DDA focuses on the detection of peptides and generates a qualitative list of proteins. Quanti- tative analysis is more useful since it enables us to determine differential expression of proteins within cells. In discovery proteomics greater than 3000 proteins can be routinely detected in a single experiment with only 1D fractionation, and 80% of these can generally be quantitated. Traditionally quantitation was only performed through metabolic or chemical labelling of peptides. Despite labelled samples being multiplexed thereby reducing measuring time, the high purity required for metabolic and chemical labels often makes their use prohibitively expensive.
HPLC has increased LC resolution and retention time reproducibility. While mass accuracy and speed have increased in modern MSs. These advances have enabled the development of new algorithms for label free quantitation (LFQ). LFQ is becoming more and more prevalent. LFQ does not require expensive labels, but does however require each sample to be run at least once. Measuring time on MSs is expensive, and is often the limiting factor on the number of samples run in an experiment.
Labelling
Metabolic Labelling Peptides can be labelled in a number of different ways, the most
robust form of labelling being metabolic labelling. Cells or whole organisms can be labelled. Cells are grown in tissue culture medium containing either heavy or light isotopes of the amino acids arginine and lysine. These amino acids contain isotopes of carbon or nitrogen with an additional neutron, which increases their mass by one Da, but does not alter the chemistry of heavy peptides. Heavy arginine and lysine are used because trypsin, the most common used protease in proteomics, cleaves peptides after arginine or lysine. This ensures that every
peptide always has a heavy labelled amino acid. Almost all protein’s in the heavy labelled cells need to be replaced for full incorporation to occur. Five cell doublings, results in more than 95% of all arginine and lysine amino acids being heavy labelled. Heavy and light labelled cells are harvested counted and combined in equal ratios. All downstream processing is done on this combined sample, so sample processing bias does not affect quantitation. Upon MS analysis peptides from the heavy and light labelled cells, elute at the same retention times, since they have the same chemical properties. The heavy labelled peptides masses have been altered by the number of neutrons added to the heavy amino acid. This can range from 6-10 Da depending on the number of carbons and nitrogens in the amino acid. These mass shifts are easily distinguished by modern MSs. The difference in intensity between the two peaks is used to generate a measure of relative abundance for the peptides, which is later conferred to the proteins.
Only one of the peptide peaks needs to be sequenced for identification. The MS detection software be can programmed to ignore peaks of a specific mass shift. Stable isotope labeling by amino acids in cell culture (SILAC) is generally used in 2 plex experiments but can also
be scaled up to 3 plex or more. This increase in complexity of the ms1 spectra can reduce the
depth of analysis.
Chemical Labelling Chemical labeling involves the addition of reporter groups onto pro-
teins or peptides. In order to achieve complete enzymatic peptide cleavage, or digestion, it is important to first reduce disulphide bonds and block the cysteines with iodoacetamide. This process was easily amenable to labelling, through the use of isotope coded affinity tag (ICAT) used in place of iodoacetamide. Samples are combined after labelling and digested, with trypsin. The tags contain biotin, which is used to enrich the peptides containing ICAT.
The tags fragment during ms2. The relative intensities of the reporter fragments are used to
deduce relative abundance of peptides and proteins from the respective samples. One of the drawbacks of the ICAT process is the small number of cysteine residues in proteins, so only a few peptides can be quantitated. However this is also an advantage since sample complexity
was also reduced in the process. In theory as long as a protein contained a cystein it could be quantitated.
The most commonly used forms of chemical labelling are isobaric tags for relative and
absolute quantitation (iTRAQ R) or Thermo scientificTMtandem mass tag (TMT). Cells are
lysed, proteins extracted and peptides digested. Labels are then covalently bound to the amino groups of the amino acids, either to the terminal amino group or other free amino groups.
The labels consist of a reported group and a balance group. The labels have been designed to fragment between the reported group and the balance group during CID. The resultant reported groups can then be found in the lower end of the spectrum, below the masses of any amino acids. Each reporter group’s mass is separated by 1 Da. Their relative abundance is
determined from the ratio between their peaks in the ms2 spectrum and relates to the relative
abundances of proteins from the originating samples. iTRAQ R comes in either four or eight
plex, while TMT has up to 12 labels. This multiplexing allows more samples to be compared simultaneously than with SILAC. Sample processing bias however may be introduced into the quantitation since the labelling only occurs after digestion. This variance can be measured by duplicating one of the samples or a pool of the samples, and using the deviation within this sample as a technical variance of the experiment, making any variance observed beyond this
level biological variance.28
Label Free Quantitation
In SILAC quantitation is done by comparing the integrated area under the curve of the light
vs heavy labelled peptides within the same ms1 scan. In a label free discovery experiment
a similar integration is done for every peptide and translated to an intensity value. These intensities values for peptides are then compared across different LC/MS/MS runs. The ac- curacy of LFQ quantitation is less than for labelled quantitation, due to increased variance caused by the measurements being taken across different LC/MS/MS runs. Some of the vari- ance is accounted for by normalisation with algorithms like MaxQuant label free quantification algorithm (MaxLFQ).
Here the assumption is made that most of the proteins within a cell state are responsible for homoeostatic control and therefore do not vary greatly. Most peptides are then expected to be expressed in a 1:1 manner. MaxLFQ compares individual peptides in a pairwise manner and determines which peptides should have 1:1 ratios. It uses the offset in the ratios of these peptides to normalise the remaining peptides. A more reasonable distribution with a mean of zero is then achieved. This enables the peptides and proteins that are truly differentially expressed to be more clearly discriminated.
Accuracy can be improved by technical replicates, multiple LC/MS/MS runs of the same sample, which however increases measuring time and cost.