Ghost bands on plasmid preps

Michael R. Elliott melliott at
Tue Apr 3 14:49:50 EST 2001


I have experienced something similar in the past, and I might be able to help.
First, I don't know that the bands you are seeing are ssDNA, as EtBr
intercalates very inefficiently to ssDNA, and whatever EtBr that does
intercalate is thought to be the result of duplex DNA formation.  You can test
this ssDNA theory by running some of the undigested plasmid prep on a gel next
to the Bam/RI cut DNA-- if there's not a band in the uncut DNA that corresponds
to the digested DNA lane, then you know it is something that has occured during
the restriction digest.  You should call the company that makes the enzymes and
see if any complaints have been logged for the EcoRI and BamHI.  A few months
ago we had a similar problem with EcoRI from Promega. They sent us a new batch
and the problem went away.  The extra bands may be due to inappropriate cleavage
of your plasmid (called "starring").

One last suggestion: the migration of  DNA thru the gel is effected by the salt
concentration of the sample prior to loading in the gel.  I know that when I
elute purified linear DNA in H2O, I often get smaller bands than expected.  I
can correct the problem by adding 1x of any restriction enzyme buffer and
letting the sample sit for 2-3min on the bench before loading on the gel.  Good

Rusty Elliott
Wake Forest University
North Carolina.

Daniel MacArthur wrote:

> Dear group,
> I'm an undergraduate biochemistry student currently attempting to write up a
> practical report on a plasmid miniprep practical. Briefly, we took two E.
> coli colonies transfected with pBluescript IISK+ vectors (one containing a
> fragment from an EcoRI digest of lambda phage DNA and one with no insert),
> purified the plasmids using a Quantum Miniprep from Bio-Rad, digested the
> plasmid DNA using EcoRI, BamHI, and a combination of the two, and then ran
> out the products on an agarose gel.
> The experiment worked beautifully, but I'm having a little trouble with
> interpreting the results. In several of the lanes there are faint bands
> running at roughly half the size of the full-length linear plasmid, which my
> demonstrator called "ghost bands". She told us that these bands were
> actually single-stranded plasmid molecules formed during the cell lysis step
> of the miniprep, which seems reasonable given their size in relation to the
> full-length plasmid. However, we were also referred to two articles by Paul
> Hengen (who I believe posts/posted on this newsgroup) which seem to be
> putting forward considerably more complex explanations for the existence of
> these "ghost bands" (e.g. that the bands represent "double-stranded, cyclic,
> coiled DNA composed of two intertwined, but permanently denatured, single
> strands of plasmid DNA").
> Does anyone have any suggestions regarding the nature of these bands? My
> demonstrator's explanation is attractively straightforward and seems to fit
> the data with respect to the plasmid without an insert (the linear plasmid
> runs at ~3 kb and the ghost band at ~1.5 kb), but it doesn't quite fit with
> respect to the plasmid with an insert (linear at ~7.3 kb, ghost at ~2.7 kb),
> and seems to be contrary to Paul Hengen's explanations. Any advice on this
> matter would be greatly appreciated.
> Daniel MacArthur.
> 3rd Year Medical Science
> University of Sydney, Australia

More information about the Methods mailing list