I believe that succinate dehydrogenase and cytochrome C genes would be most ideal markers for mitochondrial function. You may also look for genes for other components of mitochondrial electron transport.

Another important factor is mitochondrial biogenesis. You can measure expression of PGC-1a, NRF1/2 or Tfam. For direct mitochondrial function, expression of subunits of the electron transport chain complexes (take care if you measure mt coded or nucleus coded subunits) or citrate synthase.

Changes in biochemical or molecular traits specific to mitochondria are a good surrogate for its function. However, mitochondria develop occasionally compensatory mechanisms which do follow the above changes. Therefore, measurements of maximal rates of mitochondrial ATP production with different substrates is the best indicative of its function.

As Andrea said, PGC-1a would be best for biogenesis.  For measuring mitochondrial content, you could measure mtDNA or citrate synthase activity.  I think the best indication of function could be enhanced from in frozen tissue is measuring OXPHOS complexes at the protein level.  Abcam sells a great antibody that includes all five:

http://www.abcam.com/total-oxphos-rodent-wb-antibody-cocktail-ab110413.html

I would suggest:

-PGC-1a for mitochondrial biogenesis

-cytochrome c, SDH and other genes of the electron transport chain for mitochondrial function

-CPT1, TFAM

These are all genes that typically show pronounced changes.