Hyg, Neo, and $$

David Huen s.d.huen at bham.ac.uk
Fri Jun 17 03:06:15 EST 1994

In article <1994Jun17.000602.4613 at emba.uvm.edu>, brianf at med.uvm.edu (Brian
Foley) wrote:
> down-n-out (eatkinso at gpu.srv.ualberta.ca) wrote:
<stuff deleted>

> 	Yes, the bottle of hygromycin I got from Calbiochem says it is 
> suspected of being carcinogenic and a teratogen.  This would cause you
> to beleive that it is also mutagenic.  The only thing I have done 
> about this is to transfect a control lot of cells with the same plasmid
> carrying no insert DNA.  These cells are maintained on hygromycin
> right beside the cells transfected with the plasmid carrying my gene
> of interest.  I have not yet seen any mutations in the system I screen for.

Since hyg targets the nucleotide synthesis pathway, it could potentially be

> : I guess the reason that some people prefer to maintain the selective 
> : pressure is to ensure that their inserted DNA is not kicked out of the 
> : cell's genome, but how common is this really?  
> 	I think it has more to do with reversible amplification than
> excision from the genome.  Gene amplification seems to be quite 
> reversible over relatively short periods of time.
> 	The genetic markers most often cited for gene amplification
> are HGPRT and GPT.  I don't have the references handy right now, but
> write to me if you have trouble finding them.
> : ES cells that are targeted 
> : with a neo insertion are removed from drug ASAP and the neo cassette 
> : remains stable.  When you think about the number of cell divisions 
> : required to generate a mouse, that's a lot of generations.  This may be a 
> : bad example, but I'm wondering if the frequency of foreign DNA 
> : "rejection" has anything to do with the type of gene product.  It seems 
> : reasonable to me that in the absence of drug selection, the introduced 
> : DNA would be likely to be eliminated or selected against if it encoded 
> : something that the cells were not entirely happy with.  
ES cells, at least, are fully totipotent and "normal". I have serious
doubts as to the genomic stability of many lab lines though given that they
are already abnormal with regard to "immortality" etc. Certainly, the
spontaneous transformation rate of rodent fibroblast lines and the very
abnormal karyotypes of the typical ATCC line gives me the shivers as to
their genomic stability. Nevertheless, the point made about periodic
selection rather than continuous selection seems to make sense, esp.

> : If, however, the 
> : foreign DNA is relatively benign, would this be more or less stable?
> : From our own experience, we have been given a cell line that was 
> : transfected with an expression vector encoding a cell surface marker.  
> : The original selection was with G418, but we have never grown the cells 
> : in the presence of drug, and through countless generations now, the 
> : marker remains.  Obviously drug selection is not always necessary.
> 	Was it a clonal line?  Or a population of transfectants?

<stuff deleted>

Two points: First, there was a paper in Oncogene that reported that the neo
gene used in G418 selection has effects on cell phenotype in its own right.
Presumably, it phosphorylates something else other that G418 only. The
other is that with non-adherent cell lines, the usual selection procedure
doesn't really guarantee you clonal lines, esp. since some of the lines
seem to crash at low cell densities and limiting dilution cloning is often
unsuccessful in these circumstances (cross-feeding effects?).

 > 	Yes, I would think that the whole field of molecular biology would
> be better off if as much attention was paid to transfections as is paid 
> to ligations and other DNA manipulations.  If I go to the library I can 
> find all kinds of data about ligation efficiency and theory, but there
> is very little on what happens to a plasmid after it is put into a
> mammalian cell.  I'd like to see Southern blots or other data to indicate
> how many copies per cell typically integrate.  I'd like to know why
> the methods of transfection that give the best results in transient
> transfections do not work well for stable transfections.  This does not
> make sense to me...
I suppose that the DNA entry event and the integration events are
independently influenced. Presumably DEAE-dex gets stuff in fine but
inhibits whatever recombination steps that are required to get integration
(poly ADP-ribosylase or whatever is supposed be important in this I seem to
remember, some stuff in PNAS about maintenance of retroviral integrations
some time back). I seem to recall also (and I don't know the source for
this) that of the stable transfection techniques, CaPO4 gives the highest
copy no., and lipofection and electroporation give much lower levels.

Incidentally, some years ago, for the sheer heck of it a colleague and I
tested a way of getting better integration after transfection. Basically,
you transfect with whatever technique you see fit, then give the 100-300
rads of gammas the next day to tickle DNA repair. As expected, you get 3-4x
more transfectants come thru'. We can't use it routinely in our case as
it's a bit difficult to explain to the referees of a cancer paper that the
effects you see aren't from having kicked the metaphorical butt of the
lines with radiation while making them. But it should be OK if all you want
are producing lines.

Just my 0.02 worth.

More information about the Methods mailing list