A Thesis Entitled Detection of Oxidized Lipids by Laser

Home , Hydrazide, Nitrogen laser

A Thesis

Entitled

Detection of Oxidized Lipids by Laser Desorption Ionization Time of Flight Mass

Spectrometry Using Hydrazine Containing Reagents

Mohammed Abdullah A Alyami

Submitted to the Graduate Faculty as partial fulfillment of the requirements for the

Master of Science Degree in Pharmaceutical Sciences

______Dr. Hermann von Grafenstein, Committee Chair

______Dr. Youssef Sari, Committee Member

______Dr. Zahoor Shah, Committee Member

______Dr. Amanda Bryant-Friedrich, Dean College of Graduate Studies

The University of Toledo

May-2018

This document is copyrighted material. Under copyright law, no parts of this document may be reproduced without the expressed permission of the author. An Abstract of

Detection of Oxidized Lipids by Laser Desorption Ionization Time of Flight Mass Spectrometry Using Hydrazine Containing Reagents

Mohammed Abdullah A Alyami

Submitted to the Graduate Faculty as partial fulfillment of the requirements for the Master of Science Degree in Pharmaceutical Sciences

The University of Toledo

May-2018

Since matrix assisted laser desorption ionization-time of flight mass spectrometry

(MALDI-TOF MS) was established, it has been advanced as an analytical technique to detect and identify a large range of molecules. Lipids have been analyzed intensively by

MALDI-TOF MS for several years. Because lipids are essential components of the body, any change in their structures can produce dramatic effects, either beneficial or detrimental. Oxidation of phospholipids is one of these types of changes. In the brain, it is found that formation of amyloid fibrils has a strong relationship to production of oxidized phospholipids (OxPLs) such as 1-palmitoyl-2-(9-oxo-nonanoyl)-sn-Glycero-3- phosphocholine (PoxnoPC). Due to addition of an aldehyde group in oxidized lipids, it may be possible to selectivity detect them by using fluorescent reagents containing a hydrazine functional group which was the ability to react with the carbonyl group. In the context of the MALDI-TOF MS analysis technique, 2,4-dinitrophenylhydrazine (DNPH) has been used as a reactive matrix holding the promise that this might allow selective detection oxidized lipids among un-oxidized lipids. However, in these studies unreacted excess DNPH was never removed, preventing definitive conclusions to be made as iii regarding the role of DNPH as a fluorescent tag as opposed to functioning as a matrix.

Oxidized phosphatidylcholine is derivatized at its carbonyl group with DNPH to form

2,4-dinitrophenylhydrazone (DNPhydrazone). To evaluate the “reactive matrix” concept implying that DNPhydrazone might be used as a selective detection of the analyte, we removed an excess amount of unreacted DNPH by using glyoxylic acid (GA) and added back different concentrations of DNPH. We found that the signal of DNPhydrazone was directly proportional to the DNPH concentration. Moreover, signals for both 1,2- dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) mixed with DNPH were detected.

DNPH functioned as a non-reactive matrix molecule as well as an analyte tag, although

DNPhydrazone demonstrated a higher signal than DPPC in all experiments. Utilizing

DNPH as a matrix molecule, we examined the difference of laser power required for compared with 2,5-dihydroxybenzoic acid (DHB). In the presence of DNPH, our results showed that DPPC needed lower laser intensity to be recorded. In contrast, DHB required more energy for detection of DPPC. Interestingly, we observed that a complex mixture of lipids influenced the DNPhydrazone and DPPC signals significantly. Further experiments showed that although DNPH did not allow matrix-free detection of oxidized lipids and was not selective in positive ion mode, selective detection of oxidized lipids is possible in negative ion mode. Here, DNPH appears to act as a charge-tagging reagent, adding a negative charge to the hydrazine moiety of DNPH.

Acknowledgements

I believe that words are not always enough to thank and appreciate Dr.

Grafenstein. Albeit he has much knowledge, he accepted and welcomed all my objections and debating. In fact, we worked as a team listening to each other's ideas, respectively, so it was more than a professor and student relationship. I am proud that I was one of his students.

I would like to thank Kirschbaum and Hanson for all of the support that helped me to finish my thesis. Dr. Kirschbaum fought for all students to have a new instrument, which gave all of us hope to complete our experiments. Dr. Hanson never has hesitated to help me and answered all my questions.

I appreciate that Dr. Sari and Shah accepted to be members of the committee.

They gave helpful comments that I am sure will help me in my future research.

Finally, without doubt, I would not be at this level of education and career success career without support from my mother and wife.

Table of Contents

Abstract ...... iii

Acknowledgements ...... v

Table of Contents ...... vi

List of Figures ...... viii

List of Abbreviations ...... x

1. Introduction……...... 1

1.1 Overview ……………………………………………………………………………1

1.2 MALDI ……………………………………………………………………………..2

1.2.1 Ionization Zone …………………………………………………...... 3

1.2.1.1 Hard ionization ……………………………………………………………3

1.2.1.2 Soft ionization …………………………………………………………….4

1.2.1.3 Matrices……………………………………………………………….…..5

1.2.2 Acceleration and flight zone…………………………………………….….…8

1.2.3 Detectors……………………………………………………………….……...8

1.3 Derivatized oxidized lipids ……...………………………………………………...10

1.3.1 Lipid ………………………………………………………………….……...10

1.3.2 DNPH …………………………………………………………………..……11

2. Materials and Method……………………………………………….…………….…12

2.1 Chemicals……………………………………………………...……………...…..12 vi

2.2 Preparations……………………………...……………………………..………....12

2.2.1 Phospholipids…………………………………………………………..…….12

2.2.2 Extracted lipid……………………………………………………………..…13

2.3 MALDI-TOF……………………………………………………………………...13

3. Results and Discussion ……………………………………………………………...14

3.1 Importance of additional amount of DNPH to work as a matrix………………....15

3.2 Comparison DNPH to DHB in consumed laser power………………………...…20

3.3 Effect of the lipid mixture on improving the signal of DNPhydrazone and

DPPC…………………………………………………………………………...... 21

3.4 The possibility of selected oxidized lipids when utilizing DNPhydrazone as a

charge tag…………………………………………………………………………23

4. Conclusions…………………………………………………………………………. 24

References ...... 26

Appendix ...………………………………………………………………………………29

vii

List of Figures

1-1 MALDI-TOF zones ...... 3

1-2 Hard ionization mechanism ...... 4

1-3 Chemical structure of 2,5-DHB ...... 7

1-4 Chemical structures of PoxnoPC and DNPH...... 10

3-1 The mechanism for reaction between DNPH and PoxnoPC ...... 15

3-2 Positive ion mode MALDI-TOF of PoxnoPC and DNPhydrazone ...... 15

3-3 Peak intensity average of DNPhydrazone in different DNPH concentrations ...... 16

3-4 Positive ion mode MALDI-TOF spectra of DNPhydrazone ...... 17

3-5 Scheme representation steps to remove non-react DNPH ...... 18

3-6 Mass spectra of aqueous and organic phases after adding with GA ...... 18

3-7 The mechanism for reaction between DNPH and GA ...... 18

3-8 Peak intensity average of DPPC and DNPhydrazone after adding GA ...... 19

3-9 Peak intensity average of DPPC and DNPhydrazone in the separate experiment .19

3-10 Positive ion mode MALDI-TOF spectrum of DPPC mixed with 50mM DNPH ..19

3-11 Different laser power levels needed to detect DPPC ...... 20

3-12 Positive ion mode spectrum of DPPC mixed with different components ...... 20

3-13 Effect of the lipid mixture on the signal of the DNPhydrazone ...... 21

3-14 Effect of the lipid mixture on the signal of the DPPC ...... 22 viii

3-15 MALDI spectra of DPPC and DNPhydrazone in the lipid mixture...... 22

3-16 Negative charge formed in hydrazone moiety after removing H+ ………………23

3-17 Negative ion MALDI-TOF mass spectrum of DNPhydrazone………………….24

3-18 Negative ion MALDI-TOF mass spectrum of DPPC……………………………24

List of Abbreviations

amu ……………………atomic mass units

COX……………………Cyclooxygenases

DHB…………………...2,5-Dihydroxybenzoic acid DNPH………………….2,4-dinitrophenylhydrazine DNPhydrazone………...2,4-dinitrophenylhydrazone DPPC…………………..1,2-dipalmitoyl-sn-glycero-3-phosphocholine

EI ………………………Electron ionization

GA ……………………..Glyoxylic acid

IR……………………….Infrared

LOX…………………....Lipoxygenases m/z ……………………..Mass-to- charge ratios MS ……………………..Mass spectrometry MALDI ……………...... Matrix assisted laser desorption ionization MCP ………………...... Microchannel plate

N2………………………Nitrogen

OxPLs………………….Oxidized phospholipids

PLs …………………….Phospholipids PoxnoPC……………….1-palmitoyl-2-(9-oxo-nonanoyl)-sn-Glycero-3-phosphocholine

ROS ……………………Reactive oxygen species

S/N …………………….Signal to noise x

UV ………………….....Ultraviolet

Nd:YAG……………….Neodymium-doped yttrium aluminum garnet

Chapter 1

Introduction

1.1.Overview

Mass spectrometry (MS) is an analytical technique applied in many fields of science to detect or identify samples relying on measuring mass-to-charge ratios (m/z).

However, in the past, MS had the disadvantage of producing abundant fragments when using hard ionization such as electron ionization (EI). Fortunately, Hillenkamp and Karas laid the foundation to advance MS to what was later called matrix assisted laser desorption ionization (MALDI). After shooting an analyte with 266 nm laser light, they successfully ionized a peptide by mixing the analyte with a compound named a “matrix.”

Moreover, they noticed that their procedure was a form of soft ionization technique, protecting the analyzed sample from fragmentation. In the same year, Koichi Tanaka with his fellows improved previous techniques by using a nitrogen laser allowing them to ionize biological macromolecules (Hosseini and Martinez-Chapa 2017).

Although MALDI-MS has been utilized for decades chiefly for proteomic analysis, lipid analysis has only recently been advanced by using MALDI MS. The oxidation of lipids, especially phospholipids (PLs), which are located mainly in cellular

membranes, takes place by changing their structures resulting in the addition of aldehyde groups (Engel and Schiller 2017; Spener et al. 2003). Because hydrazine agents derivatize aldehydes and ketones, they can be used to analyze oxidized lipids by MALDI

(Teuber et al. 2012; Nimptsch et al. 2013). However, it is not clear whether a derivatizing reagent such 2,4-dinitrophenylhydrazine (DNPH) could be utilized to selectively tag and desorb analytes or if it acts as a matrix due to a large excess of unreacted reagent. Hence, in this study, we investigated in detail whether DNPH can be used as a “reactive matrix” molecule for the selective detection of carbonyl-containing analytes or if it works primarily as “non-reactive matrix” molecule.

1.2. MALDI

Accelerating the time of analysis, offering high sensitivity measurement of an analyte in the attomolar range, and providing mass accuracy of target compounds are remarkable features of MS (Wang et al. 2016; Fuchs and Schiller 2009).

In fact, the ability of MALDI-MS to detect small molecules, macromolecules, non-volatile and labile molecules has made it as a universal technique (Marvin, Roberts, and Fay 2003; Wang et al. 2016; Hillenkamp 2007). In 1991, a commercial instrument became available on the market; proteins and other biochemical molecules, not lipids, were mainly investigated by MALDI (Stults 1995).

Interestingly, before MALDI was established, the time-of-flight (TOF) technique was invented and used with MS. TOF measured the time that charged ions take to fly from a source to a detector after acceleration in an electric field. Because the speed of the ions depends on their m/z ratio, it should be possible to identify the molecular weights of 2

the analytes by making time measurements. Nevertheless, combining MS with TOF did not show advantages when comparing it with conventional MS. The main drawback of that combination was the insufficient resolution. Lately, MALDI MS - commonly conjoined with a TOF mass analyzer - has exhibited very high sensitivity, when using reflector mode.

MALDI-TOF-MS can be divided up into three zones: ionization zone, acceleration zone, and flight zone. Figure (1:1) portrays the components with details of

MALDI-TOF-MS (Hosseini and Martinez-Chapa 2017).

Figure 1-1 MALDI-TOF zones

1.2.1. Ionization Zone

1.2.1.1. Hard Ionization

Molecules can be ionized by removing one electron from the analyte molecules through hitting them with fast electrons. This process takes place in the gas phase; hence, electron ionization (EI) is not desirable for use with high and low-volatile compounds

(Fuchs, Süß, and Schiller 2010). Furthermore, the detection of the molecular ions would

Figure 1-2 Hard ionization mechanism be unachievable it plentiful fragment ions are generated; therefore, EI MS is not appropriate for analysis of lipid mixtures (Fuchs, Süß, and Schiller 2010).

1.2.1.2. Soft Ionization

As mentioned previously, soft ionization techniques provide a significant feature which minimizes the fragmentation of the analytes of interest during the ionization

(Hosseini and Martinez-Chapa 2017). Many ion formation mechanisms have been thoroughly discussed in the literature (see (Zenobi and Knochenmuss 1998)). One of these pathways is referred to as the pre-charging mode. The analytes could be ionized after co-crystallization with the matrices on the target plate, and the laser converts the analyzed sample from the solid to the gas phase. Unlike pre-charging mode, in the photoreaction mode, the laser irradiation plays a pivotal role in the ionization mechanism

(Wang et al. 2016). The sample crystal is exposed to the laser pulse; the matrix carrying the analyte absorbs the photon energy, sublimating in the process. At the same time, the analyte gains ions such as H+ after collision with an acidic matrix molecule to achieve a form called “cationized molecules (Hosseini and Martinez-Chapa 2017; Fuchs, Süß, and

Schiller 2010; Fuchs and Schiller 2009). Alternatively, the analyte could be losing a proton to generate an anionic analyte. The acidities of the matrix and the analyte play an important role in determining the charge of the analyte. However, positive ions are 4

widely used in MALDI-TOF- MS compared with negative ions due to the higher

sensitivity (Fuchs, Süß, and Schiller 2010).

Astonishingly, 20 years after establishing MALDI, it is still not fully understood

how the ionization process works (Hillenkamp 2007). Variations in preparing the

analyzed samples, different laser wavelengths being used, and other experimental

parameters such as sample temperature are variables complicating the identification of

the real mechanism of ionization (Zenobi and Knochenmuss 1998).

1.2.1.3. Matrices

The primary function of a matrix is to generate ions by absorbing the energy

coming from the laser. The matrix is added to the analyte at a ratio of 100 to 100000:1,

(Fuchs, Süß, and Schiller 2010; Hillenkamp 2007). In the context of choosing a suitable

matrix for experiments, there are general features that should be found in matrices. First,

the ability of the matrix to produce high signal to noise (S/N) ratios for the target peaks.

Second, minimizing the background due to matrix molecules should be achievable.

Frequently, matrix molecules can form clusters in the gas phase leading to signals farther

away from their exact molecular weights (Fuchs, Süß, and Schiller 2010; Škrášková and

Heeren 2013). Matrix isolation, prevention of the analyte from forming clusters, is the

third feature to be considered when choosing a matrix (Hosseini and Martinez-Chapa

2017). Finally, absorbing the laser light at the wavelength of the laser with a high

absorption coefficient (α) is a critical feature of matrices (Fuchs, Süß, and Schiller 2010).

The matrix absorbs the laser light according to Beer’s law (Eq.1:1);

-αz H = H0 *e

where H and H0 are the laser’s light intensity in the sample depth (z) and surface, respectively (Hillenkamp 2007). Moreover, the matrix protects the analyzed sample from the laser irradiation (Hosseini and Martinez-Chapa 2017). Different wavelengths of the lasers can be applied in MADL-TOF-MS. The typical laser used is UV such as nitrogen

(N2) at 337 nm or neodymium-doped yttrium aluminum garnet (Nd:YAG) at 355nm

(Škrášková and Heeren 2013; Hosseini and Martinez-Chapa 2017). Interestingly, infrared

(IR) lasers provide softer ionization than UV lasers and are more appropriate when applied in the absence of matrix (Hosseini and Martinez-Chapa 2017). Moreover, metastable fragmentation is less observed when using IR laser (Zenobi and Knochenmuss

1998).

Because the N2 laser is the primary laser used in MALDI, organic matrices comprising an aromatic ring are used mainly due to the ability to absorb ultraviolet (UV) light. In fact, increasing in the absorption coefficient of a matrix at the wavelength of the laser leads to improvement of the ionization process; hence, 2,5-dihydroxybenzoic acid

(DHB) is chiefly applied as a UV MALDI matrix rather than other isomers of DHB due to the significant absorption of UV light. Moreover, organic acid matrices containing carboxylic-acid groups enhance the generation of H+ to combine with the analyte (Fuchs,

Süß, and Schiller 2010).

Figure 1:3 Chemical structure of 2,5-DHB

In the context of lipid analysis, 2,5-DHB is most frequently used for detecting

PLs, and para-nitroaniline (PNA) can be an alternative matrix to 2,5-DHB in detecting some types of PLs in negative ion mode (Estrada and Yappert 2004). Interestingly, most matrices can be dissolved in organic solvents instead of having to add water, exhibiting more homogeneity of the sample analyzed compared with proteins. In fact, some lipids can form crystals when dissolved in the same solvent system as the matrix (Schiller et al.

1999; Fuchs and Schiller 2009). However, due to the lack of complete homogeneity of the analyzed sample, utilizing MALDI as a quantitative technique is questionable (Fuchs and Schiller 2009). The distribution the analyte in the matrix is not perfectly even. In other words, different zones of the analyzed sample spot could produce non-uniform signals. A multitude of experimental conditions might contribute to the signal variability

(Duncan, Roder, and Hunsucker 2008). For instance, the time that the angiotensin II peptide and matrices need to form crystals strongly influences homogeneity. Rapid crystallization leads to much more homogeneity than a slow one. Moreover, sample preparation is critical in terms of reproducibility (Nicola et al. 1995). A change of the analyte concentration increases or decreases the ionization process. The result of that is an alteration of the signal (Duncan, Roder, and Hunsucker 2008). For minimizing the 7

irreproducibility, specific recommendations might be implemented such as adding mono-

ammonium phosphate in the matrix to reduce sodium and potassium being added to the

sample ion. More details are found in (Wang and Giese 2017).

1.2.2. Acceleration and flight zone.

After ablation and ionization of the analyzed sample, ions which may be the

ionized analyte, matrix, or fragments are speeded up in an electric field of high voltage to

pass a charged grid (Hosseini and Martinez-Chapa 2017; Fuchs, Süß, and Schiller 2010).

Subsequently, they are entering a field-free space, the “flight zone,” where ions continue

flying to reach a detector. This feature allows a TOF analyzer to separate ions of different

molecular mass depending on time measurement (Wang et al. 2016). MALDI-TOF-MS is

more efficient than other MS techniques due to a theoretically limitless mass range

(Fuchs, Süß, and Schiller 2010). Because the laser pulse defines time zero at the time-of-

flight measurement, resolution critically depends on the shortness of the laser pulse

(Wang et al. 2016).

1.2.3. Detectors

The detector calculates m/z according to the equation in Eq.1:2.

m/z = 2ɋE(t/d) 2

Where m is the mass of the ionized molecule, z is the number of electrons removed or

proton added during the ionization. E is the accelerating voltage, ɋ is the elementary

charge, and t symbolizes the time that ions need to reach the detector. The distance that 8

the analyzed sample travels in the flight zone to hit the detector is denoted by the d symbol. The exact time of generation of secondary electrons depends on the molecular mass of ions. In other words, ions of low mass reach a higher velocity than ions of high mass (Wang et al. 2016); therefore, according to the above equation, low m/z ions spend a short time in the flight zone and are be detected first (Hosseini and Martinez-Chapa

2017). Detectors determine m/z of the analyte through a microchannel plate (MCP). After primary ions collide with the surface of the detector, MCP generates secondary electrons and multiplies them further (Wang et al. 2016).

In modern MALD-TOF instruments, ions move towards the detector either in the short “linear” or long “reflector” flight path (Fuchs, Süß, and Schiller 2010). Even though ions with the same m/z have theoretically the same amount of energy of motion, the ions have different capability in absorbing the laser irradiation. Hence, ions of same m/z do not necessarily have the exact same travel time to arrive at the detector. For that reason, the linear detector exhibits a distribution of m/z (Hosseini and Martinez-Chapa 2017).

Despite the insufficient resolution, it is sometimes more beneficial using the linear detector measuring large molecules because the sensitivity is more relevant in this situation. On the other hand, achieving high resolution is carried out by the reflector detector through minimizing the heterogeneity of the ions’ velocities (Fuchs, Süß, and

Schiller 2010).

1.3. Derivatized oxidized lipids

1.3.1. Lipids

PLs are mainly found in cell membranes with many significant functions in biological systems. Hence, structurally changing PLs by oxidative stress generates many products with different biological effects on the body (Domingues, Reis, and Domingues

2008). The oxidation of the PLs takes place chiefly in the unsaturated chains and is the result of two types of mechanisms: enzymatic and non-enzymatic pathways. Releasing reactive oxygen species (ROSs) such as nitrogen dioxides by non-enzymatic mechanisms are the main mediators of inflammation in diseases (O'Donnell 2011). Due to the high amount of polyunsaturated fatty acids in neurons, ROSs oxidize them such as 1- palmitoyl-2-linoleoyl-phosphocholine to 1-palmitoyl-2-(9`-oxononanoyl)-sn-glycero-3- phosphocholine (PoxnoPC). Because of the aldehyde moiety in PoxnoPC, proteins or peptides may interact with aldehydes and aggregate into amyloid fibrils. Amyloid fibrils are an important marker of neurodegenerative disorders such as Alzheimer’s disease

(Mahalka, Maury, and Kinnunen 2011).

In the enzymatic pathway, cyclooxygenases (COX) and lipoxygenases (LOX) are protein families involved in generating enzymatic oxidation products such as prostaglandins (O'Donnell 2011).

(a) (b)

Figure 1-4 Chemical structures of (a) PoxnoPC and (b) DNPH

1.3.2. DNPH

Since 1894, DNPH has been used as a derivatizing agent for the identification of aldehydes and ketones (Vogel, Buldt, and Karst 2000; Brady and Elsmie 1926). After reacting with carbonyl compounds, in particular with carbonyls in aliphatic aldehydes,

DNPH is converted to dinitrophenylhydrazone adducts which are stable and often finely crystalline. The derivative can form in the presence of low amounts of aldehyde.

However, the main disadvantage of DNPH is the low solubility in some solvents (Brady and Elsmie 1926).

During the oxidation of PLs, an aldehyde group is added (Engel and Schiller

2017). Therefore, DNPH is used to detect oxidized lipids in MALDI-TOF-MS. The properties of DNPH fit with some of desired characteristics of a matrix (see matrix section). DNPH absorbs UV light at 349 nm, so it is suitable to use with N2 and Nd: YAG lasers. Moreover, DNPH exhibits stability under high vacuum. Derivatization of oxidized

PLs (OxPLs) by 180 atomic mass units (amu) facilitates detection in the form of hydrazone (Teuber et al. 2012). Thus, identifying plasmalogens in lipid mixtures by acidic DNPH is achievable; better results were obtained in the negative mode of MALDI

(Nimptsch et al. 2013). In contrast, DNPH derivatization of carbonyls in corticosteroids shows excellent result in the positive mode (Flinders et al. 2015).

Chapter 2

Materials and Methods

2.1. Chemicals

Solvents were obtained from Fisher Chemical in the highest purity available.

DNPH in 30% water was obtained from Spectrum Chemical MFG. 2,5-DHB was purchased from Ricca Chemical Company. Glyoxylic acid monohydrate 98% (GA) was obtained from Fisher Scientific. Buffer solution pH 7, dibasic sodium phosphate, and potassium dihydrogen phosphate were obtained from Honeywell Corporation. Non- oxidized PL 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) as 25 mg powder, and oxidized lipid PoxnoPC as 1mg powder were purchased from Avanti Polar Lipids. All materials were used as received without further purification.

2.2. Preparation

2.2.1. Phospholipids

Under argon gas, 25 mg of DPPC and 1 mg of PoxnoPC were diluted with chloroform to 10 mg/ml and 1 mg/ml, respectively. Both lipids had final concentrations of 0.025 mM in all samples prepared unless indicated otherwise.

2.2.2. Extraction of lipid

Lipid extraction from adipose tissue was performed according to (Folch, Lees, and Sloane Stanley 1957). Briefly, 900 mg of mouse abdominal fat was homogenized in

6 ml of methanol, and then 12 ml of chloroform was added and the sample was agitated for 20 minutes at room temperature. The mixture was filtrated by Whatman Paper grade 1 and washed three times with 0.9% NaCl. After centrifugation for 5 minutes at 2000 rpm, the upper phase was removed, and the bottom phase was evaporated in N2 gas. 500 mg of lipid was dissolved in chloroform and methanol (2:1) to 1 mg/ml.

2.3. MALD-TOF-MS

Serial dilutions of DNPH (75, 50, 30, 20, 10, 5, 2.5, 1, and 0.5 mM) dissolved in acetonitrile were prepared without adding acid. However, at 75 mM DNPH was only partially dissolved. An equal volume of PoxnoPC or DPPC dissolved in chloroform and

DNPH dissolved in acetonitrile at indicated concentrations were applied for the whole experiment. To remove excess unreacted DNPH, 200 mM of glyoxylic acid dissolved in deionized water was added to the mixture of 0.1 mM PoxnoPC, 0.1 mM DPPC, and 50 mM of DNPH and mixed thoroughly until the color changed to yellow. Subsequently, the mixture was washed with 100 mM phosphate buffer several times until the bottom phase was close to colorless. The rationale for this procedure is that GA reacts with excess

DNPH, adding a carboxylic acid which renders DNPH water-soluble.

2,5-DHB was used as a 100 mM solution in acetonitrile, methanol, and 0.1% TFA

(40:10:1). 1 μl of 2,5-DHB was deposited on the sample plate; after the matrix dried, 1 μl

of the analyzed sample was added on top and mixed thoroughly. The mixture was introduced to MALDI-TOF after drying.

All mass spectra were recorded using a Microflex MALDI mass spectrometer

(Bruker Daltonics, Bremen, Germany). The system utilizes a pulsed 60 Hz nitrogen laser

(emission, 337 nm; delay, 120 ns; accelerating voltage 19 kV). The instrument was used in reflector positive mode and without applying gated matrix suppression. The laser fluence was used at the minimum power to obtain satisfactory spectral mass resolutions,

S/N ratios, and peak intensities. All mass spectra parameters were defined by Flex

Analysis software (Bruker Daltonics) and calibrated using PeptideCalibStandard mono.

For each spectrum, 200 shots were repeated four times on 1μl of the analyte spotted onto the MALDI target (MPS ground steel BS (MicroScoutTarget) plate; Bruker Daltonics,

Bremen, Germany) with random walk mode set to 20 shots per position.

Chapter 3

Results and Discussion

This study aimed to determine whether MALDI-TOF could detect the

DNPhydrazone without the excess of DNPH. Moreover, DNPH was under investigation as a non-reactive matrix for the lipid mixtures and non-oxidized lipids such as DPPC.

Because PoxnoPC has an aldehyde functional group, the hydrazine nucleophile attacks the carbonyl carbon to add R–NH; subsequently, H2O molecules are eliminated

(Fig 3-1) (Kadam et al. 2012).

3.1. Importance of additional amount of DNPH to work as a matrix

In Fig 3:2, PoxnoPC was detected by DHB at 650.4 m/z. After reacting with 20 mM DNPH, the peaks of DNPhydrazone were recorded at 830.5, 852.5, and 868.4 m/z corresponding to the H+, Na+, and K+ adduct. In some cases, 815 m/z peak was detected which it might be another DNPhydrazone peak (Teuber et al. 2012).

Treating PoxnoPC with different concentrations of DNPH showed a concentration-dependent signal detecting DNPhydrazone (Fig. 3- 3).

This could be due to varying amounts of DNPhydrazone or due to varying amounts of non-reacted

DNPH as a matrix. To determine if

Figure 3-3 Peak intensity average of DNPhydrazone peaks excess unreacted DNPH is working as in different concentrations of DNPH at 10% laser intensity a matrix desorbing DNPhydrazone, we first spotted DHB onto the target plate and added a mixture of lower concentration DNPH (0.5mM) and PoxnoPC. The result was an abundance of DNPhydrazone peaks observed in the presence of DHB (Fig 3-4).

To separate the effect of DNPH acting as a “non-reactive” matrix from that of

DNPhydrazone formation, we removed unreacted DNPH by adding GA and washing with buffer (Fig 3-5). GA is an aldehyde reacting with the DNPH (Fig 3-6). This adds a carboxylic acid group to DNPH, making it water-soluble and allowing it to be washed away. After examining both phases (organic and aqueous phase) by DHB, it was found that the majority of DNPhydrazone was in the organic phase “bottom phase” (Fig 3:7).

An equal volume of bottom phase was aliquoted into nine vials of different concentrations of DNPH. Then, all samples were introduced to MALDI-TOF. The strength of DNPhydrazone’s signal was directly proportional to DNPH concentration (Fig

3:8).

(a)

(b)

(c)

Figure 3-4 Positive ion MALDI-TOF mass spectra of (a) DNPhydrazone without DHB at 10% of laser intensity, (b) DNPhydrazone without DHB at 100% of laser intensity, and (c) DNPhydrazone with DHB at 43% of laser intensity

Figure 3-5 Scheme representation the steps to remove non-react DNPH (a) mixed of PoxnoPC, DPPC, and 50 mM DNPH, (b) 200 mM GA and buffer added and mixed (c) tested both phases by used DHB, and transferred the organic phase into different DNPH concentrations

Figure 3-6 The mechanism for the reaction between DNPH and GA

(a)

(b)

Figure 3-7 Mass spectra of (a) aqueous phase, and (b) organic phase with the matrix used in both

Moreover, a decreasing DNPH amount 15000 DNPhydrazone

DPPC led to weakening the ability of DNPH to 12000 act as a matrix to detect DPPC. 9000

In the absence of DNPhydrazone, 6000

detection of DPPC, which does not have Average Peak Intensity 3000

0 an aldehyde group to react with 0 5 10 15 20 25 30 35 40 Final Concentraion of DNPH (mM) hydrazine group, was affected Figure 3-8 Peak intensity averages of DPPC and DNPhydrazone after washing with GA and retreated substantially by decreasing the DNPH with DNPH

16000 concentration (Fig 3-9). At all DNPhydrazone

DPPC concentrations of DNPH, both 12000

DNPhydrazone and DPPC were detected. 8000 Although not all PoxnoPC used was fully 4000 reacted with DNPH and some of it Average Peak Intensity

0 converted to a carboxylic acid (665.1 m/z) 0 5 10 15 20 25 30 35 40 Final Concentraion of DNPH (mM) (Teuber et al. 2012), the signal of Figure 3-9 Peak intensity averages of DPPC and DNPhydrazone in separate experiment DNPhydrazone was always higher than DPPC

(Fig 3-8,9). This phenomenon could come from hydrazone formation.

As with DNPhydrazone peaks, not only the H+ adduct of DPPC (734.7 m/z) was noticeably detectable but also the Na+ and K+ adduct (757.7, and 772.7 m/z) (Fig 3-10). Figure 3-10 Positive ion mode MALDI-TOF spectrum of DPPC mixed with 50 mM DNPH

3.2. Comparison of laser power 40000 DPPC-DNPH-Tissue required for detection of DNPH DPPC-DPNH to DHB DPPC-DNPH-DHB 30000 DPPC-DHB During the experiment, we 20000 observed that DNPH as a matrix

10000 needs a lower laser power than DHB Average Peak Intensity

0 to show sufficient signal of the 0% 10% 20% 30% 40% 50% Laser intensity % analytes (Fig 3-11,12). At 10%, Figure 3-11 Different laser powers needed to detect DPPC DPPC was detectable with a signal

(a)

(b)

(c)

(d)

Figure 3-12 Positive ion MALDI-TOF mass spectra of (a) DPPC mixed with 50 mM DNPH, (b) DPPC DNPH (50 mM) in the presence of DHB, (c) DPPC with DHB only, and (d) DPPC mixed with the lipid tissue and DNPH. 20

intensity around 5429 a.u. average when using 50 mM DNPH. In the presence of 100 mM DHB only 121 a.u. average was recorded. After using DHB with the mixture of

DPPC and 50 mM DNPH, the curve was shifted toward to the matrix line (blue line). The reason for that is still unknown, but we observed that DNPH forms a very thin layer compared to DHB. However, at high laser level, the signal of DPPC mixed with DNPH went down. In contrast, a signal of DPPC increased with laser intensity when using DHB.

Interestingly, when DPPC was mixed with a complex lipid mixture, the signal was read at

12881 a.u. average at 10% (see below).

3.3. Influence of the lipid mixture in improving the signal of DNPhydrazone and

DPPC 25000 DNPhydrazone-Tissue DNPhydrazone and DPPC DNPhydrazone 20000 exhibited robust signals when they 15000 were in a complex mixture of lipids 10000 (Fig 3-13,14,15).When DPPC was Peak Intensity Average Peak Intensity 5000 mixed with the tissue lipids, DNPH 0 0 5 10 15 20 25 30 35 40 revealed a range of peaks (800-1000 Final Concentraion of DNPH (mM) Figure 3-13 Effect of the lipid tissue on raising the m/z) that were related to the lipids. signal of the DNPhydrazone

Moreover, even though these peaks were not identical with others that were exhibited when using DHB and the lipid sample, they are in the same range. However, peaks of the lipids almost faded when DNPhydrazone was present (Fig 3-15c); it might be that

DNPhydrazone peaks were in the same range of lipids peaks.

Because DPPC does not form a 25000 DPPC-Tissue DPPC hydrazone, we assumed that this 20000 phenomenon is more relevant to non- 15000 reacted DNPH and the lipid mixture 10000

characteristics. We hypothesize that an Average Peak Intensity 5000 emulsification process takes place in 0 0 5 10 15 20 25 30 35 40 the mixture leading to uniform Final Concentraion of DNPH (mM) Figure 3-14 Effect of the lipid tissue on raising the signal of the DPPC

(a)

(b)

(c)

Figure 3-15 Positive ion MALDI-TOF mass spectra of (a) the lipid tissue in the presence of DHB, (b) DPPC mixed with lipid tissue and DNPH (50 mM), and (c) PoxnoPC mixed with lipid tissue first and added DNPH (50 mM)

distribution of the analytes. According to (Weschayanwiwat, Scamehorn, and Reilly

2005), small chain of PLs (6 to 12 carbons) could work as surfactants. However, we still do not know the role of DNPH in improving the signal.

3.4. The possibility of selected oxidized lipids when utilizing DNPhydrazone as a charge tag.

Abstracting H+ from hydrazone moiety facilitates DNPhydrazone to be detected in negative ion mode (Fig 3:16) (Nimptsch et al. 2013; Berdyshev 2011). Mixing

PoxnoPC with tissue lipids and adding 50 mM DNPH exhibited a significant peak of

DNPhydrazone at 828.7 m/z (Fig 3:17). However, the nature of the 775.1 m/z peak is still unknown. In contrast, DPPC mixed with tissue lipids did not show a signal when using

50 mM DNPH (Fig 3:18).

Figure 3-16 Negative charge formed in hydrazone moiety after removing H+

Figure 3-17 negative ion MALDI-TOF mass spectrum of DNPhydrazone (828.786 m/z)

Figure 3-18 negative ion MALDI-TOF mass spectrum of DPPC

Chapter 4

Conclusion

For several decades, DNPH has been used as a tool to identify aldehydes and ketones. DNPH as a possible candidate to detect only oxidized lipids using MALDI-TOF 24

has not been studied thoroughly. In our study, DNPH could work as a matrix with DPPC and DNPhydrazone. When unreacted DNPH was removed after DNPhydrazone formation, the signal disappeared. Our results therefore suggest that it is unlikely that

DNPH acts as a “reactive matrix” allowing specific detection of aldehyde-containing lipids as it has been suggested (Teuber et al. 2012). However, DNPH appears to be a superior matrix compared with DHB, working at lower laser intensities and producing more homogeneous signal intensity. Although DNPhydarzone exhibits significant signal in the lipid mixture, it might be difficult to interpret this as selective detection because

DPPC also showed a significant peak in the lipid tissue. This study is an important step towards solving the challenging task of creating a method to detect oxidized lipids in mixtures found in the biological systems. Although DNPH was acting as a matrix in positive ion mode without selectivity for oxidized lipids, selective detection was possible in negative ion mode. Here, DNPH functions as a charge tag rather than a fluorophore facilitating selective desorption based on light absorption.

References

Berdyshev, Evgeny V. 2011. 'MASS SPECTROMETRY OF FATTY ALDEHYDES', Biochimica et biophysica acta, 1811: 680-93.

Brady, Oscar L., and Gladys V. Elsmie. 1926. 'The use of 2:4-dinitrophenylhydrazine as a reagent for aldehydes and ketones', Analyst, 51: 77-78.

Domingues, M. Rosário M., Ana Reis, and Pedro Domingues. 2008. 'Mass spectrometry analysis of oxidized phospholipids', Chemistry and Physics of Lipids, 156: 1-12.

Duncan, M. W., H. Roder, and S. W. Hunsucker. 2008. 'Quantitative matrix-assisted laser desorption/ionization mass spectrometry', Brief Funct Genomic Proteomic, 7: 355-70.

Engel, Kathrin M., and Jürgen Schiller. 2017. 'A comparison of PC oxidation products as detected by MALDI-TOF and ESI-IT mass spectrometry', Chemistry and Physics of Lipids, 203: 33-45.

Estrada, R., and M. C. Yappert. 2004. 'Alternative approaches for the detection of various phospholipid classes by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry', J Mass Spectrom, 39: 412-22.

Flinders, B., J. Morrell, P. S. Marshall, L. E. Ranshaw, and M. R. Clench. 2015. 'The use of hydrazine-based derivatization reagents for improved sensitivity and detection of carbonyl containing compounds using MALDI-MSI', Anal Bioanal Chem, 407: 2085-94.

Folch, J., M. Lees, and G. H. Sloane Stanley. 1957. 'A simple method for the isolation and purification of total lipides from animal tissues', J Biol Chem, 226: 497-509.

Fuchs, Beate, and Jürgen Schiller. 2009. 'Application of MALDI-TOF mass spectrometry in lipidomics', European Journal of Lipid Science and Technology, 111: 83-98.

Fuchs, Beate, Rosmarie Süß, and Jürgen Schiller. 2010. 'An update of MALDI-TOF mass spectrometry in lipid research', Prog Lipid Res, 49: 450-75.

Hillenkamp, F. and Karas, M. . 2007. 'The MALDI Process and Method. In MALDI MS ' in F. Hillenkamp and J. Peter‐Katalinić (ed.).

Hosseini, Samira, and Sergio O. Martinez-Chapa. 2017. 'Principles and Mechanism of MALDI-ToF-MS Analysis.' in Samira Hosseini and Sergio O. Martinez-Chapa (eds.), Fundamentals of MALDI-ToF-MS Analysis: Applications in Bio-diagnosis, Tissue Engineering and Drug Delivery (Springer Singapore: Singapore).

Kadam, Sachin S, ST Tambe, ND Grampurohit, and DD Gaikwad. 2012. 'Review article on chemical importance of Brady's reagent', Int J Res Pharm Chem, 2: 1086-92.

Mahalka, A. K., C. P. Maury, and P. K. Kinnunen. 2011. '1-Palmitoyl-2-(9'- oxononanoyl)-sn-glycero-3-phosphocholine, an oxidized phospholipid, accelerates Finnish type familial gelsolin amyloidosis in vitro', Biochemistry, 50: 4877-89.

Marvin, Laure F., Matthew A. Roberts, and Laurent B. Fay. 2003. 'Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry in clinical chemistry', Clinica Chimica Acta, 337: 11-21.

Nicola, Anthony J., Arkady I. Gusev, Andrew Proctor, Edwin K. Jackson, and David M. Hercules. 1995. 'Application of the fast‐evaporation sample preparation method for improving quantification of angiotensin II by matrix‐assisted laser desorption/ionization', Rapid Communications in Mass Spectrometry, 9: 1164-71.

Nimptsch, A., B. Fuchs, R. Suss, K. Zschornig, U. Jakop, F. Goritz, J. Schiller, and K. Muller. 2013. 'A simple method to identify ether lipids in spermatozoa samples by MALDI-TOF mass spectrometry', Anal Bioanal Chem, 405: 6675-82.

O'Donnell, Valerie B. 2011. 'Mass spectrometry analysis of oxidized phosphatidylcholine and phosphatidylethanolamine', Biochimica et Biophysica Acta (BBA) - Molecular and Cell Biology of Lipids, 1811: 818-26.

Schiller, J., J. Arnhold, S. Benard, M. Muller, S. Reichl, and K. Arnold. 1999. 'Lipid analysis by matrix-assisted laser desorption and ionization mass spectrometry: A methodological approach', Anal Biochem, 267: 46-56.

Shigeri, Y., A. Yasuda, M. Sakai, S. Ikeda, R. Arakawa, H. Sato, and T. Kinumi. 2015. 'Hydrazide and hydrazine reagents as reactive matrices for matrix-assisted laser desorption/ionization mass spectrometry to detect steroids with carbonyl groups', Eur J Mass Spectrom (Chichester), 21: 79-90.

Škrášková, Karolina, and Ron M. A. Heeren. 2013. 'A review of complementary separation methods and matrix assisted laser desorption ionization-mass

spectrometry imaging: Lowering sample complexity', Journal of Chromatography A, 1319: 1-13.

Spener, Friedrich, Michel Lagarde, Alain Géloên, and Michel Record. 2003. 'Editorial: What is lipidomics?', European Journal of Lipid Science and Technology, 105: 481-82.

Stults, John T. 1995. 'Matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS)', Current Opinion in Structural Biology, 5: 691-98.

Teuber, Kristin, Maria Fedorova, Ralf Hoffmann, and Jürgen Schiller. 2012. '2,4- Dinitrophenylhydrazine as a New Reactive Matrix to Analyze Oxidized Phospholipids by MALDI-TOF Mass Spectrometry', Analytical Letters, 45: 968- 76.

Vogel, M., A. Buldt, and U. Karst. 2000. 'Hydrazine reagents as derivatizing agents in environmental analysis--a critical review', Fresenius J Anal Chem, 366: 781-91.

Wang, Chia-Chen, Yin-Hung Lai, Yu-Meng Ou, Huan-Tsung Chang, and Yi-Sheng Wang. 2016. 'Critical factors determining the quantification capability of matrix- assisted laser desorption/ionization– time-of-flight mass spectrometry', Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 374.

Wang, Poguang, and Roger W. Giese. 2017. 'Recommendations for quantitative analysis of small molecules by matrix-assisted laser desorption ionization mass spectrometry', Journal of Chromatography A, 1486: 35-41.

Weschayanwiwat, Punjaporn, John F. Scamehorn, and Peter J. Reilly. 2005. 'Surfactant properties of low molecular weight phospholipids', Journal of Surfactants and Detergents, 8: 65-72.

Zenobi, Renato, and Richard Knochenmuss. 1998. 'Ion formation in MALDI mass spectrometry', Mass Spectrometry Reviews, 17: 337-66.

Appendix A

Positive ion mode MALDI-TOF spectra

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)

Figure A-1 DNPhydrazone peaks formed at different DNPH concentrations (a) 0.5 mM, (b) 1 mM, (c) 2.5 mM, (d) 5 mM, (e) 10 mM, (f) 20mM, (g) 30 mM, (h) 50 mM, and (i) 75mM

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)

Figure A-2 DPPC peaks mixed with different DNPH concentrations (a) 0.5 mM, (b) 1 mM, (c) 2.5 mM, (d) 5 mM, (e) 10 mM, (f) 20mM, (g) 30 mM, (h) 50 mM, and (i) 75mM

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)

Figure A-3 PoxnoPC mixed with the lipid tissue and different DNPH concentrations (a) 0.5 mM, (b) 1 mM, (c) 2.5 mM, (d) 5 mM, (e) 10 mM, (f) 20mM, (g) 30 mM, (h) 50 mM, and (i) 75mM

(a)

(b)

(c)

(d)

(e)

(f)

Figure A-4 Different laser intensities applied on DPPC in the presence of 100mM DHB (a) 0%, (b) 10%, (c) 20%, (d) 30%, (e) 40%, and (f) 50%

(a)

(b)

(c)

(d)

(e)

(f)

Figure A-5 Different laser intensities applied on DPPC mixed with 50 mM DNPH (a) 0%, (b) 10%, (c) 20%, (d) 30%, (e) 40%, and (f) 50%

(a)

(b)

(c)

(d)

(e)

(f)

Figure A-6 Different laser intensities applied on DPPC mixed with 50 mM DNPH and in the presence of 100 mM DHB (a) 0%, (b) 10%, (c) 20%, (d) 30%, (e) 40%, and (f) 50%

(a)

(b)

(c)

(d)

(e)

(f)

Figure A-7 Different laser intensities applied on DPPC mixed with the lipid tissue and 50 mM DNPH (a) 0%, (b) 10%, (c) 20%, (d) 30%, (e) 40%, and (f) 50%

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)

Figure A-8 DNPhydrazone and DPPC peaks after washing by GA and added (a) 0.5 mM, (b) 1 mM, (c) 2.5 mM, (d) 5 mM, (e) 10 mM, (f) 20mM, (g) 30 mM, (h) 50 mM, and (i) 75mM DNPH