Hardware Questions (TIRF, Cameras)

[Stephen Bunnell]

Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

Dear microscopists,

    Here's the situation:

    I need to reduce my exposure times.

    I am visualizing dynamically moving complexes (~150nm/s) that are
composed of low abundance signaling molecules, and are close to the
coverslip-water interface.

    I need to visualize these structures, in multiple colors, every 3-5
seconds- preferably faster. Presently, we use a Yokogawa spinning disc to
detect CFP, YFP, and mRFP variants. We are getting by, but would like to be
able to do the work at more physiological chimera expression levels. Typical
exposures currently range from 500 ms (good) to 3000 ms (bad) per channel.

    I am achieving adequate resolution with a 40X NA1.3 oil immersion lens
(Zeiss). I am using the Hamamatsu ORCA-ER CCD. The camera has 6µm pixels,
which we use unbinned. I do not think I can tolerate any lower resolution.

    I have two options, and a question associated with each:

    (1) Move to a TIRF system. What are your opinions about how much gain in
sensitiviy this may provide? Has anyone worked with the Zeiss TIRF module
for the Axiovert 200M?

    (2) Change cameras. I have tried a few back thinned EM-CCDs, but did not
find that they offered much benefit once the pixel size was corrected for.
Is there a better option, that offers high resolution, high sensitivity, and
low background noise? Obviously, there will be compromises. What are your
opinions of intensified CCD cameras?

    Best regards,

    -Steve Bunnell




****************************************************************************
Stephen C. Bunnell, Ph.D.
Assistant Professor
Tufts University Medical School
Department of Pathology
Jaharis Bldg., Room 512
150 Harrison Ave.
Boston, MA 02111

Phone: [phone]
Fax:   [phone]
Email: [hidden email]

[John Oreopoulos]

Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal Actually, I would recommend the combination of both if possible (both TIRF and an EMCCD camera). As long as your complexes are close to the coverslip surface (100 nm or less), you will see them easily with a huge gain in signal to noise ratio. With an objective based TIRF, you'll have to move to a higher NA objective (~1.45), and so your resolution and light collection will go up a bit as well. 

Of course, the only downside is that the pixels on the EMCCD will likely be bigger - a problem I've discussed in another thread on the confocal listserver when it comes to trying to match the Nyquist resolution afforded by the TIRF objective. Only way to know for sure is to try it in a real situation, however.

John Oreopoulos, BSc,

PhD Candidate

University of Toronto

Institute For Biomaterials and Biomedical Engineering

Centre For Studies in Molecular Imaging

Tel: W:416-946-5022

On 31-Aug-07, at 12:38 PM, Stephen Bunnell wrote:

Search the CONFOCAL archive at

http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

Dear microscopists,

    Here's the situation:

    I need to reduce my exposure times.

    I am visualizing dynamically moving complexes (~150nm/s) that are

composed of low abundance signaling molecules, and are close to the

coverslip-water interface.

    I need to visualize these structures, in multiple colors, every 3-5

seconds- preferably faster. Presently, we use a Yokogawa spinning disc to

detect CFP, YFP, and mRFP variants. We are getting by, but would like to be

able to do the work at more physiological chimera expression levels. Typical

exposures currently range from 500 ms (good) to 3000 ms (bad) per channel.

    I am achieving adequate resolution with a 40X NA1.3 oil immersion lens

(Zeiss). I am using the Hamamatsu ORCA-ER CCD. The camera has 6µm pixels,

which we use unbinned. I do not think I can tolerate any lower resolution.

    I have two options, and a question associated with each:

    (1) Move to a TIRF system. What are your opinions about how much gain in

sensitiviy this may provide? Has anyone worked with the Zeiss TIRF module

for the Axiovert 200M?

    (2) Change cameras. I have tried a few back thinned EM-CCDs, but did not

find that they offered much benefit once the pixel size was corrected for.

Is there a better option, that offers high resolution, high sensitivity, and

low background noise? Obviously, there will be compromises. What are your

opinions of intensified CCD cameras?

    Best regards,

    -Steve Bunnell

****************************************************************************

Stephen C. Bunnell, Ph.D.

Assistant Professor

Tufts University Medical School

Department of Pathology

Jaharis Bldg., Room 512

150 Harrison Ave.

Boston, MA 02111

Phone: [phone]

Fax:   [phone]

Email: [hidden email]

[vb-2]

Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

Hi Steve,

I think you have the only one option  - 2x2 binning with Orca-AG (ER, Orca II).
I haven't compared -AG and -ER side-by-side but AG might be slightly better.
OrcaII will be cleraly better especially at 500 msec exposures or longer.
 
mRFP1 is a good choice in relation to SNR, having mCherry would cause less
aggregation. CFP, GFP channels are very noisy. YFP is OK.

EM-CCDs with the gigantic pixels and 40x lens will cause huge loss of signal.
Fom my experience, as 60x NA 1.4 is the brightest lens, and Orca ER (AG or II)
at 2x2 binning is the best compromise. Nikon 100x NA 1.45 TIRF lens would give
a signal of only slightly lower intensity than 60x NA 1.4 likely due to lower
undersampling.    

TIRF may not work as you would excite molecules which are up to 200 nm away
from the glass. Moreover TIRF people often use a thicker cover glass to reduce
optical artefacts.

Good luck,

Vitaly
NCI-Frederick
301-846-6575

Quoting Stephen Bunnell <[hidden email]>:

> Search the CONFOCAL archive at
> http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
>
> Dear microscopists,
>
>     Here's the situation:
>
>     I need to reduce my exposure times.
>
>     I am visualizing dynamically moving complexes (~150nm/s) that are
> composed of low abundance signaling molecules, and are close to the
> coverslip-water interface.
>
>     I need to visualize these structures, in multiple colors, every 3-5
> seconds- preferably faster. Presently, we use a Yokogawa spinning disc to
> detect CFP, YFP, and mRFP variants. We are getting by, but would like to be
> able to do the work at more physiological chimera expression levels. Typical
> exposures currently range from 500 ms (good) to 3000 ms (bad) per channel.
>
>     I am achieving adequate resolution with a 40X NA1.3 oil immersion lens
> (Zeiss). I am using the Hamamatsu ORCA-ER CCD. The camera has 6µm pixels,
> which we use unbinned. I do not think I can tolerate any lower resolution.
>
>     I have two options, and a question associated with each:
>
>     (1) Move to a TIRF system. What are your opinions about how much gain in
> sensitiviy this may provide? Has anyone worked with the Zeiss TIRF module
> for the Axiovert 200M?
>
>     (2) Change cameras. I have tried a few back thinned EM-CCDs, but did not
> find that they offered much benefit once the pixel size was corrected for.
> Is there a better option, that offers high resolution, high sensitivity, and
> low background noise? Obviously, there will be compromises. What are your
> opinions of intensified CCD cameras?
>
>     Best regards,
>
>     -Steve Bunnell
>
>
>
>
> ****************************************************************************
> Stephen C. Bunnell, Ph.D.
> Assistant Professor
> Tufts University Medical School
> Department of Pathology
> Jaharis Bldg., Room 512
> 150 Harrison Ave.
> Boston, MA 02111
>
> Phone: [phone]
> Fax:   [phone]
> Email: [hidden email]
>

-------------------------------------------------

[vb-2]

Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

Hi Steve,

I think you have the only one option  - 2x2 binning with Orca-AG (ER, Orca II).
I haven't compared -AG and -ER side-by-side but AG might be slightly better.
OrcaII will be cleraly better especially at 500 msec exposures or longer.
 
mRFP1 is a good choice in relation to SNR, having mCherry would cause less
aggregation. CFP, GFP channels are very noisy. YFP is OK.

EM-CCDs with the gigantic pixels and 40x lens will cause huge loss of signal.
Fom my experience, as 60x NA 1.4 is the brightest lens, and Orca ER (AG or II)
at 2x2 binning is the best compromise. Nikon 100x NA 1.45 TIRF lens would give
a signal of only slightly lower intensity than 60x NA 1.4 likely due to lower
undersampling.    

TIRF may not work as you would excite molecules which are up to 200 nm away
from the glass. Moreover TIRF people often use a thicker cover glass to reduce
optical artefacts.

Good luck,

Vitaly
NCI-Frederick
301-846-6575

Quoting Stephen Bunnell <[hidden email]>:


Quoting Stephen Bunnell <[hidden email]>:

> Search the CONFOCAL archive at
> http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
>
> Dear microscopists,
>
>     Here's the situation:
>
>     I need to reduce my exposure times.
>
>     I am visualizing dynamically moving complexes (~150nm/s) that are
> composed of low abundance signaling molecules, and are close to the
> coverslip-water interface.
>
>     I need to visualize these structures, in multiple colors, every 3-5
> seconds- preferably faster. Presently, we use a Yokogawa spinning disc to
> detect CFP, YFP, and mRFP variants. We are getting by, but would like to be
> able to do the work at more physiological chimera expression levels. Typical
> exposures currently range from 500 ms (good) to 3000 ms (bad) per channel.
>
>     I am achieving adequate resolution with a 40X NA1.3 oil immersion lens
> (Zeiss). I am using the Hamamatsu ORCA-ER CCD. The camera has 6µm pixels,
> which we use unbinned. I do not think I can tolerate any lower resolution.
>
>     I have two options, and a question associated with each:
>
>     (1) Move to a TIRF system. What are your opinions about how much gain in
> sensitiviy this may provide? Has anyone worked with the Zeiss TIRF module
> for the Axiovert 200M?
>
>     (2) Change cameras. I have tried a few back thinned EM-CCDs, but did not
> find that they offered much benefit once the pixel size was corrected for.
> Is there a better option, that offers high resolution, high sensitivity, and
> low background noise? Obviously, there will be compromises. What are your
> opinions of intensified CCD cameras?
>
>     Best regards,
>
>     -Steve Bunnell
>
>
>
>
> ****************************************************************************
> Stephen C. Bunnell, Ph.D.
> Assistant Professor
> Tufts University Medical School
> Department of Pathology
> Jaharis Bldg., Room 512
> 150 Harrison Ave.
> Boston, MA 02111
>
> Phone: [phone]
> Fax:   [phone]
> Email: [hidden email]
>

-------------------------------------------------

[George McNamara]

Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal Hi Stephen,

First, consider some wetware changes:

CFP (presumably ECFP?) -> mCerulean ... Piston lab
YFP (presumably EYFP?) -> mVenus ... Miyawaki lab
mRFP (presumably mRFP1?) -> mKate ... just published by Scherbo et al in Nature Methods (PubMed ID 17721542)


The mKate article compares all the far red fluorophores.

camera tips:
1. have you optimized (maximized) the gain setting on the ORCA? Your software default may be low gain. If you change the gain, you may also need to tweak the offset to keep the 'floor' above zero (I like having it at about 100 or 200 intensity levels). There are 10 gain settings on the ORCA-ER Firewire cameras. These are analog gains, before the digitizer, so output is continuous.
   Given your exposure times (3 seconds!) I suspect you are using low gain.
2. The ORCA-ER Firewire has an "IR mode", that is useful for far-red/near-infrared signals (on for mRFP1 or mKate or Cy5, off for visible fluorophores). It increases the QE in the red by grabbing photoelectons from deeper in the CCD (blue is absorbed first, then green, then red, then IR), at the cost of a slightly higher background (more volume of the CCD, results in also grabbing more thermal electrons, so keep it off for blue, green, yellow).
Zeiss AxioVision and MetaMorph (and I'm sure Compix) users can enable these options in macros/journals, as well as in the acquire dialogs.
My thanks to Butch Moomaw of Hamamatsu for explaining these features.

emission filters: do you have spectrally optimal emission filters? Are the filters in excellent shape (no fingerprints, no pinholes)?

illumination: Assuming you are using laser(s), are you getting optimum output from them to the objective lens? I'm guessing you are using an Argion ion laser for the CFP and YFP - is its power output (knob or slider for Amperage, not AOTF) adjusted for good power output? Too low Amperage results in instability, too high results in shortened tube lifetime (max power can result in 1/10th tube life - of course if the tube is under service contract and/or the experiments are worth it, cranking the Amperage may get you better data).

Photobleaching: if you have higher illumination, you may (should!) have more photobleaching. Oxyrase (www.oxyrase.com) or other additives have been used to decrease O2


New camera and/or new lenses:  Get demos!!!

Lenses: besides (getting demo of) the 1.45 NA TIRF lenses, operated in TIRF and non-TIRF modes, how well does the Zeiss 63x/1.2 NA water immersion lens work for your live cell experiments? If you are micrometers away from the coverglass, RI matching MIGHT be preferable to high NA.

If you want to go faster, simultaneous acquisition of the three emission channels is possible with multiple cameras or the dualview/quadview from Optical Insights/Photometrics (some other companies have side-by side devices, and see Prabhat et al 2007 PNAS for an interesting configuration for a TIRF/widefield system).


Vitaly wrote in another response: TIRF may not work as you would excite molecules which are up to 200 nm away from the glass. Moreover TIRF people often use a thicker cover glass to reduce optical artefacts.

Why would a thicker coverglass reduce optical artifacts in TIRF? Measuring the coverglass to make sure it is correct for the emission side (i.e. 170 um for a standard 0.17 coverglass corrected lens) should be what matters. The TIRF excitation is propagating inside the coverglass at the coverglass-cell culture medium interface - how would it know how thick the coverglass is to the objective lens side?



best wishes,



At 12:38 PM 8/31/2007, you wrote:

Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

Dear microscopists,

    Here's the situation:

    I need to reduce my exposure times.

    I am visualizing dynamically moving complexes (~150nm/s) that are
composed of low abundance signaling molecules, and are close to the
coverslip-water interface.

    I need to visualize these structures, in multiple colors, every 3-5
seconds- preferably faster. Presently, we use a Yokogawa spinning disc to
detect CFP, YFP, and mRFP variants. We are getting by, but would like to be
able to do the work at more physiological chimera expression levels. Typical
exposures currently range from 500 ms (good) to 3000 ms (bad) per channel.

    I am achieving adequate resolution with a 40X NA1.3 oil immersion lens
(Zeiss). I am using the Hamamatsu ORCA-ER CCD. The camera has 6µm pixels,
which we use unbinned. I do not think I can tolerate any lower resolution.

    I have two options, and a question associated with each:

    (1) Move to a TIRF system. What are your opinions about how much gain in
sensitiviy this may provide? Has anyone worked with the Zeiss TIRF module
for the Axiovert 200M?

    (2) Change cameras. I have tried a few back thinned EM-CCDs, but did not
find that they offered much benefit once the pixel size was corrected for.
Is there a better option, that offers high resolution, high sensitivity, and
low background noise? Obviously, there will be compromises. What are your
opinions of intensified CCD cameras?

    Best regards,

    -Steve Bunnell




****************************************************************************
Stephen C. Bunnell, Ph.D.
Assistant Professor
Tufts University Medical School
Department of Pathology
Jaharis Bldg., Room 512
150 Harrison Ave.
Boston, MA 02111

Phone: [phone]
Fax:   [phone]
Email: [hidden email]

 

George McNamara, Ph.D.
University of Miami, Miller School of Medicine
Image Core
Miami, FL 33010
[hidden email]
[hidden email]
305-243-8436 office

[Ian Dobbie-2]

Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

Stephen Bunnell <[hidden email]> writes:

[snip]
>     I have two options, and a question associated with each:
>
>     (1) Move to a TIRF system. What are your opinions about how much gain in
> sensitiviy this may provide? Has anyone worked with the Zeiss TIRF module
> for the Axiovert 200M?

This will lead to a massive gain in signal to noise. So long as your
proteins of interest are in the magic 100nm from coverslip range you
can get down to signal molecule sensitivity. Haven't a clue about the
Zeiss TIRF module I'm afraid.

>     (2) Change cameras. I have tried a few back thinned EM-CCDs, but did not
> find that they offered much benefit once the pixel size was corrected for.
> Is there a better option, that offers high resolution, high sensitivity, and
> low background noise? Obviously, there will be compromises. What are your
> opinions of intensified CCD cameras?

The only real advantage of EMCCDs is in imaging FAST. An EMCCD should
allow you to image all the colour ranges you want in a fraction of a
second (of course depending on you filter changer speed) rather than
several seconds. The two drawbacks are increased pixel size, but this
shouldn't be a factor if you use an optavar, and effectively 1/2 the
QE. A back-thinned EMCCD shouldn't be too much worse than a front
illuminated conventional CCD such as the ORCA-ER.

As mentioned before in this thread a combination of both TIRF and an
EMCCD would be ideal but you may well not have the budget for that. As
you have tried some EMCCDs and not found them too useful and you don't
want to image really quickly, it is likely that TIRF would be the best
single option to try. I would try to get a loan of some equipment for
a week or two to see how it goes.

Ian

[Beat Ludin]

Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

At 18:38 31-08-2007, you wrote:
>     (2) Change cameras. I have tried a few back thinned EM-CCDs, but did not
>find that they offered much benefit once the pixel size was corrected for.

I suspect that this is because you have relatively high background
fluorescence so you need high photon counts to get an acceptable S/N.
Under these circumstances, EM-CCDs may be not better than or even
inferior to normal CCDs. So trying a back-illuminated conventional
CCD would be a first advice although I wouldn't expect a dramatic
effect (the QE of the Orca-ER is pretty good already). But you may be
able to gain a lot more by reducing the background:

- use riboflavin-free and serum-free medium for imaging (you can use
a balanced salt solution like Earl's or Hank's for a first test to
see the effect)
- I found that culturing cells in the presence of Trolox-C can reduce
intracellular background several-fold. But you need to do the proper
controls and find the right concentration to avoid side-effects.
- optimize your filters
- minimize stray light (reduce ambient light)
- check different immersion oils, to find the best compromise between
losses due to autofluorescence and due to mismatch in refractive OR
dispersive index.

Of course, TIRF has the potential of reducing the background a lot,
if it is compatible with your application.

Regards,  Beat