Maximum plasmid insert size?

[Phillip San Miquel]

  Last week I subcloned an 18 kb BamHI fragment into 
pBluescript. It was a shotgun ligation of a sort --
I cut an 80 kb sorghum BAC clone with BamHI and ran
the digest on a 0.6% agarose TAE gel.  Four bands
above 14 kb ran together.  I cut these out and glass
milk purified them.  It could have been 0.25 ug of
DNA I isolated.  I used  about half of this in a two step
ligation with phosphatased BamHI cut pBluescript -- maybe
50 ng.  The vector had been phenol chloroform extracted
after the phosphatase treatment but not gel purified.  
Actually I got the vector from a lab mate.  I ran it
on a gel and could see it had an enormous amount of 
RNA in it and a little undigested plasmid perhaps.
  After ligation, I phenol chloroform extracted the ligation
and precipitated it.  I resuspended it in 10 ul and
used 5 ul for the transformation.  My final transformation
volume was about 1200ul after adding media and allowing
the cells to recover at 37 oC.  On a plate on which I had spread
1 ul of this transformation, 19 white colonies and
37 blue colonies grew.  I picked 12 of the white colonies.
  There were 4 bands 14 kb or larger in the BamHI digest:
18kb, 16kb and two 14kb.  I got 5 subclones with 
one of the 14 kb bands (1 in orientation, 4 in the other),
4 subclones of the 16 kb band (2 in one orientation,
2 in the other) and 3 subclones of the 18 kb fragment
(2 in one orientation, one in the other).  
  For this reason I do not feel that transformation is
a limitation for this size of plasmid.  I think the
stability of the insert DNA is critical.  I may have
only gotten one of the 14 kb bands -- if this is
the case I would attribute that to instability of that
DNA in a very high copy number plasmid.  One of the 14
kb subclones I got does appear to be undergoing rearrangement
(a 7 kb deletion).  But you don't have much control over
that -- either it is stable or it isn't...  Go to a lower
copy number vector if it is unstable at pUC19 level copy
number.
  Competant cells are critical but it is easy to make
5x10e7 to 1x10e8 cfu/ug competant cells of common E. coli
strains using Inoue, Nojima, and Okayama's technique found
in their paper: High efficiency transformation of Escherichia
coli with plasmids, Gene 96(1990)23-28.  I've kept cells
I've made using this technique months under liquid N2 and
they still work great.
  Ligation conditions, I believe, are critical when large
fragments are to be subcloned.  Standard protocols for
determining the optimum vector to insert ratios never
made any sense to me until I read Damak and Bullock's   
paper: A Simple Two-Step Method for Efficient Blunt-End
Ligation of DNA Fragments, Biotechniques 15, No. 3(1993)
448-452.  It addresses blunt end fragment cloning but 
I think the method is applicable to large fragment
cloning as well.  The idea, in a nutshell, is to first
ligate the vector to the fragment under conditions which
favor bimolecular ligations (high concentration of DNA)
then, in a second step, increase the volume to favor
intramolecular ligation to allow the vector-insert to
circularize. I add PEG to both steps to increase crowding.
  Again I don't think transformation is a limiting factor.
I can easily transform an 80 kb BAC into competant cells.
Another guy in the lab subcloned a 50 kb fragment into
a BAC vector using competant cells.
  You can use electroporation if you want, but I don't
think it is necessary.

Phillip