Investigation of molecular probes for pH determination with electrospray ionization mass spectrometry

(1)

Investigation of molecular probes for pH

determination with electrospray ionization

mass spectrometry

Elise Arnberg

March 13

th

2019

Degree Project C in Chemistry Bachelor Program in Chemistry

Analytical Chemistry - Department of Chemistry BMC - Uppsala University

(2)

1

Acknowledgment

I want to take this opportunity to say my greatest thanks to all members of Lanekoff reach group at the Analytical chemistry department at Uppsala university. Thanks for all support, fruitful discussions and everything in between.

(3)

2

Abstract

An acidic microenvironment is a well-known feature occurring in many pathologic stages such as tumors, infection and ischemic stroke. pH measurement in biologic tissue is therefore of high interest. Mass spectrometry can potentially offer simultaneous analyses of metabolites and pH though desorption directly from biological samples surface. The aim of the bachelor thesis is to take the first steps toward a more detailed analysis of chemical microenvironment by investigating if molecular probes could be used for pH measurements using electrospray ionization. Three common pH indicators were investigated where methyl orange was selected for a more thorough investigation. These investigations indicate that protonation of methyl orange relates to the solution pH. The observed trends were maintained when methyl orange was analyzed with parameters simulated to future conditions, such as lower water content and desorption from spiked biological samples. Even if further testing is necessary, these results show that methyl orange is of future potential for pH measurements using mass spectrometry.

(4)

3

Table of content

1 Abbreviation ... 4 2 Aim ... 5 3 Introduction ... 5 4 Theory ... 6 4.1 Mass spectrometry: ... 6 4.2 Electrospray ionization:... 6

4.2.1 Solvent composition and ESI: ... 7

4.3 Linear quadrupole ion trap: ... 8

4.4 Nanospray desorption electrospray ionization ... 8

5 Experimental: ... 9

5.1 Chemicals: ... 9

5.2 Sample preparation:... 9

5.2.1 First evaluation of several molecules ... 9

5.2.2 Repeatability of methyl orange: ... 10

5.2.3 Sodium’s effect on the intensity ... 11

5.2.4 Methanol and water solution (50:50) ... 11

5.2.5 Methyl orange solution for analyses of biological sample ... 11

5.2.6 Brain tissue section ... 12

5.3 Instrument... 12

6 Results and discussion ... 13

6.1 First evaluation using several molecules ... 13

6.2 Repeatability of methyl orange samples ... 19

6.3 Repeatability of the instrument ... 22

6.4 Sodium’s effect on the intensity ... 24

6.5 Change of solvent composition ... 25

6.6 Spiked brain tissue section ... 28

7 Conclusion ... 30

(5)

4

1 Abbreviation

BTB Bromothymol blue

CRM Charged residue model

ESI Electrospray ionization

IEM Ion evaporation model

LC Liquid chromatography

M.O. Methyl orange

m/z Mass-to-charge ratio

MeOH Methanol

MS Mass spectrometry

Nano-DESI Nanospray desorption electrospray ionization

phph Phenolphthalein

(6)

5

2 Aim

Find and evaluate potential molecular probes that can be used for determination of pH in biological samples such as tissue and cells using mass spectrometry.

3 Introduction

Maintaining the intra and extracellular pH is vital for many cellular processes. Promoting glucose metabolism is a common feature of cancer cell which results in increased amounts of lactic acid consequently leading to acidification of the microenvironment, the space round the cells[1]. Therefore, tumors have a lower pH than healthy tissue [2]. Other pathological states such as inflammation, infection and ischemic stroke also give rise to a more acidic microenvironment[3] [4]. Measuring pH in tissue and cells is therefore of high importance for diagnosing and understanding these diseases. Various methods for pH measurements in biological samples exist. For example, small pH electrodes with probe tips ranging down to µm size have been used for decades[2]. However, electrodes can easily puncture cells and blood vessels causing the measurements to reflect an unknown mixture of extracellular fluid, blood and cell content. Miniature electrodes has the benefit of measuring pH over smaller distances but with limited sensitivity [2] [4]. Use of pH sensitive fluorescent probes is a common way of measuring pH in cells. The probes fluoresces light depending on the pH round them[4]. Fluorescence spectroscopy is usually more sensitive and easier to operate than pH electrodes[5]. The question is then why challenge the already existing methods for pH measurements in biological samples with mass spectrometry. Mass spectrometry is far more expensive and has more complex equipment than a pH electrode. However, a mass spectrometer can give information about metabolites, proteins, lipids and other molecules by desorption directly from the surface of a biological sample, thereby offering a possibility to measure pH and metabolites simultaneously[6]. This report will focus on taking the first steps towards a more rapid and detailed analyze of the chemical microenvironment by investigating if molecular probes can be used to determine pH of a solution by electrospray ionization mass spectrometry.

(7)

6

4 Theory

4.1 Mass spectrometry:

Mass spectrometry (MS) is a rapid and sensitive analytical tool to collect information about the analyte composition and molecular formulas as well as the samples purity. It is widely used in many scientific fields and can analyses a broad range of analytes, from small inorganic ions to larger complex biomolecules. The principle relies on forming positive or negative gas-phase ions in the ion source which are then separated from each other ions in the mass analyzer, electrically, magnetically or by time, based on their mass and charge (mass-to-charge ratio, m/z). The final destination for the ions is the detector. In the spectrum is the relative abundance or ion count plotted against the m/z ratio[7].

A variety of MS instrument currently are commercially available, and the choice of MS is dependent on multiple parameters such as structure and molecular mass of the analyte and the instruments sensitivity, speed and resolution[8]. In these experiments are electrospray ionization source coupled to a linear quadrupole ion trap as a mass analyzer, a further description of these components follows below.

4.2 Electrospray ionization:

Electrospray ionization (ESI) is a soft ionization technique that converts the analytes of interest to gas phase ions with little or no fragmentation. Following description will refer to positive ionization mode. The analytes in solution are brought to the ion source in atmospheric pressure through a thin needle like capillary[9],[8]. The needle is induced with a positive electric potential. An electric field is imposed by the potential difference created by a counter electrode which is usually a large plate with an orifice to the mass analyzer[10],[9]. In negative mode is all voltage reversed. Influenced by the field, charged ions in the solution will start moving. Positive ions will move towards the meniscus causing a polarization in the solvent. Eventually, the meniscus forms a cone shape (Taylor cone) created by the polarization and the negative potential from the counter electrode. Droplets of solvent and ions will eventually emerge from the cone tip. As the droplets travels towards the orifice, solvent evaporate with the help of a warm neutral gas leading to droplets shrinking in size. With decreasing radius, repulsion between the charged ions will increase due to coulomb repulsion. When the repulsion overcomes the droplets surface tension, the droplets will break into minor droplets in a process called coulomb fission or coulomb explosion. The process

occurs multiple times eventually leading to charged ions in gas phase [9]. Figure 1 shows a

(8)

7

Figure 1: Formation of gas phase ions in ESI. Reprinted with permission from [11].

The mechanism of the ions leaving the solvent is still discussed among researchers and has not yet been exactly determined. Two main theories for this mechanism has been proposed and accepted. The ion evaporation model (IEM) introduced by Iribarne and Thomson (1976) proposes that the ions evaporate directly from the droplet surface as an answer to the repulsion force overcoming the surface tension at a certain radii. The first model accepted was described by Dole in 1968. The charged residue model (CRM) proposes that the charged

ions stay in the solvent until all solvent has evaporated[7] [8]. Both models are shown in

figure 2.

Figure 2: Ionization models in ESI. Reproduced from [7] Rights managed by Taylor &

Francis.

4.2.1 Solvent composition and ESI:

Polar volatile solvent or solvent mixes are often used in combination with ESI. Charge separation between the ions is depending on the connectivity of the solvent. Pure water as solvent is usually not optimal due to its high surface tension and viscosity making it harder for ions to move contributing to an unstable spray. Solvent evaporation process becomes

(9)

8 harder if pure water is used which could result in lower sensitivity for the analytes[12]. The best sensitivity for ESI has been detected when the analytes are already ionized in the liquid phase[13]. A basic solution for acidic analytes and vice versa. Although there has been research showing that acidic compounds are effectively ionized in acid solvent and achieving the highest sensitivity[13]. Due to redox reaction taking place at the capillary tip and solvent evaporation, pH could change significantly between the sample solution and the charged droplets[13]. The complexity of the ionization process makes it hard to predict how that affects the analytes[12] [13]. Liigand et. alt. concludes in their extensive studies that for some analytes the ionization effect is completely unaffected by pH while for other is the

sensitivity changed drastically with different pH[13]. Although the effects can be hard to

predict based on degree of ionization or pKa. The effect of solvent pH on ESI is therefore highly individual depending on various parameters such as solvent, pKa, analyte and acid/base.

4.3 Linear quadrupole ion trap:

Linear quadrupole ion traps consist of four metal rods which are applied with a constant voltage and an oscillating radio frequency voltage. The ends of the quadrupole can also be charged. If the end sections are positive compared to the center, positive ions get trapped in the center and vice versa for negative ions. The oscillating radio frequency applied at the central section acts in the xy-plan with respect to the figure 3. The voltage can be manipulated, shooting out ions with a specific m/z ratio at a specific voltage to detectors in x- or z-direction[14].

Figure 3: Basic design of two-dimensional linear ion trap. Reproduced and permission from [15] Copyright American Society for Mass Spectrometry, 2002.

4.4 Nanospray desorption electrospray ionization

Nanospray desorption electrospray ionization (nano-DESI) method is somewhat based on similar principles as electrospray ionization. The main difference is where the analyte is added. In nano-DESI is the analyte directly desorbed from the sample surface by a solvent bridge created between two capillaries. Solvent is delivered through the primary capillary and the analytes are picked up by the liquid bridge and then ionized when sprayed out of the

(10)

9 nanospray capillary. The nanospray capillary works on the same principle as electrospray ionization but as the name reveals is the capillary diameter much smaller which allows much smaller quantities of analyte. The electric field applied between the MS inlet and the primary capillary creates a “self-flow” for non-viscous solvents in Nano-DESI[16] [9]. The size of bridge can be manipulated down to sizes that sample analytes from areas of 10 µm in diameter[17].

Figure 4: Schematic nano-DESI setup in positive mode. Permitted and drawn by Hilde-Marléne Bergman.

5 Experimental:

5.1 Chemicals:

All water used during the project was MilliQ water. Buffers where made using ammonium bicarbonate. Liquid chromatography-MS (LC-MS) grade methanol (MeOH) (Merck) was used as a solvent. Analytes consisted of dissolved bromothymol blue (BTB)(Grave), methyl orange (M.O.)(Merck) and phenolphthalein (phph)(Merck). Formic acid (Sigma Aldrich, 98-100%) and ammonia adjusted the pH in the solutions. Sodium nitrate (NaNO3) was used for

extra addition of sodium ions to the samples.

5.2 Sample preparation:

5.2.1 First evaluation of several molecules

An ammonium bicarbonate buffer of 100 mM was prepared by weighing 0.77385g ammonium bicarbonate and diluting to 100 ml in a volumetric flask. 20 ml of the 100 mM solution was diluted further to 200 ml to the final buffer stock solution with a concentration of 10 mM ammonium bicarbonate. The stock solutions were divided into portions and titrated into different pH with formic acid and ammonia. Throughout the experiment was the eventual dilution factors from added formic acid and ammonium considered to be negligible. One batch of each pH 3, 6, 8 and 10 was prepared. Approximate 1 mg of each indicator was weighed up in 4 falcon tubes, giving 4 tubes with the same indicator and a total of 12 tubes. 4 ml of each pH buffer was added to each falcon tube. Although the indicators have different

(11)

10 molecular weight the concentration in all tube will roughly be 1 mM indicator. More exact concentration is not important at this point since the only purpose is to investigate how the probes are affected by different solvent pH when analyzed. Figure 5 shows a simplified image of the falcon tubes containing different indicators and pH.

Figure 5: Schematic image of the Falcon tubes with different indicators and pH.

60 µl indicator solutions, 1mM, were further diluted to a total volume of 6 ml with respectively pH buffer to a final concentration of 10 µM of each indicator.

5.2.2 Repeatability of methyl orange:

To investigate methyl orange further and test the repeatability was a large stock solution prepared to ensure that all samples contained the same concentration. The stock solution of 10 µM methyl orange in 10 mM ammonium bicarbonate was prepared by solving 0.0049 g methyl orange in 150 ml 10 mM ammonium bicarbonate solution. Approximately 5 ml of stock solution was poured into 18 falcon tubes. The tube solutions were titrated with formic acid and ammonia to get as close to each pH of 2.5, 4, 5.5, 7, 8.5 and 10 as possible. Finally, 3 samples of each pH were made. One sample from each pH was grouped together to form what will in this report be named as a series. Finally, three series were composed. This process is overviewed in figure 6.

(12)

11

Figure 6: Overview of the series composition.

5.2.3 Sodium’s effect on the intensity

To test whether the sodium ions effect the intensity, solutions with extra sodium was prepared as followed. 0.01941 g NaNO3 was diluted to 25 ml with 10 µM methyl orange

stock solution in a volumetric flask, concentration of sodium ions approximates 10 mM. 1 ml of every sample in methyl orange series used for the methyl orange repeatability tests where

mixed with 20 µl NaNO3. Final concentration of sodium ions where approximate 0.2 mM.

5.2.4 Methanol and water solution (50:50)

In attempt to get close to a real scenario, samples with less water content was prepared. 2 ml of the methyl orange stock solution, 10 µM, was measured up into six falcon tubes. Titration of each solution with formic acid and ammonia set solutions to pH 2.5, 4, 5.5,7, 8.5 and 10 with aid of a pH-meter. 1 ml of each pH solution was mixed with 1 ml MeOH.

5.2.5 Methyl orange solution for analyses of biological sample

Preparation of solution consisting of water and methyl orange for analysis on rat brain tissue section using nano-DESI was done accordingly. 0.00782 g methyl orange was weighed in and diluted with MilliQ water to 50 ml in a volumetric flask. The flask was ultrasonicated a few minutes and then filtered through OOH filter due to poor solubility. End concentration of the solution was 10 µM M.O. The solvent was then used to create the solvent bridge in the nano-DESI.

(13)

12

5.2.6 Brain tissue section

Two cryosectioned rat brain tissue sections with a thickness of 12 µm were placed on a glass-plate. The brain tissue sections were spiked with a solution of a specific pH to alter the pH on the tissue section. The spike-solution was prepared out of a 100 mM ammonia acetate buffer set to pH 5.7 and 10. 3 µl of respectively pH solution was applied to respectively brain tissue slice and left on until evaporated before analysis.

The rat brain tissue sections were a kind gift from Malin Andersson, UU. The experiment was approved by the Uppsala animal ethics committee (no. 140/8) and conducted in accordance with the guidelines of Swedish legislation on animal experimentation (Animal Welfare act SDS1998:56) and European Union Legislation (Convention ETS123 and Directive 86/609/EEC).

Although, this kind of analysis on rat brain tissue could awake ethical concerns, the possible gain is outweighed by the loss. Only a small amount of sample is needed, and these kinds of analysis could not be done differently.

5.3 Instrument

The mass spectrometer used is Finnigan™ LTQ™ MS detector from Thermo electron. Flow rate at all experiments where 10 µL/min. Different settings where used for analyzes of BTB, M.O. and phph for the first evaluation. These setting are presented in table 1:

(14)

13 Table 1: Instrumental setting for different samples.

Sample Ionization mode Sheath Gas Flow Electrospray Voltage (kV) Capillary Voltage (V) BTB pH 3 + 13 3 11 BTB pH 3 - 40 4 -40 BTB pH 6 + 13 3 11 BTB pH 6 - 36 4 -30 BTB pH 8 + 13 3 11 BTB pH 8 - 33 4 -35 BTB pH 10 + 13 3 11 BTB pH 10 - 40 4 40 M.O. pH 3 + 36 3 21 M.O. pH 3 - 36 4 -35 M.O. pH 6 + 13 3 11 M.O. pH 6 - 36 4 -36 M.O. pH 8 + 36 3 21 M.O. pH 8 - 37 4 -39 M.O. pH 10 + 40 3 30 M.O. pH 10 - 37 4 -39 phph pH 3 + 35 3 10 phph pH 3 - 20 4 -49 phph pH 6 + 37 3 12 phph pH 6 - 30 4 -35 phph pH 8 + 37 3 12 phph pH 8 - 36 4 -7 phph pH 10 + 37 3 12 phph pH 10 - 33 4 -1

Test of repeatability of M.O. as well as the extra sodium M.O. samples were analyzed at an electrospray voltage of 3kV, sheath gas flow at 36 and capillary voltage of 21 V. 2 microscans where used to get a more stable base line. Samples with 50% MeOH were analyzed with settings: capillary voltage 27.5 V sheath gas flow 68 and electrospray voltage of 3kV.

HI 208 pH-meter from Hanna instruments© was used for all pH-measurements. If no measurement was done within a week, the instrument was recalibrated before new measurements.

6 Results and discussion

6.1 First evaluation using several molecules

Chosen analytes, bromothymol blue, methyl orange and phenolphthalein, differ from each other in chemical structure and pKa, figure 7. Whereas pKa is a measure of the acid strength, ESI in positive mode responds to the ion formed when a proton is added[18]. To get a first

(15)

14 impression of the different analytes behavior was solutions with a pH range 3-10 for every indicator prepared and analyzed. From the spectra obtained was the base peak identified and the average intensity of the base peak was plotted against pH of the solution. This was done for spectra in both positive and negative mode. Resulting diagrams are shown in figure 8-12.

Figure 7: a) Bromothymol blue b) Methyl orange c) Phenolphthalein, all structures have been drawn in the program BIOVIA draw.

Figure 8: Average intensities for base peak, m/z 625, in BTB sample plotted against pH of solution analyzed in positive mode.

0 5000 10000 15000 20000 25000 30000 0 2 4 6 8 10 12 Int ens it y pH

I

_{[BTB 625]}

+

(16)

15 Figure 9: Average intensities for base peak, m/z 623, in BTB sample plotted against pH of solution analyzed in negative mode.

Plotting intensities of base peak for BTB in positive mode against pH of the solution, figure 8, a linear relation seems to exist between pH 6 and 10. Nothing could be said about the linear relation between pH 3 and 6 since there is only 2 data points. Although it´s not that clear, a linear relation may also exist between all pH in negative mode when base peak intensity is plotted against solution pH, figure 9. This could be of potential interest for further testing.

Figure 10: Average intensities for base peak, m/z 306, for methyl orange sample plotted against pH of the solution analyzed in positive mode.

0 20000 40000 60000 80000 100000 120000 0 2 4 6 8 10 12 Int ens it y pH

I

_{[BTB 623]}

-0 50000 100000 150000 200000 250000 300000 350000 400000 450000 500000 0 2 4 6 8 10 12 Int ens it y pH

I

_{[M.O. 306]}

+

(17)

16 Figure 11: Average intensities for base peak, m/z 304, for methyl orange sample plotted against pH of the solution analyzed in negative mode.

For methyl orange samples, plotting the base peak intensity against solution pH in positive mode, a clear trend could not be seen, figure 10. In negative mode, figure 11, there seems to be some kind of linear trend up till pH 8, this could also be of potential for future testing.

Figure 12: Average intensities for base peak, m/z 319, for phenolphthalein sample concentration 10 µM plotted against pH of the solution analyzed in positive mode.

For phenolphthalein, plotting base peak intensity against solution pH was more problematic. In positive mode, a base peak was found and could be plotted, figure 12. In negative mode the base peak given by the spectra differed with the pH of the solution making it problematic to find a trend and plot in a diagram. Since phenolphthalein was more problematic to analyze, decision was made to not analyze the molecule further.

0 20000 40000 60000 80000 100000 120000 140000 160000 180000 200000 0 2 4 6 8 10 12 Int ens it y pH

I

_{[M.O. 304]}

-0 5000 10000 15000 20000 25000 30000 35000 40000 45000 50000 0 2 4 6 8 10 12 Int ens it y pH

I

_{[phph 319]+}

(18)

17 Normalization of the data was done in attempt to find more possible trends. Average intensities for the base peaks of BTB in positive mode where divided by the average intensities from the base peak in negative mode, I[BTB 625]+/I[BTB 623]-. Plotted against the pH

of the solution gives the polynomial trend of second degree, this is shown in figure 13. The ratio is at highest at pH 6 which means that the protonated BTB molecule is more abundant in pH 6 when compared to the deprotonated BTB. Why not the protonated BTB is not more abundant in pH 3 can be due to the unpredictable ionization mechanism of ESI or the fact the BTB has two hydroxyl groups that can be protonated and that could affect the m/z ratio. Although the plot shows a good correlation coefficient, the parabolic shape of the curve can make it hard to distinguish between pH from a given ratio values. Practically it would be more difficult and time consuming to analyze two spectra required for plotting this ratio.

Figure 13: Intensities from base peak of positive and negative mode divided and plotted against pH of the solution.

Figure 14 shows a methyl orange mass spectrum at pH 7. The base peak showed at 306 which is the protonated methyl orange, I[M.O.+H]+. Methyl orange normally has a sodium adduct

which is detected at m/z 328, I[M.O.+Na]+. This sodium adduct is the one referred to when the

adduct peak of methyl orange is mention further in this report. The average intensities of these peaks divided by each other give a ratio, I[M.O.+Na]+/I[M.O.+H]+ or I[M.O.+H]+/I[M.O.+Na]+,

which are plotted against the pH of the solution, figure 15 and 16. By using the ratios the method is not as depending on good intensity values every time since they can vary due to various factors. R² = 0,995 0 0,05 0,1 0,15 0,2 0,25 0,3 0,35 0,4 0 2 4 6 8 10 12 I[BTB 625] +/I [B TB 623] -pH

I

_{[BTB 625]}

+

/I

_{[BTB 623]}

(19)

-18 Figure 14: Mass spectrum of methyl orange in pH 7. Protonated M.O. is the base peak at 306 and the sodium adduct of M.O. is peak at 328.

Figure 15: Average intensities of adduct peak and base peak as a ratio plotted against the pH of the solution, analyzed in positive mode.

0 0,01 0,02 0,03 0,04 0,05 0,06 0,07 0,08 0 2 4 6 8 10 12 I[M .O .+ N a] +/I [M .O .+ H ] + pH

I

_[M.O.+Na]+

_/I

[M.O.+H]+

(20)

19 Figure 16: Average intensities of base peak and adduct peak as a ratio plotted against the pH of the solution.

When the data is normalized more trends could be seen. Plotting ratio for base peak and adduct peak for methyl orange shows potential for linearly up to pH 8. Considering the facts that plotting ratio of base peak and adduct peak of methyl orange against solution pH looks most promising and more practical then BTB, methyl orange was investigated further.

6.2 Repeatability of methyl orange samples

Methyl orange sample series that was prepared as described in section “sample preparation” above was analyzed on multiple occasions to test whether the previous results and the sample analyses are repeatable. Average intensities of adduct peak and base peak as ratio is plotted against pH of the solution in figure 17 and 18.

0 5 10 15 20 25 30 35 40 0 2 4 6 8 10 12 I[M .O.+H] +/I [M .O.+N a ] + pH

I

_[M.O.+H]+

_/I

[M.O.+Na]+ y = 0,0359x - 0,0367 R² = 0,9793 y = 0,0407x - 0,1051 R² = 0,9903 y = 0,0081x - 0,004 R² = 0,9577 0 0,05 0,1 0,15 0,2 0,25 0,3 0 2 4 6 8 10 12 I[M .O.+N a ] +/I [M .O.+H] + pH

I

_[M.O.+Na]+

_/I

[M.O.+H]+ Serie1 Serie2 Serie3

(21)

20 Figure 17: Average intensities of adduct peak and base peak as a ratio plotted against the pH of the solution. Linear plot between pH 4 to 8.5

Figure 18: Average intensities of base peak and adduct peak as a ratio plotted against the pH of the solution. Polynomial trend line of second degree plotted between pH 4 and 8.5.

The most remarkable with figure 17 and 18 compared to figure 15 and 16 where the adduct ratio first was plotted is that the slope of the graphs are inverted. For the ratio adduct peak intensity over the base peak intensity, I[M.O.+Na]+/I[M.O.+H]+, is the slope negative in figure 15

and positive in figure 17. For the inverted ratio, base peak over adduct peak, I[M.O.+H]+/I[M.O.+Na]+, the slope is positive in figure 16 and negative in figure 18. The slopes of

figure 17 and 18 would be more accurate regards to pH argument where the concentration of protons increases with decreasing pH. More protons would lead to more protonated M.O. and increased intensity for the protonated base peak. Why the first plots are inverted is hard to explain. It could have something to do with the sample preparation since the samples used in the two experiment where prepared different. Something called “the wrong-way-round ionization” where analytes give high intensities in solvent condition they are not expected to be ionized should not be the case here since almost the same instrumental settings are used in both cases[13]. The instrument could also be affected in some way since various experiments was ongoing on the instrument in question at the time. The capillary leading to the mass analyzer was also taken out and cleaned during the time, which could be a contributing factor.

A linear trend could still clearly be seen between pH 4 and 8.5 when the adduct intensity over the base peak intensity is plotted with r2 value of 0.96-0.99, figure 17. Looking at the base peak plotted over the adduct peak, figure 18, a linear relation could no longer be seen as in figure 16. Instead is a polynomial plot best fitted with a r2 value of 0.98-0.99. Since pH 2.5 and 10 are so far off the plots they are excluded from regression plots from now on. Although the r2_are_good_{, both figure 17 and figure 18 shows a wide variation of slopes between the}

series, especially between series 3 and the rest. One explanation could be instrumental errors; R² = 0,98891 R² = 0,99890 R² = 0,99998 0 5 10 15 20 25 30 35 40 0 2 4 6 8 10 12 I[M .O.+H] +/I [M .O.+N a ] + pH

I

_[M.O.+H]+

_/I

[M.O.+Na]+ Serie1 Serie2 Serie3

(22)

21 the instrument was operated by multiple users with varying analytes during the period of question. The instrument also suffered from clogging during this period.

Same series injected at another occasion obtained the values that are plotted in figure 19 and 20 below. Figure 19 shows average intensities ratio for adduct peak over the base peak plotted against pH of the solutions of series 1 to 3. Figure 20 shows the average intensities for base peak over adduct peak plotted against the pH of the solution. The individual plots for all series with r2values and slope equation are presented in the appendix, figure A.1 – A.6.

Figure 19: Ratio of adduct peak and base peak average intensity plotted against pH of the solution from the different series.

Figure 20: Ratio of base peak and adduct peak average intensity against pH of the solution from the three different series.

0 0,02 0,04 0,06 0,08 0,1 0,12 0 2 4 6 8 10 12 I[M .O.+N a ] +/I [M .O.+H] + pH

I

_[M.O.+Na]+

_/I

[M.O.+H]+ Serie1 Serie2 Serie3 0 5 10 15 20 25 30 35 40 45 50 0 2 4 6 8 10 12 I[M .O.+H] +/I [M .O.+N a ] + pH

I

_[M.O.+H]+

_/I

(23)

22 Comparison between figure 19-20 and figure 17-18, the slopes are closer together in figure 19 and 20.

Disregard to the slope variation in previous tests, results from the reinjection shows that pH value 5.5 in series 2 lays way off from the rest of the datapoints in the same solution pH both in figure 19 and 20 and are therefore considered an outlier. The plots do follow the linear fit for the ratio adduct peak over base peak and a polynomial fit for base peak over the adduct peak. This proves that the results are somewhat repeatable since the same trend could be observed days later which makes it more interesting to investigate M.O. further.

6.3 Repeatability of the instrument

Even when sodium has been added the same trend can be seen in methyl orange samples with different pH. Linear trend for ratio intensities I[M.O.+Na]/I[M.O.+H] between pH 4 and 8.5 and a

polynomial trend line of second degree for ratio intensities I[M.O.+H]/I[M.O.+Na] between pH 4

and 8.5. This concludes that sodium in the solution has little or no effect on the trend given by the ratio and solution pH. This result shows that the method is of potential for the future since biological sample naturally contain salts.

Data point from series 2 at pH 5.5 differs from the other points with the same solution pH, figure 23 and 24. This point is considered an outlier. Since this sample also is consider an outlier in previous analyses, figure 19 and 20, there may be something wrong with the sample itself such as contamination.

6.5 Change of solvent composition

(28)

27 Figure 27: Ratio of average intensities for adduct peak I[M.O.+Na]+ and base peak I[M.O.+H]+

against pH of the solution for the 50% MeOH solutions and average water-based series, 10 µM methyl orange.

Figure 28: Ratio of average intensities for base peak and adduct peak against pH of the solution for 50% MeOH solutions and average water-based series.

Analyzing samples with MeOH and water buffer (50:50) gives intensities ratios I[M.O.+H]/I[M.O.+Na] close to only water-based samples prepared with the same pH of the

solutions, figure 27 and 28. Almost all values fit within the standard deviation ranges. Exceptions are two intensities ratios from samples with pH 7 and 10 with MeOH and water buffer (50:50) that differ quite a lot from the rest. Table 3 contain relative standard deviation between I[M.O.+H]/I[M.O.+Na] intensity ratios from water samples and the difference between

0 0,02 0,04 0,06 0,08 0,1 0,12 0,14 0,16 0,18 0 5 10 15 I[M .O.+N a ] +/I [M .O.+H] + pH

MeOH and water

50% MeOH solutions Average Water solutions 0 5 10 15 20 25 30 35 40 0 5 10 15 I[M .O.+H] +/I [M .O.+N a ] + pH

MeOH and water

50% MeOH solutions Average water solutions

(29)

28 MeOH and average water samples intensities as a factor. Obvious outliers from water buffer samples are excluded.

Table 3: Relative standard deviation of water buffer samples and ratio between intensities for MeOH and average water.

pH RSD water (%) Factor intensity

average water and MeOH 2.5 0.4 0.93 4 11.9 0.91 5.5 9.7 0.98 7 22.3 0.60 8.5 6.0 1.0 10 10.9 0.58

The factors in table 3 shows that almost all intensity values for MeOH samples lay close to the water sample intensities. Exceptions still is MeOH samples with pH 7 and 10 that differ a lot from the pure buffer samples. High RSD values for these pH shown in table 3 indicate differences in intensities between the water buffer series, the differ in factor may not only depend on errors from MeOH analyses. Reasons for such variations for pH 7 and 10 could be contamination of some kind and the datapoints could be considered outliers. More data would have to be obtained to draw a more valid conclusion, but this result is a very strong indication that the same trend could be seen in samples prepared with MeOH and water buffer solution as for the pure buffer solutions. This result is also a big step toward applying the method on real biological samples since nano-DESI demands a less viscous solvent than pure water to operate.

6.6 Spiked brain tissue section

Rat brain tissue section was used as a biological sample to investigate if methyl orange in the solvent would be seen when analyzed on real biological samples. A methyl orange solution with a concentration of 10 µM in water was used to form the solvent bridge in the nano-DESI instrument coupled to orbitrap mass analyzer. Since the water buffer is viscous, pneumatically assisted nano-DESI was employed. Three different spiked spots on the rat brain tissue section with pH 5.7 and 10 were then analyzed. The average intensity of the protonated methyl orange I[M.O.+H]+ over the sodium adduct methyl orange I[M.O.+Na]+ where

(30)

29 Figure 29: The average intensities for base peak and adduct peak against pH of the spiked spots of the rat brain.

Since there are no prognoses how the methyl orange will act when in contact with biological samples, low ratio values compared to ESI ratio values is nothing strange. Compared to samples analyzed using ESI is the average intensity ratio, I[M.O.+H]+/I[M.O.+Na]+, lower for pH 10

compared to pH 5.5. Ratios given from the brain sample follow the same trend, except for one spot spiked with pH 10 that have a higher ratio. Although more data from this kind of measurement would be needed to draw a conclusion, the result from the spiked brain samples show potential.

0 0,5 1 1,5 2 2,5 3 3,5 0 2 4 6 8 10 12 I[M .O.+H] +/I [M .O.+N a ] + pH

Brain sample

pH 5.7 pH 10

(31)

30

7 Conclusion

The search for suitable molecule probes for determination of pH using MS resulted in a more detailed investigation of methyl orange compared to bromothymol blue and phenolphthalein since M.O. was of more potential. Repeated analyses of M.O. showed that plotting average intensity ratios for the sodium adduct peak over the protonated base peak against solution pH show a linear trend between pH 4-8.5. The inverted ratio, intensities of base peak over intensities for adduct peak, show instead a polynomial trend of second degree. Even if the sodium adduct is used in evaluation are the trends independent of sodium concentration in the solvent. Test of repeatability for the M.O. samples resulted in acceptable RSD values indicating that the method is repeatable. Indications were obtained that the same trend could be seen when parameters change to simulate future analysis conditions, such as analyses of samples with lower water content and nano-DESI analyses on spiked biological samples. Although more tests are necessary for development of a validated method, the results obtained are a strong indication the M.O is of high potential for pH determinations using MS.

(32)

31

8 References

1. Granja S, Tavares-Valente D, Queirós O et al. Value of pH regulators in the diagnosis, prognosis and treatment of cancer. Seminars in Cancer Biology 2017;43:17–34.

2. Tannock IF. Acid pH in Tumors and Its Potential for Therapeutic Exploitation. :13.

3. Orlowski P, Chappell M, Park CS et al. Modelling of pH dynamics in brain cells after stroke. Interface focus 2011;1:408–16.

4. Zhang X-X, Wang Z, Yue X et al. pH-sensitive fluorescent dyes: are they really pH-sensitive in cells? Molecular pharmaceutics 2013;10:1910–7.

5. Han J, Burgess K. Fluorescent Indicators for Intracellular pH. Chem Rev 2010;110:2709–28. 6. Garza KY, Feider CL, Klein DR et al. Desorption Electrospray Ionization Mass Spectrometry Imaging of Proteins Directly from Biological Tissue Sections. Anal Chem 2018;90:7785–9.

7. El-Aneed A, Cohen A, Banoub J. Mass Spectrometry, Review of the Basics: Electrospray, MALDI, and Commonly Used Mass Analyzers. Applied Spectroscopy Reviews 2009;44:210–30.

8. Awad H, Khamis MM, El-Aneed A. Mass Spectrometry, Review of the Basics: Ionization. Applied Spectroscopy Reviews 2015;50:158–75.

9. Kebarle P, Verkerk UH. Electrospray: From ions in solution to ions in the gas phase, what we know now. Mass Spectrometry Reviews 2009;28:898–917.

10. Blades AT, Ikonomou MG, Kebarle P. Mechanism of electrospray mass spectrometry. Electrospray as an electrolysis cell. Analytical Chemistry 1991;63:2109–14.

11. Kebarle P, Tang L. From ions in solution to ions in the gas phase - the mechanism of electrospray mass spectrometry. Anal Chem 1993;65:972A-986A.

12. Kostiainen R, Kauppila TJ. Effect of eluent on the ionization process in liquid chromatography– mass spectrometry. Journal of Chromatography A 2009;1216:685–99.

13. Liigand J, Laaniste A, Kruve A. pH Effects on Electrospray Ionization Efficiency. Journal of The American Society for Mass Spectrometry 2017;28:461–9.

14. C. Harris D. Quantitative Chemial Analysis. 8th ed. New York: W. H. Freeman and Company, 2010.

15. Schwartz JC, Senko MW, Syka JEP. A two-dimensional quadrupole ion trap mass spectrometer. J Am Soc Mass Spectrom 2002;13:659–69.

16. J Roach P, Laskin J, Laskin A. Nanospray Desorption Electrospray Ionization: An Ambient Method for Liquid-Extraction Surface Sampling in Mass Spectrometry., 2010.

17. Nguyen SN, Sontag RL, Carson JP et al. Towards High-Resolution Tissue Imaging Using Nanospray Desorption Electrospray Ionization Mass Spectrometry Coupled to Shear Force Microscopy. J Am Soc Mass Spectrom 2018;29:316–22.

18. Svahn O, Björklund E. Increased electrospray ionization intensities and expanded

chromatographic possibilities for emerging contaminants using mobile phases of different pH. Journal of Chromatography B 2016;1033–1034:128–37.

(33)

32

Appendix

Slopes from analyses of samples series 1-3, repeatability of methyl orange samples.

Figure A.1: Ratio of adduct peak and base peak average intensity against pH of solution from series 1, concentration 10 µM methyl orange. Linear trend line plotted between pH 4 and 8.5

Figure A.2: Ratio of adduct peak and base peak average intensity against pH of solution from series 2, concentration 10 µM methyl orange.

[M.O.+H]+

(34)

33 Figure A.3: Ratio of adduct peak and base peak average intensity against pH of solution from series 3, concentration 10 µM methyl orange. Linear trend line plotted between pH 4 and 8.5

[M.O.+Na]+

(35)

34 Figure A.5: Ratio of base peak and adduct average intensity against pH of solution from series 2, concentration 10 µM methyl orange.

Figure A.6: Ratio of base peak and adduct average intensity against pH of solution from series 3, concentration 10 µM methyl orange. Polynomial trend line of second degree plotted between pH 4 and 8.5

Individual plots from analyses of the sample’s series with extra sodium. 0 5 10 15 20 25 30 35 40 0 2 4 6 8 10 12

I

[M.O.+Na]+

(38)

37 Figure A.11: Ratio of the average intensity of base peak and adduct peak against pH of the solution from series 2, concentration 10 µM methyl orange and 0.2 mM NaNO3.

Figure A.12: Ratio of the average intensity of base peak over adduct peak against pH of the solution from series 3, concentration 10 µM methyl orange and 0.2 mM NaNO3.

[M.O.+Na]+