'They might add a "R-1" or "R-2" to the far left column if there is a revision.'

You just need to extend the regex to handle that. Here's an example:

#!/usr/bin/env perl

use strict;
use warnings;

use Data::Dumper;

my $jid = '777';
my $text = 'header 777 111 777 A 222 777R-1 333 777R-2 A 444';
my $re = qr{(?x: \A (R-\d+|) \s*? (A|) \s (\d+) )};
my @lines;
my $wanted_line = 0;

for my $line (split /$jid/, $text) {
    next unless $wanted_line++;
    my @fields = $line =~ $re;
    push @lines, [ $jid . shift(@fields), @fields ];
}

print Dumper(\@lines);
[download]

Output:

$VAR1 = [
          [
            '777',
            '',
            '111'
          ],
          [
            '777',
            'A',
            '222'
          ],
          [
            '777R-1',
            '',
            '333'
          ],
          [
            '777R-2',
            'A',
            '444'
          ]
        ];
[download]

print ... $fields[1] . ",". $fields[3] . ",". $fields[4] . ",". ...

Here's an example to show a better way to handle that:

$ perl -e 'my @x = qw{a b c d e f}; print join ",", @x[0,3,4]'
a,d,e
[download]

On an unrelated note, there are problems with your open statements. Use of package variables can lead to all sorts of bugs that are hard to track down. Your six error messages are identical: how will you know which file generates "Can't open the output file ...". Look to using lexical filehandles and the 3-argument form of open. Consider the autodie pragma — you'll do less work and get better error reporting.