cloning 16S rDNA TRFLP bands

Hello all,
I'm looking to clone and sequence 16s rDNA TRFLP bands. What a
nightmare! I'm running IR-labelled PCR product digests on a Li-Cor
acrylamide based DNA sequencer. With the help of an IR viewer I can
cut out the DNA fragments from the gel. The problem is in the cloning
and sequencing of the bands. I must go through a cloning step first
for many of the bands, since I may have 2 or more distinct PCR
digestion products co-migrating at the same bp resolution. In addition
the samples have been denatured, and are at very low quantities, and
for reamplification I only know the sequence for one priming site.
Some procedural thoughts:

1. Cut band out of gel, re-amplify/second strand synthesis with
classical random priming or phi29 based random amplification (any
thoughts?), use previously engineered restriction site within primer
to cut off the 5' IR-label (the other end is already blunt), then clone
into dephosphorylated vector.

2. Cut bands out of gel, re-anneal (with the hope that similar
sequences don't form chimeras), digest then clone? I'm guessing low
nanogram to sub-nanogram quantities here.

3. Cut bands out of gel, re-anneal (with the hope that similar
sequences don't form chimeras), ligate in an adaptor to the end without
a known priming site, re-pcr, and clone?

4. Quit science, concentrate more on mountain biking here in Colorado,
let my wife deal with making money.

Gary