Melting of viral RNA by coat protein: Assembly strategies for elongated plant viruses

VXROLOGY

.

108, 225-240 (1981)

Melting of Viral RNA by Coat Protein: Assembly Strategies J. W. ERICKSON’ Department

of Plant

Sciences,

The

Plant Viruses

AND J. B. BANCROFT

University

Accepted

for Elongated

of Western

September

Ontario,

London,

Ontario, Canada

8, 1980

The coat proteins of two rod-shaped plant viruses, papaya mosaic virus (PMV) and tobacco mosaic virus (TMV), have been tested for RNA melting activity. Under reconstitution conditions, PMV protein melts RNA in a noneooperative fashion. This activity is aspecific and is inhibited by low concentrations of NaCl as is virus reconstitution. TMV protein does not melt RNA either in the absence of NaCl or under reconstitution conditions at moderate ionic strength levels. The results suggest that elongated plant viruses have evolved at least two different assembly strategies in order to satisfy the requirement that the RNA within these viruses be in a melted configuration.

A central problem for the self-assembly of RNA-containing elongated plant viruses is the drastic change in structure which the nucleic acid must undergo during assembly in vitro. The RNAs within elongated plant viruses are fully hyperchromic and do not exhibit any classical secondary structure (j-5), although under virus reconstitution conditions the viral RNAs in solution can have from 30-70% of their bases paired, depending on the ionic strength (1, S-9). The encapsidation of RNA by the viral coat protein during in vitro assembly thus requires that work be done in the form of rupturing, or melting, the hydrogen-bonded base pairs in the free RNA. The significance, in terms of energetics, that such a melting requirement might play in assembly can be obtained by comparing the average stability of a base pair involved in a double-stranded RNA segment, where AH hydrogen bonding = - 10 kcal/mole nucleotide (20 ) and AH stacking L- -5 to -8 kcal/mol nucleotide (11), with the calculated RNA-protein interaction stability term for TMV assembly, where AH = -3 kcal/mol nucleotide (12). Such a consideration need not be taken into account in proposed schemes for the selfassembly of spherical plant viruses, since 1 To whom reprint requests should be addressed. Present address: Department of Biological Sciences, Purdue University, West Lafayette, Ind. 4790’7. 235

their RNAs apparently undergo little net change in secondary structure during assembly (13-16). In this study, the coat proteins of two elongated viruses, papaya mosaic virus (PMV) and tobacco mosaic virus (TMV), were tested for the ability to melt the secondary structure of free viral RNA in solution. These viruses were chosen because of the widely different ionic strength requirements of their in vitro self-assembly reactions; TMV requiring high ionic strength (lr), and PMV being inhibited at NaCl concentrations above 0.02 M (6, 18). PMV, PMV coat protein, and PMV-RNA were prepared as before (18 ). The protein was concentrated using (NH,&KJ and stored in water at 5“ after extensive dialysis. TMV and TMV-RNA were obtained as described (19). TMV coat protein was prepared by the acetic acid method (.ti+O),The isoelectrically precipitated protein was dialyzed against and stored in 0.01 M Tris buffer, pH 8.5, at 5”. Soybean trypein inhibitor protein (STI) was a gift from Dr. R. B. van Huystee. It was dialyzed extensively against water prior to its use. To determine protein concentrations, after correction for light scattering, the following extinctions were E".'% 280nm= 0.75 for PMV protein (1); E$$$, = 1.27 for TMV protein (20); and E!$$&, = 0.94 for ST1 (X1). For the RNAs, Ei&9;:, = 25.0 in 0.10 M phos0042~6822/81/010235-06$02.00/0 Copyright D 1981 by Academic Press, Inc. All rights of reproduction in any form rm~wd.

236

SHORTCOMMUNICATIONS

phate buffer, pH 7.0, was assumed. RNA melting experiments were performed in a Pye-Unicam SP1800 dual beam spectrophotometer as before (6). For melting in the presence of protein, an equal weight of protein was added to both the reference and the RNA-containing sample cuvettes. Turbidity changes during the melting experiments both with and without protein were minimal. All RNA melting curves were fully reversible in the absence of protein. The melting of viral RNA during the in vitro assembly of an elongated virus can be envisaged as occurring in one of two principal ways: by an encapsidation-linked mechanism, in which melting is achieved locally and concomitant with encapsidation, or by an overall melting of the RNA prior to and independent of the encapsidation event. To distinguish between these possibilities, it was necessary to perform the melting experiments under conditions in which encapsidation does not normally occur. The ability of PMV protein to melt TMVRNA was tested at pH 8.5 at low ionic strength. These conditions are optimal both for PMV assembly (18) and for assembly specificity since encapsidation does not occur with heterologous viral RNAs (19). Comparison of the melting curves of TMVRNA alone and in the presence of a near stoichiometric (20:1, w:w) quantity of PMV coat protein (Fig. 1) reveals that the viral protein lowered the T, of the viral RNA by about 45% and also destroyed the weak cooperativity exhibited by its melting profile. Further, once melted, the secondary structure of the RNA was not fully recovered by cooling in the presence of PMV protein. If NaCl was added to 0.1 M to the RNA-protein complex after renaturation, the absorbance dropped to below the starting level, showing that the RNA was not hydrolyzed during denaturation in the presence of protein. The melting of TMV-RNA at 25” displayed an essentially linear dependence on the amount of added PMV protein (Fig. 2), indicating that the protein binds to RNA in

a noncooperative fashion. Maximal melting was achieved with a near stoichiometric ratio of protein:RNA. This means that during PMV assembly nearly all the protein that is required for RNA encapsidation is bound in the initial stages of the reaction, and this results in the complete melting of the RNA. As a control, TMV-RNA was titrated with STI, a protein with a similar molecular weight and isoelectric point to the PMV protein subunit (21, 92). A slight hypochromic effect was observed, indicating that the basic residues of the ST1 protein tended to shield the charged phosphates of the RNA backbone. A similar effect has been observed with denatured ribonuclease (23 >. The effect of NaCl on the melting function of PMV protein was also tested. In the presence of 0.1 M NaCl, which inhibits the encapsidation of PMV-RNA at pH 8.5 (6,18 ), PMV protein had no effect on the melting profile of PMV-RNA (Fig. 3). Complete inhibition of melting by PMV protein was observed down to 0.025 1M NaCl with both PMV and TMV-RNAs (unpublished results). TMV coat protein was tested for melting activity under conditions identical to those used for PMV protein, since TMV protein does not encapsidate RNA at low ionic strength (17’). The melting profiles of TMV-RNA alone and in the presence of TMV protein were indistinguishable, as well as reversible (Fig. 4). Therefore, under these conditions, TMV protein does not function as a melting protein. The possibility that the disc form of TMV protein is required for melting, as it is for the initiation of encapsidation (24)), was accounted for by performing the melting at pH 7.25 in 0.5 Z sodium pyrophosphate using the heterologous PMV-RNA. Under these conditions of pH and ionic strength, which are favorable for TMV assembly (25 ), most of the protein is converted to the disc polymer within 10 min (26). However, the melting profile of PMV-RNA was unaffected by the presence of TMV protein in these conditions, although discs were present. The above experiments show that, under

SHORT COMMUNICATIONS

25 -

. :: zo-

6 L p

15-

6 i IO:: Pe 5,3”0

5

10

15

20

25

30

TEMPERATURE

35

40

45

50

(“C)

FIG. 1. Melting profiles of TMV-RNA in the presence and absence of PMV coat protein. TMV-RNA alone (solid circles): 0.04 mg of TMV-RNA in 1 ml of 0.01 M ‘KS buffer, PH 8.5, was heated from 5 to 50”. TMvRNA + PMV protein (open circles): 0.04 mg of TMVRNA was equilibrated at 5” with 1.0 mg of PMV protein in 1 ml of 0.01 M Tris buffer, pH 8.5. The mixture was heated to 30” (upward arrow) and then cooled back down to 5” (downward arrow) at approximately the same rate. A separate melting profile of PMV alone showed that no significant change in A 260"In occurred over the range 5-35”.

virus reconstitution conditions, PMV coat protein can melt RNA, and that this activity is independent of encapsidation. On the other hand, TMV protein does not melt RNA even under conditions which can result in encapsidation with homologous RNA. The different results obtained with the two coat proteins may reflect the different assembly strategies employed by their viruses. PMV assembly starts with a rapid and specific nucleation event in which a 50-nm nucleoprotein helix is formed (27). At the same time, coat protein binds nonspecifically to the remaining, nonencapsidated RNA to form the extended particles (18). The results obtained from titrating TMV-RNA with PMV protein suggest that the extended particles are assembly intermediates in which the protein is bound stoichiometrically to the RNA which is melted. The binding and melting reactions, per se, are apparently distinct, since extended particles have been observed in these and in previous experiments (19, 28) under a variety of condi-

237

tions where minimal to no melting occurs. While it is not clear whether melting precedes the initiation event, it does appear that the slower elongation reaction requires that the RNA not have extensive secondary structure, at least at the site of helix elongation, since both elongation (27’) and the protein melting activity are inhibited by low levels of NaCl, which serve to stabilize the regions of RNA secondary structure. In contrast to the case for PMV, TMV assembly clearly does not require a preencapsidation melting of its RNA, since TMV protein is apparently unable to melt RNA in the absence of encapsidation. TMV assembly may involve a mechanism whereby the RNA is pulled through the central canal of the growing nucleoprotein helix by the &subunit

disc

polymer

of the

TP/IV

coat

protein (29). The quantized binding of over 110 nucleotides per disc presumably provides the energy needed to melt the nucleic acid, either locally at the binding site, or during its entry into and movement

E100E % Z 5 50 6 z f o 5 I ’

i

I 5

10

15

20

PROTEIN

RNA

25

30

35

J

RATIO (w WI

FIG. 2. Titration of TMV-RNA with PMV protein at constant temperature. Upper curve: 0.04 mg of TMV-RNA in 1 ml of 0.01 M Tris buffer, pH 8.5, was equilibrated at 25”. Increments of concentrated PMV protein were then added to the RNA and reference (buffer only) solutions to give the indicated weight ratio of protein:RNA. Melting effect is plotted as percentage maximum increase in A,,.,, after correcting the absorbance values for light scattering. Lower curve: in a separate experiment, soybean trypsin inhibitor protein was used to titrate TMV-RNA under conditions otherwise identical to those described above. For this case, the A,,,, values are plotted directly.

238

SHORT COMMUNICATIONS

Is 5

8 10

* 15

0 c 3 ( 20 25 30 35 TEMPERATURE (‘C)

II 40

FIG. 3. Effect of NaCl on the melting activity of PMV protein. PMV-RNA alone (dashed curve): 0.04 mg of PMV-RNA in 1 ml of 0.01 M Tris buffer, pH 8.5, containing 0.1 il4 NaCl was heated from 5 to 40”. PMV-RNA + PMV protein (solid curve): 0.04 mg of PMV-RNA, equilibrated at 5” with 1.0 mg of PMV protein in 1 ml of 0.01 M Tris buffer, pH 8.5, containing 0.1 M NaCl, was heated to 35” (PMV protein is denatured around 39”). The renaturation curve was not significantly different from the melting curve.

through the canal. Consideration of the detailed structure of TMV (30) suggests that the protein V helix, which lines the central canal in the virus, will give rise to a substantial net negative surface charge, and this might contribute to the destabilization of the RNA secondary structure. Thus, the melting of TMV-RNA can be considered to be an encapsidation-linked process. The proposed scheme for TMV assembly in vitro may be applicable to the in vivo On the other hand, judging situation. solely from the in v&o data and the moderate ionic strength levels presumed to exist intracellularly, the melting activity of PMV protein and, indeed, PMV assembly itself must be inhibited. To account for PMV assembly in vivo, the problem of how the RNA secondary structure is overcome must be explained. Perhaps a stronger melting protein is supplied by the host. During the life cycle of an RNA bacteriophage, three host ribosomal proteins participate in the tetrameric RNA replicase enzyme, which specifically replicates the viral RNA (31). One of these ribosomal proteins (Sl) actually possesses RNA melting activity (32). Sl melts double-stranded regions which have a T, up to about 55”, which is in the T, range expected for PMV-RNA in vivo at a monovalent cation concentration near 0.15 M and a divalent cation concentration around 0.1 mM (6). Alternatively, the problem

of melting the extensive regions of secondary structure formed by the free RNA in vivo might be circumvented by linking encapsidation to RNA plus-strand synthesis. The latter presumably proceeds in the 5’ to 3’ direction, and the encapsidation of PMVRNA starts at or near the 5’ end of the RNA (33). The in vitro self-assembly reactions of all flexuous and rigid elongated plant viruses studied, except TMV, initiate near the 5’ end of the RNA (33-34 M. AbouHaidar, personal communication), and are inhibited by low levels of NaCl (9, 18, 36-38, M. AbouHaidar, personal communication). These viruses are probably assembled in vivo via one of the strategies proposed above for PMV, since the coat proteins of these viruses are apparently unable to overcome the stabilizing effect of salt on the RNA secondary structure, whose presence ultimately inhibits encapsidation. Implicit in these strategies is that there need be no correlation between the diameter of the central canal and the assembly strategy used by the elongated plant viruses as has been suggested (9), since viruses with both larger (tobacco rattle virus, (39)) and

TEMPERATURE PC)

FIG. 4. Melting profiles of TMV-RNA in the presence and absence of TMV protein. TMV-RNA alone (dashed curve): 0.04 mg of TMV-RNA in 1 ml of 0.01 M Tris buffer, pH 8.5, was heated from 5 to 42”. TMVRNA plus TMV protein (solid curve): 0.04 mg of TMVRNA equilibrated at 5” with 0.80 mg of TMV protein in 1 ml of 0.01 M Tris buffer, pH 8.5, was heated to 42” (upward arrow), and then cooled to 5” (downward arrow).

SHORT

COMMUNICATIONS

smaller (PMV) diameter canals than that found for TMV assemble in a polar fashion starting from the 5’ end of their RNAs. On the other hand, the encapsidation of TMVRNA starts internally, about 1000 nucleotides from its 3’ end (40). This feature rules out the possibility for a replication-linked encapsidation mechanism because of the direction of RNA synthesis. Thus, the unique assembly strategy employed by TMV may be a special adaptation which the virus has evolved to solve the problem of how to overcome the extensive secondary structure which its RNA must possess prior to its encapsidation in vivo . The melting function of PMV protein apparently facilitates virus reassembly in vitro under the nonphysiological conditions of low ionic strength. However, since an in vivo role for the melting activity is unlikely, in view of the intracellular salt levels, there remains the question of why PMV coat protein possesses a melting function at all. The answer, if it not be chance, probably reflects the evolution of PMV. The melting activity may be a vestigial function from an evolutionary coat protein predecessor, possibly a host-derived melting protein, whose intrinsic activity may or may not have played an essential role in the early evolutionary stages of the virus life cycle. However, during its development into a more efficient coat protein, some of its melting activity was sacrificed. An equally attractive speculation is that the melting function of the coat protein may, itself, be in the development stage. These considerations suggest that the relationships between viral coat proteins and host proteins, particularly nucleic acid-binding proteins, might be an area worth exploring.

2. BRAKKE, M. K., Vivobgy 98, 76-87 (1979). 3. BONHOEFFER, F., and SCHACHMAN, H. K., Biochem. Biophys. Res. Commun. 2, ,366--371 (1960). R. M., B&him. 6. HOMER, R. B., and GOODMAN, Biophys. Actu 378,296-304 (1975). 5. LANINA, T. P., TYULKINA, L. G., LEDNEVA, R. K., and BOGDANOV, A. A., Mol. Biol. (USSR) 8, 591-598 (1974). M., ERICKSON, J. W., and BAN6. ABOUHAIDAR, CROFT, J. B., Virology 98, 116-126 (1979). 7. BOEIYTKER, H., J. Mol. Bill. 2, 171-188 (1969). 8. GRATZER, W. B., and RICHARDS, E. G., Biopolymers 10, 2697-2614 (1971). J. B., ABOUHAIDAR, M., snd ERICK9. BANCROFT, SON, J. W., Irirologl( 98, 121-130 (19’79). I., UHLENBECK, 0. C., and LEVINE, M. D., Nature (London) 236, 362-367 (1971). 11. RIESNER, D., and ROME& R., In “PhysicoChemical Properties of Nucleic Acids (J. Duchesne, ed.), Vol. 2, pp. 237-318. Academic Press, New York, 1973. 12. LAUFFER, M. A., In “Entropy-Driven Processes in Biology: Polymerization of Tobacco Mosaic Virus Protein and SimiIsr Resetions” (A. KIeinzeller, G. F. Springer, and H. G. Wittman, eds.), p. 151. Springer-Verlag, New York, 1975. J. I., Advan. Virus Res. 15, 19. TIKCHONENKO, 201-296 (1969). H., C~ER, R. I., and GRATZER, 14. ISENBERG, W. B., B&him. Biophys. Acta 232, 184-191 (1971). P., GUANTIERI, V., VOVLAS, C., and 15. PIAZZOLA, TAMBURRO, A. M., J. Gen. Viral. 37, 359-372 (1977). 16. DENLOYE, A. O., HOMER, R. B., and HULL, R., J. Gen. Viral. 41, 77-85 (1978).

10. TINOCO,

17. 18.

STILL-

B.,

Virology

J. B.,

Virolo.gy

H., and SINGER,

ERICKSON,

J. W., and BANCROFT,

90, 36-46 (1978). 19.

20. 21. 22. 23.

REFERENCES 1. ERICKSON, J. W., BANCROFT, J. B., and MAN, M. S., manuscript in preparation.

FRAENKEL-CONRAT,

23, 354-362 (1964).

ACKNOWLEDGMENTS This work wss supported by the National Resesreh Council of Catmda and the Academic Development Fund of the University of Western Ontario.

289

26.

ERICKSON, J. W., ABOUHAIDAR, M., and BANCROFF, J. B., Virology 99, 66-66 (1978). FRAENKEL-CONRAT, H., V~TOZO~Y 4, l-4 (1967). WV, Y. V., and SCHERAGA, H. A., Biochemistry 1, 698-705 (1962). DURHAM, A. C. H., and BANCRO~, J. B., Virology 93, 246-252 (1979). VON HIPPEL, P. H., JENSEN, D. E., KELLY, R. C., and MCGHEE, J. D., In “Nucleic Acid-Protein Recognition” (H. Vogel, ed.), pp. 65-83. Academic Press, New York, 1977. BUTLER, P. J. G., and DURHAI, A. C. H., Advan. Protein Chem. 31, 187-251(1977).

SHORT

240

COMMUNICATIONS

G., and HIRTH, L., FEBS Lett. (1973). 26. M., LEBEURIER, G., and HIRTH, L., FEBS Lett. 22, 297-300 (1972). 27. ERICKSON, J. W., and BANCROFT, J. B., Virology 90.47-53 (1978). 28. ERICKSON, J. W., and BANCRO~, J. B., In “Biological Recognition and Assembly” (C. F. Fox, ed.). Liss, New York, in press, 1980. 29. BUTLER, P. J. G., FINCH, J. T., and ZIMMERN, D., Nature (London) 265,217-219 (1977). SO. STUBBS, G., WARREN, S., and HOLMES, K., Nature (London) 267, 216-221 (1977). 31. BLUMENTHAL, T., and CARMICHAEL, G. G., Annu. Rev. B&hem. 48, 525-548 (1979). 32. SZER, W., HERMOSO, J. M., and BOUBLIK, M., 25.

LEBEURIER, 34, 19-23 LONCHAMPT,

B&hem.

33. 34. 35. 36. 37.

38. 39. 40.

Biophys. Res. Commun. 70,957-s-4 (1976). ABOUHAIDAR, M., and BANCRO~, J. B., Virology 90, 54-59 (1978). ABOUHAIDAR, M., and HIRTH, L., Virology 76, 173-185 (1977). MORRIS, T. J., and SEMANCIK, J. S., Virology 53, 215-224 (1973). GOODMAN, R. M., Virology 76, 72-78 (1977). ATABEKOV, J. G., NOVIKOV, V. K., VISHNICHENKO, V. K., and KAFTANOVA, A. S., Virology 41, 519-532 (1970). MCDONALD, J. G., and BANCROFT, J. B., J. Gen. Viral. 35, 251-263 (1977). OFFORD, R. E., J. Mol. Bid. 17,370-375 (1966). ZIMMERN, D., Cell 11, 463-482 (1977).