Here's a command-line version (see perlrun) (update: this code processes line-by-line, so it should work on any file — unless a "line" is more than a couple GB long! :). Caution: This only works with Perl versions 5.10+ because it uses the  \K regex operator (see Extended Patterns). If you have a pre-5.10 Perl version, let me know; a simple fix can be had (update: what the heck; the pre-5.10 substitution is  s{ ^ (> \d+) \n }{$1\t}xms with all else the same).

c:\@Work\Perl\monks\yueli711>type 1.fa
>1
AGTCGTAGCAT
>2
TGAGCTACG
>3
GGCATAGN
>4
CGCACNCAGCTACACC
>5
NGATAGCTACA

c:\@Work\Perl\monks\yueli711>perl -pe "s{ ^ > \d+ \K \n }{\t}xms"  1.f
+a  > 1.txt

c:\@Work\Perl\monks\yueli711>type 1.txt
>1      AGTCGTAGCAT
>2      TGAGCTACG
>3      GGCATAGN
>4      CGCACNCAGCTACACC
>5      NGATAGCTACA
[download]

This runs under Windows. I can't test this, but I think if you're running under *nix, just replace all the  " (double-quotes) in the command-line with  ' (single-quotes).