TKO miniflourometer (long)

[James Graham]

Alexander,

In my opinion, Philips comments, while very valuable to those who have 
useable TK100's, fall a little short in terms of their ability to serve as a 
"FAQ" on the TK100. Firstly, he describes relatively minor problems (ie. 
2X discrepency with UV absorbance quantitation) and an "overflourescence" 
problem, while what I have consitently seen is the tendency to intermitently
read a particular sample as a "blank", which then showed excellent yield 
following agarose gel electrophoresis, along with the good utility 
described. Philip describes problems involving rather finer points in 
quantitation, rather than the outright failures described in the 
newsgroup archive. (I am personally familiar with 5 units, three that  
are unreliable to the point of being unusuable, and two which are in 
routine reportedly successful use across two departments I have been in.)

The presence of dust, while certainly a confounding aspect (particularily
when it is suggested that it is so difficult to exclude) may be related
to some problems, but seems to me an unlikely "explanation", particularily
when far more sensitive experiments are being carried out sucessfully in
the same laboratories by the same hands. The assay procedure itself, so
often blamed, is really quite simple acutally, in comparison to many routine
laboratory procedures. I fail to find user error a likely explanation
for the widespread failure of the intstrument reported in the newsgroup
archives, barring a tenuous potential for proper beam alignment in
terms of cuvette positioning or other unacceptable design problem .

While signficant sensitivity to traces of contaminants may be at the
heart of the problem(s), the maufuacturer and others have reported that
the Hoest quantitation itself is relative insensitive to a range of
common laboratory reagents likley encountered (eg. SDS, phenol,
proteins, etc.).

Finally, Phillipe's comments neglect to mention that the problem with
the TK100 was acknowledged by Hoeffer to an extent at one time, at least
in terms of a "reconditioning" of some kind offered to users at
significantadditional relative expense. Any FAQ would have to include some
reference to whatever this still unexplained recall involved.

(Failing to observe the reported problem, is indeed negative
evidence -a type of data usually found of limited utility in
defining reported phenomena.)

Regards,
Jim
J. Graham PhD
Biology Department
Washington University of St. Louis

Philip wrote:

 I certainly can't shed much light on the problems of individual machines that
I haven't personally used, but I have used 3 different TKO miniflourometers
bought over a long period with I believe reasonable results i.e. the amounts
of DNA we estimate by the TKO seem to roughly agree with what we see on a gel
following EB staining.  So why do some people have problems?  Certainly one
problem is dust.  If you do not dispense the reagent from a closed container
i.e. Repipette it is very difficult to get good results. Pipette tips can 
be a problem.  They can be sterile and still have dust on them.  The 
problem of dust interferring with any fluorescenct technique is well known.
     A second problem is that there can be chemical interference.  The most
dramatic example I know of I discovered myself recently.  I wanted to follow
the binding of DNA in an alkali miniprep to glass beads to see if binding was
quantitative.  I got suspiciously high readings for the amount of "DNA"  left
in the supernatant. As a control I mixed NaOH and SDS with the standard GET
(or TEG as some call it) buffer used in alkali minipreps and then  added the
usual 5 M KOAc reagent to obtain the usual white precipitate which I spun 
out. I then assayed the amount of "DNA" in this supernatant.  I emphasize 
that this was a completely mock miniprep.  No cells were present.  To my 
surprise I got very high fluorescence.  So there it seems there is some 
chemical reaction between GET, NaOH, SDS, and KOAc that can masquerade as 
"DNA" byfluorescence. I only did this once, so I'd appreciate it if 
someone anyone wants to repeat this little experiment and see if they 
confirm my result.  However, I tentatively conclude that a very common 
procedure can lead to false results. I hasten to add that after washing 
and elution of the DNA from the glass beads, I got a reasonable value for 
the amount of DNA.
      A third problem is that many common miniprep methods yield DNA that is
probably contaminated with RNA.  When you consider that there is about 10x
more RNA than DNA in a cell, a procedure that gets rid of 90% of the RNA 
still leaves equal amounts of DNA and RNA.  Since plasmid DNA is only a 
fraction of the total DNA there is plenty of room for error.  When we 
first set up our automated sequencing facility we found that consistently 
the TKO gave lower readings than A260 for DNA content.  Usually the 
difference was less than 2-fold, but in some preps the differences were 
as much as 5 or 10 fold.  My impression (not a careful study!) is that as 
folks got better at making miniprep DNA the sequencing results improved 
and the discrepency between the fluorometer and A260 tended to get less.  
Many workers think that if they use RNase or CsCl the resulting DNA is 
free of RNA.  Probably not so.  EB is a very insensitive way of looking 
at RNA on a gel, and since pancreatic RNase only cleaves at U,C, 
polypurine runs will lead to rather large fragments of RNA contaminating 
your DNA.  (I assume this is the material that is so nicely removed by 
the various ion exchange and glass methods.)  Note that RNA oligos
will not pellet in CsCl, so that large pellet of RNA you may see in the 
bottom of your tube does not mean there is no RNA contamination.  I know 
of at least one paper in which the authors examined standard miniprep DNA 
by HPLC and found a large peak of A260 absorbing material that was not 
plasmid DNA.  In some organisms like yeast, you can RNase and phenol 
extract and get rediculously large amounts of RNA in preps of chromosomal 
DNA which can only be removed by a second RNAse treatment after the bulk 
of protein and RNA are removed.  So if  you demand that the TKO give you 
results consistent with UV absorbance measurements, you are are probably 
not playing fair.
      Another problem is that one should not expect these fluorescence
measurements to be +/- 10%-at least not without a lot of work.  I wonder if
people who estimate DNA concentration by running gels think that 
technique is +/- 10%.  I hope not.  I always rezero the machine just 
before taking a reading, and I'd say that results are usually +/- 20% or 
better.  Do people really need better for most molecular biology experiments?
      I do not believe the electronics of the machines I have seen are
unstable.  If they were, do you think you would be able to get consistent
results with the calf thymus DNA standards?  Just look at the display 
with no sample in the cuvette.  On the machines I use, it is quite stable for
reasonable periods.  Of course, if you let the DNA dye complex stay in the
cuvette being bombarded by UV for many minutes, you will see the display
drift, but this is probably fluorescent bleaching not instability of the
machine.
      In summary, I find the TKO to be a great time saver in the lab.  It
has to be used with care and understanding like any highly sensitive
technique.  There are traps for the unwary which I have discussed, but maybe
some of you ought to get your machines off the shelf one last time, and 
see if after considering some of the points I've raised it works for you.

Usual disclaimers.

Philip Carl
Res. Assoc. Prof. of Pharmacology
UNC Chapel Hill