The sequence of the 3'-terminal 180 and 140 nucleotides of RNAs 2 and 3, respectively, of tobacco streak virus (TSV) was deduced by reverse transcription in the presence of a specific primer and chain terminators. Homology between the two RNAs was found to be restricted to a 3-terminal region of about 45 nucleotides. The data were compared with the sequence of the homologous region of 145 nucleotides occurring at the 3'-termini of the alfalfa mosaic virus (A1MV) RNAs, which contains the specific binding site for coat protein (Koper-Zwarthoff et al., Nucleic Acids Res. 7, 1887-1900 (1979); Houwing and Jaspars, Biochemistry 17, 2927-2933 (1978)). This was done because of the evidence that the RNAs of A1MV and TSV contain specific binding sites for their own as well as each others coat protein, and that binding of coat protein to these sites is required to initiate infection (Van Vloten-Doting, Virology 65, 215-225 (1975)). The 3'-terminal homologous regions of A1MV and TSV have two features in common: the presence of several stable hairpins and the multiple occurrence of the tetranucleotide sequence AUGC. The hairpins cause the linear array of tandemly repeated AUGC-boxes. It is postulated that the primary interaction of coat protein molecules with the RNAs of AlMV and TSV is a cooperative process involving several binding sites each being composed of a hairpin flanked at its 3'-side by an AUGC-sequence.

A 226-nucleotide fragment was derived from alfalfa mosaic virus RNA 4 (ALMV RNA 4), the subgenomic messenger for viral coat protein, and its sequence was deduced by in vitro labeling with polynucleotide kinase and application of RNA sequencing techniques. The fragment contains the 3'-terminal 45 nucleotides of the coat protein cistron and the complete 3'-noncoding region of 182 nucleotides. The total length of RNA 4 was calculated to be 881 nucleotides. AlMV RNAs 1, 2 and 3 were elongated with a 3'-terminal poly(A) stretch and subjected to sequence analysis by using a specific primer, reverse transcriptase and chain terminators. This revealed and extensive homology between the 3'-terminal 140 to 150 nucleotides of all four ALMV RNAs. Despite a number of base substitutions, the secondary structure of the homologous region is highly conserved. The observed homology indicates that, as with RNA 4, the sites with a high affinity for the viral coat protein are located at the 3'-termini of the genomic RNAs.

The sequence of the 5'-terminal 74 nucleotides of alfalfa mosaic virus RNA 4, the mRNA for the viral coat protein, has been deduced by using various new techniques for abe ing the RNA at the 5' end with 32P and for sequencing the 5'-32P-labeled RNA. The sequence is NpppGUUUUUAUUU-UUAAUUUUCUUUCAAAUACUUCCAUCAUGAGUUCUUC-ACAAAAGAAAGCUGGUGGGAAAGCUGG. The AUG initiator codon is located 36 nucleotides in from the 5' end; the nucleotide sequence beyond corresponds to the amino acid sequence of the coat rotein. This 5' noncoding region is rich in U (58% U); except for the 5'-terminal G, the next G in is part of the initiator AUG codon. "Cap" structures of the type m7G(5')ppp(5')N are found at the 5' terminus of many eukaryotic mRNAs and have been implicated in ribosome binding (1). The detection of other possible common signals has been hampered by the absence, until recently, of comparative sequence data on the 5'-noncoding regions of these molecules. For some time, only the sequence of the nine nucleotides between the "cap" and the AUG initiator codon in brome mosaic virus (BMV) RNA 4 was known (2). Recently, nucleotide sequences preceding the initiator codon have been published for rabbit a-and O-globin mRNA (3-5), the simian virus 40 VP1 gene (6), oncornaviral RNA (7-9), reovirus mRNA (10), vesicular stomatitis virus mRNA (11), and tobacco mosaic virus RNA (12). Comparison of these data reveals no readily identifiable recognition signals and shows that the distance between the "cap" and the first AUG codon is variable. The synthesis of these RNA species involves various different enzymes and templates, which may be either DNA or RNA, and the sequence at their 5' end may be determined by constraints other than ribosome recognition alone.We have analyzed the sequence of the 5'-noncoding region in RNA 4 of alfalfa mosaic virus (AlMV), which is functionally equivalent to BMV-RNA 4. AIMV, like BMV, has a genome consisting of three RNA molecules, the coat protein cistron being located in RNA 3, the smallest genome fragment (13 making up the BMV genome are infectious as such, the AIMV genome has to be supplemented with a small amount of coat protein, or the RNA 4, which presumably acts as the coat protein messenger (19). Moreover, the overall sequence homology between the genomes of the two viruses is negligible (20).In the course of this study, we made use of a number of novel rapid sequencing techniques, including a gel sequencing method similar in principle to the DNA sequencing method of Maxam and Gilbert (21) and a "wandering spot" analysis that uses two-dimensional polyacrylamide gels (details to be published elsewhere). All the methods involve first the in vitro enzymatic labeling of the 5' terminus of AIMV RNA 4, after removal of the "cap" by periodate oxidation, #-elimination, and phosphomonoesterase digestion (3). The results enable us to deduce the sequence of the 5'-terminal 74 nucleotides of AIMV RNA 4. The AUG initiator codon of the coat protein cistron was found to be preceded by 36 nucleotides. The ...

Alfalfa mosaic virus RNA 4, the subgenomic messenger for viral coat protein, was partially digested with RNase T1 or RNase A and the sequence of a number of fragments was deduced by in vitro labeling with polynucleotide kinase and application of RNA sequencing techniques. From overlapping fragments, the complete primary sequence of the 881 nucleotides of RNA 4 was constructed: the coding region of 660 nucleotides (not including the initiation and termination codon) is flanked by a 5' noncoding region of 39 nucleotides and a 3' noncoding region of 182 nucleotides. The RNA sequencing data completely confirm the amino acid sequence of the coat protein as deduced by Van Beynum et al. (Fur.J. Biochem. 72, 63-78, 1977).

Nucleotide sequences at the 5'-termini of the alfalfa mosaic virus genomic RNAs and the intercistronic junction in RNA 3 were deduced and compared to identify possible common recognition signals for replicating enzymes in the corresponding minus-stranded viral RNAs. Homology between the 5'-terminal sequences is less than 11 nucleotides and no complementarity with the homologous sequence occurring at the 3'-end of the viral RNAs was observed. Homology between the 5'-terminus and intercistronic region in RNA 3 is compatible with the synthesis of subgenomic RNA 4 by internal initiation of transcription on the RNA 3 minus strands. The sequence around the intercistronic junction can be folded into a very stable secondary structure.