Hi,

all the confocal microscopes have analysis softwares, you can use one of these software to quantify the fluorescence or you can use "Image J" a software that you can take free from internet.

To see if the protein is in the nucleus, you can do double immunofluorescence with DAPI or Hoecst which are specific for nuclei.

Best regards

Hello Casadei,

often the programms from the fluorescence microscope have analysis software where you can count the fluorescence. Maybe this can help. The question with the signal in cytoplasm, I think you can focus on the plateau of the nucleus and then you don't have this problem. Or you use confocal microscopy.

Best,

Sven

In my opinion, immunofluorescence is a qualitative analysis.... you can say that a protein changes its amount in whole the cell (in dependence of the amount of light), but not in a particular cell compartment. The quantification may be done with a wastern blot after fractionation of nuclear proteins from total proteins, using a nuclear protein as inner control.

best regards

Simona

You could perform a confocal microscopy assay. In the Z plane you can distinguish between the signal from the nucleus and the signal from the cytoplasm above. To get a measure of the amount of signal you may follow Simona's advice.

Hi,

To answer your second question: the nucleus occupies most of the cell volume in many cell types. In cell culture (e.g. flattened cells), the portions of cytoplasm at the bottom and top of the nucleus are relatively "thin". As Sven suggested above, the use of confocal microscopy will certainly help you to distinguish between cytoplasmic and nuclear signal. When scanning at different optical levels within the cell depth, and if you do not have nuclear labeling, your nucleus should appear black on the monitor (when for example focusing in the middle of the cell). You could also label DNA with DAPI (or any other nuclear probes) to clearly delineate the nuclear volume.

Regards

Michael

Hi,

all the confocal microscopes have analysis softwares, you can use one of these software to quantify the fluorescence or you can use "Image J" a software that you can take free from internet.

To see if the protein is in the nucleus, you can do double immunofluorescence with DAPI or Hoecst which are specific for nuclei.

Best regards

Thanks for answer,

well, fractionation is for sure the most accurate (if enough control for the specificity).

But histology also have advantage like looking how nucleus is looking like, if cell are looking healthy...

I´m wondering if ratio are more suitable, like ratio signal in nucleus compare to total signal....

Here are links: http://www.plospathogens.org/article/info%3Adoi%2F10.1371%2Fjournal.ppat.1001186

http://www.plosone.org/article/info:doi/10.1371/journal.pone.0009806

Hi !

Absolute fluorescence quantification is indeed *quite* dangerous (and very often plain wrong), due to the various experimental conditions you can adjust during image acquisition. Assuming your read-out is linear (i.e. the captured light increases linearly with the number of fluorophores), at best you can rely on comparative analyses (one cell or cell region vs another cell *within the same image*), and average out this analysis over hundreds of cells.

As Sven mentioned, the answer to your specific question depends on the type of microscope you are using. A wide-field fluorescence microscope will capture all incoming light (i.e. from your focal plane and around it), therefore the layer of cytoplasm above the nucleus will bias your read-out. On the other hand, a confocal microscope rejects out-of-focus light, such that your read-out should in theory only correspond to your focus plane (which you then have to make sure that it cuts through the nucleus, not above). In the latter case however, you may lose some relevant signal (precisely from the other focal planes where there is no ambiguity), which could improve your statistical analysis.

As for analysis softwares, there are a number of open solutions out there that are more efficient than the ones bundled with microscopes (we use Icy in our lab, for instance, icy.bioimageanalysis.org).

Hope this helps

Alexandre

Hi Nicolas,

You can perform a double immunofluorescence with a nuclear marker (like DAPI) and your usual primary antibody against your protein, and quantify the intensity of an eventual colocalized signal with a confocal microscope. However, if you have the opportunity, the best way to answer your question will come at the electron microscopic level, after staining your protein with a gold-conjugated antibody. But of course, it will depend on the importance of answering this question and how much precise your signal need to be, as an EM experiment can be time consuming and expensive...

Best regards,

Cedric.

Confocal microscopy generates spectacular images in different planes by, 'saturating' the fluorescent signal coming from the tissue in the microscope. If by quantify you mean just see if the protein is in the nucleus, then confocal is the way to go. If your manipulation just knocks down the expression of the protein in the nucleus and you want quantify the decrease, then you need to lyse cells, obtain nuclear fraction, and use western blot to quantify protein of interest.

When I have used immunofluorescence to quantify the relative amount of protein in the nucleus vs cytoplasm, I've used confocal for the reasons mentioned already by others. A few things you need to consider that have not yet come up: be careful of the nucleoli in your cells. The protein may localize more strongly or not at all in this compartment and therefore change the pixel intensity. Using Hoechst or DAPI to counterstain will show you where the nucleoli are.

You can really only do a relative nuclear:cytoplasmic intensity in each cell and average this over many cells, as mentioned above. I took two areas from the nucleus, two areas from the cytoplasm, subtracted the background signal (from an area not covered by cells) and then calculated the nuclear/cytoplasmic ratio. This gives a number from 0-1 if there is more signal in the cytoplasm, but a number from 1-many hundreds (depending on if the image is 8 or 16-bit) if there is more signal in the nucleus. Because of this, you will likely not have normally distributed data (not a bell shape), and you need to take that into account in your stats.

If you have over-saturated the detector on your scope (most programs have a setting to show you) when taking your images, your analysis will be invalid.

To make sure you can determine the extent of the cytoplasm and the correct location of the nucleus, take each field with filters for DIC or phase contrast, hoechst and protein-of-interest.

If using the software on your scope for the analysis isn't feasible, ImageJ is a good alternative, although I found it a bit clunky at first.

Having said all that, hopefully you also have a construct with an epitope tag so you can distinguish your transfected protein from native protein, or HEK do not express your protein-of-interest. It's also likely a good idea to back up your analysis with sub-cellular fractionation and WB as others suggested.

As many useful commentary has already been noted above, I will only add another tidbit from my experience. I have used ImagePro Plus in the past, and found it very valuable for densitometry. Once you acquire your image (continued emphasis on using the same acquisition parameters/procedure, white balance, etc.) you can select your area of interest and make the aforementioned calculations. I thought the software was user-friendly, but it does require a license.

Regards,

Sara

Try subcellular fractionation to obtain intact nuclei, membranes and cytosolic fractions. Then quantify the signal by WB analysis. You even could stain isolated nuclei seeded onto coverslips.

I have done it in brain tissue (see link bellow) but should work in cultured cells as well.

http://www.sciencedirect.com/science/article/pii/S0306452212006574

Thank you so much for all these answer!s!!!

I like the idea of the confocal and the DAPI (like said Ronny Sczech or

Thank you Michaël Marie, i did not knew that the nucleus was a such big structure, that will facilitate analysis if cytoplasm is really thin compaire to the nucleus size.

As several people suggested, i will use ImageJ, mainly because i like open source project, and also because I already used it.

We already generated these data, so I will make a try and put some images online with my code (for imageJ) to see how it looks and what it gives.

To answer Alexandre Dufour, we use a wide-field fluorescence microscope and an apotome. Hope it will be good enough.

I will have a look on data handling on R when picture analysis will be good enough (thank you Eric Patterson for the trick of normalization)

And to answer Gontzal García del Caño, we will confirm results using fractionation, that is for sure needed.

Do you think that I should publish the small bach on this post or should i oppen a new question?