J. gen. Virol. (I978), 38, 241-249
241
Printed in Great Britain
The Chemical Nature o f an Antiviral Factor (AVF) from
Virus-infected Plants
By R I T A M O Z E S , Y. A N T I G N U S ,
I. S E L A AND I. H A R P A Z
Virus Laboratory, The Hebrew University, Faculty of Agriculture, Rehovot, Israel
(Accepted IO August 1977)
SUMMARY
An antiviral factor from virus-infected plants (AVF) was purified in an active
form on SDS-polyacrylamide gels. AVF binds to concanavalin A and is partially
sensitive to c~-glucosidase. It is sensitive to pronase only when incubated in conditions suitable for proteolysis of glycoproteins. Alkaline phosphatase affected
the electrophoretic mobility of AVF, but did not abolish antiviral activity. AVF
was insensitive to DNase, fl-glucosidase and pancreatic lipase. The AVF band
obtained upon electrophoresis could be stained with Coomassie blue and by the
Schiff-periodate procedure for carbohydrates. AVF is considered to be a phosphoglycoprotein with a mol. wt. of about 22000.
INTRODUCTION
The antiviral factor (AVF) from tobacco mosaic virus (TMV)-infected leaves of Nicotiana
glutinosa L. is a substance occurring in response to virus infection in plants which carry
the gene for virus localization (Antignus, Sela & Harpaz, I977). AVF was reported in the
past to be sensitive to ribonuclease (Sela, Harpaz & Birk, 1966). Later on, however, we found
that AVF activity could be restored upon removal of the ribonuclease, and that AVF does
not contain RNA (Antignus, Sela & Harpaz, 1975). It is possible to label AVF with 32p.
orthophosphate, and t o get distribution patterns of radioactivity and antiviral activity
upon polyacrylamide gel electrophoresis (Antignus et al. I977). Hence it became possible
to follow AVF integrity after various treatments as reported in the present communication.
METHODS
A VF preparation, labelling and activity. The preparation of a semi-purified AVF eluted
from DEAE-cellulose, its electrophoresis on gels, its labelling with radioactive phosphate
and the test for AVF activity, are all described elsewhere (Antignus et al. 1975, I977). To
follow the distribution of antiviral activity in a gel, individual gel slices (in groups of three)
were stirred in I ml of 0"65 M-NaC1 plus o'oI M-Na-phosphate buffer, pH 7"6, for I2 h and
centrifuged at 5000 g for IO min. The supernatant fluid was then dialysed against the same
buffer and checked for its antiviral activity. Radioactivity and antiviral activity were determined in sister gels of the same electrophoretic run.
Chemical determinations. The amount of protein in the AVF preparations was determined
as follows: proteins were precipitated in Io ~ cold trichloroacetic acid and sedimented by
low speed centrifugation. The precipitate was dissolved in o.2 ml I N-NaOH and its protein
content determined according to Lowry et al. (195 I). Carbohydrates were determined by the
16
VIR 38
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 19:08:53
242
R. M O Z E S A N D
OTHERS
anthrone method (Hassid & Neufeld, I964) and lipids were determined by the dichromate
method (Amenta, I964). Sugar analysis was by gas-liquid chromatography as described by
Phillips & Smith (r973).
Enzymes. All enzymes were applied to solutions containing 2 mg of DEAE-eluted AVF
in I ml as substrate. The integrity of AVF following enzyme digestion was checked through
its antiviral activity, prior to or after electrophoresis, or through the radioactivity patterns
on gels.
Digestion with alkaline phosphatase (EC 3- I. 3- i; Worthington Biochemical Corp.) was
carried out with 2 units of enzyme per ml in o'o5 M-tris, p H 8"4, for 3 h at 37 °C. When
applied in combination with another enzyme, incubation conditions were as specified for
the other enzyme. Snake venom phosphodiesterase (EC 3. r .4- I ; Worthington Biochemical
Corp.) was added to AVF solutions to make 0.2 mg/ml in o.ot M-tris, pH 9"3, and incubated for 3 h at 37 °C. Glucosidases (~ and fl) were both from Sigma; ~-glucosidase (EC
3- 2.1.20) was added to AVF to make I mgflnl in o.oI M-Na-phosphate, p H 6"9, and flglucosidase (EC 3.2. I .2I) was added to make the same concentration in o.ot M-Naacetate, pH 5"o. Both were incubated with AVF for 3 h at 37 °C. Pancreatic lipase (EC
3- I. x. 3; Sigma type II) was also incubated with AVF for 3 h at 37 °C under the conditions
specified by Bier (I955). DNase (EC 3. ~. 4.5; Sigma) was also incubated with AVF in the
above acetate buffer for 3 h. Pronase (Serva, no systematic nomenclature) was applied to
AVF under conditions devised for glycoproteins (Spiro, I966). Two mg of DEAE-eluted
AVF were dissolved in t ml of o.oI M-tris, p H 7-8, plus 5 mM-CaC12. Pronase (2o #g) was
added (as well as a drop of toluene to prevent microbial contamination) and incubated at
37 °C for 48 h. An additional i o #g pronase was added to the reaction mixture after 24 h of
incubation.
RESULTS
DEAE-eluted AVF was dialysed against distilled water and lyophilized. Three different
preparations of AVF were each analysed for their protein, sugar and lipid contents as
described above. The following mean values were obtained: 82. 7 ~ protein, 9 ~ carbohydrates and no lipids. Similar results (79 ~ protein, 9.2 ~ carbohydrate) were obtained with
control mock preparations, derived from non-infected plants. The large proportions of
protein were surprising in view of the almost protein-free gels obtained previously (Antignus
et al. I975). However, if most of these proteins were actually phosphorylated glycoproteins,
as suggested below, they would tend to accumulate on the acidic side upon isoelectrofocusing.
At the start of this study only the following relevant information was available to us:
(a) AVF could be effectively labelled with phosphate, (b) its activity was resistant to ribonuclease, and (c) its activity was resistant to proteolytic enzymes under standard conditions
(Sela et al. I966). To check the effect of phosphatases on AVF, alkaline phosphatase and
snake venom phosphodiesterase were each incubated with 32P-labelled DEAE-eluted AVF
as described above, and then the AVF was electrophoresed as described elsewhere
(Antignus et al. I975). As demonstrated in Fig. I, the gel pattern of AVF was noticeably affected by alkaline phosphatase. A considerable fraction of small phosphate-containing fragments was detectable far down the gel, and the main AVF peak, by now much
more uniform but still active, moved slower than before. It is apparent that the AVF contained phosphorylated end-groups in a monoester linkage that contribute considerably to its
electrophoretic mobility, and practically all the phosphate in the zone lying in front of the
main peak was accessible to the enzyme. However, most of the radioactivity of the main
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 19:08:53
Chemical nature o f A V F
12001
/ (.)
I
I ~
~I
I
z43
I
I t
/
400
25
g
0 .~
8 o o - (b)
50
~
e
0
800
4°°1-
......
-
00t
400 ,
0
10
20
30
40
Distance from top (mm)
50
Fig. I. Distribution of radioactivity and antiviral activity among polyacrylamide gel slices following electrophoresis of an AVF preparation which had been previously treated with (b) alkaline
phosphatase, (c) snake venom phosphodiesterase, (d) alkaline phosphatase + snake venom phosphodiesterase. (a) Controls were incubated under the same conditions without enzyme. Radioactivity (A) was determined for every slice (I mm). Antiviral activity (0) was extracted from
groups of three consecutive slices.
peak itself was preserved, indicating the presence of unexposed phosphate groups as well.
When snake venom phosphodiesterase was applied to the DEAE-eluted AVF, the rate of
electrophoretic migration of the radioactive material was not affected, and a consistent
reduction of only about zo ~ in the area of the main radioactive peak was noted. Hence,
the structure of the AVF could not be considered as based on a backbone of phosphodiester chain such as polyphosphate or poly ADP-ribose (Matsubara et al. I97o)- AVF was
found to be insensitive to DNase and to pancreatic lipase.
AVF activity could be extracted with phenol in the presence of a high salt concentration
(Antignus et al. I975)- Beside nucleic acids, acidic polysaccharides can also be extracted in
this manner. Because DEAE-eluted AVF, as mentioned above, contained about 9
polysaccharide, the association of polysaccharide with AVF had to be investigated. G a s liquid chromatography analysis of a DEAE-eluted AVF preparation showed quite a different
pattern of sugars as compared with a control preparation (Fig. 2).
To verify the association of carbohydrates with AVF, 2 mg of DEAE-eluted A V F was
applied to a column of Sepharose-bound concanavalin A (Pharmacia) in o.oi M-phosphate,
p H 7"6. The column was then washed in a stepwise manner with Io ml each of 2 ~ , 4 ~ and
6 ~ ~-methyl-D-glucoside. The same procedure was carried out with a mock AVF preparation. Each fraction was dialysed against water, lyophilized, dissolved in o'5 ml of the above
z6-2
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 19:08:53
244
R. M O Z E S A N D O T H E R S
Galactose
Mannose
UCOSe
Rhamnose
e.,
©
~2a
II
II
II
Arabinose
\.
II
Ribose
,!
\..
\
0
5
11
II
II
II
10
Time 0nin)
15
Fig. 2. Gas-liquid chromatography analyses of sugars in the hydrolysates of a DEAE-eluted AVF
preparation (
) and its control ( - - ) .
Table I. Antiviral activity of fractions of an A VF preparation eluted from a
concanavalin A-Sepharose column
Eluted with
o.ot M-Phosphate, pH 7"6
~-Methyl-D-glucoside (2 ~oo)
~-Methyl-o-glucoside (4 ~ )
cc-Methyl-o-glucoside (6 ~ )
Number of
lesions on
control halfleaves (mock
AVF)
Number of
lesions on
treated halfleaves (AVF)
97I
911
898
918
949
460
4O4
893
protection
2"3
49"5*
55"O*
2"7
* Significance < I 700(t-test for related samples).
phosphate buffei and tested for its capacity to inhibit TMV infection, as compared to mock
AVF, on I2 half-leaves of Datura stramonium L. AVF activity was found to be bound to
the column, but could be eluted with 2 to 4 ~ of ~-methyl-D-glucoside (Table I), indicating
the association of saccharide with AVF.
The presence of saccharides in the active fraction of AVF suggested, among other possi-
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 19:08:53
Chemical nature o f A VF
I
2000 -
(a)
I
I ~
I
245
I
A
100
U
0
10
20
30
40
Distance from top (ram)
50
Fig. 3. Radioactive profiles (A) and distribution of antiviral activity (A), as described in Fig. I,
of an AVF preparation which was electrophoresed on gels after being treated with pronase in a
manner specifiedfor glycoproteins (b) or incubated similarly without enzyme (a).
bilities, that AVF could be of a glycoprotein nature. This would still be consistent with
previous findings, as glycoproteins, under normal conditions, are insensitive to most
proteases. A major corroboration of the claim that protein is associated with AVF would
be its sensitivity to pronase under the extreme conditions specially devised for the proteolysis of glycoproteins. DEAE-eluted AVF, when treated with pronase as described above and
analysed on gels, was completely destroyed as indicated by the loss of its antiviral activity
and the change in its radioactive profile (Fig. 3)The presumed glycoprotein nature of AVF was further studied with better purified
preparations eluted from concanavalin A columns with 5 ~ c~-methyl-D-glucoside, dialysed
and lyophilized (con A-eluted AVF). Con A-eluted AVF was electrophoresed on SDSpolyacrylamide gels and stained with Coomassie blue according to Fairbanks, Steck &
Wallach (1971). The gel pattern varied in different experiments, indicating much heterogeneity. Usually it consisted of a series of bands on a background 'smear'. Nevertheless,
the antiviral activity always coincided with the stained bands. In all cases, a major band (or
a set of bands) was found near the front of the smear. This band was absent from gel patterns of control, mock preparations. One such example is given in Fig. 4- The major AVF
band, shown in Fig. 4, was eluted from an unstained sister gel, divided into two samples and
re-electrophoresed as before. One gel was stained for proteins with Coomassie blue and the
other gel was stained for carbohydrates by the Schiff-periodate procedure (Fairbanks et al.
1971). In both gels, a uniform single band was observed and the bands superimposed, again
indicating the association of both protein and carbohydrate with AVF (Fig. 5).
The susceptibility of con A-eluted AVF to alkaline phosphatase and c~-glucosidase was
tested again. The radioactive profile of AVF and its antiviral activity were partially (and
sometimes completely) destroyed by cc-glucosidase. Alkaline phosphatase again affected the
rate of migration of AVF without affecting its antiviral activity. However, the activity of
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 19:08:53
246
R. MOZES AND OTHERS
60~
40
20 ~
1
20
40
Distance from top (mm)
0
60
Fig. 4. Electropherograms of a con A-eluted AVF preparation. Gels were either stained with
Coomassie blue and scanned at 560 nm, or sliced as described in Fig. i for checking their antiviral
activities. (---) Absorbance of mock preparations; (--) absorbance of the AVF preparation;
(©) pattern ofantiviral activity eluted from an AVF gel and compared on half-leaves with eluants
of a mock gel.
8
L
I
0
I
20
40
Distance from top (ram)
-
-
i
60
Fig. 5. Electrophoresis of material eluted from slices 25 to 35 of a sister gel to that shown in Fig. 4.
(--) Absorbance (at 560 nm) of a gel stained with Coomassie blue; ( - - - ) absorbance (at 530 nm)
of a gel stained by the Schiff-periodate procedure.
a-glucosidase in destroying A V F was enhanced when applied in combination with alkaline
phosphatase (Fig. 6 and Table 2).
The mol. wt. o f A V F was determined f r o m its electrophoretic mobility in S D S - p o l y acrylamide gel (7"5 ~ ) in relation to protein mol. wt. markers (Fig. 7)- Because surface
phosphate groups contribute to its mobility, A V F was dephosphorylated with alkaline
phosphatase prior to electrophoresis. The mol. wt. o f A V F was determined as about 22oo0
(it migrates as a I 5 o o o mol. wt. protein when phosphorylated). However, it should be noted
that the above mol. wt. determination assumed a regular log-linear relationship between
mol. wt. and electrophoretic mobility which is still uncertain for glycoproteins.
Due to the difficulty in quantifying plant virus infectivity, it was not practical to assign
activity units to AVF. However, an estimate o f the extent o f purification can be assessed
from the following figures: I kg o f fresh, TMV-infected leaves yield about 4oo m g o f crude
A V F which in turn give a b o u t 4o mg o f DEAE-eluted A V F , and finally about I mg o f con
A-eluted AYF.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 19:08:53
Chemical nature of A VF
247
T a b l e 2. Antiviral activity o f con A-eluted A VF preparation following enzyme treatment
Protection ( ~ )
t
Treatment
No enzyme
Alkaline phosphatase
c~-Glucosidase
c~-Glucosidase and alkaline phosphatase
Expt. i
Expt. 2
44*
6I*
I7"
5
75*
75*
O
o
* Significance < I ~ (t-test for related samples).
Alkaline
phosphatase
(a)
7
-d
(c)
LYZ ~X
!
0
20
40
Distance from top (ram)
60
0
15
30
Distance from top (ram)
45
Fig. 7
Fig. 6
Fig. 6. Electropherograms of gels stained with Coomassie blue. The AVF preparation had been
treated with alkaline phosphatase (a), with ~-glucosidase (b) and with both enzymes combined (c).
Fig. 7. Determination of the mol. wt. of dephosphorylated AVF in SDS-polyacrylamide gels. Mol.
wt. markers were: reduced bovine serum albumin (BSA; 68 ooo), pepsin (PEP; 35 ooo) and lysozyme
(LYZ; I4ooo).
Con A-eluted AVF (0"5 mg) and its respective mock preparation were incubated in 5 ml
of o.I M-KC1-HC1, p H z-o, for I2 h at 4 °C, and then dialysed against I 1 of o-or M-trisHC1, p H 7"6, with three changes. The activity of a series of dilutions of the p H z.o-treated
AVF was tested. AVF remained active even when diluted a million times to give o-I ng
(4 x io -x~ M). The molar concentration of TMV, against which the AVF was protecting,
was Iooooo times higher (4 x io -7 M). This rules out any possibility that AVF neutralizes
T M V in a stoichiometric manner.
DISCUSSION
AVF was shown to stain like a protein in polyacrylamide gels and to be sensitive to
pronase under certain conditions. However, AVF also stains like a saccharide, binds to
concanavalin A and is sensitive to ~-glucosidase. Hence, AVF is most likely to be a glycoprotein. There are many indications that phosphorus is also associated with AVF." it can
be labelled with radioactive phosphorus, is eluted from DEAE-cellulose with o'55 u-NaC1
and can be extracted with phenol-like phosphoproteins (Lai, I976). Moreover, alkaline
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 19:08:53
248
R. MOZES AND OTHERS
phosphatase decreases the electrophoretic mobility of AVF. However, at least surface
phosphate groups are not essential for antiviral activity. The tool. wt. of unphosphorylated
AVF is about 22ooo.
The above cumulative data almost precisely fit the description of a possibly related set of
substances: the interferons. Interferons are also assumed to be glycoproteins and the
evidence for this was obtained indirectly, essentially through similar methodology as employed in the present paper (Grossberg, I972; Dorner, Scriba & Weil, I973). Many interferon preparations were found to be heterogeneous upon polyacrylamide gel electrophoresis
(Schonne, Billian & DeSomer, I97o; Weil & Dorner, I973; Stewart, I974; Reynolds &
Pitha, I975). This is probably due to different levels of glycosylation (Schonne et al. I97O;
Dorner et al. 2973) or to aggregation with a number of other proteins (Reynolds & Pitha,
I975). It also seems that by removing glycosylated end-groups, the homogeneity of the
interferon preparation is improved without a total loss in activity (Dorner et al. I973),
which resembles the present observation with AVF and oc-glucosidase. The tool. wt. of
many interferons fall in the range of 20oo0 to 28ooo with some observations of a minor
component of mol. wt. about I5OOO (Stewart et al. I977). Interferon also withstands treatment with SDS and is stable at pH 2.o.
The chemical nature of AVF and its antiviral activity clearly resembles interferon. It is
very possible that the role of AVF in plants is analogous to that of interferon in animals.
However, identifying AVF with interferon would be somewhat premature. The detailed
structure of AVF can be different from that of interferon, and the fine molecular processes
by which AVF gives protection against viruses in plants can also be different from animal
systems. It should also be noted that AVF is associated with the phenomenon of virus
confinement within local lesions in plants, a reaction which is not common in virus infections in animals.
This work has been supported in part by a grant from the Central Research Fund of the
Authority for Research and Development o f the Hebrew University, and in part by the BenGurion Foundation. Miss Rita Mozes is a recipient of a Casali Scholarship.
REFERENCES
AMENTA, J. S. (I964). A rapid chemical method for quantification of lipids separated by thin layer chromatography. Journal of Lipid Research 5, 270-272.
ANTIGNUS, Y., SELA, I. & HARPAZ, I. (1975). A phosphorus-containing fraction associated with antiviral
activity in Nicotiana spp. carrying the gene for localization of TMV infection. Physiological Plant
Pathology 6, I59-I68.
ANTIGNUS, V., SELA,I. & HARPAZ,I. (I977)- Further studies on the biology of an antiviral factor (AVF) from
virus-infected plants and its association with the N-gene of Nicotiana species. Journal t~fGeneral Virology
35, IO7-I ~6.
BIER, B. M. (I955). Lipases. In Methods in Enzymology, vol. I, pp. 627-642. Edited by S. P. Colowick and
N. O. Kaplan. London & New York: Academic Press.
DORNER, F., SCRtaA, U. & WEIL, R. (1973). Interferon: evidence for glycoprotein nature. Proceedings of the
National Academy of Sciences of the United States of America 7o, I981-1985.
FAIRBANKS, G., STECK, T. L. & WALLACH, D. F. H. (I97I). Electrophoretic analysis of the major polypeptides
of the human erythrocyte membrane. Biochemistry xo, 26o6-2619.
GROSSBERG, S.E. 0972). The interferons and their inducers. Molecular and therapeutic considerations.
New England Journal of Medicine z87, 13-I9.
HASSID, W. Z. & NEUEELD,E. F. 0964). Quantitative determination of starch in plant tissues. In Methods in
Carbohydrate Chemistry, vol. IV, pp. 33-36. Edited by R. L. Whistler. London & New York: Academic
Press.
LAI, M. M. C. (I976). Phosphoproteins of Rous sarcoma viruses. Virology 74, 287-3oi.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 19:08:53
Chemical nature of A VF
249
LOWRY, O. H., ROSEBROUGH, N. J., FARR, A. L. & RANDALL, R. J. 0 9 5 I ) . Protein measurement with the Folin
phenol reagent. Journal of Biological Chemistry I93, 265-275.
MATSUBARA, H., HASEGAWA, S., FUJIMURA, S., SHIMA, T. & SUGIMURA, Y. 0 9 7 0 ) . S t u d i e s o n p o l y ( a d e n o s i n e
diphosphate ribose). IX. Direction of hydrolysis of poly (adenosine diphosphate ribose) by rat liver
phosphodiesterase. Journal of Biological Chemistry 245, 4317-4320.
E'HILLIPS, D. V. & SMITH, A. E. (I973). A rapid method for gas chromatographic analysis of mono- and disaccharide mixtures. Analytical Biochemistry 54, 95-Ioi.
REYNOLDS, F. H. JUN. & PITHA, P.M. (I975). Molecular weight study of human fibroblast interferon. Biochemical and Biophysical Research Communications 65, Io7-112.
SCHONNE, E., BILLIAN,A. & DeSOMER, P. (197O). The properties of interferon. IV. Isoelectric focusing of rabbit
interferon (NDV-RKI3). In International Symposium on Standardisation of Interferon and Interferon
Inducers, pp. 6t-68. Edited by F. T. Perkins and R. H. Regamey. Basel & New York: Karger.
SELA, I., HARPAZ, I. & BIRK, V. 0966). Identification of the active component of an antiviral factor from virus
infected plants. Virology 28, 7I-78.
SPIRO, R.G. 0966). Characterization of carbohydrate units of glycoproteins. In Methods in Enzymology,
vol. VIII, pp. 26-52. Edited by S. P. Colowick and N. O. KapIan. London & New York: Academic
Press.
STEWART, W. E. 0974). Distinct molecular species of interferons. Virology 6i, 80-86.
STEWART, W. E., LIN, L., CHUDZIO, T. & WIRANOWSKA-STEWART, M. 0977). H u m a n leukocyte interferons:
analysis of size and charge heterogeneities by SDS-PAGE, ion exchange chromatography, isoelectric
focusing, and two-dimensional gel electrophoresis. Fifth Aharon Katzir-Katchalsky Conference, May
2-6, Rehovot, Israel (abstract).
WEIL, g. & DORNER, r. (1973). Interferon structure: facts and speculation. In Selective Inhibitors of Viral
Functions, pp. Io7-IZ2. Edited by W. A. Carter. Cleveland: CRC Press.
(Received I4 June I 9 7 7 )
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 19:08:53