Evaluation of lipid bromination

(1)

Bachelor thesis, 15 hp Life Science program, 180 hp

Spring term 2020

Evaluation of lipid bromination

For the relative measurement of a chlorine gas biomarker

Lovisa Ålander

Supervisors: Crister Åstot & Solomon Tesfalidet Examiner: Magnus Andersson

Swedish Defence Research Agency, FOI Umeå Umeå University

(2)

(3)

i

Abstract

Ever since chlorine was introduced as a chemical warfare agent, proving exposure to it has been difficult since there has been no unambiguous biomarker. However, resent research has found that phospholipids containing oleic acid (e.g. POPG) in the epithelial lining of alveoli can be chlorinated and used as exposure biomarkers. As of now, samples are taken by means of bronchoalveolar lavage, giving bronchoalveolar lavage fluid (BALF) samples. The purpose of this bachelor thesis is to find an internal reference to enable a relative measure of a biomarker for the exposure of chlorine. The aim was to label POPG by means of brominating oleic acid, to serve as an internal reference to enable a normalization of chlorinated POPG (POPG-HOCl) levels. The established sample workflow includes a derivatization of the phospholipids to produce a fatty acid ethanolamide prior to LC-MS/MS analysis. However, working with oleic acid of phospholipids as an internal reference means there is a major risk of contamination by triglycerides and other sources of oleic acid, and labelling is, therefore, crucial for biological samples. POPG-HOCl was synthesised by successive epoxidation and chlorination, and the chlorohydrin stability under bromination conditions were studied. Mixed POPG:POPG-HOCl samples in 1:1 ratios were prepared in two sets; one of which was brominated. An LC-MS method was applied on synthetic samples and a nano LC-MS/MS method for bronchoalveolar lavage fluid (BALF) samples. Our results show that bromination can be used for correction of errors that originate from BALF sampling errors, overlapping background and sample LC-MS signals, in the determination of oleic acid. However, the bromination of POPG, with the poor yield, needs a great deal of optimization.

(4)

ii

(5)

iii

List of Figures

Figure 1. A biomarker for chlorine gas exposure: POPG-HOCl... 7

Figure 2. Stearoyl-HOCl EtAmide, synthesised on SPE column. ... 8

Figure 3. Summarized workflow ...12

Figure 4. ¹H-NMR spectra; Left: 9,10-dibromostearic acid, and right: 9-chloro-10- hypoxystearic acid. ... 13

Figure 5. Circled proton groups for characteristic ¹H-NMR chemical shifts; upper right 9,10-epoxystearic acid, upper left 9,10-dibromostearic acid, lower left 9-chloro-10- hydroxystearic acid. ... 13

Figure 6. Mass spectra of dibromostearoyl MeOate Left: 20V cone voltage, loss of Br2 visible. Right: 5V cone voltage, no loss of Br2 visible...14

Figure 7. Average peak areas, replicates and linear regressions of the methyl esters. Brominated: dibromostearic acid MeOate, Oleic acid: oleic acid MeOate. ... 15

Figure 8. Chromatogram of brominated BALF sample (2h1), for stearoyl-HOCl EtAmide, oleoyl EtAmide and dibromostearoyl EtAmide, with MRM channels 378.3>324.3, 326.3>62.1 and 486.1>324.3 respectively. ... 17

Figure 9. Left: Peak areas of BALF samples and brominated BALF samples. Peak areas for different fragmentations and isotope compositions of parent ions are summed to total peak areas. Right: Peak area ratios of dibromostearoyl EtAmide and oleoyl EtAmide against conversion factor. ... 17

List of Tables

Table 1. MRM channels for background signals. ... 10

Table 2. Gradient during lipid background analysis. ... 10

Table 3. SIR channels for the two methyl esters, ammonium adducts... 11

Table 4. Gradient during the conversion factor analysis. ... 11

Table 5. SIR channels for POPG sample analysis. ...12

Table 6. Gradient during POPG sample analysis. ...12

Table 7. MRM channels for BALF sample analysis. ...12

Table 8. Average peak areas of 3 replicates and standard deviations. ...16

(7)

v

List of abbreviations

FA – Formic acid, UHPLC grade

EtOH – Ethanol, 99.7 % spectroscopy grade

MeOH – Methanol, ≥ 99.9 % analytical reagent grade EtAmine – Ethanolamine

EtAmide – Ethanolamide AcN – Acetonitrile

MQ – Milli-Q® water, Advantage A10 system Merck Millipore MeOate – Methyl ester

AmAc – Ammonium acetate

POPC – Palmitoyl-oleoyl phosphatidylcholine POPG – Palmitoyl-oleoyl phosphatidylglycerol

POPG-HOCl – Palmitoyl-oleoyl phosphatidylglycerol chlorohydrin POPG-Br2 – Palmitoyl-9,10-dibromostearoyl phosphatidylglycerol Oleic acid – (E)-octadec-9-enoic acid

Stearic acid – Octadecanoic acid HOCl – Hypochlorous acid

CHCl3 – Chloroform DCM – Dichloromethane

MTBE – tert-Butyl methyl ether NBS – N-bromosuccinimide CWA – Chemical warfare agent LC – Liquid chromatography MS – Mass spectrometry

MS/MS – Tandem mass spectrometry SPE – Solid-phase extraction

HLB – Hydrophilic lipophilic balance SIR – Selected ion recording

MRM – Multiple reaction monitoring BALF – Bronchoalveolar lavage fluid

(8)

vi

(9)

7

1. Introduction

The purpose of this thesis is to evaluate the bromination of POPG, for the relative measurement of a chlorine gas biomarker, chlorinated POPG (POPG-HOCl). The bromination was evaluated with synthetic samples where the ratio between POPG and POPG-HOCl is known, as a proof of concept, and with biomedical samples from chlorine exposed laboratory animals. When sampling for the biomarker (POPG-HOCl), the method of bronchoalveolar lavage, washing of the epithelial lining of the lungs, is used. With this method of sampling there is no way to control the amount of lung surface that is sampled. Therefore, measuring a concentration of the biomarker cannot provide any information on the level of exposure, which is why a relative measure must be applied. A relative measure can be achieved by normalizing the biomarker to a reference analyte that is present at a stable level, a method often used for normalizing biomarkers in urine [1].

1.1 Background

A biomarker is, simply put, a molecule that is present either exclusively or at elevated levels following a particular event, such as a disease [2]. Biomarkers have been used for many years to diagnose conditions such as cancer and oxidative stress [3, 4] as well as exposure to various chemicals [5]. The biomarker studied in this bachelor thesis is POPG-HOCl (see Figure 1), established by P. Hemström et al. [6], for the evaluation of chlorine exposure.

Figure 1. A biomarker for chlorine gas exposure: POPG-HOCl

Chlorine is widely used in the world today; for the synthesis of medicinal drugs, plastics such as PVC, pesticides, and as an antibacterial additive in drinking water to mention a few [6].

Sadly, due to its toxic properties, chlorine gas was used as a CWA in World War I, as well as in more recent years in Syria [7]. More than 10 min exposure to 50 ppm chlorine gas can be life- threatening [8] however it has, and still is, problematic to prove the usage of chlorine since no unambiguous biomarkers have been available. For several years chlorinated species of the amino acid tyrosine was used; however, these tyrosine species occur during inflammation [9].

During inflammation, white blood cells produce HOCl, which may react with biomolecules to form the same chlorinated products that can be formed from chlorine exposure. However, the chlorohydrin of POPG and POPC have been promising biomarkers due to their absence during inflammation [6].

In this bachelor thesis work, the analysis and sample work-up was further developed based on the work by Hemström et al. [6]. As part of their established sample work-up (not published), the phospholipids are cleaved by transesterification with EtAmine (on an HLB-SPE column) which after rearrangement yields the EtAmide of the oleic acid chlorohydrin; stearoyl-HOCl EtAmide (Figure 2). This cleavage not only facilitates MS analysis but enables a collective analysis of chlorinated POPG, POPC and other oleic acid containing phospholipids, since they all yield stearoyl-HOCl. However, a transesterification in solution degrades chlorohydrins (to epoxides), and the oleic acid derivative cannot be detected. A solution to this was to protect the

(10)

8 fatty acid tails on an HLB-SPE column; the long carbon tails of the fatty acids are surrounded by the lipophilic part of the cartridge, shielding off the EtAmine. However, it is uncertain if this protection works since no calibration curve has been made; due to possible degradation and lack of internal standard of oleoyl EtAmide.

Figure 2. Stearoyl-HOCl EtAmide, synthesised on SPE column.

One of the main issues working with phospholipids is their extensive appearance and adhesion;

they represent an essential fraction of our epidermis, and the lipid bilayer of all our cells, [10], [11] and adhere well to both glassware and plastics. These properties can result in a great deal of contamination during the entire process, producing overlapping background signals.

Background signals of the oleic acid amide were detected and traced back to lipid contamination of the EtAmine and SPE columns (not published). Because of these complications, we realised the importance and great need for an early labelling of lipids originating from the biomedical sample, to be able to distinguish between them and lipids from contamination. One such labelling that could serve this purpose is a dibromination of oleic acid in POPG/C, which would change not only mass but fragmentation and retention time of oleoyl species. A bromination could also improve the recovery after SPE work-up since there would be no alkenes left to degrade. If this recovery is sufficient, a bromination could also improve the relative measurement of the chlorine biomarker.

The major analytical techniques used during this thesis project are LC-MS, and ¹H-NMR. The reader not familiar with these techniques may be interested in sections of the book on analytical chemistry by D. C. Harris [12, pp. 668-678 & 699-700, 559, 579-587] and Solomons’

Organic Chemistry [13, pp. 392-436]. For details on the SPE technique we also refer to Harris’

Analytical Chemistry [12, pp. 785-787].

1.2 Aim

In this bachelor thesis, we aimed to find an internal reference that can be used to measure relative levels of POPG-HOCl; how much POPG-HOCl there is in relation to all precursor POPG. More specifically, we investigated whether dibrominated POPG, palmitoyl-9,10- dibromostearoyl phosphatidylglycerol, could serve this purpose. For POPG-Br2 to serve as such a reference, there are two major requirements. The chlorohydrin group of POPG-HOCl present in a mixed sample has to be stable in the bromination cocktail not to deplete the biomarker pool, and POPG-Br2 has to be produced from POPG at a high reproducible yield. Due to the extensive contamination risk of oleic acid from triglycerides, subjecting the entire sample to labelling can be an essential task when performing trace analysis in biological samples. This labelling could correct for overlapping signals of oleic acid from contamination and sample during analysis.

(11)

9

2. Methods

2.1 Organic synthesis

This section describes the different methods of synthesis used to prepare molecules that will serve as a basis for the analysis later performed. All ¹H-NMR analyses were performed on a Bruker Avance III Ultrashield 500 MHz NMR-spectrometer equipped with a Cryoprobe 5 mm BBFO head at room temperature (298 K). ¹H-NMR analysis was performed after most reactions.

2.1.1 Dibromination

Br2 of 1.1 equivalents was added to (E)-octadec-9-enoic acid in Et2O and stirred for 15min at room temperature [14], ¹H-NMR analysis confirmed complete bromination. The reaction was quenched with sat. Na2S2O3, washed with water, dried (Na2SO4) and concentrated to yield 9,10- dibromooctadecanoic acid. POPG was subject to the same reaction with Br2 present in 2 equivalents.

2.1.2 Chlorohydrin

(E)-Octadec-9-enoic acid in DCM was epoxidized by mCPBA (3 equivalents) at room temperature overnight. The reaction was quenched with sat. Na2S2O3, transferred to a separation funnel and extracted with 3x20mL Et2O. The Et2O phase was dried (Na2SO4) and concentrated to yield 9,10-epoxystearic acid. The epoxide was diluted in Et2O, and 2M HCl in Et2O was added in 1.5 equivalents, the reaction was left to stir at room temperature overnight after which the product was concentrated to yield 9-chloro-10-hydroxystearic acid followed by

1H-NMR analysis. This chlorohydrin was subject to the bromination described above.

A sodiumsalt of POPG was subject to the epoxidation and chlorination described above to yield POPG-HOCl.

2.1.3 Methyl Esterification

A spatula tip of toluenesulfonic acid was added to 9,10-dibromostearic acid and (E)-octadec- 9-enoic acid dissolved in MeOH, and let to stir over two days. The MeOH was evaporated, the products were dissolved in water, extracted three times with DCM, dried (Na2SO4) and concentrated to yield the two acids’ methyl esters.

(12)

10

2.2 Preparations and Analysis

This section includes the different analytical techniques used during this project. For the reader interested in earlier analyses on chlorine gas biomarkers we refer to the article written by P. Hemstöm et al. [6].

2.2.1 Background Signals

EtAmine was vacuum distilled (16 mmHg) where pre, main and end fractions were collected.

Samples (1 mL) of the main fraction and non-distilled EtAmine were extracted with 1mL MTBE and frozen to separate the phases. 200 µL of the MTBE phases were concentrated under N2 gas flow, diluted in 200µL 50% EtOH. The samples were analysed with 2 MRM channels (Table 1) on a Waters Xevo™ mass spectrometer (ES+) with a Waters Acquity UPLC® liquid chromatography system. The gradient in Table 2 on an ACQUITY UPLC® Protein BEH C4 300Å 1.7µm 2.1×100mm 1/pkg column was used with eluents A: AcN + 0.1% FA and B: MQ + 0.1% FA.

Table 1. MRM channels for background signals.

Molecule M⁺ Daughter ion(s)

Oleoyl EtAmide 326.3 62.1

Stearoyl EtAmide 328.3 310.3

Table 2. Gradient during lipid background analysis.

Time (min) Eluent A Eluent B

Initial 75 25

5 5 95

7 5 95

7.1 75 25

10 75 25

2.2.2 Conversion Factor

Since we did not have the EtAmides of dibromostearoyl and oleic acid, a conversion factor based on the sensitivity difference between dibromostearoyl and stearoyl analytes on the mass spectrometer was calculated. This was done using the methyl esters of dibromostearoyl and oleic acid. Several concentrations were prepared and analysed on an Acquity™ Ultra Performance LC with a Micromass Quattro micro™ mass spectrometer. Both gradient and MS settings were developed on these instruments using different voltages, eluents, etc. once gradient and MS files were developed, a series of concentrations (1:1 ester mixture) were prepared and analysed on a Waters Xevo™ to see which concentrations could be used.

The esters were diluted in EtOH to concentrations of 10, 50 and 100mM in three replicates and analysed on a Waters Xevo™ mass spectrometer (see SIR channels in Table 3) with a Waters Acquity UPLC® liquid chromatography system with gradient according to

Table 4. An ACQUITY UPLC® Protein BEH C4 300Å 1.7µm 2.1×100mm 1/pkg column with eluents A: 50mM AmAc and B: AcN:MQ + 50mM AmAc was used. The conversion factor was then taken as the slope quotient.

(13)

11 Table 3. SIR channels for the two methyl esters, ammonium adducts.

Molecule [M+NH4]⁺

Oleoyl MeOate 314.5

Dibromostearoyl MeOate 472.3 474.3 476.3

(Bromine isotope composition) (79:79) (79:81) (81:81)

Table 4. Gradient during the conversion factor analysis.

Time (min) Buffer A Buffer B

Initial 70 30

8 5 95

10 5 95

10.1 70 30

12 70 30

2.2.3 Sample Preparation

The following samples were prepared based on the assumption of 100% yield after bromination and 80% yield after SPE to final concentrations of 5ppm stearoyl EtAmide chlorohydrin (13.2µM).

Stock solutions of 1mM POPG in 5:1 MeOH:CHCl3 and 210µM POPG-HOCl in Et2O were prepared. The samples prepared from these solutions will be referred to as POPG-samples.

1.65µL POPG and 7.86µL POPG-HOCl were added to three vials, in three replicates, and let to dry, at which point they were diluted in 1.5mL Et2O. To the first vial, a stir bar and 0.2µL Br2

(>2300 equivalents) were added and left to stir for 20min, following the work-up described in Dibromination with an additional wash with water. The second vial was subject to the reaction work-up only with an additional water wash, and the third only diluted in Et2O. All vials were stored in -18°C awaiting SPE work-up.

BALF samples fromrRats exposed to 250ppm chlorine for 15min (“approved by the regional ethics committee on animal experiments in Umeå, Sweden according to directive 2010/64/EU”) are named according to sampling time after exposure; control (no exposure), 2h1, 2h2and 6h after exposure respectively. 400µL BALF, and a solvent blank of 400µL MQ, was diluted in 4mL MTBE and 1.4mL MeOH whereupon they were vortexed and allowed to settle. 800µL MQ was added for phase separation and the BALF samples were centrifuged at 2,500prm for 5min. The MTBE phases was transferred into new vials and allowed to evaporate at 40°C under N2 flow. The BALF samples were then diluted in 1mL chloroform, split into two sets and concentrated at 40°C under N2 flow. One set was subject to bromination with the additional water wash during work-up.

All samples were allowed to dry under N2 flow at 40°C after which they were dissolved in 100µL EtOH, vortexed and further diluted with 100µL MQ. OASIS® HLB 1cc/10mg columns were conditioned with 200µL 50% EtOH, whereupon the sample was applied. The sample vials were washed with an additional 200µL 50% EtOH, followed by the application of 100µL distilled EtAmine after which the columns were wrapped in foil and placed in a 40°C oven for 2h. A blank sample was prepared by adding 100µL distilled EtAmine after column conditioning, and placing it in the oven for 2h.

(14)

12 All columns were washed with 2 × 1mL 10% FA and let run dry. Thereafter, the products were eluted with 500µL EtOH. The EtOH was evaporated under N2 flow at 40°C for about 2h, and the samples were stored at 4°C awaiting analysis on LC-MS.

2.2.4 Sample Analysis

POPG samples were dissolved in 100µL 50% EtOH and analysis was performed on a Waters Xevo™ mass spectrometer (see SIR channels in Table 5) with a Waters Acquity UPLC® liquid chromatography system. An ACQUITY UPLC® Protein BEH C4 300Å 1.7µm 2.1×100mm 1/pkg column was used with eluents A: AcN + 0.1% FA and B: MQ + 0.1% FA in a gradient according to

Table 6. The sample workflow is summarized in Figure 3.

Table 5. SIR channels for POPG sample analysis.

Molecule [M+H]⁺

Stearoyl HOCl EtAmide 378.3

Oleoyl EtAmide 326.3

Dibromostearoyl EtAmide 484.1 486.1 488.1

(Bromine isotope composition) (79:79) (79:81) (81:81)

Table 6. Gradient during POPG sample analysis.

Time (min) Buffer A Buffer B

Initial 90 10

10 5 95

12 5 95

12.1 90 10

15 90 10

BALF samples were dissolved in 50µL EtOH, vortexed and further diluted in 50µL MQ and analysed on a Waters nano Acquity Ultra Performance LC™, with eluents 95% MQ + 0.1% FA and 5% can + 0.1% FA, and a Waters Xevo® TQ-XS with Zspray™ Nanoflow™, with MRM channels according to Table 7. The MRM channels monitor one, or several, specific

fragmentations from the molecular ion.

Table 7. MRM channels for BALF sample analysis.

Molecule M⁺ Daughter ion

Stearoyl-HOCl EtAmide 378.3 324.3

Oleoyl EtAmide 326.3 62.1

Stearoyl EtAmide 328.3 310.3

Dibromostearoyl EtAmide 484.1 486.1 488.1 324.3 326.3 (Bromine isotope

composition) (79:79) (79:81) (81:81)

Figure 3. Summarized workflow

(15)

13

3. Results and Discussion

3.1 Organic Synthesis

Below follows the characteristic ¹H-NMR chemical shifts of the products synthesized from oleic acid Figure 4, and the proton groups responsible for these shifts are circled in Figure 5.

9,10-Epoxystearic acid (500 MHz, CDCl3): δ = 0.88 (t, 3H), 2.38 (t, 2H), 2.9 (m, 2H), 3.5 (q, 4H).

9-Chloro-10-hydroxystearic acid (500 MHz, CDCl3): δ = 0.88 (t, 3H), 2.38 (t, 2H), 3.5 (q, 4H), 3.6 (m, 1H), 3.8 (m, 1H).

9,10-Dibromostearic acid (500 MHz, CDCl3): δ = 0.88 (t, 3H), 2.38 (t, 2H), 3.5 (q, 4H), 4.2 (dd, 2H).

Figure 4. ¹H-NMR spectra; Left: 9,10-dibromostearic acid, and right: 9-chloro-10- hypoxystearic acid.

Figure 5. Circled proton groups for characteristic ¹H-NMR chemical shifts; upper right 9,10-epoxystearic acid, upper left 9,10-dibromostearic acid, lower left 9-chloro-10- hydroxystearic acid.

Methyl esters (500 MHz, CDCl3): δ = 0.88 (t, 3H), 2.3 (t, 2H), 3.67 (s, 3H), dd

q m

(16)

14 Stearic acid HOCl showed no change in chemical shifts after being subject to the bromination reaction. Epoxidation and chlorination of POPG produced POPG-HOCl in 92% yield.

We chose to use Br2 instead of NBS for the bromination because of the intermediate bromonium ion that is formed using Br2. Even though NBS produces Br2, radical reactions can take place as well. A radical reaction in combination with a chlorohydrin signifies a risk of losing either the hydroxyl- or chloro group; due to the electron deprived carbons they are attached to.

3.1.1 Bromination; Advantages and Problems

After bromination, oleic acid yielded 85% 9,10-dibromostearic acid, while the bromination of POPG in Et2O and DCM gave yields of 53% and 38% POPG-Br2 respectively. ¹H-NMR study of these two POPG-Br2 products showed that the recovered product from Et2O contained alkenes, but not that from CH2Cl2. The presence of alkenes means that the actual POPG-Br2

yields are even more inadequate. The poor yields can be due to not only the solubility of POPG in the two solvents but also solvent polarity. The intermediate bromonium ion may require a certain degree of polarity. However, the NMR study of 9,10-dibromostearic acid (after bromination of oleic acid) indicates that this should not the main reason; the bromination was complete, since there were no signals from remaining double bonds.

The major purpose of the dibromination is to label oleic acid in the sample at an early stage, to avoid contaminant oleic acid signal overlap. The two oleic acid species will have different masses as well as retention times and fractionation patterns during analysis.

3.2 Preparations and Analysis

3.2.1 Conversion factor

When analysing the methyl esters we found that they ionized in sodium adducts ([M+Na]⁺ peaks were found) and as a consequence did not fragment to their molecular ions. We therefore chose to use eluents with ammonium. The ammonium would be in excess and it’s adduct is much more readily fragmented due to the ability do loose a charge in form of a proton (which sodium ions cannot). Another problem we encountered when analysing the dibromostearoyl MeOate was how easily bromine was lost during mass spectrometry; a cone voltage of no more than 5V had to be used. The difference in fragmentation can be seen in Figure 6. This fragmentation will however not present a problem during MRM, since the loss of bromine is a very unique feature.

Figure 6. Mass spectra of dibromostearoyl MeOate Left: 20V cone voltage, loss of Br2 visible.

Right: 5V cone voltage, no loss of Br2 visible.

Br2

(17)

15 The mass spectrometer turned out to be very insensitive to the methyl esters, leaving us no other option than to use very high concentrations. This meant that we reached the plateau of the response to dibromostearic acid MeOate, as can be seen in Figure 7. Due to this insensitivity and flattening of the curve, we assumed a linearity for concentrations of ≤ 50mM dibromostearic acid MeOate. Therefore, the conversion factor of 3.81 was based on the slopes, K, between 10 and 50mM; 𝐾𝑦_𝐵𝑟⁄𝐾𝑦 according to Equation 1 and

Equation 2.

Figure 7. Average peak areas, replicates and linear regressions of the methyl esters.

Brominated: dibromostearic acid MeOate, Oleic acid: oleic acid MeOate.

Equation 1.

𝑦_𝐵𝑟 = 1.36 ∙ 10⁵𝑥 + 1.07 ∙ 10⁶

Equation 2.

y = 3.57 ∙ 10⁴𝑥 + 2.62 ∙ 10⁵

This conversion factor will be used to estimate the relative levels of dibromostearoyl EtAmide and stearoyl EtAmide, even though it is calculated based on the response of their methyl esters.

It is very important to note that this conversion factor is based on the hypothesis that a bromination will change the relative sensitivity of the EtAmides by a comparable factor to the methyl esters. Note also that the methyl esters and amides form different adducts during ionization. The conversion factor is based on NH4+ adducts [M+NH4]⁺, whilst the sample analyses are based on proton adducts, [M+H]⁺. However, in theory the surface activity determines the electrospray MS sensitivity of analytes within the same substance group. Thus, a similar relative effect of bromination on the two substance classes can be assumed.

R² = 0,9241 R² = 0,9828 0,00

5000000,00 10000000,00 15000000,00

0 20 40 60 80 100 120

Average peak area

Concentration (mM)

Conversion Factor

Brominated Oleic acid

Repl brominated repl oleic acid

(18)

16 3.2.2 Sample Preparation and Analysis

The initial sample concentrations were based on 100% and 80% yield after bromination and SPE work-up respectively. This was done to imitate a real sample, as to not manipulate it based on how well the bromination works.

Table 8. Average peak areas of 3 replicates and standard deviations.

Sample set Dibromostearoyl

EtAmide Oleoyl

EtAmide Stearoyl-HOCl EtAmide

Brominated 260107

± 20708 5451

± 4143 16177

± 1452

Reference -

- 384086

± 44568 29755

± 12963

Based on the data from LC-MS analysis in Table 8, there was minimal oleic acid left after bromination. However, none of the brominated samples corresponded to a higher concentration of dibromostearoyl EtAmide than oleoyl EtAmide, based on the conversion factor with an average peak area ratio of 0.7. This can to some extent be explained by the poor yield of POPG-Br2 (53%), yet taking that into account we still do not reach a ratio of 3.8. A poor recovery after SPE work-up could explain the rest of the lack of response, assuming that the recovery of oleoyl EtAmide is at its maximum. Since POPG is hard to dissolve in EtOH, and POPG-Br2 is expected to be even less soluble, this may be the main issue of the poor recovery.

Also, since dibromostearoyl EtAmide is expected to be highly lipophilic, it is possible that much of it is not eluted from the SPE column.

Based on that all samples were made with POPG and POPG-HOCl in 1:1 proportions and the response of the two analytes are assumed to be equal; we expected the response of the two analytes to be more similar. This difference could be caused by dissimilar recoveries after SPE work-up, or a poor synthesis yield. The difference in response of the chlorohydrin between the first and the two last samples can be explained by the fact that they were analysed on two different days. The response of the mass spectrometer can vary from day to day, and the similarity in response pattern implies that these differences should not be of concern. That is;

the minor difference in chlorohydrin response does not signify that something went wrong in the preparation of the samples.

(19)

17 3.2.3 BALF Sample Analysis

Figure 8. Chromatogram of brominated BALF sample (2h1), for stearoyl-HOCl EtAmide, oleoyl EtAmide and dibromostearoyl EtAmide, with MRM channels 378.3>324.3, 326.3>62.1 and 486.1>324.3 respectively.

Figure 9. Left: Peak areas of BALF samples and brominated BALF samples. Peak areas for different fragmentations and isotope compositions of parent ions are summed to total peak areas. Right: Peak area ratios of dibromostearoyl EtAmide and oleoyl EtAmide against conversion factor.

0 50000000 100000000 150000000 200000000 250000000

control 2h1 2h2 6h control 2h1 2h2 6h

Normal Brominated

Peak area

BALF Sample Analysis

Oleoyl amide Chlorohydrin amide Dibromo amide

0 0,5 1 1,5 2 2,5 3 3,5 4

Peak area ratio

BALF Samples

Dibromostearoyl EtAmide Stearoyl-HOCl EtAmide

Oleoyl EtAmide

(20)

18 Analysis of BALF samples showed that dibromostearoyl EtAmide fragments in two different ways; loss of 2×HBr, and Br2, and the total response was consequently summed. Comparing these results (Figure 9) to the synthetic POPG samples (Table 8) the responses of the BALF- samples look more similar to those expected, based on the conversion factor calculations. All BALF-samples gave a higher response of dibromostearoyl EtAmide than oleoyl EtAmide while still not reaching the calculated ratio. Why the bromination seems to have worked better on the BALF samples can be because of the difference in lipid composition. BALF contains phosphatidylcholines (e.g. POPC) as well, which are zwitterions. With POPC being a zwitterion, it can possess a higher solubility in Et2O than POPG, which could enable a higher bromination yield and a higher solubility of POPC-Br2 in EtOH.

Something to also note is the difference in response of chlorohydrin EtAmide and oleoyl EtAmide in the BALF and synthetic POG samples (Figure 9 and Table 8). Considering the rats’

high exposure of chlorine we still expected there to be a big difference in response, which is reflected by the results. This result further supports that something practically went wrong with the POPG samples, such as the solubility problems during synthesis, since they were subjected to the same SPE work-up.

4. Conclusions

After this thesis project we can conclude that there is a great need of an internal reference, since the LS-MS signals from the proven contamination and the sample can be distinguished thanks to the bromination. We can also conclude that dibromostearoyl is a suitable internal reference for relative stearoyl-HOCl measurements. With the bromination of POPG in the synthetic samples we were able to distinguish between oleic acid originating from the sample and background. Bromination of BALF samples showed even more promising results than the synthetic POPG samples, however the bromination still needs to be optimized. We found that dibromostearoyl fragments in two ways; loss of 2xHBr and Br2, and to improve quantification, one should take both fragmentations in consideration. To optimize the conversion factor, samples of <10mM should be analysed.

5. Acknowledgements

I would like to give my warmest thank you to my supervisors and mentors at FOI Umeå; Crister Åstot, Petrus Hemström and Andreas Larsson. Thank you for giving me this opportunity, all of this experience, and for the battle between organic and analytical chemistry. I would also like to thank my supervisor at Umeå University, Solomon Tesfalidet, for the rewarding comments and thoughts on this thesis. Thank you especially to Rafael Velasquez and my family; thank you for being there for me and listening to my constant blabber on chemistry.

(21)

19

6. References

[1] S. S. Waikar, V. S. Sabbisetti and J. V. Bonventre, "Normalization of urinary biomarkers to creatine during change in glomerular filtration rate," Kidney international, vol. 78, p.

486.494, 2010.

[2] Nationalencyklopedin, biomarkör.

[3] H. R. Griffiths, L. Möller, G. Bartosz, A. Bast, C. Bertoni-Freddari, A. Collins, M. Cooke, S. Coolen, G. Haenen, A.-M. Hoberg, S. Loft, J. Lunec, R. Olinski, J. Parry, A. Pompella, H. Poulsen, H. Verhagen and S. B. Astley, "Biomarkers," Molecular aspects of medicine, vol. 23, pp. 101-208, 2002.

[4] X. Chen, Y. Ba and L. Ma, "Characterization of microRNAs in serum: a novel class of biomarkers for diagnosis of cancer and other diseases," Cell Research, vol. 18, pp. 997- 1006, 2008.

[5] A. M. Calafat, M. P. Longnecker and H. M. Koch, "Optimal Exposure Biomarkers for Nonpersistent CHemicals in Environmental Epidemiollogy," Environmental Health Perspectives, vol. 123, no. 7, 2015.

[6] P. Hemström, A. Larsson, L. Elfsmark and C. Åstot, “L-α-Phosphatidyglycerol Chlorohydrins as Potential Biomarkers for Chlorine Gas Exposure,” Analytical Chemistry, vol. 88, pp. 9972-9979, 2016.

[7] "Third report of the Organization for the Prohibition of Chemical Weapons-United Nations Joint Investigative Mechanism," United Nations, 2016.

[8] EPA. AEGL Chlorine. Subcomittee on Acute Exposure Guideline Levels, Committee on Toxicology, Board on Environmental Studies and Toxicology., "Acute Exposure Guideline Levels for Selected Airborne Chemicals," vol. 4, EPA, 2004.

[9] M. P. Curtis, “Chlorinated Tyrosine Species as Markers of Inflammation: A Kinetic Study,” Loma Linda University Electronic Theses, Dissertations & Projects., 2015.

[10] M. A. Lampe, M. L. Wiliams and P. M. Elias, "Human epidermal lipids: characterization and modulations during differentiation," vol. 24, pp. 131-140, 1983.

[11] E. Fahy, S. Subramaniam, H. A. Brown, C. K. Glass, A. H. Merril Jr., R. C. Murphy, C. R.

Raetz, D. W. Russell, Y. Seyama, W. Shaw, T. Shimizu, F. Spener, G. van Meer, M. S.

VanNieuwenhze, S. H. White, J. L. Witztum and E. A. Dennis, "A comprehensive classification system for lipids," Eur. J. Lipid Sci. Technol, no. 107, pp. 337-364, 2005.

[12] D. C. Harris and C. A. Lucy, Quantitative Chemical Analysis, 9 ed., W. H. Freeman and Company, 2016.

[13] T. G. Solomons, C. B. Fryhle and S. A. Snyder, Solomons' Organic Chemistry, 12 ed., John Wiley & Sons Singapore Pte. Ltd., 2016.

[14] M. S. F. Lie Ken Jie and K. Prabhavathi, “Synthesis of pyridazine fatty ester derivatives in water: a sonochemical approach,” Journal of the Chemical Society, Perkin Transactions 1, 1997.

(22)

20

7. Appendix I

Table I 1. Average peak area of 3 replicates for different concentrations of dibromostearic acid MeOate and oleic acid MeOate

Concentration (mM) Dibromostearic acid

MeOate Oleic acid MeOate

10 2 429 798 619 865

50 7 885 022 2 049 649

100 10 345 160 3 179 475

Evaluation of lipid bromination