Why A and B primers in SAGE? Why not just one?

[abaldor]

Reading about SAGE (serial analysis of gene expression). Everywhere discusses adapters having A or B primer but no one explains explicitly why dividing the pool of anchoring enzyme-restricted cDNAs is necessary or why primers A and B are necessary.

At this point, I assume the lack of detail on the A and B primers mean that the A and B primers create a primer pair for amplification just like any primer pair.

Q1: is this accurate or am I already off base?

Q2: If this is accurate, why do you need to divide the pool of anchoring enzyme-restriced cDNAs? Why not add a mix of adapters, some with primer A and some with primer B? Would they not be stochastically ligated to the cDNA fragments?

Q3.a: the above notwithstanding -- Is it possible, during the formation of ditags, that a tag complimentary with primer A is ligated to another tag complimentary with primer A? I understand the cDNA fragments are repaired into blunt ends, so I imagine this would happen just fine.

Q3.b: Would a ditag complimentary with primer A at both ends not amplify using two primer As? Why do you need two primers A and B? Why not just add complimentary tags for primer A to all cDNA fragments and amplify everything with primer A?

Thanks for any insight!

[abaldor]

Also, once the ditag has been formed and the adapter is cleaved using anchoring enzyme, how is the adapter removed before the ditags are concatenated? Thanks again.