Oct 11, 1993 - EMSA conditions were adapted from Weinberger et al. (1986). .... Aviv CD spectropolarimeter, to Alan Frankel for patient guidance during.
The EMBO Journal vol.13 no.3 pp.727-735, 1994
Specific RNA binding by amino-terminal peptides of alfalfa mosaic virus coat protein
M.L.Baer' 2, F.Houser' 2, L.S.Loesch-Fries3 and L.Gehrke' 2,4 1Division of Health Sciences and Technology, Massachusetts Institute of Technology, Building E25-545, Cambridge, MA 02139, 2Program in Cell and Developmental Biology, Harvard Medical School, Boston, MA 02115 and 3Department of Botany/Plant Pathology, Purdue University, West Lafayette, IN 47906, USA 4Corresponding author (at MIT) Communicated by A.Rich
Specific RNA-protein interactions and ribonucleoprotein complexes are essential for many biological processes, but our understanding of how ribonucleoprotein particles form and accomplish their biological functions is rudimentary. This paper describes the interaction of alfalfa mosaic virus (AIMV) coat protein or peptides with viral RNA. AIMV coat protein is necessary both for virus particle formation and for the initiation of replication of the three genomic RNAs. We have examined protein determinants required for specific RNA binding and analyzed potential structural changes elicited by complex formation. The results indicate that the amino-terminus of the viral coat protein, which lacks primary sequence homology with recognized RNA binding motifs, is both necessary and sufficient for binding to RNA. Circular dichroism spectra and electrophoretic mobility shift experiments suggest that the RNA conformation is altered when amino-terminal coat protein peptides bind to the viral RNA. The peptide-RNA interaction is functionally signifi'cant because the peptides will substitute for AIMY coat protein in initiating RNA replication. The apparent conformational change that accompanies RNA- peptide complex formation may generate a structure which, unlike the viral RNA alone, can be recognized by the viral replicase. Key words: conformation/peptide/replication/RNA-protein interactions/virus
Introduction Specific RNA -protein interactions are essential for protein biosynthesis (Hershey, 1991; Spirin and Ryazanov, 1991), RNA splicing (Moore et al., 1993), virus replication (Zapp et al., 1991; Kao and Ahlquist, 1992; Nassal, 1992) and localization of messenger RNAs (Singer, 1993). Despite the central importance of ribonucleoprotein complexes, our understanding of how sequence and structure unite to form determinants that are required for intermolecular RNA protein recognition and interaction is at an early stage of development. Although biochemical approaches have defined specific ribonucleoproteins and revealed sites of interaction (Scherly et al., 1990; Calnan et al., 199 1a; Dreyfuss et al., 1993; Moore et al., 1993), high resolution X-ray crystal Oxford University Press
structure data are currently available for only two RNA -protein complexes (Rould et al., 1989; Ruff et al., 1991). As amino acid sequence data for RNA binding proteins accumulate, families of RNA binding proteins have been proposed based upon consensus binding motifs [summarized in Mattaj (1993)]. Classification is useful for predicting functional relationships, but in many cases, little is known about how the proteins recognize and bind RNA or how RNA conformation presents a platform that is recognized by the RNA binding protein. RNA-protein interactions in alfalfa mosaic virus (AlMV) and the related ilarviruses are essential not only for virus encapsidation, but also for viral infectivity (Bol et al., 1971; Matthews, 1991). The genetic information of AIMV is contained in three positive-sense genomic RNAs (RNAs 1-3), while the coat protein (CP) is encoded on the subgenomic RNA 4. Each of these RNAs is separately encapsidated (Jaspars, 1985). Infection is dependent upon the genomic RNAs plus CP (Bol et al., 1971). The role of CP in 'genome activation', which includes the initial steps of virus replication, is poorly understood. For viruses having a positive-sense RNA genome, the 3' termini of the RNA are essential for recognition and binding by the viral RNAdependent RNA polymerase (replicase) molecules that copy the genome. Putative replicase recognition signals such as pseudoknots or transfer RNA-like structures are found in the 3' termini of a number of plant virus RNAs, including those of the bromovirus and tobamovirus groups (Hall, 1979; Pleij et al., 1985; Garcia-Arenal, 1988; Schimmel, 1989). Such structures are not found in the RNAs of AIMV or the ilarviruses; instead, strong signals for CP binding are localized in the 3' untranslated regions. Houwing and Jaspars (1978) proposed that the recognition signal for replicase binding is created by an RNA conformational change that accompanies CP binding to the RNA 3' terminus. The amino-terminus of AIMV CP is lysine-rich, which distinguishes it from the arginine-rich RNA binding proteins
(Calnan et al., 1991a,b; Kjems et al., 1991b). Trypsin digestion will cleave 25 amino-terminal amino acids from A1MV CP, resulting in the loss of genome activation potential (Bol et al., 1974). Virions of AIMV and tobacco streak virus, a related ilarvirus, contain zinc, and there is homology in the CP amino acid sequences with zinc finger domains (Sehnke et al., 1989). It has been proposed that these putative zinc finger domains are important in genome activation. We report here that the lysine-rich amino-terminus of AIMV CP is necessary and sufficient for binding 3'-terminal fragments of AlMV RNA 4 in vitro. Although CPs lacking the basic amino-terminus cannot bind RNA, a heterologous protein containing the amino-terminal 38 amino acids of AIMV CP fused to rabbit cx-globin bound to AIMV RNA. The binding of chemically synthesized amino-terminal CP peptides to the 3' terminus of A1MV RNA 4 is accompanied by an apparent RNA conformational change as shown by 727
M.L.Baer et al.
electrophoretic mobility bandshift data and circular dichroism (CD) analysis. Substitution of these peptides for AlMV CP in virus replication assays resulted in viral genome activation and the accumulation of CP, whereas addition of a heterologous peptide or an unrelated virus CP did not result in genome activation. Therefore, these data indicate that the amino-terminus of A1MV CP is necessary and sufficient for binding RNA and for activating the viral genome for replication. The results are consistent with the possibility that a CP-mediated conformational change in the virus RNA is required for RNA recognition by the viral replicase.
Results
(AIMV718-881, Figure iB) was analyzed by electrophoretic mobility shift assay (EMSA) (Figure 2A). At a CP dimer:RNA ratio of 10, three bands with diminished electrophoretic mobility are visible (Figure 2A, lane 2). The apparent Kd (at 220C) for the CP-AIMV718-881 interaction is 500 nM as determined by nitrocellulose filter binding and by analysis of EMSA patterns at varying RNA concentrations (data not shown). The specificity of the CP-AIMV718-881 was tested by competitive binding using the following unlabeled competitor -
AIMV RNA 4 and its protein translation product, AIMV (CP), are shown schematically in Figure lA. The aminoterminus of CP is rich in basic amino acids (underlined in Figure lA) and contains a trypsin-sensitive site between Arg25 and Lys26 (Bol et al., 1974; Jaspars, 1985). The flexible amino-terminus of CP extends to approximately amino acids 36-38 (van Beynum et al., 1977; Kan et al., 1982). The 3' untranslated region of A1MV RNA 4 contains a high affinity CP binding domain that was mapped by
ribonuclease protection experiments (van Boxsel, 1976; Verhagen et al., 1976; Driedonks et al., 1978; Houwing and Jaspars, 1978). A homologous CP binding domain exists in the 3'-untranslated region of the three genomic AIMV RNAs (Koper-Zwartoff and Bol, 1980).
AIMV718-881, a tobacco mosaic virus (TMV) RNA
RNAs:
A
1
881 nt
AIMV RNA 4
_
coat protein binding domain
AIMV Coat Protein
1
220 amino acids
2
38
A
C. G
sw.
G
A
U
G
G8 6 O L (G=C6 C
A
C
-
{C;
=
ll-A
C
A
A
A-UJ 87Uo=G
S
C =GC
8 X3 CJG 8 8 1 U A lz i~ q qAL)G C A U C;
A-U A-U
A-U
820U-A
A-U U-A
Fig. 2. EMSA of CP-RNA complexes indicating specificity of the C
C-G G-C860 U A 85OU-A C A
U-A
-
CG
Fig. 1. (A) Schematic representation of AIMV RNA 4 and the translation product of RNA 4, AIMV CP. The CP binding region of the AlMV RNAs is located in the 3' untranslated region. The aminoterminal region of CP is basic; the basic amino acids are underlined. (B) Sequence and secondary structure of the CP binding domain of AIMV718-881. The protein synthesis termination codon is found at nucleotide position 700, and the fragment representing nucleotides 718-881 was subcloned as a BstXI/SmaI fragment into a transcription vector containing the bacteriophage T7 promoter. Lower case letters represent vector nucleotides. Tetranucleotide repeats of the sequence AUGC are indicated in bold lettering.
728
A
CP-A1MV718-881 interaction. (A) Binding and competition using labeled AIMV718-881 RNA. Lane 1: free (unbound) AIMV718-881
A-U840 A-U 870CG 881 U-A C-C GU-A 770 A-U 810 4,~ ~ U-A-U U-A' UA 1T7gatcTGGGAUGC AUGC AGG UUGC U-A AUGC U-A AUG AUGCcc 718
___
-
C U
G
UAAUA-U CG A-U G.C C-G
C A
J
U AA C830 A-U C-G U-A G=C C-G C.G 800 U-A 750A-U 760 CGA. U G=C A A CGA U-A.G.C A-U U-A A C---yG~ A U
U3 G
8;
.Xj
A
A A
U A C A A UB -C
C
U.-isI4
6
D
-
C U C790
joij .4
1 2 3 4
cB C860 Li
B
....
ON
A A
A
B
3 4 5 6 7 8
C
ac-SSSQAGGEAGEPTJMSQNYAALMAQLPEPPALEVP 1 25
-
fragment, a 3'-terminal 39 nt fragment of AIMV RNA 4 (AlMV843-881, Figure 2C), or a mutant form of the AIMV 39 nt fragment (AIMV843-881.) that does not bind CP (Figure 2D). The characterization of RNA AlMV843-881a
A 1
X:s.
To examine CP binding, 3'-terminal nucleotides 718-881 of RNA 4 (Figure IB) were subcloned into a vector containing the bacteriophage T7 promoter to permit in vitro RNA transcription. The secondary structure of this RNA (Figure iB) was proposed by Houwing and Jaspars (1982) and later tested experimentally by enzymatic structure mapping (Quigley et al., 1984). CP binding to the 170 nucleotide (nt) 3'-untranslated region of AIMV RNA 4
RNA. Lanes 2, 5 and 8: AIMV718-881 RNA plus a 10-fold molar positions of RNA -protein complexes are marked by bullets in lane 2. Lane 3: same as lane 2 with the addition of a 100-fold molar excess of unlabeled competitor AIMV718-881 RNA. Lane 4: same as lane 2 with the addition of a 100-fold molar excess of TMV RNA fragment; lane 6: same as lane 2 but including a 24-fold molar nucleotide excess of the AIMV843-881 RNA (shown in panel C); lane 5: same as in lane 2, but containing a 24-fold molar nucleotide excess of the mutant 39 nt RNA, AIMV843-881aaa (shown in panel D). (B) Binding of CCMV CP to AIMV718-881 RNA. The unbound RNA band is indicated by an arrow. Lane 1: unbound AIMV718-88, RNA only. Lanes 2-4: AIMV718-881 RNA plus a 1-, 50- or 100-fold molar excess of CCMV CP. (C) AIMV843-881, used for competition binding experiments; (D) AIMV843-881aaw, used for competition binding experiments. Nucleotides at positions 866-868 and 877 have been changed relative to the wild-type sequence shown in panel C. Mutant RNA AlMV843-881ia does not bind CP (Houser et al., 1994). excess of virion CP. The
Alfalfa mosaic virus RNA - peptide interactions
is described elsewhere (Houser et al., 1994) where the critical role of AUGC865-868 in binding CP is described. The EMSA experiments demonstrate that the binding of CP to labeled A1MV718-881 is blocked by the cognate unlabeled 170 nt RNA, AIMV718-881 (Figure 2A, lane 3), or the 3'-terminal 39 nucleotide, A1MV843-881, (lane 6). However, binding was not blocked by the TMV RNA (lane 4) or by AlMV843-881. (lane 7). The specificity of the AIMV RNA 4 -CP interaction was also assessed by testing the binding of cowpea chlorotic mottle virus (CCMV) CP to AIMV718-881 (Figure 2B). CCMV CP is similar to AIMV CP in that its amino-terminal domain is basic. The absence of bandshifts indicates that CCMV CP does not bind to AIMV718-881 (Figure 2B). Thus, the data shown in Figure 2 indicate that binding of AlMV CP to AIMV718-881 is specific. To determine if the amino-terminus of CP, which has been described as a mobile amino-terminal arm (van Beynum et al., 1977), is necessary and sufficient for binding the AlMV RNAs, we tested the binding of several CP constructs in which the CP amino-terminus was either deleted or fused to a heterologous protein that does not bind RNA. DNA constructs were subcloned into vectors containing either an SP6 or T7 promoter for in vitro RNA synthesis. The transcribed RNAs were translated using an mRNAdependent rabbit reticulocyte system. The radiolabeled translation products (drawn schematically in Figure 3A) were analyzed by SDS -PAGE (Figure 3B). Although AIMV CP has a predicted subunit molecular weight of 24 000, it A. Pfecoec 2 4v ~~~~~~~~v~~~~~ ~
22
migrates with an apparent molecular weight of 27-28 000 in these gels (Kruseman et al., 1971) (Figure 3B, lane 1). The mobilities of the rCPANH2 and rCP/globin translated proteins were also slightly lower than predicted (Figure 3B, lanes 2 and 3). Binding of the newly translated proteins to the AIMV RNA 4 template during the in vitro protein synthesis was prevented by linearizing the rCP and rCPANH2 plasmid DNAs at nucleotide position 718, which lies downstream of the CP translation termination codon (nt 700) but upstream of the 3'-terminal CP binding domain. Therefore, the CP mRNA run-off transcripts lacked CP binding ability. An EMSA was used to test binding of the in vitro translation products to RNAs. The radiolabeled proteins were incubated without added RNA, with AIMV718-881 RNA, or with the TMV RNA fragment and analyzed by electrophoresis into a non-denaturing gel (Figure 4). No bands were present in the gel lanes containing samples of control translation products incubated either with or without added RNA (Figure 4A and B, lanes 1-3). In the absence of RNA, the radiolabeled rCP did not enter the EMSA gel (Figure 4A, lane 4). Coomassie Brilliant Blue staining of EMSA gels following electrophoresis of virion CP demonstrated that all protein remained in the wells (data not shown). Similar behavior of other nucleic acid binding protein in the absence of RNA or DNA has been reported previously (Bianchi et al., 1989). Upon incubation of unlabeled AIMV718-881 with rCP, a complex formed as indicated by the band present in Figure 4A, lane 5. This band was not detected when rCP was incubated with TMV RNA (Figure 4A, lane 6), indicating that the interaction between radiolabeled rCP and AlMV718-881 is specific. Only a single shifted band is present in Figure 4A, lane 5 because the amount of rCP incubated with AIMV718-881 was sufficient only for the formation of the first of multiple
I ci
-CP
::.-- :;c.
-t--
1 2 34 5 6l,_ . "" 'P
i
-1...
...
1 2 3 _4
5 6 1 _-
-2
3-._4 5 6
-
'^.c, p
B.
2
3
4
A
Fig. 3. (A) Schematic representation of recombinant proteins synthesized by in vitro translation and used for EMSA. CP sequences are shown as solid lines while cz-globin sequences are cross-hatched. rCP refers to CP synthesized in vitro; rCPANH2 refers to the aminoterminal deletion mutant; rCP/a-globin refers to a chimeric protein containing the amino-terminal basic domain of AIMV CP fused to rabbit a-globin; cs-globin is rabbit os-globin. (B) SDS-PAGE analysis of translated proteins. rCP (lane 1), rCPANH2 (lane 2), rCP/globin (lane 3) or ca-globin (lane 4) proteins synthesized in vitro were analyzed by electrophoresis in a 15% SDS-polyacrylamide gel. The migration of molecular weight standards (labeled in kDa) is indicated at the left.
B
C
Fig. 4. EMSA of RNA-protein complexes using radiolabeled proteins synthesized in vitro bound to viral RNA fragments. (A) Analysis of RNA binding of the radiolabeled full-length CP (rCP) to RNAs. Proteins synthesized in a control translation reaction (without added mRNA) were incubated in the absence of RNA (lane 1), or with AIMV718-881 RNA (lane 2) or TMV RNA (lane 3). rCP was incubated without RNA (lane 4), with AIMV718-88, RNA (lane 5) or with the TMV RNA fragment (lane 6). (B) Analysis of binding of the amino-tenminal deletion protein (rCPANH2) to RNAs. Lanes 1-3: control reactions as in panel A. Labeled rCPANH2 protein was incubated without added RNA (lane 4), with AIMV718-88, RNA (lane 5) or with TMV RNA (lane 6). (C) Analysis of binding CP/a-globin fusion protein (rCP/a-globin) or rabbit ca-globin to RNAs. Labeled rCP/ax-globin was incubated without RNA (lane 1), with AIMV718-881 RNA (lane 2) or with TMV RNA fragment (lane 3). Rabbit a-globin translation products were incubated without RNA (lane 4) with AIMV718-88, RNA (lane 5) or with added TMV fragment (lane 6). Each group of reactions contained equal amounts of radiolabeled protein and RNAs.
729
M.L.Baer et al.
Tat
C P25
e .,.
CP38
R52
Fig. 5. EMSA of AIMV718-881 RNA bound to peptides. Left panels: radiolabeled AIMV718-881 was incubated with peptides CP25 or CP38, which contain the amino-terminal 25 or 38 amino acids respectively of the viral CP. Lanes 1 and 8: labeled AIMV718-881 RNA only; lanes 2, 5 and 7: RNA plus a 25-fold molar excess of peptide; lane 3: same as lane 2 plus a 100-fold molar excess of competitor A1MV718-881 RNA; lane 4: same as lane 2 plus a 25-fold molar nucleotide excess of competitor AIMV843-88, RNA; lane 6: same as lane 2 plus a 25-fold molar nucleotide excess of competitor AlMV843-881an RNA. Right panels: radiolabeled AIMV718-881 was incubated with the HIV Tat peptide (YRKKRRQRRRA) or the modified Tat peptide R52 (YKKKRKKKKKA). Lanes 1 and 9: labeled AIMV718-881 RNA only; lanes 2, 5 and 8: RNA plus a 25-fold molar excess of peptide; lane 3: same as lane 2 plus a 100-fold molar excess of competitor AIMV718-881 RNA; lane 4: same as lane 2 plus a 25-fold molar nucleotide excess of competitor AIMV843-88, RNA; lane 6: same as lane 2 plus a 25-fold molar nucleotide excess of competitor AIMV843-88laaa RNA; lane 7: same as lane 2 plus a 100-fold molar excess of TMV fragment RNA.
complexes observed in Figure 2. Increasing the CP concentration by addition of unlabeled virion CP resulted in the appearance of multiple shifted complexes (data not shown). The results presented in Figure 4B, lanes 4-6 show that rCPANH2, lacking the CP amino-terminus, failed to bind AIMV RNA, thereby confirming that the basic aminoterminus is essential for binding RNA (Zuidema et al., 1983). In addition, however, we found that fusion of the amino-terminal 39 amino acids of CP to rabbit a-globin, which does not bind RNA (Figure 4C, lanes 4-6), created a chimeric protein that bound AIMV718-88, RNA (Figure 4C, lane 2). The chimeric protein also exhibited weak affinity for the TMV RNA fragment as shown in Figure 4C, lane 3, although the native CP does not bind the heterologous TMV RNA (Figure 4A, lane 6). Therefore, the behavior of chimeric CP/a-globin protein was less discriminatory, permitting weak interaction with another RNA. The data presented in Figure 4 are conclusive evidence that the AlMV CP amino-terminus is necessary for binding AIMV RNAs. To determine if the basic amino-terminal CP arm is indeed sufficient for binding RNA, peptides corresponding to the amino-terminal 25 amino acids (peptide CP25) and 38 amino acids (peptide CP38) (see Figure IA) were chemically synthesized and tested for binding to RNA. Human immunodeficiency virus (HIV) Tat peptide (Calnan et al., 1991a) was used as a control peptide to test binding specificity. Results are shown in Figure 5, where arrows mark the position of unbound (free) RNA. Binding of CP25 and CP38 to AIMV718-881 resulted in a negative rather than a positive bandshift, that is, the RNA-peptide complexes migrated more rapidly than free RNA in the non-denaturing gels (Figure 5, left panels, compare lanes 1 and 2). The specificity
730
of peptide -RNA complex formation was tested by the addition of excess amounts of an unlabeled competing RNA. The negative bandshift observed in the presence of peptides CP25 or CP38 was eliminated by the addition of cognate unlabeled AIMV718-88, RNA and AIMV843-881 (Figure 5, left panels, lanes 3 and 4), but it was not affected by adding AlMV843-881 (Figure 5, left panels, lane 6). No binding of the arginine-rich HIV Tat peptide to AIMV718-881 RNA was detected (Figure 5, upper right panel). A modified Tat peptide, R52, which is similar in charge distribution to the wild type Tat peptide but in which the majority of arginine residues have been changed to lysine (Calnan et al., 199 lb) was also tested for binding (Figure 5, lower right panel). AIMV718-881 interacted with the R52 peptide, resulting in a negative bandshift (Figure 5, lower right, compare lanes 1 and 2). However, the formation of this complex was affected by adding an excess of specific or non-specific RNAs. Specific RNA competitors AIMV718-881 (lane 3) and the shorter fragment AlMV843-881 (lane 4) interfered with R52 complex formation as did the non-specific competitors AlMV843-881a (lane 6) and the TMV RNA (lane 7). The results of these competition experiments strongly indicate that binding of peptides CP25 and CP38 to AIMV718-881 is specific, and that the R52 peptide binds non-specifically. The data presented in Figure 5 are evidence that the amino-terminus of CP is sufficient for binding RNA. The increased mobility of the AIMV718-881 -peptide complexes in the non-denaturing polyacrylamide gels (negative bandshifts) suggests that the conformation of the RNA becomes more compact as a result of RNA -peptide binding, thereby permitting accelerated migration. Several lines of evidence suggest that peptides CP25 and CP38 bind RNA with a 1: 1 stoichiometry. A single peptide -RNA band is detected in EMSA gels over a wide concentration of peptide using either the 170 nt AIMV718-881 or the 39 nt AIMV843-881 (Figure 5 and Houser et al., 1994). In addition, mutation at a single site (AUGC865-868) in both the 170 nt and the 39 nt RNAs (Figure 2C) results in the loss of peptide binding activity (Houser et al., 1994). This observation suggests strongly that the peptides bind a single site, i.e. AUGC865-868, in both the 170 nt and the 39 nt RNAs. Finally, hydroxyl radical footprinting data demonstrate that CP and the CP25/CP38 peptides all protect a region of the 39 nt AlMV843-881 RNA that extends from approximately nt 861 to 868 (F.Houser and L.Gehrke, unpublished data). It is possible that the peptides bind to RNA as a dimer, but the EMSA data suggest that this is unlikely. Circular dichroism, a sensitive but qualitative indicator of changes in RNA conformation (Hashizume and Imahori, 1967; Tritton and Crothers, 1976; Daly et al., 1990; Tan and Frankel, 1992), was used to investigate further the peptide-induced RNA structural change. The spectrum of peptide CP25, shown in Figure 6A, strongly indicates that the amino-terminus of CP assumes a random coil conformation [for representative spectra, see Creighton (1984)]. This conclusion is consistent with that of Kan and colleagues based upon proton NMR analysis of protein-RNA complexes (Kan et al., 1982). The spectra of AIMV718-881 RNA and the CP25 -AIMV718-881 complex are shown in Figure 6B. The major peak of ellipticity present at 265 am for free RNA was reduced by 15 % by the presence of a 10-fold excess of CP25 peptide. The AIMV718-881 spectrum was -
Alfalfa mosaic virus RNA - peptide interactions
A -50000 -
ffrfl]
0
subtracted from that of the complex to yield the difference spectrum shown in Figure 6C. The large difference at 265 nm probably represents changes in RNA conformation because the peptide conformation contributes minimally at this wavelength. The data in Figure 6C also show a minimum at 220 nm, which represents contributions from both the RNA and the peptide. It has been suggested that the aminoterminus of A1MV CP becomes structured upon binding RNA (van Beynum et al., 1977); however, we cannot draw conclusions about CP25 conformation from the data presented in Figure 6. The biological significance of the peptide -RNA interactions was tested by investigating the capacity of the peptides to activate AIMV replication. For successful infection by the three A1MV genomic RNAs, CP or RNA 4 (the subgenomic mRNA coding for CP) must be present. A hypothesis to explain the requirement for CP is that the RNA -CP complex provides a signal that facilitates the recognition of AlMV RNAs by the viral replicase (Houwing and Jaspars, 1978). Analysis of A1MV infection of tobacco protoplasts was used to determine the activity of the peptides. Mixtures of RNAs plus CP or peptides were introduced into protoplasts using polyethylene glycol, and the percentage of infected protoplasts was determined by an immunoassay using antibodies directed against CP (Loesch-Fries and Hall, 1980). Infected protoplasts are detected by this assay because they contain large amounts of AIMV CP; non-infected protoplasts are not detected because the amount of residual CP from the inoculum is below the level of detection of this assay. Genomic RNAs of AIMV, isolated from virions, were separated from the subgenomic RNA, which is also encapsidated, by repeated sucrose density gradient centrifugations. Inoculation of tobacco protoplasts with 0.5 pmol of the genomic RNAs resulted in a small percentage of infected protoplasts (Figure 7A, RNA only). This percentage is due to a small amount of subgenomic RNA molecules in the genomic RNA preparation. Addition of the peptides CP25 or CP38 resulted in large increases in the number of infected protoplasts (Figure 7A). CP25 was less active than CP38; this was consistent in many different experiments with various concentrations of RNAs and peptides. As shown in Figure 7A, the activity of CP38 was nearly equal to that of A1MV CP. To determine if the activity was specific to AIMV peptides, unrelated virus CP or peptide was added to the inoculum. CCMV CP is similar to AIMV CP in that it has a basic amino-terminus; however, it does not bind to AIMV RNA (Figure 2B). Addition of CCMV CP resulted in an infection level similar to that of the genomic RNAs (Figure 7A). Likewise, the arginine-rich HIV Tat peptide was also inactive (Figure 7A). The percentage of infected protoplasts reflects the number of infected protoplasts in which CP is being synthesized. To confirm that this percentage correlates with viral RNA replication, a portion of several inoculated protoplast samples was incubated in the presence of [3H]uridine, which labels newly synthesized RNA. As shown in Figure 7B (lane 1), viral RNAs were not detected in the sample inoculated with genomic RNAs alone. When protoplasts were inoculated with RNAs plus peptides or AIMV CP, significant amounts of viral RNA accumulated (Figure 7B, lanes 2-4); however, no labeled viral RNAs were observed when the protoplasts were inoculated with RNAs plus CCMV CP (Figure 7B, -
E
-5000 -
0
O N
-10000 - O
E S cm
-1 5000 - O 0
n
la
-20000 - O
-25000 200
1
220
240
260
280
300
280
300
X (nm)
B 7500-
I~ I.
5000
0
E
I
0
E
2500-
0
0-
la
*g +q
200
220
240
260
A (nm)
C 0 0
0
E
10
-500
E
0
ol0 -1500 4 -1000
.
la
0
-2000 200
1
I
220
240
O
260
280
30C .1
A (nm) Fig. 6. CD analysis of A1MV718-881 RNA and AlMV CP peptide CP25. (A) Spectrum of CP25 alone. (B) Spectra of AIMV718-881 RNA alone (+) and of the CP25/AIMV718-881 RNA complex at a 10:1 stoichiometry (diamonds). (C) Difference spectrum of the spectra presented in (B).
731
M.L.Baer et al.
A
B 1
U)
2 3 4
5
n
.......
,
RNA 1 4--- RNA 2 .* RNA 3
50 US
_-
-
R NA 4
0 o
-.,.. ....1
.... .. '..
()}
::.: :-: :.7.7.f.
_ .-.-:-
RNA ONLY
.. .,, t.t.
CP25
CP38
AIMV CP
_
CCMV CP
Tat
Fig. 7. Activity of peptides and virus CPs in genome activation. (A) Percentage of infected protoplasts following inoculation of tobacco protoplasts with 0.5 pmol AlMV genomic RNAs alone (RNA only), or RNAs plus a 120-fold molar excess of CP25 peptide, CP38 peptide, AlMV CP, CCMV CP, or Tat peptide. (B) Accumulation of newly synthesized RNAs in protoplasts. A portion of the protoplast samples described in panel A, excepting that inoculated with Tat peptide, was cultured for 24 h in medium containing 40 yCi/ml [3H]uridine. Total cellular RNAs were isolated and separated by electrophoresis in a 2.4% polyacrylamide-agarose gel. The labeled RNAs were detected by fluorography. The protoplast samples were inoculated with genomic RNAs only (lane 1), genomic RNAs plus CP25 (lane 2), genomic RNAs plus CP38 (lane 3), genomic RNAs plus AIMV CP (lane 5) or genomic RNAs plus CCMV CP (lane 6). The positions of the AIMV RNAs are indicated at the right. The position of viral RNA1 is partially obscured by a labeled cellular RNA.
lane 5). Densitometric analysis of the fluorograph shown in Figure 7B was used to determine the relative amounts of AIMV RNAs 3 and 4 present following inoculation with RNA plus CP25 or CP38. Assuming that the specific radioactivity of each virus RNA is constant from sample to sample, protoplasts inoculated with genomic RNAs plus CP38 (Figure 7B, lane 3) contained 1.7 times more RNA 3 and 4 than the sample inoculated with genomic RNAs plus CP25 (Figure 7B, lane 2). Thus, there is a direct correlation between the percentage of infected protoplasts (Figure 7A) and RNA accumulation data (Figure 7B). The data in Figure 7 demonstrate that CP25 and CP38 will substitute for AIMV CP in genome activation. -
Discussion A unique property of AIMV and the related ilarviruses is that the viral RNA -CP interaction has a dual role in virion structure/virus assembly and in initiation of virus replication (Bol et al., 1971). RNA recognition by the viral replicase may involve a CP-mediated RNA conformational change (Hinz et al., 1979; Houwing and Jaspars, 1980; Srinivasan and Jaspars, 1978, 1982); therefore, this experimental system offers potential for studying both RNA -protein interactions and the possible relationship of conformational change to the functional activity of the complex. The experimental data presented in this paper support several conclusions. First, the amino-terminus of A1MV CP is both necessary and sufficient for binding AIMV RNAs. This conclusion is based upon the results of testing RNA binding by intact CP, by the amino-terminal truncated CP, and by a fusion protein wherein the amino-terminus was fused to a-globin, resulting in the formation of a chimeric 732
RNA binding protein. To extend these findings, aminoterminal peptides were chemically synthesized, and their specific binding to RNA 4 (Figure 5) is evidence that the amino-terminus of the viral CP is also sufficient for binding the RNA. The second conclusion suggested by the data is that binding of amino-terminal peptides to the 3'-untranslated region of AIMV RNA 4 alters the conformation of the RNA. This conclusion is based upon both EMSA (Figure 5) and CD analysis (Figure 6), and the data are consistent with conformational changes proposed in the replicase model for the virus (Houwing and Jaspars, 1980). A third conclusion is that interaction of the amino-terminal peptides with RNA is functionally relevant because the peptides activated the AIMV genome in protoplast infection experiments (Figure
7).
A number of other plant viral coat proteins, including those of brome mosaic virus (Tremaine et al., 1977), CCMV (Tremaine et al., 1972), southern bean mosaic virus (Tremaine and Ronald, 1978) and turnip yellow mosaic virus (Peter et al., 1972), have an amino-terminal basic 'arm' region that is likely to be important for RNA binding. Several studies have demonstrated that removal of the amino-terminal basic domain by trypsin treatment interrupts virus assembly or genome activation (Bol et al., 1974; Tremaine et al., 1977; Tremaine and Ronald, 1978). Removal of the AIMV CP amino-terminus prevents RNA binding (Figure 4); in addition, however, it may also preclude protein-protein interactions between CP and the replicase complex. Although it is known that the AIMV replicase is encoded in genomic RNAs 1 and 2 (Nassuth et al., 1981), it is not clear if the viral replicase or host cell factors interact directly with RNA or with AIMV CP. Quadt et al. (1991) recently reported co-purification of AIMV CP as part of the AIMV replication complex, suggesting that CP may indeed interact with
Alfalfa mosaic virus RNA - peptide interactions
replicase. The higher functional activity of CP38 relative to CP25 in the virus replication assays (Figure 7) may reflect its slightly greater affinity for RNA or an enhanced ability to interact with other replicase components. Extension of the AIMV CP peptide CP25 to CP38 adds three additional basic amino acids and four prolines, the latter of which are known to diminish conformational diversity and therefore stabilize structure. Peptide-RNA interactions have also been analyzed in HIV, where it has been shown that Tat and Rev peptides bind HIV RNA specifically (Daly et al., 1990; Calnan et al., 1991a,b; Kjems et al., 1991a,b; Puglisi et al., 1992). The AIMV CP peptides CP25 and CP38 are distinguished from the HIV Tat and Rev peptides by primary structure: the HIV peptides are characterized by a close clustering of arginines, while the AIMV CP peptides have a high proportion of lysines interspersed with other amino acids. Calnan et al. (199 La) showed that a single arginine located in a basic environment was sufficient for specific Tat peptide binding to tar RNA. Although the AIMV amino-terminus is lysine-rich, we cannot conclude on the basis of current data that it is the lysines and not the arginines that are responsible for A1MV RNA binding. Despite the sequence differences, the structural consequences of binding these peptides to their target RNAs may be similar. Calnan et al. (1991a) and Kjems et al. (1991a) each described RNA conformational changes that accompanied peptide binding. Kjems et al. (199la) cite negative bandshift data (similar to the data shown in Figure 5) as evidence of a Revresponsive element (RRE) RNA conformational change. The 17 amino acid Rev peptide used in these experiments reported by Kjems et al. (199la) was found to protect the same regions of HIV- 1 env RNA as the full-length Rev protein. These results suggested that analysis of peptide binding reflects the binding properties of the intact protein. The precise nature of the RNA conformational change that is suggested by EMSA and CD data upon peptide binding to AlMV RNA is not known. It is notable that several hairpin structures at the 3' terminus of AIMV RNAs are separated by the single-stranded tetranucleotide sequence AUGC (Figure 1), which has been suggested to have a role in binding viral CP (Koper-Zwartoff and Bol, 1980). Hydroxyl radical footprinting analysis shows that the CP protects the AUGC sequences; moreover, mutations in this sequence prevent CP25/CP38 peptide binding (Houser et al., 1994). The potential of a given nucleic acid sequence to be distorted is an important criterion of sequence-specific protein binding (Steitz, 1990), and it is possible that replication of AIMV RNAs requires both specific protein/peptide binding and perturbation of the RNA conformation. Although removal of the amino-terminus of AIMV CP prevents RNA binding and genome activation [this paper; see also Bol et al. (1974)], the specific determinants of CP sequence and structure have not been elucidated. It is striking that CPs of AIMV and a number of other viruses in the ilarvirus group will cross-activate their respective genomes despite a lack of primary sequence homology or serological relationship (Gonsalves and Garnsey, 1975; van VlotenDoting, 1975). Characterization of determinants for RNA binding and genome activation in these functionally equivalent proteins that lack amino acid homology will facilitate testing the possibility of a higher order consensus RNA binding motif among plant viral CPs.
Materials and methods DNA plasmid constructs Plasmid pSP65A4, representing the complete 881 bp AlMV RNA 4 cDNA, has been described previously (Loesch-Fries et al., 1985b). A 170 bp 3-terminal fragment of the AIMV RNA 4 cDNA that contains high affinity CP binding sites was subcloned as a BstXI-SnaI fragment into a transcription vector that minimizes the number of transcribed vector nucleotides (Jobling et al., 1988). The rCPANH2 construct was made by digesting pSP65A4 DNA with Ncil, which cleaves the DNA in the protein coding region near amino acid 38. The truncated coding region was ligated to an oligonucleotide containing the AIMV RNA 4 untranslated leader sequence plus nucleotides that both provided an initiation codon for protein synthesis and re-established the open reading frame. The amino-terminal amino acid sequence of the rCPANH2 protein is Met-Ser-His-Ala-Val40Val41 (CP sequences are in bold letters). The chimeric rCP/globin construct was generated by first digesting pSP65A4 DNA with Ncil and SmaI to remove coding sequences downstream of the amino-terminal basic domain. The full-length rabbit a-globin coding sequence was then subcloned downstream of the AIMV RNA 4 untranslated leader sequence and amino-terminal 38 amino acids. The control TMV RNA fragment (Goelet et al., 1982) used in competition binding experiments was constructed by deleting the CP coding region and fusing the 5' and 3' untranslated regions to generate a 220 nt TMV fragment. Construction and characterization of the A1MV843-881 and AlMV843- 88aaa RNAs are described elsewhere (Houser et al., 1994). In vitro transcription RNAs were transcribed in vitro using bacteriophage SP6 or T7 RNA polymerase (Gibco-BRL Life Sciences or Promega) according to the manufacturer's suggested protocols. When A1MV RNA 4 transcripts were to be used as templates for in vitro protein synthesis, the plasmid DNAs were linearized in the 3' untranslated region at the BstXI site so that the run-off transcripts lacked the 3' CP binding domain. Radiolabeled RNAs for use in EMSA were generated by supplementing the transcription reactions with [35S]GTP (Dupont/New England Nuclear). The 3-terminal 39 nt fragments of AlMV RNA 4 (Figure 2C and D) were synthesized by in vitro transcription using synthetic DNA templates
(Milligan
et
al., 1987).
Virion coat protein Purified virion CP was prepared from virions essentially according to the protocol described by Kruseman et al. (1971). The isolated form of the protein is a stable dimer (Kruseman et al., 1971).
Electrophoretic mobility shift assay EMSA conditions were adapted from Weinberger et al. (1986). Prior to use, all AIMV RNAs were heated to 65°C for 2 min in REN buffer (50 mM NaCl, 10 mM MgCl2, 1 mM EDTA, 10 mM Tris-HCl, pH 7.5) to dissociate aggregates (Srinivasan and Jaspars, 1978) and then cooled slowly to room temperature over 15 min to permit renaturation. One picomole of radiolabeled, gel-purified RNA and 0-100 pmol of purified A1MV CP were incubated in a total volume of 10 Al of binding buffer (10 mM Tris-HCl pH 7.5, 50 mM NaCl, 1 mM DTT, 1 mM EDTA, 5% glycerol, 1 1 RNAsin and 55 pmol yeast phenylalanine transfer RNA at room temperature for 10 min. The molar concentration of CP was calculated on the basis of its isolated form as a dimer (Kruseman et al., 1971). When unlabeled competitor RNA was used, labeled and unlabeled RNAs were mixed before the addition of CP. For assays involving radiolabeled CP synthesized by in vitro translation, aliquots of the protein synthesis reaction containing 50 000-100 000 c.p.m. were incubated with 2 pmol of unlabeled, gel-purified RNA. The tracking dyes xylene cyanol FF and bromophenol blue were added, and the mixtures were analyzed by electrophoresis in a 7.5% polyacrylamide gel (acrylamide:bis ratio of 46:1) in 0.5 x TBE (44.5 mM Tris, 44.5 mM boric acid, 1.25 mM EDTA, pH 8.3) at constant power of 1 W (7 cm x 10 cm x 0.75 mm in a Bio-Rad Mini-PROTEAN II electrophoresis cell) at room temperature. The gels were pre-run at the 1 h prior to loading the samples. Dried gels were exposed same power for to X-ray film (Kodak) or placed in cassettes for phosphorimaging analysis. -
Protein translation and purification
Messenger RNAs were translated at a concentration of 30 Ag/mil in an mRNA-dependent rabbit reticulocyte lysate system (Promega Biotec) in the presence of [35S]methionine (Dupont/New England Nuclear). After incubating the protein synthesis reaction (100 1d) at 30°C for 1 hour, hemoglobin was separated from CP translation products by batch ion exchange chromatography using
DEAE Bio-Gel A
(Bio-Rad Laboratories)
733
M.L.Baer et al. equilibrated in 50 mM Tris, pH 7.5. Hemoglobin was removed by washing the agarose beads three times with 50 mM Tris, pH 7.5. The bound radioactive translation products were eluted in two successive 400 u1 washes of elution buffer (1.5 mM NaCl, 50 mM Tris pH 7.5). The eluates were pooled, diluted 10-fold with 50mM Tris, pH 7.5, and pipetted into a Centricon-10 microconcentrator (Amicon) that had been pre-wetted with a buffer containing 50 mM Tris pH 7.5, 150 mM NaCl. The microconcentrator was then centrifuged at 5000 g,4°C in a Sorvall RC5B centrifuge fitted with an SS34 fixed-angle rotor until the volume was reduced to - 50-100 A1 ( -1 h). The amount of acid-precipitable radioactivity (Mans and Novelli, 1961) in the final concentrated protein was - 10 000-50 000 c.p.m./ul.
Peptide synthesis Peptides were synthesized using T-Boc chemistry on an Applied Biosystems Model 430A Solid Phase Peptide Synthesizer and purified by high pressure liquid chromatography. The amino-terminus of virion CP is acetyl-serine (van Beynum et al., 1977); therefore, peptides were synthesized with the corresponding N-acetyl group.
Circular dichroism CD spectra were determined using an Aviv model 60DS spectropolarimeter equipped with a Hewlett-Packard Peltier temperature controller. Samples were prepared in 20mM potassium phosphate, 100 mM potassium fluoride, pH 7.3. The concentration of RNA was determined by absorption spectra while the concentration of peptide was determined by amino acid analysis. CD spectra were collected from 300 nm to 200 nm and averaged over five scans. The averaging time at each wavelength was 10 s and the scans were run at 5°C. Mean molar residue ellipticity was calculated using a molecular mass of 2618 for CP25. Virus replication in tobacco protoplasts Protoplasts were prepared from Nicotiana tabacum L. cv. Xanthi-nc as previously described (Loesch-Fries et al., 1985a). The protoplasts were inoculated with 0.3 -0.5 pmol AIMV genomic RNA per 105 protoplasts or with mixtures of genomic RNAs plus virus CPs or peptides by a polyethylene glycol procedure (Loesch-Fries et al., 1985a). Protoplasts were then incubated at 105 protoplasts per ml of medium (Samac et al., 1983) for 25 h at 26°C under constant illumination of -1000 lux. Following incubation, viable protoplasts were isolated by flotation on 20% sucrose. The percentage of infected protoplasts was determined by an indirect fluorescent antibody assay procedure (Loesch-Fries and Hall, 1980) using rabbit antiserum directed against AIMV followed by addition of fluorescein isothiocyanate-conjugated anti-rabbit IgG (Cappel Laboratories). AIMV RNAs synthesized during infection were labeled by incubating protoplasts in medium containing 40 /Ci/ml [3H]uridine. Total RNA was isolated from the protoplasts and analyzed by PAGE and fluorography as described previously (Loesch-Fries and Hall, 1980). The relative amounts of radioactivity in AIMV RNAs 3 and 4 were determined by measuring the absorbance of the bands in the fluorograph with a densitometer equipped with peak integrator (LKB Ultroscan Laser Densitometer).
Acknowledgements We thank Edward Halk for providing AIMV coat protein, Mark Young for providing the CCMV coat protein, and Monsanto Company for providing the TMV coat protein clone. We are indebted to Peter Kim for use of the Aviv CD spectropolarimeter, to Alan Frankel for patient guidance during the CD analyses and data interpretation, and to Phillip Sharp for use of the Phosphorimaging apparatus. Alan Frankel, Roger Kaspar, Jane-Jane Chen, Louane Hann, Patty Ansel, Jamie Williamson, Jody Puglisi and Philippe LeBoulch participated in many useful discussions, and Judith Wu, Jun-Ming Cai and Peter Chefalo provided expert technical assistance. Peptides were prepared at the Biopolymers Laboratory, Howard Hughes Medical Institute, Massachusetts Institute of Technology. This work was supported by awards from the Whitaker Health Sciences Fund and the National Institutes of Health (GM 42504).
References Bianchi,M.E., Beltrame,M. and Paonessa,G. (1989) Science, 243, 1056-1059. Bol,J.F., Van Vloten-Doting,L. and Jaspars,E.M.J. (1971) Virology, 46, 73 -85. Bol,J.F., Kraal,B. and Brederode,F.T. (1974) Virology, 58, 101-110. Calnan,B.J., Biancalana,S., Hudson,D. and Frankel,A.D. (1991a) Genes Dev., 5, 201-210.
734
Calnan,B.J., Tidor,B., Biancalana,S., Hudson,D.
and Frankel,A.D. (1991b) Science, 252, 1167-1171. Creighton,T.E. (1984) Proteins: Structures and Molecular Principles. W.H.Freeman and Company, New York. Daly,T.J., Rusche,J.R., Maione,T.E. and Frankel,A.D. (1990) Biochemistry, 29, 9791-9795. Dreyfuiss,G., Matunis,M.J., Pinolroma,S. and Burd,C.G. (1993) Annu. Rev. Biochem., 62, 289-321. Driedonks,R.A., Krijgsman,P.C.J. and Mellema,J.E. (1978) Eur. J. Biochem., 82, 405-417. Garcia-Arenal,F. (1988) Virology, 167, 201-206. Goelet,P., Lomonossoff,G.P., Butler,P.J.G., Akam,M.E., Gait,M.J. and Karn,J. (1982) Proc. NatlAcad. Sci. USA, 79, 5818-5822. Gonsalves,D. and Garnsey,S.M. (1975) Virology, 67, 319-326. Hall,T.C. (1979) Int. Rev. Cytol., 60, 1-26.
Hashizume,H. and Imahori,K. (1967) J. Biochem. (Tokyo), 61, 738-749. Hershey,J.W.B. (1991) Annu. Rev. Biochem., 60, 717-755. Hinz,H., Srinivasan,S. and Jaspars,E.M.J. (1979) Eur. J. Biochem., 95, 107- 114. Houser,F., Baer,M.L., Liem,K.F.,Jr, Cai,J.-M. and Gehrke,L. (1994) J. Virol., in press. Houwing,C.J. and Jaspars,E.M.J. (1978) Biochemistry, 17, 2927-2933. Houwing,C.J. and Jaspars,E.M.J. (1980) Biochemistry, 19, 5244-5260. Houwing,C.J. and Jaspars,E.M.J. (1982) Biochemistry, 21, 3408-3414. Jaspars,E.M.J. (1985) In Davies,J.W. (ed.), Molecular Plant Virology. CRC Press, New York, pp. 155-221. Jobling,S.A., Cuthbert,C.A., Rogers,S.G., Fraley,R.T. and Gehrke,L. (1988) Nucleic Acids Res., 16, 4483-4498. Kan,J.J., Andree,P.J., Kouijzer,L.C. and Mellema,J.E. (1982) Eur. J. Biochem., 126, 29. Kao,C.C. and Ahlquist,P. (1992) J. Virol., 66, 7293-7302. Kjems,J., Brown,M., Chang,D.D. and Sharp,P.A. (1991a) Proc. NatlAcad. Sci. USA, 88, 683-687. Kjems,J., Calnan,B.J., Frankel,A.D. and Sharp,P.A. (1991b) EMBO J., 11, 1119-1129. Koper-Zwartoff,E.C. and Bol,J.F. (1980) Nucleic Acids Res., 8, 3307-3318. Kruseman,J., Kraal,B., Jaspars,E., Bol,J., Brederode,F. and Veldstra,H. (1971) Biochemistry, 10, 447-455. Loesch-Fries,L.S. and Hall,T.C. (1980) J. Gen. Virol., 47, 323-332. Loesch-Fries,L.S., Halk,E.L., Nelson,S.E. and Krahn,K.J. (1985) Virology, 143, 626-629. Loesch-Fries,L.S., Jarvis,N.P., Krahn,K.J., Nelson,S.E. and Hall,T.C. (1985) Virology, 146, 177-187. Mans,R.J. and Novelli,G.D. (1961) Arch. Biochem. Biophys., 94, 48-53. Mattaj,I.W. (1993) Cell, 73, 837-840. Matthews,R.E.F. (1991) Plant Virology. Harcourt Brace Jovanovich, New York. Milligan,J.F., Groebe,D.R., Witherell,G.W. and Uhlenbeck,O.C. (1987) Nucleic Acids Res., 15, 8783-8798. Moore,M.J., Query,C.C. and Sharp,P.A. (1993) In Gesteland,R. and Atkins,J.F. (eds), The RNA World. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 303 -357. Nassal,M. (1992) J. Virol., 66, 4107-4116. Nassuth,A., Alblas,F. and Bol,J.F. (1981) J. Gen. Virol., 53, 207-214. Peter,R., Stehelin,D., Reinbolt,J., Collot,D. and Duranton,H. (1972) Virology, 49, 615-617. Pleij,C.W.A., Rietveld,K. and Bosch,L. (1985) Nucleic Acids Res., 13, 1717- 1731. Puglisi,J.D., Tan,R., Calnan,B.J., Frankel,A.D. and Williamson,J.R. (1992) Science, 257, 76-80. Quadt,R., Rosdorff,H.J.M., Hunt,T.W. and Jaspars,E.M.J. (1991) Virology, 182, 309-315. Quigley,G.J., Gehrke,L., Roth,D.A. and Auron,P.E. (1984) NucleicAcids Res., 12, 347-366. Rould,M.A., Perona,J.J., Soll,D. and Steitz,T.A. (1989) Science, 246, 1135-1142. Ruff,M., Krishnaswamy,S., Boeglin,M., Poterszman,A., Mitschler,A., Podjarny,A., Rees,B., Thierry,J.C. and Moras,D. (1991) Science, 252, 1682-1689. Samac,D.A., Nelson,S.E. and Loesch-Fries,L.S. (1983) Virology, 131, 455-462. Scherly,D., Boelens,W., Dathan,N.A., Vanvenrooij,W.J. and Mattaj,I.W. (1990) Nature, 345, 502-506. Schimmel,P. (1989) Cell, 58, 9-12. Sehnke,P.C., Mason,A.M., Hood,S.J., Lister,R.M. and Johnson,J.E. (1989) Virology, 168, 48-56.
Alfalfa mosaic virus RNA - peptide interactions Singer,R.H. (1993) J. Cell. Biochem., 52, 125-126. Spirin,A.S. and Ryazanov,A.G. (1991) In Trachsel,H. (ed.), Translation in Eukaryotes. CRC Press, London, pp. 325-349. Srinivasan,S. and Jaspars,E.M.J. (1978) Biochim. Biophys. Acta, 520, 237-241. Srinivasan,S. and Jaspars,E.M.J. (1982) Biochim. Biophys. Acta, 696, 260-266. Steitz,T.A. (1990) Q. Rev. Biophys., 23, 205. Tan,R.Y. and Frankel,A.D. (1992) Biochemistry, 31, 10288-10294. Tremaine,J.H. and Ronald,W.P. (1978) Virology, 91, 164-172. Tremaine,J.H., Agrawal,H.O. and Chidlow,J. (1972) Virology, 48, 245-254. Tremaine,J.H., Ronald,W.P. and Agrawal,H.O. (1977) Virology, 91, 164-172. Tritton,T.R. and Crothers,D.M. (1976) Biochemistry, 15, 4377-4385. van Beynum,G.M.A., de Graaf,J.M., Castel,A., Kraal,B. and Bosch,L. (1977) Eur. J. Biochem., 72, 63-78. van Boxsel,J.A.M. (1976) High-affinity Sites for Coat Protein on Alfalfa Mosaic Virus. Ph.D. Thesis. University of Leiden, The Netherlands. van Vloten-Doting,L. (1975) Virology, 65, 215-225. Verhagen,W., van Boxsel,J.A.M., Bol,J.F., van Vloten-Doting,L. and Jaspars,E.M.J. (1976) Ann. Microbiol. (Inst. Pasteur), 127a, 165. Weinberger,J., Battimore,D. and Sharp,P. (1986) Nature, 322, 846-848. Zapp,M.L., Hope,T.J., Parslow,T.G. and Green,M.R. (1991) Proc. Natl Acad. Sci. USA, 88, 7734-7738. Zuidema,D., Bierhuizen,M.F.A. and Jaspars,E.M.J. (1983) Virology, 129, 255-260. Received on June 7, 1993; revised on October 11, 1993
735
