Data Generated by Quantitative Liquid Chromatography-Mass Spectrometry Proteomics Are Only the Start and Not the Endpoint: Optimization of Quantitative Concatemer-Based Measurement of Hepatic Uridine-5′-Diphosphate–Glucuronosyltransferase Enzymes with Reference to Catalytic Activity

Brahim Achour; Alyssa Dantonio; Mark Niosi; Jonathan J. Novak; Zubida M. Al-Majdoub; Theunis C. Goosen; Amin Rostami-Hodjegan; Jill Barber

doi:10.1124/dmd.117.079475

Article Figures & Data

Download figure
Open in new tab
Download powerpoint
Fig. 1.
Schematic of the methodological approach used to assess abundance levels of UGT enzymes using raw LC-MS/MS data. Representative peptides are selected using criteria outlined in Supplemental Methods. This selection process applies to targeted (MRM) and untargeted/global studies. Selection from peptides that are detected consistently in an LC-MS/MS experiment should take into account the uniqueness and the stability of the peptides. Selected fragments should be stable and representative of the peptide (of sufficient length) to return consistent quantification. Correction factors should be applied for label incorporation, especially when low abundance proteins are analyzed. The spike ratio should be consistent with the dynamic range of expression of the target proteins. Normalization should be consistent across all samples and measured parameters. An example of this process is shown in Fig. 2.
Download figure
Open in new tab
Download powerpoint
Fig. 2.
An example of the assessment process applied to UGT2B15 in sample HH06: choice of peptide standard and correction for label incorporation efficiency (A and B), featuring elution profiles of QconCAT alone (dashed lines), and QconCAT and analyte sample (continuous lines) for heavy (blue) and light (red) peptides (peptide 1: WIYGVSK; peptide 2: SVINDPVYK). Differences between total protein mass measurements using Bradford and BCA assays (C). Calculation of UGT2B15 abundance in sample HH06 using the outlined correction factors and their contributions to the change in reported abundance (D). In (C), the arrow shows sample HH06, and data points in red reflect a difference in content higher than a cut-off relative error () of 15% for each sample j between the two protein content assays. Overall differences in mean and distribution between data from the two assays were nonsignificant according to Mann-Whitney U test; however, individual values were poorly correlated. In (D), the overall shift in abundance was −19.8% () for enzyme i = UGT2B15 before and after optimization, with the main contributing factor being the selection of peptide/fragment transitions (%RE = −29.8%), followed by total HLM protein content (+18.3%) and correction for label incorporation (+6.1%). Text in purple font reflects corrected values. BA, Bradford assay; QIS, QconCAT-based internal standard.
Download figure
Open in new tab
Download powerpoint
Fig. 3.
Correlation between individual UGT enzyme abundances and activity rates (n = 24) using the original dataset (A) and the reassessed data on the basis of the proposed strategy (B). Moderate to strong, statistically significant correlations are shown in blue and weak correlations in gray. Units of abundance-measurement are picomoles per milligram of HLM protein, and units of catalytic activity are nanomoles (glucuronide) per minute per milligram of HLM protein. Substrates used for activity measurement are: β-estradiol (UGT1A1), chenodeoxycholic acid (UGT1A3), trifluoperazine (UGT1A4), 5-hydroxytryptophol (UGT1A6), propofol (UGT1A9), zidovudine (UGT2B7), S-oxazepam (UGT2B15). Rs, Spearman correlation coefficient. Dashed lines represent lines of regression.
Download figure
Open in new tab
Download powerpoint
Fig. 4.
Correlation matrix of QconCAT-derived individual UGT enzyme abundances (n = 24) and activity rates (abundance vs. activity). Strong, statistically significant correlations are shown in blue. Units of abundance-measurement are picomoles per milligram of HLM protein, and units of catalytic activity are nanomoles (glucuronide) per minute per milligram of HLM protein. AZT, zidovudine; CDCA, chenodeoxycholic acid; EST, β-estradiol; 5HTOL, 5-hydroxytryptophol; OXAZ, S-oxazepam; PRO, propofol; TFP, trifluoperazine. Supplemental Table 2 shows the statistical analysis used to generate the abundance-activity correlation matrix.

Tables

Figures
Additional Files

View popup

TABLE 1

Assessment of peptides and fragments used to quantify each of the eight UGT enzymes

Protein	Peptide Sequence^{^a}	In Silico Score (0 to 1)^{^b}	Theoretical Assessment^{^c}	Incorporation Correction Factor (%)^{^d}	MRM Transitions Monitored  [(m/z)^z]_peptide → [(m/z)^z]_fragment^{^e}
UGT1A1	D₇₀GAFYTLK₇₇^{^f}	0.354	++	4.0	457.73⁺²/524.31⁺¹ (y₄)	457.73⁺²/671.38⁺¹ (y₅)	457.73⁺²/742.41⁺¹ (y₆)
UGT1A1	T₇₈YPVPFQR₈₅	0.363	++	2.0	504.27²⁺/547.30⁺¹ (y₄)	504.27²⁺/646.37⁺¹ (y₅)	504.27²⁺/743.42⁺¹ (y₆)
UGT1A3	Y₁₆₄LSIPTVFFLR₁₇₄^f	0.514	+++	1.0	678.39²⁺/681.41⁺¹ (y₅)	678.39²⁺/782.46⁺¹ (y₆)	678.39²⁺/879.51⁺¹ (y₇)
UGT1A4	Y₁₇₅IPCDLDFK₁₈₃^{^f}^,^{^g}	0.432	+++	5.0	585.78²⁺/637.32⁺¹ (y₅)	585.78²⁺/797.35⁺¹ (y₆)	585.78²⁺/894.40⁺¹ (y₇)
UGT1A4	G₁₈₄TQCPNPSSYIPK₁₉₆^g	0.522	++	5.0	724.85²⁺/791.43⁺¹ (y₇)	724.85²⁺/905.47⁺¹ (y₈)	724.85²⁺/1002.53⁺¹ (y₉)
UGT1A6	S₁₀₃FLTAPQTEYR₁₁₃^f	0.554	+++	2.0	656.83²⁺/793.38⁺¹ (y₆)	656.83²⁺/864.42⁺¹ (y₇)	656.83²⁺/965.47⁺¹ (y₈)
UGT1A6	V₂₅₀SVWLLR₂₅₆^{^h}^,^ⁱ	0.354	+	1.5	436.77²⁺/587.37⁺¹ (y₄)	436.77²⁺/686.43⁺¹ (y₅)
UGT1A9	E₁₃₉SSFDAVFLDPFDNCGLIVAK₁₅₉^f,g	0.553	++	4.0	1172.56²⁺/1233.63⁺¹ (y₁₁)	1172.56²⁺/1348.66⁺¹ (y₁₂)	1172.56²⁺/1461.74⁺¹ (y₁₃)
UGT2B4	F₁₇₄SPGYAIEK₁₈₂	0.391	+++	5.0	506.26²⁺/623.34⁺¹ (y₅)	506.26²⁺/680.36⁺¹ (y₆)	506.26²⁺/777.41⁺¹ (y₇)
UGT2B4	A₃₂₁NVIASALAK₃₃₀^f	0.412	+++	4.0	479.29²⁺/560.34⁺¹ (y₆)	479.29²⁺/673.42⁺¹ (y₇)	479.29²⁺/772.49⁺¹ (y₈)
UGT2B7	T₄₁ILDELIQR₄₉^{^f}	0.426	++	2.0	550.82²⁺/658.39⁺¹ (y₅)	550.82²⁺/773.42⁺¹ (y₆)	550.82²⁺/886.50⁺¹ (y₇)
UGT2B7	A₂₅₃DVWLIR₂₅₉^ⁱ	0.300	+	1.5	436.75²⁺/587.37⁺¹ (y₄)	436.75²⁺/686.43⁺¹ (y₅)	436.75²⁺/801.46⁺¹ (y₆)
UGT2B15	W₉₇IYGVSK₁₀₃^{^h}	0.325	+	3.0	426.73²⁺/553.30⁺¹ (y₅)	426.73²⁺/666.38⁺¹ (y₆)
UGT2B15	S₄₃₂VINDPVYK₄₄₀^f	0.406	++	5.0	517.78²⁺/506.30⁺¹ (y₄)	517.78²⁺/735.37⁺¹ (y₆)	517.78²⁺/848.45⁺¹ (y₇)
MetCAT^{^j}	GVNDNEEGFFSAR^{^k}	0.561	++++	2.5	721.32²⁺/813.39⁺¹ (y₇)	721.32²⁺/942.43⁺¹ (y₈)	721.32²⁺/1056.47⁺¹ (y₉)

↵^a Peptide sequences as defined by the human UniProtKB database (http://www.uniprot.org). Subscript number labels on the C- and N-terminal amino acids of peptide sequences denote their positions in the UGT protein sequences on the basis of their database entries. The terminal lysine (K) and arginine (R) residues were labeled using [¹³C₆] stable isotopes in the QconCAT standard.
↵^b In silico assessment was carried out using CONSeQuence algorithm on the basis of charge, hydrophobicity and secondary structure (Eyers et al., 2011).
↵^c Theoretical assessment on the basis of criteria outlined in Supplemental Information; arbitrarily, +, ++, +++, and ++++ scores were assigned to peptides under assessment (highest score, +++++) by two independent analysts.
↵^d The proportion of light to heavy peptide owing to inefficient incorporation of the ¹³C label needed to correct quantification ratios; this can be variable from batch to batch.
↵^e Up to three transitions for each peptide were designed in silico using Skyline (superscript indicating charge states, z); selected fragments were then appraised on the basis of unique sequences, m/z, quality of elution profiles, and the CV of the returned quantitative ratios. In this table, only the light (native) peptide transitions are listed, where the y-ions (subscript indicates the length of the sequence) were used.
↵^f Peptide selected for quantification of each UGT enzyme on the basis of the selection criteria outlined in Supplemental Information and in silico appraisal.
↵^g Cysteine residues were alkylated (by carbamidomethylation), necessitating an increment of +57.0215 Da in monoisotopic mass of peptides and certain fragments.
↵^h Two transitions were designed and monitored for peptides VSVWLLR (UGT1A6) and WIYGVSK (UGT2B15), which returned low scores on the basis of theoretical, in silico, and fragment assessments. These peptides were excluded from analysis.
↵ⁱ The isobaric sequences VSVWLLR (UGT1A6) and ADVWLIR (UGT2B7) were overlapped on the chromatogram owing to close retention times (with the same m/z of parents and fragments). These peptides were excluded from analysis.
↵^j MetCAT: QconCAT used as a standard for the quantification of human liver P450 and UGT enzymes (Russell et al., 2013; Achour et al., 2015).
↵^k Sequence of QconCAT-based internal standard used for quantification of the QconCAT.

Additional Files

Figures
Tables

Data Supplemental

Supplemental Data -
Supplemental Methods

Supplemental Figure 1 - Elution profiles of UGT Q-peptides analyzed in the same LC-MS/MS assay (sample HH06)

Supplemental Figure 2 - Correlation of UGT abundance values measured using the two QconCAT peptides representing UGT1A1 (A), UGT1A4 (B), UGT1A6 (C), UGT2B4 (D), UGT2B7 (E) and UGT2B15 (F) after optimization

Supplemental Figure 3 - Cross-laboratory evaluation of UGT abundance measurements using stable isotope-labeled
(SIL) peptide standards and quantification concatemer (QconCAT) standard: Box and whiskers plot of the abundance
measurements (n=24) of UGT enzymes quantified by the two methods (A) with the individual values shown in panel
(B)

Supplemental Figure 4 - Correlation between individual protein abundance measurements (n=24) of UGTs 1A1 (A), 1A3 (B), 1A4 (C), 1A6 (D), 1A9 (E), 2B4 (F), 2B7 (G) and 2B15 (H) using two proteomic methodologies (SIL vs QconCAT) after optimization of QconCAT data analysis

Supplemental Figure 5 - Correlation matrix of individual protein abundances of UGT enzymes (abundance vs abundance) using QconCAT methodology (n=24)

Supplemental Table 1 - Comparison of QconCAT-based UGT measurements before and after optimization of data analysis

Supplemental Table 2 - Correlation matrix of QconCAT-derived individual UGT enzyme abundances (n=24) with activity rates (abundance vs activity)

Supplemental Table 3 - Correlation matrix of individual protein abundances of UGT enzymes (abundance vs abundance) using QconCAT methodology (n=24)

Supplemental Table 4 - Enzymes with established QconCAT-based quantification methods after optimization

References