Evaluation and Automization of Phosphopeptide Enrichment via TiO2 Affinity Chromotography

(1)

Bachelor Thesis Scheikunde

Evaluation and Automization of Phosphopeptide

Enrichment via TiO2 Affinity Chromotography

door

Matthew Delport

28

th

March 2016

Studentnummer

10452540

Onderzoeksinstituut

Verantwoordelijk docent

Van’t Hoff Institute for Molecular Sciences (HIMS)

Prof. Garry Corthals

Onderzoeksgroep

Begeleider

(2)

2

Populair Wetenschappelijke Samenvatting

Het lichaam gebruikt eiwit fosforylering reacties om bepaalde processen te controleren. Je kunt het zien als een lichtknopje, wanneer fosforylering plaats vindt word het lichtknopje aangezet en kan er een bepaald proces plaatsvinden. Dit proces kan ook weer gestopt worden door de reactie terug te draaien en zo wordt het knopje weer uitgezet. Een voorbeeld van waar fosforylering een grote rol speelt binnen onze lichaam is bij het ontwikkelen van kanker. Bepaalde eiwitten worden dan aangezet en vervolgens niet goed uitgezet waardoor een persoon kanker kan krijgen. Fosforylering speelt niet alleen een rol bij velen andere ziektes, maar ook bij het algemeen functioneren van het lichaam. Het kunnen inzien van waar en hoe deze reacties plaats vinden binnen het lichaam is dus bijzonder interessant.

Methodes om eiwit fosforylering te kunnen begrijpen door analyses uit te voeren is over de laatste decennia sterk ontwikkeld maar alles gebeurd wel nog offline, wat inhoudt dat men de analyses handmatig uitvoeren in een lab. Deze methodes zijn tijdrovend waardoor er simpelweg niet genoeg data verkregen kan worden. Ook maken mensen altijd foutjes waardoor deze methodes slecht reproduceerbaar zijn. Deze redenen zorgen ervoor dat er niet genoeg betrouwbare data verkregen kan worden om nuttige conclusies te trekken.

Mijn project richt zich op het ontwikkelen en implementeren van een online proces waardoor de eerdergenoemde uitdagingen verholpen kunnen worden. Een online proces houdt in dat de analyse d.m.v. geautomatiseerde machines uitgevoerd wordt. Deze machines kunnen dan geprogrammeerd worden om de hele tijd door analyses uit te voeren waardoor er niet altijd iemand in het lab hoeft te zijn. Dit maakt het ook mogelijk om veel meer data te verkrijgen in een kortere tijd. Ook zijn machines vele malen nauwkeuriger dan de meesten van ons, wat resulteert in betrouwbare data wat door wetenschappers overal ter wereld gereproduceerd kan worden. Dit is een belangrijke stap in het volledig begrijpen van hoe eiwit fosforylering precies werkt zodat nare ziektes zoals kanker bestreden kunnen worden.

(4)

4

Summary

Cells constantly have to react to signals received by their surroundings, an example being growth-factor signals which are responsible for cell growth, differentiation and proliferation. The degradation of the capability to effectively regulate these cellular responses is one of the main causes of many human diseases such as cancers6. Since the isolation of phosphoserine by Phoebus A. Levene and Fritz A. Lipmann in 1932, post-translational modifications (PTMs) have been found to play a major role in the regulation of cellular responses as they strongly influence and control enzymatic activity, protein conformation, protein-protein interactions, and cellular localization.1,2 In particular protein phosphorylations have been found to be one of the most important PTMs as it is believed that there are over 105 phosphorylated sites in the mammalian proteome3. At any given moment 30-50% of all proteins are phosphorylated4 and the number of genes involved in phosphorylation processes may be as much 2-3% of the whole eukaryotic genome.5 It is therefore of great interest to develop and improve methods for the rapid detection of phosphorylation, phosphorylation site mapping along with protein kinase and phosphatase substrate identification. This will lead to a better understanding of signalling networks and the proteins involved making it possible to develop potent and specific pharmacological signalling modulators for therapeutic use.6 In the quest of finding a capable tool for such research mass spectrometry (MS) has proved to be a powerful option in phosphorylation analysis, however it is still far from being robust due to challenges in phosphoproteomics such as proteins being present in low-abundance and sub-stoichiometric phosphorylation.1 Phosphopeptide enrichment prior to MS analysis is therefore of the essence and an effective method for this is offline titanium dioxide affinity chromatography. This has led to the question of whether titanium dioxide affinity chromatography can be improved and implemented into an online analysis method, which is the aim of this project.

This thesis will begin with a complete discussion of protein phosphorylations by explaining their biological purpose, the chemistry involved and the practical importance of understanding and mapping phosphorylations. Methods currently used in phosphoproteomics including their strengths and weaknesses will be described and compared to the results of our in-house developed methods used in our research. Concluding remarks based on our experimental observations will be presented along with perspectives for future research.

(5)

5

1. Introduction

1.1. The biological importance of phosphorylation

Every form of life is made up of countless gene-encoded proteins which all have their own essential task in maintaining the organism they are situated in. These proteins are even further diversified via so called PTMs which enzymatically and covalently modify proteins during or after their biosynthesis, expanding the number of distinct human protein species by at least one order of magnitude. PTMs are comprised of many types of reactions such as glycosylation, lipidation, cleavage of peptide bonds and many more but one of the most important and prevalent PTM is reversible phosphorylation, catalysed by kinases and phosphatases. Besides covalent modifications, protein functionality is even further broadened by the formation noncovalent complexes resulting in unique functionality, activity or cellular localization. Proteins are able to form homo- and heteromeric complexes as well as complexes with drugs, metabolites or metals. The covalent and noncovalent interactions work side-by-side as a noncovalent complex’s lifetime, function and localization may all be affected by certain covalent modifications (Fig. 1)7. This intricate interplay is what makes the proteome extremely dynamic, in turn providing for the capability of effective sensing and regulation within cells.

Figure 1: Visualization of the expansion of the human proteome by the principles of covalent protein modifications and noncovalent interaction (with proteins, cofactors, metals). The estimated number of human genes (22000 – 25000)

(6)

6

1.2. A short history of protein phosphorylation

The first signs of the proteins containing phosphate were made evident in 1906 with the characterization of the acidic nutritional proteins, casein and vitellin.8 The following advances took place nearly 30 years later when phosphoserine was detected in vitellin,9 a protein isolated from egg yolk. The high levels of phosphorylation within casein grant it calcium binding properties, allowing phosphate and calcium to be readily available for bone formation. The main component of bone is calcium hydroxyapatite, a mixture of calcium hydroxide phosphate. By concluding that phosphoproteins acted as nutritional sources of phosphate and calcium, the first known function of phosphorylated proteins was discovered.

Following these observations, Carl and Gerty Cori in the late 1940’s, and Earl Sutherland, Edwin Krebs and Edmond Fischer all contributed to the revolutionary understanding that phosphate attachment was not only an add-on to nutritional proteins, but that it also controls the function of glycogen phosphorylase, the key enzyme of energy metabolism. This enzyme cuts α-D-glucose from glycogen, forming α-D-glucose-1-phosphate after concomitant attachment of inorganic phosphate. The discovery that glycogen phosphorylase’s activity is regulated by hormones and exists in an active and inactive form was made by the Coris10. Sutherland set out to study the effects of hormones on glycogen metabolism and then made the discovery of cAMP being the first ‘second messenger’ in transmembrane signal transduction. In the same time frame, Krebs and Fischer made the Nobel Prize winning discovery of reversible phosphorylation being a control mechanism for the activation and deactivation of glycogen phosphorylase. They observed that without a phosphorylated Ser15, glycogen phosphorylase was inactive and once phosphorylated, activity was restored.11

The evolution of our understanding of phosphorylations did not stop here, and with the use of cAMP-dependant protein kinase A as a model enzyme, two breakthroughs were achieved. Firstly, it was the first kinase to be fully sequenced12 and visualized with a complete 3D structure.13 This also led to the discovery of tyrosine phosphorylation14 in addition to the previously known serine and threonine phosphorylation. Certain tyrosine kinases were also found to be viral oncogene products and a class of receptors with intracellular tyrosine kinase activity was found which paved the road to discovering a common mechanism for transmembrane signal transduction which will shortly be described. The biological importance of protein phosphorylation has become extremely evident over the last century resulting in an ever increasing interest to expand our knowledge in the field and more revolutionary findings will surely surface.

(7)

7

1.3. Mechanism of Phosphorylation

Protein phosphorylation being as prevailing as it is, it may come as a surprise that it is only the sidechain of three specific amino acids which get reversibly phosphorylated in eukaryotic cells namely; serine, threonine and tyrosine. These three amino acids all have a nucleophilic hydroxyl group enabling the nucleophilic attack of the terminal phosphate group (γ-PO32-) on the ubiquitous

phosphoryl donor adenosine triphosphate (ATP), followed by the transfer of the phosphate group to the amino acid side chain and formation of adenosine diphosphate (ADP). Magnesium (Mg2+) catalyses this transfer by chelating ATP’s γ- and β-phosphate group, resulting in a lower reaction threshold. A large amount of free energy is released once the phosphate-phosphate bond in ATP is broken, making the unidirectional. The discussed mechanism is visualized in Scheme 1.15

Scheme 1: Diagram of serine phosphorylation. Enzyme-catalyzed proton transfer from the hydroxyl group on serine stimulates the nucleophilic attack of the γ-phosphate group on ATP, resulting in transfer of the phosphate group to

(8)

8

1.4. Protein Kinases

As all reactions in the body are driven by enzymes, so is protein phosphorylation which is catalysed by kinases. To date there is believed to be more than 500 kinases in the human proteome; this class of proteins is also referred to as the human kinome16 and substrates are found in various forms such as carbohydrates, lipids, nucleotides and proteins. With ATP being such an efficient phosphate donor, nearly all protein kinases use ATP as their co-substrate, with a few exceptions which make use of guanosine triphosphate. ATP has the ideal structure for the transfer of a α-, β-, γ-phosphate group in nucleotidyl-, pyrophosphoryl- or phosphoryl transfer, respectively.17 Although kinases vary greatly in substrate specificity, they do display a highly conserved ATP-binding site.18

The kinome can be divided into subfamilies which display variable catalytic domain specificity, including tyrosine kinases or serine/threonine kinases. The mammalian kinome is comprised of roughly 80% serine/threonine kinases, and in line with this statistic, over 90% of the phosphoproteome consists of pS and pT. This has been backed by studies which have stated the relative abundance ratio of pS:pT:pY in cells to be 1800:200:1.19 Despite the low abundance of tyrosine phosphorylation, it is a heavily researched in the biomedical field due to its association to human diseases by dysregulation of receptor tyrosine kinases. Besides protein kinases showing specificity towards target amino acids, they also take adjacent consensus sequences in to account.20 These consensus sequences play a role in determining whether a kinase is able to phosphorylate one or more substrates and also the multiplicity in which a single protein can be phosphorylated.

Kinases themselves also have regulatory subunits that are often controlled by phosphorylation21. The majority of protein kinases are found to be in their inactive basal state when dephosphorylated and activated via phosphorylation and the opposite is only true in rare cases. Other kinases may even require a combination of phosphorylated and dephosphorylated sites in order to be activated, the proto-oncogene Src serves as an example of this complexly regulated system22.

(9)

9

1.5. Signal Transduction Cascades

The key to protein phosphorylations being extremely efficient in signal transduction lies in their reversibility which makes them capable of swiftly responding to intra- and extracellular stimuli15. Signal transduction cascades are set off by one or more proteins physically sensing cues through cleavage, ligand binding or other responses. These proteins then submit a signal to second messengers and signalling enzymes, which in case of phosphorylation, activate downstream kinases, resulting in phosphorylation and activation of linked downstream substrates, including additional kinases. The described process is repeated until the desired physiological response is achieved. Signal transduction cascades have been found to proceed in two fashions. One being a linear fashion whereby kinase I activates kinase II which then activates kinase III and so forth. The second form is the amplification of an initial signal whereby kinase I activates multiple kinases, which in turn also all activate other kinases (Fig.2)15. The second form of signalling makes it possible for a single molecule, for example a growth factor, to activate global cellular programs including proliferation.23

Figure 2: Signal transduction cascades amplify the signal output. External and internal stimuli induce a broad range of cellular responses through a series of second messengers and enzymes. Linear signal transduction pathways yield the sequential activation of a discrete number of downstream effectors, while other stimuli elicit signal cascades that

amplify the initial stimulus for large-scale or global cellular responses.15

An example of how meaningful such cascades are lies in one of the molecular mechanisms of cancer development in which genomic mutations hijack a signal transduction pathway, resulting in a dysregulated surplus of cellular proliferation. Recognizing viral oncogene products as tyrosine kinases or products directly related to the interference of intracellular signal transduction triggered the widespread acceptance of kinases and phosphatases being involved in cellular growth control. A recent and rather impressive example is the finding of the chromosomal translocation leading to the Philadelphia chromosome is the molecular cause of chronic myelogenous leukaemia (CML).24 Fully understanding these mechanisms and pathways plays an essential role in pharmacology and in order to do so, effective analysis methods have to be established.

(10)

10

1.6. MS of Phosphopeptides

With data claiming that 1/3 of all proteins in a mammalian cell are phosphorylated25, the human genome contains 500 genome-encoded protein kinases and that phosphorous to sulphur ratios point towards a high degree of protein phosphorylation26, it is safe to say that phosphorylation appears to be the most abundant and essential protein modification. This being said, phosphorylations are only intermittently detected in proteomic analyses not specially adapted for phosphopeptide detection. Additionally, a method capable of efficiently detecting multiply phosphorylated peptides has yet to be developed as it is only singly and some doubly phosphorylated peptides which are readily detected. The inconsistency between what is expected and what is actually detected regarding protein phosphorylations has led to the general belief that phosphopeptide analysis is an extremely challenging practice. Researchers interested in the matter have somewhat united and agreed upon some of the underlying factors that are apparently making this analysis so challenging. Most reports on this topic tend to attribute undesired results to these factors which are: (i) the widely sub-stoichiometric phosphorylation of proteins, (ii) the suppression of phosphopeptide ionization in the presence of an excess of unmodified peptides, (iii) a generally lower ionization efficiency of phosphopeptides versus non-phosphopeptides (in the generally applied positive ion mode), and (iv) the application of data-directed MS/MS analysis which leads to a notorious under sampling of peptides of low abundance. A recent article has focused on sorting these blindly used arguments into facts and myths27. The article makes the following conclusion regarding each argument. Argument (i) is likely true as the majority of phosphopeptide signals detected in digests are accompanied by their nonphosphorylated analogues which are also more abundant. Argument (ii) is likely false or at least not always true in the strict sense as stated. LC-MS experiments were executed with a complex protein digest which was also spiked with variable amounts of phosphopeptides. These experiments were not able to confirm a decrease of the absolute phosphopeptide signals intensities with the increasing excess of unmodified peptides28. This being said, it is true that in LC-MS, with electrospray ionization (ESI), the MS detection usually involves relatively simple mixtures which have already passed LC separation. However one of ESI’s true weaknesses is its limited dynamic range due to all analytes competing for the limited number of excess charges on the spray droplets. This makes the conclusion ‘that when in the presence of a large number of analytes the less abundant species will fail to be ionized’, plausible. However concrete evidence supporting this general hypothesis is yet to be developed. Argument (iii) is true for many cases with a couple exceptions, although this effect doesn’t seem to be highly significant. Finally, argument (iv) was deemed true and can certainly result in reduced detection of phosphopeptides when combined with a lower ionization efficiency and phosphopeptide abundance.

Factors leading to phosphopeptide analysis being notoriously challenging which were not discussed in the article mentioned above but however do play an important role are phosphoprotein degradation during sample work up, for example by phosphatases, resistance of phosphoproteins to digestion as well as metal-ion mediated adsorption of multiply phosphorylated phosphoproteins or phosphopeptides on surfaces, to which the sample is exposed in the analytical process7.

(11)

11

1.7. Phosphopeptide Enrichment

Due to the challenging nature of phosphopeptide analysis it is of the essence to make use of phosphopeptide enrichment techniques. Usually using phosphoprotein rich samples followed by protease-specific digestion and MS-analysis does not suffice and therefore complex samples require a second enrichment step in order to accurately determine the sites of phosphorylation. There are commercial kits available on the market making the enrichment process relatively easy, fast, and reproducible however is has been shown that these various methods differ in specificity of isolation and in the set of phosphoproteins and phosphopeptides isolated29, indicating that the use of one single method may not be sufficient for global phosphoproteome analysis. Various enrichment methods will be discussed below.

1.7.1. Immunoprecipitation

Total proteins are immunoprecipitated via the use of antibodies specific to phosphorylated residues. Immunoprecipitation of proteins containing phosphotyrosine is used more often than that of proteins containing phosphoserine or phosphothreonine as the antibodies specific to the former tend to be much more reliable than those specific to the latter30,31,32,33. Phosphotyrosine (pTyr) residues have therefore been studied much more intensely over recent years, despite their low abundancy compared to phosphoserine (pSer) and phosphotheronine (pThr) residues and specific pTyr binding domains (PTB) have resulted in the determination of global tyrosine phosphorylation states of the cell34,35. It is not yet fully understood why pSer- and pThr-specific antibodies fail to work efficiently in phosphoprotein immunoprecipitation, however these antibodies are renowned to be extremely specific towards certain consensus motifs. In reality, immunoprecipitation of pSer and pThr phosphoproteins can only be achieved with an expensive mixture of multiple antibodies, resulting in only a minor amount of research being reported in this area36.

1.7.2. Immobilized Metal-Ion Affinity Chromatography (IMAC)

Originally being introduced for purification of His-tagged proteins37, IMAC has evolved into one of the most commonly used technique for phosphopeptide enrichment. The reason for this technique showing improved success rates is due to the fact that it reduces ion suppression effects that would otherwise occur in untreated complex samples38. Negatively charged phosphate groups on phosphopeptides bind to the IMAC stationary phase via electrostatic interactions with positively charged metal ions, which are bound to the column material via linkers such as nitriloacetic acid (NTA), iminodiacetic acid (IDA), and Tris(carboxymethyl)ethylenediamine (TED). It was initially found that immobilized metal ions such as Ni2+, Co2+, and Mn2+ preferentially bind to proteins with a high His density but more relevantly, immobilized metal ions of Fe3+, Ga3+, and Al3+ have higher specificity for phosphopeptides. Zr4+ has also recently been reported to bind phosphopeptides with a high specificity39. A major disadvantage in IMAC methods is they tend to non-specifically bind peptides containing the acidic amino acids glutamine and aspartate and the strong binding of multiply phosphorylated peptides. The issue of binding acidic amino acids can be countered via esterification of the carboxylic acids to methyl esters using hydrochloric acid-saturated, dried methanol40, however reaction conditions must be chosen with extreme care to prevent the occurrence of side reactions

(12)

12

which would increase the complexity of the sample. The experimental conditions (for example; pH, ionic strength, or organic composition of the solvents) of any IMAC procedure have to also be carefully considered as slight variations are known to have radical effects on the specificity of the IMAC stationary phase. Despite these drawbacks, IMAC has become increasingly popular, and one of the reasons for this is the exceptional compatibility with subsequent separation and detection techniques such as LC-ESI-MS/MS41.

1.7.3. Titanium Dioxide (TiO2)

Pinkse et al. have recently reported the use of titanium dioxide as a promising alternative to IMAC42. The method is based on the selective interaction of water-soluble phosphates with porous titanium dioxide microspheres via bidentate binding at the TiO2 surface. A TiO2 precolumn is used to trap

phosphopeptides under acidic conditions, which are then subsequently desorbed under alkaline conditions. Even though TiO2 based columns still trap nonphosphorylated acidic peptides, it has been

reported to display increased specificity towards the enrichment of phosphopeptides. The nonspecific binding of acidic peptides has been found to be able to be diminished to a certain extent when 2,5-dihydroxy-benzoic acid was used during peptide loading43. This novel procedure has proven to be more effective than IMAC. An in-depth analysis has recently been reported describing the relationship between the occurrence of some amino acids and the phospho-specific and nonspecific binding of peptides using TiO2-based enrichment44. Two well characterized peptide mixtures, each

consisting of either 33 or 8 synthetic phosphopeptides or their non-phosphorylated analogue and varying in charge and hydrophobicity were tested. The acquired data confirmed that TiO2 is indeed

highly selective for phosphopeptide enrichment. It is noteworthy to mention that a drastic decrease in phosphopeptide recovery was found in samples containing phosphopeptides with multiple basic amino acids. Once again being able to be coupled to a LC—ESI-MS/MS workflow has been a major advantage of this technique.

Figure 3: Schematic discription of the interaction of a phosphopeptide with a TiO2 surface. The directly interacting partners are oxygen atoms of the phosphate group and positively polarized titanium (IV) atoms. It is assumed that the two negatively charged oxygen atoms of a phosphate group may interact either with a single Ti(IV) centre or with two

(13)

13

1.7.4. Online Phosphopeptide Enrichment

The enrichment of phosphopeptides is often executed manually as automated procedures have not yet been adequately developed. These manual offline procedures involve the use hand-made columns by packing the desired solid phase into pipette tips, which are then used to desalt and enrich the phosphopeptides. Offline procedures are not only highly labour intensive, but also result extremely poor reproducibility due to factors such as loading speed which cannot be precisely controlled45. It is therefore of great interest to establish efficient online enrichment methods if we are to move forward in the field of phosphoproteomics. Online methods have many advantages which make it an attractive route to follow. Analytical instrumentation can carry out procedures with high precision factors such as loading speed and sample handling being easily controlled which would greatly improve reproducibility. The possibility of automation also serves as a solution for the labour insensitivity of offline methods as unattended operation and ultimately high throughput analysis is made possible. In the journey to fully understanding protein phosphorylations, it is essential to improve the factors mentioned and this can be achieved by creating online enrichment methods.

1.8. The Aim of This Project

The main goal of this project was to establish a fully online method for enrichment of phosphopeptides using a TiO2 pre-column. Recent attempts have been made however these

procedures were not fully online as they involved offline desalting and sub-fractionation of the sample prior to the online phosphopeptide enrichment45. On the road to achieving the goal, offline and semi-online methods will also be executed and analysed in order to predict conditions which will be successful in online enrichment.

(14)

14

2. Methodology

2.1. Column Packing & Fritting

2.1.1. Chemicals, Materials and Apparatus Used

The following chemicals, material and apparatus were used for column packing Table 1: Chemicals, materials and apparatus used for column fritting and packing

Chemical, material, apparatus Manufacturer

Milli-Q water, 18 MΩcm Merck Millipore

TiO2 particles, 5 μm particle diameter Zirchrom

Ethanol absolute for analysis Emsure

Fused silica cappilaries Polymicro Technologies

Kasil 1 potassium silicate NEXT Advance Frit Kit

Formamide 25% NEXT Advance Frit Kit

GF/C glass microfiber filters, 25 mm diameter Whatman

In-house built pressure bomb

2.2.1. General Procedure

Fused silica capillaries were cut with a diamond cutter and inspected under a microscope to ensure that the ends were smooth, which were otherwise smoothen with a ceramic file. A mixture of 25 % formamide (100 μL) and milli-q water (75 μL) was made and vortexed. Kasil 1 (100 μL) was then added to the mixture and vortexed once more. One end of the cut capillaries was fritted by gently poking holes in the GF/C filters after wetting the penetration area with the prepared kasil / formamide solution (2 μL). This process was repeated 5 times per frit. Finally the capillaries were left to bake overnight at 85 °C with the frit-end down.

Once fritted, column packing could be executed. A slurry of TiO2 (4 mg) and EtOH AAS (200 μL) was

made and vortexed. The slurry was then placed into the pressure bomb while being magnetically stirred. Columns were cleaned by wetting an analytical tissue with EtOH AAS and gently wiping the column before being placed into the pressure bomb with the frit-end up. The TiO2 particles were

packed under a pressure of 1000 psi. Once the desired bed length had been reached the bomb was closed and the column was swiftly removed. The stationary phase bed was consolidated to form a stable bed formation by passing 100% Milli-Q water through the column at a flow rate of 1 uL/min.

(15)

15

2.2. Trypsin Digestion of α-casein From Bovine Milk

All percentages used in the methodology are specified as volume/volume (v/v).

2.2.1. Chemicals used

The following chemicals were used for the digestion of alpha-casein.

Table 2: Chemicals used for Trypsin Digestion of α-casein From Bovine Milk

Chemical Purity Manufacturer

Milli-Q water, 18 MΩcm Merck Millipore

Ammonium bicarbonate (NH4HCO3) ≥ 99.5 % Fluka

DL-Dithiolthreitol (DTT) ≥ 99.0 % Sigma

Iodoacetamide (IAA) ≥ 99 % Sigma

α-casein from bovine milk ≥ 70 % Sigma

Urea Sigma

Trypsin BRP European Pharmacopoeia reference Standard

2.2.2. Solutions Prepared

For the digestion of α -casein, the following solutions were prepared:

 2,5 M ammonium bicarbonate was prepared by dissolving x g in 20 μL milli-q water

 6 M urea solution was prepared by dissolving urea (3.9990 g) in NH4HCO3 (2.5 M, 2.5 mL)

 Reducing reagent was prepared by dissolving DTT (59.8 mg) in NH4HCO3 (2.5 M, 1.0 mL)

 Alkylating reagent was prepared by dissolving IAA (72.0 mg) in NH4HCO3 (2.5 M, 1.0 mL)

 Trypsin solution was prepared by dissolving trypsin (3.4 mg) in NH4HCO3 (2.5 M, 3.4 mL)

α-casein was reconstructed in urea solution (100 μL) in a 2 mL Eppendorf tube. Reducing reagent (5 μL) was added and the mixture was left to reduce for 1 hour at 37 °C. Once reduced, alkylating reagent (20 μL) was added and left to react for 1 hour in the dark (wrapped in aluminium foil). After 1 hour, reducing reagent (20 μL) was added to consume leftover alkylating reagent present. NH4HCO3

(2.5 M, 900 μL) was then added to dilute the urea and finally trypsin (1:30 trypsin to protein ratio by weight) was added and left to digest over night at 37 °C. The sample was stored at -20 °C

See Error! Reference source not found. for amounts of α-casein and trypsin solution used. Table 3: Variable amounts of chemicals used

Date

Sample Name 29-04-2016 20160429MDα1 29-04-2016 20160429MDα2

α-casein from bovine milk 3.6 mg 2.7 mg

(16)

16

2.3. Off-line Phosphopeptide Enrichment of Digested α-casein Sample

2.3.1. Chemicals, materials and apparatus used

The following chemicals, materials and apparatus were used for the desalting and phosphopeptide enrichment by TiO2 affinity chromatography (Table 4):

Table 4: Chemicals, materials and apparatus used for Off-line Phosphopeptide Enrichment of Digested α-casein Sample

Furthermore, in-house made GF/C and C18 spin columns were made via the following procedures: The GF/C spin column was made by cutting 3 small discs (approximate radius of 0.5 mm) from a GF/C sic with the end of a glass pipette. The cut discs were placed inside a 200 μL pipette tip to serve as a plug.A small hole was then pricked into the lid of a 2 mL Eppendorf tube and the pipette tip was placed into this hole so that it was firmly held above the 1 mL mark.

The C18 spin column was made by cutting 1 small disc (approximate radius of 0.5 mm) from a C18 disc with the end of a glass pipette. The cut disc was placed inside a 200 μL pipette tip to serve as a plug.A small hole was then pricked into the lid of a 2 mL Eppendorf tube and the pipette tip was placed into this hole so that it was firmly held above the 1 mL mark.

(17)

17

2.3.2.1. Desalting prior to enrichment:

A previously digested sample 20160429MDα1 thawed at 4 °C and then handled at room temperature. An empore C18 cartridge was placed in to a 2 mL Eppendorf tube after cutting a small hole in the tubes lid. The cartridge was then wet with ACN (1 mL) via centrifugation (2000 rpm, 60 s) and then placed in a new 2 mL eppendorf tube. The digest sample was then loaded (2000 rpm, 60 s) and the flow-through was recovered and loaded once more. The cartridge was then washed 3 times with 0.1% TFA (1 mL, 2000 rpm, 60 s) and eluted with 6% TFA / 80% ACN (1 mL, 2000 rpm, 60 s). The desalted digest was stored at 4 °C for an hour before proceeding to the enrichment.

2.3.2.2 Phosphopeptide enrichment via TiO2 affinity chromatography:

The centrifugation times noted were in insufficient in certain cases. In the case of incomplete flow-through, centrifuge until complete.

TiO2 particles (5 μm, 11.3 mg) were added to milli-Q water (226 μL) to form slurry. Theslurry was

loaded onto the in-house made GF/C spin column via centrifugation (4000 rpm, 120 s). The column was then equilibrated with 6% TFA / 80% ACN (200 μL, 4000 rpm, 120 s) after which the desalted sample was loaded in steps of 200 μL (4000 rpm, 180 s). The column was then washed twice with 6% TFA / 80% ACN (200 μL, 4000 rpm, 180 s) flowed by a 2 washes with 0.1% TFA (200 μL, 4000 rpm, 180 s). After washing, 5 % NH4OH was added to the column and the phosphopeptides were manually

eluted with a plastic syringe (PTFE tape was wrapped around the syringe tip to ensure a tight connection with the pipette tip.) into a new Eppendorf tube. The large amount of back-pressure resulted in a lot of man-force to be applied. After elution 10% FA (400 μL) was added to acidify the sample. The following desalting should be immediately performed to avoid methionine oxidation. Centrifugation times should be lengthened if incomplete flow-through is observed.

2.3.2.3. Desalting after phosphopeptide enrichment:

A C18 spin column was wet with ACN (50 μL, 4000 rpm) and equilibrated with 0.1% TFA (50 μL, 4000 rpm). The enriched phosphopeptide sample was then loaded onto the column in 200 μL (4000 rpm). The column was then washed 3 times with 0.1% TFA (50 μL, 4000 rpm). Finally the phosphopeptides were manually eluted with a syringe into a new tube. As explained in the enrichment step, tape was used for a secure fit and a large amount of man-power was needed. The eluate was then evaporated to dryness in a GeneVac DNA concentrator for 30 minutes and then stored at -20 °C. The sample was named; 20160502 MD2.

Note that centrifugation was executed until complete flow-through was observed which usually took approximately 10 minutes.

(18)

18

2.3.3. LC-MS Method for Phosphopeptide Analysis

LC-MS analysis was performed on an inline setup of a Eksigent nanoLC quarto serie coupled to a TripleTOF 5600 + (Sciex, Framingham MA, USA).

2.3.3.1. Instrumental conditions

After calibration of the LC-MS system, the C18 trap column (NanoLC, ChromXP trap, C18 3 μm fully porous particles, 350 μm i.d. x 0.5 mm) was attached to the trapping/injection valve. After purging the system for 20 cycles in the loading buffer (0.1% TFA 3% acetonitrile, 97% LC-MS grade water v/v) the trap was equilibrated at 1 μL/min for 45 min. The analytical column was in-house packed with Magic AQ (New Objective, Woburn MA, USA) C18 stationary phase (5 μm fully porous particles, 100 angstrom pore size). The C18 column was equilibrated at 300 nL/min for 1 hr after purging the A and B reservoirs for 20 cycles in 0,1% formic acid in MS grade water for A and 0,1% formic acid in LC-MS grade ACN for B. After equilibration the sample (10 μL) was loaded onto the trap via the LC-LC-MS autosampler. After washing with loading buffer for 10 min, the sample was eluted off the trap column by the system automatically switching the valve so that the trap column was inline with the analytical column, this allowed the gradient mobile phase flow through the trap column, eluting the phosphopeptides. The gradient conditions were as follows, the total runtime was 100 min:

Time (min) %A %B 0 40 60 50 1 99 60 1 99 62 40 60 90 40 60 2.3.3.2. MS parameters:

Accumulation time: 250 mS based on Sciex recommendations 400-250 m/z mass range

Multiple charged ions enabled

Max number of candidate ions: 30 (relates to fragmentation) TOF mass range 200 to 1800 m/z

2.3.3.3. Sample preparation

α -casein digestions were prepared and enriched using the in-solution digestion and enrichment procedures described previously. After digestion and enrichment, these samples were stored for 1 week at -20 C until analysis. Prior to analysis, the enriched digests were removed from -20 C and diluted to 10 μL with the loading buffer before injection.

(19)

19

2.4. Semi On-line Phosphopeptide Enrichment of Digested α-casein

Sample

2.4.1. Chemicals, Materials and Equipment Used

The same chemicals, materials and apparatus were used as in section 2.2.1., with the addition of the following apparatus (Table 5: Additional appartus used for the semi-online phosphopeptide

enrichment):

Table 5: Additional appartus used for the semi-online phosphopeptide enrichment

Apparatus

Acquity UPLC Binary solvent manager, M-class, Waters Acquity UPLC μSample manager – FL, Waters

Acquity UPLC TUV detector, Waters

The desalting procedures performed on sample 20160429MDα2 were the same as those in section 2.3.2.

2.4.2.1. Online enrichment:

A home packed TiO2 column (75 μm ID, 300 Å pore size) was attached to the M-Class UPLC. 0.1% TFA

was placed in the A1 solvent position and 6% TFA / 80% ACN in the B1 position. The system was purged and then slowly ramped up to a flow-rate of 0.500 μL/min under a flow of 100% solvent B1. The pressure started stabilizing around 1000 psi. The column was left to equilibrate for 45 minutes. The loading method was made and applied which comprised of 40 injections of 5 μL. The column was then washed with 100% of solvent B1 for 30 minutes followed by a 45 minute wash with 100% of solvent A1. During this wash, the 6% TFA / 80% ACN in the B1 position was switched with 5% NH4OH.

After washing, the elution took place with 100% of the new B1 solvent for 1 hour. The eluent was collected in a 2 mL Eppendorf tube. Finally, 70 μL was added to the enriched phosphopeptide sample and subsequent desalting was executed.

2.4.3. LC-MS Method for Phosphopeptide Analysis

The LC-MS analysis of the sample acquired in the semi on-line phosphopeptide enrichment procedure was analysed in the same manner as described in section 2.3.3.

(20)

20

2.5. Fully On-line Phosphopeptide Enrichment of Digested α-casein

Sample

2.5.1. Instrumental Conditions and MS Parameters

The instrumental conditions and MS parameters of the fully on-line phosphopeptide enrichment were the same as those described in section 2.3.3. with the following exceptions;

An in-house packed trap column (20 μm TiO2 particles, 300 Å, 200 μm ID) was attached to the

trapping/injection valve instead of the C18 trap column. 5% ammonium hydroxide in LC-MS grade water was attached to line A and 45% water, 50 % acetonitrile for line B.

2.5.1. Sample Preparation

Separate digestions of yeast and alpha-casein were prepared using the in-solution digestion procedure described previously. After digestion, these digests were stored for 1 week at -20 C until analysis. Prior to analysis, the digests were removed from -20 C and injected without dilution to be in agreement with the offline enrichment procedure where the digest was enriched without dilution.

(21)

21

3. Results & discussion

3.1. Offline Phosphopeptide TiO

2

Enrichment

3.1.1. Condition Performance

A common drawback of TiO2 enrichment techniques is nonspecific isolation of acidic peptides along

with the desired phosphopeptides. Current techniques devoted to countering this challenge exist, such as O-methyl esterification of carboxyl-groups40, however result in incomplete yields and unwanted side reactions such as deamidation and subsequent methylation of asparagine and glutamine residues43. Another method is the addition of dihydroxybenzoic acid (DHB) during phosphopeptide binding which leads to a decrease in the amount of non-phosphorylated peptides being detected43. The addition of DHB may cause problems with LC-MS systems and contaminate the MS’ ion source making it incompatible with online methods46,47. Due to the drawbacks mentioned above, these techniques were not explored. Rather, TFA was used as mobile phase during phosphopeptide binding and washing and NH4OH was used to elute the phosphopeptides. 0.1% TFA

is known to have pH of 1.8 – 2.0 resulting in acidic residues remaining in their neutral state during phosphopeptide loading minimizing the amount of acidic residues binding to the TiO2 stationary

phase opposed to when acetic acid (pH 2.7-2.9) is used. As the pH of phosphoric acid decreases from 1.8 to 1.1 when methylated48, a similar decrease is expected when bound to a peptide which would mean that under the acidic conditions provided by a 0.1% TFA mobile phase, the phosphate groups would be negatively charged and readily bind to the TiO2 stationary phase. Recently presented work

stated that elution with ammonium bicarbonate (pH 9) was not sufficient to elute all phosphopeptides and that a pH of 10.5 is optimal43. This is why the stronger base ammonium hydroxide was used as to elute the phosphopeptides.

3.1.2. Peptides Detected

The conditions used resulted in the detection of 537 peptides of which 354 were phosphorylated. 273 of these peptides belonged to α-casein S1 (86.5 % sequence coverage) with 117 phosphopeptides. Mainly single phosphopeptides were isolated with a few doubles and multiples (Fig. 4). 168 peptides belonged to the α-casein S2 protein (71.17 % sequence coverage), of which 42 were phosphorylated.

(22)

22

Figure 4: Number of singly, doubly and multiply phosphopeptides isolated belonging to α-casein S1 The programme searched against the Uniprot database (not species specific). False discovery rate: 1%. Total number of distinct peptides

detected was 273, total number of distinct phosphopeptides detected was 117.

As α-casein is a well-studied and understood protein, it may be used as a benchmark to compare our method to other studies. According to the literature49 nine phosphorylation sites exist on α-casein S1 and the method used detected five of these sites. This result confirms that the method used is capable of effectively determining phosphosites. However, α-casein S2 contains ten phosphosites, of which only two were detected50. This could simply mean that the protein sample predominantly contains α-casein S1.

The phosphosites detected can be can be validated by manually analysing the obtained spectra for each peptide. The mass spectra contain b- and y-ion peaks which correlate to cleavages resulting in the charge being retained by either the amino- or carboxy-terminus respectively. By validating that cleavage has taken place and a b- or y-ion peak is present after and before a suggested phosphorylated site, respectively, one can confirm whether that specific phosphorylation is indeed present. In the case of a confirmation, the preceding ion will have the mass of the corresponding peptide fragment excluding a phosphate group and the mass of the ion after the phosphosites will correlate to the mass of the corresponding peptide fragment including a phosphate group51. When dealing with suggested pThr and pTyr sites, one must take into account that these are much less common than pSer sites and therefore confirmation of such sites must be done so with caution, especially when the suggested phosphosite is neighbouring a Ser. The method used to confirm phosphosites is discussed below using pSer130 as an example:

The mass spectrum (Figure 5: see next page) indicating phosphorylation at Ser 130 corresponds to the following peptide fragment:

PNS[Pho]AEER

The obtained y6-ion with a mass of 785.2825 corresponds to the peptide fragment NSAEER with water and a phosphorylation. The y4-ion with a mass of 504.2413 corresponds to the peptide fragment AEER with water. This data indicates that Ser130 is indeed phosphorylated. This conclusion can be further validated by also analysing the b2- and b5-ion which also support the existence of pSer130

(23)

23

Figure 5: Mass spectrum and masses of b- and y-ion peaks corresponding to peptide fragment PNS[Pho]AEER. Screenshot taken from ProteinPilot database search output (Sciex, USA). The programme searched against the Uniprot database (not

(24)

24

3.1.3. Novel Phosphosites on α-casein S1

Of the 117 α-casein S1 phosphopeptides detected, according to ProteinPilot, 10 were distinct and 6 of these have not yet been reported in literature (Table 6). Since the phosphorylation of α-casein S1 has been fully characterised, it is expected that the newly observed phosphorylation sites seen in this study are most likely due to the high concentrations we used, which may not reflect realistic conditions. In any case, it is prudent to examine the phosphorylation sites reported by ProteinPilot. The validity of the reported phosphorylation sites were examined by manually interpreting the mass spectra as shown in the previous section. The results are shown below in Table 1.

Table 6: Summary novel distinct phosphopeptides detected belonging to α-casein S1 and whether they can be confirmed or not.

Position Notes

Thr64 Possible but cannot be confirmed

Ser103 Possible but cannot be confirmed

Ser195 Data suggests existence

(25)

25

3.1.3.1. pSer195

pSer195 belongs to the peptide with the following sequence: TDAPSFS[PHO]DIPNPIGSENSEK

Spectra obtained suggesting that pSer195 is phosphorylated have a b8-ion present with a mass of 901.2975 Da which indicates that the TDAPSFSD fragment is indeed phosphorylated, this could be on Ser193 or Ser195. Furthermore the b6-ion is also present with a mass of 619.2722 Da which indicates that the fragment TDAPSF is not phosphorylated meaning that Ser195 is phosphorylated instead of Ser 193. The y17+2-ion present can only confirm that either Ser193 or Ser195 is phosphorylated, however b-ions present are sufficient to suggest that pSer195 exists.

Figure 6: Mass spectrum and masses of b- and y-ion peaks corresponding to peptide fragment

TDAPSFS[Pho]DIPNPIGSENSEK. Screenshot from output of a database search using ProteinPilot. The programme searched against the Uniprot database (not species specific). False discovery rate: 1%

(26)

26

3.1.4.

Novel Phosphosites on α-casein S2

Of the 42 phosphorylated peptides belonging to the α-casein S2 protein, 7 were distinct and 5 were novel phosphorylation not reported in earlier literature (Table 7). These phosphorylations can be verified in the same manner utilized in section 3.1.2. However all spectra obtained for these novel phosphosites had insufficient data present to make any confirmations.

Table 7: Summary novel phosphopeptides detected belonging to α-casein S1 and whether they can be confirmed or not.

Position Notes

3.1.5. Conclusion: Offline TiO

2

Enrichment

The results acquired under the chosen conditions seemed promising to apply to an online method. A large amount of peptides were detected, of which 66 % were phosphorylated showing that this technique is effective in trapping phosphopeptides. The method also proved effective by determining the new phosphorylation site pSer195 and many other possible novel phosphorylations were found but lacked sufficient data for confirmation. These results were effective therefore we decided to choose these conditions as a starting point to carry out the semi-online phosphopeptide TiO2

enrichment.

3.2. Semi-Online Phosphopeptide TiO

2

Enrichment

3.2.1. Detected Peptides

After following the semi-online phosphopeptide enrichment procedure 148 distinct phosphopeptides were detected of which 22 were phosphorylated. 83 of these distinct peptides belonged to α-casein S1 (80.84 % sequence coverage) and 31 to α-casein S2 (49.55 % sequence coverage). Of these peptides, 12 belonging to α-casein S1 were phosphorylated and 7 belonging to α-casein S2 were phosphorylated. From the known phosphosites in α-casein S1 and α-casein S2, only 3 sites were detected for α-casein S1 and 2 were detected for α-casein S2. The decrease in peptides detected is related to two factors; the amount of α-casein digested and the amount of sample loaded during enrichment. In both cases these amounts were lower during the semi-online enrichment. 2.7 mg of α-casein was digested in the semi-online method opposed to 3.6 mg for the off-line method and 0.200 ml of the desalted sample was used to load the TiO2 column online opposed to 1 ml during the

(27)

27

and instrument flow rate limitations: the maximum flow rate was 2.0 uL/min which would require 200 mins of elution time. Considering the lower amounts used it is no surprise that the number of detected proteins are lower, if 1 ml was to be loaded and if the number of peptides detected correlate linearly to sample volume, we could expect that 5 x 148 = 740 peptides would be detected. It is remarkable that only 15 % of the detected peptides were phosphorylated. This decrease may be due to the use of an in-house packed column and further column optimization would have to be executed to improve the phosphopeptide selectivity. The online method has still proven to specifically isolate singly phosphorylated peptides (Figure 7: Number of singly and doubly phosphopeptides isolated belonging to α-casein S1).

Figure 7: Number of singly and doubly phosphopeptides isolated belonging to α-casein S1. ProteinPilot searched against the Uniprot database (not species specific). False discovery rate: 1%. Total number of distinct peptides detected was 83,

total number of distinct phosphorylated peptides detected was 12.

3.2.2. Novel Phosphosites on α-casein S2

The semi-online method detected far fewer novel phosphosites compared to that of the offline method. This could mean that is less efficient but as many of the novel phosphosites in the off-line method could not be confirmed and are likely false identifications, it can be argued that the semi-online method is more efficient by displaying less false identification.

One novel phosphosite was however suggested by ProteinPilot; α-casein S2 at Thr159. The corresponding mass spectra did not display enough data to make a confirmation and pThr159 was regarded as a false identification.

3.2.3. Conclusion: Semi-online TiO

2

Enrichment

Although the phosphopeptide specificity of the semi-online method was not as highly observed for the off-line method, it still proved selective for phosphopeptide enrichment. Possible reasons lie in the decreased mass loading capabilities due to instrument restrictions. Less mass was loaded onto the column since a single injection could only load 20 μL of sample and took approximately 3 minutes for one injection cycle. The low flow rate also resulted in lengthy analysis times and reducing these times can lead to incomplete washing of non-phosphorylated peptides and elution of phosphopeptides. Increasing the washing time could also improve the final phosphopeptide

11 1

# of single # of double

(28)

28

specificity by eliminating more of the non-phosphorylated peptides from the column prior to elution. However, this may also increase dilution which would decrease sensitivity.

The results under the chosen conditions showed effectiveness to a certain degree and were therefor also used for the fully on-line method. However, we packed a shorter column and used larger TiO2

particles for increased porosity, both facilitating the use of higher flow rates by reducing the system backpressure.

3.3. Online Phosphopeptide TiO

2

Enrichment

The attempt at on-line failed and no data was able to be obtained. The main cause for this was the high pH of the two mobile phases which were both approximately 12. This proved to be far too basic as it was observed that the silica tubing started dissolving. The pH range compatible with a typical LC-MS system is known to be 2-8 however there are systems which are compatible with a broader range of pH values. These systems make use of stainless steel tubing which unfortunately would not be compatible with phosphopeptide enrichment processes as metals are known to have interactions with phosphopeptides. Additional challenges in online phosphopeptide were already mentions in 3.2.3., such as low injection volumes resulting in low mass loading and low flow rates resulting in lengthy analysis times or insufficient washing and elution when analysis times are lowered. To determine whether it was possible to lower the pH to more LC-MS compatible values while preserving selectivity for phosphopeptide enrichment, a series of offline phosphopeptide enrichments were carried out at different pH values. Unfortunately, the MS sensitivity was low to produce any meaningful conclusions.

(29)

29

3. Conclusion & Outlook

The results obtained from the offline and semi-online methods clearly show that TiO2 is effective at

enriching phosphopeptides. The offline method did however have better results than the semi-online method which can be due to numerous factors. These factors include lower mass loading and lower flow rates. This can be countered by increasing analysis times however one of the reasons for seeking an online method is to decrease analysis time. The in-house packed TiO2 column could also undergo

optimization. During the semi-online method the column used had an inner diameter of 75 μm and the particles used had a diameter of 5 μm. By increasing the particle diameter the porosity of the bed would increase resulting in less backpressure which would facilitate higher flow rates.

Although the in-house packed columns proved promising, it was not able to implement them in a fully online method. The main reason for failure was the extremely high pH caused by the

ammonium hydroxide mobile phase. As a high pH is needed to specifically elute the phosphopeptides this is a worrying factor however the pH used was around 12 which led to the exploration of

phosphopeptide elution using different pH values ranging from 9.6 to 11.4. Unfortunately no results were obtained when analysed via MS due to complications and this is still an area worth addressing in the future.

(30)

30

Acknowledgements

I would like to thank Prof. Garry Corthals for welcoming me into his research group and making it possible for me to work on the project. I am also very grateful for Dr. Michelle Camenzuli being such a great day-to-day mentor and supervisor which supported me whenever needed. Furthermore I would like to thank the whole Corthals group and it was a pleasure working with each and every one of them. Finally I would like to thank Dr. Leo de Koning for his time and effort in being my second assessor.

References

1_{A. Paradela, J.P. Albar, Advances in the analysis of protein phosphorylation, Journal of proteome} research 7 (2008) 1809.

2_{J.V. Olsen, B. Blagoev, F. Gnad, B. Macek, C. Kumar, P. Mortensen, M. Mann, Global, In Vivo, and} Site- Specific Phosphorylation Dynamics in Signaling Networks, Cell 127 (2006) 635-648.

3_{http://www.phosphosite.org}

4_{D.E. Kalume, H. Molina, A. Pandey, Tackling the phosphoproteome: tools and strategies, Curr.} Opin. Chem. Biol. 7 (2003) 64-69.

5_{M.J. Hubbard, P. Cohen, On target with a new mechanism for the regulation of protein} phosphorylation, Trends Biochem. Sci. 18 (1993) 172-177.

6_{T. Hunter, Signaling— 2000 and Beyond, Cell 100 (2000) 113-127.}

7_{W.D. Lehmann, Protein phosphorylation analysis by electrospray mass spectrometry a guide to} concepts and practice / / Wolf D. Lehmann. Cambridge : : Royal Society of Chemistry, Cambridge, (2010).

8_{P.A. Levene, C.L. Alsberg, The cleavage products of vitellin, J. Biol. Chem. 2 (1906) 127–133.} 9_{F.A. Lipmann, P.A. Levene, Serinephosphoric acid obtained on hydrolysis of vitellinic acid, J. Biol.} Chem. 98 (1932) 109–114.

10_{C. F. Cori, G. T. Cori, Polysaccharide phosphorylase, In Nobel lectures, physiology or medicine}

1942-1962, Amsterdam: Elsevier, (1947) 186-206.

11_{E. H. Fischer, D. J. Graces, E. R. Crittender, E. G. Krebs, Structure of the site phosphorylated in} the phosphorylase b to a reaction, J. Biol. Chem. 234 (1959) 1698–1704.

12_{S. Shoji, D. C. Parmelee, R. D. Wade, S. Kumar, L. H. Ericsson, K. A. Walsh, H. Neurath, G. L.} Long, J. G. DeMaille, E. H. Fischer, K. Titani, Complete amino acid sequence of the catalytic subunit of bovine cardiac muscle cyclic AMP-dependent protein kinase, Proc. Natl. Acad. Sci. USA 78 (1981) 848–851.

13_{D. R. Knighton, J. Zheng, L. F. T. Eyck, V. A. Ashford, N. H. Xuong,}

S. S. Taylor, J. M. Sowadski, Crystal structure of the catalytic subunit of cyclic adenosine monophosphate-dependent protein kinase, Science 253 (1991) 407–414.

14_{T. Hunter, B. M. Sefton, Transforming gene product of Rous sarcoma virus phosphorylates} tyrosine, Proc. Natl. Acad. Sci. USA 77 (1980) 1311–1315.

15 https://www.thermofisher.com/nl/en/home/life-science/protein-biology/protein-biology-learning-center/protein-biology-resource-library/pierce-protein-methods/phosphorylation.html# 16_{G. Manning et al., The protein kinase complement of the human genome, Science 298 (2002)} 1912-34.

17_{C. Walsh, Enzymatic reaction mechanisms, San Francisco: W. H. Freeman xv (1979) 978.}

18_{C. Walsh, Posttranslational modification of proteins: Expanding nature's inventory, Englewood,} Colo.: Roberts and Co. Publishers xxi (2006) 490.

19_{M. Mann et al., Analysis of protein phosphorylation using mass spectrometry: Deciphering the} phosphoproteome, Trends Biotechnol. 20 (2002) 261-8.

20_{T. Pawson, P. Nash, Assembly of cell regulatory systems through protein interaction domains,} Science 300 (2003) 445-52

21_{L. N. Johnson, R. J. Lewis Structural basis for control by phosphorylation, Chem Rev. 101 (2001)} 2209-4

(31)

31

22_{M. B. Yaffe, Phosphotyrosine-binding domains in signal transduction. Nat. Rev. Mol. Cell Biol. 3}

(2002) 177-86.

23_{G. L. Johnson, R. Lapadat, Mitogen-activated protein kinase pathways mediated by ERK, JNK,} and p38 protein kinases, Science 298 (2002) 1911-2

24_{P. C. Nowell, J. Clin, Discovery of the Philadelphia chromosome: a personal perspective, Invest.} 117 (2007) 2033–2035.

25_{P. Cohen, The origins of protein phosphorylation, Nat. Cell Biol. 4 (2002) E127–E130.}

26 _{R. Krϋger, D. Kϋbler, R. Pallisse´ , A. Burkovski, W. D. Lehmann, Protein and proteome}

phosphorylation stoichiometry analysis by element mass spectrometry, Anal. Chem. 78 (2006) 1987–1994.

27_{H. Steen, J.A. Jebanathirajah, J. Rush, N. Morrice, M.W. Kirschner, Phosphorylation analysis by} mass spectrometry: myths, facts, and the consequences for qualitative and quantitative measurements, Molecular & cellular proteomics : MCP 5 (2006) 172.

28 _{H. Steen, J. A. Jebanathirajah, M. Springer,M.W. Kirschner, Stable isotope-free relative and} absolute quantitation of protein phosphorylation stoichiometry by MS, Proc. Natl. Acad. Sci. USA 102 (2005) 3948–3953.

29 _{B. Bodenmiller, N.M. Lukas, M. Mueller, B. Domon, R. Aebersold, Reproducible isolation of} distinct, overlapping segments of the phosphoproteome, Nature Methods 4 (2007) 231.

30_{A. Pandey, M.M. Fernandez, H. Steen, B. Blagoev, M.M. Nielsen, S. Roche, M. Mann, H.F. Lodish,} Identification of a novel immunoreceptor tyrosine- based activation motif- containing molecule, STAM2, by mass spectrometry and its involvement in growth factor and cytokine receptor signaling pathways, The Journal of biological chemistry 275 (2000) 38633.

31 _{A. Pandey, A.V. Podtelejnikov, B. Blagoev, X.R. Bustelo, M. Mann, H.F. Lodish, Analysis of} receptor signaling pathways by mass spectrometry: Identification of Vav- 2 as a substrate of the epidermal and platelet- derived growth factor receptors, Proceedings of the National Academy of Sciences of the United States 97 (2000) 179.

32_{H. Steen, B. Kuster, M. Fernandez, A. Pandey, M. Mann, Tyrosine phosphorylation mapping of the} epidermal growth factor receptor signaling pathway, The Journal of biological chemistry 277 (2002) 1031.

33 _{Y.G. Yeung, Y. Wang, D.B. Einstein, P.S. Lee, E.R. Stanley, Colony- stimulating factor- 1} stimulates the formation of multimeric cytosolic complexes of signaling proteins and cytoskeletal components in macrophages, The Journal of biological chemistry 273 (1998) 17128.

34_{G. Caratù, D. Allegra, M. Bimonte, G.G. Schiattarella, C. D'Ambrosio, A. Scaloni, M. Napolitano,} T. Russo, N. Zambrano, Identification of the ligands of protein interaction domains through a functional approach, Molecular & cellular proteomics : MCP 6 (2007) 333.

35 _{K. Machida, C.M. Thompson, K. Dierck, K. Jablonowski, S. Kärkkäinen, B. Liu, H. Zhang, P.D.} Nash, D.K. Newman, P. Nollau, T. Pawson, G.H. Renkema, K. Saksela, M.R. Schiller, D. Shin, B.J. Mayer, High- Throughput Phosphotyrosine Profiling Using SH2 Domains, Mol. Cell 26 (2007) 899-915.

36_{M. Grønborg, T.Z. Kristiansen, A. Stensballe, J.S. Andersen, O. Ohara, M. Mann, O.N. Jensen, A.} Pandey, A mass spectrometry-based proteomic approach for identification of serine/threonine-phosphorylated proteins by enrichment with phospho- specific antibodies: identification of a novel protein, Frigg, as a protein kinase A substrate, Molecular & cellular proteomics : MCP 1 (2002) 517. 37 _{J. Porath, J. Carlsson, I. Olsson, G. Belfrage, Metal chelate affinity chromatography, a new} approach to protein fractionation, Nature 258 (1975) 598.

38 _{G.L. Corthals, R. Aebersold, D.R. Goodlett, Identification of Phosphorylation Sites Using} Microimmobilized Metal Affinity Chromatography, Meth. Enzymol. 405 (2005) 66-81.

39_{S. Feng, M. Ye, H. Zhou, X. Jiang, X. Jiang, H. Zou, B. Gong, Immobilized zirconium ion affinity} chromatography for specific enrichment of phosphopeptides in phosphoproteome analysis, Molecular & cellular proteomics : MCP 6 (2007) 1656.

40_{S.B. Ficarro, M.L. Mccleland, P. Todd Stukenberg, D.J. Burke, M.M. Ross, J. Shabanowitz, D.F.} Hunt, F.M. White, Phosphoproteome analysis by mass spectrometry and its application to Saccharomyces cerevisiae, Nat. Biotechnol. 20 (2002) 301.

41_{K. Moser, F.M. White, Phosphoproteomic analysis of rat liver by high capacity IMAC and LC- MS/} MS, Journal of proteome research 5 (2006) 98.

42 _{M.W.H. Pinkse, P.M. Uitto, M.J. Hilhorst, B. Ooms, A.J.R. Heck, Selective isolation at the} femtomole level of phosphopeptides from proteolytic digests using 2D- NanoLC- ESI- MS/ MS and titanium oxide precolumns, Anal. Chem. 76 (2004) 3935.

(32)

32

43 _{M.R. Larsen, T.E. Thingholm, O.N. Jensen, P. Roepstorff, T. Jørgensen J.D., Highly selective} enrichment of phosphorylated peptides from peptide mixtures using titanium dioxide microcolumns, Molecular & cellular proteomics : MCP 4 (2005) 873.

44_{C. Klemm, S. Otto, C. Wolf, R.F. Haseloff, M. Beyermann, E. Krause, Evaluation of the titanium} dioxide approach for MS analysis of phosphopeptides, Journal of Mass Spectrometry 41 (2006) 1623-1632.

45_{M.W.H. Pinkse, S. Mohammed, J.W. Gouw, B. van Breukelen, H.R. Vos, A.J.R. Heck, Highly} robust, automated, and sensitive online TiO2-based phosphoproteomics applied to study endogenous phosphorylation in Drosophila melanogaster, Journal of proteome research 7 (2008) 687.

46_{T. E. Thingholm, T. J. D. Jorgensen, O. N. Jensen, M. R. Larsen, Highly selective enrichment of} phosphorylated peptides using titanium dioxide, Nat. Protocols, 1, 4 (2006) 1929–1935.

47 _{N. Sugiyama, T. Masuda, K. Shinoda, A. Nakamura, M. Tomita, Y. Ishihama, Phosphopeptide}

enrichment by aliphatic hydroxyl acid-modified metal oxide chromatography for NanoLC-MS/MS in proteomics applications, Mol. Cell. Proteomics, 6 (2007) 1103–1109.

48_{A. Saha, N. Saha, L. N. Ji, J. Zhao, F. Gregan, S. A. A. Sajadi, B. Song, and H. Sigel, Stability of}

metal ion complexes formed with methyl phosphate and hydrogen phosphate, J. Biol. Inorg. Chem. 1 (1996) 231–238

49_{http://www.uniprot.org/uniprot/P02662} 50_{http://www.uniprot.org/uniprot/P02662}

51_{I.A. Papayannopoulos, ChemInform Abstract: The Interpretation of Collision‐ Induced Dissociation} Tandem Mass Spectra of Peptides, ChemInform 27 (1996) no-no.

Evaluation and Automization of Phosphopeptide Enrichment via TiO2 Affinity Chromotography

Bachelor Thesis Scheikunde