screening phage library with oligo

HARDIES at THORIN.UTHSCSA.EDU HARDIES at THORIN.UTHSCSA.EDU
Fri May 12 11:35:37 EST 1995


Steve Goldberg writes:

> I have started to probe a pig liver cDNA library using a labeled
oligonucleotide.  There appears to be inumerable variations on
the different steps of the protocol.  I would be interested to receive
comments or tips on the following steps (especially those that were
successful!):

> 1.  Transferring and fixing phage DNA:  Some procedures say immerse
the filters in the denaturation and neutralization; others claim this
gives diffuse signals.

Heavy signals can give some smearing or streaking when the filters are 
dunked.  It's usually not bad enough to obscure the result, but the
effort to float or put on a wet stack of paper towels is minimal, so
we do that.  On the other hand, if there are so many filters to process
that this becomes a bottleneck, we dunk them and get on with it.

> 2.  End-labeling oligonucleotides:  How much DNA is needed, how much
label (again, I've seen anywhere from 5 t0 25 ul of gamma-dATP rec-
commended).  

The key is not to have a large molar excess of oligo over molar amount
of rATP.  Otherwise, you're necessarily competing with an excess of
cold oligo to no good purpose.  Generally 5-10 pmole of oligo & rATP
is enough; but for optimal performance you would have to increase the
amount of both in proportion to the volume of hybridization mix to 
avoid becoming kinetically limited.

> How about the ECL 3' labeling kit or similar?  

This can give higher sensitivity by incorporating multiple labelled
bases/oligo.  In this case you need 6-10 x molar excess of label over
oligo.  If you go in with more oligo than label, you blow the point
of the procedure.

However, if you have any sensitivity problem, I recommend the plaque
amplification procedure of Salvio and Woo.  This dramatically amplifies
your signal much more than any of the fine tuning of the labelling
reaction can; and makes the experiment very forgiving of technical
problems.  It does introduce a new background problem:  you have
to scrub the lawn off of the filters before hybridization to 
avoid a splotchy background.  If you do this, really scrub them against
each other hard; you can't scrap off the DNA.

> 3.  Pre-hybridization and hybridization conditions.  Some use SSPE,
others SSC; some include pyrophosphate; others leave out Denhardt's
in the hybridization.

We've slogged through these and a bunch more noise reduction recommendations
and have a hard time seeing that any of them make a difference.  SSPE and
SSC seem to perform the same.  However, a purist would note that SSC is
a lousy buffer in this pH range.  Pyrophosphate is supposed  to compete
for rATP that carries over into the hybridization.  However, I can't
seem to raise a background with rATP even if I directly load microcuries
of it into the hybridization.  Prehybridization with something seems important,
minimally either Denhardt's or SDS.  Different filters have different
properties, so if the manufacturer calls for SDS, do it.  Don't put
competitor DNA or RNA in the hybridization mix.  At best it doesn't
have any complementarity to the oligo and hence is irrelevant.  At worst
it'll compete out your signal.  It might help to put competitor DNA/RNA in
the prehyb., where it may cover up unbound sites on the filter.  The 
temp. and duration of the prehyb. seem less important than just doing
it.  We find that the best defense against getting fooled by spurious
spots on the filters is just to do duplicate plaque lifts.  If you
get a spotty background the problem is in the hyb. mix; if you get a
splotchy background the problem is too much stuff (rseg. aga stuck 
to the filter;
if you get a solid grey or black background, you probably need more
SDS.  Some filters, particularly nitrocellulose, generate a background
when they get old; and the only solution is to get new filters.

> 4.  Wash conditions.  

Wash conditions are critically dependent on your oligo.  If you can
predict the Tm, stay about 10 degrees under it [for a margin of safety].
If you can't predict the Tm (because you have an unknown number of 
mismatches)  then start washing at a low temp. (~40 unless you think you're
match could be particularly unstable), expose wet, and
repeat at 5 degree increments until there are only a few spots left.
Then hope for the best.  Note that most Tm prediction equations are
designed to work in 1 M Na, which is the concentration found in the
typical wash solution for oligos (6xSSC).  If you choose a different
wash solution, make sure you maintain this salt concentration, or the Tm
could be way off.



More information about the Methods mailing list