70I
J. gen. Virol. 0979), 43, 7oi-7o6
Printed in Great Britain
Copper Chelate Affinity Chromatography
Fibroblast
of Human
and Leucocyte Interferons
(Accepted I6 January I979)
SUMMARY
Human fibroblast and human leucocyte interferons display a strong affinity for
the copper chelate of bis-carboxymethyl amino agarose, binding tenaciously over
a wide pH range (7"4 to 4.o). Their binding is apparently irreversible on a sorbent
saturated with copper (24"8/~mol of Cu~+/ml of column bed). However, both
interferons can be partially recovered from sorbents of lower copper content,
prepared by leaching the columns with sodium citrate at pH 9"o. The recovery of
fibroblast interferon from a leached sorbent (5"8 #tool of Cu2+/ml of column bed)
is about 3° % and that of leucocyte interferon about 60 %. Moreover, the strength
of binding of leucocyte interferon can be modulated by leaching copper chelateagarose with citrate of varying concentration.
Several proteins have an affinity for transition metal chelates, and this property has been
utilized for their purification (Porath et al. r975; L6nnerdal et al. r977; Lebreton, I977).
Recently, human fibroblast interferon was highly purified on Zn ~+ chelate-agarose with
a good (approx. 6o %) recovery of activity (Edy et al. I977). Syrian hamster interferon was
chromatographed on Cu 2+ chelate-agarose with excellent (about 9o %) recovery of activity
(Bollin & Sulkowski, I978). Our preliminary attempts to chromatograph human fibroblast
and human leucocyte interferons on Cu ~+ chelate-agarose were unsuccessful due to their
tenacious binding. A similar, apparently irreversible, binding of some other proteins may also
occur. We now report that the binding strength of Cu 2+ chelate-agarose can be modulated by
varying its metal content. Both human interferons can be displaced from Cu 2+ chelateagarose columns which have been appropriately leached with citrate at pH 9-o.
Iminodiacetic Acid-//-Sepharose CL-4B (iminodiacetic acid coupled to Sepharose
CL-4B matrix via a molecular arm) was purchased from Pierce Chemical Co. (Rockford, Ill.,
U.S.A.). Concanavalin (Con A)-Sepharose was obtained from Pharmacia Fine Chemicals
(Piscataway, N.J., U.S.A.). Fluorescamine was purchased from Roche Diagnostics
(Nutley, N.J.) and ethylene glycol from Baker Chemical Co. (Phillipsburg, N.J.). All
other chemicals were of analytical grade. Human fibroblast interferon was prepared in
human diploid fibroblasts, essentially according to Havell & Vil~ek (I972) and purified as
previously described (Davey et al. I976). Its specific activity was about 2.7 x Io e reference
units/rag protein. Human leucocyte interferon was prepared according to Pidot et al.
(I972). Its specific activity was 5 to 6 x Io~ reference units/mg protein. Disc SDS-PAGE
was performed as described by Summers et al. (I965). Interferon activity was recovered
from gel slices as described elsewhere (Chadha et al. I978). The columns of iminodiacetic
acid-agarose were charged with copper as follows. Columns (9 x 5o to !oo mm) were
equilibrated with o.I M-sodium acetate, I M-NaCI, pH 4"o at 4 °C. Volumes of 5o ml of
CuSO4.5H20 or ZnCI~ at 5 mg/ml in the equilibrating buffer were applied and the columns
were rinsed with ~oo ml of the same buffer. The columns were charged with metal ions in the
equilibrating buffer because the same buffer was used as the terminal eluant of all columns.
Downloaded from www.microbiologyresearch.org by
oozz-I317/79/oooo-3479 $02.00©
t979 SGM
IP: 88.99.165.207
On: Sat, 17 Jun 2017 17:35:00
702
Short communications
The copper content of Cu 2+ chelate-agaroses was measured by displacing the Cu 2+ ions
with 2o mM-EDTA in o.o2 M-sodium phosphate, I M-NaCI, pH 7"4, and measuring the
absorption at 675 nm. The calibration curve was prepared with CuSO4.5H20 in the same
solvent. A solution of 3 mg ofCuSO4.5H20 per ml gave a reading ofo'7 A675(Iicm light path)
in a Gilford spectrophotometer 24o. The copper content of a saturated sorbent prepared as
described above, was approx. 24"8/zmol/ml of column bed. Less than saturated sorbents
could be prepared by rapidly admixing the sorbent cake into a CuSO~ solution of an
appropriate concentration. However, partially charged sorbents, prepared in such a manner,
gave rather poorly reproducible results. Partial saturation with Cu 2+ ions was therefore
achieved by leaching a column as follows. A column containing sorbent saturated with
Cu 2+ was sequentially washed with: (I) o.o2 M sodium phosphate, I M-NaC1, pH 9"o,
IOO ml; (2) o.I to o'5 M-sodium citrate, I M-NaC1 in o.o2 M-sodium phosphate, pH 9"o,
25o ml; (3) 0.o2 M-sodium phosphate, I M-NaCI, pH 9"o, lOO ml and (4) o.2o M-sodium
phosphate, I M-NaC1, pH 7"4, IOO ml. All washings were performed at 4 °C. The copper
contents of sorbents leached with citrate of different concentrations were as follows:
9"3/zmol of Cu2+/ml (o.l u-citrate) and 5"8 t~mol of Cu2÷/ml (0"5 M-citrate).
All columns were equilibrated, charged and developed at 4 °C. Columns were eluted at
a flow rate of approx. 3o ml/cm2/h using a peristaltic pump. When the interferon preparations were chromatographed, the eluate was divided by a stream splitting device. Larger
fractions were used for protein determination; smaller fractions were collected into a 1%
solution of bovine serum albumin in o.oz M-sodium phosphate, pH 7"4 (o'15 M-NaC1), and
used for interferon assays. Interferon activity was assayed on human fibroblasts by the
colorimetric technique of Finter (I969) and titres are given in research reference units.
Protein concentration was measured by fluorometric assay (B6hlen et al. I973) with bovine
serum albumin as a standard.
Human fibroblast interferon (HF-IF) preparations, obtained by chromatography on
Con A-Sepharose, were dialysed and applied to Zn e+ chelate-agarose columns in o.o2 Msodium phosphate, I M-NaC1, pH 7'4, or in o.I M-sodium acetate, 1 M-NaC1, pH 6-o. All
HF-IF was retained and was subsequently eluted (7o% recovery) from the columns at
pH 5"o (not illustrated). However, when an eluate from a Con A-Sepharose column containing 5o % ethylene glycol was diluted twofold and applied on a Zn 2+ chelate-agarose column,
all HF-IF activity was found in the breakthrough fraction. Apparently, H F - I F does not
bind to Zn 2+ chelate-agarose in the presence of ethylene glycol. As dialysis can result in
a partial loss of HF-IF activity, an attempt was made to chromatograph HF-IF preparations (containing 25 % ethylene glycol) on copper chelate-agarose, which is a stronger
sorbent than zinc chelate-agarose (Porath et al. 1975; Bollin & Sulkowski, r978 ). All
HF-IF was retained on a Cu 2+ chelate-agarose column, regardless of the presence of ethylene
glycol. However, the binding strength of copper chelate-agarose (saturated sorbent) was
found to be excessive, as HF-IF could not be recovered even after a prolonged elution of the
column at pH 4"o (not illustrated).
Fig. I illustrates the chromatography of an HF-IF preparation, partially purified on
Con A-Sepharose (containing 2 5 ~o ethylene glycol), on a Cu 2+ chelate-agarose column
which was leached with o'5 M-citrate, pH 9"0. The retention of HF-IF activity, despite
a much lower copper content (5"8 #tool Cu2+/ml) was still complete, but only 3o% of the
activity was recovered by elution of the column at p H 4"o. There was also a concomitant
purification of HF-IF to an extent which cannot reliably be estimated from such a small
scale experiment but probably of at least tenfold. Chromatography on Cu 2+ agarose can
also eliminate traces of concanavalin A found as contaminant (Berthold et al. I978) in
Con A-Sepharose purified HF-IF (not illustrated).
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 17:35:00
703
Short communications
pH6
pH4
i
O
?.0 x
0, I0
>"
= 0.08
1-0 ~.
0.04
0
0
10
20
30
40
50
Fraction number
60
70
Fig. I. Chromatography of human fibroblast interferon on copper chelate-agarose. An interferon
preparation, purified on Con A-Sepharose, was dialysed against an equilibrating buffer (o.I Msodium acetate, I M-NaCI, pH 6'0, containing 25 % ethylene glycol). A sample of 2 ml, containing
t'z x io n units of activity and 0"45 mg of protein was applied to the column (9 x IOO mm) leached
with 0"5 M-sodium citrate, pH 9"0, which was initially washed with the equilibrating buffer, and then
with buffer at pH 4"0, as indicated by arrows. The recovery of interferon activity was about 30 %
and that of protein about 57 %. (2)--0, interferon; 0 - - 0 , protein.
Having shown that human fibroblast interferon can be chromatographed on Cu 2+
chelate-agarose, we attempted to extend our observations to human leucocyte interferon
(HL-IF). In a preliminary set of experiments, we observed that HL-IF did not bind to zinc
chelate-agarose, thus confirming the earlier observations of Edy et al. (I977). Copper
chelate-agarose was then tested as a possible metal chelate sorbent. It was soon observed
that HL-IF had a high affinity for this metal chelate. Fig. 2(a) illustrates the chromatography of HL-IF on a Cu ~+ chelate-agarose column (saturated sorbent). HL-IF bound
completely and only an insignificant amount, about 5 %, of the applied activity could be
recovered from the column with a falling pH gradient. By contrast, when the chromatography was performed on a leached sorbent (o'5 M-citrate, pH 9"o), as shown in Fig. 2(b),
the recovery of HL-IF was approx. 6o %. However, a major portion of HL-IF recovered
from the column was found, together with other proteins in the breakthrough fraction.
Moreover, the bound and then displaced portion of HL-IF was eluted from the column
along with large amounts of other proteins, rendering this particular sorbent (leached with
o'5 M-citrate) valueless for the purification of HL-IF. Clearly, a sorbent with binding
properties intermediate between fully charged and thoroughly leached could be more useful.
To this end, we explored the sorptive properties of copper chelate-agarose leached with
citrate of varying concentration. Copper chelate-agarose, prepared by leaching a fully
charged column with o-I u-citrate, pH 9"o (9"3/~mol/ml of column bed), proved to be
a sorbent with the desired properties: there was a complete retention of HL-IF and its
displacement, although partial (about 6o % of the applied amount), was rather selective as
shown in Fig. z (c).
The experiments illustrated in Fig. 2 were performed in the presence of ethylene glycol
(25 %). When ethylene glycol was omitted from all solvents, the results were essentially the
same (not illustrated).
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 17:35:00
Short communications
7o4
(a)
EL
E2
"6.0
0-4
5"5
~__~
5.0
0.2
- 4.5
4.0
90.47_
6o
60
Y"
o
50
X
40
"~
5.0
©
0.2
4-5
30
20
va
-7,
4-0
0 I
J
c)
,
gI
i
E~
60
0.4
0.2
0
10
20
30 40 50 60
Fraction number
70
80
oo
50
5.5
40
5.0
30
4.5
20
4.0
l0
22_
Fig. 2. Modulation of the binding strength of copper chelate-agarose for human leucocyte interferon. (a) Copper chelate-agarose saturated with copper. An interferon preparation was dialysed
against o-o2 M-sodium phosphate, I M-NaC1, pH 7"4, containing 25 % ethylene glycol. A sample
of 3 ml containing 4oooo units of activity and 8' 1 mg of protein was applied to a column (9 x 5o mm)
equilibrated with the dialysis buffer. The column was initially washed with equilibrating buffer
then with o.t M-sodium acetate, I M-NaCI, pH 6.o, 25 % ethylene glycol. The column was then
developed with a pH 6 to 4 gradient (El). Finally, the column was developed with the terminal
buffer (E~). The recovery of interferon was about 5 % and that of protein about lo %. (b) Copper
chelate-agarose leached with o'5 M-sodium citrate, pH 9"o. A sample of 5 ml containing 7oooo
units of activity and 1o5 mg of protein was applied to the column. The development of the column
was identical to that described in (a). The recovery of interferon was about 6o % and that of
protein about 55 %. (c) Copper chelate-agarose leached with oq M-sodium citrate, pH 9"o. A sample
of 5 ml, containing 62 50o units of activity and 12 mg of protein was chromatographed as described
in (a). The recovery of interferon Was about 6o % and that of protein about 48 %. C)--O, Interferon; •
• , protein; - - ,
pH.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 17:35:00
Short communications
705
The partial retention of HL-IF on a column leached with 05 M-citrate (Fig. 2b) could
reflect the molecular size heterogeneity of HL-IF (Stewart & Desmyter, 1975), or be due to
other factors of trivial nature (column capacity, complexing with other proteins, etc.). In
order to check whether the first alternative is true, we determined the mol. wt. of the HL-IF
portion found in the breakthrough fraction and the HL-IF portion retained and subsequently displaced from a column (leached with o'5 M-citrate, pH 9.o). SDS-PAGE
analysis revealed that both molecular subcomponents of HL-IF, mol. wt. 2Iooo and
I5OOO, were present not only in the breakthrough fraction but also in the fraction eluted by
a falling pH gradient. The proportion of 2IOOO mol. wt. component to r5ooo tool. wt.
component was 3 : 7 in the breakthrough fraction and 6: 4 in the retained and subsequently
eluted fraction. Thus it seems that there is threefold enrichment rather than selective
retention of the higher tool. wt. form of HL-IF on the Cu 2+ chelate-agarose column (leached
with o'5 M-citrate, pH 9.o).
Metal chelate affinity chromatography of proteins offers a novel purification principle
which has already been successfully exploited with crude HF-IF (Edy et al. I977). Our aim
is to use this technique to follow the initial purification of HF-IF(Davey et al. I976) and for
the purification of HL-IF. Zinc chelate chromatography of HF-IF cannot be performed in
the presence of ethylene glycol. As the dialysis of a Con A-Sepharose eluate, containing
HF-IF and ethylene glycol, may result in some loss of activity, we tried to avoid it by
employing copper chelate-agarose rather than zinc chelate-agarose. The data reported here
show that copper chelate affinity chromatography can be performed with both HF-IF and
HL-IF, although the recovery of the interferon activity, particularly in the case of HF-IF,
is not completely satisfactory.
Excessive binding of both interferons to copper chelate-agarose, in comparison to that
on zinc chelate-agarose, may be due to the formation of additional linkages; this would
occur if some amino acid side chains on human interferons coordinate to Cu 2+ chelates but
not to Zn 2+ chelates. Another explanation also has to be considered, namely that at a certain
metal density of the sorbent, not all possible bonds are formed between percolating protein
molecules and the sorbent. An increase in the metal density, above a certain critical level,
would result in the formation of multiple coordination bonds. Whichever alternative is
correct, and they are not mutually exclusive, a low metal density should result in a decrease
in the number of attachment points and a fall in the strength of binding. The change in the
sorptive properties of Cu 2+ chelate-agarose, as a result of its leaching with citrate, is congruent with the above interpretation. A further substantial contribution to the understanding
of binding of both interferons to metal chelates could be made by preparing a series of
sorbents with different degrees of substitution with iminodiacetate and evaluating their
sorptive properties. The chelates of other transition metals (Co 2+, Ni 2+) should also be
explored as potential ligands and such an investigation is under way.
This work was supported in part by a Center Grant in Viral Chemotherapy CA 14801-06;
by Public Health Service Grant AI 12933-o3 and the State of New York, Department of
Health.
Department of Viral Oncology
Roswell Park Memorial Institute
Buffalo, New York I4263 , U.S.A.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 17:35:00
K. C. CHADHA
P. M. GROB
A. J. MIKULSKI
L. R. DAVIS, JUN.
E. SULKOWSKI
706
Short communications
REFERENCES
BERTHOLD, W., TAN, C. & TAN, Y. H. (t978). Purification and in vitro labeling of interferon from a h u m a n
fibroblastoid ceU line. Journal of Biological Chemistry z53, 5206-521 I.
B()HLEN, P., STEIN, S., DAIRMAN, W. & UDENERIEND, S. (1973). F l u o r i m e t r i c a s s a y o f p r o t e i n s in the nanogram
range. Archives of Biochemistry and Biophysics x55 , 213-220.
BOLLIN, E. JUN. & SULKOWSKI, E. (1978). Metal chelate affinity chromatography of hamster interferon.
Archives of Virology 58, 149-152.
CHADHA, K. C., SCLAIR, M., SULKOWSKI, E. & CARTER, W. A. (1978). Molecular size heterogeneity of human
leukocyte interferon. Biochemistry x7, I96-2oo.
DAVEY, M. W., SULKOWSKI, E. & CARTER, W. A. (I976). Binding of h u m a n fibroblast interferon to concanavalin
A-agarose. Involvement of carbohydrate recognition and hydrophobic interaction. Biochemistry IS,
7o4-713.
EDY, V. G., BILLIAU, g. & DE SOMER, P. (1977). Purification of human fibrobIast interferon by zinc chelate
affÉnity chromatography. Journal of Biological Chemistry zSz, 5934-5935.
FINTER, N. B. (1969). Dye uptake methods for assessing viral cytopathogenicity and their application to
interferon assays. Journal of General Virology 5, 419-427.
HAVELL, E. A. & VILCEK, J. (1972). Production of high-titred interferon in cultures of human diploid cells.
Antimicrobial Agents and Chemotherapy z, 476-484.
LEBRETON, J.P. (I977)- Purification of human plasma alpha2-SH glycoprotein by zinc chelate affinity
chromatography. FEBS Letters 8o, 351-354.
LONNERDAL, B., CARLSSON,J. & PORATH, J. 0977). Isolation of lactoferrin from h u m a n milk by metal-chelate
affinity chromatography. FEBS Letters 75, 89-92.
PIDOT, A. L. R., O'KEEFE n, G., MCMANUS, N. & MclNTYRE, O. R. (1972). H u m a n leukocyte interferon: the variation in normals and correlation with P H A transformation. Proceedingsof the Society for Experimental
Biology and Medicine i4o , I263-I269.
PORATH, J., CARLSSON, J., OLSSON, I. & BELFRAGE, G. (I975). M e t a l c h e l a t e affinity c h r o m a t o g r a p h y , a n e w
approach to protein fractionation. Nature, London z$8, 598-599.
STEWART II, W.E. & DESMYTER, .I. (1975). Molecular heterogeneity of h u m a n leukocyte interferon: two
populations differing in molecular weights, requirements for renaturation, and cross-species antiviral
activity. Virology 67, 68-73.
SUMMERS, D. E., MAIZEL, J.V. & DARN'ELL, .L E. (1965). Evidence for virus specific non-capsid proteins in
poliovirus infected HeLa cells. Proceedingsof the National A cademy of Sciences of the United States of
America 54, 5o5-513.
(Received 25 September I 9 7 8 )
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sat, 17 Jun 2017 17:35:00