01 August 1992
RE: SOUTHERN SMILES
Last 30 July Dr. Raymond Dalgleish of the U. of Leiceste
lanation on the influence of ethidium bromide
on the smiling of bands during agarose gel electrophoresis.
My labmates and I find his hypothesis quite plausible. Dr.
Dalgleish stressed that in gels cast with EtBr, DNA samples
at each lane are not uniformly intercalated with the dye
particularly at high DNA concentration. Those fragments
located at the periphery of the lane awould be exposed to a
more dense dye environment while those at the middle would
tend to have fewer dye molecules intercalated in them.
Consequently, the effective weights of DNA fragments of
supposedly equal sizes but of diverse extent of
intercalation with EtBr will be different and likewise their
Just recently after Dr. Dalgleish posted his message, we
received another on this topic but it tries to dispute
Dalgeish's hypothesis. The sender pointed out that it is
just overloading which contributes to smiling of bands. I
definitely agree. But this sender continued that ethidium
ion which can move oppositely relative to DNA migration is
"replenished" such that the limiting effect of the dye does
not hold. Well I suppose that if the cation ethidium moves
considerably during electrophoresis, then the dye would
eventually run off from the gel and would actually be
depleted and made more limiting.
Our team does a lot of RFLP analysis on the rice blast
fungus. We use a probe consisting of a highly repeated DNA
sequence that hybridizes to more than 50 closely migrating
EcoRI fragments of the fungal genomic DNA. We use 18x30 cm
gels for our runs and we find it convenient to cast our gels
with EtBr. But we fail to get decent bands. In fact smiling
could sometimes obscure our RFLP patterns making it
impossible to score the bands. Now we thought before that
maybe it was just DNA overloading. Then we tried to work at
progressively lower load size without much success. In fact
we have gone to as low as 0.5 fg load per lane but smiling
of bands was still apparent and worse each time we work with
lower sample size, we fail to detect increasingly certain
hybridization signals. Then one of us last month tried to
run his gels without EtBr just for kicks and surprisingly
the smiles disappeared. This was repeated several times and
even up to 1.5 fg of DNA per lane, bands were miraculously
straight. At this level, we typically get ghastly smiling
bands when the gels were laced with EtBr. I am not saying
though that overloading has become history. I expect that
somewhere above 1.5 fg, our fragments would start grinning.
But we are quite glad now that we can manage with a
satisfactory level loading size that will give easily
scoreable bands but also won't compromise the sensitivity of
detecting hybrids with our probe.
G. J. Boulnois in his book Gene Cloning & Analysis
(Blackwell, Oxford, 1987) suggested the addition of a
microliter of 10 mg/ml EtBr in each lane after sample
loading to preclude smiling. I have this phobia really with
EtBr so I never tried his suggestion. But if it really works
then I suppose it supports Dalgleish's hypothesis.
Supplementing EtBr in each lane would tend to saturate the
DNA binding capacity creating a more homogenous effective
weight distribution per fragment size. Now I remember, we
only fing smiling in our heavy fragments. It must be that
the lower fragments are much easily saturated with EtBr. Or
it could also be that with ethidium gradually depleted
(assuming it runs off due to the electric field) the lower
fragments become less influenced by the dye by moving
towards a gradient of increasingly lower dye conc.
Grinning bands could be quite disgusting and a serious
nuisance and it is unfortunate that published protocols
hardly discuss this matter in length. I appreciate Dr.
Dalgleish's effort to explain the situation. I also hope
others would find our experience helpful and worth trying.
R. P. Scott
Plant Pathology Division
International Rice Research Institute