Goals of the work - 開發質譜法偵測生物代謝體之方法及應用

CHAPTER 1: Introduction

1.6 Goals of the work

In this study, we aim to develop mass spectrometric methods facilitating the analysis of metabolites in different kinds of biological samples: (i) single eggs of fruit fly, and (ii) human sweat. In the following chapaters, we provide information about the experiments conducted, and discuss technical obstacles that had to be overcome.

In order to measure metabolic rates in microscale samples obtained from individual fruit flies, we use MALDI-MS. However, MALDI-MS has poor quantitative capabilities. To solve this problem, we aimed to implement in-vivo labelling of fruit flies with ¹³C-labelled glucose, followed by subsequent monitoring of the labelling yields of a selected target metabolite by mass spectrometry.

In another study, in order to detect metabolites secreted by skin with sweat, we propose strategy which combines passive sampling with direct mass spectrometric analysis. This method should does not require any sample pretreatment. Since sweat samples are generally hard to collect and analyze by conventional analytical tools, we believe that setting up this method will open new possibilities for clinical analysis and doping control.

CHAPTER 2:

Isotope label-aided mass spectrometry reveals the influence of environmental factors

on metabolism in single eggs of fruit fly

2.1 Background

Circadian clock helps biological organisms to control their physiological and developmental processes.²⁴Biological rhythmicity is common in nature, with environmental factors – such as light and temperature – synchronizing internal time of the organisms to the 24-h cycle.²⁵ Although daily rhythms are often measured as activity vs. rest, other parameters – including the level of behaviour or gene expression – are also changing with a circadian period.²⁵ In humans, any disruption to the circadian rhythm can affect physical and mental performance. For example, insomnia is a common problem known to many travelers who experience jet lag.²⁶ Disruption of circadian rhythms can also lead to metabolic disorders.²⁷ Thus, studying relationships between circadian rhythms and metabolism may contribute to a better understanding of the mechanism and robustness of the biological clock. So far, most concepts in molecular regulation of circadian rhythms in eukaryotic cells have been based on transcription-translation feedback loops; however, a recent study demonstrated the existence

of a phenotypic circadian clock in red blood cells – i.e. in a biological system where no transcription occurs.²⁸

Small metazoa are convenient models for studying circadian rhythms; one of them is fruit fly (Drosophila melanogaster).²⁹ In fact, fruit fly is the most studied invertebrate, and a large amount of scientific information about this species is currently available.^30,31 Although fruit fly has served as a model organism in several studies of circadian rhythms,^32,33,34 limited data is currently available on the influence of circadian adaptation of fruit fly on primary metabolism in this species. The small volumes of samples obtained from individual flies disable the possibility of analyzing relevant metabolites using conventional analytical tools.

There exist a number of analytical techniques applicable to the analysis of metabolites in biological samples; two prominent examples include the coupling of liquid chromatography, or gas chromatography with mass spectrometry. The GC-MS platform enables analysis of volatile analytes in large numbers of samples; it offers high sensitivity, reproducibility, and can easily be automated.³ Implementation of chromatographic techniques usually necessitates pre-treatment of the samples. In most cases, this renders LC-MS and GC-MS inadequate to analysis of samples smaller than  1 mm. In the past few years, matrix-assisted laser desorption/ionization¹⁵ mass spectrometry has emerged as an enabling tool for the metabolic profiling of microscale samples;^35,36 it provides what the other analytical tools cannot offer: compatibility with micrometer-scale samples, high sensitivity, and it does not require complicated sample preparation. MALDI-MS takes advantage of laser beam to volatalize and ionize compounds embedded in a chemical matrix.³⁷ The matrix molecules absorb energy from the laser light, and transfer it onto the analyte molecules, which become ionized in the gas phase. A mass spectrum can be obtained right after a few shots of laser light impinge on a sample/matrix deposit. A major disadvantage of MALDI-MS is its

limited quantitative capability – a characteristic attributed to the limitations of sample preparation protocols, and ion suppression effects. When using MALDI-MS in the studies of metabolism, one possible solution to this problem is the implementation of stable-isotope labels.^38,39 The in-vivo labelling of fruit flies with stable isotopes – in combination with LC-MS – has already gained an insight into the proteome of fruit fly.⁴⁰ Therefore, to mitigate the problems related to the signal variability in MALDI-MS, here we have also implemented in-vivo isotopic labelling of metabolites in fruit flies.

In order to investigate the influence of circadian adaptation on egg metabolism, we have implemented the following protocol (Figure 2.1): First, fruit flies (D. melanogaster) are adapted to the day/night cycle using artificial white light. Second, female flies are incubated with ¹³C-labeled glucose (the only carbon source) for 12 h – either during the circadian day or circadian night, either at light or at dark. The labelled carbohydrate is ingested by the flies, and metabolized. Third, samples of eggs are obtained from the incubated flies, and the relative abundances of metabolite isotopologues present in individual eggs are determined by MALDI-MS. We sought answers to the following questions: (i) Will ¹³C-labelled glucose be used as a carbon source in primary metabolism, and will the ¹³C atoms be incorporated into metabolites in individual eggs? (ii) Can MALDI-MS provide useful quasi-quantitative results (i.e. without performing absolute quantification), which would reflect the treatment applied to the fly stocks (e.g. variation of temperature or light)? (iii) Does the circadian clock affect metabolite labelling in female fruit flies?

Figure 2.1. Experimental design and chemical analysis workflow. (1) pre-conditioning (entrainment) of the culture stock (adaptation to the 12/12-h (light/dark) cycle); (2) incubation of female fruit flies with

13C6-glucose solution; (3) dissection of the anesthetized flies; (4a/4b) preparation of individual eggs for mass spectrometric analysis; (5) mass spectrometry, and (6) data analysis.

We have opted for measuring metabolite levels in the samples composed of single eggs.

This choice was made due to several reasons: (i) Eggs can be considered as a sink for the absorbed nutrients and primary metabolites. (ii) Eggs occupy substantial volume in the fly abdomen, and the amount of the biological material contained in single egg is sufficient for the analysis by MALDI-MS. (iii) It is relatively easy to obtain multiple eggs from individual flies through manual dissection. (iv) Eggs are more compact, less vulnerable to osmolarity changes and mechanical stress, as compared with other fruit fly organs (e.g. ovarioles, gut, brain) which can be sampled for chemical analysis. Fruit fly eggs measure approximately half

millimeter, which can be regarded as an adequate size of a sample for the analysis by MALDI-MS following careful sample preparation.

2.2 Experimental section

2.2.1 Fly stocks

Fruit flies (Drosophila melanogaster; w¹¹¹⁸ – a normal control strain purchased from the Drosophila Stock Center in the Department of Biology at Indiana University, USA) were reared on a standard medium (water, yeast, soy flour, yellow cornmeal, agar, light corn syrup, propionic acid) loaded into plastic vials. Typically, the stock culture was maintained at room temperature. The default photoperiod was 16-hr day / 8-hr night; however, during the entrainment period (before the experiments related to the circadian rhythms), it was changed to 12-hr day / 12-hr night.

2.2.2 Isotopic labelling

Female and male fruit flies were separated under stereomicroscope (Zeiss, Munich, Germany), and the female individuals were subsequently transferred into 100-mL glass vials (Richiden-Rika Glass Company, Kobe City, Japan). A plastic cap with a stripe of filter paper wetted with 1% ¹³C₆-glucose solution in water was inserted to each of the vials, so that the flies were exposed to the ¹³C6-labelled glucose during the following hours/days. In most experiments, the vials were put inside an incubator in order to control the temperature. Most flies survived at least 7 days under these conditions. During the labelling experiments, illumination was provided by a light-emitting diode (LED) lamp (white light; Aliiv, Taipei, Taiwan), which ensured the illuminance of 4000 lux. During the entrainment period, weaker

17 light ( 150 lux) was used.

2.2.3 Dissection of flies and sample preparation

Before the dissection, flies were anesthetized with carbon dioxide gas. Heads and abdomens were separated from thoraxes using miniature scissors (Vannas-Tübingen Spring;

FST, Foster City, California). A set of precise tweezers (Dumont, Munich, Germany) was used to remove ovaries, and obtain unfertilized eggs (for a reference to the dissection protocol, see, for example: ref. 41). After brief washing in phosphate buffered saline solution, the eggs were transferred – one-by-one – onto separate recipient sites (i.d. 0.4 m) of a metal AnchorChip plate (Bruker Daltonics, Bremen, Germany). Following the deposition of the eggs onto the target plate, an aliquot of 0.5 L 1:1 (v/v) acetonitrile/water solution (after initial optimization) was pipetted to initiate extraction of metabolites from the egg; then, a 0.5-L aliquot of 6 mg mL^-1 9-aminoacridine (Sigma-Aldrich, St Louis, USA; MALDI matrix) solution in acetone (Merck, Darmstadt, Germany) was added. Following the evaporation of the solvents, the resulting sample/matrix deposits were ready for analysis by mass spectrometry. Note that 9-aminoacridine is a carcinogen, and personal safety equipment must be used when handling preparations of this compound. Animal tissue residues are disposed off as biological waste.

2.2.4 Mass spectrometry

During the mass spectrometric analysis, we used the AutoFlex III MALDI-time-of-flight (TOF)-MS from Bruker Daltonics, which is equipped with a solid-state laser (λ = 355 nm). Negative-ion mode was used with the following settings: ion source 1, -19.0 kV; ion source 2, -16.6 kV; lens, -8.45 kV; delay time, 0 ns. The

Smartbeam-laser focus was set to “small” (40 m); 100 shots were fired at each sample spot with the preset frequency of 50 Hz. The mass range was set to 400-1000 Da, and the suppression threshold was set to 400 Da, so that the low-m/z ions (including the matrix ions) could not reach the detector.

2.2.5 Data treatment

The MS data were acquired using the FlexControl software (version 3.0; Bruker Daltonics), and further analyzed with the FlexAnalysis software (version 3.0; Bruker Daltonics). The output data were further used to calculate the ratios of peak areas at the m/z:

571/(565+571). These ratio values were used to plot histograms with the bin width of 0.1.

Statistical analysis (one- and two-sample Kolmogorov-Smirnov test, Wilcoxon rank sum test) was conducted using the MATLAB software (version 7.6.0.324 (R2008a), MathWorks, Natick, USA). Curve fitting was conducted using the SPSS software (version 19, IBM Corp., New York, USA). Other data were treated and displayed using the Origin Pro software (version 8; Origin Lab Corporation, Northampton, USA).

2.3 Results and discussion

2.3.1 Preliminary experiments and optimization

Implementation of the proposed analytical workflow (Figure 2.1) has been preceded by a series of preliminary experiments. Initially, the in-situ extraction of metabolites from samples was tested and optimized. Metabolites were passively extracted from eggs on the MALDI target (AnchorChip, Bruker Daltonics), and co-crystallized with the matrix compound (9-aminoacridine). We chose to use acetonitrile as the extraction solvent since it has

widely been used for extraction of cellular metabolites (e.g. ref. ⁴²). During the optimization step, we tested mixtures of acetonitrile with water at different volume ratios. 9-Aminoacridine is a suitable matrix for MALDI-MS analysis of metabolites in the negative-ion mode.⁴³ Signals corresponding to several primary metabolites – which leaked out of the egg – could readily be identified in the MALDI mass spectra (e.g. Figure 2.2).

Figure 2.2. Wide m/z-range negative-ion MALDI mass spectrum of a single egg obtained from fruit fly.

Metabolites were extracted in situ using 50% acetonitrile solution. MALDI matrix: 9-aminoacridine.

Labels of the most prominent peaks: ADP, adenosine diphosphate (m/z 426); GDP, guanosine diphosphate (m/z 442); ATP, adenosine triphosphate (m/z 506); GTP, guanosine triphosphate (m/z 522);

UDP-Glc, uridine diphosphate glucose (m/z 565); UDP-GlcNAc, uridine diphosphate N-acetyl glucosamine (m/z 606).

The optimization of the sample preparation involved experiments in which the percentage of acetonitrile in the extraction solution as well as the concentration of 9-aminoacridine in the matrix cocktail were varied. Based on the measurements of signal-to-noise (S/N) ratios in the resulting spectra (Figures 2.3 and 2.4), we chose acetonitrile mixed with water at the ratio 1:1 (v/v) as the extraction solvent, and 6 mg mL^-1 9-aminoacridine solution in acetone as the MALDI matrix cocktail. It is believed that at low percentage of acetonitrile, biological membranes are not destabilized/degraded sufficiently to support the leakage of the contents of the cells. On the other hand, at high percentage of acetonitrile, evaporation of the extraction solution is too fast, and – as a result – the extraction time is too short, and the amounts of extracted metabolites are insufficient to produce intense MS signals. It should also be pointed out that – unlike the common extraction protocols used in metabolite analysis – the on-target extraction is almost completely passive, i.e. without prior sample degradation, grinding, shaking, or stirring. Metabolites need to be extracted despite the presence of the outer chorion layer protecting the egg. Although many standard protocols include the removal of the outer layer, the in-situ extraction process was conducted without prior dechorionation of the eggs. Based on the preliminary experiments, the outer protective layer of the eggs did not stop extraction of metabolites, and relatively high MS signals could be recorded (Figure 2.2).

Figure 2.3. Influence of the concentration of acetonitrile (ACN) in the extraction solution on the signal-to-noise (S/N) ratios. This experiment included several metabolites (ATP, m/z 506; GTP, m/z 522; UDP-glucose, m/z 565; and a phospholipid, m/z 833) extracted from single eggs, and analyzed by MALDI-MS.

Figure 2.4. Influence of the concentration of the 9-aminoacridine matrix solution (in acetone) on the signal-to-noise (S/N) ratios of the peaks of various standard compounds. We found that at 6 mg mL-1 9-aminoacridine (in acetone), the S/N value is highest. Therefore, we chose this concentration of 9-aminoacridine for further experiments.

22 2.3.2 Isotopic labelling of fruit flies

When ¹³C₆-glucose is administered to fruit flies as the only carbon source, the ¹³C atoms are gradually incorporated into cellular metabolites. Figure 2.5A shows the outcome of labelling over 1, 6 and 19 days. We found that – using 1% ¹³C-glucose solution as the only carbon source – uridine diphosphate glucose (UDP-Glc) is promptly labelled with ¹³C, while the labelling of other metabolites – for example, adenosine triphosphate (ATP) – is much slower, and it does not reach completion during several days of incubation. The latter is concluded based on the presence of the non-labelled form of ATP after 19 days (the peak at the m/z 506 in Figure 2.5A). Conversely, ¹³C-labeled glucose can readily replace the unlabelled glucose moiety in the molecule of UDP-Glc. Initially, the labelling of UDP-Glc is limited to the glucose (C6) moiety, while the UDP moiety remains unlabeled. Biosynthesis of UDP-Glc using UTP and glucose as substrates involves only two reactions, which are catalyzed by two enzymes: hexokinase (EC 2.7.1.1) and UTP-glucose-1-phosphate uridylyltransferase (EC 2.7.7.9; Figure 2.5B).^44,45,46 This gives the possibility of using partial isotopic labelling of UDP-Glc as an indicator of the early stages of primary metabolism while

13C6-glucose is the only carbon source available. On the other hand, de-novo biosynthesis of adenosine triphosphate (ATP) involves multiple biotransformations, thus only several carbons can be replaced in most ATP molecules during several days of incubation.

In order to confirm the incorporation of carbon-13 to the UDP-glucose molecule, we implemented MALDI-MS/MS: As shown in Figure 2.6, the fragment of the non-labelled UDP-glucose (MS m/z 565) was found to be non-labelled glucose phosphate (MS/MS m/z 241).

Figure 2.5. Isotopic labelling of ATP and UDP-Glc in single eggs. (A) MALDI mass spectra obtained from fruit flies following 1, 6, and 19 days of incubation with the ¹³C₆-glucose solution. (B) Reaction scheme of labelling glucose moiety in UDP-Glc with ¹³C₆-glucose.

The tandem MS analysis of the parent ion at the m/z 571 (in a sample obtained from a labelled fly) revealed a shift of the fragment peak to the m/z 247, which is due to the substitution of 6

12C atoms with ¹³C atoms in the glucose moiety of UDP-Glc. Since the mortality rate was high when the glucose solution was used as the only source of nutrients, we opted for short-term incubations ( 1 day), and – in further experiments – we focused on the measurement of the labelling of glucose moiety in UDP-Glc molecules. It should be noted that UDP-Glc is represented by an MS peak (m/z 565 or 571) that does not suffer from spectral/matrix interference.

Figure 2.6. MALDI-MS and MS/MS analysis of metabolites in the eggs of fruit flies incubated with ¹³C-glucose.

Left: spectra for eggs obtained from flies incubated with ¹²C₆-glucose. Right: spectra for eggs obtained from flies incubated with ¹³C₆-glucose for 12 h. Evaluation of the MS/MS data was aided by the METLIN database (Scripps Center for Metabolomics, La Jolla, USA).

Using 9-aminoacridine matrix, it is also possible to detect other metabolites (e.g.

adenosine diphosphate (m/z 426), guanosine diphosphate (m/z 442), adenosine triphosphate (m/z 506), guanosine triphosphate (m/z 522), uridine diphosphate glucose (m/z 565), uridine diphosphate N-acetyl glucosamine (m/z 606); Figure 2.2). Selecting UDP-Glc (m/z 565) as a target analyte, we further want to show that MALDI-MS is readily applicable to detection of metabolic effects of environmental factors in fruit fly eggs. In future studies, one may also consider studying the labelling patterns of other metabolites by MALDI-MS. However, one can anticipate that the interpretation of results of such studies may not be as straightforward as in the case of UDP-Glc. Another point to consider is that when ¹³C atoms replace ¹²C atoms, the signal-to-noise ratios of the main peaks (corresponding to the unlabelled metabolites) are decreased due to isotopic dilution. Therefore, the experimental strategy described here is most applicable to analysis of metabolites represented by peaks with high signal-to-noise ratios.

2.3.3 Time course of UDP-Glc labelling

Subsequently, the time progress of the incorporation of 6 ¹³C atoms to the glucose moiety of UDP-Glc in fruit fly eggs was studied. The incorporation of ¹³C to UDP-glucose takes approximately 20-30 h: during this time, a gradual increase of the ratio of the labelled UDP-Glc to the total (labelled + unlabelled) UDP-Glc can be observed (Figure 2.7). Based on the sigmoid function fitted to all the data points, the time (t1/2) after which the labelling of the studied population of eggs was 50% is estimated to be  13 h. A considerable variability of the labelling progress can be observed within the population of eggs analyzed at every time point (Figure 2.7). This variability will be discussed later on.

Figure 2.7. Progressive incorporation of the ¹³C-label to UDP-Glc in fruit fly eggs. The data points were fitted with a sigmoid function: y = 1.013·(1/(1+e(-0.135x+1.792))). During this experiment, the flies were incubated with ¹³C₆-glucose and illuminated with white light (~ 4000 lux). Note that the sampling intervals (as projected onto the x-axis) are coincidentally not constant.

It should also be pointed out that the application of matrix solution to the biological sample, and subsequent analysis in the vacuum compartment of MALDI mass spectrometer may cause some bias to the analytical result. However, this bias is greatly reduced by using isotopic labels: it is unlikely that the enzyme-catalyzed labelling process will proceed with high rate following the quneching of metabolism with acetonitrile used in the course of sample preparation.

2.3.4 Influence of the incubation conditions on the isotopic labelling of eggs

In the following experiment, we investigated the influence of temperature on the labelling of UDP-Glc. Two batches of female fruit flies were incubated under low (21 C) and high (28 C) temperature for 24 h; in both cases, light was on. Following the analysis by MALDI-MS, we found that the labelling level was significantly higher in the flies incubated at 28 C, as compared with the flies incubated at 21 C (Figure 2.8).

Figure 2.8. Influence of temperature (21 vs. 28 C) and illumination (dark vs. light) on the labelling of glucose

moiety in UDP-Glc molecules extracted from individual eggs. Female flies were incubated with

13C₆-glucose solution during 24 h. The default conditions were: white light on, ~ 4000 lux (in the study involving the change of temperature); temperature, 28 C (in the study involving the change of illumination).

In another experiment, we also found that illuminating the batch of female flies during the incubation with ¹³C6-glucose had strong effect on the labelling of UDP-Glc in eggs: the flies incubated in the dark metabolized much less ¹³C₆-glucose than the flies incubated under

在文檔中開發質譜法偵測生物代謝體之方法及應用 (頁 25-0)