropriate
adjuvant suspension. The samples were allowed to
equilibrate with agitation for 30 min and then centrifuged.
The concentration of protein in the supernatant was
measured using the bicinchoninic acid protein assay
procedure (Pierce, Rockford, IL). The standard protocol
(20-1200pgm1-1) was used. The amount of adsorbed
protein was determined by mass balance. The adsorption
data were plotted according to the Langmuir equation 7.
The commercial aluminium phosphate adjuvant (2%
equivalent A1PO4) was suspended in 0.15 M NaC1. The
same adsorption procedure was followed as above except
that the NaCI concentration of the dilution solution was
adjusted so that the final concentration was 0.06, 0.15 or
0.25 M NaCI.
The surface charge was characterized by measuring the
electrophoretic mobility (DELSA 440, Coulter, Hialeah,
FL) of sediment aliquots diluted 1:80 with supernatant.
The isoelectric point (pI) of the aluminium-containing
adjuvants and proteins was determined based on the
principle that the electrophoretic mobility is zero when
the pH is at the isoelectric point. The pls of the aluminium
hydroxide and aluminium phosphate adjuvants were 1I. 1
and 5.0, respectively. The pls of lysozyme and BSA were
9.6 and 5.0, respectively. Thus, aluminium hydroxide
adjuvant and lysozyme will be positively charged while
aluminium phosphate adjuvant and BSA will be
negatively charged at pH 7.4.
The adsorption of negatively charged BSA by
positively charged aluminium hydroxide adjuvant at
pH 7.4 follows the Langmuir equation at ionic strengths
of 0.06, 0.25 and 0.75 M as the R 2 value for the linear
form of the Langmuir equation is 0.99 (Table 1). The
adsorptive capacity was taken from the plateau of the
isotherm shown in Figure 1. The adsorptive capacity
decreased from 1.71 to 1.43 mg BSA/mg AI as the ionic
strength increased from 0.06 to 0.75 M. This behaviour
indicates that electrostatic forces are contributing to
adsorption.
The adsorption of positively charged lysozyme by
positively charged aluminium hydroxide adjuvant at
0.06 M ionic strength was much less than was observed
for BSA. Only ~0.1 mg lysozyme was adsorbed per mg
A1 even when the concentration of lysozyme in
the supernatant was 4 - 6 m g m l - l . Thus, electrostatic
repulsive forces in these systems are apparently strong
enough to prevent substantial adsorption of lysozyme by
other attractive forces. In addition, the adsorption
isotherm did not fit the Langmuir equation (Table 1).
The adsorption of positively charged lysozyme by
negatively charged aluminium phosphate adjuvant was
strongly dependent on the ionic strength (Figure 2). At
low ionic strength, 0.06 M, the adsorption followed the
Langmuir equation (Table 1). However, at higher ionic
strengths adsorption was reduced and did not follow the
Langmuir equation. This behaviour indicates that

Table 1 Effect of ionic strength and ethylene glycol concentration on the adsorptive capacity ot bovine serum albumin and lysozyme by
aluminium-containing adjuvants at pH 7.4 and 25°C

Adsorptive capacity Langmuir

_+standard error equation
,Adjuvant Protein Adsorption solution (mg mg -1 AI) Ft 2

Aluminium hydroxide BSA 0.06 M NaCI 1.71___0.03 0.99

0.25 M NaCI 1.61 -t-0.04 0.99
0.75 M NaCI 1.43 _ 0.02 0.99
Aluminium hydroxide Lysozyme 0.06 M NaCI - 0.35
Aluminium phosphate Lysozyme 0.06 M NaCI 0.66_+0.10 0.99
0.15 M NaCI - 0.74
0.25 M NaCI - 0.69
Aluminium phosphate BSA 0.06 M NaCI
Aluminium hydroxide BSA 0% Ethylene glycol 1.71 + 0.03 0.99
m 0.06 M NaCI
20% Ethylene glycol 1.70 -t- 0.05 0.99
in 0.06 M NaCI
40% Ethylene glycol 1.68 _+0.03 0.99
in 0.06 M NaCI
Aluminium phosphate Lysozyme 0% Ethylene glycol 0.66+0.10 0.99
in 0.06 M NaCI
20% Ethylene glycol 0.57 + 0.02 0.99
m 0.06 M NaCI
40% Ethylene glycol 0.47 ___0.03 0.98
in 0.06 M NaCI

42 V a c c i n e 1995 V o l u m e 13 N u m b e r 1
 Protein adsorption by aluminium-containing adjuvants: R.H. AI-Shakhshir et al.

z.oT Figure 4. The adsorption followed the Langmuir equation

s e3 in the presence of ethylene glycol (Table 1). However, the
o
adsorptive capacity was inversely related to the ethylene
r
1.5
F/ glycol concentration. This indicates that hydrophobic
interactions play a role in the adsorption of lysozyme by
aluminium phosphate adjuvant.
1.0 The number of binding sites is a factor that may
o
contribute to the sensitivity of lysozyme in comparison
with BSA to the ionic strength or the ethylene glycol
<, concentration. In macromolecules the number of binding
0.5
sites is generally proportional to the molecular weight s.
Therefore, the adsorption of lysozyme (M, 14300) is
expected to be more susceptible than BSA (Mw 66 300)
0.0 t , , , 1,,, ~1 ~ E, , I, ~, , I ,t ~ ,t
to changes in the adsorption conditions 9.
0.0 1.0 2.0 3.0 4.0 5.0 The observation that hydrophobic interactions con-
BSA in solution (mg/ml) tribute to the adsorption of lysozyme but not BSA is
Figure 1 Effect of electrolyte concentration on the Langmuir adsorption consistent with the report ~° that lysozyme has a longer
isotherm for bovine serum albumin by aluminium hydroxide adjuvant at retention time than BSA during hydrophobic interaction
pH 7.4 and 25°C. (©) 0.06 M NaCI; (~) 0.25 M NaCl; (17) 0.75 M NaCI chromatography.

0.8 I 2.0
t-

~" 1.5
v 0.6 /

S g
0.4 1.0
..Q
S
.<
os

0.¢
0.0 1.0 2.0 3.0 4.0 5.0 6.0 0.0
Lysozyrne in solution (mg/rnl)
0.0 1.0 2.0 3.0 4.0 5.0
Figure 2 Effect of electrolyte concentration on the Langmuir adsorption BSA in solution (mg/ml)
isotherm for lysozyme by aluminium phosphate adjuvant at pH 7.4 and
25~C. (O) 0.06M NaCI; (~) 0.15M NaCl; ([-I) 0.25M NaCI Figure 3 Effect of ethylene glycol on the Langmuir adsorption isotherm
for bovine serum albumin by aluminium hydroxide adjuvant at ionic
strength 0.06M, pH7.4 and 25°C. (O) 0% Ethylene glycol; (~) 20%
ethylene glycol; (E]) 40% ethylene glycol

electrostatic adsorption forces are important in the 1"0 I o

adsorption of lysozyme by aluminium phosphate
adjuvant. <
Virtually no adsorption of negatively charged BSA by 0.8 ~ o o
negatively charged aluminium phosphate adjuvant was vE 0.6 ~x ~
observed. Apparently, the electrostatic repulsive force .¢1
prevents adsorption by other forces. Q
Adsorption studies were conducted in solutions 0.4
containing ethylene glycol to determine the role of
hydrophobic interactions 5. The adsorption of BSA by
0.2
aluminium hydroxide adjuvants in solutions containing
0, 20 or 40% ethylene glycol and 0.06 M NaCI followed
the Langmuir equation (Table 1). However, the ethylene 0.0 ~ r~ I~~ '' I' 'I' i ~~ =+ I E'''I~' 'I [
glycol did not affect the adsorptive capacity (Figure 3, 0.0 1.0 2.0 3.0 4.0 5.0 6.0
Table 1). This indicates that hydrophobic interactions do Lysozymeinsolution(mg/ml)
not contribute to the adsorption of BSA by aluminium
hydroxide adjuvant at pH 7.4. Figure 4 Effect of ethylene glycol on the Langmuir adsorption isotherm
for lysozyme by aluminium phosphate adjuvant at ionic strength 0.06 M,
The adsorption of lysozyme by aluminium phosphate pH 7.4 and 25°C. (O) 0% Ethylene glycol; (A) 20% ethylene glycol; (I--I)
adjuvant was affected by ethylene glycol, as seen in 40% ethylene glycol

V a c c i n e 1995 V o l u m e 13 N u m b e r 1 43
Protein adsorption by aluminium-containing adjuvants: R.H. AI-Shakhshir et al.
Proteins with high conformational adaptability, such
as BSA and to a lesser extent lysozyme, can to a certain
degree adsorb at most surfaces, hydrophilic or
hydrophobic, even if conditions are not favourable 11.
This is consistent with the observation that small
amounts of BSA or lysozyme were adsorbed by the
hydrophilic surfaces of aluminium phosphate adjuvant
or aluminium hydroxide adjuvant, respectively, even
though adsorption is electrostatically unfavourable.
This study has demonstrated that electrostatic
attractive forces and hydrophobic forces contribute to
the adsorption of antigens by aluminium-containing
adjuvants. It is likely that van der Waals forces and
hydrogen bonding also contribute to adsorption in these
systems. However, the attractive adsorption forces may
not be sufficient to produce adsorption of an antigen
by an aluminium-containing adjuvant if electrostatic
repulsive forces are present, especially when the protein
has high structural stability 11.
ACKNOWLEDGEMENTS
This study was supported in part by a Purdue Research
Foundation Fellowship (R.H.A.-S.) and Merck Research
Laboratories.
44 V a c c i n e 1995 V o l u m e 13 N u m b e r 1
REFERENCES
1 Edelman, R. Vaccine adjuvants. Rev. Infect. Dis. 1980, 2, 370-383
2 Seeber, S.J., White, J.L. and Hem, S.L. Predicting the adsorption of
proteins by aluminium-containing adjuvants. Vaccine 1991, g,
201-203
3 Callahan, P.M., Shorter, A.L. and Hem, S.L. The importance of
surface charge in the optimization of antigen-adjuvant interactions.
Pharm. Res. 1991, 8, 851-858
4 Shirodkar, S., Hutchinson, R.L., Perry, D., White, J.L. and Hem, S.L.
Aluminum compounds used as adjuvants in vaccines. Pharm. Res.
1990, 7, 1282-1288
5 Srinivasan, R. and Ruckenstein, E. Role of physical forces in
hydrophobic interaction chromatography. Sep. Purif. Methods 1980,
g, 267-370
6 May, J.C., Progar, J.J. and Chin, R. The aluminum content of
biological products containing aluminum adjuvants: determination
by atomic absorption spectrometry. J. Biol. Stand. 1984,12, 175-183
7 Adamson, A.W. Physical Chemistry of Surfaces, 3rd Edn, Wiley,
New York, 1976, pp. 386-389
8 Arai, T. and Norde, W. The behavior of some model proteins at
solid-liquid interfaces: 2. Sequential and competitive adsorption.
Colloids Surf. 1990, 51, 17-28
9 Squires, P.G. and Himmel, M.H. Hydrodynamics and .protein
hydration. Arch. Biochem. Biophys. 1979, lg6, 165-177
10 Gooding, D.L., Schmuck, M.N. and Gooding, K.M. Analysis of
proteins with new, mildly hydrophobic high performance liquid
chromatography packing materials. J. Chromatogr. 1984, 296,
107-114
11 Norde, W. and Lyklema, J. Why proteins prefer interfaces.
J. Biomater. Sci. Polym. Ed. 1991, 2, 183-202