cDNA for pyrosequencing

The question that I get most is how to prepare cDNA for pyrosequencing. Pyrosequencing of complementary DNA (cDNA) poses a unique set of drawbacks, but is ultimately related to the issues of pyrosequencing homopolymers. Eukaryotic cDNA is transcribed from messenger RNA (mRNA) typically using commercially available kits that use a Poly-T primer (i.e., Invitrogen SuperScript). Because eukaryotic mRNA is poly-adenylated on the 3’ end, the poly-T primer is used to target this region, as T is the complement of A for pairing. The one drawback of pyrosequencing cDNA libraries is the presence of long regions of homopolymers. In our experience, the accuracy of Roche/454 Life Sciences base calling on homopolymers diminishes markedly after the 5th or 6th base in a homopolymer stretch. If by the third flow no nucleotides are incorporated, that particular read is discarded. This problem of data loss by the A homopolymers is exasperated by the fact that the intensity of the bead with the 30+ As is also obscuring neighboring beads, thus causing additional loss of data. Finally, the faulty reads obscure the control reads such that even the data collected cannot be used with much confidence. Fortunately, a recent, simple modification solves the polyA dilemma with Roche/454 Life Sciences pyrosequencing.
Since the confounding issue is the presence of the polyA tail on the 3’ end of the cDNA, Novaes et al (2008) have incorporated a modified 3’ primer that has a rare restriction site (SfiI) built in the 3’ cDNA adapter primer. Following relatively standard cDNA protocols, the additional step of digesting the cDNA with SfiI and removal of the polyA tail via column or Ampure purification greatly increases the percent passing reads and data yield. A similar solution has been proposed by Frias-Lopez et al (2008) for bacterial cDNA. Bacterial cDNA (which lack adenylation on their mRNA) are polyadenylated via ligation, thus making them effectively eukaryotic cDNA for downstream applications. A rare restriction enzyme (BpmI) recognition sequence is inserted between the polyT and the T7 promoter sequence thus allowing removal of the confounding polyA region prior to pyrosequencing library construction.