"high stringency" washes -NOT?

David L. Haviland, Ph.D. dhavilan at IMM2.IMM.UTH.TMC.EDU
Wed May 7 09:40:34 EST 1997

At 12:09 5/4/97, Jim wrote:
>Apparently, "(mismatched hybrids) are more stable at high salt
concentration than low". This is contrary to the calculated effect of
increased salt concentration when using Bolton and McCarthy's: 
>Tm = 81.5 - 16.6 (log[Na+]) + 0.41(G/C) -0.63(form%) -600/L
>determination of melting temperature (which I described above).

Jim (Dr. Graham):

Me thinks I have an answer.  Where did you get the formula?   Knowing you
and desire to be precise, I don't think you typed it wrong.   However, it
may have been printed wrong...  

A Log10 of a number less than 1, produces a negative number.  So from some
rusty rules of algebra, if 16.6 times a negative number is negative, its
subtraction would be positive and would add the number to the 81.5.  With
the formula above, I would agree with your conclusion that the more salt
drops the Tm, so at a certain temp, more salt would mean less hybridization
and cross species hybridizations would be hard to do.  This isn't how it
works in practice.

I think where you got the formula is wrong.  I know of this one from Davis
et al.  The little black spiral bound mehods book that has a few recipes
that are outright wrong (PBS & TBE).  However, this formula appears correct
as I've seen it elsewhere and I've used it religously for RNAse protection
assays.  I believe it is around pp 75-76.

I have it as:

Tm = 16.6Log[M] + 0.41(%G/C) + 81.5 - Pm - B/L - 0.65(%F)


M is molar concentration of Na+ (1XSSC is .165M Na+).  %G/C is the percent
G and/or C bases in the probe.  Pm is the % mismatched bases.  %F is the
percent formamide.  B is 675 for probes up to 100 bases and L is the length
of the probe in bases.  Notice the 16.6Log will end up being a negative
number by itself for Na+ concentrations less than 1M.  For long probes,
over 700 bases, in practice the formula reduces to:

Tm = 16.6Log[M] +  81.5 

For large probes the %G/C becomes a ever smaller factor.  For Southerns (at
least for me) formamide isn't used.  The % mismatch isn't usually known in
cross species (or homologue) hybridizations.  Finally the B/L is moot with
large probes over 675 as it approaches 1 as L increases.

Now,  1XSSC is (.165M):   Tm = 16.6Log[.165] + 81.5
				 = 68.5 'C

With  5XSSC (.825M):	  Tm = 16.6Log[.825] + 81.5
				= 80.1 'C

The way this one is written, the rule (and practice) hold. 

Hope this helps,

More information about the Methods mailing list