It sounds likely that the "per mg protein" is relating CAT[alase]
activity to total protein, probably obtained through running a
Bradford, Lowry or Biruet assay on a crude protein extract. This gives
an approximation of protein content in solution, but is largely only
reactive to the aromatic residues, so the effect depends not only on
the protein concentration but also on the protein's amino acid
content.
If the "per mg protein" numbers were obtained via ELISA, or something
like HIS-tag purification, then the "protein" would be mostly/highly
pure CAT[alase].
Seems that fresh weight is not preferred due to things like water
content variation. If you know the fresh weight activity, and the
activity of extracted protein, along with the activity of
highly/mostly pure CAT[alase] then you would know what the
concentration of CAT[alase] in fresh tissue, and you'd know the ratio
of activity of the enzymes. The latter measurement would best be made
with highly pure protein, but that's harder to do than crude protein
extraction, and knowing the activity of CAT[alase] to total protein
might often be enough to get your answer (say for example, is my GMO
E.coli producing more insulin-production enzyme than pre-GMO). You
don't need highly pure protein to answer that question because E.coli
to E.coli will have similar total protein content, given they're the
same or similar strain. So if you assume the pure enzyme's activity is
constant, you can compare the activity amongst samples, if you don't
know if the activity is the same (maybe one CAT[alase] gene is mutated
and works slower than the other) you need to do a crude protein
extraction and compare the activities... but ideally you'd purify the
enzyme for each sample and compare highly-pure sample activities to
determine if their rate was different or not.
"
Catalase activity was expressed per mg of extractable protein
(specific activity) rather than on a per seed basis since the amount
of extractable protein did not vary among the various seed samples (it
was always close to 1.3 mg seed–1) whereas dry and fresh seed weights
markedly changed during seed development. Indeed, sunflower seeds
contain c. 15–20% albumins, which are extractable with the buffer used
in this study, and c. 80% reserve protein (globulins, prolamins, and
glutelins) (Lusas, 1985), which accumulate during seed‐filling, but
were not extracted by the unsalted potassium phosphate buffer
"
http://jxb.oxfordjournals.org/content/55/396/475.full
Here's a highly-cited paper on determining CAT activity:
Catalase in vitro
Hugo Aebi, Methods in Enzymology
Volume 105, 1984, Pages 121–126
http://libgen.org/scimag5/10.1016/S0076-6879%252884%252905016-3.pdf
>
https://groups.google.com/d/msgid/diybio/CAK73W77Mxp9S%2BQ5-KB2LzxOdtxNP4wRi-xPhRHQybunahQgjjA%40mail.gmail.com.
-Nathan