Absolute quantification by Multiple Reaction Monitoring (MRM)

The use of a triple quadrupole mass analyzer allows the detection of the specific transition from a given precursor to a user-defined fragment ion (single reaction monitoring). The precursor mass is selected in quadrupole Q1, fragmentation takes place in q2, and the fragment ion is detected in Q3. Multiple reaction monitoring (MRM)allows the detection of multiple fragment ions specific for one precursor. Signals from MRM transitions of endogenous and standard peptides are well suited for absolute quantification of the peptide/

protein under investigation. In previous studies, MRM with standard peptides has proven to be a suitable method for absolute quantification of proteins in a mixture (Abbatiello et al., 2008; Langenfeld et al., 2009). In LC-coupled MRM experiments, peak overlaps caused by co-eluting peptides can be neglected under defined conditions, namely by choosing several MRM transitions specific for a certain precursor. Operating the first and the third quadrupole as a mass filters guarantees that only the specific MRM transitions are monitored and co-eluting peptides do not influence the quantitative signal.

For absolute quantification of the hPrp19/CDC5L complex, three MRM transitions for each selected peptide sequence were designed. In all cases, the doubly charged precursor was chosen as Q1 mass, and the three most intense y-type fragment ions with an m/z above that of the precursor were chosen as Q3 masses. For the selected MRM transitions, the declustering potential (DP), entrance potential (EP), collision energy (CE), and collision cell exit potential (CXP) of the instrument were first optimized (for information about MRM transitions and optimized instrument parameters see Table A.3 in the Appendix) and then tested by analyzing the standard peptides and the endogenous peptides separately. For this

purpose, 70 ng of hydrolyzed hPrp19/CDC5L complex or 100 fmol of standard peptides were separated by LC, and the MRM transitions were monitored using the optimized parameters.

All MRM transitions were well separated and showed sufficient intensity for quantitative analysis (Figure 4.6). The analysis of the hydrolyzed hPrp19/ CDC5L complex showed no MRM transition for the heavy counterparts (Figure 4.6 A) and vice versa (Figure 4.6 B). We infer that the chosen MRM transitions are highly peptide-specific and can therefore be used for their investigation.

Figure 4.6: Specificity of the designed MRM transitions. Labeled standard and endogenous peptides were analyzed separately. The detected MRM transitions show sufficient intensity for quantitative analysis. (A) 70 ng of hydrolyzed hPrp19/CDC5L complex were separated by LC and analyzed by MRM. Traces for MRM transitions of standard peptides are empty. (B) 100 fmol of standard peptides were separated by LC and analyzed by MRM.

Traces for MRM transitions of endogenous peptides are empty. ** TVPEELVKPEELSK was not used for absolute quantification. (*) labeled amino acid (standard peptides).

As described for MALDI and ESI analyses, various amounts of hydrolyzed hPrp19/CDC5L complex were supplemented with equal amounts of all standard peptides and vice versa. In the different experiments, three MRM transitions were monitored for each standard and endogenous peptide. Figure 4.7 shows an example of the total of six MRM transitions for the coeluting standard and endogenous peptides (LGLLGLPAPK derived from CDC5L). Ratios between standard and endogenous peptides were obtained by integration of the peak areas of the corresponding transitions and the values were then used to calculate protein ratios from three technical replicates.

Figure 4.7: Example of the MRM transitions for an endogenous and the corresponding standard peptide (LGLLGLPAPK derived from CDC5L protein). The doubly charged precursor mass was selected as Q1 mass and the y6, y7, and y9 fragment ions were chosen as Q3 masses. Peptide ratios were calculated from integrated peak areas of the corresponding transitions.

The peptide ratios based on the three MRM transitions for the individual peptides are compared in bar Figure 4.8. Clearly, peptide ratios obtained for a given protein differed significantly when hydrolysis was performed in the presence of urea as compared with the corresponding ratios as obtained from hydrolysis in the presence of acetonitrile (hPrp19, SPF27, and PRL1, Figure 4.8 A). Peptide ratios obtained from hydrolysis in acetonitrile were more consistent for the same protein, i.e. peptide ratios obtained for peptides generated from the same protein show comparable values (Figure 4.8 B). The standard deviations of the protein ratios derived after hydrolysis in the presence of acetonitrile were lower than the corresponding values after digestion in the presence of urea (Table 4.6).

Figure 4.8: Peptide ratios obtained from MRM analysis of the hPrp19/CDC5L complex. Peptide ratios from three replicates of three different MRM transitions for each peptide are plotted in bar diagrams for hydrolysis in the presence of urea (A) or acetonitrile (B). (A) Peptide ratios obtained from hydrolysis in urea show no clear protein stoichiometry. (B) Peptide ratios obtained from hydrolysis in acetonitrile are more consistent than those obtained from hydrolysis in urea.

Two peptides (EAAAALVEEETR derived from SPF27 after hydrolysis in the presence of urea, Figure 4.8 A; and TGYNFQR derived from PRL1 after hydrolysis in the presence of acetonitrile, Figure 4.8 B) consistently revealed a very low peptide ratio. The low value for this peptide is probably caused by low abundance of the endogenous peptide when compared with the isotopically labeled standard peptide suggesting that the endogenous peptide might be underrepresented. Indeed, both of these peptides were also hardly detectable in MALDI-TOF/TOF and ESI-LC-MS/MS experiments and could thus not be used for calculation of the protein ratio in the previous experiments. The low values are indeed consistent with the presence of a slightly longer form of the peptide EAAAALVEEETR (SPF27) containing a missed cleavage site after hydrolysis in the presence of urea as monitored in the initial analysis (see above). The adjacent protein sequence of this particular SPF27 peptide shows several tryptic C-terminal cleavage sites (EAAAALVEEETR/R/YR/) that might increase the possibility of missed cleavages in urea, but not in acetonitrile (compare peptide ratios for SPF27 in Figure 4.8 A and B). Conversely, the peptide derived from PRL1 (TGYNFQR), which shows a low value after hydrolysis using acetonitrile, does not have any additional adjacent tryptic sites. The nearest tryptic sites are 43 positions in the N-terminal and 27 positions in the C-terminal direction. Thus, the selected peptide sequence is located in a protein region that contains only very few tryptic cleavage sites (Figure 4.9). A very long tryptic peptide containing a missed cleavage might have been generated in the presence of acetonitrile (consistently with the observation in early experiments that longer peptides with missed cleavages are generated preferentially when acetonitrile is used for digestion; see above, Figure 4.2), and its detection might be hampered in ESI and MALDI analyses. As the presence of missed cleavage site-containing peptides causes underrepresentation of the endogenous peptide to be quantified, both these peptides were excluded from the calculation of the protein stoichiometry of the hPrp19/CDC5L complex after hydrolysis in urea and acetonitrile, respectively. Since one of the peptides (TGYNFQR derived from PRL1) represents a proteotypic peptide (Table 4.3), the results for the missed cleavage sites of tryptic peptides in the presence of different denaturing agents highlight the need for thorough evaluation of experimental data before standard peptides for absolute quantification are selected.

Figure 4.9: PRL 1 sequence. The peptide TGYNFQR (underlined) is located in a protein region that contains no further tryptic cleavage sites. The nearest tryptic sites are 43 positions in N-terminal and 27 positions in C-terminal direction.

As described above, protein ratios were calculated from average peptide ratios of three technical replicates and three different MRM transitions for each peptide. The MRM experiments after hydrolysis of the hPrp19/CDC5L complex in the presence of urea revealed a stoichiometry of approximately 4:1 for hPrp19 relative to CDC5L and SPF27, 5:1 relative to PRL1, and 9:1 relative to CTNNBL1. Hydrolysis in the presence of acetonitrile and subsequent analysis by MRM resulted in a stoichiometry of 2:1 for hPrp19 relative to CDC5L, 5:1 relative to SPF27, and 4:1 relative to PRL1 and CTNNBL1 (Table 4.6).

Table 4.6: Relative protein stoichiometries within the hPrp19/CDC5L complex as determined by MRM analyses. Peptide ratios were calculated from MRM signals of endogenous and standard peptides. Average peptide ratios from replicates from three different MRM transitions for each peptide were used to calculate protein ratios. Protein stoichiometries are displayed by the ratio of hPrp19 to the relevant protein showing the stoichiometry of hPrp19 within the hPrp19/CDC5L complex. Values in parentheses were omitted for calculation of the average values. For SPF27 no ratio could be determined for dilution of the standard peptides.

Protein complex [ng] 70 70 35 17.5 70 70 Average

standard peptides [fmol] 100 100 100 100 50 25

   8M/2M urea   

Protein stoichiometry

hPrp19/CDC5 3.54 3.37 3.46 3.73 3.09 3.70 3.48 ± 0.236

hPrp19/SPF27 3.32 2.99 5.44 (11.92) / / 3.92 ± 1.330

hPrp19/PRL1 5.35 4.80 5.21 6.62 4.37 4.96 5.22 ± 0.767

hPrp19/CTNNBL1 10.00 6.78 13.03 9.03 6.39 (3.94) 9.05 ± 2.691

   80 % (v/v) acetonitrile   

Protein stoichiometry

hPrp19/CDC5 1.97 1.96 2.01 1.95 1.87 1.82 1.93 ± 0.071

hPrp19/SPF27 4.69 4.83 5.58 6.11 / / 5.30 ± 0.665

hPrp19/PRL1 4.31 3.94 3.77 3.37 3.82 3.72 3.82 ± 0.306

hPrp19/CTNNBL1 4.24 4.03 3.74 3.59 (1.97) (2.33) 3.90 ± 0.106

4.1.7 Comparison of results for absolute quantification in determining the