Anion-Exchange Chromatography Coupled to High-Resolution

Subscriber access provided by Binghamton University | Libraries

Article

Anion-exchange chromatography coupled to high resolution mass spectrometry: a powerful tool for merging targeted and non-targeted metabolomics Michaela Schwaiger, Evelyn Rampler, Gerrit Hermann, Walter Miklos, Walter Berger, and Gunda Koellensperger Anal. Chem., Just Accepted Manuscript • Publication Date (Web): 05 Jun 2017 Downloaded from http://pubs.acs.org on June 10, 2017

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 18

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

Anion-exchange chromatography coupled to high resolution mass spectrometry: a powerful tool for merging targeted and non-targeted metabolomics Michaela Schwaiger†, Evelyn Rampler†, Gerrit Hermann†,‡, Walter Miklos§, Walter Berger§, Gunda Koellensperger†,∫,* † Department of Analytical Chemistry, Faculty of Chemistry, University of Vienna, Waehringer Str. 38, 1090

Vienna, Austria

‡ ISOtopic solutions, Waehringerstr. 38, 1090 Vienna, Austria §

Institute of Cancer Research, Department of Internal Medicine I, Medical University of Vienna, Borschkegasse 8a, 1090

Vienna, Austria ∫

Vienna Metabolomics Center (VIME), University of Vienna, Althanstrasse 14, 1090 Vienna, Austria

* Corresponding author: E-Mail: [email protected]

ABSTRACT In this work, simultaneous targeted metabolic profiling by isotope dilution and non-targeted fingerprinting is proposed for cancer cell studies. The novel streamlined metabolomics workflow was established using anion-exchange chromatography (IC) coupled to high resolution mass spectrometry (MS). The separation time of strong anion-exchange (2 mm column, flow rate 380 µL min-1, injection volume 5 µL) could be decreased to 25 min for a target list comprising organic acids, sugars, sugar phosphates and nucleotides. Internal standardization by fully

13C

labeled Pichia

pastoris extracts enabled absolute quantification of the primary metabolites in adherent cancer cell models. Limits of detection (LODs) in the low nM range and excellent intermediate precisions of the isotopologue ratios (on average < 5% N=5 over 40 hours) were observed. As a result of internal standardization, linear dynamic ranges over 4 orders of magnitude (5 nM – 50 µM, R2 > 0.99) were obtained. Experiments on drug-sensitive versus resistant SW480 cancer cells showed the feasibility of merging analytical tasks into one analytical run. Comparing fingerprinting with and without internal standard proved that the presence of the

13C

labeled yeast extract required for absolute

quantification was not detrimental to non-targeted data evaluation. Several interesting metabolites were discovered by accurate mass and comparing MS2 spectra (acquired in ddMS2 mode) with 1 ACS Paragon Plus Environment


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

spectral libraries. Significant differences revealed distinct metabolic phenotypes of drug-sensitive and resistant SW480 cells.

Today, high resolution mass spectrometry (MS) has reached a point that hypothesis driven absolute quantification studies can be performed using accurate mass profiling methods rather than using multiple reaction monitoring by triple quadrupole MS. In fact, there is a paradigmatic shift recognizing the capabilities of modern high resolution MS as quantitative tools.1 High resolution MS and simplified workflows,2 will be the key to improve the analytical throughput in metabolomics. Starting point of the establishment of any metabolomics toolbox should be the targeted approach. Accuracy of quantification demands validation of sample preparation procedures regarding extraction efficiency and recovery, quenching efficiency, and cell leakage upon quenching.3,4 In turn, this facilitates the successful establishment of non-targeted procedures, as the developed sample preparations routinely address metabolites of widely differing chemical and physical properties.5 Apart from this fact, comprehensive metabolomics workflows involve both hypothesis driven targeted quantification in parallel to non-targeted analysis for the discovery of metabolic rearrangements drawing a clear line between the two essential tasks. Up to now, hypothesis generation and hypothesis validation is addressed as separate measurement tasks on different MS platforms.2,6 As a consequence, state of the art metabolomic experiments require multiple analytical runs on one sample. Typically, hydrophilic interaction liquid chromatography (HILIC) and reversed phase (RP) chromatography (MS measurements in positive and negative mode using acidic and basic pH) in combination with high resolution MS are implemented as discovery tools.7 As both separation methods show poor selectivity for the central carbon metabolome, comprehensive workflows often comprised targeted analysis dedicated to exactly this metabolite panel next to nontargeted analysis.8 Only recently, a seminal critical review indicated the potential of merging different analytical tasks into simplified workflows.2 Reducing the complexity of the sample by on line derivatization with p-toluenesulfonylhydrazine, merging non-targeted and targeted analysis of aldehydes and ketones in one analytical run was enabled.9 In this work, we address a merged workflow based on anion-exchange chromatography (IC) in combination with high resolution MS detection. Anion chromatography proved to be a valuable separation method for key metabolites of the primary carbon metabolome,10–17 offering an alternative to the state of the art strategies based on gas chromatography (GC)18 and ion pairing LC.19–22 GC is unrivalled regarding the selectivity of the separation of sugars and sugar phosphates, 2 ACS Paragon Plus Environment

Page 2 of 18

Page 3 of 18

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


however, falls short when nucleotides are of interest. Compared to IC-MS, LODs in GC-MS/MS are on average somewhat higher (factor of 2-10). E.g. for sugar phosphates LODs ranging between 1200 fmol were obtained by GC-MS/MS.18 As a drawback, derivatization required in GC-MS is depending on the sample matrix and on the degree of automation (just-in-time derivatization by a robotic system is desirable). Ion pairing chromatography offers the highest coverage of the primary metabolome.22,23 Nevertheless, it has been only rarely coupled to high resolution MS,20,24 as the use of an ion pairing reagent demands for the establishment of a dedicated instrumentation.25 In fact, the contamination of ion pairing reagent is detrimental for other applications. Moreover, the general sensitivity is compromised due to suppression effects. Especially for nucleoside diphosphates and triphosphates rather high LODs ranging from 50 to 200 ng mL-1 (sub-µM) have been reported.20 Ion chromatography, more specifically anion-exchange chromatography, provides comparable separation selectivity to ion pairing chromatography. However, molecules that can be protonated (e.g. amino acids) are lost in the suppression system required for MS detection and are hence not amenable to subsequent MS analysis. As a matter of fact, many metabolites of the central carbon cycle occur in multiple isomers and are prone to in-source fragmentation. Accordingly, their separation becomes an absolute prerequisite for the two essential tasks accurate quantification and compound identification from MS fingerprints. Anion-exchange chromatography is a powerful separation method for metabolomic applications. Table S-1 gives an overview on the state of the art anion-exchange MS couplings in the field.10–17 So far, few studies covered the different compound classes amenable for anion-exchange by normal flow chromatography comprehensively. If the selected metabolite panel was small, the run time was reduced to 20 min. E.g. a recent work15 addressed IC-high resolution MS for absolute quantification of 6 organic acids based on standard addition via

13C

labeled standards was

implemented as quantification strategy. Moreover, several sugar phosphates were identified by their established non-targeted evaluation pipeline using the Metlin database (La Jolla, CA, USA).15 Moreover, a dedicated method for the quantification of 28 organic acids with run times of 19 min was published recently.17 Capillary IC shows longer run times (45 – 75 min, see Table S-1) but was propagated with the idea of increasing robustness and sensitivity, as microflow would decrease the potential salt contamination of the ionization source. The increased sensitivity results from maximizing the injection volume as typically 5 µL are injected into a flow of 25 µL min-1.11,14 In this work, we propose a streamlined workflow by IC-high resolution MS (Q Exactive HF) integrating targeted quantification and non-targeted fingerprinting into one analytical run. We addressed internal standardization based on fully

13C

labeled Pichia pastoris extracts. The

3 ACS Paragon Plus Environment


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

capabilities of this versatile internal standardization strategy have been shown in several studies.23,26–28 In order to ensure absolute quantification in metabolomics, isotope dilution is required. As a consequence, in this work we investigated whether the addition of the internal standard impeded non-targeted fingerprinting with regard to compound annotations and differential analysis between different sample groups. A proof of principle study addressed the metabolome of drug sensitive colon cancer cell as compared to the model of acquired resistance.

EXPERIMENTAL SECTION Metabolite standards, internal standard and solvents The metabolite standards for organic acids, sugar phosphates and nucleotides were purchased from Sigma-Aldrich or Fluka (Vienna, Austria) except of malic acid, which was purchased from Merck (Vienna, Austria). Standard stock solutions of 1 or 10 mM were prepared in water and used for the preparation of a multi-component standard of 0.5, 1, 5, 10, 50, 100, 500 nM and 1, 5, 10, 50 µM. Additionally, a quality control (QC) of 1 µM was prepared. A fully 13C labeled yeast extract of Pichia pastoris (2 billion cells) from ISOtopic solutions e.U., (Vienna, Austria) was reconstituted in 2 mL water and added in same amounts to the calibration standards, the QC as well as to the samples. The final dilution of the internal standard for the measurement was 1:10 (v/v). Deionized water from an ultra-pure water system (resistance 18.2 MΩ) was used for ion chromatography. Furthermore, LC-MS grade methanol from Fluka (Vienna, Austria) was used as make-up flow.

Cancer cell culture and sample preparation The human colon carcinoma cell line SW480 was purchased from the American Tissue Collection Center (ATCC). The sensitive cells were compared to its subline of acquired triapine resistance.29 Five times 1 x 106 of each cell type were seeded in 6-well plates in minimal essential medium (MEM) containing 10% fetal calf serum without antibiotics and incubated for 24 h (37 °C, 5% CO2). The sampling procedure was based on the “direct solvent scraping” method.30 After 24 h, the medium was removed and the wells were washed three times with 1 mL of PBS solution. Then, 50 µL internal standard and 1 mL ice-cold methanol (80% methanol, 20% water, v/v) were added. The cells were scraped into the extraction solvent and transferred to an Eppendorf tube. Then, the cell scraper and the wells were washed two times with 475 µL extraction solvent to reach a final volume of 2 mL extract. After thorough mixing and centrifugation (20,000 RCF, 5 min, 4 °C), the 4 ACS Paragon Plus Environment

Page 4 of 18

Page 5 of 18

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


supernatant was aliquoted into HPLC vials (400 µL were dried under reduced pressure and reconstituted in 100 µL water) whereas the cell pellet was used for protein determination with the 2D-Quant Kit (GE Healthcare, Munich, Germany). For the non-targeted approach, additionally samples without internal standard were prepared, thus the wells were washed with 2 x 500 µL to reach the final volume of 2 mL. Moreover, a pooled sample was prepared by mixing same aliquots of all samples (separately for extracts with and without internal standard). Samples were measured in a randomized way. After each tenth injection, the QC sample was analyzed. Additionally, calibration standards were measured in the beginning, the middle and the end of the entire sequence. For method validation, samples without internal standard were measured in a separate block.

Ion chromatography A Dionex Integrion HPIC System (Thermo Scientific) was used for anion-exchange chromatography. The separation was conducted on a Dionex IonPac AS11-HC column (2 x 250 mm, 4 µm particle size, Thermo Scientific) equipped with a Dionex IonPac AG11-HC guard column (2 x 50 mm, 4 µm, Thermo Scientific) at 30 °C. A potassium hydroxide gradient was produced by an eluent generator with a potassium hydroxide cartridge that was supplied with deionized water. The separation was carried out with a step gradient at a flow rate of 0.380 mL min-1 beginning with 10 mM KOH over 3 min, 10-50 mM from 3 to 12 min, 50-100 mM from 12 to 19 min, held at 100 mM from 19-21 min and re-equilibrated at 10 mM for 4 min. This resulted in a total run time of 25 min. A Dionex AERS 500, 2 mm suppressor was used to exchange potassium ions against protons in order to produce water instead of potassium hydroxide before entering the mass spectrometer. It was operated with 95 mA at a temperature of 15 °C. Furthermore, methanol was provided as make-up flow at a flow rate of 0.150 mL min-1. The samples were introduced via a Dionex AS-AP autosampler and full loop injection on a 5 µL loop (overfill factor 3). The temperature of the autosampler was set to 6 °C.

Mass spectrometry High resolution mass spectrometry was conducted on a high field Thermo Scientific™ Q Exactive HF™ quadrupole-Orbitrap mass spectrometer equipped with an electrospray source. Full mass scan (full MS, 50 – 750 m/z) in negative mode was used at a resolution of 120,000 for all calibration standards and samples. The automatic gain control (AGC) target was set to 1 x 106 ions and the maximum injection time (IT) was 200 ms. The ESI source parameters were the following: sheath gas 50, auxilary gas 14, sweep gas 3, spray voltage 2.75 kV, capillary temperature 230 °C, S-Lens RF level 45 and auxiliary gas heater 380 °C. Spectrum data were acquired in profile mode. For the nontargeted approach, also data-dependent MS2 (ddMS2) fragmentation spectra of the 1 µM standard mix spiked with internal standard and of a pooled sample (prepared by mixing aliquots of all 5 ACS Paragon Plus Environment


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

samples) were acquired. A top 5 method with an AGC target of 1 x 105, a maximum injection time of 50 ms and a minimum AGC target of 1 x 103 was used. The ions were isolated with 1 m/z, fragmented with HCD energy NCE 30 and detected with a resolution of 30,000.

Data processing Targeted data evaluation for the quantification of organic acids, sugar, sugar phosphates and nucleotides was performed in TraceFinderTM 3.3 from Thermo Scientific™ with internal standardization. All calibration curves were linear and weighted 1/x. For non-targeted data processing, Thermo Scientific™ Compound Discoverer™ 2.0 software was used. This software combines feature detection with statistical data evaluation. Retention time alignment within 0.15 min and 5 ppm mass tolerance was performed. For the detection of unknown features on MS1 level, a minimum peak intensity of 10,000, 3 ppm mass tolerance and a minimum number of 2 isotopes were used. Unknown compounds were grouped according to a mass tolerance of 3 ppm and 0.15 min. The “Fill Gaps” node was used with 3 ppm and 0.1 min. mzCloud search was performed with 5 ppm mass tolerance and an assignment threshold for compound annotation of 70.

RESULTS AND DISCUSSION IC high-resolution MS of metabolites In this work, merging absolute quantification based on isotope dilution and non-targeted fingerprinting by normal flow IC-high resolution MS (Q Exactive HF) was evaluated for a 45 metabolite panel. Starting point of the study was the separation of a 45 metabolite standards including organic acids, nucleotides and sugar phosphates on the IonPac AS11-HC column using normal flow IC and relatively short separation times of 25 min. Establishing a gradient up to 100 mM KOH, nucleoside triphosphates could be successfully eluted obtaining good peak shapes (see Figure S-1). All mono-, di- and triphosphates of the nucleosides and the isomers deoxyguanosine-triphosphate (dGTP) and adenosine-triphosphate (ATP) were baseline separated. Due to the short gradient, only limited selectivity was obtained regarding sugar phosphates. In the case of hexose-monophosphates, G1P and F1P were baseline separated, whereas the 6-phosphates of glucose, fructose and mannose were not. Fructose-1,6-bisphosphate was successfully separated from all hexose-monophosphates enabling absolute quantification despite in source fragmentation (see Figure S-1). In fact, significant loss of one phosphate group was observed resulting in the isobar fragment at m/z 259.0224 otherwise overlapping with the hexose-monophosphates. No separation could be achieved for the pentose-5-phosphates (ribose-5-phosphate, ribulose-5-phosphate). The 6 ACS Paragon Plus Environment

Page 6 of 18

Page 7 of 18

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


two isomers citric acid and isocitric acid were almost baseline separated from each other and separated from cis-aconitic acid. The latter was crucial, as citric and isocitric showed in source fragmentation which could not be resolved from cis-aconitic acid by mass selectivity (m/z of 173.0092 [M-H]-). Oxaloacetic acid was separated from pyruvic acid and malic acid from fumaric acid, respectively. Again this was crucial since pyruvic and fumaric acid are produced as fragments of oxaloacetic and malic acid, respectively, in the ESI source (see Figure S-1). Table 1 gives an overview on the selected metabolite panel and their respective retention times.

Internal standardization with uniformly 13C labeled metabolites produced from Pichia pastoris It is a commonly accepted fact that an isotope ratio based approach is a prerequisite for accurate quantification in metabolomics.23 In this study, we implemented a fully

13C

labeled cell extract

derived from the yeast Pichia pastoris as compound specific internal standard providing nearly full coverage of the selected metabolite panel (see Table 1). In the past, this internal standardization strategy enabled accurate quantification of primary metabolites as proven by interlaboratory31 and inter-platform comparisons.32 For the selected metabolite panel of this study, the concentrations of the uniformly following

13C

13C

labeled analogs were found in the low nM up to several µM range. Only the

compounds revealed concentrations too low for internal standardization: the four

deoxytriphosphates present in the DNA (dATP, dCTP, dGTP, dTTP), dCMP, the cyclic forms cAMP and cGMP, CMP, isocitric acid, oxaloacetic acid and mannitol-1-phosphate. In the case of the poorly separated hexose-6-phosphates, the sum of the 13C peaks was used to correct for G6P, F6P and M6P. The internal standardization strategy for all compounds is given in Table 1.

Table 1. List of 45 metabolite standards, retention times and the uniformly 13C labeled metabolites used for internal standardization. Compound

2-Phosphoglyceric acid (2PG) and 3PG AMP 6-Phosphogluconic acid ADP alpha-Ketoglutaric acid ATP cAMPa

RT [min]

12C

metabolite

Uniformly 13C labeled internal standard ISTD [M-H]-

Formula

[M-H]-

ISTD compound

ISTD formula

13.65

C3H7O7P

184.9857

U13C 2PG + 3PG

13C3H7O7P

187.9957

11.77

C10H14N5O7P

346.0558

13C10H14N5O7P

356.0894

12.67

C6H13O10P

275.0174

13C6H13O10P

281.0375

16.90

C10H15N5O10P2

426.0221

13C10H15N5O10P2

436.0557

10.68

C5H6O5

145.0143

13C5H6O5

150.0310

19.82

C10H16N5O13P3

505.9885

U13C AMP U13C 6Phosphogluconic acid U13C ADP U13C alphaKetoglutaric acid U13C ATP

13C10H16N5O13P3

516.0220

11.95

C10H12N5O6P

328.0452

U13C AMP



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

Compound

RT [min]

12C

metabolite

Page 8 of 18

Uniformly 13C labeled internal standard

CDP

14.29

C9H15N3O11P2

402.0109

U13C CDP

13C9H15N3O11P2

ISTD [M-H]411.0411

cGMPa

19.40

C10H12N5O7P

344.0402

cis-Aconitic acid

15.49

C6H6O6

173.0092

13C6H6O6

179.0293

Citric acid

14.45

C6H8O7

191.0197

U13C GMP U13C cis-Aconitic acid U13C Citric acid

13C6H8O7

197.0399

C9H14N3O8P

322.0446

U13C GMP 13C9H16N3O14P3

491.0074

13C6H14O12P2

345.0089

13C6H13O9P

265.0426

13C6H13O9P

265.0426

Formula

[M-H]-

ISTD compound

ISTD formula

CMPa

9.02

CTP

17.72

C9H16N3O14P3

481.9772

U13C CTP

dATPa

19.58

C10H16N5O12P3

489.9936

U13C ATP

dCMPa

7.91

C9H14N3O7P

306.0497

U13C GMP

dCTPa

17.12

C9H16N3O13P3

465.9823

U13C CTP

dGTPa Fructose-1,6bisphosphate (FBP) Fructose-1phosphate (F1P) Fructose-6phosphateb (F6P) Fumaric acid

22.14

C10H16N5O13P3

505.9885

17.29

C6H14O12P2

338.9888

9.81

C6H13O9P

259.0224

10.59

C6H13O9P

259.0224

11.28

C4H4O4

115.0037

U13C GTP U13C Fructose-1,6bisphosphate U13C Fructose-1phosphate U13C Hexose-6phosphates U13C Fumaric acid

GDP Glucose-1phosphate (G1P) Glucose-6phosphateb (G6P) GMP

21.20

C10H15N5O11P2

442.0171

7.60

C6H13O9P

259.0224

10.32

C6H13O9P

259.0224

18.35

C10H14N5O8P

GTP

22.67

Hexoses

2.27

IMP Isocitric

acida

13C4H4O4

119.0171

13C10H15N5O11P2

452.0506

13C6H13O9P

265.0426

13C6H13O9P

265.0426

362.0507

U13C GDP U13C Glucose-1phosphate U13C Hexose-6phosphates U13C GMP

13C10H14N5O8P

372.0843

C10H16N5O14P3

521.9834

U13C GTP

13C10H16N5O14P3

532.0169

C6H12O6

179.0561

U13C Hexoses

13C6H12O6

185.0762

13C10H13N4O8P

357.0734

18.08

C10H13N4O8P

347.0398

U13C IMP

15.00

C6H8O7

191.0197

U13C Citric acid

Lactic acid

2.70

C3H6O3

89.0244

U13C Lactic acid

13C3H6O3

92.0345

Malic acid

9.12

C4H6O5

133.0143

U13C Malic acid

13C4H6O5

137.0277

13C6H14O6

187.0919

13C6H13O9P

265.0426

13C5H11O8P

234.0287

13C3H5O6P

169.9852

Mannitol Mannitol-1phosphatea Mannose-6phosphateb (M6P) Oxaloacetic acida Pentose-5phosphates Phosphoenolpyruvic acid Pyruvic acid Sedoheptulose-7phosphate Succinic acid

2.12

C6H14O6

181.0718

7.47

C6H15O9P

261.0381

10.75

C6H13O9P

259.0224

12.47

C4H4O5

130.9986

11.13

C5H11O8P

229.0119

15.40

C3H5O6P

166.9751

3.57

C3H4O3

87.0088

C7H15O10P

289.0330

C4H6O4

117.0193

U13C Mannitol U13C Hexose-6phosphates U13C Hexose-6phosphates U13C Fumaric acid U13C Pentose-5phosphates U13C Phosphoenolpyruvic acid U13C Pyruvic acid U13C Sedoheptulose7-phosphate U13C Succinic acid

11.28 9.07

TMP

14.63

C10H15N2O8P

321.0493

U13C TMP

TTPa

20.76

C10H17N2O14P3

480.9820

U13C ATP

UDP

19.34

C9H14N2O12P2

402.9949

U13C UDP


13C3H4O3

90.0188

13C7H15O10P

296.0565

13C4H6O4

121.0328

13C10H15N2O8P

331.0829

13C9H14N2O12P2

412.0251

Page 9 of 18

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


Compound

RT [min]

12C

metabolite

Uniformly 13C labeled internal standard

UMP

15.80

C9H13N2O9P

323.0286

U13C UMP

13C9H13N2O9P

ISTD [M-H]332.0588

UTP

21.11

C9H15N2O15P3

482.9613

U13C UTP

13C9H15N2O15P3

491.9914

Formula

[M-H]-

ISTD compound

ISTD formula

In case no 13C equivalent was available, a similar compound was used as internal standard (italic font). b G6P, F6P, and M6P were not baseline separated. For internal standardization the sum of the 13C peaks (U13C Hexose-6-phosphates) was used. a

Analytical figures of merit of targeted metabolomics A 1 µM multi-metabolite standard spiked with the yeast based

13C

internal standard was injected

five times over a period of 40 h in order to assess the intermediate repeatability of retention time, calculated concentrations and peak areas. The obtained results based on full MS data acquisition are shown in Table 2. The internal standard derived from yeast added to the calibration standard is at the same time a perfect mimic of a biological matrix. In spite of a gradient up to 100 mM and the high flow rate of 380 µL min-1, the intermediate repeatability regarding retention time, peak area and determined metabolite concentration was excellent. Except for GTP (1.3%), the relative standard deviation (RSD) of the retention time was below 1% (Table 2). 29 out of 45 metabolites could be quantified with an experimental repeatability RSD < 3%. Being based on area ratio measurement, the RSDs for quantification were significantly smaller than the RSDs observed for the peak areas, respectively. The implemented normal flow set-up showed a superior performance compared to capillary IC showing typically RSDs between 5 and 8% for intensities and retention times.14 Finally, in this work excellent LODs and LOQs in the low nM range were obtained (Table S-2). Overall, the analytical figures of merit of isotope dilution based on IC-high resolution MS were comparable to an earlier study based on LC-MS/MS.33 Table S-3 compares LODs, LOQs, correlation coefficients and the linear dynamic range of calibration for 8 metabolites investigated by both MS platforms using internal standardization by uniformly

13C

labeled yeast. A recent IC-triple

quadrupole MS study reported LOQs between 0.25 µM and 50 µM regarding the quantitative analysis of organic acids17 proving once more the validity of high resolution MS as alternative to MS/MS quantification providing an internal standard is used. Typical metabolomics experiments involve quantification of different metabolites varying over 4 orders of magnitude in complex matrices requiring internal standardization.



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 10 of 18

Table 2. RSDs (N=5) of retention times, calculated concentrations and peak areas for all compounds present in the 1 µM metabolite mix spiked with uniformly 13C labeled yeast cell extract as internal standard and measured five times over 40 hours.

Retention time

Calculated concentration

Peak area

RSD # of [%] metabo.



≤ 0.1

21

≤1

11

≤ 5.0

12

≤ 0.2

12

≤3

18

≤ 7.5

23

≤ 0.4

8

≤5

11

≤ 10

9

≤ 1.0

3

≤ 7.5

5

≤ 12

1

≤ 1.3

1

Non-targeted data evaluation IC-high resolution MS in combination with isotope dilution proofed to be a suitable platform for absolute quantification. The excellent figures of merit such as retention time stability (repeatability of < 1% as shown before) and mass accuracy (< 1.5 ppm for the target metabolite panel) qualify the method for simultaneous non-targeted analysis. Given that, the next step was the thorough investigation of non-targeted fingerprinting in samples spiked with yeast based

13C

internal

standards. First, compound annotation was tested by measuring a calibration solution containing the 45 metabolite panel (see Table 1) at a concentration of 1 µM spiked with

13C

internal standards

(concentrations ranging from low nM to several µM) in the ddMS2 mode. On the basis of MS2 spectra, a library search in mzCloud (HighChem LLC, Slovakia) was performed (level 2 compound annotation after nomenclature proposed by the Chemical Analysis Working Group of the Metabolomics Standards Initiative).34 All compounds present in the standard mix and contained in the mzCloud library at the time of analysis (Jan 11, 2017) were successfully annotated, showing that the acquisition of MS2 spectra was triggered in the ddMS2 mode despite the presence of internal standard. However, mzCloud annotation was restricted by four aspects: (1) co-elution (for the coeluting pentose-5-phosphates four isomers were suggested), (2) ambiguous isomeric compounds despite chromatographic separation (for the hexose-monophosphates six possible isomers were proposed with matching scores > 85 but higher scores are conveniently correlated with the correct isomeric structure), (3) in-source fragmentation (seven metabolites were annotated twice due to insource fragmentation resulting in the loss of a phosphate unit) and (4) library coverage (10 10 ACS Paragon Plus Environment

Page 11 of 18

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


metabolites were not contained in mzCloud: FBP, isocitric acid, 6-phosphogluconic acid, CMP, cGMP, CDP, CTP, UTP, FBP, and 2PG). Summarizing, compound identification by comparison of MS2 spectra only (level 2), was not straightforward and demanded manual curation. Finally, only by chromatographic selectivity and measurement of standards level 1 identification can be achieved as level 1 metabolite identification requires at least two orthogonal data (e.g. retention time and mass spectrum) compared to a reference standard analyzed under the same conditions. Without reference standards, metabolites can only be putatively annotated (level 2) based on spectral comparison with libraries.34 The presence of

13C

labeled internal standards was not detrimental to the number of annotated

compounds in standards, despite the fact that some

13C

labeled components were present at µM

concentration. Next, compound annotation was considered in a proof of principle study addressing an in vitro cancer model. Comparative non-targeted analysis was performed using SW480 cancer cell preparations (1 x 106 cells seeded and extracted after 24 h incubation) with and without

13C

labeled internal standards. The added amount of yeast based 13C labeled standard was equivalent to the prior investigated multi-metabolite standard. As can be readily observed in Table 3, 41 compounds were identically annotated in samples with and without internal standard. The number of annotated compounds (mass accuracy 5 ppm, assignment threshold 70) that were not found either in the spiked cancer cell extract or in the pure extract or vice versa was very small. Table 3. Comparison of a cancer cell extract (preparation of 106 SW480 cells) with and without internal standard regarding putatively annotated compounds (level 2 annotation34). SW480 ISTD

SW480 no ISTD

# of mzCloud annotations

50

46

MS2 available in one sample type only

5

4

Features and MS2 available in one sample type only

4

1

Event

Based on MS/MS spectra acquired in the pooled sample, mzCloud search revealed a multitude of putatively annotated metabolites that were not included in the targeted analysis. These comprised e.g. sugar nucleotides (such as guanosine 5'-diphospho-ß-L-fucose, uridine 5'-diphosphogalactose, UDP-N-acetylglucosamine),

different

other

carboxylic

acids

(such

as

galacturonic acid,

pantothenic acid, 3,3-dimethylglutaric acid) and numerous modified amino acids (such as N-acetyl11 ACS Paragon Plus Environment


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 18

methionine, N-formyl-methionine, N-acetyl-1-aspartylglutamic acid). Independent of the presence of 13C internal standard, half of the metabolites that were putatively annotated (level 2) was affected by either in-source fragmentation or isomers (see Figure 1). These results confirm an earlier study.35 Thus, the importance of chromatographic selectivity and retention time as additional qualifier for peak annotation is evident.

Figure 1. Reliability of putatively annotated compounds of a cancer cell extracts (preparation of 106 SW480 cells) without internal standard found by mzCloud search. Comparison with standards revealed a number of identified false positives due to in-source fragmentation and due to isomers. Isomeric interference leads to features with more than one proposed structure in case the MS2 spectrum is not unique for one isomer.

Merging targeted metabolomics with non-targeted analysis The cancer cell line SW480 and its subline of acquired resistance was studied by the proposed workflow merging targeted absolute quantification with non-targeted fingerprinting. Five biological replicates of sensitive versus resistant SW480 cancer cell extracts (1 x 106 cells seeded and extracted after 24 h incubation) spiked with yeast based fully

13C

labeled internal standard upon

preparation were investigated. In order to prove the validity of the approach, additionally, samples prepared without internal standards were studied by non-targeted analysis. Figure S-2 gives a general schematic overview on measurements required for non-targeted data evaluation by the commercially available software Compound Discoverer 2.0. Feature detection and fold change (FC) calculations between the two sample groups were based on Full MS measurements. Data dependent acquisition of MS/MS spectra was performed on a pooled sample to allow spectral comparison with the library for compound annotation. The same Full MS data files of the samples with internal


Page 13 of 18

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


standard were also used for the targeted quantification based on isotope dilution. Thereby, the measurement time and required sample volume could be significantly decreased. Isotope dilution IC-MS revealed concentrations ranging over nearly 4 orders of magnitude, from nM to several µM (5 nM – 50 µM, R2> 0.99) for the investigated metabolite panel. Scaled to the protein content, typical concentrations for organic acids were between 1 and 10 nmol mg-1 protein (approx. 0.8 to 8 µM in the measurement solution) whereas the concentration of sugar phosphates was in general lower (20 – 200 pmol mg-1 protein, 10 to 130 nM measured). A wide concentration range was observed in the case of nucleotides with several deoxynucleotides < LOD, up to 20 nmol mg-1 protein (14 µM measured) found for ATP (Table S-4). 16 metabolites showed an RSD ≤ 10% (N=5) and further 10 metabolites RSDs below 20% (N=5). The obtained biological repeatability was in the expected range for replicates of adherent cancer cell cultures extracted consecutively following an established protocol.30 Differences in absolute concentrations obtained by targeted IC-MS between the two cancer cell models are shown in Figure 2 and detailed results on all investigated metabolites are given in Table S-4.

Figure 2. Different concentrations in resistant and sensitive SW480 cancer cells measured by targeted IC-high resolution MS and quantified with yeast based fully 13C labeled internal standard. Error bars denote standard deviations and statistical significance (t-test) is indicated as follows: * p < 0.05, ** p < 0.01.



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 14 of 18

Figure 3. Principal component analysis of sensitive versus resistant SW480 cancer cell extracts (A) without internal standard and (B) with fully 13C labeled yeast cell extract used as internal standard (ISTD) in the same amount in all samples analyzed by anion-exchange chromatography high resolution mass spectrometry. For both preparations a clear separation could be obtained.

In non-targeted data evaluation, a clear clustering between resistant and sensitive SW480 cancer cells in principal component analysis (PCA) could be obtained for the cell extractions with and without internal standard (Figure 3). The median of the RSDs of all peak areas resulting from the fully 13C labeled isotopologues selected for the targeted quantification approach in all samples was 14%. Hence, successful clustering of the cancer cell groups was possible as all samples contain the same amount of internal standard and the internal standard is equally affected by sample preparation. Setting the criteria of peak area RSD < 15% in at least one of the sample groups, 459 and 574 aligned features (consisting of different isotopologues and adducts) were found in the samples without and with internal standard, respectively. 153 and 129 showed differences with an adjusted p-value of < 0.05 (Benjamini-Hochberg), respectively. These differences are displayed as heat maps in Figure S-3. Out of these differential compounds, several could be putatively annotated in both experimental settings, such as e.g. acetylated amino acids and other organic acids. From the putatively annotated compounds, which were not already included in the targeted analysis, Nacetyl-methionine and N-acetyl-glutamic acid showed the most significant decrease in resistant cells with fold changes (FC, ratio resistant/sensitive) of 0.3. On the other hand, galacturonic acid showed 14 ACS Paragon Plus Environment

Page 15 of 18

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


a significant upregulation in resistant cells (FC 2). For pentose-phosphates (m/z 229.0119) more than one peak was putatively annotated indicating that ribose-1-phosphate is separated from the pentose-5-phosphates which were analyzed in the standard mix. All of those peaks are decreased in the resistant cells (FC 0.4 – 0.6). Using our novel IC-high resolution MS workflow for merging targeted and non-targeted metabolomics distinct metabolic phenotypes of drug-sensitive and resistant SW480 cells were observed.

CONCLUSIONS In this work, we introduced a novel IC-MS workflow that allows merging targeted quantification based on isotope dilution with non-targeted data analysis. By using the same data files for both approaches, the measurement time and sample amounts could be considerably decreased supporting high-throughput analysis. Fully 13C labeled extracts of the yeast Pichia pastoris proved to be ideally suitable for compound specific internal standardization of human derived samples extending the linear dynamic range to 4 orders of magnitude. The presented strategy based on anion chromatography and high resolution MS facilitating internal standardization is a promising approach to unravel changes in the central carbon metabolism and nucleotides in different biological systems.

ASSOCIATED CONTENT Supporting information Additional information as noted in the text. The Supporting Information is available free of charge on the ACS Publications website (http://pubs.acs.org).

AUTHOR INFORMATION Corresponding Author * E-mail: [email protected]

Author Contributions The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. 15 ACS Paragon Plus Environment


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 16 of 18

Notes The authors declare no competing financial interest.

ACKNOWLEDGEMENTS We want to thank Petra Volejnik and Yasin El Abiead for continuous support and fruitful discussions. Thermo Fisher Scientific is acknowledged for providing the Dionex Integrion HPIC System and the Mass Spectrometry Center (MSC), Faculty of Chemistry, University of Vienna for providing mass spectrometric instrumentation. Lastly, Fellinger Krebsforschung is gratefully acknowledged for financial support.

REFERENCES (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18)

Rochat, B. TrAC Trends Anal. Chem. 2016, 84, Part B, 151–164. Cajka, T.; Fiehn, O. Anal. Chem. 2016, 88 (1), 524–545. León, Z.; García-Cañaveras, J. C.; Donato, M. T.; Lahoz, A. ELECTROPHORESIS 2013, 34 (19), 2762–2775. Vuckovic, D. Anal. Bioanal. Chem. 2012, 403 (6), 1523–1548. Ritter, J. B.; Genzel, Y.; Reichl, U. Anal. Biochem. 2008, 373 (2), 349–369. Armitage, E. G.; Barbas, C. J. Pharm. Biomed. Anal. 2014, 87, 1–11. Ivanisevic, J.; Zhu, Z.-J.; Plate, L.; Tautenhahn, R.; Chen, S.; O’Brien, P. J.; Johnson, C. H.; Marletta, M. A.; Patti, G. J.; Siuzdak, G. Anal. Chem. 2013, 85 (14), 6876–6884. Dunn, W. B.; Broadhurst, D.; Begley, P.; Zelena, E.; Francis-McIntyre, S.; Anderson, N.; Brown, M.; Knowles, J. D.; Halsall, A.; Haselden, J. N.; Nicholls, A. W.; Wilson, I. D.; Kell, D. B.; Goodacre, R.; Consortium, T. H. S. M. (HUSERMET). Nat. Protoc. 2011, 6 (7), 1060–1083. Siegel, D.; Meinema, A. C.; Permentier, H.; Hopfgartner, G.; Bischoff, R. Anal. Chem. 2014, 86 (10), 5089–5100. van Dam, J. C.; Eman, M. R.; Frank, J.; Lange, H. C.; van Dedem, G. W. K.; Heijnen, S. J. Anal. Chim. Acta 2002, 460 (2), 209–218. Kvitvang, H. F. N.; Kristiansen, K. A.; Bruheim, P. J. Chromatogr. A 2014, 1370, 70–79. Kiefer, P.; Nicolas, C.; Letisse, F.; Portais, J.-C. Anal. Biochem. 2007, 360 (2), 182–188. Alonso, A. P.; Piasecki, R. J.; Wang, Y.; LaClair, R. W.; Shachar-Hill, Y. Plant Physiol. 2010, 153 (3), 915–924. Wang, J.; Christison, T. T.; Misuno, K.; Lopez, L.; Huhmer, A. F.; Huang, Y.; Hu, S. Anal. Chem. 2014, 86 (10), 5116–5124. Hu, S.; Wang, J.; Ji, E. H.; Christison, T.; Lopez, L.; Huang, Y. Anal. Chem. 2015, 87 (12), 6371– 6379. Burgess, K.; Creek, D.; Dewsbury, P.; Cook, K.; Barrett, M. P. Rapid Commun. Mass Spectrom. 2011, 25 (22), 3447–3452. Petucci, C.; Zelenin, A.; Culver, J. A.; Gabriel, M.; Kirkbride, K.; Christison, T. T.; Gardell, S. J. Anal. Chem. 2016. Cipollina, C.; ten Pierick, A.; Canelas, A. B.; Seifar, R. M.; van Maris, A. J. A.; van Dam, J. C.; Heijnen, J. J. J. Chromatogr. B 2009, 877 (27), 3231–3236. 16 ACS Paragon Plus Environment

Page 17 of 18

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


(19) Coulier, L.; Bas, R.; Jespersen, S.; Verheij, E.; van der Werf, M. J.; Hankemeier, T. Anal. Chem. 2006, 78 (18), 6573–6582. (20) Lu, W.; Clasquin, M. F.; Melamud, E.; Amador-Noguez, D.; Caudy, A. A.; Rabinowitz, J. D. Anal. Chem. 2010, 82 (8), 3212–3221. (21) Wu, J.; Zhang, Y.; Wiegand, R.; Wang, J.; Bepler, G.; Li, J. J. Chromatogr. B 2015, 1006, 167–178. (22) Luo, B.; Groenke, K.; Takors, R.; Wandrey, C.; Oldiges, M. J. Chromatogr. A 2007, 1147 (2), 153–164. (23) Büscher, J. M.; Czernik, D.; Ewald, J. C.; Sauer, U.; Zamboni, N. Anal. Chem. 2009, 81 (6), 2135– 2143. (24) Kiefer, P.; Delmotte, N.; Vorholt, J. A. Anal. Chem. 2011, 83 (3), 850–855. (25) Yanes, O.; Tautenhahn, R.; Patti, G. J.; Siuzdak, G. Anal. Chem. 2011, 83 (6), 2152–2161. (26) Neubauer, S.; Haberhauer-Troyer, C.; Klavins, K.; Russmayer, H.; Steiger, M. G.; Gasser, B.; Sauer, M.; Mattanovich, D.; Hann, S.; Koellensperger, G. J. Sep. Sci. 2012, 35 (22), 3091–3105. (27) Wahl, S. A.; Seifar, R.; ten Pierick, A.; Ras, C.; van Dam, J.; Heijnen, J.; van Gulik, W. In Metabolic Flux Analysis; Krömer, J. O., Nielsen, L. K., Blank, L. M., Eds.; Methods in Molecular Biology; Springer New York, 2014; pp 91–105. (28) Bennett, B. D.; Yuan, J.; Kimball, E. H.; Rabinowitz, J. D. Nat. Protoc. 2008, 3 (8), 1299–1311. (29) Miklos, W.; Pelivan, K.; Kowol, C. R.; Pirker, C.; Dornetshuber-Fleiss, R.; Spitzwieser, M.; Englinger, B.; van Schoonhoven, S.; Cichna-Markl, M.; Koellensperger, G.; Keppler, B. K.; Berger, W.; Heffeter, P. Cancer Lett. 2015, 361 (1), 112–120. (30) Dettmer, K.; Nürnberger, N.; Kaspar, H.; Gruber, M. A.; Almstetter, M. F.; Oefner, P. J. Anal. Bioanal. Chem. 2011, 399 (3), 1127–1139. (31) Klavins, K.; Neubauer, S.; Chalabi, A. A.; Sonntag, D.; Haberhauer-Troyer, C.; Russmayer, H.; Sauer, M.; Mattanovich, D.; Hann, S.; Koellensperger, G. Anal. Bioanal. Chem. 2013, 405 (15), 5159–5169. (32) Ortmayr, K.; Hann, S.; Koellensperger, G. Analyst 2015, 3465–3473. (33) Klavins, K.; Drexler, H.; Hann, S.; Koellensperger, G. Anal. Chem. 2014, 86 (9), 4145–4150. (34) Sumner, L. W.; Amberg, A.; Barrett, D.; Beale, M. H.; Beger, R.; Daykin, C. A.; Fan, T. W.-M.; Fiehn, O.; Goodacre, R.; Griffin, J. L.; Hankemeier, T.; Hardy, N.; Harnly, J.; Higashi, R.; Kopka, J.; Lane, A. N.; Lindon, J. C.; Marriott, P.; Nicholls, A. W.; Reily, M. D.; Thaden, J. J.; Viant, M. R. Metabolomics Off. J. Metabolomic Soc. 2007, 3 (3), 211–221. (35) Xu, Y.-F.; Lu, W.; Rabinowitz, J. D. Anal. Chem. 2015, 87 (4), 2273–2281.



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

FOR TOC ONLY


Page 18 of 18

Anion-Exchange Chromatography Coupled to High-Resolution

Recommend Documents