summary of one-step gene disruption

Dave Eide deide at UA.D.UMN.EDU
Wed Apr 20 18:54:35 EST 1994

Thanks to all who responded.  The discussion was very helpful  I have assembled 
a summary of responses for your archives.

Dave Eide

And I quote...

>We would like to do some one-step gene disruptions using flanking regions 
>smaller than the recommended minimum.  Rothstein (Meth. Enzymol. vol. 194) 
>suggests using flanking regions of 100 bp or more on each side of the 
>marker and the method we want to try gives us only about 50 bp on each side.  
>Does anyone have information regarding homologous recombination of fragments of
>this size?  Any input will be greatly appreciated.
>Dave Eide 

	Actually it works with even shorter homologous regions. Check a 
protocol that uses PCR in NAR: 21(14), 3329-3330  (1993).  This protocol have
been used in our lab with good results.  (used just once).

Good luck,

Jaime E. Arenas
Division of Biology


It is possible to get targeting with sequences that short. See Baudin et al.
Nucl. Acid Res. 21: 3329 1993. Also, Fred Sherman had several papers a few
years ago describing transformation with oligos that restored function to 
mutant cytochrome c genes. The efficiency of targeting to the intended site is 
lower than you get with larger homologous regions.

Preston Garrison
Biochem. Dept.
San Antonio, Tx 78284-7760
garrisonp at


we've tried replacements using PCR generated fragments which had about 25 nt
of our gene flanking a PCRd HIS3 gene, and it really didn't work.  Clearly,
there's something to Rothstein's limit, but some have gotten what they needed
with much shorter constructs, either because they did everything else a
certain way, or they were lucky.  I'd say--go for it, and if it doesn't work
reasonably quickly (you should know very soon), drop back and try longer
flanking regions.
Chuck Cole 


I have gotten homologous recombination to work with a 153 bp PCR
product co-transformed with a linearized plasmid in an application
called One-Step Site-Directed Mutagenesis.

Here is a schematic diagram of DNA samples co-transformed:

       ====X============== 153 bp PCR product (mutations at X)
       24nt       <-40nt->
 //============== ============\\
 ||                           ||
 ||linearized 15 kbYEp plasmid||
 ||                           ||

To regenerate the plasmid, one needed homologous recombination on
either side of the site at which the plasmid was linearized (110 bp
and 40 bp).  However, to get site-directed mutants, recombination on
the left side had to be in the distal 24 nt.  

It turns out that linearized plasmids transform about as well as
closed circulars but addition of the PCR product boosted
transformation frequency by 3-fold.  The frequency of site-directed
mutants among the transformants was consistent with recombination at
an unbiased rate all the way to the ends of the PCR product.

Site-directed mutagenesis and ordinary gene disruptions should,
therefore, be possible with PCR-generated samples that have only tens
of nt of identity at each end.

The reference for One-Step Site-Directed Mutagenesis is
Brenner, Bevan & Fuller, Current Biology 3, pp. 498-506, 1993.

Best of luck,
Charlie Brenner

Charles Brenner, Ph.D.		Fellow of the Leukemia Society of America
Rosenstiel Center, rm 610	phone (617) 736-4944
Brandeis University		fax   (617) 736-2405
Waltham, MA 02254-9110		brenner at

More information about the Yeast mailing list