Insert alone gives colonies?

Malay curiouser at
Tue Aug 22 15:03:03 EST 2000

Dear Dr Barrett:

Your problem seems to be very interesting. May be you should pursue it
further and you might land up something very interesting. But skepticism is
the primary criterion of any science. And I document below my explanation of
this queer phenomenon (if all the possibility of contamination is ruled out
and you are confident that so simple answer will explain the phenomenon) and
may be some possible solution. I am also sending this message to the
newsgroup in the hope that someone more experienced than me will shed some

----- Original Message -----
From: "Peter Barrett" <BARRETT at>
To: <curiouser>
Sent: Tuesday, August 22, 2000 9:27 PM
Subject: Re: Insert alone gives colonies?

> Dear Malay,
> I am Ben's supervisor.  I can respond directly to your comments:
> MIME-Version: 1.0
> Content-type: text/plain; charset=Windows-1252
> Content-transfer-encoding: 7BIT
> > No User <no.user at> wrote:
> > : Anybody ever seen something like this before? On a
> > : ligation of a long PCR fragment, much troubleshooting
> > : has shown that colonies come up on an insert alone +
> > : ligase control plate.  (Minipreps on colonies from the
> > : vector + insert plate are either vector without insert
> > : (rarely) or (mostly) some small garbage plasmid,
> > : smaller than the starting vector.  Also, a vector +
> > : insert without ligase control plate gives few or no
> > : colonies, suggesting that the colonies on the insert +
> > : ligase plate arise through ligation.)
> >Give some numbers. How many colonies? Did you try isolating plasmids from
> >the isert only plates? What do you mean by garbage plasmid?
> Numbers: on the vector no insert plate, get maybe 5-10 colonies.  On the
> plus insert plate, get maybe 50-100 colonies.  On the insert alone plate,
> maybe 50 colonies.  No, we have not tried isolating plasmid from the
insert only
> plate (though my suspicion is that, if we were to do so, they would be all
> "garbage" plasmid).  Here's how it went.  He did four ligations.  The
first two
> were with the same insert (long PCR'd from genomic DNA) and two different
> vectors (with compatible ends).  The second two were with two different
> (again, long PCR'd from genomic) and the same vector for both (again,
> ends).  For each of the ligations, he did the vector-alone control.  For
> four ligations, the background from the controls was low (see above).  He
> minipreps from all four ligations: for the first ligation, he did 24
> of 24, 7 were vector plus insert (i.e., the desired clone); 3 were vector
> without insert; 1 was a plasmid of slightly smaller size than the vector
> deletion derivative?) and the remainder (i.e., 13) were the "garbage"
> a plasmid of approximately 2.3-2.4 kb, much smaller than the vector
plasmid (in
> this case, vector = 7 kb).  For the other three ligations, he did 18
> for each.  In the case of the second ligation, all of the minipreps were
> "garbage" plasmid.  In the third and fourth ligations, approximately 1/4
and 1/2
> (respectively) of the preps were the vector without insert; the remainder
> the "garbage" plasmid.

What you say is a "garbage" plasmid is a deletion derivative of the original
one. It has been seen that DH5alpha is notorious in maintaining big
plasmids. You may wish to switch to DH10B for maintaining big plasmid
construct. Lowering the growth temperature will also help.
> We have not tested the "garbage" plasmid in any way, but upon further
> examination we see that it has come up frequently over time in the past,
> ligations previously.

I guess these derivatives comes only when you use tthis particular plasmid.

> So we thought it must be a contaminant in one of the
> reagents, solutions, etc. ... but obviously it must be in one of the
> reagents.  So we did control ligation/transformations: mock (no DNA +
> mock -ligase (no DNA, no ligase), no ligase (vector + insert DNA, no
> and insert alone (insert + ligase).  Of these, the insert + ligase gave
lots of
> colonies (see above), but all others had few if any colonies.

This is the most queer phenomenon:insert alone giving rise to clones. Did
you do (insert, no ligase) transformation? This will rule out any
contaminants. You must have done this, in that case this phenomenon is
inexplainable except two possibilities:

# Your amplified fragment contains a plasmid contruct. I guess the
probability of such a thing happening is almost  zero.
# The more probable solution is if you are using a cloned polymerase
(recombinant) then the possibility is very high that your enzyme is
contaminated. It has been seen in case of Taq that it has very high tendency
to stick to DNA. I guess you should give a try by changing using a different
stock of enzyme.

(One could argue
> that there is trace amounts of uncut vector in the no ligase control, and
> that some ligation will be done by the bacteria themselves in the absence
> ligase.)  We also did other controls (see below).
> > It is most likely contamination of the PCR product with a linearised
> > plasmid -- perhaps the template itself -- that you co-isolated in the
> > gel slice.
> >Possible but rare to see.
> Yes, and unlikely in this case since the PCR was done from genomic DNA.
> >Electrophoresis never gives perfect resolution, and
> > virtually any slice you take from a lane will contain at least a trace
> > representation of all the DNA you loaded (or in addition, if you didn't
> > clean out the tank carefully, a representation of DNAs loaded by your
> > colleagues in the recent past!).  That is why two or three rounds of gel
> > purification is often necessary.
> >Again possible but 3/4 rounds of purification is not necessary.
> Yes, I agree-- it's of course possible as he says that the gel box,
> buffers, etc. are the source of the contamination.  To be safe, I have had
> make up everything new.
> >I guess your competent cells are not good and contaminated. Try plating
> >cells alone in a proper Amp plate. See that you get no clones there.
> Yes, we too were suspicious of this as one of the first possibilities to
> consider.  So we did the following additional controls: 1) no
> (i.e., plate directly on Amp); 2) transform with nothing; 3) transform
> water.  All were basically zero.  Since essentially all the controls
> pretty clean, I also had him do a control of ligating and transforming the
> (we resuspend most all of our DNA's in TE following an EtOH precipitation,
> just use from the low melt agarose directly).  This was also zero.  So I'm
> pretty sure at this point that it's either a) some contaminant of a linear
> in one of the solutions somewhere, but not the TE; or b) the inserts can
> inherently produce colonies.
The correct answer in science is the most simplest one, following this rule,
you must rule out all possibilites of contamination before thinking in terms
of my more  "esoteric" explanation.

Best of luck,

Malay Kumar Basu
Centre for Cellular and Molecular Biology
Hyderabad 500007

Fax: (00-91)40-7171195
Phone: (00-91)40-7172241
Smoke me a kipper, I'll be back for breakfast.
curiouser at


More information about the Methods mailing list