Biology Reference
In-Depth Information
To determine the mitochondrial or non-mitochondrial origin of a
protein we use a quantitative measure of protein abundance across
the differentially enriched mitochondria isolations. In its simplest
form this involves spectral counting, in which the number of pep-
tides identifi ed for a given protein is used as a measure of its abun-
dance. For example if for a particular protein 6 peptides were
matched in the low-purity sample, 9 in the medium-purity sample,
and 12 in the high-purity sample, then this protein would be
confi rmed as a mitochondrial protein. As the number of peptides
identifi ed for this protein increases in abundance with increasing
mitochondrial purity we can confi rm its mitochondrial origin.
A further extension of this method is a statistical analysis of the
changes observed for a particular protein. To do this using stan-
dard statistical such as the
t
-test, a number of statistical assump-
tions must be met. To meet these Zybailov et al. [
52
] developed
the normalized spectral abundance factor (NSAF), which includes
a parameter that also compensates for variations in different pro-
tein lengths. They showed that the natural log of each NSAF
results in a Gaussian distribution of a dataset, permitting analysis
using the
t
-tests. However it becomes essential to avoid ln(0) and
thus zero spectral counts need to be replaced by a fractional value
empirically derived for each data set analyzed. This aim is to fi nd
the smallest value that maintains a normal distribution of the
ln(NSAF) data [
52
] and replace all zero values with this fractional
value. Once these statistical parameters are implemented it becomes
possible to determine the statistical signifi cance of a particular pro-
tein enrichment/depletion, confi rming or refuting its mitochon-
drial origin. However in some cases, diffi culties with abundant
mitochondrial proteins can occur where all possible peptides for a
protein are found in all three samples and thus no enrichment is
observed; fortunately in most cases we have encountered these
proteins to be generally well-known mitochondrial proteins that
have had their subcellular localization previously confi rmed by
other techniques.
Data Analysis
4
Notes
1. 0.3-0.4 M of osmoticun is required using either sucrose or
mannitol to maintain mitochondrial structure by preventing
excessive swelling and rupturing of membranes.
2. 2-5 mM of divalent cation chelator (EDTA or EGTA) acts to
inhibit the activity of various phospholipases and proteases.
3. 25-50 mM of a basic pH buffer system (MOPS, TES, or
Na-pyrophosphate) prevents media acidifi cation during vacu-
ole rupture.
