Extraction and Quantitation of Ketones and Aldehydes from

Jun 26, 2015 - Alterations of these carbonyls are linked to diabetes,(1) epilepsy,(2) ..... PDF. ac5b01000_si_001.pdf (208.68 kB). The authors declare...
0 downloads 0 Views 2MB Size
Subscriber access provided by UB + Fachbibliothek Chemie | (FU-Bibliothekssystem)

Article

Extraction and Quantitation of Ketones and Aldehydes from Mammalian Cells using Fluorous Tagging and Capillary LC-MS WEI YUAN, Shuwei Li, and James L Edwards Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.5b01000 • Publication Date (Web): 26 Jun 2015 Downloaded from http://pubs.acs.org on July 5, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Extraction and Quantitation of Ketones and Aldehydes from Mammalian Cells using Fluorous Tagging and Capillary LC-MS Wei Yuan1,2, Shuwei Li1,2, James L. Edwards3* 1.Institute for Bioscience and Biotechnology Research, University of Maryland, 9600 Gudelsky Drive, Rockville, MD 20850, 2.Department of Chemistry and Biochemistry, University of Maryland, College Park, Maryland 20742 3. Department of Chemistry, Saint Louis University, St. Louis, Missouri, 63103

*To whom correspondence should be addressed. Phone +1 314 977 3624 E-mail: [email protected]

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 25

Abstract The extraction and quantitation of carbonyl metabolites from cell lysate was accomplished using a carbonyl-reactive fluorous tag and capillary liquid chromatography coupled to mass spectrometry (capLC-MS). Selective fluorous tagging for ketones and aldehydes provided a thirty-fold increase in sensitivity using electrospray ionization MS.

Separation of fluorous

tagged carbonyl resulted in good separation of all components and tandem MS was able to differentiate structural carbonyl isomers. The average limit of detection for carbonyl standards was 37 nM (range 1.5-250 nM) with linearity of R2>0.99. Reproducibility for metabolites in cell lysate averaged 9% RSD. Human aortic endothelial cells (HAECs) were exposed to varying levels of glucose and their carbonyl metabolite levels were quantified. Significant metabolite changes were seen in glycolysis and the propanoate pathway from a glucose challenge. Using an untargeted approach 120 carbonyl metabolites were found to change in hyperglycemic HAECs. From this list of compounds, multiple metabolites from the pentose phosphate and tryptophan metabolic pathways were discovered.

This system provides

excellent sensitivity and quantitation of carbonyl metabolites without the need of isotope standards or labels.

ACS Paragon Plus Environment

Page 3 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Introduction Endogenous ketone and aldehyde levels are well dispersed throughout cellular metabolic cycles.

Alterations of these carbonyls are linked to diabetes1, epilepsy2,

anorexia/malnutrition3and alcoholism4.

In particular, carbonyls are critical components of

sugar metabolism as monosaccharides are reducing sugars with a carbonyl group. Oxidative stress is also known to catalyze carbonyl formation5.

Despite their relevance to disease

states, ketone and aldehyde metabolites are under analyzed.

This likely stems from low

concentrations under basal conditions, reactivity with amines and high polarity.

These

characteristics complicate separation based analyses and indicate a need for high sensitivity, semi-targeted analytical systems to uncover novel metabolic changes. Tagging ketones and aldehydes through hydrazide or reductive amination chemistries to enhance signal/ionization in mass spectrometry (MS) has previously been effectively employed. In particular, Girard P and T reagents, carbonyl reactive groups with a quartenary amine have been successful in specific tagging of aldehydes for ESI-MS analysis6 and for carbonyls using MALDI-MS7. Tagging of glycans with hydrophobic tags have also been used to effectively enhance ionization efficiency for LC-MS analysis8,9. In a recent report, signal enhancement of carbonyls with anionic tagging achieved sub micromolar LODs10. Though these methods have been utilized for signal enhancement, the tagging itself is not used to extract the targeted analytes from the sample matrix. Extracting ketones/aldehydes is a promising but underdeveloped approach for quantitative analyses11-13.

Using amino-oxy reactive beads to extract intracellular

ketones/aldehydes in combination with LC-MS, has successfully purified carbonyls but detected only fourteen analytes.12 The sparse detection of metabolites is likely due to slow kinetics of the solid phase reaction14. In a recent carbonyl extraction with ESI-FTICR-MS 3 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 25

analysis, 30 metabolites were detected but quantitation was demonstrated for only one analyte15. As a result, enhancements in reaction kinetics, separation and sensitivity for metabolite extraction based on functional groups are needed to produce a higher yield of interesting metabolites. Fluorous tags have been shown to increase kinetics compared to solid phase reactions yet do not have the binding capacity limitation that biotin/avidin tags hold. Fluorous tags allow for extraction based on fluorous interactions rather than the more conventional polar or nonpolar interactions. One such example was demonstrated by reacting glucose with an amine containing fluorous tag through reductive amidation which was then analyzed by MALDI-MS13. This work suggests a role for fluorous tags in extracting and quantitating metabolites, though MALDI-MS is unable to differentiate isomers and suffers from difficulty in quantitation. In this work, we sought to develop an LC-MS method to target carbonyls for extraction, signal enhancement and quantitation from an in vitro model of diabetic complications. Ketones and aldehydes were reacted with fluorous tags and extracted from the intracellular metabolome. Extracts were analyzed using capillary LC-MS, showing good differentiation of structural isomers by both chromatographic separation and MS/MS fragmentation. Analyses of carbonyls from human aortic endothelial cells (HAECs) treated under low and high glucose environments were compared. Targeted and untargeted approaches demonstrated carbonyl metabolomic changes were consistent with hyperglycemic stimuli.

Materials and Methods Chemicals: All ketone and aldehyde standards and triethylammonium bicarbonate buffer (1M, pH 8.5) were purchased from Sigma Aldrich (St. Louis, MO). Aminooxy-N-(3perflurooctyl-propyl) acetamide and fluorous silica gel were purchased from Fluorous 4 ACS Paragon Plus Environment

Page 5 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Technologies Inc. (discontinued, Pittsburgh, PA). HPLC-grade solvents were purchased from Honeywell Burdick & Jackson (Muskegon, MI). Cell culture conditions HAECs were obtained from Lonza (Walkersville, MD) and maintained in endothelial growth medium (EGM) containing 2% fetal bovine serum (FBS), growth factors and 5 mM glucose in 6 cm tissue culture dishes. Cells were incubated with 5 or 30 mM glucose for 7 days in EGM containing 2% FBS. Metabolite extraction Metabolites were extracted as previously reported with slight modifications. In brief, cells were rapidly rinsed by warm PBS twice and quenched with 500 µl of ice cold 80:20 methanol -water (v/v) containing 2.5 µM 4-oxo-heptanedioic acid as internal standard followed by freezing in a dry ice/ethanol bath (-70°C). Cells were then scraped and lysed with a sonic dismembranator on ice using 10 1-second bursts at low power. Cell supernatants were collected for fluorous labeling after lysates were centrifuged at 14,000 rpm for 10 min at 4°C. Fluorous labeling of carbonyl containing metabolites The reaction of carbonyl fluorous labeling is shown in Figure 1. The amino-oxy group on aminooxy-N-(3-perflurooctyl-propyl) acetamide selectively reacts with carbonyl group on metabolites to form their carbonyl oxime derivatives at pH 5. For cell lysate labeling, the cell metabolites were extracted as described above at the end of 7 days treatment. Three hundred microliter of trimethylammonium bicarbonate buffer (30 mM, pH 5) was added to each lysate to reach the final concentration of trimethylammonium bicarbonate at 10 mM and the methanol at 50% (v/v). Twenty microliter of Aminooxy-N-(3-perflurooctyl-propyl) acetamide (50 mM) was then added to each reaction. The labeling reaction was conducted at 70°C overnight. The next

5 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 25

day, the samples were vacuum dried and reconstituted in 300 µl of 20% MeOH/H2O (v/v) for fluorous tagged carbonyl metabolites isolation. Fluorous tagged carbonyl containing metabolites isolation Fluorous silica gel was loaded in a spin column and prewashed by 1 ml MeOH three times followed by 1ml 20% MeOH/H2O (v/v) three times. Samples in 20% MeOH/H2O (v/v) were then slowly loaded on the fluorous silica gel. After washing by 1 ml 20% MeOH/H2O four times, the labeled metabolites were eluted by 2 ml MeOH. The sample was then dried by speedvac and reconstituted in 65% MeOH/H2O (v/v) containing 5 mM ammonium formate for LC/MS analysis. Flow injection analysis Different concentrations of ketone and aldehyde standards were incubated with excess amount of Aminooxy-N-(3-perflurooctyl-propyl) acetamide in 10 mM trimethylammonium bicarbonate in 50% methanol/H2O (v/v) (pH 5). pH of the reaction was adjusted to 5 by 2 M acetic acid. Labeled and non-labeled ketone and aldehyde standards with concentrations ranging from 0.1 to 100 µM were analyzed by flow injection ESI-MS. Flow analyses were performed using the divert/inject valve of the Finnigan LTQ Orbitrap discovery ion trap mass spectrometer fitted with a 5 µl sample loop and the Agilent 1200 HPLC. The mass spectrometer was equipped with electrospray ionization (ESI) interface. MS analyses were performed in negative mode. Spray voltage was 4 kV, sheath gas was 25 psi, auxiliary gas was 5 psi and capillary temperature was 300°C. Flow rate was 0.15 ml/min using 50% methanol/H2O(v/v) containing 10 mM formic acid. Selective ion monitoring (SIM) was used for quantitation.

6 ACS Paragon Plus Environment

Page 7 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Liquid chromatography (LC) / mass spectrometry (MS) conditions LC/MS analyses were performed in negative ion mode with a Thermo LTQ Orbitrap Discovery ion trap mass spectrometer (San Jose, CA) equipped with a nanospray ionization (NSI) interface coupled with an Agilent 1200 HPLC (Palo Alto, CA). The flow rate from the Agilent pump was 0.80 ml/min. A micro cross (Upchurch, Oakharbor, WA) was placed between the Agilent pump and capillary column to reduce the flow rate directed to the capillary column at 20-30 nl/min using a 75 µm I.D. silica capillary as the flow splitter. Samples were injected offline by removing the capillary column from the system and placing it in a enclosed chamber containing the sample. Pressure (400 psi) is applied to inject the sample directly onto the capillary column after which the column is reattached to the system and gradient started. This method allows for discrete, small sample plugs to be injected without the excess dead-volume seen in conventional six-port injection valves. Separations were performed on 50 µm I.D. silica capillary (Polymicro Technology, Phoenix, AZ) columns with in-house made frits packed with 3 µm Luna C8 reversed phase particles (Phenomenex, Torrance, CA). All columns had 20-30 cm packed bed length. Mobile phase A was 5 mM ammonium formate in H2O and mobile phase B was 5 mM ammonium formate in methanol. Analytes were eluted with a 25 min gradient from 80 to 100% solvent B. The temperature of the heated capillary was 200°C. Full mass scan was acquired in the Orbitrap with a resolution of 15,000. Fragmentation was activated by collision induced dissociation (CID) with a collision energy of 35%. Three most intense peaks in each survey MS scan was selected for data-dependent acquisition. The instrument control, data acquisition, and data analysis were performed by Xcalibur software (Thermo Electron Corporation, version 2.0.7 SP1). Data processing 7 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 25

Data was processed by XCMS on line version provided by Scripps Center for Metabolomics. Spectra were calibrated offline using 705.0893407529 m/z for oxo-heptadioic acid.

Mass

accuracy of known carbonyls in the HAEC lysate were between 0.1 and 5 ppm. Identification of untargeted full MS analysis was based on mass accuracy of 100 fold) compared to tagged glucose. 9 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Capillary Separation of F-Tagged Carbonyl Standards.

Page 10 of 25

The fluorous tag imparts a high

degree of hydrophobicity on the analyte. Tagged analytes were found to require a minimum of 20% organic solvent composition to remain soluble. Separation conditions were explored to optimize the resolution of small organic carbonyls. The standard carbonyls are found in Table 1. Initial work on this front utilized hydrophilic interaction chromatography (HILIC) so as to resolve the tagged analytes based on the organic portion rather than the ubiquitous fluorous tag. HILIC was attempted but had limited sample loading capacity and generally yielded poor resolution for the compounds attempted here. RPLC was used hereafter. RPLC was attempted for the nine standards with the expectation that the hydrophobic tag would yield stronger retention, larger sample loading and increased resolution. Figure 2A shows the separation of aldehyde and ketone standards using a C8 reverse phase column (50 µm i.d. × 20 cm). All compounds were well retained and gave good resolution. The width of the peaks is due to the extraordinarily strong interactions of the alkyl chains of the stationary phase and the analytes fluorous chain. As the addition of the fluorous tag adds 531 Da to the mass of each analyte whose molecular weight is typically below 200 Da, the resolution achieved for these compounds with 80% homology was remarkable. Previous work examining many of these same untagged compounds on a C18 reverse phase column yielded dramatically different retention order17.

This suggests the possibility of other retention

mechanisms involved in this system, with one possibility being differential folding of the long fluorocarbon tail depending on the metabolite. Further work is needed to dissect the exact mechanism(s) which are at play in this system. The performance of this system was substantial enough to provide resolution of the structural isomers acetoacetate and oxybutyrate. In fact, the resolution of these equimolar standards was 1.3. To gain further differentiation between these isomers, collision induced ACS Paragon Plus Environment

10

Page 11 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

dissociation was explored.

Despite their structural homology, substantial differences in

fragmentation was found (Figure 2B). Amino-oxy-fluoro-acetoacetate yielded an initial loss of H2CO2 from the analyte portion of the molecule followed by sequential losses of 20 Da (HF neutral losses). This fragmentation pattern was unique to acetoacetate as the remainder of the carbonyl standards all showed fragmentation similar to oxybutyrate which cleaved between the nitrogen and oxygen atom of the oxime bond. The nearly universal fragmentation of this N-O cleavage suggests precursor ion scanning using a triple quadrupole MS could allow for even higher sensitivity. Calibration curves were generated for these seven standards to demonstrate feasibility of quantitation in a real sample.

Standards were analyzed from 25 nM to 10 µM.

All

calibration curves showed excellent linearity with R2 > 0.99 (Table 1). The sensitivities for all compounds were high with the average being 1×107 counts/µM. Lower limits of detection were 2 nM for most of the compounds.

Completeness of the reaction was found to be

generally greater than 93%, except for glucose-6- phosphate which showed only 75% reactivity (Supporting Information, Table S2). This likely resulted in the glucose-6-phosphate as an outlier for both sensitivity and LOD, yielding considerably lower sensitivity and a higher LOD. While the exact reason for this low reaction efficiency is unknown, one possibility is that the unfolding of the aldehyde sugar is hindered by the presence of the phosphate group. Capillary Separation of F-Tagged Carbonyl of Mammalian Cells.

Fluorous tagging of

mammalian cells was undertaken to show the feasibility of this system to quantitate carbonyl levels in a biological system in response to an environmental stimulus. In order to simplify analysis and remove noncarbonyl compounds from the sample, an extraction of fluorous tagged species was undertaken. In this way, only carbonyls would be injected into the capLCMS system. Injection volume was optimized to 160 nL of HAEC sample which yielded the 11 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 25

highest number of metabolite peaks (over 3,000 features) as determined by the data analysis program XCMS. Separation of tagged carbonyls in biological samples showed good resolution of most targeted compounds (Figure 3). Glyceraldehyde-3-phospahte (Gly-3-P) was detected with high signal to noise next to an adjacent peak with similar nominal mass. This peak at 701 m/z was further examined for the possibility of being the isomer dihydroxyacetone phosphate (DHAP). Spiking the HAEC sample with DHAP showed a peak between the two existing peaks, indicating that the Gly-3-P peak was not co-eluting with DHAP and that endogenous DHAP was below the detection limits of our system.

Separation of acetoacetate and

oxobutyrate was significantly diminished in our sample (Rs 0.7) when compared to the standard (Rs 1.3). This is attributed to the difference in peak heights, as the standards had approximately a 1:1 signal ratio, whereas in the HAEC sample, signal intensity ratios were 1:8. To ensure proper identification and quantitation, targeted MS/MS for these isomers was performed. Investigations of reproducibility in HAEC lysate showed detection of all previous standards except for phenylpyruvate (Figure 5). The % RSD of peak height for the sample lysate was between 2.5% and 16.3% with glucose-6-phosphate again yielding the highest variance (Table 1). The capillary columns are manufactured in-house and inevitably hold variation between them. This resulted in some shifts in retention time. Upon normalizing retention time to the internal standard, targeted analytes in HAEC samples show good alignment with standards (Supporting Information Table S3) The low variance shown in biological samples indicated the ability to quantify biological samples. Targeted Carbonyls in Biological Samples. 12 ACS Paragon Plus Environment

Page 13 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

HAECs were treated with 30 mM glucose or 5 mM glucose for 7 days to simulate a diabetic and non-diabetic environment respectively. Six replicates of each condition were performed. As cell populations differed according to glucose concentrations, data were normalized to cell number18. Metabolite extraction and sample preparation was very reproducible as the internal standard data showed a ratio 1.02 ± 0.06 for 30 mM HAECs vs. 5 mM HAECs (Figure 5). Compounds matching standards on the basis of retention time, mass and fragmentation pattern were quantified in both glucose conditions.

As expected, changes in glycolysis

carbonyl metabolites (see Figure 4A) were the most pronounced19. Glucose, G-6-P, Gly-3-P and pyruvate all showed significant increases in response to high glucose. While the retention time and m/z match for derivatized glucose it cannot be excluded that such a peak may contain alternate structural isomers (hexoses). A predominant theory in diabetic complications is

that

an

enzyme

involved

in

glucose

metabolism,

glyceraldehyde

3-phosphate

dehydrogenase, becomes damaged during hyperglycemia and causes a build-up of the Gly-3P19. Flux analysis experiments are needed to confirm the role of these metabolites in such a hypothesis. α-Ketoglutarate showed no change in response to hyperglycemia. One possibility is that excess levels of ketoglutarate can be converted to glutamate.

Elevated levels of

glutamate have been previously shown under the same conditions as those presented here20,21.

Of particular interest was the elevation of acetoacetate and oxobutyrate (fold

change of 1.9 and 1.7 respectively), both components of the propanoate pathway (Figure 5). The propanoate pathway is scarcely investigated in the literature but –omics investigations have found this pathway up-reregulated in diabetic tissue22,23. Untargeted Carbonyls in Biological Samples. Analysis of these data in an untargeted format yielded >120 carbonyl metabolites with a significance of p