Panels of monoclonal antibodies were raised against viral non-structural proteins of encephalomyocarditis virus (EMCV) and used to probe infected cells in laser confocal microscopy experiments and Western analyses. Surprisingly, all Mengovirus and EMCV-infected cells showed strong targeting of protein 2A, 3B(VPg), 3C(pro), and 3D(pol) signals to cellular nuclei, in particular to nucleoli, from the earliest times of infection. Viral capsid proteins (1AB, 1C, and 1D) and other non-structural proteins (2B, 2C, and 3A) did not target nuclei and remained cytoplasmic throughout the infection. The cardioviral 2A protein (subject of this article) has a novel 143 amino acid sequence, terminating in a 19 amino acid COOH-terminal processing cassette (PCC) that participates in autocatalytic, co-translational primary cleavage of the viral polyprotein. The remainder of the 2A protein shares only limited similarity with other viral or cellular sequences, except for a short motif (KRvRPFRLP) near PCC resembling the nuclear localization signals (NLS) common to many yeast ribosomal proteins. Deletions within the EMCV 2A protein that impinge on this region have been reported to diminish the ability of virus to inhibit cap-dependent translation of cellular mRNAs. We have now observed that these same deletions prevented nuclear localization. Cellular expression of 2A protein from RNA transcripts or cDNAs confirmed that it does not require other viral proteins or activities for nuclear transport; even when expressed as a single protein, 2A protein effectively shuts off translation from capped reporter mRNAs. Within infected, transfected, or DNA vector-transformed cells, the 2A protein was always found in close association with the nucleolar ribosomal chaperone protein B23, which may help the traffic 2A into nucleoli like a surrogate ribosomal protein, by virtue of the putative nucleolar localization signal (NoLS). The data are consistent with a novel mechanism for virus-induced host protein shut off in cardioviruses, whereby 2A helps to upregulate the synthesis of new and modified ribosomes that have an inherent preference for internal ribosomal entry site (IRES)-dependent viral genome translation over cap-dependent host mRNA translation.