Multiple proteins and subgenomic mRNAs may be derived from a single open reading frame on tobacco mosaic virus RNA.

It has previously been shown that messenger activity for a protein of Mr = ca. 30k exists in RNA fractions extracted from particles of either native or alkali stripped U1 TMV, or from cowpea strain TMV, that are smaller than full genomic length. Analysis of sucrose gradient fractions containing this activity reveals a number of slightly smaller template activities directing synthesis of proteins between 18.5k and 29k in size. All of these messenger activities, including that for the 30k protein, respond to cap analogues in anomalous ways. Discrete RNA species that include active mRNAs for these proteins can be demonstrated in the same fractions by labelling with preparations of vaccinia capping enzyme and [alpha-32P] GTP without prior beta-elimination. Detailed analysis of three of these proteins (of Mr's ca. 30k, 29k and 23k) by peptide mapping and translation of purified vaccinia-labelled RNA demonstrates that all three are unrelated to the large early TMV proteins, but are related to each other in such a way as to form a nested set with staggered N termini and identical C termini. mRNAs of chain lengths ca. 1900 and 1500 bases direct synthesis of the 30k and 23k proteins respectively, an mRNA of about 1850 bases directs both 29k and (perhaps because of cross-contamination) 30k synthesis. Initiation codons for the 29k and 23k proteins have been mapped at positions 4960-4962 and 5191-5193 respectively on TMV RNA. Since all three encapsidated templates have similar properties we conclude that either there is a family of 30k-related proteins with unusual mRNAs, or that none of these in vitro translation products are directed by physiological templates.