white colonies have no inserts, why?

Hiranya Roychowdhury hroychow at NMSU.EDU
Thu Jul 16 11:39:42 EST 1998

At 04:08 AM 7/16/98 -0700, beisz at SUN5.IBP.AC.CN wrote:
>I made a library of an unidentified DNA, using pGEM-3Zf(+) as vector, EcoRI
>and PstI as
>cloning site, and E. coli JM109 as host. Two thirds of the transformants are
>white, but only
>5% of these white colonies contain inserts. Who can tell me why?
>You may post here, or mail to beisz at ibp.ac.cn.
>The detailed procedure:
>1) cut the vector with EcoRI and PstI, recover the large fragments fron gel.
>2) cut the exogenous DNA with EcoRI and PstI, phenol-chloroform extraction,
>3) sticky-end ligation, 15-25C, overnight.
>4) transformation.
>5) isolation plasmids with alkaline method, enzyme cut, electrophoresis.

Here are a couple of possible explanations:

A. The vector (or a percentage of it) did not get digested with one of the
enzymes. This is not unusual. Hence you get intramolecular ligation.

Solution: While you recover the vector fragment from the gel, make sure that
you see the RI-PI "stuffer" running on the gel (approximately where RNA
would run if you had any). The best way to make sure you see that, run a
minimum of 1% gel (preferably 1.2%) and, after the fastest tracking dye has
moved half-way, quickly take a look at the gel. If both RE's have cut, you
should be able to locate the ~40bp RI-PI fragment.

B. Both enzymes did cut, but concatenated vector molecules are transforming
cells. These would yield white colonies.

To save a lot of aggravation, what you need to do is to simply
dephoshorylate the digested vector (yes, even if it is digested with two
different enzymes) before you purify it. This would ensure that no matter
what, the only ligations occuring are between the intended insert and the
vector, irrespective of the possibility that only one of the enzymes cut
part of your sample.

Dr. Hiranya Sankar Roychowdhury
Plant Genetic Engineering Lab.
New Mexico State University
Las Cruces, NM 88003
Ph. (505) 646-5785
hroychow at nmsu.edu

More information about the Methods mailing list