Is there a way of doing this ..., using only substr function?

As has been shown, you can. But why would you?

tr *ISN'T* a regular expression. It is a compile time constructed table lookup; designed for exactly this purpose; and it will do the job anywhere from 100x to 400 78x & 270 times faster than the other methods offered here:

#! perl -slw
use strict;
use Benchmark qw[ cmpthese ];

our $seq = "ACGGGAGGACGGGAAAATTACTACGGCATTAGCacgggaggacgggaaaattactacg
+gcattagc";

our %xref = (
  A => 'T', C => 'G', G => 'C', T => 'A',
  a => 't', c => 'g', g => 'c', t => 'a',
);
our %rc = ( A => q{T}, T => q{A}, C => q{G}, G => q{C}, );

cmpthese -1, {
    tr => q[
        ( my $revcmp = reverse $seq ) =~tr[ACGTacgt][tgcaTGCA];
    ],
    davido => q[
        my $seq = $seq;
        for(my $ix = 0; $ix < length $seq; ++$ix ) {
            substr( $seq, $ix, 1, $xref{ substr( $seq, $ix, 1 ) } );
        }
        my $reverse = reverse($seq);
    ],
    atcroft => q[
        my $complement;
        my @letters = split //, $seq;
        while ( @letters ) {
            my $l = uc pop @letters;
            $complement .= $rc{$l};
        }
    ],
    hdb => q[
        my $rev='';
        my $n = length $seq;
        while( $n-- ){
            $rev .= $_ for map chr( $_ & 2 ? $_^4: $_^21 ), ord substr
+ $seq, $n, 1
        }
    ],
}

__END__
C:\test>junk60
             Rate     hdb atcroft  davido      tr
hdb        9827/s      --    -35%    -71%   -100%
atcroft   15077/s     53%      --    -55%    -99%
davido    33822/s    244%    124%      --    -99%
tr      2679052/s  27163%  17669%   7821%      --
[download]

(I know you guys know this; but does the OP??? And if they do, why are they asking for this?)