Choosing a primary antibody for immunofluorescence?

Immunofluorescence is the labeling of antibodies or antigens with fluorescent dyes. This technique is often used to visualize subcellular distribution of biomolecules of interest. Immunofluorescent-labelled tissue sections are studied using a fluorescence microscope or by confocal microscopy.

Most commonly, immunofluorescence employs two sets of antibodies: a primary antibody is used against the antigen of interest; a subsequent, secondary, dye-coupled antibody is introduced that recognizes the primary antibody. In this fashion the researcher may create several primary antibodies that recognize various antigens, but, because they all share a common constant region, may be recognized by a single dye-coupled antibody. Typically this is done by using antibodies made in different species. For example, a researcher might create antibodies in a goat that recognize several antigens, and then employ dye-coupled rabbit antibodies that recognize the goat antibody constant region (denoted rabbit anti-goat). This allows re-use of the difficult-to-make dye-coupled antibodies in multiple experiments.

In some cases, it is advantageous to use primary antibodies directly labelled with a fluorophore. This direct labelling decreases the number of steps in the staining procedure and, more importantly, often avoids cross-reactivity and high background problems. Fluorescent labelling can be performed in less than one hour with readily available labeling kits.

As with most fluorescence techniques, a significant problem with immunofluorescence is photobleaching. Loss of activity caused by photobleaching can be controlled by reducing the intensity or time-span of light exposure, by increasing the concentration of fluorophores, or by employing more robust fluorophores that are less prone to bleaching (e.g. Alexa Fluors or DyLight Fluors).

Many uses of immunofluorescence have been outmoded by the development of recombinant proteins containing fluorescent protein domains, e.g. green fluorescent protein (GFP). Use of such "tagged" proteins allows much better localization and less disruption of protein function.

Not sure I understand the question, because you say to "confirm" a mitochondrial protein.  Is the important part that it be MITOCHONDRIAL, or that you confirm a particular protein.  If the latter, you need a specific antibody for that protein, obviously.  If just mitochondrial, then I would think malate dehydrogenase because that is a mito marker, whereas ATP synthase and hexokinase are not selective to mitochondria.