- Study protocol
- Open Access
A study protocol for quantitative targeted absolute proteomics (QTAP) by LC-MS/MS: application for inter-strain differences in protein expression levels of transporters, receptors, claudin-5, and marker proteins at the blood–brain barrier in ddY, FVB, and C57BL/6J mice
Fluids and Barriers of the CNS volume 10, Article number: 21 (2013)
Proteomics has opened a new horizon in biological sciences. Global proteomic analysis is a promising technology for the discovery of thousands of proteins, post-translational modifications, polymorphisms, and molecular interactions in a variety of biological systems. The activities and roles of the identified proteins must also be elucidated, but this is complicated by the inability of conventional proteomic methods to yield quantitative information for protein expression. Thus, a variety of biological systems remain “black boxes”. Quantitative targeted absolute proteomics (QTAP) enables the determination of absolute expression levels (mol) of any target protein, including low-abundance functional proteins, such as transporters and receptors. Therefore, QTAP will be useful for understanding the activities and roles of individual proteins and their differences, including normal/disease, human/animal, or in vitro/in vivo. Here, we describe the study protocols and precautions for QTAP experiments including in silico target peptide selection, determination of peptide concentration by amino acid analysis, setup of selected/multiple reaction monitoring (SRM/MRM) analysis in liquid chromatography–tandem mass spectrometry, preparation of protein samples (brain capillaries and plasma membrane fractions) followed by the preparation of peptide samples, simultaneous absolute quantification of target proteins by SRM/MRM analysis, data analysis, and troubleshooting. An application of QTAP in biological sciences was introduced that utilizes data from inter-strain differences in the protein expression levels of transporters, receptors, tight junction proteins and marker proteins at the blood–brain barrier in ddY, FVB, and C57BL/6J mice. Among 18 molecules, 13 (abcb1a/mdr1a/P-gp, abcc4/mrp4, abcg2/bcrp, slc2a1/glut1, slc7a5/lat1, slc16a1/mct1, slc22a8/oat3, insr, lrp1, tfr1, claudin-5, Na+/K+-ATPase, and γ-gtp) were detected in the isolated brain capillaries, and their protein expression levels were within a range of 0.637-101 fmol/μg protein. The largest difference in the levels between the three strains was 2.2-fold for 13 molecules, although bcrp and mct1 displayed statistically significant differences between C57BL/6J and the other strain(s). Highly sensitive simultaneous absolute quantification achieved by QTAP will increase the usefulness of proteomics in biological sciences and is expected to advance the new research field of pharmacoproteomics (PPx).
Proteomics by name is almost 20 years old and has rapidly grown into one of the most active research areas in biological sciences. Proteomics has had tremendous impacts on a variety of biological fields. Mass spectrometry (MS)-based protein identification is now widely adopted, and recent advances in MS and global proteomics (Figure 1), including protein sequence databases, have enabled the identification of hundreds to thousands of proteins in biological materials in a single analysis [1–3]. However, the proteome coverage attainable with available global proteomic approaches remains insufficient. Highly abundant proteins are easy to identify, but low-abundance proteins are difficult to detect due to high background noise when analyzing complex samples (Figure 1). Physiologically relevant molecules with low protein expression levels, such as transporters and receptors, are not readily identified by current global proteomic technologies. Improvements in fractionation, purification, and separation techniques in sample preparation and liquid chromatography (LC) and increased resolution and sensitivity of MS devices are required but remain challenging. Furthermore, the activities and roles of individual proteins must be elucidated, but this is hampered by the lack of quantitative information for protein expression in global proteomics.
Quantitative targeted absolute proteomics (QTAP) represents a new generation of proteomic methods that has permitted the determination of absolute protein expression levels (mol) of target proteins in biological materials by liquid chromatography-linked tandem mass spectrometry (LC-MS/MS) (Figure 1) . Selected/multiple reaction monitoring (SRM/MRM) in the MS/MS devices is an essential analytical mode in QTAP that allows the target peptides of target proteins to be distinguished in significantly complex samples and provides high selectivity and a high signal-to-noise ratio (Figure 1). Because triple quadrupole (QqQ) MS has excellent sensitivity and a wide dynamic range, SRM/MRM analysis with QqQ MS permits highly sensitive quantification of target proteins and is the most widely used method for QTAP (Figure 1). For these reasons, QTAP is useful to understand the activities of functional proteins, including low-abundance molecules.
The selection of target peptides for target proteins is a critical but rate-limiting step to achieve highly sensitive and reliable protein quantification in QTAP. To solve this problem, we have established a method to design appropriate target peptides in silico from sequence information in protein databases (Table 1) . Hence, we have succeeded in quickly developing LC-MS/MS quantification methods for several proteins. We have quantified more than 100 molecules, including transporters and receptors in the human blood–brain barrier (BBB), and have elucidated inter-species differences in protein expression levels between humans, the cynomolgus monkey, and the ddY mouse [4–6]. Furthermore, the quantitative protein expression profiles of many molecules have been applied to validate the use of the human BBB model cell line (hCMEC/D3) in vitro by comparison with in vivo human BBB cells .
Quantitative assays using antibodies, such as quantitative western blotting and ELISA, are widely used for protein quantification. However, these assays have significant disadvantages, including the lack of suitable specific antibodies for many proteins and the difficulty in obtaining these antibodies. By contrast, QTAP permits the development of appropriate LC-MS/MS-based absolute quantification methods for almost any target protein if sequence information is registered in the protein databases. The dynamic range of quantification in QTAP is significantly wider than that in antibody-based quantification. Furthermore, QTAP method development requires only 1 month, which is considerably more rapid than antibody development .
Another advantage of QTAP is the ability to reconstruct the in vivo activities of individual target molecules by integrating the molecular activities measured in vitro with in vitro/in vivo differences in protein expression levels . The major limitation of in vivo functional analysis using imaging technologies such as positron emission tomography (PET) and single photon emission computed tomography (SPECT) is the difficulty of accurately evaluating the specific activity of target proteins because the specificities for tracers are often similar among protein molecules, including functionally unknown proteins. QTAP can solve this problem based on the in vitro-to-in vivo reconstruction theory and is expected to advance the new research field of pharmacoproteomics (PPx).
The purpose of the present manuscript is to provide detailed protocols and precautions for QTAP experiments. To demonstrate the usefulness and limitations of QTAP, an application of QTAP in biological sciences is introduced that utilizes data from inter-strain differences in the protein expression levels of transporters, receptors, tight junction proteins, and marker proteins at the blood–brain barrier (BBB) in ddY, FVB, and C57BL/6J mice.
Methods and design
Workflow of QTAP
The QTAP experiment consists of 9 steps, and the basic workflow is outlined in Figure 2. Step 1 is the selection of the target proteins to be quantified. Global proteomics, protein chip technology, mRNA analysis using PCR or DNA chip technology, and other methodologies can be used to conduct global screening of the proteins that will be targeted in QTAP.
Step 2 is the selection of the target peptide sequences for the target proteins. The target peptides are selected in silico based on the peptide selection criteria (Table 1). The peptide should have the following features: unique amino acid sequence, efficient protease digestion, appropriate LC retention time, and good MS sensitivity. It is important to predict the MS sensitivity of the peptide before analysis because sensitivity can vary by 1 million-fold, depending on the amino acid sequences of the peptides .
Step 3 is the synthesis of a stable-isotope labeled peptide that will be used as an internal standard (IS) and an unlabeled 95% pure peptide that will be used to establish a calibration curve.
Step 4 is the accurate determination of the concentration of the peptide solution by quantitative amino acid analysis (AAA) and the optimization of the LC-MS/MS conditions, including SRM/MRM transitions, declustering potentials (DP), and collision energies (CE).
Step 5 is the construction of the calibration curve using a mixture of a dilution series of the unlabeled peptide and a fixed amount of the labeled peptide. The peptide mixture is injected onto the C18 column of the LC coupled with MS/MS to confirm the sensitivity and accuracy of the optimized SRM/MRM analysis and appropriate peptide separation on the column.
Step 6 is the preparation of the protein samples. QTAP is applicable to several types of protein samples, including those used in ELISA or immunoblotting. We have already applied QTAP for whole tissue lysates of human tissues; monkey and mouse brain capillaries; whole cell lysates of hCMEC/D3 and human breast cancer cell lines; the microsomal fraction of liver; crude membrane fractions of human breast and stomach cancer cell lines; plasma membrane fractions of liver, kidney, platelets, meningioma, hCMEC/D3 cells, and HUVECs; cytosolic fractions of human pancreatic adenocarcinoma cell lines; and the plasma of pancreatic cancer patients [4–7, 11–18]. The minimal sample requirement for QTAP experiments is 50 μg protein. The preparation procedures for brain capillaries and plasma membrane fractions are described in section “Preparation of protein samples”.
Step 7 is the preparation of the peptide samples. The protein samples are first denatured with 7 M guanidine hydrochloride or 8 M urea, followed by reduction and S-carbamoylmethylation. When guanidine hydrochloride is used, the alkylated proteins are precipitated with a mixture of methanol, chloroform, and water and dissolved in 6 M urea. When urea is used in the first denaturation step, the precipitation step is omitted. The alkylated proteins are digested with lysyl endopeptidase (LysC), followed by trypsin. Then, a fixed amount of stable-isotope-labeled IS peptides is added, and the sample is acidified. Desalting is performed if needed.
Steps 8 and 9 are the determination of the protein expression levels of the target proteins in the biological samples by LC-MS/MS. The procedures for peak recognition, which is one of the most important steps for accurate quantification, are described in section “Data analysis”.
Method setup for QTAP
Selection of an appropriate peptide from the peptide mixture produced by trypsin digestion of the target protein
The selection of the probe peptide for the target protein is essential to achieve highly reliable and sensitive protein quantification by SRM/MRM analysis. Global proteomic approaches have been applied for peptides obtained by trypsin digestion of biological samples that express high levels of the target protein in which peptides with a high signal intensity are selected for quantification. However, this strategy is not only time-consuming but also requires the preparation of samples that express high levels of the target protein; therefore, the number of proteins for which quantification methods can be established is quite limited. Furthermore, trypsin digestion efficiency, peptide specificity, post-translational modification (PTM), and polymorphisms should be considered for accurate quantification, but the small number of peptides identified by global proteomics does not necessarily allow this.
To solve these problems, we have established a theory to predict the appropriate peptides for SRM/MRM quantification based on previous proteomic data and experience (Table 1). This allowed us to design highly sensitive and highly accurate target peptides in silico from sequence information registered in protein databases including UniProtKB . We successfully quantified more than 100 transporter proteins in isolated human brain microvessels [6, 19]. We have previously established an LC-MS/MS quantification method for more than 500 proteins including human, monkey, and mouse transporters, enzymes, and receptors, and we have reported quantitative protein expression profiles in brain capillaries, livers, kidneys, platelets, plasma, meningioma, human BBB model cell lines (hCMEC/D3), human umbilical vein endothelial cell lines (HUVECs), human pancreatic adenocarcinoma cell lines, and human breast and stomach cancer cell lines [4–7, 11–18]. In contrast to global proteomics, in QTAP, it is important that the target peptides are completely digested by proteases to enable the estimation of the absolute expression levels of the target proteins. Therefore, sequences that may cause incomplete digestion, such as continuous sequences of arginine (R) or lysine (K) (RR, KK, RK, KR) and a proline (P) at the C-terminal side of R or K (RP, KP), and transmembrane regions should be avoided (Table 1). The selection and use of multiple peptides for a target protein is useful for increasing the credibility of the absolute quantification.
Some protein families have a high similarity of amino acid sequences between subtypes, which sometimes makes it difficult to select the specific peptide for each subtype, particularly for short proteins. LysC digestion (digestion at K) is sometimes useful to obtain specific peptides if they cannot be chosen from trypsin-digested peptides (digestion at R and K). Alternatively, a common peptide can be selected for the target and similar proteins, and a specific peptide can be selected for the similar protein. The absolute expression level of the target protein is obtained by subtracting that of the similar protein from the total level, which is determined using the common peptide.
Preparation of the peptide solution for the calibration curve and internal standard
For absolute quantification of target proteins, the peptide concentrations of the stock solutions of the non-labeled (standard, St) and stable-isotope-labeled (internal standard, IS) peptides synthesized with > 95% purity must be determined by amino acid analysis (AAA), an accurate quantification method for peptides and proteins. A portion of the stock solution is hydrolyzed for 24 hours at 110°C in 5.7 N HCl to digest the peptide into free amino acids. After the HCl is dried in a vacuum-centrifuge or with N2 gas, the resulting amino acids are dissolved in 0.02 N HCl and quantified by an amino acid analyzer (e.g., HPLC-UV system with post-column ninhydrin derivatization) to determine the peptide concentration of the stock solution. The accuracy of the concentration determined by AAA is one of the key factors for accurate absolute quantification of target proteins. Therefore, pipette handling of less than 100 μL volumes should be avoided throughout the experiment, and the experiment should be conducted in quadruplicate at least. After concentration determination, the peptide solution is stored at -80°C. Freeze-thaw cycles should be minimized.
Selection of a suitable configuration of mass spectrometer and liquid chromatograph
Table 2 compares the performance of the different types of MS devices available on the market. In tissues and cells, abundant proteins as well as low-abundance proteins play an important functional role. Membrane proteins, such as transporters and receptors, generally have low expression levels. To quantify as many functional proteins as possible, it is necessary to use a MS device with high sensitivity and reliable quantification. Triple quadrupole mass spectrometers (QqQ) enable highly sensitive (attomole) and reliable quantification with a wide dynamic range (six orders of magnitude) by employing the SRM/MRM mode. Recent advances in mass spectrometry technologies have improved the sensitivity and quantification reliability of LTQ, Orbitrap and TOF, so that new generations of these MS instruments can also quantify attomole levels of proteins in SRM/MRM mode. The short dwell time (time spent to acquire the specific SRM/MRM transition) and fast switching of transitions in QqQ permit the simultaneous quantification of hundreds of targets while maintaining acceptable sensitivity. Therefore, we have selected QqQ for QTAP.
Recent MS instrument development has improved the quantification performance of high-resolution MS devices. The latest Q-TOF (AB Sciex TripleTOF 5600) can achieve highly sensitive target quantification with a dynamic range of four to five orders of magnitude while maintaining high resolution, which is referred to as high-resolution SRM/MRM (HR-SRM/MRM) analysis. Therefore, this MS device significantly reduces the noise level observed in the traditional SRM/MRM mode and can quantify target peptides with small peaks that are masked by background noise in QqQ. Furthermore, the improved scan speed of the MS device permits multiple transition analysis for the quantification of several target peptides. Therefore, this MS device would be useful for the quantification of target proteins in highly complex protein samples, such as whole tissue lysates, that have high levels of background noise.
LC selection is also important for QTAP. Conventional HPLC and nanoLC separation prior to MS analysis are the most widely used separation methods. The advantages of conventional HPLC are that it is more robust and easier to use, even for beginners, and that larger amounts of sample can be analyzed compared to nanoLC. The disadvantage of conventional HPLC is that it is less sensitive than nanoLC separation, thereby requiring a small ID column such as a 1.0 mm ID C18 column with a flow rate of 50 μL/min. By contrast, because nanoLC separation is sensitive, the analysis can be performed with low amounts of sample. For example 1 μg trypsin-digested peptides can be analyzed in sample-limited situations; however, large amounts of sample cannot be injected due to the small column ID and flow path. However, the nanoLC system is less robust and difficult to use, which requires training and patience, leading to low reproducibility. Recently, the introduction of a nanoLC system coupled with a chip column, e.g., the Eksigent NanoLC-Ultra and Ekspert nanoLC400, has led to high robustness and ease-of-use for nanoLC-MS/MS. Chip columns can be exchanged in seconds and provide reproducible results day-to-day, column-to-column and lab-to-lab.
The extension of the upper pressure limit of HPLC instrumentation to 1300 bar (ultra-high pressure liquid chromatography, UHPLC) and the introduction of columns that are packed with porous sub-2-μm and superficially porous (fused-core or core-shell) particles have opened new frontiers in the resolution of target peaks and analysis speed. Small-diameter ESI electrodes, such as 25 and 50 μm ID, have been developed and can minimize post-column dispersion, leading to sharper peaks. Therefore, microflow UHPLC-MS/MS with a superficially porous column and a small-diameter ESI electrode achieves more sensitive and higher throughput quantification than conventional HPLC-MS/MS. Furthermore, the microflow UHPLC system is as robust as conventional HPLC. However, the narrow peaks that are produced by microflow UHPLC (peak width less than 10 seconds) require a fast duty cycle in the MS device, which is only available in the latest generations of MS devices. Certain MS analyzers with fast acquisition speed (e.g., QqQ or TOF) are more compatible with microflow UHPLC than others (e.g., ion trap, Orbitrap, or FT-ICR).
Setup of LC-MS/MS quantification system for SRM/MRM analysis
Here, we describe the principle of SRM/MRM analysis and how to optimize the analytical conditions for peptide quantification in the SRM/MRM mode of HPLC-QqQ, which is one of the most routinely used configurations of LC and MS for quantification.
The target peptide is quantified by SRM/MRM mode using QqQ to achieve highly selective and sensitive quantification (Figure 3). QqQ employs three chambers; the 1st Q (Q1) and 3rd Q (Q3) are mass filters that pass the peptide ion with the target mass. In the 2nd Q (Q2), the peptide ion is fragmented by collision with N2 gas. The use of two mass filters provides high selectivity and a high S/N ratio. The combination of Q1 and Q3 mass filters is called an SRM/MRM transition, which can be changed every few msec, and hundreds of SRM/MRM transitions can be simultaneously monitored in a single analysis.
In our approach, each target peptide is quantified by measuring four different SRM/MRM transitions, which consist of a parent ion (Q1) and four different daughter ions (Q3). This allows us to increase the selectivity for the target peptide by monitoring the chromatographic coelution of eight transitions of the target and internal standard peptides, thereby ensuring the reliable identification of signal peaks (Figure 4). Furthermore, this SRM/MRM analysis provides four quantitative values for a target peptide from four transition sets by using the four corresponding calibration curves (Figure 5), increasing the accuracy of the quantification. By comparing the four quantitative values, the overlap of the noise peaks with the target peptide peak can be determined, and when necessary, SRM/MRM transitions can be changed to appropriate transitions that are not affected by noise peaks. A total of 8 SRM/MRM transitions (four transitions for the target peptide and four corresponding transitions for the internal standard peptide) are required for the quantification of one protein. Therefore, 37 different proteins can be simultaneously quantified in a single analysis using the currently available maximum of 300 SRM/MRM transitions (multiplexed SRM/MRM analysis).
To achieve a highly sensitive SRM/MRM analysis for individual peptides, it is essential to select four highly sensitive transitions and to optimize the declustering potentials (DP) and collision energies (CE). These conditions are determined from MS/MS spectra that are obtained by the direct infusion with a syringe pump of a 0.1-1 μM peptide solution at a flow rate of 5 μL/min into the MS device. Typically, doubly charged precursor ions (singly or triply charged for some peptides) are selected (Q1). A total of four transitions per peptide (Q3-1, -2, -3 and -4) that correspond to high intensity daughter ions are selected. The DP and CE are optimized to maximize signal strength. For the internal standard peptides that are labeled with 13C and/or 15N, the 4 transitions corresponding to those of the standard peptides are selected with the same DP and CE as the standard peptides. A dilution series of standard peptide (blank, 1, 5, 10, 50, 100, 500 and 1000 fmol) with a fixed amount of internal standard peptide (500 fmol for AB Sciex API5000; 100 fmol for AB Sciex QTRAP5500) is injected onto the C18 column of the LC coupled with the QqQ to confirm the appropriate separation of the peptide by the column and the sensitivity and accuracy under the optimized analytical conditions of SRM/MRM analysis (Figures 4 and 5). Table 3 shows an example of the optimized conditions for SRM/MRM analysis using an API5000 and a QTRAP5500. Peptides can be simultaneously quantified under optimized conditions in SRM/MRM mode.
The other MS parameters, including those that are related to the ESI source, are common between peptides, and the optimized values for HPLC-API5000 or QTRAP5500 with a flow rate of 50 μL/min are listed in Table 4. Polarity switching between positive and negative ESI is useful to prevent decreased MS sensitivity caused by the accumulation of ions inside of the analyzer. Thus, a positive mode is followed by a negative mode for a short period after the elution of the target peptides (Table 4). The HPLC conditions are also listed in Table 4. The gradient profile for HPLC is set at approximately 1%B/min or less than 1%B/min to avoid significant ion suppression. The switching valve enables the removal of undesired substances, including salts. Therefore, the valve is essential to keep the MS analyzer clean and to avoid decreased sensitivity.
Preparation of protein samples
Preparation of brain capillary-rich fraction (whole tissue lysate)
Figure 6 shows a preparation procedure for the brain capillary-rich fraction. The details for this procedure have been described previously [6, 20, 21]. Here, we describe the precautions that should be followed for this experiment:
Homogenization: A Potter–Elvehjem homogenizer with a medium clearance should be used to avoid disrupting endothelial cells. Manual up-and-down strokes without rotation are essential to prevent the aggregation of brain capillaries. The number of strokes should be optimized in advance to achieve a high purity and recovery of brain capillary.
Centrifugation with dextran: the appropriate concentration of dextran to isolate brain capillaries with high purity and recovery varies depending on the animal, the state of the brain (fresh, frozen, different degrees of freshness) or dextran lot [5, 6]. Therefore, the concentration should be optimized in advance.
Different sizes of nylon meshes (210, 85, 45, 20 μm) are used to fractionate brain vessels. The vessels passing through a 85-μm mesh are brain capillaries. Nylon mesh or glass beads should be washed well after trapping brain capillaries to ensure blood cell removal.
The nylon mesh method is more appropriate than the glass bead column method for isolating capillaries from frozen brain.
After collecting capillaries from mesh or beads, the frequency of pipetting should be minimized to avoid the adsorption of capillaries to the pipette tip.
Freeze-thaw cycles of capillaries should be minimized.
The recovery of brain capillary is approximately 50–100 μg protein from whole capillary lysate/g brain. Because the minimal requirement for QTAP is 50 μg protein/sample, it is recommended to prepare brain capillaries from at least 3 g brain (10 mouse or 3 rat cerebrums) because a significant loss of capillaries could occur throughout the experiment if the starting amount of brain is small.
Preparation of plasma membrane fraction
One of the advantages of quantification at the protein level is that the expression levels of target proteins in certain subcellular compartments can be determined by subcellular fractionation. This is crucially different from quantification at the mRNA level. Recently, we demonstrated that the protein expression levels of transporters in the plasma membrane fraction do not correlate with their mRNA expression levels in 17 human liver biopsies , thereby highlighting the advantage of quantification at the protein level, particularly in the plasma membrane fraction, to understand the transport activities of membrane transporters. For QTAP analysis using the plasma membrane fraction, it is important to prepare the sample with highly reproducible purity and recovery. Figure 7 shows a plasma membrane fraction preparation procedure. The details for the procedure have been described previously [4, 7, 16]. Here, we describe the precautions that should be followed to achieve high reproducibility:
At least 0.5 g tissue or 5.0 × 107 cells are needed to stably obtain more than 100 μg protein from the plasma membrane fraction. To prepare the plasma membrane fraction of isolated brain capillaries, at least 50 g brain tissue is needed.
After nitrogen cavitation, cell disruption should be confirmed by microscopy. If disruption is insufficient, the samples should be homogenized or subjected to nitrogen cavitation again. The pressure that is used for nitrogen cavitation can also be increased.
The sample suspension should be gently stacked on top of a pre-established 38% sucrose solution to avoid disturbing the interface between the sample suspension and the 38% sucrose solution. After ultra-centrifugation, the turbid layer should be carefully recovered with a syringe.
Protease inhibitors should be used throughout the experiment. For phosphoproteomics, phosphatase inhibitors should be used in addition to protease inhibitors.
Freeze-thaw cycles of the plasma membrane fraction should be minimized.
Absolute quantification by LC-MS/MS
The materials that are used for QTAP are listed in Table 5.
The reagents that are used for QTAP are listed in Table 6 with compositions, storage temperature, and other information.
Preparation of peptide samples (proteins to peptides)
Peptide samples are prepared according to the procedure described in Table 7. First, proteins (50–100 μg) are solubilized, denatured with denaturing buffer, reduced by DTT, and S-carbamoylmethylated by IAA. Second, the alkylated proteins are precipitated with a mixture of methanol, chloroform and water. Third, the protein precipitates are dissolved in 6 M urea in 100 mM Tris–HCl (pH 8.5), diluted 5-fold with 100 mM Tris–HCl (pH 8.5), and digested with LysC followed by digestion with TPCK-treated trypsin.
The efficiency of enzyme digestion is one of the key points for the absolute quantification of target proteins. We have confirmed the efficient digestion of glut1 in mouse brain microvessels and human MDR1 in MDR1-overexpressing cells by comparing the absolute amounts of digested peptides with that determined by quantitative binding assay and immunoblotting . Furthermore, we have also confirmed that no bands greater than 20 kDa were observed by SDS-PAGE after trypsin digestion . These results suggest that enzyme digestion proceeds efficiently; however, the results do not necessarily indicate complete digestion for all molecules other than glut1 and MDR1. As shown in Figure 8, the digestion speed differs between molecules, suggesting that the time dependency of trypsin digestion should be examined before the absolute quantification of target proteins to determine the digestion efficiency. In case of inefficient digestion, it is necessary to change target peptides and/or optimize the digestion reaction condition. Figure 8 also shows a dramatic improvement of the digestion rate of monocarboxylate transporter 1 (Mct1/Slc16a1) and Na+/taurocholate co-transporting polypeptide (Ntcp/Slc10a1) by a combination of LysC, trypsin, and a trypsin enhancer (ProteaseMax) compared to single digestion with trypsin. Therefore, the use of LysC and ProteaseMax in addition to trypsin is useful to facilitate enzyme digestion.
After trypsin digestion, a fixed amount of IS peptides is added to the digested peptide samples and a dilution series of St peptides (blank 1, 5, 10, 50, 100, 500 and 1000 fmol as injected amounts) prepared from a 100 nM St peptide mixture; then, the digested peptide samples and dilution series are acidified and centrifuged (Table 7). The supernatants are analyzed under the optimized analytical condition (Tables 3 and 4). Each target peptide is measured by 4 different SRM/MRM transitions (Figure 4), and up to 37 proteins can be simultaneously quantified. If a scheduled SRM/MRM mode is applied, it is possible to simultaneously quantify more than 37 proteins. The time required for a single analysis is 2 hours using a traditional HPLC-MS/MS system because a slow gradient is necessary to avoid significant ion suppression, and thus an entire analysis, including the dilution series, target protein samples, and quality controls, would be finished within a few days.
The sample solution obtained by the trypsin digestion of biological samples includes not only peptides but also various substances, including salts from “dirty” samples, that can contaminate the MS device and decrease sensitivity. Therefore, the clean-up of samples prior to introduction to the MS device is crucial to maintain its performance. Valve switching after the C18 column and before the MS device is useful to automatically remove the salts and hydrophilic substances (Table 4). Furthermore, the clean-up of samples with desalting tips prior to LC-MS/MS injection is also useful to avoid sample clogging in the column and flow path, particularly in nanoLC, while keeping the MS device cleaner and concentrating peptide samples before injection. GL-Tip SDB (GL Sciences Inc., 200 μL, 7820–11200) is superior to conventional C18 desalting tips for the retention of more hydrophilic peptides; in addition, this device eliminates the loss in the desalting step of low hydrophobic peptides that are retained on the C18 HPLC analytical column and not retained in C18 desalting tip. The GL-Tip GC (GL Sciences Inc., 200 μL, 7820–11201) can cover a wider hydrophilic range of peptides than the GL-Tip SDB. Therefore, the combined use of the GL-Tip SDB and the GL-Tip GC is beneficial for covering a wide range of hydrophilic and hydrophobic peptides.
Data analysis includes peak recognition, calibration curve preparation, and protein expression level and limit of quantification calculations. Here, we describe the precautions that should be followed for data analysis.
A positive peak for the target peptide is defined as one that is detected at the same retention time (± 0.10 min) as the IS peptide under the HPLC conditions that are listed in Table 4. Because a target peptide and its corresponding IS peptide are monitored by 4 transitions, the peak recognition is confirmed by the chromatographic coelution of the 8 transitions (Figure 4). Figure 9 illustrates an example of an ambiguous chromatogram. The retention time of the peak in Figure 9C is more than 0.10 min different from that of the IS peptide in Figure 9G. Therefore, the peak is not derived from the target peptide and should not be recognized. Furthermore, the peak area ratios of the target and IS peptides are theoretically identical between the four transition sets (Figure 4). As shown in transition sets 3 and 4 (Figure 9C, D, G and H), if certain transition sets display larger peak area ratios than other transition sets, it is possible that the recognized peaks include noise peaks. We have previously established that the coefficients of variation for protein expression levels of various proteins were within 20% when determined from three transition sets . Therefore, in principle, we consider that coefficients of variation of over 20% among the four transition sets are likely to indicate the presence of noise peaks, and the corresponding transitions should not be used for accurate quantification. To solve these problems, it is beneficial to use a high-resolution MS device, such as a TripleTOF5600 (Table 2), or to change the target peptide.
Preparation of the calibration curve
Calibration curves are prepared with every transition by plotting the peak area ratios of the St and IS peptides (y-values) against the St peptide amounts (x-values) (Figure 5). The correlation coefficient (R2) of the regression line should be greater than 0.99 for accurate quantification. The slope of the line should be [1/IS peptide amount]. Otherwise, the peptide concentration determined by AAA may be incorrect, and/or the peptides may have adsorbed to the tubes and pipette tips during the preparation of the dilution series. The precision of the St peptide concentration is more critical than that of the IS peptide because an incorrect concentration of the St peptide will result in the incorrect calculation of the protein expression levels of the target proteins.
Calculation of protein expression level
Kamiie et al.  have validated that protein expression levels yield coefficients of variation of less than 20.0% when determined from three peaks with area counts greater than 5000. Therefore, signal peaks with an area count greater than 5000 that are detected at the same retention time (± 0.10 min) as an IS peptide are defined as positive. When positive peaks are observed in three or four sets of transitions, the proteins are considered to be expressed in the target protein samples. The absolute amount (fmol) of each target peptide is determined as the average of three or four quantitative values that are calculated from the target-to-IS peptide peak area ratios in the target samples and the calibration curve. The protein expression level (fmol/μg protein) of the target protein is obtained by dividing the determined absolute amount (fmol) of the target peptide by the total protein amount (μg protein) of the analyzed samples. Biological samples, such as tissues and cells, suffer from significant ion suppression compared to authentic samples, not including the matrix, resulting in a 2- to 10-fold decrease in the peak area of peptides in the biological samples compared with the authentic samples. The degree of ion suppression also differs between samples. Therefore, sensitivity correction using IS peptides is important for the accurate quantification of target proteins.
Calculation of the limit of quantification (LQ)
The LQ of non-detected molecules in protein samples is defined as the protein concentration (fmol/μg protein) that yields a peak area count of 5000 in the chro-matogram of the target protein samples. When the calibration curve is obtained using Eq. 1, the amount (fmol) of target protein that is equivalent to a peak area count of 5000 (ATarget eq 5000) is calculated using Eq. 2 from the peak area (counts) of the IS peptide in the target protein samples (PAISin sample) and the values of Slope and Intercept in Eq. 1. Then, the LQ is obtained with Eq. 3 by dividing ATarget eq 5000 by the total protein amount (μg protein) of the target protein samples analyzed (ASample).
where PASt in Authentic and PAIS in Authentic are the peak areas (counts) of the St peptide and IS peptide in the authentic samples, respectively, and ASt in Authentic is the amount (fmol) of the St peptide in the authentic samples. In silico peptide selection criteria enable the selection of highly sensitive peptides for any proteins. As a result, the LQs have been reported to be less than 1 fmol/μg protein for 99% of the transporters and receptors in the analysis of human brain microvessels .
Software for data analysis
The simultaneous quantification of many proteins results in a significant amount of SRM/MRM data. For example, the simultaneous quantification of 37 proteins uses 296 SRM/MRM transitions per analysis (37 × 8 (4 transitions for the target peptide and 4 transitions for the IS peptide)), and the SRM/MRM analysis of 20 samples including eight standards (blank, 1, 5, 10, 50, 100, 500 and 1000 fmol) and 12 biological samples results in 5920 SRM/MRM chromatograms (296 × 20). Therefore, the rapid recognition of target peaks and the rapid calculation of protein expression levels are essential. Conventional software for data processing includes an automatic peak recognition function that is based on the retention time of a signal peak in certain sample analyses. However, the different retention times of signal peaks for peptides between samples make it difficult to properly recognize the target peaks. As a result, the target peaks must be manually recognized in most chromatograms, which is labor-intensive and requires approximately one week for data analysis. To solve this problem, we have developed automatic analysis software that is specialized for simultaneous absolute quantification when stable-isotope labeled peptides are used as internal standards. This software identifies the target peaks based on the retention time of the signal peaks of the IS peptides in each analysis. Therefore, the software can overcome the problem of different retention times of target peaks between samples and can properly automatically recognize target peaks in most chromatograms, accelerating data processing compared to conventional software.
Typical problems and troubleshooting in QTAP experiments are listed in Table 8. Proteins and peptides, particularly hydrophobic ones, are likely to adsorb to the upper walls of the tubes and pipette tips and are partially lost during sample preparation, resulting in an underestimation of the protein expression levels of target proteins. Immature technique also causes irreproducible quantification results because the degrees of loss are different between samples. To improve the technique, it is important to carefully handle sample solutions to minimize the adsorption of proteins and peptides to tube walls and tips and to practice several times to smoothly execute the experiment. An artificial protein known as a “monitoring protein” is a useful tool for evaluating the proficiency of experimenters; a fixed amount of the artificial protein (A fmol) is added into every protein sample before sample preparation, the digested peptides of the artificial protein are quantified by LC-MS/MS after sample preparation (B fmol), and the recovery rate is calculated for every sample preparation procedure by determining the ratio of B/A.
Example: Comparison of protein expression levels of transporters, receptors, claudin-5, and marker proteins in brain capillaries in different mouse strains: ddY, FVB, and C57BL/6J
The availability of mouse models has facilitated significant progress in the study of CNS-related diseases because genetic engineering technologies, such as gene knockout and transgenic mice, enable the elucidation of the role of specific genes in these diseases. Several types of mouse strains are widely used in these studies, and distinct differences between the strains have been reported in behavioral and neural parameters . However, thus far, there is no information regarding inter-strain differences in BBB permeability. To clarify these differences, QTAP can be used to analyze functional proteins, such as transporters, receptors, and tight junction proteins, that are involved in BBB transport. We have previously reported the absolute expression levels of transporters, receptors, claudin-5, and marker proteins in isolated brain capillaries of ddY and FVB mice [4, 6, 23]. However, in the present study, these were quantified again, together with the corresponding levels in C57BL/6J mice, in order to eliminate any experimental bias due to differences in experimental day and experimenter.
Table 9 shows the direct comparison of the protein expression levels in the brain capillaries of ddY, FVB, and C57BL/6J mice. A total of 13 molecules, including 7 transporters, 3 receptors, 1 tight junction protein, and 2 marker proteins, were detected among 18 molecules in all 3 mouse strains, and the differences in the protein expression levels between the 3 strains were within a 2.2-fold range for the 13 molecules. This result suggests that inter-strain differences in BBB permeability in mice are small, unlike behavioral and neural parameters.
The expression levels of breast cancer resistance protein (Bcrp/Abcg2) and monocarboxylate transporter 1 (Mct1/Slc16a1) were statistically significantly different between C57BL/6J mice and the other strain(s). The expression level of Bcrp/Abcg2 in C57BL/6J mice was significantly greater than that in ddY and FVB mice, whereas the Mct1 expression level in C57BL/6J mice was significantly less than that in FVB mice. At the BBB, Bcrp forms a functional barrier against drug entry into the brain by pumping drugs out of brain capillary endothelial cells. The higher level of Bcrp at the BBB of C57BL/6J mice thus implies reduced BBB drug permeability. At the BBB, Mct1 contributes to the supply of ketone bodies as an alternative source of energy to the brain, and thus the lower expression of Mct1 in C57BL/6J mice suggests a difference in brain energy metabolism compared to FVB mice.
Most of the data in ddY and FVB mice were quantified again in the present study to ensure comparability. These data were in good agreement the originally reported data [4, 6, 23], within ±30% in almost all cases, demonstrating the good day-to-day and experimenter-to-experimenter reproducibility of QTAP.
Possible application and perspective
Clarifying the physiological role of the BBB and the regulation of its function are crucial for the diagnosis and prevention of CNS disease and the development of new CNS-targeted drugs. Because proteins play pivotal roles in cellular function and are the minimum unit of cellular function, it is important to clarify the absolute protein expression level in cells, the mechanism of regulation, and the modification of protein function. In this study protocol, we have introduced a detailed procedure for establishing the method, preparing the sample, and quantifying protein expression by LC-MS/MS for highly sensitive, selective, and simultaneous protein quantification. This methodology is applicable for BBB research as well as any research involving proteins. QTAP-based research is novel and will enable us to clarify several important subjects. Table 10 summarizes what were achieved by QTAP. Tables 11 and 12 summarize the potential applications of QTAP for in vitro and in vivo studies, respectively. When the assay sensitivity of MS/MS increases 10-fold compared to the current level of sensitivity and reaches 1 attomole peptide per analysis, the progress of QTAP-based research will be significantly enhanced. The quantitative analysis of modified proteins is also one of the most important subjects to clarify the mechanism of signal transduction and to identify a solution to regulate the associated mechanisms. In the future, QTAP-based studies will revolutionize the progress of BBB research.
YU: Assistant professor in Graduate School of Pharmaceutical Sciences, Tohoku University, Japan.
MT: Associate professor in Graduate School of Pharmaceutical Sciences, Tohoku University, Japan.
WO, YH, and YT: Graduate students in Graduate School of Pharmaceutical Sciences, Tohoku University, Japan.
SO: Full professor in Faculty of Life Sciences, Kumamoto University, Japan.
TT: Full professor in Graduate School of Pharmaceutical Sciences, Tohoku University, Japan.
Amino acid analysis
Central nervous system
Liquid chromatography–tandem mass spectrometry
Limit of quantification
Lysyl endopeptidase C
Triple quadrupole mass spectrometer
Quantitative targeted absolute proteomics
Selected/multiple reaction monitoring
Ultra-high pressure liquid chromatography
Under limit of quantification.
Chen P, Li X, Sun Y, Liu Z, Cao R, He Q, Wang M, Xiong J, Xie J, Wang X, Liang S: Proteomic analysis of rat hippocampal plasma membrane: characterization of potential neuronal-specific plasma membrane proteins. J Neurochem. 2006, 98: 1126-1140. 10.1111/j.1471-4159.2006.03934.x.
Nagano K, Taoka M, Yamauchi Y, Itagaki C, Shinkawa T, Nunomura K, Okamura N, Takahashi N, Izumi T, Isobe T: Large-scale identification of proteins expressed in mouse embryonic stem cells. Proteomics. 2005, 5: 1346-1361. 10.1002/pmic.200400990.
Pshezhetsky AV, Fedjaev M, Ashmarina L, Mazur A, Budman L, Sinnett D, Labuda D, Beaulieu JF, Menard D, Nifant’ev I, Levy E: Subcellular proteomics of cell differentiation: quantitative analysis of the plasma membrane proteome of Caco-2 cells. Proteomics. 2007, 7: 2201-2215. 10.1002/pmic.200600956.
Kamiie J, Ohtsuki S, Iwase R, Ohmine K, Katsukura Y, Yanai K, Sekine Y, Uchida Y, Ito S, Terasaki T: Quantitative atlas of membrane transporter proteins: development and application of a highly sensitive simultaneous LC/MS/MS method combined with novel in-silico peptide selection criteria. Pharm Res. 2008, 25: 1469-1483. 10.1007/s11095-008-9532-4.
Ito K, Uchida Y, Ohtsuki S, Aizawa S, Kawakami H, Katsukura Y, Kamiie J, Terasaki T: Quantitative membrane protein expression at the blood–brain barrier of adult and younger cynomolgus monkeys. J Pharm Sci. 2011, 100: 3939-3950. 10.1002/jps.22487.
Uchida Y, Ohtsuki S, Katsukura Y, Ikeda C, Suzuki T, Kamiie J, Terasaki T: Quantitative targeted absolute proteomics of human blood–brain barrier transporters and receptors. J Neurochem. 2011, 117: 333-345. 10.1111/j.1471-4159.2011.07208.x.
Ohtsuki S, Ikeda C, Uchida Y, Sakamoto Y, Miller F, Glacial F, Decleves X, Scherrmann JM, Couraud PO, Kubo Y, Tachikawa M, Terasaki T: Quantitative targeted absolute proteomic analysis of transporters, receptors and junction proteins for validation of human cerebral microvascular endothelial cell line hCMEC/D3 as a human blood–brain barrier model. Mol Pharm. 2013, 10: 289-296. 10.1021/mp3004308.
Ohtsuki S, Uchida Y, Kubo Y, Terasaki T: Quantitative targeted absolute proteomics-based ADME research as a new path to drug discovery and development: methodology, advantages, strategy, and prospects. J Pharm Sci. 2011, 100: 3547-3559. 10.1002/jps.22612.
Uchida Y, Ohtsuki S, Kamiie J, Terasaki T: Blood–brain barrier (BBB) pharmacoproteomics: reconstruction of in vivo brain distribution of 11 P-glycoprotein substrates based on the BBB transporter protein concentration, in vitro intrinsic transport activity, and unbound fraction in plasma and brain in mice. J Pharmacol Exp Ther. 2011, 339: 579-588. 10.1124/jpet.111.184200.
Anderson L, Hunter CL: Quantitative mass spectrometric multiple reaction monitoring assays for major plasma proteins. Mol Cell Proteomics. 2006, 5: 573-588.
Kawakami H, Ohtsuki S, Kamiie J, Suzuki T, Abe T, Terasaki T: Simultaneous absolute quantification of 11 cytochrome P450 isoforms in human liver microsomes by liquid chromatography tandem mass spectrometry with In silico target peptide selection. J Pharm Sci. 2011, 100: 341-352. 10.1002/jps.22255.
Niessen J, Jedlitschky G, Grube M, Bien S, Schwertz H, Ohtsuki S, Kawakami H, Kamiie J, Oswald S, Starke K, Strobel U, Siegmund W, Rosskopf D, Greinacher A, Terasaki T, Kroemer HK: Human platelets express organic anion-transporting peptide 2B1, an uptake transporter for atorvastatin. Drug Metab Dispos. 2009, 37: 1129-1137. 10.1124/dmd.108.024570.
Niessen J, Jedlitschky G, Grube M, Kawakami H, Kamiie J, Ohtsuki S, Schwertz H, Bien S, Starke K, Ritter C, Strobel U, Greinacher A, Terasaki T, Kroemer HK: Expression of ABC-type transport proteins in human platelets. Pharmacogenet Genomics. 2010, 20: 396-400. 10.1097/FPC.0b013e32833997b0.
Obuchi W, Ohtsuki S, Uchida Y, Ohmine K, Yamori T, Terasaki T: Identification of transporters associated with Etoposide sensitivity of stomach cancer cell lines and methotrexate sensitivity of breast cancer cell lines by quantitative targeted absolute proteomics. Mol Pharmacol. 2013, 83: 490-500. 10.1124/mol.112.081083.
Ohmine K, Kawaguchi K, Ohtsuki S, Motoi F, Egawa S, Unno M, Terasaki T: Attenuation of phosphorylation by deoxycytidine kinase is key to acquired gemcitabine resistance in a pancreatic cancer cell line: targeted proteomic and metabolomic analyses in PK9 cells. Pharm Res. 2012, 29: 2006-2016. 10.1007/s11095-012-0728-2.
Ohtsuki S, Schaefer O, Kawakami H, Inoue T, Liehner S, Saito A, Ishiguro N, Kishimoto W, Ludwig-Schwellinger E, Ebner T, Terasaki T: Simultaneous absolute protein quantification of transporters, cytochromes P450, and UDP-glucuronosyltransferases as a novel approach for the characterization of individual human liver: comparison with mRNA levels and activities. Drug Metab Dispos. 2012, 40: 83-92. 10.1124/dmd.111.042259.
Yoneyama T, Ohtsuki S, Ono M, Ohmine K, Uchida Y, Yamada T, Tachikawa M, Terasaki T: Quantitative targeted absolute proteomics-based large-scale quantification of proline-hydroxylated alpha-fibrinogen in plasma for pancreatic cancer diagnosis. J Proteome Res. 2013, 12: 753-762. 10.1021/pr3008144.
Yoshikawa A, Nakada M, Ohtsuki S, Hayashi Y, Obuchi W, Sato Y, Ikeda C, Watanabe T, Kawahara Y, Hasegawa T, Sabit H, Kita D, Nakanuma Y, Terasaki T, Hamada JI: Recurrent anaplastic meningioma treated by sunitinib based on results from quantitative proteomics. Neuropathol Appl Neurobiol. 2012, 38: 105-110. 10.1111/j.1365-2990.2011.01197.x.
Shawahna R, Uchida Y, Decleves X, Ohtsuki S, Yousif S, Dauchy S, Jacob A, Chassoux F, Daumas-Duport C, Couraud PO, Terasaki T, Scherrmann JM: Transcriptomic and quantitative proteomic analysis of transporters and drug metabolizing enzymes in freshly isolated human brain microvessels. Mol Pharm. 2011, 8: 1332-1341. 10.1021/mp200129p.
Ohtsuki S, Yamaguchi H, Asashima T, Terasaki T: Establishing a method to isolate rat brain capillary endothelial cells by magnetic cell sorting and dominant mRNA expression of multidrug resistance-associated protein 1 and 4 in highly purified rat brain capillary endothelial cells. Pharm Res. 2007, 24: 688-694. 10.1007/s11095-006-9188-x.
Pardridge WM, Yang J, Eisenberg J, Tourtellotte WW: Isolation of intact capillaries and capillary plasma membranes from frozen human brain. J Neurosci Res. 1987, 18: 352-357. 10.1002/jnr.490180213.
Ingram DK, Jucker M: Developing mouse models of aging: a consideration of strain differences in age-related behavioral and neural parameters. Neurobiol Aging. 1999, 20: 137-145. 10.1016/S0197-4580(99)00033-0.
Agarwal S, Uchida Y, Mittapalli RK, Sane R, Terasaki T, Elmquist WF: Quantitative proteomics of transporter expression in brain capillary endothelial cells isolated from P-glycoprotein (P-gp), breast cancer resistance protein (Bcrp), and P-gp/Bcrp knockout mice. Drug Metab Dispos. 2012, 40: 1164-1169. 10.1124/dmd.112.044719.
We thank Ms. N. Handa and A. Niitomi for secretarial assistance. The studies were supported in part by four Grants-in-Aid for Scientific Research (S) [KAKENHI: 18109002], Scientific Research (A) [KAKENHI: 24249011], Young Scientists (B) [KAKENHI: 23790170], and the Japan Society for the Promotion of Science (JSPS) Fellows [KAKENHI: 207291] from the JSPS. The studies were also supported in part by a Grant-in-Aid for Scientific Research on Priority Area [KAKENHI: 17081002] from The Ministry of Education, Culture, Sports, Science and Technology (MEXT), 3 Grants for the Development of Creative Technology Seeds Supporting Program for Creating University Ventures, the Creation of Strategic Innovation Project, and the Revitalization Promotion Program (A-STEP) from the Japan Science and Technology Agency (JST). Furthermore, the studies were also supported in part by an Industrial Technology Research Grant Program from New Energy and the Industrial Technology Development Organization (NEDO) of Japan, a Health and Labour Sciences Research Grant from The Ministry of Health, Labour and Welfare, the Funding Program for Next Generation World-Leading Researchers by the Cabinet Office, Government of Japan, and Proteomedix Frontiers Co. Ltd.
Tetsuya Terasaki and Sumio Ohtsuki are full professors of Tohoku University, and Kumamoto University, respectively, and are also directors of Proteomedix Frontiers Co. Ltd. Their positions at Proteomedix Frontiers do not affect the design of the study, collection of the data, analysis or interpretation of the data, decision to submit the manuscript for publication, or writing of the manuscript. The other authors declare no competing interest.
Conception: YU, MT, SO, and TT. Design: YU, MT, YH, YT, SO, and TT. Data acquisition: YU, YH, and YT. Analysis: YU, YH, and YT. Data interpretation: All authors. Wrote manuscript: YU, MT, WO, YH, YT, and TT. Reviewed manuscript: YU, MT, SO, and TT. Approval of final version of manuscript: All authors.
Authors’ original submitted files for images
Below are the links to the authors’ original submitted files for images.
About this article
Cite this article
Uchida, Y., Tachikawa, M., Obuchi, W. et al. A study protocol for quantitative targeted absolute proteomics (QTAP) by LC-MS/MS: application for inter-strain differences in protein expression levels of transporters, receptors, claudin-5, and marker proteins at the blood–brain barrier in ddY, FVB, and C57BL/6J mice. Fluids Barriers CNS 10, 21 (2013). https://doi.org/10.1186/2045-8118-10-21
- Quantitative targeted absolute proteomics (QTAP)
- Pharmacoproteomics (PPx)
- Absolute expression level
- In silico peptide selection criteria
- Blood–brain barrier
- Strain difference
- Tight junction protein