NMR-Guided Mass Spectrometry for Absolute ... - ACS Publications

Subscriber access provided by UNIVERSITY OF ADELAIDE LIBRARIES

Article

NMR-Guided Mass Spectrometry for Absolute Quantitation of Human Blood Metabolites G. A. Nagana Gowda, Danijel Djukovic, Lisa Fan Bettcher, Haiwei Gu, and Daniel Raftery Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.7b04089 • Publication Date (Web): 02 Jan 2018 Downloaded from http://pubs.acs.org on January 3, 2018

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 24 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

NMR-Guided Mass Spectrometry for Absolute Quantitation of Human Blood Metabolites

G. A. Nagana Gowda,1* Danijel Djukovic,1 Lisa Fan Bettcher,1 Haiwei Gu,1 and Daniel Raftery1,2,3*

1

Northwest Metabolomics Research Center, Anesthesiology and Pain Medicine, and 2

Department of Chemistry, University of Washington, Seattle, WA 98109, 3

Fred Hutchinson Cancer Research Center, Seattle, WA 98109

*Address correspondence to G. A. Nagana Gowda ([email protected]); ORCID: 0000-0002-0544-7464 Daniel Raftery ([email protected]); ORCID: 0000-0003-2467-8118

1 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

ABSTRACT Broad-based, targeted metabolite profiling using mass spectrometry (MS) has become a major platform used in the field of metabolomics for a variety of applications. However, quantitative MS analysis is challenging owing to numerous factors including: 1) the need for, ideally, isotope labeled internal standards for each metabolite, 2) the fact that such standards may be unavailable or prohibitively costly, 3) the need to maintain their concentrations close to the target metabolites, or 4) the alternative use of time consuming calibration curves for each target metabolite. Here, we introduce a new method in which metabolites from a single serum specimen are quantified based on a recently developed NMR method [Anal. Chem. 2015, 87, 706-15] and then used as reference for absolute metabolite quantitation using MS. The MS concentrations of 30 metabolites thus derived for test serum samples exhibited excellent correlation with NMR (R2 > 0.99) with a median CV of 3.2%. This NMR guided MS quantitation approach is simple, easy to implement and offers new avenues for routine quantification of blood metabolites using MS. The demonstration that NMR and MS data can be compared and correlated when using identical sample preparations allows improved opportunities to exploit their combined strengths for biomarker discovery and unknown metabolite identification. Intriguingly, however, metabolites including glutamine, pyroglutamic acid, glucose and sarcosine correlated poorly with NMR due to stability issues in their MS analysis or weak/overlapping signals. Such information is potentially important for improving biomarker discovery and biological interpretations. Further, the new quantitation method demonstrated here for human blood serum can in principle be extended to a variety of biological mixtures. Keywords: Quantitation, NMR spectroscopy, mass spectrometry, blood serum/plasma,


Page 2 of 24

Page 3 of 24 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


metabolomics, protein precipitation INTRODUCTION A large number of investigations in metabolomics is focused on profiling human blood serum/plasma, owing to its relevance in diagnosing, managing and understanding virtually all human diseases.1-3 Mass spectrometry (MS) and nuclear magnetic resonance (NMR) spectroscopy are the two major analytical platforms used in the field for the analysis of blood serum/plasma. They are complementary methods; while MS is highly sensitive, NMR is highly quantitative. However, while NMR enables the absolute quantitation of metabolites in a biological mixture using a single internal standard, MS lacks such capability, which is due to a number of factors, including ionization efficiencies, ion suppression and/or matrix effects. Common methods for absolute quantitation using MS include the use of internal isotope labeled standards or structural analogues, external calibration using standard curves, and standard addition. However, numerous factors including the often high cost or unavailability of internal standards, the need to maintain their concentrations close to target metabolites, or alternatively, the requirement of calibration curves for each target metabolite pose a major challenge, especially, for routine metabolomics applications. There have been numerous developments to exploit in vivo isotope labeled metabolites using bacteria, algae or fungi for quantitation in metabolomics and lipidomics.4-10 However, the use of uniformly labeled compound mixtures is in a developmental phase, and their utility for blood metabolite analysis has not been practical so far. The limited ability to quantitate a large number of metabolites, routinely, has often restricted MS-based metabolomics studies to the comparison of relative peak areas and rendered the vast body of published MS data difficult to compare across multiple studies. Comparison and



correlation of metabolite data generated using MS and NMR to exploit their combined strength in the metabolomics field has also been limited as a result. For many years, a major factor that affected comparison and correlation of NMR and MS data was the fact that NMR-based metabolomics analysis traditionally involved the use of intact blood serum/plasma, while MS analysis involved prior removal of the abundant serum/plasma proteins. The major drawbacks of intact serum analysis using NMR are 1) the peak intensities of many low to moderate concentration metabolites are hard to distinguish from the large and variable protein/lipid signal background, and 2) the concentrations of many metabolites are grossly underestimated due to attenuation caused by binding to serum/plasma proteins,11-15 which makes comparison with MS data very challenging. More recently, we developed an optimized protein removal method, resulting in the quantitation of a large number of metabolites including amino acids, organic acids, carbohydrates and heterocyclic compounds.14,15 The use of an identical sample preparation method for both NMR and MS provided a common platform for analysis of blood metabolites such that data can be compared directly. This is important since the performance of sample processing for MS analysis is typically evaluated based on the total number of ions detected and not on the comprehensive quantitative analysis of metabolites.16-23 In the present study, we describe a new NMR-guided method for absolute quantitation of blood metabolites using MS. Here, NMR derived concentrations for a single serum sample were used as reference values for quantitation of metabolites in the rest of the samples using MS. Excellent correlations were observed between NMR guided MS concentrations and those derived from NMR, and the results were duplicated on a second targeted LC-MS platform. The NMRguided MS quantitation approach is simple, easy to implement and, importantly, by obviating the need for internal standards and tedious procedures, offers new avenues for MS-based


Page 4 of 24

Page 5 of 24 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


quantitative blood metabolomics. A few metabolites, however, correlated poorly with NMR and, interestingly, such results challenge the implicit assumption that metabolites are largely stable during MS analysis. This approach therefore opens new avenues for identification of potentially unstable metabolites during MS analysis, the knowledge of which may be critical for biomarker discovery and biological interpretations.

MATERIALS AND METHODS Methanol, sodium phosphate, monobasic (NaH2PO4), sodium phosphate, dibasic (Na2HPO4), and 3-(trimethylsilyl)propionic acid-2,2,3,3-d4 sodium salt (TSP) were obtained from Sigma-Aldrich (St. Louis, MO). Deuterium oxide (D2O) and a mixture of uniformly13

C,15N-labeled (97-99% enrichment) standard amino acids that included leucine, methionine,

tryptophan and tyrosine were obtained from Cambridge Isotope laboratories, Inc. (Andover, MA). Human serum samples from 8 healthy subjects (4 Male; 4 Female; Table S1) were obtained from Innovative Research, Inc (Novi, MI). A commercial pooled human serum was also obtained from Innovative Research, Inc. Deionized (DI) water was purified using an in-house Synergy Ultrapure Water System from Millipore (Billerica, MA). All chemicals were used with no further purification. Preparation of phosphate buffer: A buffer solution (0.1 M, pH=7.4) was prepared by dissolving 249.9 mg anhydrous NaH2PO4 and 1124.0 mg anhydrous Na2HPO4 in 100 g D2O and used without further pH correction. Serum protein precipitation: Frozen serum samples were thawed at room temperature (25 °C), homogenized using a vortex mixer, and then 550 µL was pipetted into a 2 mL Eppendorf vial (Fisher Scientific). For protein removal, serum samples were mixed with methanol in a 1:2 ratio



Page 6 of 24

(v/v), which was recently shown by NMR to extract metabolites optimally.14,15 The resulting mixtures were vortexed for 30 s, incubated at -20 °C for 20 min and centrifuged at 13,400 rcf for 30 min to pellet proteins. The supernatants were filtered using 0.45 µm PVDF filters (Phenomenex, Torrance, CA) and the filtrate for each sample was divided into two parts. One part (1050 µL) was dried using an Eppendorf Vacufuge-Plus vacuum concentrator; the resulting residue was mixed with 300 µL phosphate buffer (0.1M) in D2O containing 50 µM TSP. The solution volume was increased to 600 µL using phosphate buffer in D2O (0.1M) with no TSP and transferred to 5 mm NMR tube for analysis; the other part (150 µL) was diluted to 500 µL using a mixture of deionized water and methanol (1:2 v/v) and used for targeted LC−MS/MS analysis.

NMR Spectroscopy: NMR experiments were performed at 25 oC on a Bruker Avance III 800 MHz spectrometer equipped with a cryogenically cooled probe and Z-gradients suitable for inverse detection as described previously.15 Briefly, the CPMG (Carr-Purcell-Meiboom-Gill) pulse sequence with water suppression using presaturation was used for

1

H 1D NMR

experiments. To enable absolute quantitation of metabolites using the internal reference TSP, the CPMG experiments were performed with 128 transients and a sufficiently long recycle delay (D1=15 s). Chemical shifts were referenced to the internal TSP signal. The raw data was Fourier transformed after zero filling once to a total spectrum size of 32K points. Bruker Topspin versions 3.0 or 3.1 software packages were used for NMR data acquisition, processing and analyses. NMR Peak assignment and metabolite quantitation: Assignment of NMR peaks to specific metabolites was based on the recent study on blood metabolite quantitation using NMR, where the characteristic peaks for metabolites are annotated to enable their easy identification and


Page 7 of 24 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


routine analysis.15 Chenomx NMR Suite Professional Software package (version 5.1; Chenomx Inc., Edmonton, Alberta, Canada) was used to quantitate metabolites. This software allows fitting spectral lines using the standard metabolite library for 800 MHz 1H NMR spectra and the determination of concentrations. Peak fitting with reference to the internal TSP signal enabled the determination of absolute concentrations for metabolites.15 Mass Spectrometry: Targeted LC-MS Analysis: Blood serum metabolites were analyzed by mass spectrometry using an AB Sciex QTrap 5500 or AB Sciex QTrap 6500+ mass spectrometers (AB Sciex, Toronto, ON, Canada). The AB Sciex QTrap 5500 spectrometer was coupled to an LC system composed of two Agilent 1260 binary pumps, an Agilent 1260 autosampler, and Agilent 1290 column compartment containing a column-switching valve (Agilent Technologies, Santa Clara, CA). For chromatography using the Agilent LC system, the mobile phase was composed of Solvent A: 5 mM ammonium acetate in 90% H2O/10% acetonitrile with 0.2% acetic acid, and Solvent B: 5 mM ammonium acetate in 90% acetonitrile/10% H2O with 0.2% acetic acid. The AB Sciex QTrap 6500+ mass spectrometer was coupled to an LC system composed of four Shimadzu Nexera LC-20 pumps, an AB Sciex autosampler and AB Sciex column compartment containing a column-switching valve (AB Sciex, Toronto, ON, Canada). For chromatography using the Shimadzu LC system, the mobile phase was composed of Solvent A: 10 mM ammonium acetate in 95% H2O/3% acetonitrile/2% methanol/0.2% acetic acid, and Solvent B: 10 mM ammonium acetate in 93% acetonitrile/2% methanol/5% H2O /0.2% acetic acid. The eluted metabolites were ionized using a Sciex Turbo electrospray ionization (ESI) source and targeted data acquisition was performed in multiplereaction-monitoring (MRM) mode using optimized parameters as described earlier.24 Metabolites were analyzed in positive or negative ionization mode by injecting each sample twice, 2 µL (for



Page 8 of 24

Qtrap 5500) or 5 µL (for Qtrap 6500+) for analysis using positive ionization mode and 10 µL for analysis using negative ionization mode. Samples for Qtrap 6500+ were diluted by a factor of 2.3 relative to those used for Qtrap 5500. Two hydrophilic interaction chromatography (HILIC) columns (SeQuant ZIC-cHILIC; 150 × 2.1 mm, 3.0 µm particle size, Merck KGaA, Darmstadt, Germany for Agilent LC; Waters XBridge Amide; 150 × 2.1 mm, 2.5 µm particle size for Shimadzu LC), connected in parallel were run under identical conditions for positive and negative ionization modes. MRM peaks for metabolites were integrated using MultiQuant 2.1 or 3.02 softwares (AB Sciex). To ensure the instrument’s stability during the analysis, a quality control sample prepared using a pooled serum was analyzed under identical conditions, before and after the analysis of the serum samples. Separately, MS analysis was also performed using the AB Sciex QTrap 5500 mass spectrometer after spiking the samples with

13

C,

15

N-isotope

labeled amino acid mixture (10µL; 30.55 mg/50mL in water : methanol (1:1 v/v)).

Absolute metabolite quantitation using targeted LC-MS/MS guided by NMR: One of the serum samples (Table S1) was randomly designated as the reference sample. Absolute concentration of each metabolite in the reference serum obtained by NMR was set as the MS concentration for the corresponding metabolite’s MRM peak. Absolute concentrations for metabolites in the remaining (test) serum samples were then derived by multiplying MRM peak area of a metabolite by the ratio of absolute concentration to MRM peak area of the same metabolite in the reference serum as shown in equation (1). () =

( )

………………….. (1)

Where C(MS)mn is the absolute concentration of metabolite m in sample n derived from NMR guided mass spectrometry; MRMmn is the MS peak area of metabolite m in sample n; C(NMR)mr is 8 ACS Paragon Plus Environment

Page 9 of 24 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


the absolute concentration of metabolite m derived from NMR in reference sample r; and MRMmr is the MS peak area of metabolite m in reference sample r. Further, to comprehensively evaluate thus derived concentrations based on a single reference sample, concentrations for each metabolite were recalculated by switching the reference serum to a different sample. This procedure was repeated until each of the 8 serum samples (Table S1) was used as the reference once. Separately, metabolite concentrations for the 8 serum samples were obtained using a pooled serum from all 8 samples or a commercial pooled serum sample as reference. Absolute concentrations thus derived using MS were correlated with NMR concentrations for the same samples to evaluate the performance of NMR-guided MS for absolute quantitation. Errors in concentration across the sample set, computed using the MS/NMR data, were used to evaluate relative matrix effects across samples.

RESULTS AND DISCUSSION Protein precipitation using methanol provided well resolved 1H NMR spectra for all serum samples. Mass spectrometry analysis in MRM mode combining positive and negative ionization mode targeted 198 metabolites. Thirty-eight metabolites were detected by both MS and NMR, and were selected as the initial set of molecules for quantitation (Table S2). However, NMR/MS peaks for eight metabolites were either overlapped in the NMR spectrum, co-eluted off the LC column but had the same MRM transition in MS, or displayed poor chromatographic peak shapes, which made them less reliable for evaluation of this quantitation approach and hence were omitted from further analysis. Concentrations for the remaining 30 metabolites were obtained by NMR using a single internal reference, TSP, and the obtained values are listed in Table S3 for all serum samples. Figure 1 shows typical NMR spectrum of a serum sample with



expanded regions and peak annotations for quantitated metabolites and Figure S1 shows typical MS/MS ion chromatograms for the metabolites. Metabolite concentrations corresponding to MRM peak areas for one of the serum samples (initially, Sample 1) were set equal to those obtained by NMR (i.e., the NMR reference concentration). Using these values, absolute concentrations corresponding to each metabolite in the other seven samples based on MRM peak areas were obtained (Table S4). A comparison of the determined concentrations for the same metabolite in the same serum sample using different NMR reference concentrations indicated that most of the metabolite concentrations agreed well irrespective of the reference sample used. Metabolite concentrations thus derived by MS showed a median CV of 4.0, 3.4, 3.5, 4.0, 3.2, 5.8, 3.2 and 5.7 % for 25 metabolites when serum samples 1-8 was used as the reference, respectively. As an example, Figure 2 shows a comparison of metabolite concentrations derived either by NMR or NMR guided MS, for a typical serum sample. The data, shown here indicate the excellent agreement between the two methods of quantitation. Almost identical data were obtained even when the pooled serum from all 8 samples or commercial pooled serum was used as the reference sample (Figure S2). Correlations of metabolite concentrations for the 8 serum samples are shown in Figure 3 and Figure S3, using Sample 5 as the reference; the data show excellent correlation between NMR and NMR-guided MS concentrations (R2>0.99) for each of the other 7 serum samples. In the figures, Sample 5 was used as the reference is also shown for comparison (R2 = 1). Excellent correlations were obtained irrespective of the serum sample was used as the reference. In addition, a comparison of concentrations for amino acids, leucine, methionine, tryptophan and tyrosine, obtained by NMRguided approach and isotope labeled internal standards showed a good agreement with a median CV of 6.2 % between the two approaches (Figure 4).


Page 10 of 24

Page 11 of 24 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


To further evaluate this quantitation approach, concentrations for the same metabolites in all the serum samples derived by NMR and NMR-guided MS were compared, individually. As shown in Figure 5 and Figure S4, most of the metabolites exhibited very good correlation between NMR and MS (R2 > 0.9). However, metabolites including glutamine, pyroglutamic acid, glucose and sarcosine showed very poor correlations (Figure 6). For glutamine and pyroglutamic acid the observed correlations are in accordance with the known phenomena of glutamine cyclization to pyroglutamic acid.15,25,26 Cyclization occurs in the electrospray ionization (ESI) source during LC-MS analysis; the magnitude of cyclization in the ESI source can range from 33 to 100%, and depends on fragmentor voltage.26 In view of such effects by the ESI source, as well as those caused by sample preparation,15, 25 it is important to be careful with the general assumption that the vast majority of metabolites are stable when evaluated in metabolomics. These findings also open new avenues for identification of potentially unstable metabolites during MS analysis, the knowledge of which is potentially important for improved biomarker discovery and biological interpretations. The poor correlation of glucose may be because it exhibits multiple MRM peaks (Figure S1) and sugars such as mannose with identical molecular weight and similar retention time overlap. On the other hand, the poor correlation of sarcosine may arise from the measurement errors due to its weak MRM peak (Figure S1). In order to assess the relative magnitude of the matrix effect on MS ionization across samples, concentrations of metabolites from the same sample determined, separately, using different samples as NMR reference were evaluated. As shown in Figure S5, the average CV for more than 50% of the metabolites was 5.6%, while the average CV for all metabolites except three was 7.2%. Three metabolites including glutamine, pyroglutamic acid and sarcosine, however, each exhibited a CV of more than 15%, which is in accordance with the poor



correlations shown between NMR and NMR-guided MS concentrations (Figure 6). As an additional evaluation for the matrix effect, the concentrations for metabolites in the 8 serum samples obtained using the pooled serum from all 8 samples or the commercial pooled serum as the reference showed almost identical data to using one of the 8 eight serum samples as the reference (Figure S2). These results indicate that the sample matrix effect on quantitation is tolerable. Absolute quantitation of blood metabolites using MS on a routine basis is impeded by numerous factors and challenges, as described in the Introduction. In vivo isotope labeled metabolites especially using yeast (Pichia pastoris) is promising to be a reliable tool for quantitation in metabolomics and lipidomics.7-9 More than 99% labeling efficiency has been achieved for hundreds of already identified metabolites using this approach, the extracts show a high dynamic range for the use in quantitative experiments and addition of this extract to samples has demonstrated not to alter untargeted metabolomics analyses. This approach, however, is still in a developmental phase and its utility for blood metabolite analysis has not been demonstrated. NMR spectroscopy draws its strength from being highly reproducible and quantitative; a single internal reference can in principle be used to determine absolute concentrations for all NMR measurable metabolites in a biological mixture. In this study, we have exploited this strength of NMR to guide absolute quantitation of metabolites using MS, in biological samples with similar matrices. In particular, the study exploits recent advances in blood metabolite quantitation using NMR, which provides an optimized and common protocol for analysis of the same samples using both NMR and MS.15 Using this approach and identical sample processing for both NMR and MS methods, linear correlation between NMR and MS data could be achieved


Page 12 of 24

Page 13 of 24 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 metabolites in all the serum samples (Figures 2, 3 and 5). Due to this linear relationship, NMR derived metabolite concentrations for a single serum is suffice to obtain absolute concentrations for the same metabolites in other serum samples using MS. Ultimately, the use of internal standards for MS quantitation is preferred, however there appear to be a number of additional metabolites that can be quantified using this simple and cost effective alternative approach. In this proof-of-concept study, we used human serum to demonstrate NMR-guided metabolite quantitation in MS because absolute quantitation of human blood metabolites is of great interest in biomedicine. It may be interesting to note that blood serum or plasma is particularly suited for the NMR guided quantitation since variation among samples is very small (typically less than a factor of 2). Application of this approach to biological systems that exhibit greater variation among samples than serum or plasma will likely be more challenging. In particular, factors that contribute to ion suppression as well as MS linearity can become key issues to address when using the NMR guided approach for measurement of such biological systems.

In conclusion, we present a new NMR-guided MS method for absolute metabolite quantitation in human blood serum using mass spectrometry, which takes advantage of recent advances in NMR-based blood metabolite quantitation. The method employs quantitative data for a single serum specimen derived from NMR for MS-based absolute quantitation of the same metabolites in other serum samples. It uses an optimized protein removal method common to both MS and NMR to ensure that the detected metabolites retain a linear relationship between NMR and MS. Absolute quantitation of blood metabolites provides the capability to compare data across metabolomics studies and to clinical data obtained using other platforms. The only requirement is to use the same NMR calibrated reference sample, periodically, to correct for changes in the MS peak intensities. From the data presented, relative sample matrix effects 13 ACS Paragon Plus Environment


across samples appear to be limited, at least in this small study. Validation of this observation needs to be performed, and ultimately the use of internal standards when available and cost effective will normally provide the most accurate results. However, as a potential replacement for time consuming external calibration using multiple standards, this approach may be appealing. Additionally, it is both interesting and a cause for concern that some metabolites correlate very poorly between NMR and MS. Because such correlations were not anticipated and challenge the widely accepted implicit assumption that the vast majority of metabolites are generally stable during MS analysis, this approach provides a mechanism to identify potentially unstable metabolites. This knowledge is highly useful for biomarker discovery and biological interpretations. In sum the described NMR-guided quantitative method is simple, easy to implement and, importantly, can reduce or obviate the need for internal standards and tedious procedures to thereby offer new avenues for MS based quantitative blood metabolomics on a routine basis. These findings provide an opportunity for quantitative analysis of an enhanced pool of metabolites in blood and ultimately a variety of other biological mixtures. While this work constitutes a proof of concept study with a relatively small number of quantified metabolites, the use of sensitivity enhancement methods including smaller sample volume probes, enhanced sample concentration and/or signal averaging, can increase the limit of quantitation for NMR and expand significantly the NMR-guided MS quantifiable metabolite pool. It should be noted that for both NMR and MS, a weak or an overlapping peak, or a peak with a poor line shape can deleteriously affect peak integration and thereby affect reliable quantitation. Hence, high quality spectra are critical for accurate quantitation. In addition, the CPMG NMR experiment used in this study, while provides a much cleaner baseline compared to the one-pulse or 1D NOESY


Page 14 of 24

Page 15 of 24 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


experiments, can underestimate serum metabolite concentrations albeit marginally due to relaxation effects. A comparison of metabolite concentrations in the two experiments showed an underestimation in CPMG experiment by an average of 4.6%.14 Such relaxation effects, therefore,

need to be accounted for to improve the quantitation accuracy using this experiment.

Acknowledgments We acknowledge financial support from the NIH (National Institute of General Medical Sciences 2R01GM085291). Notes The authors declare the following financial interest: Daniel Raftery reports holding equity and an executive position at Matrix Bio, Inc.

Supporting Information: Tables S1 to S4, and Figures S1 to S5.



References 1. Beckonert, O.; Keun, H. C.; Ebbels, T. M.; Bundy, J.; Holmes, E.; Lindon, J. C.; Nicholson, J. K. Nat. Protoc. 2007, 2(11), 2692-703. 2. Psychogios, N.; Hau, D. D.; Peng, J.; Guo, A. C.; Mandal, R.; Bouatra, S.; Sinelnikov, I.; Krishnamurthy, R.; Eisner, R.; Gautam, B.; Young, N.; Xia, J.; Knox, C.; Dong, E.; Huang, P.; Hollander, Z.; Pedersen, T. L.; Smith, S. R.; Bamforth, F.; Greiner, R.; McManus, B.; Newman, J. W.; Goodfriend, T.; Wishart, D. S. PLoS One 2011, 6 (2), e16957. 3. Nagana Gowda, G. A.; Raftery, D. Curr. Metabolomics 2013, 1 (3), 227-240. 4. Bennett, B. D.; Yuan, J.; Kimball, E. H.; Rabinowitz, J. D. Nat. Protoc. 2008, 3, 1299-311. 5. Vielhauer, O.; Zakhartsev, M.; Horn, T.; Takors, R.; Reuss, M. J. Chromatogr B Analyt. Technol. Biomed Life Sci. 2011, 879(32), 3859-70. 6. Mairinger T.; Hann, S. Anal. Bioanal. Chem. 2017, 409(15), 3713-3718. 7. 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), 3091-105. 8. Rampler, E.; Coman, C.; Hermann, G.; Sickmann, A.; Ahrends, R.; Koellensperger, G. Analyst. 2017, 142(11), 1891-1899. 9. Schwaiger, M.; Rampler, E.; Hermann, G.; Miklos, W.; Berger, W.; Koellensperger, G. Anal. Chem. 2017, 89(14), 7667-7674. 10. Weiner, M.; Tröndle, J.; Schmideder, A.; Albermann, C.; Binder, K.; Sprenger, G.A.; WeusterBotz, D. Anal. Biochem. 2015, 478, 134-40. 11. Nicholson, J. K.; Gartland, K. P. NMR Biomed. 1989, 2 (2), 77-82. 12. Chatham, J. C.; Forder, J. R. Biochim. Biophys. Acta 1999, 1426 (1), 177-184. 13. Bell, J. D.; Brown, J. C.; Kubal, G.; Sadler, P. J. FEBS Lett. 1988, 235, 81-86. 14. Nagana Gowda, G. A.; Raftery, D. Anal. Chem. 2014, 86(11), 5433-40. 15. Nagana Gowda, G. A.; Gowda, Y.; Raftery, D. Anal. Chem. 2015, 87(1), 706-15. 16. Vuckovic, D. Anal. Bioanal. Chem. 2012, 403 (6), 1523-1548. 17. 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.; Human Serum Metabolome (HUSERMET) Consortium. Nat. Protoc. 2011, 6(7), 1060-83. 16 ACS Paragon Plus Environment

Page 16 of 24

Page 17 of 24 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


18. 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. 19. Bruce, S. J.; Tavazzi, I.; Parisod, V.; Rezzi, S.; Kochhar, S.; Guy, P. A. Anal. Chem. 2009, 81 (9), 3285-3296. 20. Jiye, A.;, Trygg, J.; Gullberg, J.; Johansson, A. I.; Jonsson, P.; Antti, H.; Marklund, S. L.; Moritz, T. Anal. Chem. 2005, 77 (24), 8086-8094. 21. Zelena, E.; Dunn, W. B.; Broadhurst, D.; Francis-McIntyre, S.; Carroll, K. M.; Begley, P.; O'Hagan, S.; Knowles, J. D.; Halsall, A. Anal. Chem. 2009, 81 (4), 1357-1364. 22. Tulipani, S.; Llorach, R.; Urpi-Sarda, M.; Andres-Lacueva, C. Anal. Chem. 2013, 85 (1), 341348. Want, E. J.; O'Maille, G.; Smith, C. A.; Brandon, T. R.; Uritboonthai, W.; Qin, C.; Trauger, S. A.; Siuzdak, G. Anal. Chem. 2006, 78 (3), 743-752. 23. Want, E. J.; O'Maille, G.; Smith, C. A.; Brandon, T. R.; Uritboonthai, W.; Qin, C.; Trauger, S. A.; Siuzdak, G. Anal. Chem. 2006, 78 (3), 743-752. 24. Zhu J, Djukovic D, Deng L, Gu H, Himmati F, Chiorean EG, Raftery D. Colorectal cancer detection using targeted serum metabolic profiling. J. Proteome Res. 2014, 13(9), 4120-30. 25. Nagana Gowda, G. A.; Gowda, Y.; Raftery, D. Anal. Chem. 2015, 87(7), 3800-3805. 26. Purwaha, P.; Silva, L. P.; Hawke, D. H.; Weinstein, J. N.; Lorenzi, P. L. Anal. Chem. 2014, 86(12), 5633-5637.



Phenylalanine Tyrosine

7.8

7.6

7.4

7.2

7.0

4.0

Betaine

Threonine

Serine

3.9

3.8

3.7

3.6

ppm

Carnitine

Glycine

4.1

6.0

ppm

3.5

3.3

3.4

3.2

Ornithine Creatinine Creatine Lysine

8.0

Choline Acetylcarnitine

8.2

Uridine

Tryptophan

Histidine

Pyroglutamic acid Proline

3.1

ppm

Glutamine

Aspartic acid Asparagine Dimethylglycine

2.95

2.90

Methionine

Sarcosine

2.85

2.80

2.75

2.70

2.65

2.60

2.55

2.50

2.45

2.40

ppm

Valine

ppm

Internal Reference (TSP)

0.95

Lactic acid

1.00

Alanine

1.05

Glucose

Leucine

Isoleucine

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 18 of 24

8 7 6 5 4 3 2 1 ppm 1 Figure 1: A typical 800 MHz (cryo-probe) 1D CPMG H NMR spectrum of human serum obtained after protein precipitation using methanol (1:2 serum:methanol, v/v) with expanded regions and peak annotations for 30 metabolites, which were quantitated using NMR to guide18 their absolute quantitation using MS. ACS Paragon Plus Environment

Page 19 of 24


Concentration (µM)

25 NMR

20

MS

15 10 5

Concentration (µM)

0

70 60 50 40 30 20 10 0

NMR

MS

350

Concentration (µM)

300

NMR

MS

250 200 150 100 50 0

4000

Concentration (µM)

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

3000

NMR

MS

2000 1000 0 Lactic Acid

Glucose

Figure 2: Comparison of absolute concentrations of 28 metabolites derived from NMR and NMR-guided MS for the same blood serum sample (average CV=4.1%). MS concentrations were derived based on one of the serum samples that was used as the reference sample. Similar results were obtained irrespective of the serum sample that was used as the reference (see text). Glutamine and pyroglutamic acid were omitted due to their 19 otherwise noted, the MS data were obtained using an errors arising from glutamine cyclization. 15,25,26 Unless AB Sciex QTrap 5500 mass spectrometer.ACS Paragon Plus Environment


REFERENCE y=x R² = 1

300 200 100

200 100

(a)

(b)

Male, 30 Yrs., African American 0

400

100

200 NMR

300

400

0

100

Male, 28 Yrs., Caucasian 100

200 NMR

400

300

200 100

(c) 0

200 300 NMR

y = 1.0453x + 0.4427 R² = 0.9947

300

200

0

100

400

y = 0.9885x + 1.364 R² = 0.9937

300

Female, 41 Yrs., African American

0

MS

Concentration (µM) MS

y = 0.997x - 0.4514 R² = 0.9992

300

0

(d) Female, 26 Yrs., Caucasian

0

400

0

400

100

200 300 NMR

400

300 y = 0.9852x + 0.9786 R² = 0.9966

300

y = 0.97x - 1.0914 R² = 0.9913 200

MS


400

MS


400

200

100

100

(e)

(f)

Male, 31 Yrs., Caucasian

0

Male, 36 Yrs., Caucasian

0 0

100

200 NMR

300

400

0

400 300

600 500

200

(g)

100

Female, 45 Yrs., Caucasian

0

100

200 300 NMR Concentration (µM)

200 NMR

300

400

y = 0.8935x + 2.4552 R² = 0.9958

400 300 200 100 0

0

100

700

y = 1.0156x + 2.823 R² = 0.9923

MS


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 20 of 24

(h) Female, 21 Yrs., Caucasian 0 100 200 300 400 500 600 700 NMR Concentration (µM)

Figure 3: Correlation of concentrations for metabolites derived from NMR and NMR-guided MS for 8 healthy human serum samples. Here, MS concentrations for one serum sample were set equal to those derived from NMR for the same serum sample (Reference #5, R2=1; Fig. 3a); these concentrations, along with their MRM peak areas were used as references for quantitation of the same metabolites in the remaining serum samples (Figs. 3b-3h). Virtually identical results were obtained irrespective20 of the serum sample that was used as the reference. Glutamine and pyroglutamic acid were omitted due to their errors arising from glutamine cyclization;15,25,26 ACS Paragon Plus Environment lactate and glucose were also omitted for clarity (see Figure S3 for data for all 30 metabolites).

Page 21 of 24

50 45 40 35 30 25 20 15 10 5 0

100

Concentration (µM)

U-13C,15N-Tryprophan NMR guided MS

Leucine

U-13C,15N-Leucine

90 80 70 60 50 40 30 20 10 0

40

NMR guided MS

Tyrosine U-13C,15N-Tyrosine NMR guided MS

Methionine

35

80

Concentration (µM)

Concentration (µM)

Tryptophan

Concentration (µM)

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


60 40 20

U-13C,15N-Methionine NMR guided MS

30 25 20 15 10 5

0

0

Figure 4: Comparison of absolute concentrations for four metabolites derived from NMR-guided MS and targeted MS using 13C, 15N isotope labeled (uniformly) internal standards. Here, NMR guided MS concentration for each metabolite was determined separately using each of the 8 serum samples as reference and the differences in values for each reference are shown with an error bar. The median CV between the values from NMR-guided MS method and isotope labeled internal standards was 6.2%.



Uridine 5

MS

R² = 0.9746

45

y=0.7902x+0.675

y=0.7415x+3.1053

1

y=0.7982x+2.9592

10

5 3

5

7

5

45

25 45 NMR

10

65

Tryptophan

Methionine 40

MS 30

25

20

15 15

25 35 NMR


100

30 40 NMR

50

75

85

R² = 0.9163

y=0.914x+3.3544

y=0.9652x+0.1208

y=0.7836x+9.1716

45

25 60

80 NMR

100

25

105

50

75 NMR

45

100

MS

MS

180

120 100

y=0.8684 x+6.1988

45 65

85 NMR

105

R² = 0.9891

350 250

120 140 NMR

100

160

220

R² = 0.9361

1500 1000

y=0.906x+18.519

y=0.7693x+66.749

200 150 250 350 450 550 650 NMR Concentration (µM)

180 NMR

Lactic Acid 2000

R² = 0.9529

400 300

150

140

2500

MS

450

100

140

Alanine

Proline

MS

550

80

R² = 0.9189

100

500

650

105

y=0.9419 x+6.0696

y=0.9725 x+0.3367

80 45

85 NMR

Threonine

Serine R² = 0.9542

140

65

65

220

160 Leucine R² = 0.9577

85

R² = 0.9744

65

50

40

70

Phenylalanine MS

MS

60

40 50 60 NMR

105 Tyrosine

R² = 0.9649

40

y=0.8496x+4.9842

20 30

100 Asparagine

80

40

20 20

45

40

R² = 0.9676

50

30

y=0.7438x+7.4471

y=0.7956x+2.9294

30 NMR

Isoleucine

60

R² = 0.9347

MS

R² = 0.9309

20

70

50

35

R² = 0.9834

20

25

3


30

R² = 0.9826

NMR


Aspartic Acid

Creatine

MS


40

65

7

1


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 22 of 24

200

300 400 500 NMR Concentration (µM)

y=0.83x+383.13

500 500 1000 1500 2000 2500 NMR Concentration (µM)

Figure 5: Correlation of absolute concentrations for metabolites derived from NMR and NMRguided MS for 8 healthy human (4 male; 4 female) serum samples. Here, MS concentrations for one serum sample (Sample #1) were set equal to those derived from NMR and used as the 22 in the remaining serum samples. Virtually references for quantitation of the same metabolites identical results were obtained irrespective of the serum sample used as the reference. ACS Paragon Plus Environment

Glutamine

500

400

R2 = 0.3141

400

MS

500

300

Pyroglutamic Acid R2 = 0.2072

300 200 y=0.2198x+281.01

y=0.207x+198.87

200 200

300 400 NMR

2

100

500

100

5000

Sarcosine

1.8

200

300 NMR

400

500

Glucose

2

R2 = 0.4471

R = 0.269

1.6

MS


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 23 of 24

4000

1.4 1.2

y=0.4569x+2025.2

y=0.4697x+0.728

1 1

3000 3000

1.5 2 NMR Concentration (µM)

4000 5000 NMR Concentration (µM)

Figure 6: Correlations of concentrations for individual metabolites derived from MS and NMR for 8 healthy human serum samples. Here, MS concentrations for one serum sample (Sample #1) were set equal to those derived from NMR and used as the references for quantitation of the same metabolites in the remaining serum samples. The poor correlation of glutamine and pyroglutamic acid is as anticipated based on massive glutamine cyclization discovered recently by NMR15,25 as well as MS.26



Concentration (µM)

Figure for TOC

Concentration (µM)

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

375 300 225 150 75 0

NMR

Single Reference Serum Sample

MS/MS 375 Absolute Concentrations for Test Serum Samples 300 225 150 75 0


Page 24 of 24

NMR-Guided Mass Spectrometry for Absolute ... - ACS Publications

Recommend Documents