I use the following snippet from http://www.perlmonks.org/?abspart=1;displaytype=displaycode;node_id=308179;part=1
for parsing fasta format files: Unfortunately, I can't find this script's author. I just modified it slightly to try your problem.

use strict;
use Data::Dumper;

my ($counter, $line_count, @rec);
{
  local $/ = '>';
  while (<DATA>) {
    
    s/^>//g; # strip out '>' from beginning
    s/>$//g; # and end of line
    next if !length($_);           # ignore empty lines

    my ($header_info) = /^(.*)\n/; # capture the header
    s/^(.*)\n//;                   # and strip it out

    push @rec, $header_info;
    s/\n//mg;                      # join the sequence strings
    $counter++ if $_=~/ACGACACGTAT/;

  }
}

$line_count = scalar @rec;
print "$line_count  seqeuences were tested and $counter seqeuences hav
+e a tag\n";


__DATA__
>GJVIMO101AUT0H length=45 xy=0234_0223 region=1 run=R_2010_07_01_11_09
+_50_
ACGACACGTATACGTGCGTGTCGCGTCTCTCAGCACACAGAGTAG
>GJVIMO101ANKZK length=45 xy=0151_1902 region=1 run=R_2010_07_01_11_09
+_50_
ACGACACGTATCGCGCGCGNGCGCGCGCGCGCGCGCGCGCGCGCG
>GJVIMO101AOIE9 length=41 xy=0162_0179 region=1 run=R_2010_07_01_11_09
+_50_
ACGACACGTATCTCATTGTGCTCAAGGCCTGAGCACAATGA
>GJVIMO101ALCLG length=100 xy=0126_0114 region=1 run=R_2010_07_01_11_0
+9_50_
ACGACACGTATGCTGCTGGTGCTGCTGTAACAGTTCCTGCTGATGCTGCAAGTGCTGCTG
CTGTAACTGTTGCTGCTGTAATCTCTGCTGCTGCTGCTGT
>GJVIMO111ALCLG length=100 xy=0126_0114 region=1 run=R_2010_07_01_11_0
+9_50_
GCTGCTGGTGCTGCTGTAACAGTTCCTGCTGATGCTGCAAGTGCTGCTG
CTGTAACTGTTGCTGCTGTAATCTCTGCTGCTGCTGCTGT
[download]