"high stringency" washes -NOT?
Jim
Greetings colleagues,
Anyone else find it puzzling that "high stringency" washes feature a reduction in salt, a component used during hybridization to lower the annealing midpoint temperature (Tm)? One might reasonably expect high salt in washes to help melt less stable mismatched duplexes by lowering the Tm of all duplexes. Such a wash would be an undescribed "high salt -stringent wash".
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).
The principle of charge screening in bringing together similarly charged poly-ions seems to me to support the notion that increased salt should make it easier for perfectly matched or mismatched duplexes to form (salt lowering Tm), as well as assisting any temperature in melting either type of duplex. (High stringency = high salt). This is apparently taken into consideration when determining hybridization temperatures, yet disregarded in washing -in fact reversed in practice.
If one wanted to "increase stringency" by making washing conditions say Tm -5 when using Tm -20 in hybridization reactions, one would need to maintain the same salt concentration and increase the wash temperature 15 degrees, or more conveniently, simply INCREASE salt concentration at the same temperature, thereby lowering the calculated Tm 15 degrees. :)
Theory and practice colliding perhaps?
(Perhaps mismatched hybrids are differentially sensitive to the effect of salt on melting to an extent where lowering salt becomes more effective than increasing salt to melt less stable duplexes ?)
E-mail me copies of responses directly please,
Jim
J. Graham PhD
Biology Department
Washington University of St. Louis
Peter Nilsson
Hello Jim.
I can only add a biochemical explanation on the effect of ionic strength ([Na]) on the stability of the double helix: The negative charges on
opposite strands in the backbone, the phosphate-groups, destabilize the helix by repulsing the the strands in the double helix from each other.
This intrinsic instability is balanced by positivly charged ions (Na+) wich shields the negative charges on the backbone. With very high ionic
strength one get not only shielding but also a ionic chemical bound between Na+ and P-.
Also high ionic strength increases the hydrofobic interactions between the bases "inside" the helix, the base stacking, which also stabilize
the interactions between the oppossite strands. Therefore, these two effects acting together gives a stabilization (higher Tm) with increased
ionic strength. Decreasing the [Na] means higher stringency.
Regards, peter.
--
Peter Nilsson, Molecular Biologist
Clinical Chemistry Department,
Umeå University Hospital,
S-901 85 Umeå,
Sweden.
gra...@biodec.wustl.edu
As Dave Haviland kindly points out, the Bolton and Mcarthy equation
given in Sambrook (Mantiatis Manual 2nd ed.) and propagated elsewhere is
dead wrong. They give:
Tm = 81.5 - 16.6 (log[Na+]) + 0.41(G/C) -0.63(form%) -600/L
rather than the correct:
Tm = 81.5 + 16.6 (log[Na+]) + 0.41(G/C) -0.63(form%) -600/L
What is it with this equation that editors and apparently scientists
find so difficult to state correcly (absolute value assumed, %GC verses
decimal)? Is it wrong in the orginal Bolton and McCarthy paper as well?
Anyone?
Apparently, adding salt raises the Tm, therefore washing at reduced salt
ands same temperature would ideed move conditions closer to the melting
temperature of hybrids.
Carry on, :)
Jim
gra...@biodec.wustl.edu
Wrapped =
Greetings colleagues,
Anyone else find it puzzling that "high stringency" washes feature a
reduction in salt, a component used during hybridization to lower the
annealing midpoint temperature (Tm)? One might reasonably expect high
salt in washes to help melt less stable mismatched duplexes by lowering
the Tm of all duplexes. Such a wash would be an undescribed "high salt
-stringent wash".
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).
Gene Huh
In article <MacWeb04M...@clarkcentris650.wustl.edu>,
gra...@bionet.wustl.edu (Jim) wrote:
# Greetings colleagues,
#
# Anyone else find it puzzling that "high stringency" washes feature a
reduction in salt, a component used during hybridization to lower the
annealing midpoint temperature (Tm)? One might reasonably expect high salt
in washes to help melt less stable mismatched duplexes by lowering the Tm
of all duplexes. Such a wash would be an undescribed "high salt -stringent
wash".
#
# 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).
.
.
.
# E-mail me copies of responses directly please,
#
# Jim
# J. Graham PhD
# Biology Department
# Washington University of St. Louis
According to the equations for Tm estimates of DNA-DNA, DNA/RNA and
RNA/RNA quoted by Sambrook and/or by the RedBook, the above equation has a
sign switched. I think it should be "+16.6(log[M+]", since the
Bolton-McCarthy equation as quoted by Sambrook and several other equations
all have the cation component added rather than subtracted. Needless to
say, the apparent paradox is no longer.
Hope this helps
Gene
--
Gene S. Huh, Ph.D.
Department of Molecular and Cell Biology
Life Sciences Addition, Room 221
University of California at Berkeley
Berkeley, California 94720-3200
Email: gs...@socrates.berkeley.edu
Kristian H. Jensen
Peter Nilsson (Peter....@klinkemi.umu.se) wrote:
: Hello Jim.
Hi all you cyber dna netter!
Exactly, there are 3 major contributions to DNA hybridisation;
backbone repulsion; base stacing and hydrogen bonding between bases.
The first two are responsible for the increase in Tm with salt, as
repulsion is shielded and hydrophobic base stacking is enhanced.
Base hydrogen bondig only becomes "important" at low ionic strength,
hence the low salt high stringency Washes.
regards Kristian.
David F. Spencer
> As Dave Haviland kindly points out, the Bolton and Mcarthy equation
> given in Sambrook (Mantiatis Manual 2nd ed.) and propagated elsewhere is
> dead wrong. They give:
>
> Tm = 81.5 - 16.6 (log[Na+]) + 0.41(G/C) -0.63(form%) -600/L
>
> rather than the correct:
>
> Tm = 81.5 + 16.6 (log[Na+]) + 0.41(G/C) -0.63(form%) -600/L
>
> What is it with this equation that editors and apparently scientists
> find so difficult to state correcly (absolute value assumed, %GC verses
> decimal)? Is it wrong in the orginal Bolton and McCarthy paper as well?
> Anyone?
The foulup on the sign is hardly the major problem with this formula from
Sambruck et al. The Bolton and McCarthy citation is wrong - that paper doesn't
even have an equation and doesn't address melting temperature at all. The
seminal paper for all of this discussion is the paper of Marmur and Doty in
JMB vol.5:109-118, 1962 (which, it would appear, like most of the work of that
critical era is now long forgotten by most in this field). The formula that
they derived was, for DNA in SSC, that Tm = 69.3 + 0.41 (%GC), where 69.3 is the
extrapolated Tm for 0% GC and 0.41 is the slope. The paper also determined the
Tm for a few DNAs in 10 mM phosphate and found that going from SSC to the 10 mM
lowered the Tm by almost exactly 20 C. Now you would think that this information
would be adequate to build a reasonable equation (for long DNA and without
formamide) but the Sambruck equation starts with '81.5' (where the devil did
that come from?). It also follows from Marmur and Doty, and which I have used
in various test equations of my own, that the effect of salt would be properly
allowed for by a formula such as '+19.25(log[Na+] - log (0.165))' where I have
substituted 19.25 for Sambruck et al.s 16.6 (the 19.25 can be calculated from
M + D's data) and 0.165 is, of course the Na+ concentration in SSC. That
would give Tm = 69.3 + 0.41(%GC) + 19.25 (log[Na+] - log(0.165)); this will
possibly fail at very high [Na+]. The only paper I could quickly find for
effect of formamide on Tm is McConaughy, Laird and McCarthy, Biochemistry
8:3289, 1969, where they give the number (about) 0.72 C per 1% formamide. So
where did the 0.63 come from in Sambruck et al.?
My guess is that the famous Sambruck-Maniatis equation was pretty much thrown
together arbitrarily and like all too many procedures in the molecular biology
lab of today, its origins have never been questioned. It also indicates that
these types of formulas are not all that useful for this type of work anyway,
because, whether for PCR or hybridization, ultimately temperatures are
determined empirically.
-Dave Spencer
David F. Spencer
As a followup to my own followup I played around more with the arithmetic and
derived the equation [presumably applicable to DNA:DNA liquid hybrids]:
Tm = 84.3 + 19.2(log[Na+]) + 0.41(%GC) - 0.72(%form.) - (600 or 650 / length)
This is consistent with Marmur and Doty (and their's is the only hard data that
I'm aware of), plus the different factor for formamide. The length correction
seems to be a bit of a crap shoot; I've not seen a derivation of this number.
The formula under discussion in Sambruck et al.
Tm = 81.5 + 16.6(log[Na+] + 0.41(%GC) - 0.63(%formamide) - (600/l)
presumably was taken from Meinkoth and Wahl, Analytical Biochem. 138 : 267-284
(1984) where the equation is given as :
Tm = 81.5 + 16.6 log M + 0.41 (%GC) - 500/n - 0.61(%formamide)
where they claim the formula is only useable above about 50 nucleotides. This
formula was apparently derived by Meinkoth and Wahl but the attribution is not
clearly stated in the paper. They also give the so-called Wallace formula
Td = 4(G+C) + 2(A+T)
supposedly used for 14-20 mers, but also, it would appear, applicable only for
solid substrate hybridization and not for PCR, etc. Just what you're supposed
to use for primers between 20 and 50, and in liquid, is unclear.
These approximations are obviously flawed because they assume composition alone,
and not context, determines the Tm of short or medium length oligos. The proper
method for determining Tm in oligos was given in Breslauer et al. PNAS 83 :
3746-3750 (1986) and uses thermodynamics and something akin to real math.
Rychlik and Rhoads in NAR 17:8543(1989) describe a program that calculates Tm
and Td for oligos using Breslauer et al. data. This program (called OLIGO) may
be available in an early version at the usual molbio archives (Indiana, etc.)
but I believe the program went commercial several years ago. [This program seems
to calculate for 1 M salt and does not appear to have corrections for salt or
formamide concentrations].
Dave
--
David F. Spencer, PhD
Dept. of Biochemistry
Dalhousie University
Halifax, Nova Scotia, Canada
dspe...@is.dal.ca
dspe...@rsu.biochem.dal.ca
David F. Spencer
In article <dspencer-ya023580...@News.Dal.Ca>,
dspe...@is.dal.ca (David F. Spencer) wrote:
This thread has deteriorated to me arguing with myself but this will be my last
contribution and then I shall let it die (unless something stupendously exciting
develops which is unlikely).
The formula that started it all, from Sambrook et al., incorrectly attributed to
Bolton and McCarthy, for DNA in solution:
Tm = 81.5 + 16.6(log10[Na+]) + 0.41(% G+C) - 0.63(%formamide) - (600/l)
^ The typo in Sambrook has this as '-'
In my calculations attempting to derive "the" equation I failed to calculate the
[Na+] correctly for SSC, which is really 0.195 (0.15 for NaCl + 3 X 0.015 for
Na citrate). The log10 of 0.195 is -0.71. I had used the correct value for
0.010 M sodium phosphate, pH 7.0, which has a [Na+] of 0.015.
The first number in the Tm equation (almost universally given as 81.5)
should be
the sum of 69.3 (the extrapolated Tm for 0% AT in SSC, from Marmur and
Doty) and
the number x(-log10[Na+] in SSC) where x is usually 16.6 but I've also seen
18.5. In Marmur and Doty the best approximation seems to be (from their Fig.4)
that the Tm drops by nearly exactly 20 C when going from SSC to 0.010 M
phosphate. The factor x above would come from dividing 20 by the difference of
1.82 (the log10 of 0.015) and .71 (the log10 of 0.195). This gives 18.0 rather
than 16.6 and the first constant becomes 82.0, which is probably close enough
to 81.5 but requires the critical coefficient to be 18.0, not 16.6 (and I never
found the origin of the 16.6 before abandoning the hunt).
The 0.41(%GC) is undisputed (thank heavens).
The formamide correction is slightly shakey; the first number I found in print
was 0.72 (McConaughy, Laird and McCarthy, Biochemistry 8:3289(1969) but later
Casey and Davidson (NAR 4:1539 (1977) gave 0.75 C for poly dA:dT and 0.5 C for
poly dG:dC; the 0.63 above is clearly the average and the two DNAs graphed in
Casey and Davidson give coefficients of 0.65 (E.coli) and 0.62 (for human). It
is clear that the 0.63 +- value technically should have some composition
correction factor; I leave that exercise up to the reader, if any are left.
The last value in the equation, which is either 600/length (or sometimes 500/l
or 650/l, etc.) turns out to be a bit of a can of worms. The constant (600, 500)
has apparently been taken from Britten, Graham and Neufeld , Methods Enzymol.
volume 29:363(1974) on page 367. Although they cite some other papers for length
corrections they conclude a best compromise of 650, technically for 0.18 M
[Na+]. Earlier papers they cite gave about 500 (for 20 mM Na+) or 750 for SSC.
They acknowledge that the 650 value "is somewhat uncertain" and it seems that
the length correction needs a [Na+] correction.
None of this allows for the 2-2.5 mM Mg++ in the PCR mix, or the fact that
modern PCR buffers often/frequently are NH4+ rather than K+ based.
When all is said and done, if you're using TAQ try your first PCR annealing at
55 C, or maybe 50 C if its kind of AT rich, or maybe 45 if your using inosine or
expect some mismatches. It's what everyone does in the end even after mucking
around with silly equations of questionable origins and accuracy.
Cheers,