• No results found

Protein mass spectrometry in the drug discovery process

N/A
N/A
Protected

Academic year: 2023

Share "Protein mass spectrometry in the drug discovery process"

Copied!
53
0
0

Loading.... (view fulltext now)

Full text

(1)

From the Department of Medical Biochemistry and Biophysics Karolinska Institutet, Stockholm, Sweden

PROTEIN MASS

SPECTROMETRY IN THE DRUG DISCOVERY

PROCESS

Agneta Tjernberg

Stockholm 2005

(2)
(3)

"What cannot be understood is no object of belief."

Sir Isaac Newton

(4)

All previously published papers were reproduced with permission from the respective publisher.

Published and printed by Karolinska University Press Box 200, SE-171 77 Stockholm, Sweden

© Agneta Tjernberg, 2005 ISBN 91-7140-251-9

(5)

To my family

(6)

ABSTRACT

Most targets for drug discovery are proteins. Drug discovery includes target identification and validation, structural biology, lead generation, lead optimisation, pre-clinical development and clinical trials. The overall aim of this thesis is to apply protein mass spectrometry (MS) in the early drug discovery process. Matrix-assisted laser desorption/ionisation time-of-flight MS (MALDI-TOF MS) and electrospray ionisation-MS (ESI-MS), as well as other biophysical methods, have been used in different stages in this process.

Both MALDI-TOF MS and ESI-MS/MS were used in the first study to identify a polypeptide from rat liver cytosol, found to be active during glycopeptide export. The 23- kDa glycopeptide of interest was identified as rat guanylate kinase (GK). GK was found to fully support glycopeptide export from the rat liver microsomes in the presence of ATP.

The ability of GK to substitute for complete cytosol was also confirmed.

In the second study, two model peptides, KFFEAAAKKFFE and KFFEYNGKKFFE, were created and compared in order to elucidate the molecular determinants for amyloid formation. By the use of several biophysical methods it was shown that the amyloidogenic properties were strongly dependent on the structural environment of the amyloidogenic sequence. Freshly dissolved KFFEAAAKKFFE appeared as a monomeric, unstructured conformation and formed thick fibril bundles over time. In contrast, KFFEYNGKKFFE did not form fibrils but underwent oligomerization into a structure of defined size. Results from analytical ultracentrifugation and nondenaturing ESI-MS strongly suggest that this peptide formed 12-mers.

In the third study, non-denaturing ESI-MS was used for the development of a binding assay for the determination of dissociation constants between proteins and low-molecular-mass compounds. By the introduction of an MS response factor, which is a measure of the stability of the complex in the gas-phase, the method became valid both for ionic and non- polar interactions. As a model system, we used the extracellular soluble domain of the human growth hormone receptor, a drug target for the treatment of growth hormone disorders.

In the fourth study, MALDI-TOF MS and ESI-MS/MS were used in combination with other biophysical techniques to characterize how pyridazine analogues inhibit protein tyrosine phosphatase 1B (PTP1B). PTP1B is a drug target for non-insulin dependent diabetes mellitus and obesity. This study showed that pyridazine analogues inhibit PTP1B via an indirect mechanism, which generated hydrogen peroxide. The hydrogen peroxide oxidised the active site cysteine, leading to enzyme inactivation.

Finally, in the fifth study, different biophysical techniques were used to investigate the effect of dimethyl sulfoxide (DMSO) with regard to protein stability, protein aggregation and binding affinities of drug compounds. DMSO is the standard solvent for the preparation of stock solutions of compounds from chemical libraries. The assay concentration of DMSO can be as high as 5% (v/v), or 715 mM. Our study revealed significant differences in the behaviour of the proteins in the presence and absence of low amounts of DMSO (0.1- 3%). In addition, we showed that low DMSO concentrations influence the ionisation process in ESI-MS.

(7)

LIST OF ORIGINAL ARTICLES

This thesis is based on the following articles, referred to in the text by their Roman numerals. The articles are reprinted by permission from the respective copyright holders.

I. Ali B.R.S, Tjernberg A, Chait B.T, and Field M.C.

A microsomal GTPase is required for glycopeptide export from the mammalian endoplasmic reticulum.

J. Biol. Chem. 2000; 275: 33222-33230.

II. Hosia W, Bark N, Liepinsh E, Tjernberg A, Persson B, Hallén D, Thyberg J, Johansson J, and Tjernberg L.

Folding into a β-hairpin can prevent amyloid fibril formation.

Biochemistry 2004; 43: 4655-4661.

III. Tjernberg A, Carnö S, Oliv F, Benkestock K, Edlund P.O, Griffiths W.J, and Hallén D.

Determination of dissociation constants for protein – inhibitor complexes by electrospray ionization mass spectrometry.

Anal. Chem. 2004; 76: 4325-4331.

IV. Tjernberg A, Hallèn D, Schultz J, James S, Benkestock K, Byström S, and Weigelt J.

Mechanism of action of pyridazine analogues on protein tyrosine phosphatase 1B (PTP1B).

Bioorg. Med. Chem. Lett. 2004; 14: 891-895.

V. Tjernberg A, Markova N, Griffiths W.J, and Hallén D.

Dimethyl sulfoxide related effects in protein characterization.

Submitted.

(8)

CONTENTS

1 ABBREVIATIONS...1

2 INTRODUCTION...3

2.1 General background ... 3

2.2 Mass spectrometry ... 4

2.2.1 Background ... 4

2.2.2 Matrix-assisted laser desorption/ionisation (MALDI) ... 5

2.2.3 Electrospray ionisation (ESI) ... 6

2.2.4 Mass analysers ... 8

2.3 The drug discovery process... 11

2.3.1 Target identification and validation ... 11

2.3.2 Structural studies... 12

2.3.3 Lead generation... 15

2.3.4 Lead optimization ... 17

2.3.5 Pre-clinical development... 17

2.3.6 Clinical development ... 17

2.4 Other methods ... 17

2.4.1 Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS- PAGE) ... 17

2.4.2 Reversed phase high performance liquid chromatography (RP-HPLC). ... 18

2.4.3 Nuclear magnetic resonance (NMR)... 18

2.4.4 Isothermal titration calorimetry (ITC) ... 18

2.4.5 Differential scanning calorimetry (DSC) ... 19

2.4.6 Dynamic light scattering (DLS) ... 19

2.4.7 Circular dicroism (CD) ... 19

2.4.8 Analytical ultracentrifugation (AUC) ... 19

3 AIM OF THE THESIS...21

4 RESULTS AND DISCUSSION...22

4.1 Identification of guanylate kinase (GK) (paper I)... 22

4.2 two polypeptide structures in amyloid fibril formation (Paper II)... 25

4.3 Determination of dissociation constants for protein-ligand complexes by ESI-MS (paper III) ... 26

4.4 Mechanism of action of pyridazine analogues on protein tyrosine phosphatase 1B (PTP1B) (paper IV)... 30

4.5 Dimethyl Sulfoxide related Effects in Protein Characterization (paper V) 33 5 CONCLUSIONS...35

6 ACKNOWLEDGEMENTS...37

7 REFERENCES...40

(9)

1 ABBREVIATIONS

Three- and one-letter codes for the 20 commonly occurring amino acids and their monoisotopic residue masses in Da.

Amino acid 3-letter 1-letter MW (Da)

Alanine Ala A 71.037

Arginine Arg R 156.037

Asparagine Asn N 114.043

Aspartic acid Asp D 115.027

Cysteine Cys C 103.009

Glutamic acid Glu E 129.043

Glutamine Gln Q 128.059

Glycine Gly G 57.021

Histidine His H 137.059

Isoleucine Ile I 113.084

Leucine Leu L 113.084

Lysine Lys K 128.095

Methionine Met M 131.040

Phenylalanine Phe F 147.068

Proline Pro P 97.053

Serine Ser S 87.032

Threonine Thr T 101.048

Tryptophan Trp W 186.079

Tyrosine Tyr Y 163.063

Valine Val V 99.068

ATP Adenosine triphosphate

AUC Analytical ultracentrifugation

CD Circular dichroism

CID Collision induced dissociation Da Dalton

DC Direct current

DIOS Desorption/ionisation on silicon DLS Dynamic light scattering

DMSO Dimethyl sulfoxide

DSC Differential scanning calorimetry

DTT Dithiothreitol ECD Electron capture dissociation

ER Endoplasmic reticulum

ESI Electrospray ionisation

FP Fluorescence polarization

(10)

FRET Fluorescence resonance energy transfer FTICR Fourier-transform ion-cyclotron resonance

GK Guanylate kinase

H/D Hydrogen/deuterium HPLC High performance liquid chromatography HTS High throughput screening

IC50 Inhibitory concentration 50%

ITC Isothermal titration calorimetry

MALDI Matrix-assisted laser desorption/ionisation

KD Dissociation constant

LC/MS Liquid chromatography/mass spectrometry LSIMS Liquid matrix secondary ion mass spectrometry

MS Mass spectrometry

MS/MS Tandem mass spectrometry

m/z Mass-to-charge ratio

NMR Nuclear magnetic resonance

PTP1B Protein tyrosine phosphatase 1B

PTMs Post-translational modifications

QTOF Quadrupole time-of-flight

RF Radio frequency

RP Reversed phase

SDS-PAGE Sodium dodecylsulphate polyacrylamide gel electrophoresis SEC Size exclusion chromatography

S/N Signal-to-noise SPA Scintillation proximity assay SPR Surface plasmon resonance TBS Tris buffered saline

TCEP Tris(2-carboxyethyl)phosphine hydrochloride

TFA Trifluoroacetic acid

TIC Total ion current

TOF Time-of-flight

TR-FRET Time resolved fluorescence resonance energy transfer UV Ultraviolet

(11)

2 INTRODUCTION

2.1 GENERAL BACKGROUND

Drug discovery and development are the processes of identifying and evaluating drugs for the safe and effective treatment of human disease. Drug discovery includes target identification and validation, structural biology, lead generation, lead optimisation, pre-clinical development and clinical development:

Target id and validation

Lead generation

Lead optimization

Pre-clinical development

Clinical development Structural

biology Target id

and validation

Lead generation

Lead optimization

Pre-clinical development

Clinical development Structural

biology Target id

and validation

Lead generation

Lead optimization

Pre-clinical development

Clinical development Structural

biology

Proteins are the primary targets for most drug discovery efforts. The recent advances in mass spectrometry (MS) with “soft” ionization methods that enable studies of macromolecules make MS a useful tool in the drug discovery process. The role of protein MS in pharmaceutical discovery continues to expand for both functional characterization and structural analysis. Matrix-assisted laser desorption/ionisation time-of-flight MS (MALDI-TOF MS) and electrospray ionisation-MS (ESI-MS) are frequently used in different stages of the early drug discovery process:

- MS is a key technique in proteomic studies for the identification of novel target proteins.

- MS methods are used in order to investigate the quality of target proteins employed in screening and in structural studies driving drug design.

- MS is an excellent tool for the elucidation of mechanisms of action between target proteins and drug compounds.

- MS is useful for the assessment of noncovalent interactions, protein folding and dynamics.

- MS can be used as a screening tool for the determination of dissociation constants between target proteins and compounds.

In the five papers included in this thesis, MS is employed for protein identification, for the study of protein folding, for the determination of binding affinities, for the study of mechanism of action, and for biophysical characterization.

(12)

2.2 MASS SPECTROMETRY

2.2.1 Background

MS allows for the mass measurement of ions. All MS instruments consist of three basic components (Fig. 1). First, an ion source ionises the molecules, then a mass analyser separates the ions according to their mass-to-charge ratio (m/z) and finally, a detector measures the ion beam current. Each of these elements exists in many forms and is combined to produce a wide variety of MS instruments with specialized characteristics.

Detector

e.g. Photomultiplier, Microchannel plate, Electron multiplier

Analyser

e.g. Quadrupole, time-of-flight,

ion trap

Ion source

e.g. ESI, MALDI

Data system

Detector

e.g. Photomultiplier, Microchannel plate, Electron multiplier

Analyser

e.g. Quadrupole, time-of-flight,

ion trap

Ion source

e.g. ESI, MALDI

Data system Data system

Figure 1. The mass spectrometer.

The mass spectrometer is based on the work performed by the English scientist Sir J.

J. Thomson. His research, which led to the discovery of the electron in 1897, also led to the first mass spectrometer1. While he measured the effects of electric and magnetic fields on ions generated by gases present in cathode ray tubes, Thomson noticed that the ions moved through parabolic trajectories proportional to their "mass- to-charge" ratios.

The period from the late 1930's to the early 1970's was a time of great achievements in the field of MS, especially for mass analysers. In 1946, William E. Stephens proposed the concept of time-of-flight (TOF) analysers2, and in the mid-1950’s the quadrupole analyser, as well as the quadrupole ion trap, were developed by Wolfgang Paul3,4.

A new type of ion source, chemical ionisation (CI)5 was developed in 1966 and made it possible to ionise thermo-labile biomolecules for the first time. In CI, abundant reagent gas ions are first formed by electric ionisation of a reagent gas, and the reagent ions then ionise analytes present in the gas phase. Plasma desorption (PD)6, introduced in 1976, used high-energy ions (252Cf) to desorb and ionise molecules bound to a thin foil. PD achieved some success for the analysis of proteins, but was

(13)

never shown to be reliable for molecular masses higher than 10 kDa. A milestone was achieved when M. Barber et al7, 8 described the successful use of a non-volatile chemical protection environment (i.e. glycerol) next to the molecules to enable polar and thermally labile compounds to survive the ionisation process. This work showed that accelerated (fast) atoms, typically argon or xenon, could be used for ionisation of small biomolecules (i.e. masses below 10 kDa) combined with on-line fragmentation for structure determination9. This technique, termed fast atom bombardment (FAB), and the closely related method liquid matrix secondary ion MS (LSIMS), did not solve the problem of reaching higher masses but had a major impact on expectations on future success.

ESI and soft laser desorption, i.e. MALDI, were developed in the mid 1980's and had a significant impact on the capabilities of MS. Both these ionisation techniques had the ability to analyse high molecular-mass compounds, and they are described in more detail below.

2.2.2 Matrix-assisted laser desorption/ionisation (MALDI)

MALDI was developed by Tanaka et al.10 and by Hillenkamp and Karas11, 12, 13, 14. MALDI provides non-destructive vaporization and ionisation of both large and small biomolecules. In MALDI analysis, the analyte is first co-crystallized with a large molar excess of a matrix compound, usually an UV-absorbing weak organic acid15. The most frequently used matrix materials are derivatives of cinnamic acid, introduced by Beavis and Chait (e.g. α-cyano-4-hydroxy-cinnamic acid16 and sinapinic acid17), and benzoic acid derivatives, introduced by Karas and Hillenkamp (i.e. 2,5-dihydroxybenzoic acid18). The laser used in MALDI instruments is generally a nitrogen laser emitting at 337 nm. A pulse from the UV laser results in the vaporization of the matrix, which carries the analyte. The matrix therefore plays a key role by absorbing the laser energy and by transferring the energy, causing the analyte to vaporize (Fig 2). The matrix also serves as a proton donor and proton acceptor and ionises the analyte in both positive and negative ionisation modes.

Very large proteins can be ionised by MALDI. Nelson et al19 have demonstrated the analysis of human IGM at m/z approximately 1 MDa. The MALDI technique is very sensitive and we have used it in proteomics-based research for the identification of proteins (paper I). The protein of interest is digested and the masses obtained by MALDI of the proteolytic peptides are compared with sequence information from protein databases (peptide mass fingerprinting20, 21, 22). Peptide mass fingerprinting

(14)

can also be used for the detection of covalent modifications of proteins since the peptide fragment containing the modified amino acid will have an increased mass as compared to the non-modified fragment (paper IV). MALDI in combination with hydrogen/deuterium (H/D) exchange was used to study the flexibility of short peptides (paper II).

Non-covalent complexes are normally not observed by MALDI-TOF MS, since the proteins are denatured when analysed. However, the use of MALDI-TOF-MS for the study of non-covalent protein-protein interactions has been reported. The results have shown the technique to be effective with some, but not all complexes23.

+ +

+ +

+

M A L D I ta rg e t

Pulsed laser beam

Matrix analyte

+ +

+ +

+

M A L D I ta rg e t M A L D I ta rg e t

Pulsed laser beam

Matrix analyte Matrix analyte

Figure 2. The MALDI ion source. The matrix plays a key role by absorbing the laser light energy and indirectly causing the analyte to vaporize.

2.2.3 Electrospray ionisation (ESI)

The principle of ESI was first described by Dole et al24, 25 and the successful combination of ESI and MS was demonstrated by Fenn et al26, 27, 28. ESI generates ions directly from solution by creating a fine spray of highly charged droplets in the presence of a strong electric field (typically 3.5 kV). As the solvent evaporates and the droplet decreases in size, the charge density increases. When the repulsion between charges on the surface exceeds the surface tension, ions are drawn out to form what is known as a Taylor cone29. If the surface tension at the tip of the Taylor cone is exceeded by the applied electrostatic force, smaller charged droplets are formed. The droplets are then transported through a pressure gradient and an electric potential gradient towards the mass analyzer. During that transition the droplets reduce in size. Two theories exist on the mechanism by which desolvated ions are formed from charged droplets. The ‘charge residue model’ (CRM)30 proposes that ions originate from extremely small droplets containing one single molecule of the

(15)

analyte. The ‘ion desorption model’ (IDM)31,32 assumes that before a droplet reaches the ultimate stage, its surface electric field becomes sufficiently large to lift an analyte ion at the surface of the droplet over the energy barrier and into the gas-phase. An illustration of the electrospray ionisation process is shown below (Fig. 3). The general opinion is that gaseous protein ions are formed via the CRM mechanism and that small ions are produced by the IDM 33,34,35.

Counter electrode

Sample Solution

- +

A

Enlargement of the spray

B

++ + +

+ +

+ + + +

+

+ + ++

+ +

+ +

+ ++

+ ++ + +

Ion desorption model

Charge residue model

Surface electric field ’lifts’ analyte

ion from droplet

Solvent evaporation from a droplet containing a single

analyte molecule + ++

+ + + ++ + +

+ +

+ ++ +

++ + +

+ +

+ + + +

+ ++ + +

+ +

+ + + +

+ ++ + +

+ +

+ + + +

+

+ + ++

+ +

+ +

+ ++ + + ++

+ +

+ +

+ ++ + + ++

+ +

+ +

+ ++

+ ++ + + + ++ + + + ++ + +

Ion desorption model

Charge residue model

Surface electric field ’lifts’ analyte

ion from droplet

Solvent evaporation from a droplet containing a single

analyte molecule + ++

+ ++ + ++ ++ + ++ +

+ ++ + + + ++ + + + ++ + +

+ +

+ ++ + + +

+ ++ + + ++ + + + ++ + + + ++ + +

Counter electrode

Sample Solution

- +

A

Counter electrode

Sample Solution

- +

Counter electrode

Sample Solution

- + Sample

Solution Sample Solution Sample Solution

- +

A

Enlargement of the spray

B

++ + +

+ +

+ + + +

+

+ + ++

+ +

+ +

+ ++

+ ++ + +

Ion desorption model

Charge residue model

Surface electric field ’lifts’ analyte

ion from droplet

Solvent evaporation from a droplet containing a single

analyte molecule + ++

+ + + ++ + +

+ +

+ ++ +

++ + +

+ +

+ + + +

+ ++ + +

+ +

+ + + +

+ ++ + +

+ +

+ + + +

+

+ + ++

+ +

+ +

+ ++ + + ++

+ +

+ +

+ ++ + + ++

+ +

+ +

+ ++

+ ++ + + + ++ + + + ++ + +

Ion desorption model

Charge residue model

Surface electric field ’lifts’ analyte

ion from droplet

Solvent evaporation from a droplet containing a single

analyte molecule + ++

+ ++ + ++ ++ + ++ +

+ ++ + + + ++ + + + ++ + +

+ +

+ ++ + + +

+ ++ + + ++ + + + ++ + + + ++ + +

Enlargement of the spray

B

++ + +

+ +

+ + + +

+

+ + ++

+ +

+ +

+ ++

+ ++ + +

Ion desorption model

Charge residue model

Surface electric field ’lifts’ analyte

ion from droplet

Solvent evaporation from a droplet containing a single

analyte molecule + ++

+ + + ++ + +

+ +

+ ++ +

++ + +

+ +

+ + + +

+ ++ + +

+ +

+ + + +

+ ++ + +

+ +

+ + + +

+

+ + ++

+ +

+ +

+ ++ + + ++

+ +

+ +

+ ++ + + ++

+ +

+ +

+ ++

+ ++ + + + ++ + + + ++ + +

Ion desorption model

Charge residue model

Surface electric field ’lifts’ analyte

ion from droplet

Solvent evaporation from a droplet containing a single

analyte molecule + ++

+ ++ + ++ ++ + ++ +

+ ++ + + + ++ + + + ++ + +

+ +

+ ++ + + +

+ ++ + + ++ + + + ++ + + + ++ + +

Figure 3. A. The ESI process (positive mode). Positive ions in the sample solution are directed towards the liquid surface emitting a spray of fine droplets at the tip of the spray capillary. The charged droplets are attracted by the counter electrode and the low pressure behind the orifice. B. The droplets are rapidly reduced in size either according to the ion desorption model or according to the charge residue model.

ESI-MS can be used under denaturing or non-denaturing conditions. For analyses such as protein identification based on the amino acid sequence of proteolytic peptides (paper I), we used denaturing conditions in order to maximize sensitivity.

Both the sensitivity and the resolution of the peaks are higher under denaturing conditions, where organic solvents at low pH are used. Non-denaturing ESI-MS allows the transfer of non-covalently bound complexes from a non-denaturing solution (e.g. ammonium acetate pH 7) into the gas-phase. This technique was employed in papers II, III, and V for the characterization of peptides and native proteins. In paper III, dissociation constants for protein-ligand complexes were determined by non-denaturing ESI-MS. The first articles that describe the identification of non-covalent protein-ligand interactions by MS were published 199136, 37, 38. This technique is now widely used for detecting interactions between proteins or between proteins and low molecular mass compounds.

In order to maintain non-covalent complexes during the transfer from the solution into the gas phase, it is important to create a less steep pressure gradient in the interface

(16)

between the atmosphere and the vacuum. The way this is achieved varies between different types of MS instruments. In general, reduced source pumping and collisional cooling will help to preserve the non-covalent complexes39, 40, 41, 42. In order to optimise our Q-TOF Ultima API instrument for this type of analysis, an extra valve was installed on the source backing pump, so that the pump could be throttled to give a reading of ∼2 mbar on the backing line pirani gauge. Ar was introduced into the collision cell to give a reading on the analyser penning gauge of 1.8e-5 mbar.

The major questions concerning non-covalent binding have been whether the interactions in the gas-phase resemble those in solution, and if the interactions are possible to quantify. Many studies confirm that the affinities obtained from ESI-MS reflect the situation in solution43, 44, 45, 46, 47, 48, 49, 50. In paper III we present a new method for measuring and calculating dissociation constants that can be applied both for polar and non-polar interactions.

2.2.4 Mass analysers

In this thesis three different mass analysers have been used: the quadrupole, the quadrupole ion trap, and the time-of-flight (TOF) analyser. Our Q-TOF Ultima instrument consists of a combination of quadrupoles and a TOF analyser.

2.2.4.1 Quadrupoles

The quadrupole mass analyser consists of four parallel rods to which is applied a direct current (DC) voltage and a superimposed radio-frequency (RF) potential3, 4. By adjusting the RF to DC ratio, only ions of a specific m/z can pass the electric field.

The m/z is directly proportional to the voltage applied to the rods. Thus, by changing the voltages, the m/z of interest are allowed to pass to the collector where they are detected. The quadrupole is the most widely used analyser due to its ease of use, large m/z range, and good linearity for quantitative work. A disadvantage is the limited resolution.

2.2.4.2 Ion traps

The physics behind the quadrupole ion trap is very similar to the quadrupole mass analyser3, 4. In an ion trap the ions are trapped in an RF quadrupole field. A mass spectrum is obtained by changing the RF field to eject the ions from the trap. It is also possible to isolate one ion species by ejecting all others from the trap. The isolated ions can subsequently be fragmented by collisional activation. The fragments

(17)

detected generate a fragmentation spectrum. The primary advantage of quadrupole ion traps is that multiple collision-induced dissociation (CID)51, 52, 53 experiments (MSn) can be performed without having multiple analysers. Other important advantages include the compact size and the ability to trap and accumulate ions to increase the signal-to-noise ratio of a measurement.

2.2.4.3 Time-of-flight

The TOF analyser is one of the simplest mass analysing devices2 and is often used in combination with MALDI. TOF analysis is based on accelerating the ions with the same amount of energy to a detector. Since the ions have the same energy, but different masses, the ions reach the detector at different times. The smaller ions reach the detector first because of their higher velocity. The analyser measures time-of-flight and the mass is determined from the time of arrival of the ion.

The arrival time of an ion at the detector is dependent upon the mass (m), charge (z), and kinetic energy (KE) of the ion. Since the KE is equal to ½ (mv2), or the velocity v =√(2KE/m), ions will travel a given distance, d, within a time, t, where t is dependent on their m/z.

2.2.4.4 Other analysers

In addition to the three analysers used in this thesis, there are other analysers worth mentioning. The first is the magnetic sector analyser54. The double focusing magnetic sector mass analysers are the "classical" model to which other mass analysers are compared55. The ions pass through a magnetic sector in which a magnetic field is applied in a direction perpendicular to the direction of ion motion. To achieve better resolution, an electric sector focuses the ions according to their kinetic energy. The magnetic sector analysers give a very high reproducibility, and they also give the best quantitative results of all analysers. A drawback with this type of analyser is that it is very large and expensive. Furthermore, it is not well suited for pulsed ionisation methods (e.g. MALDI).

Another analyser is the Fourier-transform ion-cyclotron resonance (FTICR), where ions rotate around a magnetic field with a cyclotron frequency related to their m/z value. Due to the very accurate mass determination and resolving power achieved with this analyser56, 57 FTICR is a valuable tool in protein identification, reducing the risk of false positives or negative identifications58, 59. In addition to CID, the use of

(18)

electron capture dissociation (ECD)60 to fragment and achieve structural information of proteins further extends the capabilities of FTICR-MS.

2.2.4.5 Tandem mass spectrometry (MSMS)

MS/MS 52,61 is a crucial technique in many applications related to drug discovery and came to the fore with the development of the CID procedure. MS/MS enables structural information to be obtained for components of a mixture using multiple stages of mass analysis. An MS/MS instrument consists of more than one analyser, in practice usually two in series. The two analysers are separated by a collision cell, where an inert gas, e.g. argon, collides with the selected sample ions and induces fragmentation. Common combinations of analysers are the quadrupole ion traps, in which the stages are separated in time, and the quadrupole/TOF analysers, which perform MS/MS using sequential analysers.

Automated MS/MS coupled to online microcapillary LC is now a commonly used technique for sequence analysis of proteins62,63. This technique involves digestion of proteins followed by LC separation of the peptide fragments. Each eluted peptide is automatically selected for CID using automated data-dependent scan functions.

The tandem mass spectrum represents amino acid sequences of the peptide. The protein identification is performed by computer algorithms that correlate the MS data to sequences in protein databases. In paper I, we have applied this technique using a quadrupole ion trap. The protein identification from the resulting MS/MS spectra was performed by searching the NCBI nonredundant database with the program PEPFRAG64.

CID causes in principal two types of fragments along the peptide backbone. A nomenclature has been developed in order to describe product ions656667. According to this nomenclature, fragment ions containing the C-terminal end are classed as either x, y, or z, while fragment ions containing the N-terminal end are classed a, b, or c. A subscript indicates the number of residues in the fragment (Fig 4). In addition to the proton carrying the charge, c ions and y ions abstract an additional proton from the precursor peptide. CID spectra often result in the dominant fragmentation at the amide bonds in the polyamide backbone, producing ions of the type b or y.

(19)

A

B

Figure 4. A. Nomenclature of MS/MS fragment ions of peptides. B. The structures of the singly charged sequence ions b2 and y2.

2.3 THE DRUG DISCOVERY PROCESS

Below follows a description of the drug discovery process, with an emphasis on the involvement of MS. It is demonstrated how MS techniques can be employed in the different steps of the early drug discovery process in order to improve protein characterization as well as being an aid in protein-ligand binding studies.

2.3.1 Target identification and validation

Drug targets are macromolecules, mostly proteins, which play an important role in the onset or progression of a particular disease. The number of identified biological targets has been vastly expanded through genomics and proteomics. Pharmaceutical companies are advancing many of these newly identified potential targets into drug discovery. Many other potential drug targets have yet to be validated, and their roles in disease are not completely understood. As already mentioned, MS is a key technique in proteome studies for the identification of proteins. Historically the term proteomics, “protein complement expressed by a specific genome”, appeared in 1995 after the introduction of new methods in MS allowing ionization of biological macromolecules68. Proteins linked to a specific cellular phenotype, disease or metabolic state could now be screened on a large-scale basis. One effective approach for identifying proteins, employed in paper I, involves 1) protein separation by electrophoresis on either one or two-dimensional gels; 2) in-gel enzymatic digestion

(20)

of bands, or spots, of interest; 3) extraction of the resulting proteolytic peptides; 4) LC separation of the peptide mixture with on-line LC-MS/MS analysis of the peptide ions. The ion mass of each peptide and its fragment masses obtained are used to search the protein database. An alternative is to use MALDI mass fingerprinting, a technique where the masses of the resulting proteolytic peptides are used for the database search. There is extensive literature available on methods of MS and their application in proteomic studies69,70,71,72,73,74,75,76,77,78. Furthermore, the location of post-translational modifications (PTMs) can be determined by the MS techniques described above. In eukaryotic cells, PTMs are frequently occurring after translation, leading to an increased protein diversity. Common PTMs are glycosylation, phosphorylation, acetylation, methylation, and many more79, 80. In paper I, we discovered a PTM on the N-terminus of the identified protein. The initial Met had been removed and the following Ala residue was acetylated.

In 2001 the human proteome organization, HUPO81, was founded. Its major goals were to make an inventory of all human proteins, create protein atlases of cells, organs, tissues, define schemes of protein-protein interactions, develop special informational databases, and to search for specific markers of pathological processes.

In contrast to the human genome organization, HUGO82, founded for complete nucleotide sequencing of the human genome, HUPO did not announce precise aims and dates for termination of these projects, since it is difficult to predict the exact number of different proteins in cells under normal conditions and in various diseases.

2.3.2 Structural studies

Structural biology is the process of cloning, expressing, purifying and characterizing proteins in order to generate information about their structures, functions and interactions with drug candidates. The preservation of the native three-dimensional structure during analysis is critical to successful drug development. Methods to improve the understanding of protein structure are increasingly important for the pharmaceutical industry. In structure-based drug design, the three-dimensional structure of a drug target interacting with small molecules is used to guide drug discovery. The dominating technique to obtain high-resolution structural information on proteins is X-ray crystallography. The use of other biophysical methods, such as MS, dynamic light scattering (DLS), circular dicroism (CD), differential scanning calorimetry (DSC), analytical ultra centrifugation (AUC), and nuclear magnetic resonance (NMR) to rigorously measure characteristic protein properties, is a key step

(21)

to improve the quality of data on target proteins both before performing crystallisation experiments and before high throughput screening (HTS).

Thus, MS is an important tool for protein characterization83,84. For instance, ESI-MS is used to distinguish between folded and unfolded proteins. ESI charge-state distributions and the envelopes of charges provide information on solution conformations85 (Fig 5). The availability of basic sites on the protein is determined by the conformation of the protein as a more folded protein structure has less basic sites available for protonation than an unfolded protein.

BPase BB386 5 uM

10000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 3200 3400 3600 3800 m/z

100

%

2327

2133

1423 1600

1280 1707 1829 1970

2560

2844 3207

M11+

M10+

M12+

M13+

M16+

M20+

M9+

Unfolded protein

Folded monomer

Dimer

BPase BB386 5 uM

10000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 3200 3400 3600 3800 m/z

100

%

2327

2133

1423 1600

1280 1707 1829 1970

2560

2844 3207

M11+

M10+

M12+

M13+

M16+

M20+

M9+

Unfolded protein

Folded monomer

Dimer

Figure 5. ESI-MS on a 25.6 kDa enzyme in 20 mM ammonium acetate, pH 7. The charge states 20+ to 13+ correspond to unfolded protein, while the charge states 12+ to 9+ correspond to folded monomer. The presence of dimers can also be observed.

Another application for ESI-MS in protein characterization is to determine the stoichiometry of noncovalent complexes86. In paper II, we use nondenaturing ESI-MS to investigate the oligomerization of the peptide KFFEYNGKKFFE, containing two amyloidogenic sequences (KFFE). It was found that the peptide did not polymerise into fibrils but formed a stable 12-mer and multiples thereof. Noncovalent binding of ligands to proteins can also be observed by ESI-MS. It is not unusual to find fortuitous ligands, derived e.g. from the E. coli medium, bound to recombinant proteins (Fig. 6). These molecules may not be the physiological ligands, but they might be essentialfor stabilizing the active conformation of the protein, enabling its crystallization. Crystallization of proteins can be spoiled due to unfolding, degradation, or aggregation. The lack of crystallisation can also be due to the

(22)

presence of unstructured segments in the protein87 and a major challenge is to reveal these flexible sites. Although not exclusively true, structured domains are often the functional elements within proteins, and determination of the structure of these domains can therefore provide insights into structure-function relationships. Sequence alignment and secondary structure prediction of protein families may be of help to define structured domains. Although these methods are sufficient to permit an initial design of protein constructs for crystallization, questions often remain about the precise boundaries of folded domains.

28 28 28 28 28 28 28 28 28 28 28

28 28500 29000 29500 30000 30500

0 100

%

29715 29690

28969

29261

30010 30067

Free Protein 28970

Protein:720 Da ligand complex

Protein:745 Da ligand complex

0 10 20 30 40 50 60 70 80 90 100

000 28600 29200 29800 30400 3

0 10 20 30 40 50 60 70 80 90 100

000 28600 29200 29800 30400 3

0 10 20 30 40 50 60 70 80 90 100

000 28600 29200 29800 30400 3

Free Protein 28970

No complex

A. B.

28500 29000 29500 30000 30500

0 100

%

29715 29690

28969

29261

30010 30067

Free Protein 28970

Protein:720 Da ligand complex

Protein:745 Da ligand complex

0 10 20 30 40 50 60 70 80 90 100

000 28600 29200 29800 30400 3

0 10 20 30 40 50 60 70 80 90 100

000 28600 29200 29800 30400 3

0 10 20 30 40 50 60 70 80 90 100

000 28600 29200 29800 30400 3

Free Protein 28970

No complex

28500 29000 29500 30000 30500

0 100

%

29715 29690

28969

29261

30010 30067

Free Protein 28970

Protein:720 Da ligand complex

Protein:745 Da ligand complex

28500 29000 29500 30000 30500

0 100

%

29715 29690

28969

29261

30010 30067

Free Protein 28970

Protein:720 Da ligand complex

Protein:745 Da ligand complex

0 10 20 30 40 50 60 70 80 90 100

000 28600 29200 29800 30400 3

0 10 20 30 40 50 60 70 80 90 100

000 28600 29200 29800 30400 3

0 10 20 30 40 50 60 70 80 90 100

000 28600 29200 29800 30400 3

Free Protein 28970

No complex

0 10 20 30 40 50 60 70 80 90 100

000 28600 29200 29800 30400 3

0 10 20 30 40 50 60 70 80 90 100

000 28600 29200 29800 30400 3

0 10 20 30 40 50 60 70 80 90 100

000 28600 29200 29800 30400 3

Free Protein 28970

No complex

A. B.

Figure 6. MS characterization of the recombinant human nuclear receptor, NR5A2.

A) MALDI-TOF MS. Only the free, denatured protein can be observed. B) Non- denaturing ESI-MS of the same sample. Two additional peaks can be observed, corresponding to two protein-ligand complexes at masses ca 720 and 745 Da higher than the mass of the free protein. The ligands were identified as two phospholipids (phosphatidylglycerol and phosphatidyletanolamine).

Exposed protein surfaces and flexible sites are sensitive to proteolytic cleavage, while stably folded domains remain intact under protease-limited reaction conditions. Thus, limited proteolysis of the protein combined with MS can provide important complementary information to the alignment and prediction procedures obtained by computational chemistry88, 89, 90. MS analysis of the resulting fragments yields a cleavage ‘map’ that provides information on secondary, tertiary and, in multicomponent assemblies, quaternary structure.

MS-based hydrogen/deuterium (H/D) exchange is another MS technique with the ability to locate protein-binding sites and structure changes and can quantify dynamic and energetic parameters91, 92, 93, 94. Initial studies demonstrated that amide H/D exchange rates could be measured by MS95, 96. Katta and Chait96 demonstrated that folded and unfolded ubiquitin could be detected by their different H/D exchange rates. In paper II we used H/D exchange and MALDI-TOF MS to study the flexibility of two peptides.

(23)

2.3.3 Lead generation

Lead generation is the process of identifying potential drug compounds, or leads, that interact with a target with high potency and selectivity. During lead generation, assays for screening compound libraries are developed in order to identify biologically active compounds. This process of screening large numbers of compounds (100,000-300,000 or more compounds per screen) is called high throughput screening (HTS). Homogeneous assay formats such as scintillation proximity assay (SPA)97, fluorescence resonance energy transfer (FRET)98, time- resolved FRET (TR-FRET)99, and fluorescence polarization (FP)100 are today frequently used drug-screening techniques. They enable automation and assay miniaturization, resulting in labour and cost savings and are therefore ideal choices for HTS strategies. These types of assays measure the extent of inhibition caused by the compound. The measured IC50 value is the amount of compound resulting in a 50% inhibition of the protein activity. However, as a direct interaction between the drug compound and the target protein is not measured in these studies, it is advisable to let a compound undergo a screen-to-hit process before it can be considered to be a lead. In this process compounds that are considered amenable for chemical development are selected from the compounds identified as hits in a HTS effort.

Methods such as NMR, nondenaturing ESI-MS, and isothermal titration calorimetry (ITC) are employed to ensure that the HTS hits interact with the target through relevant mechanisms (e.g. by direct binding). For many reasons reversible binders are preferred over irreversible binders and it is hence important to investigate the mode of binding. The combination of proteolytic cleavage and MALDI-TOF MS is an excellent tool for detecting covalent binders and to identify the binding site (Fig. 7).

850 940 1030 1120 1210

0 5 10 15 20 25 30 35

850 940 1030 1120 1210

0 5 10 15 20 25 30

35 Protein + ligand A

Protein control

LVVECVMK

LVVECVMK

LVVECVMK + ligand

850 940 1030 1120 1210

0 5 10 15 20 25 30 35

850 940 1030 1120 1210

0 5 10 15 20 25 30

35 Protein + ligand A

Protein control

LVVECVMK LVVECVMK

LVVECVMK

LVVECVMK LVVECVMK + ligandLVVECVMK + ligand

Figure 7. MALDI-TOF MS on a tryptic digest of a human fatty acid binding protein, incubated in the absence (A) and presence (B) of a 311 Da compound (ligand A).

13000 1.7E+4

13000 1.7E+4

Mass (m/z)

% Intensity

Mass (m/z)

% Intensity

Mass (m/z)

% Intensity

Mass (m/z)

% Intensity

1026.52

1076.49 920.55

1068.52 957.53

920.51 1026.48

1231.45

957.50 1076.44

1060.57 993.47

917.52

1026.52

1076.49 920.55

1068.52 957.53

920.51 1026.48

1231.45

957.50 1076.44

1060.57 993.47

917.52

A

B

13000 1.3E+4

A

13000 1.3E+4

A A

(24)

If a ligand is covalently bound to the protein, it will remain attached to the protein after denaturation and proteolytic digestion. The proteolytic peptide containing the amino acid to which the ligand is bound will increase in mass corresponding to the mass of the ligand. In the example above, the protein was in excess of the ligand during incubation. This is why the proteolytic peptide LVVECVMK was observed both with and without ligand (Fig. 7B).

In the screen-to-hit process it should also be investigated whether the compound affects the stability of the protein in any unfavourable way. CD spectroscopy, DSC and MS are often used for this purpose. In paper IV, we describe a process for finding the mechanism of a drug-induced protein inactivation. The drug-inhibited protein was digested and analysed by MALDI-TOF MS, and we found that the destabilization of the protein was due to oxidation of the active site cysteine.

Many biophysical techniques can be used for the direct study of binding affinities of the protein-drug interaction. Surface plasmon resonance (SPR), NMR, isothermal titration calorimetry (ITC) and ESI-MS, are preferred methods that do not need labelling. In binding studies the dissociation constant (KD) is measured. KD indicates the strength of binding between the protein (M) and the ligand (L) at equilibrium:

KD = [M][L] / [ML]

MS plays an increasingly important role in the lead generation phase of the drug discovery process. Many different MS approaches for measuring direct binding affinities are presented in the literature43, 44, 45, 46, 47, 48, 49, 50. A drawback of many of the existing ESI-MS methods is that they are based on relative MS responses, which makes them valid only for complexes dominated by polar or ionic interactions. As MS measurements are performed in the gas phase, the difference in complex stability between the gas phase and solution must be considered. While the dissociation of polar or ionic complexes in the gas phase can be neglected, nonpolar complexes can be completely dissociated in the gas phase. In paper III, we demonstrate a novel approach using nondenaturing ESI-MS as a rapid tool for drug screening, where noncovalent complexes can be investigated regardless of the type of interactions involved. This is possible due to the introduction of an MS response factor, which reflects the MS response of a ligand-saturated protein.

(25)

2.3.4 Lead optimization

Lead optimization is a complex, multi-step process for refining the chemical structure of a compound to improve its drug characteristics. The goal is to produce a preclinical drug candidate. For optimizing a drug candidate, there are some important drug characteristics to be considered: potency, selectivity, toxicity, metabolism, and formulation.

2.3.5 Pre-clinical development

Prior to human clinical testing, a potential drug candidate must undergo extensive in vitro and in vivo studies to predict human drug safety. These studies investigate toxicity over a wide range of doses and the mechanism by which the drug is metabolized. Based on these results it is decided whether the drug candidate should enter clinical trials.

2.3.6 Clinical development

Clinical trials are performed on the drug candidates that are considered viable after the pre-clinical studies. The clinical evaluation on human subjects is performed on a variety of aspects of the drug, such as safety, effectiveness, and dosage.

Lead optimisation, pre-clinical development and clinical trials mainly involve analyses of the drug compound. Many MS techniques are used in these phases of the drug discovery process for the analysis of the drug or its metabolites.

In this thesis, however, only MS methods involving proteins are presented.

2.4 OTHER METHODS

As a complement to the described MS methods, several different techniques have been used in the present studies. A brief description of them is given below.

2.4.1 Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS- PAGE)

The purpose of SDS-PAGE is to separate proteins according to their size101,102. SDS is an anionic detergent that dissociates and unfolds proteins. The addition of a reducing agent such as DTT ensures that disulfide-dependent oligomers are separated

References

Related documents

Omvendt er projektet ikke blevet forsinket af klager mv., som det potentielt kunne have været, fordi det danske plan- og reguleringssystem er indrettet til at afværge

I Team Finlands nätverksliknande struktur betonas strävan till samarbete mellan den nationella och lokala nivån och sektorexpertis för att locka investeringar till Finland.. För

För att uppskatta den totala effekten av reformerna måste dock hänsyn tas till såväl samt- liga priseffekter som sammansättningseffekter, till följd av ökad försäljningsandel

Generella styrmedel kan ha varit mindre verksamma än man har trott De generella styrmedlen, till skillnad från de specifika styrmedlen, har kommit att användas i större

På många små orter i gles- och landsbygder, där varken några nya apotek eller försälj- ningsställen för receptfria läkemedel har tillkommit, är nätet av

Figur 11 återger komponenternas medelvärden för de fem senaste åren, och vi ser att Sveriges bidrag från TFP är lägre än både Tysklands och Schweiz men högre än i de

Det har inte varit möjligt att skapa en tydlig överblick över hur FoI-verksamheten på Energimyndigheten bidrar till målet, det vill säga hur målen påverkar resursprioriteringar

 Påbörjad testverksamhet med externa användare/kunder Anmärkning: Ur utlysningstexterna 2015, 2016 och 2017. Tillväxtanalys noterar, baserat på de utlysningstexter och