IUBio GIL .. BIOSCI/Bionet News .. Biosequences .. Software .. FTP

Single molecule optically viewing?

Doug McDonald mcdonald at aries.scs.uiuc.edu
Sat May 16 18:35:25 EST 1998

Eyal Peleg wrote:
> Hello
> Can anyone tell me how to view a single fluorescent molecule by an OPTICAL
> microscope.

> What is the technical information of his microscope (objective, light
> source)

One normally uses a scanning confocal microscope for this. Ours
is home-made based on a standard Zeiss Axiovert 100 TV. 

The objective of ours is a Zeiss Fluar 100X NA 1.3 .... but
just about any NA 1.25 to 1.4 objective will work OK. 

Our light source is a 13 milliwatt Nd:YVO4 green 532 nm laser (don't
buy a lower power one, the more powerful in the range 0-50mw,
the more stable). It is attenuated to about 1 mw with a pinhole and
a set of BG-39 and KG-5 glass filters that remove oodles of stray light
at 800 and 1.06 microns.

It is reflected at 45 degrees off a CaF2 flat and this gives us
about 3 microwatts of green light. This is sent through the
tube lens and the objective to the sample. (You will get lots
of autofluorescence from the lenses at first,
due to fluorescent dust on them, but this will eventually
get photobleached away after a few days of hard use.) 

The fluorescent light comes back through the objective and beamsplitter
(97 % goers through) and then through **two** Kaiser Scientific
Raman-notch filters (10-6 transmission rating at 532 nm) tilted
to get the best possible rejection of the reflected green. 
You really need the 10-14 rejection we get. 
Then a small lens focuses the fluorescent light on an EG&G Canada
SPCM silicon avalanche photodiode photon counting module. (Quantum
efficiency about 75%). The output of that goes to the counters in the
controlling computer.

Our microscope scans the sample using an "inverted beetle"
scan head as used in STMs (i.e. the sample sits in a holder
which rests on three piezotubes that scan in synchronism.)

>Are there any instructions for the specimen preparations?
> (fluorescent dye etc.).

Well, that depends on the experiment you are doing! We use
cover slips as sample holders. These are then cleaned
and coated with covalently bound films of various silanes
(e.g. aminopropyldimethylmethoxysilane is good). The dyes 
are then covalently bound to the amino group by using
appropriate dyes (Chap. 1 of the Mol. Probes catalog).
We've tried many different dyes. The "bread and butter" one
is 5 (and 6) carboxytetramethylrhodamine N-hydroxysuccinimidyl
ester (C-1171)(TAMRA). The best one is Bodipy R6G SE (D-6180) 
but it is expensive as all getout. We've also used Texas Red,
various other (redder) Bodipys and ADA (A-1139). This
microscope obviously does not work with fluorescein and it's
spectral look-alikes. 

We get a background of about 225 counts per second of
which 130 is the SPCM background and the rest is autofluorescence
of the optics and (a **terrible problem**) the immersion oil.
A single TAMRA or Bodipy R6G molecule produces 30,000 to 150,000 counts
per second. The product of amine + CN + ADA gives about 8000
counts per second. 

Once the apparatus is working, this is a very very easy experiment.
We have also looked at more traditional biological samples
with this microscope and it is far far more sensitive than the
traditional scanning confocal microscope. 

Since the "reply-to" button added bionet.molbio.proteins.fluorescent
to the newsgroups list, I'll comment on those beasties. We have
plenty of sensitivity to see such things as single GFP molecules except
of course that our microscope will not excite plain GFP. We chose
the 532 nm laser to reduce problems of autofluorescence. This
was a wise choice for us. The problem of seeing single GFPs in cells
will be autofluorescence of the samples. Confocal microscopy
is a great help, of course, but, the redder the better. We've
considered trying GFP variants, but have been too busy. 

Doug McDonald

More information about the Fluorpro mailing list

Send comments to us at archive@iubioarchive.bio.net