something like:

LINE:
  while ( <PROTEIN> ) {
      if ( /^VERSION\s+(\S+)/ )
      {
      # extracts the first non-space sequence after the VERSION-token
          $protname = $1;
          next LINE;
      }
      if ( /^DBSOURCE\s+.*\s(\S+)\s*$/ )
      {
      # extracts the last non-space sequence on the DBSOURCE line
          $rna = $1;
          next LINE;
      }
  }
[download]

You first need to determine the syntax of the lines you want, and the location of the values you want to extract:
Is the "YP_001648463.1" always the first field after the VERSION? Is the RNA always last on the DBSOURCE lines?
You should consider what to do with invalid input: what happens if DBSOURCE is not present or $protname does not match the ACCESSION-value...
For more info on regular expressions, check out perldoc perlre.