- Email: [email protected]
Tobacco mosaic virus coat protein and the large subunit of the host protein ribulose-1,5-bisphosphate carboxylase share a common antigenic determinant
155.262-266
VIROLOGY
Tobacco
(1986)
Mosaic Virus 15Bisphosphate
Coat
Protein and the Large Subunit of the Host Protein Carboxylase Share a Common Antigenic Determinant
RALF Department
G. DIETZGEN’
of Plant Received
AND
MILTON
Pathology, CorneU Uniw-rsity, April
14, 1986; accepted
August
Ribulose-
ZAITLIN’ Ithaca,
New
York
l&T13
12, 1986
An immunological relationship was detected between the coat protein of the common (U,) strain of tobacco mosaic virus (TMV) and the large subunit of the ubiquitous CO,fixing host enzyme, ribulose-l&bisphosphate carboxylase (RuBisCo). When assayed by Western immunoblotting or indirect ELISA, polyclonal antisera to TMV coat protein and to RuBisCo reacted with both antigens. In addition, a monoclonal antibody specific for the C-terminal antigenie determinant of TMV coat protein reacted with RuBisCo. Conversely, several monoclonal antibodies generated to the large subunit of RuBisCo reacted with TMV coat protein. This cross-reactivity was verified by an examination of the amino acid sequences of both proteins. A region of homology was found between the carboxy proximal portion of coat protein and the sequence 60-73 residues from the amino terminus of RuBisCo large subunit. This homology was not mirrored at the nucleic acid level because of different codon usages for the two proteins. 0 1986 Academic Press, Inc.
The enzyme ribulose-l,&bisphosphate carboxylase (RuBisCo) catalyzes the photosynthetic fixation of carbon dioxide. It consists of eight large and eight small subunits with molecular weights of approximately 53,000 and 12,000, respectively (1, 2). In higher plants the large subunit, in which the amino acid sequence is highly conserved among different species, is encoded in the chloroplast genome (I). RuBisCo is the most prevalent plant protein in the world; it can be a troublesome impurity contaminating plant virus preparations (3). In the course of the immunological characterization of the H protein of TMV (4), an unexpected serological crossreactivity was observed between the large subunit of RuBisCo and the coat protein of the common (U,) strain of TMV. To investigate this cross-reactivity fur’ Present address: Department of Plant Pathology, Waite Agricultural Research Institute, University of Adelaide, Glen Osmond, South Australia 6064, Australia. ‘TO whom requests for reprints should be addressed: Department of Plant Pathology, Cornell University, Ithaca, N.Y. 14853. 0042-6822/86 Copyright All rights
$3.00
@ 1986 by Academic Press, Inc. of reproduction in any form reserved.
262
ther, several polyclonal and monoclonal antibody preparations were analyzed: (a) rabbit antiserum to RuBisCo holoenzyme from wheat (a nonhost for the common strain of TMV); (b) rabbit antiserum to UiTMV; (c) mouse monoclonal antibody TMV95 (5), which is directed against the C-terminal antigenic determinant of TMV coat protein; this monoclonal antibody which belonged to the IgG2a subclass had been produced in ascitic fluid, precipitated by 50% ammonium sulfate, and finally purified by affinity chromatography on protein ASepharose; (d) hybridoma supernatants containing monoclonal antibodies to the large subunit of RuBisCo from pea chloroplasts (6). Western immunoblotting analysis using antibodies both to TMV coat protein and to RuBisCo clearly demonstrated a reciprocal serological cross-reactivity between the two respective antigens (Fig. 1). Monoclonal antibody TMV-95 reacted with TMV coat protein (mol wt 17,500) and with the large subunit of RuBisCo to a similar extent (lanes 4 and 5). The darker band in lane 5 compared to lane 4 was due to the application of twice the molar amount of
SHORT
COMMUNICATIONS
Icp 4RS
FIG. 1. Cross-reactivity of antibodies to both TMV coat protein and the large subunit of RuBisCo, as analyzed by Western immunoblotting. Purified preparations of spinach RuBisCo (40 pg; lanes 3 and 5) and Ui-TMV coat protein (36 cg; lanes 2 and 4), and clarified extract of TMV-infected tobacco leaves (lane 1) were fractionated by 12% polyacrylamide-SDS gel electrophoresis (7). Proteins were transferred electrophoretically to nitrocellulose filters (8). Residual binding sites were blocked by incubation in a 3% bovine serum albumin-containing buffer. Immobilized antigens were reacted with (A) anti-RuBisCo polyclonal antiserum or (B) anti-TMV coat protein monoclonal antibody TMV-95 (5) followed by an amplification step using 1:loOO diluted goat anti-rabbit IgG in (A) and goat anti-mouse IgG in (B). The immune complexes were reacted with ?-labeled protein A and visualized by exposing the nitrocellulose to preflashed Kodak X-ray film for 5 hr at -80’ using an intensifying screen. RL = RuBisCo large subunit; R, = RuBisCo small subunit; CP = TMV Coat Protein.
the large subunit of RuBisCo over TMV coat protein to the gel in the respective lanes. Conversely, polyclonal antiserum to RuBisCo reacted strongly with the large subunit of RuBisCo (lane 3), but less strongly with the viral coat protein (lanes 1 and 2). Nonspecific reactivity of antibodies due to interactions of their carbohydrate side chains with plant lectins was ruled out, since identical results to those seen in Fig. 1 were obtained when the carbohydrates were removed from the antibodies beforehand. [Glycosyl groups were cleaved by treatment with lysozyme and
263
the enzyme subsequently inactivated by addition of copper sulfate (II).] Rabbit preimmune serum and affinity purified mouse IgG were used as controls. To underscore that TMV coat protein and the large subunit of RuBisCo share an antigenie determinant, monoclonal antibodies directed against the large subunit of RuBisCo were screened for cross-reactivity. In a collection of such antibodies (kindly provided by Dr. Kenneth Keegstra), one might expect to find monoclonal antibodies which also react with the homologous antigenic determinant on TMV coat protein. Indeed, 2 of 23 monoclonal antibodies assayed (which were of the IgG class) reacted with both antigens contained in TMV-infected tobacco extracts in Western immunoblotting (data not shown). In direct ELISA (5), the results obtained by immunoblotting were confirmed. Purified preparations of U1-TMV coat protein, spinach RuBisCo (Sigma) and the large subunit of tobacco RuBisCo were used at concentrations of 10 pg/ml to coat microtiter plates (Immulon 2, Dynatech, Alexandria, Va.). After blocking residual binding sites on the plastic surface, a dilution series of antibodies was applied. Bound antibodies were detected using 1:lOOO dilutions of goat anti-rabbit or goat antimouse IgG conjugated to alkaline phosphatase (Zymed Laboratories, San Francisco, Calif.) with 1 mg/ml pnitrophenyl phosphate (Sigma) as substrate. Polyclonal antiserum made to RuBisCo from the nonTMV host wheat, and Ur-TMV antiserum both reacted strongly with their respective immunogens; however, they also showed cross-reactivity which was more than 10 times lower than the respective homologous reaction (Table 1). This agrees with our observations by Western immunoblotting (Fig. 1). Low cross-reactivity is not unexpected because only a small proportion of the antibody population in the polyclonal antisera would be directed against the cross-reacting epitope of either antigen. On the other hand, monoclonal antibody TMV95 exhibited reactivity of similar strength with both TMV coat protein and the large subunit of RuBisCo (Table l), indicating that it is directed against the proposed
SHORT
264
COMMUNICATIONS TABLE
CROSS-REACTIVITY OF ANTIBODIES OF RIBLJLOSE-1,5-BISPHOSPHATE Rabbit
Ui-TMV coat protein RuBisCo large subunit (tobacco) Bovine serum albumin 0 Values are given as the solution for 30 min at 20”. Each antigen was assayed employed at approximately
1
TO TMV COAT PROTEIN AND TO THE LARGE SUBUNIT CARBOXYLASE ANALYZED BY INDIRECF ELISA” antiserum
to
Monoclonal
antibodies
U,-TMV coat protein
II,-TMV coat protein
Wheat RuBisCo
>1.506
0.093 f 0.035
0.396 f 0.071
0.027 f 0.001
0.227 k 0.006 0.013 + 0.001
>1.500 0.005 Ik 0.001
0.636 + 0.022 0.018 f 0.003
0.210 + o.ooo 0.000
Tobacco RuBisCo
absorbance at 405 nm which was measured after incubation with the substrate The average and standard deviation from three independent repeats is shown. with identical concentrations of antibodies. The large subunit of RuBisCo was twice the molar amount of TMV coat protein. Antigens were used at 10 fig/ml.
shared antigenic determinant. Controls which did not exhibit any reactivity included rabbit preimmune serum and a monoclonal antibody to potato virus A (Inotech Diagnostik, Basel, Switzerland). A monoclonal antibody to the large subunit of tobacco RuBisCo, directed to an antigenie determinant different from the crossreactive one (unpublished results), reacted with the large subunit of RuBisCo, but not with TMV coat protein when assayed by ELISA (Table 1) or by Western immunoblotting (data not shown), The amino acid sequences of Ui-TMV coat protein (12) and the large subunit of RuBisCo (2) were compared by computerassisted analysis using the Beckman MicroGenie Sequence Software (13). Both proteins were found to share two closely adjacent sequences of five consecutive amino acids each (Table 2). The C-terminal TABLE
sequence of TMV coat protein is known to represent an antigenic determinant (14). Monoclonal antibody TMV-95 is directed against this region, as it was selected by using a synthetic peptide representing a part of this antigenic determinant (5). The serological cross-reactivity observed between Ui-TMV coat protein and the large subunit of RuBisCo is a consequence of an amino acid sequence homology in immunogenic regions of both proteins. We determined that both proteins have similar secondary structure (15) and hydrophilicity profiles (16) in this region. The homology is not mirrored at the nucleic acid level as the respective nucleotide sequences for the two proteins did not show strong homology, a consequence of differences in codon usage; the longest homology is six nucleotides (Table 2). We also tested coat proteins from several 2
COMPARISON OF PARTIAL AMINO ACID SEQUENCES AND RESPECTIVE NUCLEOTIDE SEQUENCES COAT PROTEIN AND THE LARGE SUBUNIT OF RIB~L~~E-~,~-BI~PH~~PHATE CARBOX~LASE Ui-TMV RuBisCo
to
‘&Glu --@‘m
Ser
Ser
Ser
Gly”’
“im
&
&
Tbr
Gly”
@VV
BETWEEN TMV (RuBisCo)
Trp Trp -
Thr
Ser
GlyiS6
Thr
Asp
Gly”
UI-TMV
GAG
AGC
UCU
IJCU GGU
GUU
UGG
ACC UCIJ
GGIJ
RuBisCo”
-A
UCU
UCU
ACU
=A
UGG
ACC GAIJ
=A
GGU
o The nucleotide sequence segment of the large subunit of RuBisCo nucleotides homologous between both proteins are underlined.
is presented
as RNA.
Amino
acids
or
SHORT
TABLE COMPARISON OF THE C-TERMINAL AMINO TMV STRAINS WITH A SELEIXED
Protein
265
COMMUNICATIONS 3
ACID SEQUENCES OF THE COAT PROTEINS OF SEVERAL REGION OF THE LARGE SUBUNIT OF RuBisCo
Amino
Proportional homology
acid sequences
RuBisCo
“Glu
Ser Ser -
Thr
Gly -
&l
-Trp Thr
Asp
Glyn’
I&-TMV
‘“Glu --ifiGlu
Ser Ser
Ser Gly
Leu &l
Trp T&
Ser
3’S
Thr Ala
Ser ?$
‘166111 Ser Met -“&Glu Ser Ala --
Ser cy
Leu -Val Trp -Thr Ser Th?’ Leu -Val G Thr Ser Ala’% Leu Trp G Val Thr Thr’”
Ua-TMV Dahlemense Cowpea
Ser -?$
Note. Amino acids homologous between TMV coat proteins and the large subunit of RuBisCo Amino acid sequences for strains Uz and dahlemense are from (9), cowpea from (10).
8/10 5/10 6/10 3/10 are underlined.
strains of TMV against the cross-reacting quences (23), our observed homology might monoclonal antibody (TMV 95). We found actually be expected. A common sequence faint reactions (compared with the reaction of six amino acids seems to be sufficient to of Ui coat protein) with strains Uz and induce immunological cross-reactivity (24). dahlemense but not with the cowpea strain A sequence of four amino acids has been on Western immunoblots (not shown). found to be sufficient for defining an antiAmino acid sequence comparison between genie determinant on a protein (25). It has the coat proteins of these strains and the been demonstrated by amino acid sequence RuBisCo large subunit show a more limited comparisons and monoclonal antibodies homology than that seen between U1 and that many animal and human viruses share RuBisCo (Table 3). Since we do not know antigenic determinants with host cell conthe conformation of the coat proteins of stituents (24). This phenomenon might be these strains, we are unable to ascertain connected to the occurrence of autoimmune which part of this region might be reac- diseases (w), a situation which of course tive-if indeed it is this region at all. does not apply here. To our knowledge, this The coat protein sequences of other plant is the first report of molecular mimicry beviruses contained in the databank (IS), al- tween a plant virus and a protein of its falfa mosaic virus (Jr), brome mosaic virus host. For the moment we have no data (18), cowpea chlorotic mottle virus (18), which allow us to postulate a particular cucumber green mottle mosaic virus (19), function for this homology. It may be of significance, however, that a small amount southern bean mosaic virus (20), tobacco streak virus (21), and turnip yellow mosaic of TMV coat protein-normally synthevirus (TYMV, 22) did not exhibit amino acid sized on cytoplasmic ribosomes-is found homologies to the TMV-homologous region in the chloroplast (26) where it is a comof the large subunit of RuBisCo. In the case ponent of pseudovirions. These virus-like tranof TYMV, lack of cross-reactivity was rods are composed of chloroplast confirmed by testing by Western imscripts encapsidated in TMV coat protein munoblotting with RuBisCo polyclonal (27). How does TMV coat protein find its antiserum. way into the chloroplast, or is it syntheIt is unexpected that the coat protein of sized there? TMV and the major protein of its host would share an antigenic determinant. ACKNOWLEDGMENTS Therefore, the observed homology might be the result of mere chance. Assuming This paper is dedicated to Professor S. G. Wildman that even unrelated proteins would have who in his career made significant contributions to of both RuBisCo and TMV, to some an average of 5% identical amino acid se- our understanding
266
SHORT COMMUNICATIONS
degree these separate efforts have now come together. We thank Beth Hammer and Dr. Kenneth Keegstra, University of Wisconsin, for generously supplying monoclonal antibodies to RuBisCo, Dr. Andre Jagendorf, Cornell University, for the gift of polyclonal antiserum to RuBisCo, and Dr. Ed Halk (Agrigenetics Inc.) and Dr. Sandor Pongor (Boyce Thompson Institute for Plant Research) for valuable discussions. Generous preparations of the large subunit of tobacco RuBisCo was a kind gift from Dr. Don Bourque, University of Arizona. R.D. was supported by a Fellowship of the Scientific Committee of the North Atlantic Treaty Organization by way of the German Academic Exchange Service. This work was supported in part by Grant 8500279 from the Competitive Grants Program of the United States Department of Agriculture and Grant 84-09851 from the National Science Foundation. REFERENCES 1. AKAZAWA, T., TAKEBE, T., and KOBAYASHI, H., TIBS 9,380-383 (1984). 2. SHINOZAKI, K., and SUGIURA, M., Gene 20,91-102 (1982). 3. VAN REGENMORTEL, M. H. V., In “Serology and Immunochemistry of Plant Viruses,” pp. 2932. Academic Press, New York, 1982. .4. COLLMER, C. W., VOGT, V. M., and ZAITLIN, M., virology 126,429-448 (1983). 5. DIETZGEN, R. G., Arch, Vid 87.73-86 (1986). 6. WERNER-WASHBURNE, M., CLINE, K., and KEEGSTRA, K., Plant Physd 73,569-575 (1983). 7. LAEMMLI, U. K., Nature (London) 277, 680-685 (1970).
8. TOWBIN, H., and GORDON,J., J. Immunol 72.313340 (1984). ~ ,
Methods
9. WIITMANN-LIEBOLD, B., and WI~ANN, H. G., Md Gm. Genet. 100,358-363 (1967). 10. REES, M. W., and SHORT, M. N., B&him. Biophvs. Actu 393,15-23 (1975). 11. ESSINK, A. W. G., ARKESTEIJN, G. J. M. W., and NOTERMANS, S., J. Immunol. Methods 80.91-96 (1985). 12. TSUGITA, A., GISH, D. T., FRAENKEL-CONRAT, H., KNIGHT, C. A., and STANLEY, W. M., Proc Nat1 Acad Sci. USA 46,1463-1469 (1960). 13. QUEEN, C., and KORN, L. J., NuckicAci& Res. 12, 581-599 (1984). 14. ANDERER, F. A., and SCHLUMBERGER, H. D., Biochim Biophys. Ada 97,503-509 (1965). 15. GARNIER, J., OSGUTHORPE,D. J., and ROBSON,B., J. Mol. Biol. 120,97-120 (1978). 16. HOPP, T. P., and WOODS, K. R., Proc. Natl. Ad Sci. USA 78,3824-3828 (1981). 17. BREDERODE,F. T., KOPER-ZWARTHOFF, E. C., and BOL, J. F., Nucleic Acids Rea 8,2213-2223 (1980). 18. DASGUPTA, R., and KAESBERG, P., Nucleic Acids Res. 10,703-713 (1982). 19. MESHI, T., KIYAMA, R., OHNO, T., and OKADA, Y., Vim&/~ 127,54-64 (1983). 20. MANG, K., GHOSH, A., and KAESBERG, P., virdogy 116,264-274 (1982). 21. CORNELISSEN,B. J. C., JANSEEN, H., ZUIDEMA, D., and BOL, J. F., Nucleic Acids Res. 12,2427-2437 (1984). 22. GUILLEY, H., and BRIAND, J. P., Cell 15, 113-122 (1978). 23. DOOLITTLE, R. F., Science 214,149-159 (1981). 2k FUJINAMI, R. S., and OLDSTONE, M. B. A., Science 230,1043-1045 (1985). 25. SELA, M., Science 166,1365-1374 (1969). 26. REINERO, A., and BEACHY, R. N., Plant Mol Biol 6,291-301 (1986). 27. ROCHON,D., and SIEGEL, A., Proc. Natl Ad Sci USA 81,1’719-1723 (1984).
VIROLOGY
Tobacco
(1986)
Mosaic Virus 15Bisphosphate
Coat
Protein and the Large Subunit of the Host Protein Carboxylase Share a Common Antigenic Determinant
RALF Department
G. DIETZGEN’
of Plant Received
AND
MILTON
Pathology, CorneU Uniw-rsity, April
14, 1986; accepted
August
Ribulose-
ZAITLIN’ Ithaca,
New
York
l&T13
12, 1986
An immunological relationship was detected between the coat protein of the common (U,) strain of tobacco mosaic virus (TMV) and the large subunit of the ubiquitous CO,fixing host enzyme, ribulose-l&bisphosphate carboxylase (RuBisCo). When assayed by Western immunoblotting or indirect ELISA, polyclonal antisera to TMV coat protein and to RuBisCo reacted with both antigens. In addition, a monoclonal antibody specific for the C-terminal antigenie determinant of TMV coat protein reacted with RuBisCo. Conversely, several monoclonal antibodies generated to the large subunit of RuBisCo reacted with TMV coat protein. This cross-reactivity was verified by an examination of the amino acid sequences of both proteins. A region of homology was found between the carboxy proximal portion of coat protein and the sequence 60-73 residues from the amino terminus of RuBisCo large subunit. This homology was not mirrored at the nucleic acid level because of different codon usages for the two proteins. 0 1986 Academic Press, Inc.
The enzyme ribulose-l,&bisphosphate carboxylase (RuBisCo) catalyzes the photosynthetic fixation of carbon dioxide. It consists of eight large and eight small subunits with molecular weights of approximately 53,000 and 12,000, respectively (1, 2). In higher plants the large subunit, in which the amino acid sequence is highly conserved among different species, is encoded in the chloroplast genome (I). RuBisCo is the most prevalent plant protein in the world; it can be a troublesome impurity contaminating plant virus preparations (3). In the course of the immunological characterization of the H protein of TMV (4), an unexpected serological crossreactivity was observed between the large subunit of RuBisCo and the coat protein of the common (U,) strain of TMV. To investigate this cross-reactivity fur’ Present address: Department of Plant Pathology, Waite Agricultural Research Institute, University of Adelaide, Glen Osmond, South Australia 6064, Australia. ‘TO whom requests for reprints should be addressed: Department of Plant Pathology, Cornell University, Ithaca, N.Y. 14853. 0042-6822/86 Copyright All rights
$3.00
@ 1986 by Academic Press, Inc. of reproduction in any form reserved.
262
ther, several polyclonal and monoclonal antibody preparations were analyzed: (a) rabbit antiserum to RuBisCo holoenzyme from wheat (a nonhost for the common strain of TMV); (b) rabbit antiserum to UiTMV; (c) mouse monoclonal antibody TMV95 (5), which is directed against the C-terminal antigenic determinant of TMV coat protein; this monoclonal antibody which belonged to the IgG2a subclass had been produced in ascitic fluid, precipitated by 50% ammonium sulfate, and finally purified by affinity chromatography on protein ASepharose; (d) hybridoma supernatants containing monoclonal antibodies to the large subunit of RuBisCo from pea chloroplasts (6). Western immunoblotting analysis using antibodies both to TMV coat protein and to RuBisCo clearly demonstrated a reciprocal serological cross-reactivity between the two respective antigens (Fig. 1). Monoclonal antibody TMV-95 reacted with TMV coat protein (mol wt 17,500) and with the large subunit of RuBisCo to a similar extent (lanes 4 and 5). The darker band in lane 5 compared to lane 4 was due to the application of twice the molar amount of
SHORT
COMMUNICATIONS
Icp 4RS
FIG. 1. Cross-reactivity of antibodies to both TMV coat protein and the large subunit of RuBisCo, as analyzed by Western immunoblotting. Purified preparations of spinach RuBisCo (40 pg; lanes 3 and 5) and Ui-TMV coat protein (36 cg; lanes 2 and 4), and clarified extract of TMV-infected tobacco leaves (lane 1) were fractionated by 12% polyacrylamide-SDS gel electrophoresis (7). Proteins were transferred electrophoretically to nitrocellulose filters (8). Residual binding sites were blocked by incubation in a 3% bovine serum albumin-containing buffer. Immobilized antigens were reacted with (A) anti-RuBisCo polyclonal antiserum or (B) anti-TMV coat protein monoclonal antibody TMV-95 (5) followed by an amplification step using 1:loOO diluted goat anti-rabbit IgG in (A) and goat anti-mouse IgG in (B). The immune complexes were reacted with ?-labeled protein A and visualized by exposing the nitrocellulose to preflashed Kodak X-ray film for 5 hr at -80’ using an intensifying screen. RL = RuBisCo large subunit; R, = RuBisCo small subunit; CP = TMV Coat Protein.
the large subunit of RuBisCo over TMV coat protein to the gel in the respective lanes. Conversely, polyclonal antiserum to RuBisCo reacted strongly with the large subunit of RuBisCo (lane 3), but less strongly with the viral coat protein (lanes 1 and 2). Nonspecific reactivity of antibodies due to interactions of their carbohydrate side chains with plant lectins was ruled out, since identical results to those seen in Fig. 1 were obtained when the carbohydrates were removed from the antibodies beforehand. [Glycosyl groups were cleaved by treatment with lysozyme and
263
the enzyme subsequently inactivated by addition of copper sulfate (II).] Rabbit preimmune serum and affinity purified mouse IgG were used as controls. To underscore that TMV coat protein and the large subunit of RuBisCo share an antigenie determinant, monoclonal antibodies directed against the large subunit of RuBisCo were screened for cross-reactivity. In a collection of such antibodies (kindly provided by Dr. Kenneth Keegstra), one might expect to find monoclonal antibodies which also react with the homologous antigenic determinant on TMV coat protein. Indeed, 2 of 23 monoclonal antibodies assayed (which were of the IgG class) reacted with both antigens contained in TMV-infected tobacco extracts in Western immunoblotting (data not shown). In direct ELISA (5), the results obtained by immunoblotting were confirmed. Purified preparations of U1-TMV coat protein, spinach RuBisCo (Sigma) and the large subunit of tobacco RuBisCo were used at concentrations of 10 pg/ml to coat microtiter plates (Immulon 2, Dynatech, Alexandria, Va.). After blocking residual binding sites on the plastic surface, a dilution series of antibodies was applied. Bound antibodies were detected using 1:lOOO dilutions of goat anti-rabbit or goat antimouse IgG conjugated to alkaline phosphatase (Zymed Laboratories, San Francisco, Calif.) with 1 mg/ml pnitrophenyl phosphate (Sigma) as substrate. Polyclonal antiserum made to RuBisCo from the nonTMV host wheat, and Ur-TMV antiserum both reacted strongly with their respective immunogens; however, they also showed cross-reactivity which was more than 10 times lower than the respective homologous reaction (Table 1). This agrees with our observations by Western immunoblotting (Fig. 1). Low cross-reactivity is not unexpected because only a small proportion of the antibody population in the polyclonal antisera would be directed against the cross-reacting epitope of either antigen. On the other hand, monoclonal antibody TMV95 exhibited reactivity of similar strength with both TMV coat protein and the large subunit of RuBisCo (Table l), indicating that it is directed against the proposed
SHORT
264
COMMUNICATIONS TABLE
CROSS-REACTIVITY OF ANTIBODIES OF RIBLJLOSE-1,5-BISPHOSPHATE Rabbit
Ui-TMV coat protein RuBisCo large subunit (tobacco) Bovine serum albumin 0 Values are given as the solution for 30 min at 20”. Each antigen was assayed employed at approximately
1
TO TMV COAT PROTEIN AND TO THE LARGE SUBUNIT CARBOXYLASE ANALYZED BY INDIRECF ELISA” antiserum
to
Monoclonal
antibodies
U,-TMV coat protein
II,-TMV coat protein
Wheat RuBisCo
>1.506
0.093 f 0.035
0.396 f 0.071
0.027 f 0.001
0.227 k 0.006 0.013 + 0.001
>1.500 0.005 Ik 0.001
0.636 + 0.022 0.018 f 0.003
0.210 + o.ooo 0.000
Tobacco RuBisCo
absorbance at 405 nm which was measured after incubation with the substrate The average and standard deviation from three independent repeats is shown. with identical concentrations of antibodies. The large subunit of RuBisCo was twice the molar amount of TMV coat protein. Antigens were used at 10 fig/ml.
shared antigenic determinant. Controls which did not exhibit any reactivity included rabbit preimmune serum and a monoclonal antibody to potato virus A (Inotech Diagnostik, Basel, Switzerland). A monoclonal antibody to the large subunit of tobacco RuBisCo, directed to an antigenie determinant different from the crossreactive one (unpublished results), reacted with the large subunit of RuBisCo, but not with TMV coat protein when assayed by ELISA (Table 1) or by Western immunoblotting (data not shown), The amino acid sequences of Ui-TMV coat protein (12) and the large subunit of RuBisCo (2) were compared by computerassisted analysis using the Beckman MicroGenie Sequence Software (13). Both proteins were found to share two closely adjacent sequences of five consecutive amino acids each (Table 2). The C-terminal TABLE
sequence of TMV coat protein is known to represent an antigenic determinant (14). Monoclonal antibody TMV-95 is directed against this region, as it was selected by using a synthetic peptide representing a part of this antigenic determinant (5). The serological cross-reactivity observed between Ui-TMV coat protein and the large subunit of RuBisCo is a consequence of an amino acid sequence homology in immunogenic regions of both proteins. We determined that both proteins have similar secondary structure (15) and hydrophilicity profiles (16) in this region. The homology is not mirrored at the nucleic acid level as the respective nucleotide sequences for the two proteins did not show strong homology, a consequence of differences in codon usage; the longest homology is six nucleotides (Table 2). We also tested coat proteins from several 2
COMPARISON OF PARTIAL AMINO ACID SEQUENCES AND RESPECTIVE NUCLEOTIDE SEQUENCES COAT PROTEIN AND THE LARGE SUBUNIT OF RIB~L~~E-~,~-BI~PH~~PHATE CARBOX~LASE Ui-TMV RuBisCo
to
‘&Glu --@‘m
Ser
Ser
Ser
Gly”’
“im
&
&
Tbr
Gly”
@VV
BETWEEN TMV (RuBisCo)
Trp Trp -
Thr
Ser
GlyiS6
Thr
Asp
Gly”
UI-TMV
GAG
AGC
UCU
IJCU GGU
GUU
UGG
ACC UCIJ
GGIJ
RuBisCo”
-A
UCU
UCU
ACU
=A
UGG
ACC GAIJ
=A
GGU
o The nucleotide sequence segment of the large subunit of RuBisCo nucleotides homologous between both proteins are underlined.
is presented
as RNA.
Amino
acids
or
SHORT
TABLE COMPARISON OF THE C-TERMINAL AMINO TMV STRAINS WITH A SELEIXED
Protein
265
COMMUNICATIONS 3
ACID SEQUENCES OF THE COAT PROTEINS OF SEVERAL REGION OF THE LARGE SUBUNIT OF RuBisCo
Amino
Proportional homology
acid sequences
RuBisCo
“Glu
Ser Ser -
Thr
Gly -
&l
-Trp Thr
Asp
Glyn’
I&-TMV
‘“Glu --ifiGlu
Ser Ser
Ser Gly
Leu &l
Trp T&
Ser
3’S
Thr Ala
Ser ?$
‘166111 Ser Met -“&Glu Ser Ala --
Ser cy
Leu -Val Trp -Thr Ser Th?’ Leu -Val G Thr Ser Ala’% Leu Trp G Val Thr Thr’”
Ua-TMV Dahlemense Cowpea
Ser -?$
Note. Amino acids homologous between TMV coat proteins and the large subunit of RuBisCo Amino acid sequences for strains Uz and dahlemense are from (9), cowpea from (10).
8/10 5/10 6/10 3/10 are underlined.
strains of TMV against the cross-reacting quences (23), our observed homology might monoclonal antibody (TMV 95). We found actually be expected. A common sequence faint reactions (compared with the reaction of six amino acids seems to be sufficient to of Ui coat protein) with strains Uz and induce immunological cross-reactivity (24). dahlemense but not with the cowpea strain A sequence of four amino acids has been on Western immunoblots (not shown). found to be sufficient for defining an antiAmino acid sequence comparison between genie determinant on a protein (25). It has the coat proteins of these strains and the been demonstrated by amino acid sequence RuBisCo large subunit show a more limited comparisons and monoclonal antibodies homology than that seen between U1 and that many animal and human viruses share RuBisCo (Table 3). Since we do not know antigenic determinants with host cell conthe conformation of the coat proteins of stituents (24). This phenomenon might be these strains, we are unable to ascertain connected to the occurrence of autoimmune which part of this region might be reac- diseases (w), a situation which of course tive-if indeed it is this region at all. does not apply here. To our knowledge, this The coat protein sequences of other plant is the first report of molecular mimicry beviruses contained in the databank (IS), al- tween a plant virus and a protein of its falfa mosaic virus (Jr), brome mosaic virus host. For the moment we have no data (18), cowpea chlorotic mottle virus (18), which allow us to postulate a particular cucumber green mottle mosaic virus (19), function for this homology. It may be of significance, however, that a small amount southern bean mosaic virus (20), tobacco streak virus (21), and turnip yellow mosaic of TMV coat protein-normally synthevirus (TYMV, 22) did not exhibit amino acid sized on cytoplasmic ribosomes-is found homologies to the TMV-homologous region in the chloroplast (26) where it is a comof the large subunit of RuBisCo. In the case ponent of pseudovirions. These virus-like tranof TYMV, lack of cross-reactivity was rods are composed of chloroplast confirmed by testing by Western imscripts encapsidated in TMV coat protein munoblotting with RuBisCo polyclonal (27). How does TMV coat protein find its antiserum. way into the chloroplast, or is it syntheIt is unexpected that the coat protein of sized there? TMV and the major protein of its host would share an antigenic determinant. ACKNOWLEDGMENTS Therefore, the observed homology might be the result of mere chance. Assuming This paper is dedicated to Professor S. G. Wildman that even unrelated proteins would have who in his career made significant contributions to of both RuBisCo and TMV, to some an average of 5% identical amino acid se- our understanding
266
SHORT COMMUNICATIONS
degree these separate efforts have now come together. We thank Beth Hammer and Dr. Kenneth Keegstra, University of Wisconsin, for generously supplying monoclonal antibodies to RuBisCo, Dr. Andre Jagendorf, Cornell University, for the gift of polyclonal antiserum to RuBisCo, and Dr. Ed Halk (Agrigenetics Inc.) and Dr. Sandor Pongor (Boyce Thompson Institute for Plant Research) for valuable discussions. Generous preparations of the large subunit of tobacco RuBisCo was a kind gift from Dr. Don Bourque, University of Arizona. R.D. was supported by a Fellowship of the Scientific Committee of the North Atlantic Treaty Organization by way of the German Academic Exchange Service. This work was supported in part by Grant 8500279 from the Competitive Grants Program of the United States Department of Agriculture and Grant 84-09851 from the National Science Foundation. REFERENCES 1. AKAZAWA, T., TAKEBE, T., and KOBAYASHI, H., TIBS 9,380-383 (1984). 2. SHINOZAKI, K., and SUGIURA, M., Gene 20,91-102 (1982). 3. VAN REGENMORTEL, M. H. V., In “Serology and Immunochemistry of Plant Viruses,” pp. 2932. Academic Press, New York, 1982. .4. COLLMER, C. W., VOGT, V. M., and ZAITLIN, M., virology 126,429-448 (1983). 5. DIETZGEN, R. G., Arch, Vid 87.73-86 (1986). 6. WERNER-WASHBURNE, M., CLINE, K., and KEEGSTRA, K., Plant Physd 73,569-575 (1983). 7. LAEMMLI, U. K., Nature (London) 277, 680-685 (1970).
8. TOWBIN, H., and GORDON,J., J. Immunol 72.313340 (1984). ~ ,
Methods
9. WIITMANN-LIEBOLD, B., and WI~ANN, H. G., Md Gm. Genet. 100,358-363 (1967). 10. REES, M. W., and SHORT, M. N., B&him. Biophvs. Actu 393,15-23 (1975). 11. ESSINK, A. W. G., ARKESTEIJN, G. J. M. W., and NOTERMANS, S., J. Immunol. Methods 80.91-96 (1985). 12. TSUGITA, A., GISH, D. T., FRAENKEL-CONRAT, H., KNIGHT, C. A., and STANLEY, W. M., Proc Nat1 Acad Sci. USA 46,1463-1469 (1960). 13. QUEEN, C., and KORN, L. J., NuckicAci& Res. 12, 581-599 (1984). 14. ANDERER, F. A., and SCHLUMBERGER, H. D., Biochim Biophys. Ada 97,503-509 (1965). 15. GARNIER, J., OSGUTHORPE,D. J., and ROBSON,B., J. Mol. Biol. 120,97-120 (1978). 16. HOPP, T. P., and WOODS, K. R., Proc. Natl. Ad Sci. USA 78,3824-3828 (1981). 17. BREDERODE,F. T., KOPER-ZWARTHOFF, E. C., and BOL, J. F., Nucleic Acids Rea 8,2213-2223 (1980). 18. DASGUPTA, R., and KAESBERG, P., Nucleic Acids Res. 10,703-713 (1982). 19. MESHI, T., KIYAMA, R., OHNO, T., and OKADA, Y., Vim&/~ 127,54-64 (1983). 20. MANG, K., GHOSH, A., and KAESBERG, P., virdogy 116,264-274 (1982). 21. CORNELISSEN,B. J. C., JANSEEN, H., ZUIDEMA, D., and BOL, J. F., Nucleic Acids Res. 12,2427-2437 (1984). 22. GUILLEY, H., and BRIAND, J. P., Cell 15, 113-122 (1978). 23. DOOLITTLE, R. F., Science 214,149-159 (1981). 2k FUJINAMI, R. S., and OLDSTONE, M. B. A., Science 230,1043-1045 (1985). 25. SELA, M., Science 166,1365-1374 (1969). 26. REINERO, A., and BEACHY, R. N., Plant Mol Biol 6,291-301 (1986). 27. ROCHON,D., and SIEGEL, A., Proc. Natl Ad Sci USA 81,1’719-1723 (1984).