Unusual 3' RACE artifact?

David J. Meyer meyerdj at biovx1.biology.ucla.edu
Thu Nov 3 12:44:05 EST 1994

Hello, Netters!

   This is a query about the prevalence of deletion artifacts in PCR and a
solicitation of discussion of such artifacts. Nucleotide changes in PCR
are an often-discussed problem, but I am unaware of how common deletion
artifacts are in PCR. I would like also to learn how such artifacts could

Here's the story-

   I have a cDNA from a library which has a very long 3' UTR; a shorter
mRNA (which hybridizations suggest lacks the 3' UTR, although changes
elsewhere are possible) is also identified on northerns. Using
gene-specific primers based on the sequence of the long cDNA, I performed
3' RACE and identified products arising from the long cDNA (~2 kb), and
shorter species which have been cloned and are being sequenced.

   One of the shorter RACE products (~0.7 kb) corresponds exactly to the
sequence of the long cDNA but is truncated at the 3' end (so this is the
hypothesized product). However, some still shorter products were
identified: these are truncated at the 3' end relative to the long cDNA
*at the same point as the 0.7 kb product* but contain deletions within the
ORF. One of the deletions is 237 nt and so maintains the ORF. Another
deletion is ~187 nt and does not retain the ORF. Note that the shorter
deletion appears to begin at the same postion in the sequence as does the
other deletion.

   Of course, the big question is whether these deletions are 1) actually
present in some mRNAs or 2) result from an artifact during reverse
transcription or 3' RACE (the explanation I am inclined to favor).

   *Does anyone have similar observations of such artifacts from 3' RACE
or other RNA-PCR reactions? I would be very interested in hearing about
your experiences!*

   I have run RNA folding programs to see if there are secondary
structures in the RNA which lie in the region which was deleted; of course
there are some such structures in the region, but they don't jump out and
hit you over the head. (I would need to be hit over the head with an
extremely stable secondary structure with endpoints near the deletion
endpoints to belive that it is responsible for the deletions 8-)  I plan
S1 nuclease mapping in the region to address the possible existence of
different classes of messenger RNAs‹ this way I avoid more PCR, the
technique which may give rise to the potential artifact. Another
possibility I have considered is to probe northern blots with sequences
corresponding to the region possibly absent in the shorter mRNA. I am also
considering PCR (with two gene-specific primers) from cDNA and genomic DNA
around this region to see if I get multiple products or indications of
introns in the region, respectively. I would appreciate other ideas to
address the veracity of these PCR products, and your thoughts on how these
apparent deletion products could arise by an artifact in the 3' RACE or
how they could arise in the mRNAs (other that alternate exon splicing?) if
they are, in fact, real.

Thank you for your input!


David J. Meyer
Department of Biological Chemistry
University of California-Los Angeles
meyerdj at biovx1.biology.ucla.edu

More information about the Methods mailing list