3-2 Phosphoproteomic Analysis of Human Jurkat T Cell

Chapter 3 – Results and Discussion

3-1 Immobilized Metal Affinity Chromatography

3-1.1 Phosphopeptides Purification with MALDI–TOF MS Analysis

To identify the protein phosphorylation, the proteins were digested with trypsin, followed by peptide mass mapping analysis using MALDI–TOF MS. The PMF served to identify protein by sequence database searches, and the candidate phosphopeptides would be identified by inspecting the peptide mass spectra for peptide ion signals which match tryptic peptides as predicted from retrieved protein sequence by taking account that addition of phosphoryl groups giving rise to mass increments of 80 Da.²⁵ Although this method would be fast and easy to identify phosphoprotein, the ion signal of phosphopeptides are suppressed with existence of unphosphylated peptide during the ionization process. To remove the unphosphorylated peptide, the Fe (III) IMAC were used to selectively purify the of phosphopeptides prior to the MALDI–TOF MS analysis, the analytical scheme is shown in Figure 1.

To optimize the selectivity and recovery of Fe (III) IMAC for the phosphopeptides the tryptic peptide mixture of β-casein was used. The MALDI–TOF MS analysis of tryptic peptide mixture of β-casein was shown in Figure 2 (a). The phosphopeptide of β-casein observed in the Figure 2(a), corresponded to the monophosphorylated peptide

of m/z 2062, elicited with low signal responses. Other phosphopeptides were too weak to be detected by MALDI–TOF MS. This is commonly referred to as the "suppression effect". With the use of IMAC purification, the nonphosphorylated peptides were removed, and the purified phosphopeptides were analyzed by MALDI –TOF MS as shown in Figure 2 (b). Phosphopeptides of β-casein by Fe (III) IMAC purification were significantly enriched and the base peak at m/z 2062 corresponded to the phosphopeptide of FQpSEEQQQTEDELQDK. Another phosphopeptide observed in the MALDI-TOF MS Figure 2 (b) with the m/z 3122 correspond to the tetraphosphorylated peptide ELEELNVPGEIVEpSLpSpSpSEESITR.

3-1.2 Optimization of IMAC Procedures

3-1.2.1 Elution Volume

For the IMAC purification, it is important to optimize collection time for eluted phosphopeptides. If collection too long, the sample volume may significantly increase and thus dilute the phosphopeptide. However, if collected time too short, the sample recovery may reduce. In order to optimize the collected time, the tryptic peptides of β-casein was utilized for the determination of elution time. The IMAC eluent was collected every minute and the eluted phosphopeptide was quantified by the LC-MS.

The extracted ion chromatograms (XIC) which were ion chromatogram created by taking intensity values at a signal, discrete mass value or mass range from a series of mass spectral scans. Consequently, the XIC can only be based on a single experiment and must derived from a spectrum as long as a mass range can be defined by means of a selection. The XIC of m/z 1031.42 from double charge of β-casein phosphopeptide

FQpSEEQQQTEDELQDK were used to determine the profile of elution, and spectral data for each chromatographic peak can be integrated peak area to calculate the quantities of the phosphopeptide. In Figure 3, there was the elution profile from IMAC purification β-casein phosphopeptide. Apparently, most phosphopeptides would be eluted within 5 minutes. Therefore, the collected time would be settled on 5 minutes in this study.

3-1.2.2 Elution Solution

The IMAC bound phosphopeptides were generally eluted from IMAC columns using Na2HPO4 at pH=9. However, there was a serious problem with sodium ions.

The sodium adduct formation is a major limitation of ESI-MS, which would reduce the sensitivity of sample. As shown in Figure 4 (a), sodium adduction was observed with the use of Na2HPO4. In order to eliminate the formation of sodium adduct , the Na2HPO4 was replaced by NH4H2PO4. As shown Figure 4 (b), sodium adduction to the phosphopeptides was completely eliminated by using NH4H2PO4.

3-1.2.3 Reusability

There was another important issue which is important for the large scale analysis of phosphopeptide. The reusability of IMAC column was tested by tryptic peptide mixture of β-casein. In Figure 5, there was intensity of β-casein phosphopeptide FQpSEEQQQTEDELQDK from MALDI-TOF-MS spectra. The IMAC reused efficiency was determinated over 10 times. The mean is 1.06E4, and the standard

deviation is 9.65E2. The major problem which made IMAC column ineffective was the particles blocked in IMAC column cacused the pressure increasing. Therefore, the sample would be filtered before IMAC purification.

3-1.2.4 Sensitivity

Sensitivity of IMAC column was determined by β-casein tryptic peptides. The packed IMAC column would be loaded 0.1 µg β-casein tryptic peptides. In Figure 6, there was β-casein phosphopeptide FQpSEEQQQTEDELQDK from MALDI-TOF MS spectra which still revealed the selectivity of IMAC. The sensitivity of IMAC column was determinated below 0.1 µg.

3-1.2.5 Column Capacity

The IMAC column capacity for phosphopeptides is importance for estimation of real sample such as phosphopeptides of cells. The IMAC column was 500 µm id PEEK tubing which was constructed by a 5 cm packed with Fe (III)-NTA resins. Each equilibrated Fe (Ⅲ) IMAC column was loaded with β-casein tryptic peptides, and the flow through of Fe (Ⅲ) IMAC column was collected and monitored by MALDI-TOF MS. When loaded sample had not reached the saturation of Fe (Ⅲ) IMAC column , phosphopeptides will be adsorbed by Fe (Ⅲ) IMAC column. When loaded sample had reached the saturation of Fe (Ⅲ) IMAC column, phosphopeptides will be eluted by Fe (Ⅲ) IMAC column and detected by MALDI-TOF MS. In Figure 7, there were four spectra represented 5, 10, 20, and 30 µg β-casein of IMAC flow through. The

phosphopeptides were not detected in 5 and 10 µg β-casein of IMAC flow through, indicating that they remained bound to Fe (Ⅲ) IMAC column, and meant that this condition had not reached the saturation of Fe (Ⅲ) IMAC column. In the spectrum of 20 µg β-casein of IMAC flow through, it seemed to be detected no signal of phosphopeptides, but detailed view of the region m/z 2000 ~2100 was found a weak peak of the β-casein phosphopeptide of m/z 2062. In addition, the spectrum of 30 µg β-casein of IMAC flow through was detected apparent signal of this phosphopeptide.

Therefore, this IMAC column capacity for phosphopeptides could be determined about 800 pmole which translated form 20 µg β-casein.

3-1.2.6 Sample Recovery

However, there are two issues about use of IMAC. One is that not all phosphopeptides bind to IMAC resins, and another is that the fraction eluted from IMAC is allegedly strongly contaminated with nonphosphorylated peptides. Despite these issues, no study clearly demonstrated what percentage of phosphopeptides is recovered from complex mixture by IMAC . In order to determinate the sample recovery of Fe ( Ⅲ ) IMAC column, the the purified phosphopeptide FQpSEEQQQTEDELQDK was used. In Figure 8, there were two extracted ion chromatograms (XIC). These two XIC of m/z 1031.42 from double charged of phosphopeptide were used to determine the sample recovery of Fe (Ⅲ) IMAC column, and spectral for each chromatographic peak can be integrated peak area to calculate the quantities of the phosphopeptide represented the spectra before and after IMAC experiment, respectively. The peak areas of these two spectra were 1.47E5 and 1.07E5, and sample recovery of the Fe (Ⅲ) IMAC column was 72.8 %.

3-1.2.7 Specificity

Specificity of IMAC is a key element for analysis of low abundant protein phosphorylation in the complex protein mixtures. Although the selectivity of Fe (III) IMAC combined with the specificity of trypsin cleavage and the peptide mass determination by MS enabled facile assignment of the phosphopeptides, the signals from the nonspecific binding of high abundant nonphosphorylated peptides presents in the sample may still interfere the sensitivity of phosphopeptide. However, IMAC may be effective for simple phosphoproteins like β-casein, but the selectivity of IMAC may reduce for the complex protein samples. In order to test the selectivity of IMAC, the β-casein phosphopeptide and BSA with 1:200 (w/w) were used as a test sample. This sample was loaded into Fe (Ⅲ) IMAC column, and the eluent was analyzed by MALDI-TOF MS as shown in Figure 9. The phosphopeptides could not be detected in the total digested samples due to the existence of abundance nonphosphorylated peptides. Analysis of the sample eluted from the Fe (Ⅲ) IMAC column with 200 mM NH4H2PO4 showed that the phosphopeptide can be purified from the nonphosphorylated peptide compeletely, with no contamination with nonphosphorylated peptides identified. In order to test the IMAC procedures under more rigorous condition, the β-casein phosphopeptide and BSA with 1:1000 (w/w) were used as a test sample. The phosphopeptide was detected by MALDI-TOF MS, but there were occurred nonspecific binding as show in Figure 10, and demonstrated the potential of Fe (Ⅲ) IMAC for selective enrichment of phosphopeptides when using more complex samples.

In order to test IMAC purification for the complex peptide mixture, the β-casein

was mixed with cell proteins. In Figure 11, there were two lanes on SDS-PAGE. The selected region was β-casein. The complex sample which is mixed with cell proteins was in the righthand which followed by using in-gel digestion. The phosphopeptides could still be detected successfully after IMAC purification as shown in Figure 12 (b).

3-1.3 Methylation of phosphopeptides

The major disadvantage of IMAC is nonspecific binding of negative charged amino acid residues, like glutamic and aspartic acid. In order to avoid nonspecific binding of IMAC by carboxyl groups in the peptide digest, a methyl esterification procedure based on the study by Ficarro et al. was used.²⁸ The tryptic peptides of β-casein were methyl-esterified with methanolic-HCl at room temperature for 30 minutes. In Figure 12, the spectrum of MALDI–TOF MS was the phosphopeptide with heptamethyl ester after IMAC enrichment. In Table 1, the spectral data for each chromatographic peak from XIC as shown in Figure 13 can be integrated peak area to calculate the quantities of the phosphopeptide represented whether methylation experiment after IMAC enrichment as determined by searching with (C-terminal) and (Asp/Glu) methyl ester as variable modifications. It was apparently observed that there were great differences to integrated area of phosphopeptide peak from the same quantity of phosphopeptide in methylation or not. This result had confirmed the limitation of recovery whether the reason is inefficient binding of the esterified phosphopeptides on the IMAC column or intrinsic loss of this methylation reaction.

Methylation of peptides was reported to enhance selective enrichment of phosphopeptides by IMAC in some systems, but not necessarily in others.^32-35 Sample recovery remains a potential limitation of this method due to inefficient binding of the

esterified phosphopeptides on the IMAC column. Another potential limitation is the difficulty in identifying the exact site of phosphorylation by tandem MS due to unexpected changes in peptide masses and further increasing complexity of the analyzed samples. Therefore, this method had needed to confirm further.

3-1.4 Identification of Phosphorylation sites of β-Caseins Using nLC-MS/MS

Coupling of Fe (III) IMAC with nLC-MS/MS permitted good performance in detection and analysis of phosphorylated peptides.³⁷ Separation of tryptic peptides using nLC is an excellent approach to decrease the complexity of the sample, and the eluents from nLC were analyzed by MS/MS. The MS/MS data were collected on several of the most intense ions per unit time and incorporate an exclusion list of ions that have already been selected for CID. On fragmentation by CID in a tandem mass spectrometer, phosphopeptides not only produce sequence-specific fragments but also fragments that are specific for phosphoryl group. A mixture of bovine casein containing β-casein, αS1-casein, αS2-casein, and κ-casein was digested by trypsin. The Fe (III) IMAC was employed for enrichment of phosphorylated peptides from crude peptide mixtures prior to analysis by nano-liquid chromatography tandem mass spectrometry (nLC-MS/MS). Figure 14 shows the MS and MS/MS spectra acquired from nLC-MS/MS analysis of the IMAC purified phosphopeptides. Doubly charged ion of FQpSEEQQQTEDELQDK from β-casein was observed in MS spectrum, and the sequence and phosphorylation site MS/MS spectrum was identified by Mascot to determine the. An ion series corresponding to N-terminal peptide fragments (b-ions) identified the phosphopeptide FQpSEEQQQTEDELQDK from β-casein. A mass difference of 167 Da between the b2 and b3 ion signals corresponded to the

phosphorylation site of Ser-50. β-elimination of phosphoric acid from phosphoserine was evidenced by a -98 satellite ion signal accompanying the b3 ion. According this method, ten phosphopeptides were identified as belonging to the mixture of bovine casein containing β-casein, αS1-casein, and αS2-casein shown in Table 2.

3-2 Phosphoproteomic Analysis of Human Jurkat T Cell

3-2.1 Strategic Separated Jurkat Human T Cell

3-2.1.1 Subcellular Fractionation of Human Jurkat T cell

Subcellular fractionation is the first and essential step among enrichment techment techniques in proteomics research, which is of special importance for analysis of multiprotein complexes such as cells.²⁶ Subcellular fractionation is a flexible and adjustable approach resulting in reduced sample complexity and facilitates the analysis.

Nuclear and cytoplasmic extraction reagents enabled stepwise separation and preparation of cytoplasmic and nulear extracts from human Jurkat T-cell. Addition of the first reagent to the cell pellet caused disruption of cell membranes and release of cytoplasmic contents. After recovering the intact nuclei from the cytoplasmic extract by centrifugation, the nuclei are lysed with the second reagent to yield the nuclear extract.

About 10⁸ cells were lysed to obtain the cytoplasmic extract and used BCA protein assay to quantify the protein concentration. The cytoplasmic protein extract from 10⁸ cells was determined to be 20 mg.

3-2.1.2 Separation of Subcellular Fractionation Using SDS-PAGE

SDS-PAGE is one of the most reliable methods available for separation of proteins in complex mixtures. A big advantage of this approach is that SDS-PAGE is a well established and highly reproducible method for protein separation under denaturing

conditions in a broad molecular mass range. Gels can be sliced corresponding to molecular mass makers into several fractions. Typically, this approach for fractionation of proteins is a combination of protein separation by SDS-PAGE with subsequent gel slicing, digestion of gel slices with trypsin, and analysis of the resulting peptides using LC-MS/MS. Based on this approach, Fe (Ⅲ) IMAC would be coupled with in order to enrich large-scale phosphopeptides.

In Figure 18, there is the scheme for the analyzing Jurkat T-cell in this study.

Phosphopeptides were enriched and identification from the cytoplasmic protein fraction of Jurkat T-cell. In addition, the use of IMAC-based pre-fractionation as a step in the study of the phosphoproteome has to be considered. Its major advantage is certainly that it can allow an efficient separation for proteins containing phosphorylation. Coupling of advanced separation methodologies to highly sensitive mass spectrometers provides powerful means to detect and analyze dynamic changes of low abundant phosphoproteins in cells on subcellular level.³⁶ SDS-PAGE which is the powerful alternative to size-exclusion chromatography for fractionation of proteins according to molecular mass was utilized for the complex sample from human Jurkat T-cell in order to increase the efficiency of IMAC. Therefore, by using the strategic application of separated methods with IMAC-nLC-MS/MS analysis, this approach has been used for determining large-scale phosphorylation sites on human Jurkat T-cell.

3-2.2 Analysis Human Jurkat T cell Using IMAC with nLC-MS/MS

In Figure 15, there was result of nLC- MS/MS from human Jurlat T cell without fractionation. There were 196 proteins contained 13 phosphoproteins which were

identified. After IMAC purification, there were 166 proteins contained 153 phosphoproteins which were identified as shown in Figure 16. The results described that coupled IMAC with nLC-MS/MS has demonstrated great success in the identification of phosphorylation sites from complex human Jurkat T-cell. About 200 µg sample was used, and the phosphopeptides were successfully identificated by IMAC enrichment strategy.

3-2.3 The Effect of pH on the IMAC Selectivity

In Figure 17, there were two statistical results from different pH value conditions.

In Figure 17 (a), from 185 peptides in 142 proteins with significant Mascot scores, 114 were phosphopeptides in 93 phsphoproteins, indicating 61.6 % selectivity at pH 3.5. In Figure 17 (b), from 223 peptides in 190 proteins with significant Mascot scores, 200 were phosphopeptides in 173 phsphoproteins, indicating 89.7 % selectivity at pH 3.0. Apparently, the percentage of phosphorylated peptides selected from tryptic digestion was increased from 61.6 % at pH 3.5 to 89.7 % at pH 3.0. It is known that nonspecific binding of IMAC is due to carboxyl groups, which would mainly occur via acidic residues such as glutamic and aspartic acid residues. In order to avoid nonspecific binding, pH value of IMAC system must be adjusted. Carboxyl groups are weakly coordinated with Fe-NTA resins because of negative charges from oxygen atoms. When pH value of IMAC system decreased, carboxyl groups were protonated.

The negative charges from carboxyl groups were removed; nonspecific binding of IMAC decreased and selectivity of IMAC increased. However, pH value of IMAC system would not be too low because phosphoryl groups were also protonated.

Therefore, it needed to be compromised pH value of IMAC system in order to avoid

nonspecific binding and maintain phosphopeptide recovery.

When the weak acid HA is dissolved in water, partial dissociation occurs. Thus, we can write

− +

+

↔ H A

HA

] [

] ][

[

HA A K

H

−

=

where Ka is the acid dissociation constant. To obtain the indicative pH value, this equation can be taken the negative logarithms:

] [

] log [

HA pH A

pK

−

=

This equation can be solved for [A^-] / [HA] ratio for known pKa and pH value.

The pKa of aspartic acid, glutamic acid, and phosphory residues are 3.65, 4.25, and 1.52, respectively. In Table 3, there were [A^-] / [HA] ratio of aspartic acid, glutamic acid, and phosphory residues at pH 2 ~ 5. Furthermore, [A^-] / [HA] ratio can be calculated to [A^-] / [HA] + [A^-] ratio shown in Table 3. As a result, the IMAC system was most suitable at pH 3.0 because recovery of phosphopeptides is less than 5 % and nonspecific binding is minimized.

3-2.4 Strategic Separated Jurkat Human T Cell Using SDS-PAGE with

IMAC-nLC-MS/MS Analysis

The pre-treated sample from human Jurkat T-cell was redissolved in 100 µl acetic acid, and was loaded by HPLC pump and autosampler to run IMAC procedures. The eluent from IMAC procedures was collected, and was cleaned up ZipTip. This system utilized two independently controlled HPLC pumps, autosamples, and microvalves to prepare and introduce samples into mass spectrometer. The entire system enabled the reproducible detection and identification of low quantities of phosphopeptides. To identify phosphopeptides in the above sample, MS/MS spectra obtained for human Jurkat T-cell were search with Mascot. In these searches, the parameters specified differential modifications of 80 Da to serine, threonine, and tyrosine residues.

The database search and statistical results were demonstrated the advantage of SDS-PAGE which is the powerful alternative to size-exclusion chromatography for fractionation of proteins according to molecular mass as shown in Figure 20. In Figure 21, there was statistical result of total gel fractions at pH 3.0. About 11,920 MS/MS spectra were acquired during the course of the experiment. A total of 782 peptide sequences defining 891 sites of phosphorylation were determined from the human Jurkat T-cell by IMAC with nLC- MS/MS. From 782 peptides in 603 proteins with significant Mascot scores, 688 were phosphopeptides in 529 phsphoproteins, indicating 88.0 % selectivity at pH 3.0. Classification of identified phosphorylation sites and amino acid frequencies surrounding phosphorylated serine and threonine residues, from 891 detected sites, 823 could be localized to a specific serine or threonine. As a result, without using methylation, about 90% of the identified peptides were found to be phosphorylated. This robust technology platform should be broadly

applicable to profiling the dynamics of phosphorylation.