• Search by keyword
  • Search by citation

Page 1 of 12

Application of WST-8 based colorimetric NAD(P)H detection for quantitative dehydrogenase assays

The reduction of tetrazolium salts by NAD(P)H to formazan product has been widely used to determine the metabolic activity of cells, and as an indicator of cell viability. However, the application of a WST-8 b...

  • View Full Text

Association of TM6SF2 rs58542926 gene polymorphism with the risk of non-alcoholic fatty liver disease and colorectal adenoma in Chinese Han population

Genetic factors affect the risk of non-alcoholic fatty liver disease (NAFLD) and colorectal adenoma (CRA) importantly. Transmembrane protein 6 superfamily member 2 (TM6SF2) rs58542926 is a significant genetic ...

The active role of the transcription factor Sp1 in NFATc2-mediated gene regulation in pancreatic cancer

Adenocarcinoma of the pancreas is one of the most aggressive tumor diseases affecting the human body. The oncogenic potential of pancreatic cancer is mainly characterized by extremely rapid growth triggered by...

Role of the highly conserved G68 residue in the yeast phosphorelay protein Ypd1: implications for interactions between histidine phosphotransfer (HPt) and response regulator proteins

Many bacteria and certain eukaryotes utilize multi-step His-to-Asp phosphorelays for adaptive responses to their extracellular environments. Histidine phosphotransfer (HPt) proteins function as key components ...

Up-regulation of DcR3 in microbial toxins-stimulated HUVECs involves NF-κB signalling

Sepsis is a severe condition characterised by the body’s systemic inflammatory response to infection. The specific sepsis-related biomarkers should be used in clinical diagnosis, therapeutic response monitorin...

Soluble expression of recombinant midgut zymogen (native propeptide) proteases from the Aedes aegypti Mosquito Utilizing E. coli as a host

Studying proteins and enzymes involved in important biological processes in the Aedes aegypti mosquito is limited by the quantity that can be directly isolated from the mosquito. Adding to this difficulty, digest...

Identification of the major diacylglycerol acyltransferase mRNA in mouse adipocytes and macrophages

Triacylglycerols (TAGs) are the major form of energy storage in eukaryotes. Diacylglycerol acyltransferases (DGATs) catalyze the final and rate-limiting step of TAG biosynthesis. Mammalian DGATs are classified...

Purification and characterization of cysteine protease from miswak Salvadora persica

Generally, proteases in medicinal plants had different therapeutic effects such as anti-inflammatory effect; modulate the immune response and inhibitory effect toward tumor growth. In this study, protease was ...

Effect of IAPP on the proteome of cultured Rin-5F cells

Islet amyloid polypeptide (IAPP) or amylin deposits can be found in the islets of type 2 diabetes patients. The peptide is suggested to be involved in the etiology of the disease through formation of amyloid d...

Structure and function of a lignostilbene-α,β-dioxygenase orthologue from Pseudomonas brassicacearum

Stilbene cleaving oxygenases (SCOs), also known as lignostilbene-α,β-dioxygenases (LSDs) mediate the oxidative cleavage of the olefinic double bonds of lignin-derived intermediate phenolic stilbenes, yielding ...

Simple spectrophotometric assay for measuring catalase activity in biological tissues

The details of a precise, accurate, and sensitive spectrophotometric method for measuring catalase activity are presented here. The assay was established for biological samples and depends on the rapid formati...

Establishment of a detection assay for DNA endonuclease activity and its application in the screening and prognosis of malignant lymphoma

Endonucleases play critical roles in maintaining genomic stability and regulating cell growth. The purpose of this study was to evaluate detection of endonuclease activity as an indicator in the early diagnosi...

Isolation of quercetin and mandelic acid from Aesculus indica fruit and their biological activities

In this study Aesculus indica fruit was subjected to isolation of phytochemicals. Two antioxidants quercetin and Mandelic acid were isolated in pure state. The free radical scavenging and acetyl choline esterase ...

Purification and characterization of α-amylase from Trichoderma pseudokoningii

Previous studies have demonstrated that members of Trichoderma are able to generate appreciable amount of extracellular amylase and glucoamylase on soluble potato starch. In this study the α-amylase was purified ...

Identification and characterization of smallest pore-forming protein in the cell wall of pathogenic Corynebacterium urealyticum DSM 7109

Corynebacterium urealyticum , a pathogenic, multidrug resistant member of the mycolata, is known as causative agent of urinary tract infections although it is a bacterium of the skin flora. This pathogenic bacteri...

Characterization of association of human mitochondrial lysyl-tRNA synthetase with HIV-1 Pol and tRNA 3 Lys

An important step in human immunodeficiency virus type 1 (HIV-1) replication is the packaging of tRNA 3 Lys from the host cell, which plays the role of primer RNA in the process of initiation of reverse transcripti...

Feruloyl esterase immobilization in mesoporous silica particles and characterization in hydrolysis and transesterification

Enzymes display high reactivity and selectivity under natural conditions, but may suffer from decreased efficiency in industrial applications. A strategy to address this limitation is to immobilize the enzyme....

Characterization of sulfhydryl oxidase from Aspergillus tubingensis

Despite of the presence of sulfhydryl oxidases (SOXs) in the secretomes of industrially relevant organisms and their many potential applications, only few of these enzymes have been biochemically characterized...

Glycyl-alanyl-histidine protects PC12 cells against hydrogen peroxide toxicity

Peptides with cytoprotective functions, including antioxidants and anti-infectives, could be useful therapeutics. Carnosine, β-alanine-histidine, is a dipeptide with anti-oxidant properties. Tripeptides of Ala...

Antiproliferative factor (APF) binds specifically to sites within the cytoskeleton-associated protein 4 (CKAP4) extracellular domain

Antiproliferative factor (APF) is a sialoglycopeptide elevated in the urine of patients with interstitial cystitis—a chronic, painful bladder disease. APF inhibits the proliferation of normal bladder epithelia...

Use of a special Brazilian red-light emitting railroad worm Luciferase in bioassays of NEK7 protein Kinase and Creatine Kinase

Luciferases, enzymes that catalyze bioluminescent reactions in different organisms, have been extensively used for bioanalytical purposes. The most well studied bioluminescent system is that of firefly and oth...

Dacin, one metalloproteinase from Deinagkistrodon acutus venom inhibiting contraction of mouse ileum muscle

Mice were bitten by five-pace vipers ( Deinagkistrodon acutus ), and then envenomed. It was well-known that the snake venom mainly disturbed the blood homeostasis of the envenomed victims. Ocassionally, we found th...

Serendipitous discovery of light-induced ( In Situ ) formation of an Azo-bridged dimeric sulfonated naphthol as a potent PTP1B inhibitor

Protein tyrosine phosphatases (PTPs) like dual specificity phosphatase 5 (DUSP5) and protein tyrosine phosphatase 1B (PTP1B) are drug targets for diseases that include cancer, diabetes, and vascular disorders ...

Chemical and structural characterization of α-N-acetylgalactosaminidase I and II from starfish, asterina amurensis

The marine invertebrate starfish was found to contain a novel α-N-acetylgalactosaminidase, α-GalNAcase II, which catalyzes removal of terminal α-N-acetylgalactosamine (α-GalNAc), in addition to a typical α-N-a...

Structural insight into the inactivation of Mycobacterium tuberculosis non-classical transpeptidase Ldt Mt2 by biapenem and tebipenem

The carbapenem subclass of β-lactams is among the most potent antibiotics available today. Emerging evidence shows that, unlike other subclasses of β-lactams, carbapenems bind to and inhibit non-classical tran...

Structural similarities and functional differences clarify evolutionary relationships between tRNA healing enzymes and the myelin enzyme CNPase

Eukaryotic tRNA splicing is an essential process in the transformation of a primary tRNA transcript into a mature functional tRNA molecule. 5′-phosphate ligation involves two steps: a healing reaction catalyze...

CrMAPK3 regulates the expression of iron-deficiency-responsive genes in Chlamydomonas reinhardtii

Under iron-deficient conditions, Chlamydomonas exhibits high affinity for iron absorption. Nevertheless, the response, transmission, and regulation of downstream gene expression in algae cells have not to be inve...

Binding of smoothelin-like 1 to tropomyosin and calmodulin is mutually exclusive and regulated by phosphorylation

The smoothelin-like 1 protein (SMTNL1) can associate with tropomyosin (Tpm) and calmodulin (CaM), two proteins essential to the smooth muscle contractile process. SMTNL1 is phosphorylated at Ser301 by protein ...

Increased enzymatic hydrolysis of sugarcane bagasse by a novel glucose- and xylose-stimulated β-glucosidase from Anoxybacillus flavithermus subsp. yunnanensis E13 T

β-Glucosidase is claimed as a key enzyme in cellulose hydrolysis. The cellulosic fibers are usually entrapped with hemicelluloses containing xylose. So there is ongoing interest in searching for glucose- and x...

Interactions of histatin-3 and histatin-5 with actin

Histatins are histidine rich polypeptides produced in the parotid and submandibular gland and secreted into the saliva. Histatin-3 and −5 are the most important polycationic histatins. They possess antimicrobi...

Systematic substitutions at BLIP position 50 result in changes in binding specificity for class A β-lactamases

The production of β-lactamases by bacteria is the most common mechanism of resistance to the widely prescribed β-lactam antibiotics. β-lactamase inhibitory protein (BLIP) competitively inhibits class A β-lacta...

Copper chelation and interleukin-6 proinflammatory cytokine effects on expression of different proteins involved in iron metabolism in HepG2 cell line

In vertebrates, there is an intimate relationship between copper and iron homeostasis. Copper deficiency, which leads to a defect in ceruloplasmin enzymatic activity, has a strong effect on iron homeostasis re...

Knockdown of GGCT inhibits cell proliferation and induces late apoptosis in human gastric cancer

Gamma glutamylcyclotransferase (GGCT) has been proved to be involved in various cancers, but the biological function of GGCT in gastric cancer is still largely unknown.

Identification of replication-dependent and replication-independent linker histone complexes: Tpr specifically promotes replication-dependent linker histone stability

There are 11 variants of linker histone H1 in mammalian cells. Beyond their shared abilities to stabilize and condense chromatin, the H1 variants have been found to have non-redundant functions, the mechanisms...

Avoiding false discovery in biomarker research

Human tyrosine-protein phosphatase non-receptor type substrate 1α (SIRPA) is a surface marker identified in cardiomyocytes differentiated from human embryonic stem cells. Our objective was to determine if circ...

Computational studies of human class V alcohol dehydrogenase - the odd sibling

All known attempts to isolate and characterize mammalian class V alcohol dehydrogenase (class V ADH), a member of the large ADH protein family, at the protein level have failed. This indicates that the class V...

Effect of mutations to amino acid A301 and F361 in thermostability and catalytic activity of the β-galactosidase from Bacillus subtilis VTCC-DVN-12-01

Beta-galactosidase (EC, a commercially important enzyme, catalyses the hydrolysis of β-1,3- and β-1,4-galactosyl bonds of polymer or oligosaccharidesas well as transglycosylation of β-galactopyranosi...

Sustained activation of mTORC1 in macrophages increases AMPKα-dependent autophagy to maintain cellular homeostasis

The mechanistic target of rapamycin complex 1 (mTORC1) is a well-conserved serine/threonine protein kinase that controls autophagy as well as many other processes such as protein synthesis, cell growth, and me...

Auto-thiophosphorylation activity of Src tyrosine kinase

Intermolecular autophosphorylation at Tyr416 is a conserved mechanism of activation among the members of the Src family of nonreceptor tyrosine kinases. Like several other tyrosine kinases, Src can catalyze th...

Effects of protonation on the hydrolysis of triphosphate in vacuum and the implications for catalysis by nucleotide hydrolyzing enzymes

Nucleoside triphosphate (NTP) hydrolysis is a key reaction in biology. It involves breaking two very stable bonds (one P–O bond and one O–H bond of water), in either a concurrent or a sequential way. Here, we ...

Molecular characterization of protein kinase C delta (PKCδ)-Smac interactions

Protein kinase C δ (PKCδ) is known to be an important regulator of apoptosis, having mainly pro- but also anti-apoptotic effects depending on context. In a previous study, we found that PKCδ interacts with the...

Identification and characterization of the novel nuclease activity of human phospholipid scramblase 1

Human phospholipid scramblase 1 (hPLSCR1) was initially identified as a Ca 2+ dependent phospholipid translocator involved in disrupting membrane asymmetry. Recent reports revealed that hPLSCR1 acts as a multifunc...

Tubulin is a molecular target of the Wnt-activating chemical probe

In drug discovery research, cell-based phenotypic screening is an essential method for obtaining potential drug candidates. Revealing the mechanism of action is a key step on the path to drug discovery. Howeve...

Altered activity patterns of transcription factors induced by endoplasmic reticulum stress

The endoplasmic-reticulum (ER) responds to the burden of unfolded proteins in its lumen by activating intracellular signal transduction pathways, also known as the unfolded protein response (UPR). Many signal ...

Conserved motif VIII of murine DNA methyltransferase Dnmt3a is essential for methylation activity

Dnmt3a is a DNA methyltransferase that establishes de novo DNA methylation in mammals. The structure of the Dnmt3a C-terminal domain is similar to the bacterial M. HhaI enzyme, a well-studied prokaryotic DNA meth...

Linalool isomerase, a membrane-anchored enzyme in the anaerobic monoterpene degradation in Thauera linaloolentis 47Lol

Thauera linaloolentis 47Lol uses the tertiary monoterpene alcohol ( R,S )-linalool as sole carbon and energy source under denitrifying conditions. The conversion of linalool to geraniol ...

1,2-Dichlorobenzene affects the formation of the phosphoenzyme stage during the catalytic cycle of the Ca 2+ -ATPase from sarcoplasmic reticulum

1,2-Dichlorobenzene (1,2-DCB) is a benzene-derived molecule with two Cl atoms that is commonly utilized in the synthesis of pesticides. 1,2-DCB can be absorbed by living creatures and its effects on naturally-...

Substrate specificity and function of acetylpolyamine amidohydrolases from Pseudomonas aeruginosa

Pseudomonas aeruginosa , a Gram-negative, aerobic coccobacillus bacterium is an opportunistic human pathogen and worldwide the fourth most common cause of hospital-acquired infections w...

Erratum to: Purification and characterization of a cytochrome c with novel caspase-3 activation activity from the pathogenic fungus Rhizopus arrhizus

The original article was published in BMC Biochemistry 2015 16 :21

BMC Biochemistry Reviewer Acknowledgement, 2015

The editors of BMC Biochemistry would like to thank all our reviewers who have contributed their time to the journal in Volume 16 (2015).

BMC Biochemistry

ISSN: 1471-2091

  • Search Menu
  • Advance articles
  • Author Guidelines
  • Submission Site
  • Open Access
  • About The Journal of Biochemistry
  • About the Japanese Biochemical Society
  • Editorial Board
  • Advertising and Corporate Services
  • Self-Archiving Policy
  • Dispatch Dates
  • Journals on Oxford Academic
  • Books on Oxford Academic

Japanese Biochemical Society

High-Impact Research from The Journal of Biochemistry

Explore a collection of the most read and most cited articles making an impact in the Journal of Biochemistry  published within the past two years. This collection will be continuously updated with the journal's leading articles so be sure to revisit periodically to see what is being read and cited.

Also discover the articles being discussed the most on digital media by  exploring this Altmetric report  pulling the most discussed articles from the past year.

research paper of biochemistry

Recognizing and Celebrating Women in Science

Oxford University Press is proud to support diverse voices across our publishing. In this collection, we shine a spotlight on the representation of women in scientific fields, the gains that have been made in their fields, from research and major discoveries to advocacy and outreach, and amplify the voices of women who have made a career in scientific research.

Delve into the Women in Science collection

  • Recommend to your Library


  • Online ISSN 1756-2651
  • Print ISSN 0021-924X
  • Copyright © 2024 The Japanese Biochemistry Society
  • About Oxford Academic
  • Publish journals with us
  • University press partners
  • What we publish
  • New features  
  • Open access
  • Institutional account management
  • Rights and permissions
  • Get help with access
  • Accessibility
  • Advertising
  • Media enquiries
  • Oxford University Press
  • Oxford Languages
  • University of Oxford

Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide

  • Copyright © 2024 Oxford University Press
  • Cookie settings
  • Cookie policy
  • Privacy policy
  • Legal notice

This Feature Is Available To Subscribers Only

Sign In or Create an Account

This PDF is available to Subscribers Only

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

research paper of biochemistry

Biochemistry (Moscow)

  • Coverage extends to biochemical aspects of molecular biology, bioorganic chemistry, microbiology, immunology, physiology, and biomedical sciences.
  • Encompasses experimental methods in biochemistry and theoretical contributions of biochemical importance.
  • Publishes original research articles, critical reviews, and mini-reviews.
  • Welcomes submissions in English from all countries.
  • Olga A. Dontsova

research paper of biochemistry

Latest issue

Volume 89, Issue 3

Latest articles

Protective activity of inactivated rabies vaccine using flagellin-based adjuvant.

  • Olga O. Sokol
  • Nikolai A. Nikitin
  • Irina Y. Litenkova

research paper of biochemistry

Association of Increased Homocysteine Levels with Impaired Folate Metabolism and Vitamin B Deficiency in Early-Onset Multiple Sclerosis

  • Victoria I. Lioudyno
  • Evgenia A. Tsymbalova
  • Irina N. Abdurasulova

research paper of biochemistry

Investigation of Potential Effects of Some Indole Compounds on the Glutathione S-Transferase Enzyme

  • Muhammet Serhat Özaslan

research paper of biochemistry

Novel Gain-of-Function Mutation in the Kv11.1 Channel Found in the Patient with Brugada Syndrome and Mild QTc Shortening

  • Denis Abramochkin
  • Olga S. Sokolova

research paper of biochemistry

Role of the Gut Microbiome and Bacterial Amyloids in the Development of Synucleinopathies

  • Nina P. Trubitsina
  • Anton B. Matiiv
  • Stanislav A. Bondarev

research paper of biochemistry

Journal information

  • Biological Abstracts
  • Chemical Abstracts Service (CAS)
  • Current Contents/Life Sciences
  • Google Scholar
  • Japanese Science and Technology Agency (JST)
  • OCLC WorldCat Discovery Service
  • Pathway Studio
  • Science Citation Index Expanded (SCIE)
  • TD Net Discovery Service
  • UGC-CARE List (India)

Rights and permissions

Springer policies

© Pleiades Publishing, Ltd.

  • Find a journal
  • Publish with us
  • Track your research

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • View all journals
  • Explore content
  • About the journal
  • Publish with us
  • Sign up for alerts

Collection  15 August 2018

Top 10 in Biochemistry and Molecular Biology

Explore our most highly accessed biochemistry and molecular biology articles in the first quarter of 2018. Featuring authors from around the World, these papers highlight valuable research within biochemistry and molecular biology from an international community.

research paper of biochemistry

Peptide ion channel toxins from the bootlace worm, the longest animal on Earth

  • Erik Jacobsson
  • Håkan S. Andersson
  • Ulf Göransson

research paper of biochemistry

Plasmodium dihydrofolate reductase is a second enzyme target for the antimalarial action of triclosan

  • Elizabeth Bilsland
  • Liisa van Vliet
  • Stephen G. Oliver

research paper of biochemistry

Abundant fish protein inhibits α-synuclein amyloid formation

  • Tony Werner
  • Ranjeet Kumar
  • Pernilla Wittung-Stafshede

research paper of biochemistry

Natural polyphenols as sirtuin 6 modulators

  • Minna Rahnasto-Rilla
  • Ruin Moaddel

research paper of biochemistry

In vivo fluorescence bioimaging of ascorbic acid in mice: Development of an efficient probe consisting of phthalocyanine, TEMPO, and albumin

  • Takanori Yokoi
  • Takayuki Otani
  • Kazuyuki Ishii

research paper of biochemistry

Towards high throughput GPCR crystallography: In Meso soaking of Adenosine A 2A Receptor crystals

  • Prakash Rucktooa
  • Robert K. Y. Cheng
  • Andrew S. Doré

research paper of biochemistry

Structural basis of thalidomide enantiomer binding to cereblon

  • Tomoyuki Mori
  • Toshio Hakoshima

research paper of biochemistry

Reduction in hepatic secondary bile acids caused by short-term antibiotic-induced dysbiosis decreases mouse serum glucose and triglyceride levels

  • Takuya Kuno
  • Mio Hirayama-Kurogi
  • Sumio Ohtsuki

research paper of biochemistry

Development and characterization of a novel luciferase based cytotoxicity assay

  • Hittu Matta
  • Ramakrishnan Gopalakrishnan
  • Preet M. Chaudhary

research paper of biochemistry

Malaria infected red blood cells release small regulatory RNAs through extracellular vesicles

  • Kehinde Adebayo Babatunde
  • Smart Mbagwu
  • Pierre-Yves Mantel

Quick links

  • Explore articles by subject
  • Guide to authors
  • Editorial policies

research paper of biochemistry

U.S. flag

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

  • Publications
  • Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

  • Advanced Search
  • Journal List
  • HHS Author Manuscripts

Logo of nihpa

Applications of NMR spectroscopy to systems biochemistry

The past decades of advancements in NMR have made it a very powerful tool for metabolic research. Despite its limitations in sensitivity relative to mass spectrometric techniques, NMR has a number of unparalleled advantages for metabolic studies, most notably the rigor and versatility in structure elucidation, isotope-filtered selection of molecules, and analysis of positional isotopomer distributions in complex mixtures afforded by multinuclear and multidimensional experiments. In addition, NMR has the capacity for spatially selective in vivo imaging and dynamical analysis of metabolism in tissues of living organisms. In conjunction with the use of stable isotope tracers, NMR is a method of choice for exploring the dynamics and compartmentation of metabolic pathways and networks, for which our current understanding is grossly insufficient. In this review, we describe how various direct and isotope-edited 1D and 2D NMR methods can be employed to profile metabolites and their isotopomer distributions by stable isotope-resolved metabolomic (SIRM) analysis. We also highlight the importance of sample preparation methods including rapid cryoquenching, efficient extraction, and chemoselective derivatization to facilitate robust and reproducible NMR-based metabolomic analysis. We further illustrate how NMR has been applied in vitro , ex vivo , or in vivo in various stable isotope tracer-based metabolic studies, to gain systematic and novel metabolic insights in different biological systems, including human subjects. The pathway and network knowledge generated from NMR- and MS-based tracing of isotopically enriched substrates will be invaluable for directing functional analysis of other ‘omics data to achieve understanding of regulation of biochemical systems, as demonstrated in a case study. Future developments in NMR technologies and reagents to enhance both detection sensitivity and resolution should further empower NMR in systems biochemical research.

Graphical abstract

An external file that holds a picture, illustration, etc.
Object name is nihms763208u1.jpg

1. Introduction

Since the discovery of nuclear magnetic resonance phenomenon in solids and liquids by Bloch and Purcell in 1945 [ 1 , 2 ], NMR spectroscopy has proven to be a powerful and versatile tool for structure elucidation for organic chemists, followed by structural and dynamic determination of macromolecules for the structural biologists, and more recently metabolite profiling for the field of metabolomics. The value of NMR for molecular structural and quantitative analysis is attributable to its element-selective detection and sensitivity of nuclear spin properties to the intra- and inter-molecular environment, as well as to the robust and quantitative nature of NMR measurement. These advantages have made NMR an early choice for metabolite profiling efforts [ 3 – 6 ] and an excellent partner for mass spectrometry (MS)-based metabolite profiling. For example, NMR analysis provides crucial structural parameters including functional groups, covalent linkages, and non-covalent interactions including stereochemistry, which are difficult to acquire by MS methods. On the other hand, high-resolution MS provides molecular formula information, which is an important parameter for NMR structural analysis, particularly when NMR-invisible elements ( 32 S, 16 O) are present.

Since the advent of metabolomics development, the superior sensitivity, resolution of MS (including chromatography-based MS and ultra high resolution MS), and its relative ease of data analysis, have contributed to its popularity in steady-state metabolite profiling to date [ 7 – 10 ]. In a PubMed search for metabolomic articles, only 30% of the studies involve NMR analysis. However, for the next-generation metabolomic applications involving stable isotope tracers for robust reconstruction of metabolic pathways and networks, NMR methodologies afford some unique advantages such as detailed positional isotopomer analysis for enriched metabolites, de novo structure determination of unknown metabolites (both unenriched and enriched) without the need for standards, and in situ analysis of pathway dynamics from cells to whole organisms [ 11 ].

This review will focus on the utility of NMR spectroscopy in metabolomics, with an emphasis on NMR applications in stable isotope-enriched tracer research for elucidating biochemical pathways and networks such as those shown in Figs. 18 and ​ and19. 19 . The knowledge gained from this area of research provides a ready link to genomic, epigenomic, transcriptomic, and proteomic information to achieve systems biochemical understanding of living cells and organisms.

An external file that holds a picture, illustration, etc.
Object name is nihms763208f18.jpg

NSCLC tissue slices maintain distinct metabolic reprogramming. Paired cancerous (CA) and non-cancerous (NC) lung tissues (A) freshly resected from an NSCLC patient were treated with U- 13 C 6 -glucose for 24 h at 37 °C/5% CO 2 . The tissue extracts obtained were analyzed by 1D 1 H{ 13 C}-HSQC at 14.1 T (B). The elevated buildup of various 13 C metabolites in CA versus NC tissues is consistent with enhanced glycolysis (cf. 13 C-Lac), the Krebs cycle initiated by PDH ( 13 C-4-Glu) or PCB ( 13 C-2 or 3-Glu), onecarbon metabolism ( 13 C-2-Gly), and pentose phosphate pathway/nucleotide biosynthesis ( 13 C-1′-AXP, ( 13 C-N1′ or A1′-NAD + ), as depicted in 13 C tracing of these pathways in (B). Such distinct metabolic reprogramming in CA lung tissue slices recapitulates that of human lung cancer tissues in vivo (cf. Fig. V.B.2). See Fig. V.B.2 for abbreviations.

2. NMR is a powerful tool for metabolite identification and quantification

2.1. identification and determination of metabolite structures.

NMR has long been employed to determine the chemical structures and conformations of molecules [ 12 , 13 ] and is second only to single crystal X-ray diffraction in determination of absolute stereochemistry. Because of the rich structural information and multinuclear capability of NMR, it is practical to determine the structure of a purified small molecule ab initio . In metabolomic applications, however, it is customary to work with complex extracts from cells, tissues or biofluids, which comprise mixtures of hundreds to thousands of molecules covering a very wide range of concentrations. Given the finite practical dynamic range, spectral overlap and the low sensitivity of NMR compared with mass spectrometry (MS), only a subset of metabolites in a sample can be identified and quantified (see below).

In metabolomics-based studies, the goal is to determine as many metabolites as possible (“coverage”) that represent many pathways within the vast metabolic network. For example, in a non-targeted metabolomic or biomarker study, one generally determines which NMR or MS spectral features are responding most to the varying conditions [ 14 – 21 ], followed by identification of responsive unknown compounds. For mapping known pathways, identifying all of the intermediates is not essential but is desirable. The reality is that a sizable fraction of the spectral features in metabolomic studies remain unidentified. Although one can quantify relative changes of unknown metabolites by NMR without the need for standards, this provides no useful biochemical information.

The assignment of known metabolites in complex mixtures by NMR can be based on a number of NMR parameters, including chemical shifts and their pH dependence, spin multiplicity, homonuclear and heteronuclear covalent connectivity, and the nuclear Overhauser effect. These aspects have been reviewed by us previously and will not be elaborated here [ 22 , 23 ]. The past decade of development in metabolite databases and database search tools has made it practical to assign metabolites by comparing the NMR spectral features of given samples against those of standards compiled in databases. There are now quite a few public ( http://www.eurocarbdb.org/databases/nmr-database ; http://www.bmrb.wisc.edu/metabolomics/ ; http://www.hmdb.ca/ ; http://www.bml-nmr.org/ ; http://spin.ccic.ohio-state.edu/index.php/toccata2/index ) [ 24 , 25 ] and proprietary (e.g. ChenomX, Bruker) databases available for metabolite assignment purpose. As NMR parameters can be sensitive to pH, temperature, concentration, and ionic conditions [ 26 , 27 ], the sample spectra will need to be acquired under matched conditions with the standard spectra acquired in the databases. This is a crucial aspect of NMR database search for metabolite assignment, since many laboratories differ in their extraction protocols. In addition, at some pH values, the chemical shifts can be degenerate for common metabolites, such as lactate and threonine at neutral pH [ 26 ]. Furthermore, 1D databases that rely primarily on chemical shifts can be unreliable and only a limited number of features are identified, due to spectral crowding and lack of more rigorous indication of structural features such as covalently linked partners. For a group of similar samples in a study, it is advisable to verify assignments using a set of multidimensional NMR experiments that map the carbon skeleton, including TOCSY, HSQC, HMBC and/or a combination of the two (HSQC–TOCSY).

For assigning unknown spectral features, multidimensional NMR is essential [ 28 , 29 ], and may need to include heteronuclear NMR methods based on nuclei other than 13 C, such as 15 N and 31 P. One of the significant advantages of small molecules is that long range heteronuclear couplings are sufficiently large to enable assignments via remote couplings from 1 H or 13 C to these atoms [ 30 – 32 ], which define much of the covalent atom connectivity. The combination of heteronuclear scalar coupling with chemical shift information is often sufficient to identify a compound by reference to databases and by comparison with compiled chemical shifts of compounds with different functional groups. However, the functional classes of compounds may not be clear unless additional experiments are carried out, for example by pH titration to ascertain the presence of amino or carboxylate functional groups [ 26 ].

An additional means of determining functional groups is to make use of derivatizing reagents that target specific functional groups (i.e. chemically selective, or CS, probes). By introducing an enriched atom ( 13 C or 15 N) at a strategic position in the CS probe, one can take advantage of the isotope-editing feature of NMR to detect only those metabolites with the targeted function groups (cf. Section 5.3). For example, 15 N ethanolamine and 15 N cholamine have been introduced to assign carboxylate-containing metabolites via a 2D 1 H{ 15 N}-HSQC experiment with high resolving power. The assignment can be based on the functional group and the combined 1 H and 15 N chemical shifts [ 33 , 34 ]. Other CS probes have been developed by us that select for functional groups including thiols [ 35 ] and carbonyls (aldehydes and ketones) [ 36 , 37 ]. The resulting adducts in biological extracts are optimized for both MS analysis, and 15 N edited 1 H NMR analysis that makes use of long range 2 and 3-bond couplings [ 31 ]. This has been illustrated for the carbonyl-selective aminooxy probe, for which 2D 1 H{ 15 N}-HSQC spectra (cf. Section 4.1.4) have been acquired for assignment purpose [ 31 ].

13 C enrichment in metabolites has also been developed for assignment purposes, as 13 C-based NMR analysis at natural abundance is time-consuming, particularly for low-level metabolites [ 38 , 39 ]. Several groups have suggested large-scale enrichments with 13 C to enhance detection sensitivity, while enabling 2D HCCH-TOCSY or 13 C INADEQUATE types of experiments to be carried out in a reasonable time frame [ 40 – 43 ]. A somewhat different approach has been developed by Beecher, called Isotope Ratio Outlier Analysis, or IROA [ 44 ] in which two samples are enriched to 5% and 95% with 13 C (which requires growing cells for 5 generations in the respective tracers) and mixed for NMR or MS analysis. This strategy enables the assignment of 13 C-enriched metabolites due to biosynthesis (instead of artifacts) and determination of the relative concentrations of detected metabolites [ 45 ]. While useful for cultured cell-based assignment, IROA applications to tissue-based analysis appear to be limited.

Finally, it should be noted that the experimentally acquired NMR information can be combined with ab initio calculations of NMR parameters [ 46 – 50 ] to inform stereochemistry of metabolites, which can ultimately be confirmed by chemical synthesis of the correct enantiomer for biological testing. This is critical for the correct assignment of metabolite stereoisomers with different biological activity, as illustrated by the observation of the two enantiomers of 2-hydroxyglutarate with different biological functions [ 51 – 56 ].

2.2. Metabolite quantification

2.2.1. quantification by 1-d nmr.

In addition to metabolite identification, 1-D NMR analysis via common pulse sequences such as 1 H PRESAT or NOESY is useful for quantification in metabolomic applications by calibrating the areas of metabolite resonances against those of standards with known concentrations. One important advantage of modern NMR for metabolite quantification is its large linear dynamic range and signal response (10 6 ) with a digitizer of 20 effective bits. This often obviates the need for standard curves in NMR analysis, which are commonly required for MS-based quantification. However, it should be noted that despite the high linear dynamic range of NMR digitizers, in practice this can be compromised by the Lorentzian line shape and non-optimal shimming and therefore be considerably lower.

It is practical to use a single standard for both quantification and internal chemical shift referencing without the need for authentic metabolite standards. Common choices of 1 H or 13 C NMR standards include DSS (2,2-dimethyl-4-silapentane-4-sulfonic acid)/DSS-d 6 (to eliminate the CH 2 signals that can interfere with analysis of some metabolites), or TSP (3-(trimethylsilyl) propanoic acid)/TSP-2,2,3,3,-d 4 for D 2 O and tetramethylsilane (TMS) for CDCl 3 . When metabolite resonances are well-resolved, 1D NMR quantification can be robust if appropriate signal corrections are made for partial relaxation [ 23 ]. This is particularly relevant for compounds with singlet resonances, including the reference compounds, where the T 1 values tend to be long [ 27 , 57 ]. We have chosen a recycle time (acquisition time + relaxation delay) of 6 s for 1D 1 H NMR analysis, which is a compromise between speed of acquisition or sample throughput and the magnitude of correction needed for partial saturation. Under these conditions, the majority of metabolite resonances are >99% relaxed. However, for the methyl resonances of DSS or TMS where T 1 values are 2–3 s, the resonances may be 5–15% saturated, and require a correction.

There is an often unappreciated problem with employing internal standards for quantifying metabolites in biofluids (e.g. plasma or urine) or extracts that contain appreciable amounts of protein and/or lipids. Their resonances can be broadened significantly by binding to macromolecules, thereby affecting their signal intensity. Fig. 1 illustrates such a problem, where the 1D 1 H NMR spectra of a mouse tumor extract are compared before and after treatment with 80% acetone for efficient protein removal. It is clear that signal intensity increased upon protein removal for the internal DSS standard ( Fig. 1 inset) and a number of metabolites. The extent of the signal intensity distortion varies for different compounds, ranging from negative change for glutathiones (GSH + GSSG) to high enhancement for glucose and lactate. This is because when mobile metabolites are in exchange with macromolecules, a fraction of the metabolite signal is lost to NMR detection, and the extent of signal distortion depends on the binding strength, exchange dynamics, number of binding sites, and metabolite concentrations. If strong binding occurs, metabolites can be precipitated along with proteins, which then lead to signal loss, instead of enhancement with protein removal. The problem for distortion of metabolite signal intensity by interaction with macromolecules is exacerbated when CPMG spin-echo pulse trains are used to suppress broad macromolecular resonances. Human serum contains 35–50 mg/mL or 0.5–0.75 mM of albumin, with several binding sites for different metabolites [ 58 ]. Additional binding sites can come from several other abundant serum proteins for a total protein content of 60–80 mg/mL [ 59 ]. The NMR-detectable metabolites in plasma or serum also depend on dietary history, as dietary components may compete for the same binding sites [ 60 , 61 ]. Moreover, those metabolites (e.g. Ala) that are in fast exchange on the chemical shift timescale may display chemical shifts notably different from those observed in reference databases, giving rise to problems with metabolite assignment.

An external file that holds a picture, illustration, etc.
Object name is nihms763208f1.jpg

Influence of extraction procedure on NMR spectral quality of tissue extracts. Patient-derived tumor tissues were implanted into Nod/SCID Gamma mice and let grow to ca. 1 cm in diameter before harvest and pulverization in liquid N 2 for metabolite extraction using our CH 3 CN:H 2 O:CHCl 3 (2:1.5:1 v/v) solvent partitioning method [ 384 ]. The resulting polar extract was dissolved in 50% D 2 O for 1D 1 H NMR measurement at 14.1 T (black line), followed by 80% acetone treatment to remove residual proteins and analysis again by 1 H NMR (red line). The two 1 H spectra were normalized by the peak area of the H3 resonance of lactate (3-Lac). The spectral region from 0.8 to 4.7 ppm is shown, along with the CH 3 resonance of the DSS-d 6 standard in the inset. The latter was much lower in intensity and broader in linewidth without acetone than with acetone precipitation of proteins, which indicates the influence of DSS binding to residual proteins on resonance quality. In contrast, the signal intensities of most metabolites were enhanced by acetone precipitation (e.g. H1-βglucose, H2-Gly), which is again consistent with signal attenuation by protein binding. In contrast, the reduced glutathione (GSH) resonances (i.e. H2-Cys and H4-Glu resonances of GSH) were absent from the spectrum after acetone treatment, which could be attributed to insolubility of GSH in 80% acetone. Re-extraction of the protein pellet with 60% acetonitrile restored the GSH signals.

One way to circumvent the problem of macromolecule-induced signal intensity distortion of internal standards is to implement direct spiking of the metabolite of interest. However, this is impractical for quantifying a large number of metabolites [ 26 ], and is still problematic when metabolite interactions with macromolecules are severe. Another alternative is electronic spiking (ERETIC), where a quantified digital signal is added to the spectrum, which does not suffer from any solute interferences and any frequency can be synthesized to avoid spectral overlaps [ 62 ]. The main disadvantage appears to be the calibration of such electronic signal against actual metabolite concentration, which depends on a large number of factors. Nevertheless, this is an attractive approach where high absolute quantitative accuracy is not needed.

Signal intensity distortion problems can be greatly reduced by efficient protein and lipid removal, via a combination of extraction with deproteination and delipidation. However, the nature of deproteination via precipitation is important. If the precipitation method also unfolds the proteins, then any tightly bound metabolites will be released, allowing the total amounts of metabolites to be measured. A method that precipitates or removes protein without denaturation (e.g. ultrafiltration) will remove the tightly bound metabolites, leaving only those that were unbound, or in fast exchange on the time scale of the precipitation process to be analyzed. These considerations may in part account for the differential metabolite yield observed for different sample processing techniques [ 63 , 64 ]. Furthermore, some well established protein precipitation methods can lead to destruction of metabolites, especially those using strong acids such as perchloric acid or trifluoracetic acid ([ 65 ], Fan, unpublished data).

Macromolecules can also Interfere with metabolite analysis, due to their multiple and broad resonances. These signals can be removed during data acquisition, using for example CPMG pulse trains [ 66 , 67 ]. In practice, these methods differentially distort signal intensities via the lengthy pulse train, do not always eliminate the lipoprotein contributions, and cannot overcome signal intensity distortion due to differential binding.

Another notable issue with 1D 1 H NMR quantification of metabolites in metabolomic applications is spectral crowding. Conventional peak integration for quantification purposes can only be applied to well-resolved 1 H resonances. Peak fitting or deconvolution assuming a Lorentz-Gaussian lineshape can provide accurate peak areas in crowded region [ 27 , 38 ], provided that the peak overlapping is not too severe and that the baseline is correctly determined. However, the assumption of a Lorentz-Gaussian lineshape is not always correct for unresolved features, leading to less reliable peak area determination. Multi-dimensional NMR can circumvent this issue by improving spectral resolution. One complication with 2D NMR quantification is that the peak volumes depend on the appropriate transfer functions, in addition to metabolite concentration. Approaches to metabolite quantification using 2D TOCSY and HSQC data have been described that compare peak volumes or heights of target analytes with those of a given standard acquired under identical spectral conditions [ 69 – 71 ], which helps circumvent such complications.

2.2.2. Quantification in 2D NMR

Fractional distribution of 13 C or 15 N isotopomers can be determined with acceptable precision in 2D 1 H TOCSY spectra without the need for standards. This is achieved by integrating the volumes of the two 13 C or 15 N satellite cross peaks and normalizing to the sum of the volumes of all cross peaks (cf. Fig. 4 ). We found that this can achieve an accuracy of 5–10% at natural abundance, and better for isolated resonances at higher enrichment [ 57 , 72 , 73 ]. This level of accuracy is sufficient for metabolic pathway analysis [ 74 ].

An external file that holds a picture, illustration, etc.
Object name is nihms763208f4.jpg

2.3. Sensitivity of NMR quantification

With modern cryogenically cooled probes, raw proton sensitivity in a 5 mm tube of 0.1% ethylbenzene (EtBz) can be expected to be 6000–8000:1, or 800–1200:1 for 2 mM sucrose in salt-free water, at 14–18.8 T. In practice the sensitivity (signal height compared with root mean square noise level, which is quoted by the instrument manufacturers) may scale linearly with B 0 (Lane, unpublished observations) rather than the more optimistic B 0 7 / 4 [ 75 ]. By using matched susceptibility plugs such as Shigemi tubes, the volume needed for a 5 mm probe is about 300 μL, or 3 μmol 0.1% EtBz, 0.6 μmol 2 mM sucrose. Thus to reach a signal-to-noise ratio (SNR) of 10:1 in 10 min acquisition, about 0.6 nmol of material are needed under ideal conditions. Significantly lower quantities are needed in smaller diameter coils such as 1.7 mm or even 1 mm diameter (<10 μL sample needed) where a performance boost of up to 2–3-fold for the same amount of material can be expected, despite the considerably poorer filling factor of a very small diameter tube [ 76 , 77 ]. These probes are best suited for analyzing mass limited samples. However, they do not necessarily perform as well when the sample is too concentrated, which may degrade line shape and alter chemical shifts due to high viscosity, high salt and solute–solute interactions.

Compromised cryoprobe performance, particularly at high magnetic field strength, can be expected for biofluids and some tissue extracts with high ionic strengths. These include human serum, which has an ionic strength of approximately 145 mM and an osmolality of ca. 300 mOsMl [ 78 , 79 ]. Based on Bruker NMR specifications, and observed in practice [ 80 ], the signal-to-noise ratio for 5 mm cryoprobes at 18.8 T shows a decrease in sensitivity of at least 2-fold on increasing the concentration of NaCl from 0 to 150 mM. Roughly speaking, the sensitivity of a 5 mm cryogenic probe and a conventional room temperature probe are expected to become equal at around 300 mM sodium chloride, and presumably at lower concentrations at ultrahigh field strengths (e.g. 23 T or higher). Such effects are much smaller in small diameter coils, and can be alleviated to some extent by placing a smaller diameter tube in a larger coil, e.g. 3 mm in a 5 mm probe or a 1.7 mm tube in a 3 mm coil, provided that the increased solute concentration does not degrade the spectral quality through increased viscosity or aggregation.

Sensitivity enhancement for 13 C in cryoprobes is better in larger diameter probes than microprobes, if they are optimized for 1 H detection, largely because of the poorer filling factor in the outer coil. When 13 C direct detection is desired, 13 C optimized inner coils give the best performance, albeit at lower sensitivity than 1 H optimized indirect detection for the same sample mass.

3. NMR is uniquely suited for tracking stable isotope labeling at specific atomic positions (isotopomer analysis)

The common isotopes for analysis in metabolic NMR are 13 C (1.1%), 15 N (0.37%) and 31 P (100%). Most metabolites contain carbon atoms with directly attached 1 H and/or 1 H 2–3 bonds apart, which give rise to 13 C- 1 H couplings. 13 C- 1 H coupling constants ( J CH ) in common linear metabolites typically scale with the 13 C chemical shift, which is exploited in some adiabatic decoupling schemes. 1 J CH ranges from ca. 125 Hz for alkyl methyl groups to >200 Hz for sp 2 aromatic carbons [ 30 , 82 ]. In addition to 13 C- 1 H coupling, 13 C- 13 C coupling is practical to utilize for 13 C isotopomer analysis by NMR. For the common amino acids, 3 J CC and 2 J CC range from <1 to ca. 7 Hz [ 30 ].

Nitrogen atoms can also have directly attached protons, as in amino and amide groups, or can be two or more covalent bonds from a hydrogen atom (e.g. in cholines). However, amino and amido protons can exchange readily in aqueous solution so may require special conditions to observe 15 N- 1 H coupling [ 13 , 81 ]. Phosphorus atoms with directly attached protons (e.g. phosphonates or phosphines) are not usually natural metabolites, while those with more distantly attached protons are common among nucleotides and sugar phosphate metabolites. Thus, detection of 13 P- 1 H interactions in natural metabolites typically uses long-range coupling. Moreover, as C-C, C-N, and C-P bonds are common in metabolites, it is important to consider also 13 C- 13 C, 13 C- 15 N, and 31 P- 13 C couplings for isotopomer analysis.

One of the great advantages of NMR is the ability to determine elements independently by direct detection, or by the influence of one element on another through coupling interactions, which together can be used for an enormous number of isotope-editing experiments to observe specific isotopomer distributions [ 28 , 29 , 83 , 84 ].

There are two general approaches to NMR-based isotopomer analysis, direct observation of the nucleus of interest [ 85 ], and indirect detection via an attached spectator spin, usually the proton. Direct detection has the advantage that all sites are generally observed, whereas indirect detection infers the presence of a coupled spin that does not have a directly attached proton. Although carbonyl and quaternary carbons can be observed directly, they do not show a significant NOE, cannot be enhanced by one-bond INEPT or DEPT, and typically have long T 1 values, which means that their signals are likely to be substantially attenuated owing to incomplete relaxation. Direct 13 C detection also benefits from a wider chemical shift range than indirect proton detection, which can be important in crowded 1D spectra of extracts.

However, indirect detection is considerably more sensitive than direct detection. For 13 C, the SNR advantage of 1 H detection over 13 C detection is theoretically 4 5/2 = 32 for equal numbers of nuclei (and notably higher when comparing CH 3 or CH 2 ) under fully relaxed conditions, without NOE or INEPT. This advantage decreases to 8–10-fold if maximal NOE or INEPT transfers are included. Indirect detection through one-bond couplings does not observe carbonyl and quaternary carbons, but they can be inferred from their couplings to neighboring carbon atoms or detected through multiple bond couplings by HMBC experiments for example. For small molecules with long T 2 values, indirect detection can make use of small long-range couplings for more versatile analysis [ 83 , 86 , 87 ]. The advantages of indirect detection for 15 N are very much greater than 13 C, due to its much lower γ .

In stable isotope tracer studies, tracers enriched in biologically relevant nuclei such as 13 C, 15 N, or 2 H are introduced to cells, tissues, or whole organisms for transformation into various metabolites. Depending on the transformation pathway(s), isotope labels are incorporated into specific atomic position(s) of given metabolites. For example, 13 C-2-glucose is metabolized into 13 C-2-lactate via glycolysis and into 13 C-1-ribose-5-phosphate via the pentose phosphate pathway (PPP). By determining the positions (positional isotopomers) and the number of the stable isotope labels (mass isotopologues) in transformed metabolites, the transformation pathways can be reconstructed with confidence. It is also practical to discern compartment-specific pathways with this approach, which is not possible with steady-state metabolite profiling alone. NMR is excellently suited for positional isotopomer analysis, while MS is tailor-made for mass isotopologue analysis. A number of 1D and 2D NMR methods have been utilized for this purpose, as described below.

It should also be noted that stable isotope enrichment patterns in metabolites are not only essential for reconstructing metabolic networks, but are also fundamental inputs for flux modeling, which depending on the approach may require isotopic steady state, or be compatible with dynamic conditions, as necessitated by many studies in vivo (cf. Section 5.2) [ 88 – 90 , 85 , 91 – 95 ].

3.1. Direct observe methods

3.1.1. 1d methods.

Both 1D 1 H and 13 C NMR methods have been employed to analyze for 13 C positional isotopomers [ 96 – 139 ]. The covalent coupling network of 1 H and 13 C atoms in given metabolites gives rise to pairs of satellite peaks in the 1D 1 H NMR spectra. The splitting patterns of the satellites depend on the number and positions of 13 C atoms present in the metabolite. For example, the methyl (Me) or H-3 resonance of 12 C-lactate (no 13 C present) is a doublet with no 13 C satellites ( Fig. 2A ). When C-3 or C-2 of lactate is enriched in 13 C, a pair each of doublet satellites arises equidistant from the central methyl doublet of 12 C-lactate but with different J coupling constants ( Fig. 2B ). For the lactate isotopomer labeled with 13 C at both C-2 and C-3, a pair of doublet of doublet satellites is expected ( Fig. 2C ), while a pair of doublet of doublet of doublet satellites arises from uniformly 13 C labeled lactate ([U- 13 C]-Lac) ( Fig. 2D ). The very similar coupling constants, as shown in the simulation, result in such observed multiplets. The latter pattern is evident in a polar extract of medium obtained from culturing human cancer cells in [U- 13 C]-glucose ( Fig. 2E ), which indicates the transformation of [U- 13 C]-glucose into [U- 13 C]-Lac via glycolysis.

An external file that holds a picture, illustration, etc.
Object name is nihms763208f2.jpg

13 C-lactate satellite patterns in 1 H NMR. Panels (A–D) display the simulated splitting patterns of the 1 H-3 resonance of lactate with these methyl protons coupled to 12 C-3 (A), to 13 C-3 or 13 C-2 (B), to 13 C 2 -2,3 (C), and 13 C 3 -1,2,3 (D). An example 1 H spectrum of cell culture medium after 24 h of lung adenocarcinoma A549 cell growth in U- 13 C 6 -glucose is shown in (E), where the complex splitting pattern of 13 C satellites of H-3 present in U- 13 C 3 -lactate is evident.

Despite the much lower sensitivity of detection than 1 H NMR, 1D 13 C NMR is more commonly employed for delineating different 13 C positional isotopomers due to its superior resolution and interpretability of 13 C label position(s) based on the spin coupling patterns within the covalent network. For example, 13 C enrichment simultaneously at C-3 and C-4, C-1 and C-2, C-2 and C-3, as well as C-1, C-2, and C-3 can be deduced from a 1D 13 C NMR spectrum of methylacetone glucose, a chemically modified plasma glucose ( Fig. 3A ). The 13 C positional isotopomers of fatty acyl chains of lipids can be ascertained from the 1D 13 C NMR spectrum of a cell lipid extract, as shown in Fig. 3B . The consecutive 13 C labeling of terminal (ω) CH 3 and penultimate ω-1 CH 2 ( 13 CH 3 - 13 CH 2 -) and of ω-CH 3 , ω-1 CH 2 , and ω-2 CH 2 ( 13 CH 3 - 13 CH 2 - 13 CH 2 -) are evident, respectively based on the doublet pattern of ω-CH 3 and ω-1 CH 2 and on the triplet pattern of ω-1 CH 2 .

An external file that holds a picture, illustration, etc.
Object name is nihms763208f3.jpg

13 C coupling patterns of 13 C-enriched metabolites in 13 C NMR. (A) 13 C NMR spectrum of monoacetone glucose derived from plasma glucose of rat with infusion of 13 C 2 -3,4-glucose and IP injection of [U- 13 C]proprionate and 2 H 2 O; S: singlet; D34, D12 or D23: doublet from C-3 and C-4, C-1 and C-2 or C-2 and C-3 coupling, respectively; Q: doublet of doublet from C-2 coupling to both C-1 and C-3 [ 140 ]. Reprinted from Anal Biochem, 327, E.S. Jin, J.G. Jones, M. Merritt, S.C. Burgess, C.R. Malloy, A.D. Sherry. Glucose production, gluconeogenesis, and hepatic tricarboxylic acid cycle fluxes measured by nuclear magnetic resonance analysis of a single glucose derivative, pp. 149–155, (2004), with permission from Elsevier. (B) 13 C NMR spectrum of lipids extracted from glioblastoma cultured in [U- 13 C 6 ]glucose. The inset shows the expansion of the 23 ppm region. ω, ω-1, and ω-2: terminal CH 3 , penultimate CH 2 , and CH 2 neighbor to ω-1 CH 2 ; d: doublet; t: triplet; FA-(CH 2 )n-: internal CH 2 of fatty acyl chains. Excerpted with permission from R.J. DeBerardinis, A. Mancuso, E. Daikhin, I. Nissim, M. Yudkoff, S. Wehrli, C.B. Thompson, Beyond aerobic glycolysis: transformed cells can engage in glutamine metabolism that exceeds the requirement for protein and nucleotide synthesis, Proc Natl Acad Sci USA, 104 (2007) 19345–19350 [ 129 ]. Copyright (2007) National Academy of Sciences, U.S.A.

3.1.2. 2D methods

2D 1 H NMR methods offer even more versatile and rigorous means for positional isotopomer analysis. This is illustrated by 1 H- 1 H total correlation spectroscopy (TOCSY) detection of three different 13 C positional isotopomers of glutamate in a 13 C-enriched cell extract ( Fig. 4B ), when compared with the corresponding unlabeled data ( Fig. 4A ). Namely, three pairs of 13 C satellite cross-peaks distributed at equidistance to the cross-peak of Glu-2H to Glu-4H (both 12 C-attached, Fig. 4B ) indicate the occurrence of 12 C-2,4-Glu, 13 C 1 -2-Glu, 13 C 1 -4-Glu, and 13 C 2 -2,4-Glu (structures shown in Fig. 4C ) (cf. [ 22 ] for additional satellite cross-peak patterns). The TOCSY analysis not only resolves these isotopomers, but also rigorously confirms their identity based on the chemical shifts, covalent connectivity, and 1 H- 13 C coupling constants.

A 13 C-edited version of TOCSY, 1 H- 13 C HCCH-TOCSY, enables selective detection of covalently linked 1 H’s attached to consecutively 13 C labeled carbon atoms (cf. Fig. 5E ). Fig. 5D [ 108 ] illustrates such a 2D spectrum, acquired from the polar extract of a lung adenocarcinoma A549 cell line grown in 13 C 6 -glucose under control conditions. Cross-peaks arising from lactate and glutamate isotopomers with 13 C labeled at both C-2 and C-3 positions are clearly observed, along with those derived from 13 C 5 -1,2,3,4,5-ribose containing uracil nucleotides (UXP). Also shown are the 1D projection spectra of 13 C labeled A549 cell extracts obtained from selenite (Se) and control treatments ( Fig. 5A and C , respectively), along with the high-resolution 1 H NMR spectrum of the control extract ( Fig. 5B ). Fig. 5C illustrates the 13 C editing capacity of HCCH-TOCSY for enhancing the detection of minor 13 C labeled components such as glycerophosphocholine (GPC), and unresolved components such as uracil nucleotides (UXP) (denoted by *), which is difficult to ascertain in the unedited spectrum ( Fig. 5B ). It is clear from the projection spectra ( Fig. 5A versus C ) that the selenite treatment blocks the synthesis of 13 C 2 -2,3-lactate and Glu as well as 13 C 5 -ribose-containing UXP, but enhances the production of 13 C 3 -1,2,3-glycerol-containing GPC. These data are consistent with reduced glycolytic capacity ( 13 C 2 -2,3-lactate as marker) and anaplerotic pyruvate carboxylation ( 13 C 2 -2,3-Glu as marker) (cf. Fig. 18 ) as well as increased degradation of phosphatidylcholine lipids (PCL), elicited by selenite (cf. Fig. 17 , [ 108 ].

An external file that holds a picture, illustration, etc.
Object name is nihms763208f5.jpg

2-D 1 H– 13 C HCCH–TOCSY analysis detects positional isotopomers with consecutive 13 C labels. Lung adenocarcinoma A549 cells were grown in 13 C 6 -glucose under control or selenite (Se) treatment. HCCH–TOCSY analysis was performed on the polar cell extracts to enable selective detection of protons attached to consecutively bonded 13 C atoms, as illustrated in (E). The 2D spectrum in (D) reveals Glu and lactate isotopomers consecutively 13 C labeled at C-2 and C-3 positions, as well as adenine nucleotide (AXP) isotopomer consecutively 13 C labeled at C-1 to C-5 of the ribose unit. Also shown are the 1D high-resolution 1 H NMR spectrum of the control cell extract (B) as well as the 2D projection spectra of the control (C) and selenite-treated (A) cell extracts. The 13 C editing also enhances the detection of minor 13 C labeled components such as glycerophosphocholine (GPC) and unresolved 13 C labeled components such as uracil nucleotides (UXP) (denoted by *) (cf. B and C). Such a capability is unique in revealing the decreased synthesis of 13 C 5 -1,2,3,4,5-ribose containing UXP and increased production of 13 C 3 -1,2,3-glycerol containing GPC by the selenite treatment (cf. A and C) (adapted from [ 108 ]).

An external file that holds a picture, illustration, etc.
Object name is nihms763208f17.jpg

Metabolomics-edited transcriptomic analysis for selenite-induced perturbations of glycolysis, pentose phosphate pathway, TCA cycle, glutaminolysis, and fatty acid metabolism in A549 cells. Panel (A) shows the pathways perturbed in selenite-treated A549 cells, based on SIRM analysis, which was used to interrogate parallel transcriptomic data for relevant gene expression changes. The signal log ratio of selenite treated to control mRNA expression is listed below each protein into which the mRNA was translated. Negative and positive values indicate a decrease and increase in expression, respectively. → indicates down regulation of genes, inhibition of protein activity, or depletion of 13 C-labeled metabolites (denoted by *). → indicates up regulation of genes, activation of protein activity, or accumulation of labeled metabolites. → indicates trace reactions, transport processes, and changes in concentration of unlabeled metabolites. The open green arrows indicate glutaminolysis of unlabeled Gln taken up from the medium and subsequent oxidation to generate unlabeled malate. Dashed arrows indicate multiple reaction steps. Acetyl-CoA C: Acetyl-CoA carboxylase (ACC); AMPK: 5′-AMP-activated protein kinase; DAG: diacylglycerol; DHAP: dihydroxyacetone phosphate; FPBase: fructose bisphosphatase; Glnase IP: glutaminase interacting protein; G-6-P DH: glucose-6-phosphate dehydrogenase; GAP, glyceraldehyde-3-phosphate; GPC, glycerophosphorylcholine; α-KG DH: α-ketoglutarate dehydrogenase; LDH: lactate dehydrogenase; LysoPLase: lysophospholipase; ME: malic enzyme; Malate DH: malate dehydrogenase; N transporter: neutral amino acid transporter; OAA: oxaloacetate; PC: phosphorylcholine; PFK2: phosphofructokinase 2; PLases: phospholipases; Pyruvate DH: pyruvate dehydrogenase (adapted from [ 386 ] with permission). Selenite-reduced fatty acyl synthesis is evident in the 1D 13 C projection spectrum of 2D 1 H{ 13 C}-HSQC analysis of A549 lipids shown in (B). PCL: phosphatidylcholine lipids; PEL: phosphatidylethanolamine lipids. The Western blot analysis for the protein level of phosphorylated or inactivated ACC (pACC) and AMPK (pAMPK) in response to selenite treatment is shown in (C), where selenite-induced increase in pACC and decrease in pAMPK levels relative are evident. This data suggests inactivation of AMPK by phosphorylation underlies enhanced ACC phosphorylation or inactivation by AMP.

3.2. Indirect observe methods

Indirect detection of 1 H attached to 13 C, 15 N, or other NMR-observable heteroatoms presents a far more sensitive means for analyzing labeled positional isotopomers of metabolites than direct detection of the enriched heteroatoms, due to the larger gyromagnetic ratio of 1 H ( γ H ) than those of the heteroatoms ( γ X ). A sensitivity gain of up to 32-fold is to be expected for 1 H{ 13 C} and 1 H{ 15 N} heteronuclear single quantum coherence spectroscopy (HSQC), as the sensitivity gain is proportional to ( γ H / γ X ) 5/2 (cf. also [ 22 ]). Usable 1D 1 H{ 13 C}-HSQC spectra can be acquired with as little as 0.1 to 0.2 mg of dry cell mass (Fan et al., unpublished data), which is currently not achievable with direct 13 C detection methods.

3.2.1. 1D methods

We have routinely employed 1D 1 H{ 13 C}-HSQC analysis for a quick overview of metabolites labeled with 13 C at different carbon positions [ 98 , 101 – 103 , 105 – 107 , 141 ]. Fig. 6 shows an example 1D 1 H{ 13 C}-HSQC spectrum (the first slice of a 2-D experiment, i.e. a 1D 13 C-edited proton spectrum) of a polar extract of lung adenocarcinoma PC9 cells grown in 13 C 6 -glucose for 24 h. Except for phosphocholine, of which the N-methyl resonance (NMe-PCholine) arises from the naturally abundant 13 C (cf. unlabeled versus 13 C 6 -glucose spectra), all other metabolites exhibit 13 C enrichment at various carbon positions, indicating transformations of 13 C 6 -glucose into lactate, Ala (from glycolysis), Glu, Asp, glutathiones (GSH + GSSG) (via the Krebs cycle), Gly (via glycolysis and the Ser-Gly pathway), purine (AXP, NAD + ) and pyrimidine (UXP, CXP, UDPG, UDP-GlcNAc) nucleotides (via the pentose phosphate and nucleotide ring synthesis pathways) (cf. [ 98 , 106 ]).

An external file that holds a picture, illustration, etc.
Object name is nihms763208f6.jpg

1D 1 H{ 13 C}-HSQC analysis provides an overview of 13 C positional isotopomers of metabolites in a 13 C labeled PC9 cell extract. Lung adenocarcinoma PC9 cells were grown in 13 C 6 -glucose or unlabeled glucose for 24 h before extraction. The two polar extracts were subjected to 1D 1 H{ 13 C}-HSQC analysis for 13 C positional isotopomers of various metabolites. Lac: lactate; GSH, GSSH: reduced and oxidized glutathiones; PCholine: phosphocholine; Me: methyl; UXP, CXP, AXP: uracil, cytosine, and adenine nucleotides, respectively; UDPGlcNAc: UDP-N-acetylglucosamine; UDPG: UDP-glucose.

Similarly, 1D 1 H{ 15 N}-HSQC analysis selectively and sensitively detects 15 N-labeled adducts of carbonyl-containing metabolites, as shown in Fig. 7 . The 15 N label is introduced by reacting metabolites such acetone, pyruvate, and sugar aldehydes with an 15 N-aminooxy probe, 15 N-QDA, to form the oxime adducts ( Fig. 7A ). The resulting adducts have 15 N coupled to metabolite protons at 2 or 3 bond distance with 2 J NH or 3 J NH in the 2–3 Hz range [ 142 ]. This coupling enables sensitive detection of 15 N via its covalently linked 1 H by 1 H{ 15 N}-HSQC. The 15 N editing eliminates interfering signals from unlabeled metabolites, allowing highly selective detection of low abundance carbonyl-containing metabolites such as pyruvate and acetone (cf. unedited 1 H versus 15 N-edited HSQC; Fig. 7B ). It should also be noted that volatile acetone is detected as the 15 N-QDA adduct but will be lost via lyophilization in the untreated extract.

An external file that holds a picture, illustration, etc.
Object name is nihms763208f7.jpg

1D 1 H{ 15 N}-HSQC analysis provides sensitive detection of 15 N-enriched metabolite derivatives. A549 cell extract was derivatized with an 15 N-enriched aminooxy probe 15 N-QDA (N-(2- 15 N-aminooxyethyl)-N,N-dimethyl-1-dodecylammonium), which form oxime adducts with carbonyl-containing metabolites such as acetone, pyruvate, and sugar aldehydes (A). The 2 or 3-bond coupling between 15 N of the probe and 1 H of the metabolites enables detection of 15 N via its covalently linked 1 H by 1 H{ 15 N}-HSQC analysis (B).

3.2.2. 2D methods

Since the development of multichannel high field spectrometers and of a variety of labeling schemes for macromolecules, there is an enormous number of possible experiments that can select for particular groups of atoms via scalar couplings [ 143 – 150 ]. These are important for analyzing metabolite mixtures, where scalar interactions are key to metabolite identification. In particular, isotope editing experiments that rely on scalar coupling between protons and isotopically enriched X-nuclei are most attractive for profiling labeling patterns of metabolites in mixtures, for which HSQC (or HMQC) is the simplest experiment available. An important side benefit of such experiment is the ability to determine positional isotopomer distributions in multiple isotope-enriched molecules in a complex mixture. For example, the simultaneous presence of 13 C at a specific position in 15 N-enriched metabolites in a complex mixture can be readily observed using 1 H{ 15 N} HSQC to select for 15 N containing molecules, which also give rise to 13 C satellites in the 1 H dimension of the 2D data [ 28 , 81 ]. A similar scheme could be applied to select 31 P-containing metabolites which are enriched in 13 C, for example. Moreover, the topology of a 13 C or 15 N coupled spin network can be defined by appropriate 2D and 3D heteronuclear experiments. Consider a molecule such as lactate, which can have up to 2 3 = 8 isotopomers when enriched in 13 C (cf. Fig. 2 ). The uniformly enriched isotopomer would show a very characteristic HSQC spectrum, including the 1-bond 1 H- 13 C coupling and distant coupling to C2 and C3 of lactate. Other editing experiments can be performed to better define the coupling partners, such as HMBC (cf. Fig. 10 ) [ 28 ] and HCACO (cf. Fig. 11 ) [ 84 , 151 , 152 ] for linking aliphatic protons to 13 C-carbonyl carbons, as well as the HACAN triple resonance experiment for detect 15 N amino group. The various 2D or 3D experiments that make use of the spin topology pathways of labeled metabolites are depicted for 13 C 5 , 15 N-glutamate in Fig. 8 . Some of these 2D experiments are illustrated below.

An external file that holds a picture, illustration, etc.
Object name is nihms763208f8.jpg

Spin topology pathways for 13 C 5 , 15 N 1 -glutamate in 2 or 3D NMR experiments. Various 3D NMR experiments that can be performed as the 2D version for 13 C and 15 N labeled Glu are depicted. Arrows denote magnetization transfer pathways in 13 C 5 , 15 N 1 -Glu via homonuclear and heteronuclear scalar couplings. C: 13 C; N: 15 N.

An external file that holds a picture, illustration, etc.
Object name is nihms763208f10.jpg

1 H{ 13 C}HMBC detects multiple bond 1 H– 13 C connectivity including that between carbons with no directly attached proton(s) or carbons bonded to heteroatoms in the covalent network of given metabolites. The 2D HMBC contour plots of aliphatic (A) and carbonyl/double bonded carbon regions (B) are shown along with the corresponding 1D high-resolution 13 C NMR spectral regions of a polar extract obtained from lung adenocarcinoma PC9 cells grown in 13 C 6 -glucose for 24 h. The cross-peak for C-5 (C=O) to H-4 of Glu is evident, along with those for C-4 to H-3, as well as C-2 and C-3 to H-4. Also evident are the cross-peaks for C-1 (C=O) to H-3 and C-2 to H-3 of Ala. The expanded 1D 13 C spectra for C-2, C-3, C-4, and C-5 of Glu as well as C-1 and C-2 of Ala illustrate more clearly the 13 C– 13 C coupling patterns. C: 13 C.

An external file that holds a picture, illustration, etc.
Object name is nihms763208f11.jpg

2D 1 H{ 13 C}HCACO analysis provides rigorous identification of 13 C positional isotopomers of carbonyl-containing metabolites in the polar extract of A549 cells. This 2D experiment detects covalent connectivity as cross-peaks from 13 C-carbonyl carbon to aliphatic protons of metabolites via 13 C– 13 C coupling, as illustrated in the example structure diagram (top right). The cross-peaks of 13 C-1 to H-3 and 13 C-1 to H-2 of lactate (Lac) denote the presence of 13 C 3 -1,2,3-lactate, those of 13 C-1 to H-2 and 13 C-4 to H-3 of Asp indicate the presence of 13 C 2 -1,2- and 13 C 2 -3,4-Asp, respectively, and those of 13 C-1 to H-2, 13 C-1 to H-3, and 13 C-5 to H-4 of Glu are consistent with the presence of 13 C 2 -1,2-, 13 C 3 -1,2,3-, and 13 C 2 -4,5-Glu, respectively.

2D 1 H{ 13 C}-HSQC acquired in high-resolution (HR) mode represent a powerful technique for resolving and identifying 13 C-enriched isotopomers, not only for carbons with attached protons but also for quaternary carbons, which are normally not observable by HSQC. 2D HSQC also provides more rigorous resonance assignment than its 1D version, as 13 C chemical shifts and coupling patterns can be obtained, in addition to 1 H chemical shifts. Fig. 9 illustrates the 2D 1 H{ 13 C}-HR-HSQC spectrum (A) of a 13 C-enriched A549 cell extract, along with the 1D projection spectrum in the 13 C dimension (B) for a better visualization of the 13 C- 13 C coupling patterns. Highlighted are the coupling patterns for the ribose subunit attached to the adenine ring of NAD + , where the doublet of C-1′, triplet of C-2′, and doublet of C-5′ patterns indicate the presence of 13 C 5 -1,2,3,4,5-ribose-containing NAD + isotopomer. Also illustrated is the presence of 13 C 3 -1,2,3-lactate (Lac), deduced from the triplet pattern of C-2. Even though the carboxylate carbon (C-1) of lactate is normally not observable by HSQC, its 13 C enrichment is inferred from the splitting pattern of C-2.

An external file that holds a picture, illustration, etc.
Object name is nihms763208f9.jpg

2D 1 H{ 13 C}HR-HSQC analysis provides rigorous identification of 13 C positional isotopomers of metabolites via 13 C– 13 C coupling patterns. A polar extract of A549 cells grown in 13 C 6 -glucose was analyzed by high-resolution (hr) 1 H{ 13 C}-HSQC at 18.8 T. The 2D HSQC contour plot (A) is shown along with the 1D projection spectrum in the 13 C dimension (B), where the 13 C– 13 C coupling patterns are better visualized. The doublet of C-1′, triplet of C-2′, and doublet of C-5′ of ribose attached to the adenine ring (A) of NAD + indicate the presence of 13 C 5 -1,2,3,4,5-ribose-containing NAD + isotopomer, while the triplet of C-2-lactate (Lac) identifies the presence of 13 C 3 -1,2,3-lactate.

To observe 13 C- 1 H connectivity for carbons with no attached proton(s) (e.g. carbonyls or quaternary carbons) or carbons bonded to heteroatoms (e.g. C-N), 2D 1 H{ 13 C}-HMBC (heteronuclear multiple bond correlation spectroscopy) analysis can be performed, as illustrated in Fig. 10 . This experiment complements 2D 1 H{ 13 C}-HSQC analysis to provide 13 C- 1 H covalent linkages missing from the latter, thereby enabling more rigorous assignment of 13 C positional isotopomers of metabolites. For example, the cross-peaks for C-1 to H-3 of Ala and C-5 to H-4 of Glu in the 2D HMBC plot confirm the 1D 13 C resonance at 176.5 ppm as C-1 of Ala, and that at 181.9 ppm as C-5 of Glu. The doublet for the C-1 and doublet of doublets for the C-2 resonances of Ala indicates the presence of 13 C 3 -1,2,3-Ala, while the doublet for C-5 of Glu arises from 13 C 2 -4,5-Glu. 13 C 3 -1,2,3-Ala is derived via glycolysis, while 13 C 2 -4,5-Glu is synthesized via glycolysis + canonical Krebs cycle (PDH-initiated) from the tracer 13 C 6 -glucose (cf. Fig. 10 ).

For carbonyl-containing metabolites, 2D HCACO analysis reveals 13 C carbon linkages to aliphatic protons via 13 C– 13 C coupling. As shown in Fig. 11 for an A549 cell extract, the cross-peak patterns indicate the presence of different 13 C positional isotopomers for Asp, Glu, citrate (Cit), lactate (Lac), reduced (GSH) and oxidized (GSSG) glutathiones, and UDP-glucose (UDPG). In particular, the synthesis of 13 C 2 -4,5-Glu and 13 C 2 -1,2- or 13 C 2 -3,4-Asp from 13 C 6 -glucose is indicative of transformations via glycolysis, followed by PDH-initiated Krebs cycle activity. The production of 13 C 3 -1,2,3-lactate and 13 C 2 -1,2-Gly points to glycolysis, and glycolysis plus the Ser-Gly pathway, respectively.

In addition to 13 C-edited 2D methods, the 2D 13 C- 13 C INADE-QUATE (Incredible Natural Abundance DoublE QUAntum Transfer Experiment) method [ 153 ] can be performed on 13 C-enriched extracts of low mg amounts (dry mass) of biological samples, which is normally not practical with unenriched samples. Instead of utilizing the attached 1 H for detection, this experiment analyzes 13 C- 13 C one-bond coupling in the carbon covalent network with the direct ( F 2 ) dimension as 13 C chemical shift axis and the indirect ( F 1 ) dimension as the double quantum frequency axis. It is therefore a powerful technique for tracing the entire 13 C carbon skeleton of individual metabolites. Fig. 12 illustrates the 13 C- 13 C cross-peak patterns for 13 C-enriched metabolites in the polar extract of a lung adenocarcinoma PC9 cell line grown in 13 C 6 -glucose for 24 h. The cross-peaks of 13 C-1 to 13 C-2 to 13 C-3 of Ala and lactate (Lac), 13 C-1 to 13 C-2 of Gly, and 13 C-4 to 13 C-5 of Glu, as well as their splitting patterns unambiguously indicate the presence of 13 C 3 -1,2,3-Ala and -lactate, 13 C 2 -4,5-Glu, and 13 C 2 -1,2-Gly, respectively.

An external file that holds a picture, illustration, etc.
Object name is nihms763208f12.jpg

2D 13 C– 13 C INADEQUATE analysis provides rigorous identification of 13 C positional isotopomers of 13 C-enriched metabolites in the polar extract of PC9 cells grown in 13 C 6 -glucose. This 2D experiment detects carbon covalent connectivity as cross-peaks via 13 C– 13 C coupling, as illustrated in the example structure diagram (top left). The cross-peaks of 13 C-1 to 13 C-2 to 13 C-3 of Ala and lactate (Lac), 13 C-1 to 13 C-2 of Gly, and 13 C-4 to 13 C-5 of Glu as well as their splitting patterns unambiguously indicate the presence of 13 C 3 -1,2,3-Ala and -lactate, 13 C 2 -4,5-Glu, and 13 C 2 -1,2-Gly, respectively. Shown above the 2D contour plots are the corresponding 1D high-resolution 13 C NMR spectral regions.

For analysis of metabolites with severely overlapping resonances in a narrow spectral region, as exemplified by sugars and phosphosugars, selective 3D experiments, such as a 31 P HSQC–TOCSY experiment, can be employed to resolve and identify the components and their isotopomer distributions. The main impediments to the use of 3D experiments are sensitivity and experiment time, though non-uniform sampling in two indirect dimensions [ 154 ], GFT or PR reconstructions [ 155 , 156 ] may make such applications practical in some cases.

4. Sample preparation considerations

4.1. metabolic quenching.

At the time of sampling, metabolism, and thus metabolite profiles, can change quickly if samples are not quenched immediately. A common means for metabolic quenching of biological samples is flash-freezing in dry ice or liquid N 2 . Liquid N 2 freezing is far more efficient than dry ice freezing, as it is a liquid rather than solid and has a much colder temperature (−196 °C) than dry ice (−78.5 °C). Slow freezing (even in the 10s of second range) can lead to changes in labile metabolites [ 158 ]. This aspect must be considered to minimize sampling artifacts, particularly for metabolically active thick specimens such as brain and liver tissues. For bulk tissue freezing, freeze-clamping in liquid N 2 within seconds is the method of choice for arresting metabolism. Example freeze-clamping devices are shown in Fig. 13 . These devices press fresh tissues into 1–2 mm thin layers so that they can be frozen evenly within seconds in between two aluminum plates (with excellent thermal conductivity) pre-equilibrated in liquid N 2 or other cryogens [ 159 ]. This process eliminates differential freezing from the surface to the core of thick tissue specimens with simple immersion in cryogens, which can lead to changes in labile metabolites in response to hypoxia [ 158 ].

An external file that holds a picture, illustration, etc.
Object name is nihms763208f13.jpg

Example freeze-clamping devices. A Wollenberger-like device [ 385 ] is routinely used in our laboratory for rapid freezing of human and animal tissues to quench metabolism. A commercially available freeze-clamping device for small tissue freezing (BioSqueezer) is also shown ( http://www.biosciencetechnology.com/product-releases/2010/12/snap-freeze-tissue-clamp ).

Metabolic quenching by quick homogenization of bulk tissues in ice-cold 0.6 M perchloric acid (PCA) has been shown to give metabolite profiles comparable to quenching and pulverization in liquid N 2 [ 160 ]. Although convenient, this method excludes the possibility of sharing tissues for RNA or DNA extraction for transcriptomic or genomic analysis. The extremely low pH of PCA will degrade acid-labile metabolites such as fructose-2,6-bisphosphate, NAD(P)H, ascorbate, high energy phosphate metabolites, and sugar nucleotides, etc. [ 65 ]. The process of removing PCA by titration with KOH or K 2 CO 3 is also tedious and time-consuming [ 65 ].

For adherent mammalian cells on culture plates, quenching by direct freezing is impractical. Cells need to be cleared from the medium components before quenching, which can be achieved by quickly rinsing cells in ice-cold PBS, followed by solvent quenching in cold acetonitrile (e.g. −20 °C) or methanol (e.g. −80 °C) [ 65 , 161 ]. However, this cell quenching method is unsuited to quenching fast-growing microbes, which requires much quicker manipulation before freezing. Direct whole broth quenching with cold solvents, quenching with cold or hot solvents (e.g. cold 60% methanol, hot 90% ethanol, or −23 °C glycerol-saline) prior to cell separation [ 162 – 166 ], and fast filtration methods [ 162 , 163 , 167 ] have been developed for such difficult cases. Direct whole broth quenching is quick but has the disadvantage of contaminating low levels of cellular metabolites with medium broth components [ 167 ]. Methanol quenching coupled with cell separation can circumvent this problem, but suffers a significant loss (>60%) of cellular metabolites due to non-specific leakage. Use of buffered quenching solvents [ 164 , 165 , 168 ] or isotonic wash solution coupled with fast filtration [ 167 ] has been reported to recover relatively high levels of cellular metabolites.

4.2. Choice of extraction solvent

Extraction buffers for NMR-based metabolomic analysis should meet the following criteria: (1) be readily removable; (2) have low ionic strength; (3) be efficient for inactivation and removal of macromolecules; (4) be non-reactive. In fact, the solvent used for quenching metabolism is invariably a component of the extraction buffer. Perchloric acid (PCA) or trichloroacetic acid (TCA) is often used in the literature to extract polar metabolites while quenching the cell or tissue metabolism by denaturing active proteins [ 3 , 4 , 96 , 97 , 106 , 108 , 114 , 169 – 177 ]. However, as indicated above, these strong acids can lead to metabolite degradation during extraction, and increase the ionic strength of the final extract. Organic solvent-based extraction methods are now widely accepted for efficient extraction of polar and non-polar metabolites while eliminating proteins by precipitation without increasing the ionic strength of the extract. Varying combinations of methanol, H 2 O, acetonitrile, CHCl 3 , and methyl-tert-butyl ether (MTBE) have been used to quench and extract a wide variety of biological samples from biofluids, cells, to tissues [ 37 , 98 , 101 , 103 – 105 , 142 , 162 – 164 , 167 , 168 , 175 , 178 – 196 ]. Most of the solvent-based procedures share commonalities with the Folch method [ 197 ], in that they involve separation of aqueous and water-immiscible CHCl 3 layers for simultaneous extraction of polar and non-polar metabolites, with protein precipitates situated in between the aqueous and CHCl 3 layers. A more recent method replaces CHCl 3 with the less toxic and lower density MTBE, which gives comparable lipid recovery while enabling protein precipitates to be pelleted at the bottom of the extraction tube [ 194 – 196 ]. Such an improvement allows for the development of automation of the extraction process, to greatly enhance sample throughput [ 195 ]. We also found that the MTBE method gave comparable polar metabolite profiles by 1 H NMR to those from our CH 3 CN–H 2 O–CHCl 3 method [ 65 ] ( Fig. 14 ).

An external file that holds a picture, illustration, etc.
Object name is nihms763208f14.jpg

MTBE-based extraction gives comparable 1 H NMR profile of polar metabolites as CHCl 3 -based extraction. An NSCLC patient-derived lung cancer cell line (PDLC216) was quenched in cold CH 3 CN before extraction of polar and non-polar metabolites by the addition of H 2 O and CHCl 3 or MTBE. The solvent composition for the CHCl 3 extraction method was CH 3 CN:H 2 O:CHCl 3 (2:1.5:1) for the 1st extraction and CHCl 3 :methanol:butylated hydroxytoluene (2:1:1 mM) for the 2nd extraction. The solvent composition for the MTBE extraction method was CH 3 CN:H 2 O:MTBE 3 (2:1:3) for the 1st extraction and MTBE:methanol:H 2 O (10:3:2.5) for the 2nd extraction.

4.3. Selective detection via chemical derivatization

To take advantage of the superior sensitivity of 1 H NMR for selective detection of specific functional groups of metabolites, 15 N labeled carboxyl-selective tags such as 15 N-ethanolamine [ 198 ] or 15 N-cholamine [ 199 ] have been developed (cf. Fig. 15A for reaction) and demonstrated for 1 H{ 15 N}-HSQC analysis of biofluids. An example spectrum of 15 N-cholamine-tagged human serum is shown in Fig. 15B , where a number of amino and organic acids are resolved. The permanent positive charge of the tagged metabolites also enhances their detection by mass spectrometry [ 199 ]. Similarly, 13 C-formylation of the amine group of metabolites coupled with 1 H{ 13 C}-HSQC analysis has been applied to biofluid analysis for amino metabolites at the low μM range [ 200 ]. Reliable quantitation by 2D HSQC is also achievable with this tagging method.

An external file that holds a picture, illustration, etc.
Object name is nihms763208f15.jpg

15 N-cholamine tagging of carboxyl groups enables their selective analysis in human serum by 1 H{ 15 N}-HSQC. The reaction of 15 N-cholamine with carboxylates is shown in (A) and the 1 H{ 15 N}-HSQC contour plot of a tagged human serum is illustrated in (B). (1) Aconitic acid; (2) adipic acid; (3) alanine; (7) aspartic acid; (8) betaine; (9) citric acid; (11) cystine; (12) formic acid; (15) glutamic acid; (17) glycine; (20) histidine; (21) 3-hydroxybutyric acid; (24) isocitric acid; (28) lactic acid; (29) leucine; (32) malic acid; (37) phenylalanine; (40) pyroglutamic acid; (45) threonine; (46) tryptophan; (47) tyrosine; (48) valine. Reprinted with permission from F. Tayyari, G.A. Gowda, H. Gu, D. Raftery, 15 N-cholamine–a smart isotope tag for combining NMR- and MS-based metabolite profiling, Anal Chem, 85, 8715–8721. Copyright (2013) American Chemical Society [ 199 ].

For tagging carbonyl and sulfhydryl functional groups, we have respectively developed QDA ( Q uaternary ammonium D odecyl dimethyl A minooxy) [ 37 ] and QDE ( Q uaternary ammonium D odecyl dimethyl E thyl iodoacetamide) [ 201 ] probes. The 15 N labeled QDA was employed to tag carbonyl-containing metabolites such as pyruvate ( Fig. 16A ) in A549 cell extracts to enable their detection by 1D and 2D 1 H{ 15 N}-HSQC via 2-bond ( 2 J NH ) or 3-bond ( 3 J NH ) 1 H– 15 N coupling [ 142 ]. The 15 N labeled oxime adducts such as 15 N-QDA-pyruvate and 15 N-QDA-acetone are clearly seen in the 15 N-edited 1D 1 H spectrum (red line, Fig. 16B ), while they are overwhelmed or missing (acetone removed by lyophilization) in the unedited 1 H spectrum (black line, Fig. 16B ). The 2D HSQC spectrum resolves quite a few more 15 N-QDA adducts including dihydroxyacetone-3-phosphate (DHAP) and ribose ( Fig. 16C ).

An external file that holds a picture, illustration, etc.
Object name is nihms763208f16.jpg

Carbonyl tagging by 15 N-QDA enables their selective analysis by 1 H{ 15 N}-HSQC. A549 cells were quenched and reacted in cold acetonitrile containing 0.3 mM 15 N-QDA (N-(2- 15 Naminooxyethyl)-N,N-dimethyl-1-dodecylammonium) to form QDA adducts of carbonylated metabolites such as pyruvate in (A). The tagged extract was analyzed by both 1D (B) and 2D 15 N-edited HSQC (C, adapted from Fig. 5 of A.N. Lane, S. Arumugam, P.K. Lorkiewicz, R.M. Higashi, S. Laulhé, M.H. Nantz, H.N.B. Moseley, T.W. M. Fan, Chemoselective detection and discrimination of carbonyl-containing compounds in metabolite mixtures by 1 H-detected 15 N nuclear magnetic resonance, Magnetic Resonance in Chemistry, 53 (2015) 337–343 with permission from John Wiley & Sons [ 142 ]) at 14.1 Tesla. DHAP: dihydroxyacetone-3-phosphate; Pyr: pyruvate; Ara: arabinose.

4.4. Metabolite detection without extraction

The non-invasive nature of NMR analysis has enabled metabolite profiling of intact tissues without extraction. This involves the High Resolution Magic Angle Spinning (HRMAS) approach, where tissue sample (mg quantities) is placed in a small rotor at 4 °C and spun at the magic angle at ca. 4 kHz during NMR measurements. Standard water suppression techniques such as NOESY, presaturation or CPMG are applied, so 2D rotor synchronized adiabatic TOCSY and 1 H/ 13 C HSQC experiments can be implemented [ 202 ]. The averaging of the spatial variation in magnetic susceptibilities is often excellent, giving very high quality 1D and 2D 1 H NMR spectra for a wide variety of tissues [ 203 – 212 ].

The main advantages of HRMAS analysis are that no tissue extractions are required and that the amount of free metabolite is determined. There are however several uncertainties that limit HRMAS applications in tracer-based studies. As the excised tissue is frozen, thawed, squeezed and sealed in a tight space with no O 2 and nutrient supply nor provision for removing waste products, the biochemical integrity of the tissue may be severely compromised. It is also impractical to observe physiologically relevant metabolic activity under such stressful conditions, most notably hypoxia and acidosis. Furthermore, HRMAS quantification may be compromised by variable metabolite interactions with macromolecules. To date, HRMAS has not been applied in tracer-based pathway mapping studies.

5. Applications of NMR to metabolomics

5.1. metabolite profiling studies.

The focus of the present review is on the use of stable isotope tracers in conjunction with NMR analysis for advancing metabolomic understanding. This is a relatively less popular area of NMR applications. The majority of NMR applications in metabolomics research are metabolite profiling-based, and do not involve tracer use. These have been extensively reviewed elsewhere [ 19 , 213 – 222 ] and will not be elaborated here. We emphasize that employing proper sample preparation techniques (cf. Section 4) is essential to realizing the highly reproducible nature of NMR-based metabolite profiling, compared with any other technique. For NMR profiling of biofluids, high reproducibility is readily achieved if standardized protocols are followed [ 49 ]. We have also observed good reproducibility of NMR analysis of cell and tissue extracts with standardized procedures [ 26 , 223 , 224 ].

5.2. Stable isotope tracer studies

The concept of linear metabolic pathways is more didactic than realistic, based on cursory examination of established metabolic charts (e.g. http://www.sigmaaldrich.com/content/dam/sigma-aldrich/docs/Sigma/General_Information/metabolic_pathways_poster.pdf ). For example, glycolysis may be defined as the linear path of glucose oxidation to pyruvate (Emden-Meyerhof-Parnas pathway). However, essentially all of the glycolytic intermediates also participate in multiple other pathways, such as lactic fermentation, the Krebs cycle, Ala metabolism (pyruvate), the pentose phosphate pathway (glucose-6-phosphate or G6P, fructose-6-phosphate or F6P, and glyceraldehyde-3-phosphate or GAP), glycogen synthesis (G6P), hexosamine biosynthesis (F6P), fructose metabolism (F6P), regulation of glycolysis (F6P via fructose-2,6-bisphosphate), lipid synthesis (DHAP), serine and Gly synthesis (3-phosphoglycerate or 3PGA), to name just a few (cf. Figs. 18 and ​ and19). 19 ). Such redundancy of pathway participation is the norm rather than an exception, with Glu as the champion that can participate in some 55 different pathways. Multiple parallel and sequential pathways in the biosynthesis of complex metabolites are common, as exemplified by UDP-hexosamine and nucleotide biosynthesis [ 90 , 225 , 178 ]. Profiling metabolite content by NMR and/or MS methods can provide an overview of the metabolic status of given biological systems. However, this information is grossly insufficient in resolving the intersecting metabolic networks that underlie the steady-state metabolite changes (cf. Fig. 2in Ref. [ 11 ]), nor can it provide the network dynamics that are important for deciphering metabolic regulation.

The use of isotopically enriched tracers is indispensible for elucidating metabolic pathways and pathway dynamics. Due to the high sensitivity and ease for detection of radioactive metabolites, the radiolabeled tracer approach has been used extensively in life sciences for probing metabolic reactions and pathways since its development in the 1950s [ 226 , 227 ]. However, the hazardous nature of radiochemicals adds undesirable experimental complications, and is unsuited for research use in human subjects. Another important disadvantage is that except for 3 H, radionuclides are NMR-silent, which compromises the opportunities for the use of NMR in the analysis of isotopically enriched metabolites labeled at different atomic position(s) (i.e. positional isotopomers). The label position(s), together with the number of isotope labels in metabolites (mass isotopologues) is crucially important for robust reconstruction of metabolic pathways and networks [ 11 ].

Despite the demonstration of stable isotope use in metabolic research since the early 1900s [ 228 ], stable isotope tracers were not widely employed in tracing metabolism until the development of metabolomics techniques. This is at least in part due to the cost of the tracers [ 229 ], and challenges in detecting stable isotope-labeled metabolites. The important advantage of the stable isotope tracer approach is that biologically relevant stable isotopes including 13 C, 15 N, and 2 H are NMR active. 13 C and 15 N in particular can be readily detected using a wealth of direct and indirect detection NMR methods, thanks to the past decades of development in biological NMR [ 230 , 143 , 144 , 146 , 148 , 231 – 233 , 229 ]. Not only the sensitivity and resolution, but also the capacity for structure elucidation, of NMR analysis have improved tremendously, such that comprehensive 13 C or 15 N labeling patterns of a large number of metabolites can now be ascertained using as little as a sub mg quantity of dry cell or tissue biomass or low μL of biofluids. These NMR advancements, together with revolutionary advances in biological mass spectrometry, have changed the landscape for the use of isotope tracers in metabolic research, with stable isotope tracers now preferred over radiotracers. We will illustrate with some recent studies how NMR in combination with stable isotope tracers has been applied to systematically uncover new metabolic knowledge. It should be noted that these studies are meant as examples, instead of a comprehensive review of the literature.

5.2.1. In vitro measurement cell-based tracing.

The bulk of the stable isotope tracer-based metabolic studies on cell systems have involved 13 C labeled metabolite analysis, using both direct and indirect detection methods. 13 C-glucose labeled at different positions has been the predominant tracer used in different cell-based studies. One study utilized 1 H NMR analysis coupled with 1- 13 C 1 -, 2- 13 C 1 -, or 6- 13 C 1 -glucose tracers to track the time course production of 13 C-Ala and -lactate isotopomers in the medium of continuously cultured hybridoma cells in a chemostat. This information was then used to model glucose fluxes through glycolysis, the Krebs cycle, and the pentose phosphate pathway (PPP). The metabolic flux analysis revealed that 20% of the glucose consumed was metabolized via the oxidative branch of the PPP, anaplerotic pyruvate carboxylation was negligible, and malic enzyme flux that supports fatty acid biosynthesis was about 10% of the glucose uptake rate. These data indicate a higher biosynthetic rate of continuously growing cells in a chemostat than adherent cells grown in perfused hollow-fiber bioreactors, thereby contributing towards improved bioengineering of glucose metabolism in mammalian cells [ 234 ].

Uniformly 13 C labeled glucose (U- 13 C 6 -glucose) coupled with 13 C NMR has been employed to track mitochondrial Krebs cycle metabolism in insulinoma cells. The existence of two separate pools of pyruvate was uncovered based on 13 C isotopomer analysis; one pool involved pyruvate oxidation via acetyl CoA formation, and the other exchange with the Krebs cycle intermediates or pyruvate cycling [ 111 ]. The latter pool was responsive to glucose-stimulated insulin secretion. Such compartmentalized metabolic pathways for regulating insulin release would be very difficult to uncover without the aid of stable isotope tracers and 13 C isotopomer analysis.

13 C NMR isotopomer analysis of INS-1 cells treated with U- 13 C 6 -glucose or U- 13 C 6 -Leu revealed a similar contribution of three different anaplerotic pathways under basal glucose treatment, including pyruvate carboxylation (PC) of pyruvate derived from glucose, non-glycolytic sources, and Gln oxidation. When exposed to high glucose (15 mM) tracer but not to Gln addition, PC flux increased several fold and was associated with increased insulin secretion. These data enabled the authors to conclude that anaplerotic pyruvate carboxylation, but not anaplerotic Gln oxidation, plays a major role in modulating insulin secretion [ 235 ].

13 C NMR coupled with various 13 C labeled glucose, acetate, and glutamate tracers was employed to probe metabolic trafficking between astrocytes and neurons, which is essential to neuronal functions. The NMR-based 13 C isotopomer analysis helped delineate the modulation of the glutamate-glutamine cycle between astrocytes and neurons by pyruvate cycling in astrocytes, and the Ala-lactate shuttle between astrocytes and neurons. These metabolic exchange processes enable ready supply of neurotransmitter Gln to neurons while maintaining carbon and nitrogen balance between the two cell types [ 236 ].

To explore how lithium chloride (a long-established therapeutic for bipolar disorder) can impact glial and neuronal metabolism as well as metabolic trafficking between them, we employed 1 H{ 13 C}-HSQC NMR and GC–MS coupled with a U- 13 C 6 -glucose, 13 C 1 -3-lactate, and 13 C 2 -2,3-Ala tracer approach [ 106 ]. We found that astrocytes grown in U- 13 C 6 -glucose released not only 13 C-lactate but also significant amounts of 13 C-Ala, -Gln, and -citrate into the culture media, and that this process along with pyruvate carboxylase (PCB)-initiated Krebs cycle activity was enhanced by the lithium treatment. The excretion of the latter metabolites is not commonly observed in other mammalian cells. We then showed that neurons had a high capacity for metabolizing externally supplied 13 C 1 -3-lactate and 13 C 2 -2,3-Ala and that lithium enhanced the neuronal metabolism of 13 C 1 -3-lactate via PCB-initiated Krebs cycle activity. These findings suggest that lithium’s therapeutic efficacy may be associated with its ability to improve metabolic trafficking to support the Gln–Glu cycle between astrocytes and neurons as well as neuronal energy metabolism. Without the use of 13 C tracers and detailed 13 C-isotopomer analysis by NMR, it would not be feasible to uncover the lithium-enhanced PC pathway and intercellular interactions between astrocytes and neurons.

We also applied various 1D and 2D NMR techniques to decipher the anti-cancer action of selenite in U- 13 C 6 -glucose-grown lung cancer cells [ 108 ]. Detailed 13 C labeling patterns of metabolites were systematically determined by NMR methods coupled with GC–MS analysis and used to reconstruct metabolic networks perturbed by the selenite treatment. This information was used to deduce the dysregulation of individual enzyme genes by interrogating the transcriptomic data acquired in parallel ( Fig. 17 ). Such integration of metabolomics and transcriptomics information (Metabolomics-Edited Transcriptomics Analysis or META) can greatly facilitate the functional annotation of transcriptomic data, while enabling the discovery of specific gene expression changes that govern metabolic perturbations elicited by the stressor without prior knowledge (see also Section 6).

By employing U- 13 C 6 -glucose tracer, 2D 1 H TOCSY NMR, and high-resolution Fourier transformion cyclotron resonance mass spectrometric (FT-ICR-MS) analyses [ 180 ], we measured the time course of 13 C incorporation into UDP-N-acetyl-D-glucosamine (UDP-GlcNAc) in prostate cancer Ln3CaP-LN3 cells, which is a key precursor for O- and N-linked glycosylation of proteins that modulate a number of cellular processes such as protein targeting to organelles and nutrient sensing. The time course data were modeled using a G enetic A lgorithm for I sotopologues in M etabolic S ystems (GAIMS) to deconvolute 13 C mass isotopologue data into 13 C positional isotopomer species, from which time-dependent 13 C incorporation into individual metabolic units (e.g. glucosamine, acetyl, uracil, and ribose) of UDP-GlcNAc was resolved. This in turn yielded the fractional contribution of glycolysis, the Krebs cycle, pentose phosphate pathway, hexosamine biosynthetic pathway, and pyrimidine biosynthetic pathway to UDP-GlcNAc synthesis via GAIMS modeling. Such information will greatly facilitate flux modeling of metabolic networks, without achieving the isotopic steadystate required by most current approaches [ 89 , 237 ].

We then employed both U- 13 C 6 -glucose and U- 13 C 5 , 15 N 2 -Gln tracers coupled with 2D 1 H TOCSY and 1 H{ 15 N}-HSQC analyses to track de novo synthesis of nucleotides and RNA in lung cancer A549 cells. We found that Gln is preferred over glucose for pyrimidine ring synthesis, via the synthesis of Asp via the Krebs cycle, and that 13 C incorporation from glucose into total RNA is a major sink for nucleotides during cell proliferation. We further demonstrated that an anti-cancer Se agent (methylseleninic acid) inhibits nucleotide turnover and incorporation into RNA, which underlies its toxicity in cancer cells [ 178 ].

Using U- 13 C 6 -glucose and U- 13 C 5 , 15 N 2 -Gln tracers and the S table I sotope- R esolved M etabolomics (SIRM) approach, the modulation of oxidative glucose and Gln pathways by a prominent oncogene c-MYC and hypoxia was investigated in a MYC-inducible human Burkitt lymphoma P493 cell model [ 98 ]. The 1 H{ 13 C}-HSQC NMR analysis of labeled isotopomers helped elucidate the functional role of c-MYC in sustaining cell survival and proliferation via a novel glucose-independent Krebs cycle oxidation of Gln under hypoxia and glucose deprivation. Such understanding revealed the promising potential for employing glutaminase (a key enzyme in Gln oxidation) inhibitors for anti-cancer therapy. Tissue-based tracing

The use of a 13 C 1 -2-acetate tracer coupled with 1D 13 C or 2D 1 H{ 13 C}-HMQC-TOCSY analysis of extracts enabled ex vivo distinction of the flux through the mitochondrial Krebs cycle ( V TCA ) from that of the exchange ( V x ) between the mitochondria and the cytoplasm in tracer perfused rat hearts [ 238 ]. Both NMR methods gave comparable V TCA and V x (denoted respectively by Asp or Glu formation from OAA or αKG transamination) under control and aminooxyacetate (AOA, a transaminase inhibitor) treatment. However, the inclusion of 13 C malate isotopomer data markedly improved the confidence limit for V x determination by HMQC-TOCSY analysis, which also provides higher sensitivity of detection than the 1D 13 C method. This study illustrates the feasibility of flux analysis of compartmentalized metabolic processes ex vivo in functioning organs.

By infusing 13 C 1 -2-glucose into live rats, followed by extract analysis of dissected cortical and hippocampal tissues using 1D 1 H{ 13 C} NMR, Chowdhury et al. were able to ascertain the action mechanism of an anti-depressant drug riluzole in rat brains [ 239 ]. The changes in 13 C isotopomer patterns of Glu, Gln, and γ -aminobutyrate revealed the enhancing effect of riluzole on neuronal oxidation of glucose via both PDH and PC-mediated Krebs cycle, as well as on Gln/Glu cycling between neurons and glia. It should be noted that total metabolite levels were not altered by the riluzole treatment, which means that it would be impractical to resolve the action of riluzole by metabolite profiling without the aid of the tracer approach. This study also demonstrate the power of NMR-based 13 C isotopomer analysis in resolving in vivo complex interactions of metabolic networks across different compartments.

In a SIRM study with a bolus injection of U- 13 C 6 -glucose into SCID mice, we compared the time course kinetics of 13 C incorporation into various metabolites in different mouse organs using a combination of 1D 1 H, 1 H{ 13 C}-HSQC, high-resolution 2D 1 H{ 13 C}-HSQC-TOCSY (for resolving 13 C coupling patterns), and GC–MS analyses [ 105 ].We found evidence for a high level of pyruvate carboxylation in brain, heart, and liver, which is consistent with their strong PC protein expression. Different organs also differed substantially in their capacity to synthesize Gln/Glu from glucose, with brain, heart, and liver among the highest. This is consistent with the presence of a Gln/Glu cycle, a high demand for Gln as neurotransmitter in brain tissues, and liver being the primary organ for Gln synthesis. We further compared the 13 C labeling patterns of metabolites in tumorous lung (orthotopically implanted with human lung cancer PC14PE6 cells) versus paired normal lung from the same mouse. We found evidence for enhanced glycolysis, increased Krebs cycle activity (both PDH and PC-initiated), and Gly biosynthesis in the tumorous lung, which is consistent with the metabolic reprogramming determined for human lung cancer tissues in vivo [ 97 , 107 ]. However, the buildup of the glucose tracer differed between the mouse xenograft and human lung tumors, with the former accumulated but the latter depleted relative to the paired normal lung. These metabolic insights gained from the SIRM approach are valuable in delineating the fidelity of the mouse model in reporting human tumor metabolism.

In a separate SIRM study of transgenic tumor mouse models, we compared the metabolic reprogramming elicited by oncogenes c-MYC and MET in liver and lung using U- 13 C 6 -glucose and U- 13 C 5 , 15 N 2 -Gln as tracers coupled with 1D 1 H, 1 H{ 13 C}-HSQC, and GC–MS analyses [ 240 ]. We found that accelerated glycolysis is not always associated with tumorigenesis, and glucose and Gln metabolism depends not only on the oncogenic genotype but also on the tissue context. C-MYC -induced liver tumors exhibited increased glucose and Gln catabolism, while MET -induced liver tumors synthesized Gln from glucose. These reprogrammed events were associated with differential protein expression of Gln synthetase (Glu and glutaminase 1 (GLS1). In contrast, C-MYC -induced lung tumors had enhanced Glul and GLS1 expression, and accumulated Gln. We further showed that a GLS1 inhibitor killed cells that overexpressed c-MYC and catabolized Gln. Such understanding of oncogene- and tissue-dependent metabolic reprogramming is valuable in designing synergistic strategies for cancer therapy.

In two SIRM studies of human non-small cell lung cancer (NSCLC) in situ , we employed bolus infusion of U- 13 C 6 -glucose into NSCLC patients, followed by surgical resection of the cancerous (CA) and the surrounding non-cancerous (NC) tissues for extraction, and different 1D/2D NMR and MS analyses of the paired tissue extracts [ 97 , 107 ]. Based on the 13 C labeling patterns of various metabolites, We found that CA tissues were more active in metabolizing glucose into lactate, the Krebs cycle metabolites, glutathiones, one carbon metabolites, and nucleotide ribose/rings that their NC counterparts, which is consistent with the activation of glycolysis, the Krebs cycle, pentose phosphate pathway, glutathione biosynthesis, and nucleotide biosynthesis ( Fig. 18 ). In particular, we uncovered enhanced pyruvate carboxylase-mediated anaplerosis in 59 patients studied, while there was no evidence for the activation of glutaminase-mediated anaplerosis. We then demonstrated that PCB expression was crucially important for lung cancer cell growth both in vitro and in vivo in mouse xenografts [ 107 ]. These two studies not only illustrate the feasibility of probing human tumor metabolism in the native tumor microenvironment (TME), but also reveal novel reprogrammed metabolic events vitally important for NSCLC development.

We also explored the utility of ex vivo culturing of thin tissue slices (pioneered by O. Warburg [ 241 ]) in comparison with in situ human studies for reprogrammed lung cancer metabolism [ 107 ]. These freshly prepared paired CA and NC lung tissue slices have the advantage of maintaining the 3D architecture of the TME without systemic influences, while enabling flexible treatments with various tracers and/or physical/chemical stressors. In addition, since CA and NC tissues of the same patient are compared, interferences from genetic, physiological, and/or environmental factors that can contribute to metabolic distinction are eliminated. Fig. 19 illustrates how 1D 1 H{ 13 C}-HSQC analysis of paired CA and NC lung tissue extracts can be used to delineate metabolic reprogramming in ex vivo CA lung tissues that recapitulates in vivo metabolic reprogramming in lung cancer acquired from patient studies described above. We then employed this ex vivo system for investigating the metabolic perturbations elicited by an experimental LDHA inhibitor using U- 13 C 6 -glucose and U- 13 C 5 , 15 N 2 -Gln as tracers [ 101 ]. We confirmed the on-target action of this inhibitor, i.e. blocking lactate synthesis, while revealing the unexpected action of inhibiting glutaminolysis in human lung cancer tissues.

Human glioblastoma (GBM) metabolism was studied by employing continuous infusion of U- 13 C 6 -glucose into individual patients, followed by surgical resection of tumor tissues and 1D 13 C NMR analysis of the tissue extracts [ 242 ]. The authors found that GBM actively metabolized glucose via glycolysis, the Krebs cycle initiated by both canonical PDH and anaplerotic PC, de novo Gln biosynthesis, and glycine biosynthesis. They also found that <50% of the acetyl CoA pool (deduced from the 13 C positional isotopomer patterns of Glu) was derived from blood-borne U- 13 C 6 -glucose, which is unexpected based on the high demand of brain tissues for glucose. These results suggest that important source(s) other than glucose contribute to brain tumor bioenergetics. The study also demonstrates a practical means for probing the metabolic functions of human brain tumor grown in its native microenvironment.

5.2.2. In vivo spectroscopic measurements

A unique combined advantage of NMR lies in its non-invasive nature along with the capacity for structure elucidation, which makes it excellently suited for tracking metabolic transformations in real time. Such use of NMR can provide detailed kinetics of metabolic pathways using only a single sample, which eliminates sample-to-sample variation while greatly reducing the cost and labor involved. Most importantly, in vivo or ex vivo NMR enables metabolic reactions and their intracellular compartmentation to be probed directly in tissues or whole organisms, including human subjects, in a tissue-specific manner, which cannot be achieved by any other analytical approach. The more recent development of hyperpolarized NMR methods further enhances 13 C NMR’s utility in real-time kinetic measurement of specific metabolic reactions in vivo . The following examples illustrate the power of ex vivo and in vivo NMR applications in gaining metabolic insights. Non-hyperpolarized applications cell studies.

In vivo NMR studies have been performed on live cells ranging from bacteria, yeast, to mammalian cells to investigate in situ specific enzyme reaction kinetics [ 243 ], protein structures and conformations [ 244 ], or metabolic pathways [ 245 ]. In the early studies, microbial cell suspensions were commonly investigated under oxygen deficient conditions. Subsequently, a number of means for maintaining cells under more physiologically relevant conditions were explored for in vivo NMR applications [ 246 ]. For example, 1D 13 C NMR was employed to track the metabolism of 13 C 1 -3-fructose in the suspension of a new strain of lactic acid bacteria, where fructose catabolism to mannitol via fructose-6-phosphate and mannose-1-phosphate, to acetate via the heterofermentative pathway, and to lactate via glycolysis was delineated [ 245 ].

13 C 1 -1-glucose tracer and 1D 13 C NMR were employed to measure the in vivo reaction kinetics of aldolase and triosephosphate isomerase, which are key to 13 C incorporation into 13 C 2 -1,6-fructose-1,6-bisphosphate (FBP) by catalyzing the production and exchange between glyceraldehyde-3-phosphate (GAP) and DHAP, in suspensions of Escherichia coli [ 247 ] and yeast cells [ 248 ]. Upon reoxygenation of an anaerobic E. coli cell suspension, the time course appearance of 13 C-Glu at C2, 3, and 4 positions indicated that glucose carbon entered the Krebs cycle mainly through acetyl CoA [ 247 ].

Using a membrane-cyclone-reactor system, Hartbrich et al. tracked the pathway dynamics of Lys biosynthesis in Corynebacterium glutamicum from 13 C 1 -6-glucose via the time course buildup of 13 C labeled pathway metabolites such as lactate, Glu and succinate, in addition to the final Lys product. C. glutamicum is an important microbe engineered for industrial Lys production. The in vivo kinetic buildup of these metabolites provides key information for evaluating flux distribution through two different Lys synthetic pathways, the classical succinyl-diaminopimelate pathway and the diaminopimelate dehydrogenase bypath [ 249 ].

Tholozan et al. studied ethanol metabolism in a strict anaerobe Clostridium neopropionicum by measuring the time course buildup of 13 C-1-acetate, 13 C-2-propionate, 13 C-2-propanol, and 13 C-3-butyrate from 13 C 1 -1-ethanol tracer using 1D 13 C NMR [ 250 ]. They were able to propose a reaction sequence from ethanol to propionate and propanol via the acrylate pathway based on the in vivo NMR data, and subsequently confirmed this pathway by assaying the specific enzymes involved in the pathway in cell-free extracts.

The metabolism of L- 13 C 1 -3-lactate was investigated using 1 H NMR in superfused biofilm of Shewanella oneidensis [ 251 ]. The 13 C labeled to unlabeled ratio of L-lactate and the acetate product suggested preferential metabolism of D-lactate over L-lactate, which is consistent with the genomic data on the presence of the D-lactate dehydrogenase gene. However, no genes for L-lactate dehydrogenase, D-lactate transport system, or lactate racemase were found in the S. oneidensis genome, which raises the question as to how L-lactate is metabolized. This study demonstrates the feasibility of tracking metabolic dynamics of low abundance microbial biofilm in situ , whose metabolism can respond rapidly to environmental gradients, thus affecting microbial resistance to antibiotics. It would be impractical to capture the temporal and spatial dependence of metabolic changes in biofilm organisms using the in vitro extraction approach.

DeBerardinis et al. employed a two-stage 13 C tracer treatment with real-time 13 C NMR measurement to investigate the metabolic dynamics that support the fatty acid biosynthesis required for cell proliferation in live human glioblastoma cells grown in a porous microcarrier [ 129 ]. 13 C 2 -1,6-glucose was administered first, followed by 13 C 1 -3-Gl,n so that the time course kinetics of glucose and Gln metabolism via glycolysis, the Krebs cycle, glutaminolysis, and fatty acid biosynthesis were tracked. They found that glioblastoma cells actively synthesize fatty acyl chains of lipids from glucose carbons, and that Gln further enhances this activity. This Gln effect was ascribed to active glutaminase-initiated glutaminolysis in glioma cells that helped replenish the Krebs cycle intermediates and generate NADPH (via the malic enzyme pathway) required for fatty acid biosynthesis. They also concluded that the pyruvate carboxylase-initiated anaplerotic pathway was suppressed in glioblastoma cells, even though this pathway is highly active in the primary astrocytes, as described in Section

A colon cancer cell line (HCT-116) was grown in microcarrier beads with serial administration of 13 C 1 -1-glucose and 13 C 1 -6-glucose tracers to measure flux through glycolysis and the pentose phosphate pathway (PPP) using 1D 13 C NMR [ 252 ]. The authors found that close to 12% of the glucose supplied enters PPP, ca. 7% of the glucose utilized is recycled to glucose via PPP, and 4.7% of the glucose metabolized via PPP forms lactate. Such a high rate of recycling via PPP suggests a significant contribution of PPP to cellular NADPH production. Ex vivo tissue studies

A 13 C-bicarbonate tracer coupled with 1D 13 C NMR and 13 C saturation transfer experiments was employed to follow ex vivo bicarbonate uptake and carbonic anhydrase activity in excised maize root tips during K 2 SO 4 treatment [ 109 ]. K 2 SO 4 stimulates dark inorganic carbon fixation in plant tissues. K 2 SO 4 was found to have little effect on bicarbonate uptake, intracellular bicarbonate concentration, or carbonic anhydrase activity. The measured carbonic anhydrase activity was far greater than that of the phosphoenolpyruvate carboxylase (PEPC) (measured by 14 C-bicarbonate); the latter activity was enhanced by K 2 SO 4 . Together, these results suggested that in vivo PEPC activation by ion uptake is not mediated by substrate availability. This study illustrates the power of 13 C tracer-based NMR analysis in determining in vivo enzyme activity and substrate uptake/availability, which are impractical to acquire by any other means.

Probing metabolic interaction between symbiotic vesicular-arbuscular mycorrhizal (VAM) fungi and a plant host (leek roots) represents another valuable application of ex vivo NMR including 1D 13 C NMR and 31 P NMR [ 253 ]. Time course 13 C incorporation from 13 C 1 -1-glucose into trehalose, glycogen, and mannitol (VAM metabolites) and sucrose (host root metabolite) was acquired in this study. The presence of VAM greatly reduced sucrose synthesis in the host roots despite markedly enhanced glucose consumption. This fungal modulation of host metabolism was not related to improved uptake of phosphate (as evidenced by polyphosphate buildup). Likewise, host interaction with VAM curtailed mannitol production in VAM. These results point to a central role of glucose in modulating host–fungus relations.

We investigated the flood tolerance mechanism of rice coleoptiles by employing 13 C 1 -2-acetate tracer with 13 C NMR time course analysis [ 254 ]. Time-dependent increase in 13 C incorporation into different carbons of Glu, Gln, γ -aminobutyrate (GAB), malate, citrate, and glucose was evident in anaerobic rice coleoptiles. The kinetics of 13 C-4-Glu and 13 C-2-GAB are consistent with GAB production from Glu via Glu decarboxylase activity. The much higher synthesis rate of 13 C-3-malate than that of 13 C-2-malate points toward an active glyoxylate cycle in the glyoxysome, which produces malate via condensation of acetate with glyoxylate, leading to succinate. This glyoxysome-derived succinate can then be transported to the mitochondria to replenish the Krebs cycle intermediates diverted to GAB synthesis under anaerobiosis. In a separate 15 N-nitrate tracer study, we have shown that rice coleoptiles are active in GAB and Ala synthesis as a result of nitrate reduction and assimilation under anaerobiosis, as revealed by 1D 1H and 2D 1 H{ 15 N}-HSQC and HSQC–TOCSY analyses [ 81 ]. These reactions consume protons and NADH, which are produced in excess via glycolysis under anaerobic conditions. Together these results suggest that rice coleoptiles can better maintain glycolysis and the Krebs cycle for energy production via the pathways of GAB and Ala synthesis and the glyoxylate cycle, in addition to the canonical ethanol fermentation.

Burgess et al. employed different 13 C tracers and 1D WET- 1 H { 13 C}-HMQC analysis to investigate the substrate-dependence of Krebs cycle kinetics and associated exchange kinetics between Glu and αKG ex vivo in perfused mouse heart [ 255 ]. They found that the indirect method afforded 6- to 10-fold time savings over direct 13 C NMR analysis, enabling time course kinetics to be acquired at 20 s intervals. By fitting the NMR kinetic data to a tca-FLUX model, V tca and V x (flux through the Krebs cycle and exchange reactions, respectively) were obtained as a function of the 13 C tracers used ( 13 C 1 -2-acetate, 13 C 1 -3-lactate, and 13 C 4 -2,4,6,8-octanoate). The order of V tca was related to the capacity of the tracers for NADH production external to the Krebs cycle. In addition, the lack of substrate dependence of V x suggests that malate-Asp shuttle does not significantly contribute to V x via Asp aminotransferase, otherwise 13 C 1 -3-lactate administration would be expected to post the highest V x .

Two ex vivo tissue studies utilized 13 C-glucose and 1D 13 C NMR analysis of 13 C isotopomers of metabolites to demonstrate functional compartmentation of glycolysis and glycogenolysis in the cytoplasm [ 256 , 257 ]. One study on perfused hog carotid arteries employed sequential administration of 13 C 1 -1-glucose to prelabel tissue glycogen, followed by 13 C 1 -2-glucose to distinguish glycogenolysis from glycolysis based on the time course of the buildup of 13 C 1 -3-lactate (glycogenolysis) and of 13 C 1 -2-lactate (glycolysis) [ 256 ]. They also found that glycogen did not contribute appreciably to lactate production when mitochondrial activity was intact, but its contribution to lactate synthesis was greatly enhanced by KCN, which inhibited mitochondrial oxidative metabolism. These data are consistent with two separate pools of cytoplasmic glucose with different fates, one pool being derived from external glucose supply to serve primarily as the substrate for glycolysis, and the other pool from glycogen degradation to fuel mitochondrial oxidation. The other study on perfused rat heart substantiated these earlier findings by sequential administration of U- 13 C 6 -glucose to prelabel tissue glycogen, followed by 13 C 1 -1-glucose to observe both glycogen and glucose metabolism [ 257 ]. This labeling scheme enabled discrimination between the pyruvate pools derived from glucose and glycogen by 13 C isotopomer analysis of lactate, as well as between the glycolytic and glycogenolytic pyruvate pools that fuel acetyl CoA synthesis based on 13 C isotopomer analysis of Glu. The authors concluded that (1) cytoplasmic glycolysis and glycogenolysis are compartmentalized; (2) tissue glycogen carbons are preferentially oxidized via the Krebs cycle while exogenous glucose carbons are the primary source of glycolytic lactate release into the medium; and (3) insulin further polarizes the preference for glycogen as substrate for mitochondrial oxidation. Animal studies

Using 13 C 1 -3-Ala as tracer coupled with 1D 13 C and 31 P NMR analyses, energy status in relation to gluconeogenesis was investigated by Morikawa et al. in vivo in rat liver [ 258 ] They found 13 C incorporation into 13 C-2 and 13 C-3 (but not 13 C-4) of Glu/Gln under prolonged fasting, which indicates that entry of Ala-derived pyruvate into the Krebs cycle, is exclusively mediated by pyruvate carboxylase. They also showed 13 C incorporation into 13 C-1 to 6-glucose and 13 C-1-glycogen, which points to active gluconeogenesis from Ala in the rat liver. Both gluconeogenesis and glycogenesis were inhibited under ischemia when liver ATP level was depleted. During recovery from ischemia, ATP level correlated positively with labeled glucose and glycogen signals and negatively with labeled Ala signals. These data provided in vivo evidence for the importance of energetics in modulating Ala-mediated gluconeogenesis and glycogenesis in rat liver.

In an in vivo study of human brain tumor glioblastoma cell xenografts in an orthotopic mouse model, Marin-Valencia et al. employed continuous infusion of different 13 C tracers coupled with 1D 13 C NMR analysis to track glucose and Gln utilization [ 259 ]. They found that U- 13 C 6 -glucose was metabolized into 13 C 2 -2,3-Glu and that 13 C 2 -3,4-glucose tracer led to a higher production of 13 C 1 -1-Glu/Gln than of 13 C 1 -5-Glu/Gln; both are consistent with active PC-mediated anaplerosis in vivo in GBM tumor. This is in contrast to the GBM cell findings, where PC activity was found to be suppressed [ 129 ]. In addition, U- 13 C 5 -Gln was not metabolized in the Krebs cycle in vivo in GBM tumors [ 259 ], but was found to be the main anaplerotic substrate for the Krebs cycle in vitro in GBM cells [ 129 ]. Such sharp contrasts between in vivo and in vitro tumor cell metabolism strongly argue for the need for in vivo tracer-based NMR studies.

To gain better spectral resolution and more rigorous metabolite assignment, various 2D NMR techniques have been applied to metabolite analysis in vivo in live animals [ 260 ]. In particular, 2D 1 H{ 13 C}-ge-HMQC was employed to investigate 13 C 1 -1-glucose metabolism in cat brain in vivo [ 261 ]. Signals from 13 C-4-Glu and 13 C-4-Gln, along with 13 C-3-Glu/Gln and 13 C-2-Glu/Gln (detected due to better water suppression), were resolved in the 2D spectrum (cf. Fig. 20 ). These 13 C metabolite isotopomers are important in delineating the pathways involved in Glu/Gln cycling, as described above.

An external file that holds a picture, illustration, etc.
Object name is nihms763208f20.jpg

In vivo 2D 1 H{ 13 C}ge-HMQC analysis of cat brain. The 13 C-edited 2D 1 H spectrum acquired in vivo from a cat brain resolves resonances of 13 C-4-Glu (4a) from 13 C-4-Gln (4b), which cannot be achieved with 1D 1 H methods. Also resolved are resonances of 13 C-3-Glu/Gln (5), 13 C-2-Glu/Gln (3), 13 C-1-α-glucose, and 13 C-1-β-glucose; the latter three would be difficult to detect without the better water suppression afforded by the 2D method [ 260 ]. Excerpted from MAGMA, 17 (2004) 317–338, In vivo 2D magnetic resonance spectroscopy of small animals, P. Meric, G. Autret, B.T. Doan, B. Gillet, C. Sebrie, J.C. Beloeil, Fig. 14a with kind permission from Springer Science and Business Media. Human studies

Perhaps one of the most powerful applications of NMR in biomedicine lies in in vivo human studies, which can provide invaluable novel metabolic insights that cannot be acquired by any other means. In vivo NMR studies of human subjects have been mainly focused on the brain [ 125 , 262 ], since human brain tissues are excellently suited for MRI and localized NMR spectroscopy. Presumably the difficulty of brain tissue procurement and the rapid distortion of its metabolism upon resection also add to this focus.

The first 13 C NMR spectra of human brain were acquired by Beckmann et al. [ 263 ] using surface coils on volunteers infused with 13 C 1 -1-glucose. The 13 C incorporation into Glu and Gln at C2, C3, and C4 positions as well as into lactate was observed in real time. Subsequent studies utilizing different 13 C tracers including 13 C 1 -1 or 2-glucose, 13 C 1 -3-lactate, or 13 C 1 -2-acetate enabled fluxes through the neuronal Krebs cycle ( V TCAn ) and the Glu/Gln cycle ( V cyc ) between neurons and astrocytes to be assessed in vivo in human brain ( Fig. 21 ). An approximately 1:1 ratio of V TCAn versus V cyc was evident by compiling data from 11 studies, where these two parameters were deduced under varying brain electrical activity [ 262 ]. These results suggest a close link between brain energy metabolism via the Krebs cycle and neurotransmitter production and signaling.

An external file that holds a picture, illustration, etc.
Object name is nihms763208f21.jpg

In vivo 1D 13 C NMR analysis of human brain metabolism. The 13 C spectra in (A) were acquired in vivo from the midline occipital/parietal lobe of a human volunteer infused with 13 C 1 -1-glucose, 13 C 1 -3-lactate, or 13 C 1 -2-acetate. The differences in 13 C incorporation from these tracer substrates into Glu and Gln are evident. With glucose and lactate as precursors, a major fraction of 13 C appeared in Glu C4, which is consistent with primary location of glucose and lactate metabolism in brain neurons. In contrast, the acetate tracer led to much higher 13 C enrichment in Gln C4, which supports acetate metabolism via the Krebs cycle in brain astrocytes. Based on the time course analysis of in vivo 13 C NMR data such as those in (A), the rate of neuronal Krebs cycle ( V TCAn ) and Glu/Gln cycle ( V cyc ) can be estimated as a function of brain electrical activity. Panel (B) plots V TCAn versus V cyc from 11 studies and shows a 1:1 linear relationship. These results indicate that the Glu/Gln cycle for the Gln neurotransmitter production constitutes a major metabolic flux for brain energy metabolism and that >80% of neuronal oxidative ATP production is coupled to neuronal signaling [ 262 ]. Reproduced from Figs. 2 and 3 of D.L. Rothman, H.M. De Feyter, R.A. de Graaf, G.F. Mason, K.L. Behar, 13 C MRS studies of neuroenergetics and neurotransmitter cycling in humans, NMR in biomedicine, 24 (2011) 943–957 with permission from John Wiley & Sons Inc.

Extensive clinical studies have been performed using 13 C enriched substrates including 13 C-glucose and acetate to study brain disorders (e.g. Alzheimer’s, Canavan’s encephalopathy, and epilepsy) as well as normal brain development in human subjects [ 264 ]. The slow rate of N-acetylaspartate synthesis from 13 C 1 -1-glucose was directly observed for the first time in human brain, and it fell further in patients with Canavan’s disease. The same tracer was administered to Alzheimer’s patients to uncover failure in Gln–Glu cycling and Glu neurotransmission. The beneficial effect of a ketogenic diet on reducing seizures in epileptic children was found to be related to improved Glu-Gln cycling based on enhanced 13 C 1 -1-acetate metabolism into Glu and Gln. The infusion of 13 C 1 -2-glucose tracer into healthy subjects enabled measurement of pyruvate carboxylase activity via 13 C incorporation into C2 and C3 of glutamine and glutamate [ 265 ]. Metabolic modeling of the labeling data indicated that pyruvate carboxylase accounted for 6 ± 4% of the rate of glutamine synthesis. Most of the pyruvate carboxylase flux appears to replace glutamate lost due to glial oxidation, and therefore can be considered to support neurotransmitter trafficking.

Continuous infusion of tracer into human subjects [ 266 , 267 , 264 ] has been commonly employed in in vivo human studies, and requires a more elaborate protocol. An alternative is to administer the tracer orally, which at least in brain studies yields similar results to intravenous administration, but with lower precision [ 267 ]. Hyperpolarized (HP) applications principles of dynamic nuclear hyperpolarization (dnp).

Hyperpolarization-based NMR analysis offers very high sensitivity enhancement that has been demonstrated in recent years for imaging in vivo dynamics of metabolic reactions in specific regions of tissues [ 268 – 290 ]. The fundamental principle of hyperpolarization is to transfer magnetization from a very highly polarized reservoir to the observed nucleus. One method, called dynamic nuclear polarization (DNP), uses the very high polarization achievable by unpaired electrons in a magnetic field, which is then pumped using a microwave field to the metabolites bearing one or more 13 C or 15 N atoms [ 291 ]. Under this mechanism, the maximal polarization enhancement is given by ( γ e / γ N ), which is about 660 for protons, 2640 for 13 C and 6600 for 15 N.

The cross-relaxation process for DNP is usually carried out near 1 K, which also provides a large enhancement of polarization via the Boltzmann distribution, i.e. about 300 fold for 1 K versus 300 K. For biological studies, the frozen polarized sample must be warmed rapidly (e.g. to 300 K) for measurements. Since the fractional polarization for 13 C at 300 K at 14.1 T is ca. 1.2E–5, and 0.00359 at 1 K, considerable polarization enhancements (>10,000) can only be achieved immediately after thawing [ 292 ], in order to benefit from the contribution of electron-mediated cross polarization and the enhanced polarization due to the Boltzmann distribution. Once at higher temperature, with magnetization no longer being pumped, the system will relax toward its equilibrium polarization; the rate at which this process occurs is determined by the T 1 value. Nuclei with long T 1 values are therefore desirable for prolonging the hyperpolarized state. For example, if the warming and thawing process takes 10 s, then about 30% of the initial polarization will be lost for 13 C with a T 1 value of 30 s. Thus, the time period for NMR observations is typically limited to ca. 4 T 1 . In addition, as this procedure creates an unstable non-equilibriumstate, once the spin has been excited, the polarization is lost, again limiting the number of NMR transients that can be acquired from the same sample while calling for a small flip angle for the excitation pulse. The magnetization available for each pulse with a flip angle of theta is approximately:

where M 0 is the magnetization just before the first excitation pulse, and θ is the flip angle.

The lifetime of the enhanced polarization must be comparable to the half-life of the reaction that is being measured. For the simple reaction A ⇔ B, the time-dependence of the production of B is

where k 1 and k −1 are the rate constants of the forward and reverse reaction, respectively. As the magnetization of the source is decaying exponentially, the observed magnetization on B following a bolus input of hyperpolarized A is described by the difference between two exponentials, which goes through a maximum, according to the Bloch–McConnell equations below [ 92 ].

where ρ A , ρ B are the spin lattice relaxation rate constants and 〈 M A 〉, 〈 M B 〉 are the equilibrium magnetizations of A and B. Fig. 22 shows the magnetization time course calculated for such a two-site exchange reaction with an initial relative polarization of A of 10,000 and an equilibrium constant of unity. The magnetization of A decays biexponentially due to the spin-lattice relaxation and the chemical exchange of A with unpolarized B, whereas that of B initially increases from its (low) equilibrium magnetization as the highly polarized A exchanges with B, followed by a decay as the polarization dissipates via longitudinal relaxation. The T 1 value thus sets the timescale of the observation; reactions that do not occur on this timescale are not observed.

An external file that holds a picture, illustration, etc.
Object name is nihms763208f22.jpg

Two site exchange. Magnetization time courses were calculated for two-site exchange A ⇔ B with k 1 = k −1 = 0.05 s −1 and T 1 values for A and B of 20 s. Blue squares = M A ; red circles = M B . The equilibrium and initial magnetizations of A were set to unit and 10,000, respectively. The modified Bloch equations (3A) and (3B) were solved as described in the text, and the curves were generated using Kaleidagraph (Synergy Software). The solutions to this system are: M B ( t ) = k 1 〈 M A 〉[exp( λ 1 t ) − exp ( λ 2 t )]/( λ 1 − λ 2 ). M A ( t ) = [ ( ε - λ 2 ( M A 0 - 〈 M A 〉 ) ) exp ( λ 1 t ) + ( λ 1 ( M A 0 - 〈 M A 〉 - ε ) exp ( λ 2 t ) ] / ( λ 1 - λ 2 ) . 2 λ 1,2 = − ρ A + ρ B + k 1 + k −1 ± [( ρ A + ρ B + k 1 + k −1 ) 2 − 4( ρ A ρ B + ρ A k 1 + ρ B k −1 )]. ε = - ( ρ A + k 1 ) M A 0 + k - 1 〈 M B 〉 + ρ A 〈 M 〉 . where M A 0 is the initial (hyperpolarized) magnetization of A.

The requirement for a long T 1 value limits the nature of the atoms that can be usefully probed. Generally, carbons such as carbonyl and quaternary carbons, that do not have an attached proton, have sufficiently long T 1 values (20–40 s). It is also possible to use perdeuterated compounds such as perdeuterated glucose (cf. Fig. 23 ) to slow down the longitudinal relaxation process [ 287 ].

An external file that holds a picture, illustration, etc.
Object name is nihms763208f23.jpg

Central Metabolism probed by hyperpolarized 13 C enriched precursors. Metabolites in red have been 13 C hyperpolarized at one or more atoms, and those in italics have been observed as products of hyperpolarized substrates via different pathways. These include glucose (Glc) oxidation to pyruvate (Pyr) via glycolysis, subsequent pyruvate metabolism via the Krebs cycle and lactic fermentation, anaplerosis via pyruvate carboxylation to oxaloacetate (OAA), and glucose metabolism via the pentose phosphate pathway from glucose-6-phosphate (G6P), 6-phosphogluconate (6PG) to ribulose-5-phosphate (Ru5P). Also shown are anaplerosis via glutaminolysis or Gln conversion to Glu and then to α-ketoglutarate (αKG) for entry into the Krebs cycle and ethanol oxidation to acetate via alcohol dehydrogenase (ADH) and aldehyde dehydrogenase (ALDH). DHAP: dihydroxyacetone phosphate; GAP: glyceraldehyde-3-phosphate; 3PGA: 3-phosphoglycerate; PEP: phosphoenolpyruvate; Lac: lactate, Succ succinate; DES: diethylsuccinate; fum: fumarate; mal: malate; HK: hexokinase; LDH: lactate dehydrogenase; ALT: alanine pyruvate aminotransferase; PDH: pyruvate dehydrogenase; SDH: succinate dehydrogenase; FH: fumarate hydratase; MDH: malate dehydrogenase; AST: aspartate/oxalacetate aminotransferase; PC: pyruvate carboxylase; GLS glutaminase.

As the measurement time window is short, the polarized precursor must be quickly introduced to the sample, for example by direct intravenous injection into an animal or human, or by direct addition to a cell growth medium. The kinetics of the metabolic process being observed then encompass the uptake (transport) rate of the substrate and the rate of its subsequent transformations via enzyme-catalyzed reactions. As the T 1 value is typically in the order 20–30 s [ 92 ], the time scale of the reactions that can be monitored is in the order of 2 min. This is comparatively long compared with the reciprocal turnover numbers of enzymes involved in central catabolism (in the range 10–1000 s −1 ). However, one also has to consider the sequential nature of many metabolic reactions, the concentrations of the enzymes, their fractional saturation, and the sizes of the metabolite pools that are being observed. For example, glycolytic enzymes are usually very abundant, with typical concentrations of 1–10 μM [ 293 ], and can approach 100 μM in skeletal muscle [ 294 , 295 ]. These concentrations are often comparable to substrate concentrations and K m values [ 296 ]. In fact, for central metabolism, substrate concentrations are typically substantially lower than their K m values [ 296 ].

Thus for a single human cell with cytoplasmic volume of 1 pL and enzyme concentrations of 1 μM, there is about 1 attomole of enzyme molecules per cell. An enzyme with a turnover number ( k cat ) of 100 s −1 , at 10% occupancy, would produce 10 attomoles of product s −1 per cell. On a high field NMR spectrometer with a probe optimized for 13 C detection, a SNR of 100:1 may be achieved in a single acquisition on ca. 1 μmol 13 C by conventional analysis, or ca. 100 pmol by DNP with an enhancement of 10,000. To produce 25 pmol in 60 s for hyperpolarized NMR detection would then need a minimum of 4 × 10 4 cells in the sensing region. In practice, many of the reactions under consideration are sequential, so that the intermediates build up to a much lower extent. Furthermore, a small flip angle must be used for kinetic experiments such that ca. 10% or less of the sample magnetization is sampled with each pulse, while the signal is decaying exponentially. By 1 min, only 13% of the initial signal is left and after 2 min, <2% is left, for T 1 = 30 s. These considerations, along with the limited availability of suitable hyperpolarizable substrates (see below), limit the depth of coverage for metabolic pathways in DNP-based experiments. Principle of Para-Hydrogen-Induced Hyperpolarization (PHIP)

An alternative to the DNP approach is to use para-hydrogen (the singlet state of H 2 which is present at 25% at thermal equilibrium) to induce hyperpolarization (PHIP). Here polarized para-hydrogen (enriched above its equilibrium value) is rapidly added across a double bond of a substrate (which may be deuterated to reduce T 1 relaxation) next to a 13 C carbonyl. This process usually requires a catalyst and the polarization is transferred to the target nucleus via scalar interactions [ 297 – 302 ]. Although the variety of substrates that can be polarized using PHIP is limited, the process is much faster than solid state DNP [ 303 ]. Polarization enhancements of >3000 have been achieved in a limited number of metabolic substrates [ 303 – 306 ]. The hyperpolarized substrate is then injected into the recipient organism or cell culture. An example of this is the reduction of 13 C-1-diethylfumarate to 13 C-1-diethylsuccinate with para-hydrogen, which is rapidly taken up by cells, and converted to 13 C-1-succinate by internal esterases [ 303 ]. The kinetics of the formation of downstream Krebs cycle metabolites 13 C-1-fumarate, -malate, and -OAA (via the surrogate aspartate) can be followed in vivo .

When coupled with magnetic resonance imaging, hyperpolarization techniques open an avenue for probing metabolic reactions dynamics in vivo in specific regions of tissues that is unparalleled by other techniques. This capability is uniquely suited to whole animal and human metabolic studies. Applications. Redox ratio determination

Many of the early studies used 13 C-1 or 13 C-2 pyruvate as substrate, which have relatively long T 1 values and produce significant polarization enhancement via DNP [ 276 , 278 , 306 , 307 ]. The enhanced conversion of pyruvate to lactate is generally considered to be important for tumor development [ 308 – 311 ]. This reaction is catalyzed by one of the lactate dehydrogenase (LDH) isoenzymes and can be analyzed according to the simple two-site exchange model (cf Eq. (3) ), where the pseudo first order rate constants are related to the enzyme concentration, the enzyme kinetic parameters, and the overall reaction equilibrium constant ( K ) as described by:

The intracellular free NAD + /NADH ratio has been estimated using hyperpolarized [ 13 C 6 , 2 H 12 ]-glucose in two cancer cell lines. This was achieved by measuring the free concentrations of pyruvate and lactate by DNP- 13 C NMR so that the NAD + /NADH ratio can be calculated from the independently measured intracellular pH and the assumed equilibrium constant. Christensen et al. reported ratios of 2550 ± 350 in prostate cancer PC3 cells and 738 ± 92 inMCF-7 breast cancer cells grown under normoxic condition, and the intracellular pH was estimated to be 7 to 7.1 from the bicarbonate/CO 2 (aqueous) ratio by NMR [ 287 ], although this value is typically higher than 7.2 in cancer cells [ 312 , 313 ]. The equilibrium constant was assumed to be 1.11 × 10 −11 M, though other estimates deviate from this value [ 314 ]. Nonetheless, this ratio is in line with other estimates of the order 1000:1 and is far higher than the ratio of total NAD + to NADH (i.e. bound + free species), which is typically <10:1 in mammalian cells [ 315 ]. Since the total concentration of dehydrogenases in cells is comparable to the total NAD + + NADH content, NAD + and NADH are largely in the bound state, and have differential affinity for different dehydrogenases. Thus, changes in the free NAD + /NADH ratio can have important influence on redox-sensitive metabolic reactions, including such central pathways as glycolysis and the Krebs cycle. Similar remarks apply to the less abundant NADP + and NADPH redox pair, where NADPH is generally thought to be much higher than NADP + , reflecting their functions in supporting anabolic reactions [ 316 ]. However, recent in vitro studies have reported ratios closer to unity or lower [ 317 , 318 ]. As NADPH is unstable and highly sensitive to extraction artifacts, the in vivo hyperpolarization NMR approach would be valuable in resolving this discrepancy. Again, accurate knowledge of the free to bound ratio of NADP + /NADPH, as well as their absolute concentrations, is crucially important in understanding the regulation of anabolic biochemical networks and oxidative stress.

Kinetic isotope effects

Using perdeuterated substrates such as [U- 13 C 6 , 2 H 12 ]-glucose for probing glycolytic dynamics, there is potentially a primary kinetic isotope effect (KIE) on individual reactions if they involve a C-H bond cleavage which contributes to the overall reaction rate. The primary KIE for C-H versus C-D bond breaking step may be of the order 7 or higher if tunneling is active [ 319 , 320 ]. However, the KIE on net pathway flux may be small as the net flux from glucose to end products such as pyruvate and lactate in glycolysis depends on the kinetic properties of all of the enzymes in the pathway, so that the control coefficient for any given reaction is small [ 321 – 323 ]. Nevertheless, significant KIEs on parts of metabolic pathways such as the oxidative branch of the pentose phosphate pathway have been observed under steady state conditions [ 322 ], which implies that under transient conditions (as with bolus hyperpolarization), such effects may become significant.

Pathway flux analysis

The dynamics of metabolic reactions measured by in vivo NMR are influenced by the rates of a number of processes. For example, the conversion of pyruvate to lactate observed by DNP-NMR is governed by the rate of pyruvate uptake via the monocarboxylate transporter, the exchange reaction between pyruvate and lactate, and as many other reactions that utilized pyruvate. A number of recent studies on cells and tissues of models animal and human subjects indicate that the observed time course conversion of 13 C-1-pyruvate or 13 C-2-pyruvate to 13 C-lactate largely reflects an exchange flux, which at the time immediately after the bolus injection is dominated by the reductive rate (i.e. lactate formation) [ 92 , 278 , 271 , 304 , 324 – 328 ].

The kinetics of other routes of pyruvate utilization have also been measured, including rapid transamination to alanine catalyzed by the alanine-α-ketoglutarate transaminases (ALT) [ 162 , 329 ], transport into the mitochondria and oxidative decarboxylation to acetyl CoA and CO 2 via pyruvate dehydrogenase, followed by downstream Krebs cycle reactions [ 160 , 161 , 270 , 330 , 331 ], ATP-dependent pyruvate carboxylation to OAA and its subsequent reactions in the Krebs cycle, nucleotide biosynthesis, the aspartate/malate shuttle, and as the entry point to gluconeogenesis in some tissues [ 68 , 332 ].

In recent years a wider variety of hyperpolarized substrates have become available, that can probe different parts of central metabolism [ 92 , 272 , 283 , 333 , 281 , 334 ], as summarized in Fig. 23 . The kinetics of conversion of deuterated 13 C glucose by glycolysis into pyruvate and then to lactate or into PPP intermediates, were obtained in situ in cancer cells and tumor tissues [ 277 , 287 ]. Meier et al. [ 335 ] observed real-time metabolic transformation of hyperpolarized glucose in E. coli into DHAP, 3-3PGA, pyruvate, lactate, ethanol, alanine, acetate and acetyl CoA, as well as the decarboxylation products CO 2 (aqueous) and HCO 3 - , over a time period of 75 s. Indeed most of the metabolic reactions occurred very rapidly in this system (half life of formation of 10 s or less). Similarly, glucose metabolism was followed in cultured breast cancer cells, in which DHAP, 3PGA and lactate were directly observed, and by saturation transfer other glycolytic intermediates including G6P, F6P, phosphoenolpyruvate (PEP), and pyruvate could be detected [ 336 ]. To probe glycolysis and gluconeogenesis in mouse liver, Moreno et al. [ 337 ] introduced hyperpolarized 13 C-2-DHAP, which was rapidly metabolized to 13 C 2 -2,5-glucose, 13 C-2-glycerol-3-phosphate (G3P), 13 C-2-PEP, 13 C-2-pyruvate, 13 C-2-Ala, and 13 C-2-lactate, showing the central role of DHAP in both glycolysis and gluconeogenesis, as well as the gateway to phospholipid synthesis via the precursor G3P (cf. Fig. 23 ). However, 13 C carbonyl resonance assignments of metabolites based only on chemical shift and biochemical intuition could be problematic. Recently it was shown that a particular signal was assigned to 6-phosphogluconate (6PG) in tumors, but 3PGA in yeast and potentially in other systems [ 338 ]. The biochemical interpretation of these assignments is distinct, as 3PGA relates to glycolytic flux while 6PG reflects flux through the oxidative branch of the pentose phosphate pathway.

Besides glucose-derived pyruvate as a carbon source, the Krebs cycle has multiple other carbon inputs, such as beta-oxidation and anaplerotic glutaminolysis ( Fig. 23 ). Glutaminase activity was measured by the conversion of hyperpolarized 13 C 5 -Gln into Glu in HepG2 liver cancer cells [ 339 ] and in prostate cancer cell lines [ 286 ]. Transformation of hyperpolarized 13 C-1-glutamate into α-ketoglutarate was also observed in HepG2 cells via the activity of ALT [ 279 ]. This glutaminase pathway is thought to be a major contributor to survival and proliferation in reprogrammed tumor metabolism [ 340 , 341 ]. Further reactions down the Krebs cycle have been probed using the hyperpolarized diethyl ester of 13 C-1-deuterosuccinate (readily taken up by cells), which was then converted into succinate, malate, fumarate, and aspartate (cf. Fig. 23 ).

Using hyperpolarized 13 C 1 -α-ketoisocaproate as a tracer substrate, the branched chain amino acid Leu metabolism via branched chain amino acid transferase (BCAT) was probed in murine lymphoma (EL4) versus rat mammary adenocarcinoma (R3230AC) tumors. BCAT is a putative marker of metastasis and a target of the proto-oncogene c -MYC . Its activity was monitored by the conversion of 13 C 1 -α-ketoisocaproate to Leu. EL4 but not R3230A tumors show a more than 7-fold higher hyperpolarized 13 C-leucine signal relative to the surrounding healthy tissue. The distinct molecular signatures of branched chain amino acid metabolism in EL4 and R3230AC tumors correlate well with ex vivo assays of BCAT activity [ 342 ].

It has been proposed that precursors could be multiplexed, such as 13 C-1-pyruvate, 13 C-sodium bicarbonate, 13 C 2 -1,4-fumarate, and 13 C-urea to measure multiple enzyme activities in vivo [ 274 ]. A recent MRI study demonstrates the feasibility of co-hyperpolarizing three 13 C labeled agents ( 13 C-urea, 13 C-1-hydroxymethylcyclopropane (HMCP), and 13 C-2- t -butanol) for obtaining rapid dynamic images of their distribution in mouse tissues using multiband frequency encoding coupled with the steadystate free precession (SSFP) method [ 343 ]. The dynamic in vivo 13 C MRI data enabled tissue perfusion and permeability to be quantified. As shown in Fig. 24 , t -butanol crossed the blood-brain barrier at a remarkable speed, unlike urea and HMCP. Prostate tumor tissues exhibited a higher permeability for urea and HMCP than normal brain and liver tissues. This capability of HP MRI opens up the possibility of in vivo measurement of drug delivery efficiency to target tissues.

An external file that holds a picture, illustration, etc.
Object name is nihms763208f24.jpg

Magnetic resonance imaging of three co-hyperpolarized 13 C agents in mouse brain. 13 C-Urea, 13 C-1-hydroxymethylcyclopropane (HMCP), and 13 C-2- t -butanol were simultaneously hyperpolarized before infusing into normal and transgenic TRAMP mice bearing prostate cancer. A multiband frequency encoding and balanced steady-state free precession (SSFP) MRI method was used to acquire 13 C images of the dynamic distributions of these three agents in mouse tissues. The axial tripolarized images of mouse brain (color) are overlaid on the T 2 images (gray), which show the remarkable speed at which t -butanol crosses the blood–brain barrier, unlike urea and HMCP A parallel set of dynamic images of prostate tumor tissues compared with those of normal tissues enabled measurement of tissue permeability for the three agents. Tumor tissues were shown to have a higher permeability to urea and HMCP than normal brain and liver. Such a capability can be readily extended to investigating the efficiency of therapeutic drug delivery to target tumor tissues [ 343 ]. Reproduced from Fig. 5 in C. von Morze, R.A. Bok, G.D. Reed, J.H. Ardenkjaer-Larsen, J. Kurhanewicz, D.B. Vigneron, Simultaneous multiagent hyperpolarized (13)C perfusion imaging, Magn Reson Med, 72 (2014) 1599–1609. With permission from John Wiley & Sons, Inc.

Cardiac metabolism using hyperpolarized substrates has also been an active area of study, including the use of hyperpolarized 13 C 1 -2-pyruvate to interrogate the role of acetyl carnitine as a buffer for acetyl CoA supply in ex vivo rat heart [ 344 ]. 13 C 1 -2- and/or 13 C 1 -1-pyruvate were employed to uncover a switch from predominant fatty acid oxidation to reliance on glucose oxidation in spontaneously hypertensive rat heart, while hyperthyroidism-induced hypertrophic rat heart had reduced glucose oxidation, which was mediated by pyruvate dehydrogenase kinase activity. This capability of 13 C 1 -1-pyruvate to discriminate modes of cardiac dysregulations in different causes of cardiac hypertrophy is unparalleled [ 345 ].

6. Integration of metabolic network information with knowledge gained from other ‘omics technologies

It is clear that the variety of NMR techniques coupled with the stable isotope tracer approach described above affords powerful means for exploring metabolic network dynamics in organisms and their perturbations in response to stressors. In particular, the ability of NMR to track metabolic changes noninvasively or in situ provides unmatched means for determining the dynamics and compartmentation of metabolic pathways directly in living organisms, including human subjects. Such knowledge has been sparse, but is required for our understanding of metabolism at the systems level. It is also important to realize that NMR alone cannot resolve the vastly complex and intersecting relationships of metabolic networks, and that mass spectrometry represents another powerful molecular structure-based platform that can both complement and confirm NMR approaches in deciphering metabolic network puzzles. This aspect has been discussed elsewhere and will not be elaborated here [ 11 , 88 , 346 ].

It should be further emphasized that the combination of stable isotope tracers with NMR and MS analytical power is capable of resolving metabolic reprogramming at specific enzyme steps such as the pyruvate carboxylase reaction, even directly in human subjects (cf. Fig. 18 ) [ 97 , 107 ]. The NMR- and MS-acquired stable isotope labeling patterns (positional isotopomers and mass isotopologues) of metabolites are also excellent constraint parameters formetabolic networkmodeling to globally deconvolute individual enzyme kinetics [ 180 , 347 – 358 ]. As illustrated in Fig. 17A , such information on perturbations of metabolic network dynamics is invaluable for directing functional analysis of other ‘omics data, including genomics, epigenomics, transcriptomics, and proteomics. For example, we learned from a SIRM study that selenite blocks glycolysis and the Krebs cycle in lung cancer A549 cells. Inquiring into the transcriptomic data acquired in parallel, we discovered that these perturbations could result from downregulation of multiple enzyme genes in glycolysis and the Krebs cycle ( Fig. 17 ) [ 108 ]. Likewise, the buildup of lipid catabolites (phosphorylcholine, glycerolphosphorylcholine, and inositols) points to enhanced lipid degradation ( Fig. 17B ). Again, transcriptomic data on phospholipases and lyosphospholipases substantiate the SIRM observation that lipid catabolism is enhanced via upregulation of lipases.

We also discovered from the transcriptomic data down regulation of the key enzyme, acetyl CoA carboxylase (ACC), in the fatty acid biosynthesis, and up regulation of its negative regulator AMPK (AMP-dependent kinase). These gene expression changes, together with heightened AMP/ATP ratio (known to activate AMPK), suggest reduced fatty acid biosynthesis in selenite-treated A549 cells. Subsequent 2D 1 H{ 13 C}-HSQC analysis of 13 C labeling patterns of A549 lipids showed reduced synthesis of fatty acyl chains, but not glycerol backbone synthesis in response to selenite treatment ( Fig. 17B ). These results in combination with the transcriptomic evidence show that the selenite attenuation of lipid biosynthesis is mediated via the AMPK-ACC axis, but not via glycerol-3-phosphate synthesis or incorporation into lipids. Protein expression analysis using Western blot confirmed inhibition of ACC activity via ACC phosphorylation by AMPK ( Fig. 17C ). However, the activation of AMPK appeared to be more complex than enhanced gene expression of AMPK. We found that AMPK phosphorylation was attenuated by selenite, which points towards AMPK activation via suppressing its inhibitor kinase LKB1 activity [ 359 ]. Thus, metabolomics-edited transcriptomic analysis (META) not only generates testable hypotheses on metabolic dysregulations at the transcript level, but together with targeted protein analysis also uncovers signaling cascade(s) that modulate the observed metabolic reprogramming.

Although performed manually, the above case study exemplifies the synergistic value of integrating SIRM with other ‘omics approaches in achieving systems biochemical understanding of biological perturbations. The study also illustrates that optimal gain is attained by bidirectional and cross-correlational integration of ‘omics data. Future bioinformatics development on such integration will present an unprecedented opportunity for advancing systems biochemistry in organisms of interest, whether related to pathogenesis, environmental stresses or organismal interactions.

7. Conclusions and future directions

NMR has a number of unparalleled advantages for metabolic studies, most notably the rigor and versatility in structure elucidation, and in analysis of positional isotopomer distribution, afforded by multinuclear and multidimensional experiments, as well as the capacity for in vivo spatially selective analysis of tissues in living organisms. There are also limitations on NMR in metabolic studies, including (1) its relatively low sensitivity [ 360 ] (ca. 1 nmol without hyperpolarization or in the 10 pmol range with hyperpolarization for kinetic studies), compared with LC-MS, where femtomole detection is now routine; and (2) limited resolution in terms of the number of metabolites relative to high-resolution MS. Although fmol detection is routine in LC-MS, the amount of material needed to achieve this detection limit is much larger, as only a small fraction of the LC eluate can be introduced into the mass spectrometer. Although MS-based profiling is now more popular in metabolomic studies, NMR has been and will continue to be a key player in stable isotope-resolved metabolomic studies. It synergizes with MS in such applications by providing complementary and indispensible isotopomer distribution information, as illustrated in this review. With resurgent interest in probing metabolism in vivo , particularly in human subjects, NMR’s ability to track metabolic dynamics in situ in living cells and tissues is unmatched by any other technique. For these reasons, we expect to see future expansion of NMR applications in metabolic research. Further development in NMR techniques to improve detection limits and resolution is therefore highly desirable to facilitate this expansion. Below, we list a few areas of development that are under way.

There are several approaches to improving sensitivity. The most expensive option is to develop ultrahigh magnetic field strength, of the order 23–28 T [ 361 , 362 ], which may benefit metabolic NMR more than macromolecular NMR. However the performance of cryogenic probes and field stability at such high magnetic field strength is presently problematic, and the sensitivity enhancement may not approach the expected theoretical field dependence (see above). Cryogenic probes can provide a substantial sensitivity boost, typically 4-fold over a comparable RT probe for non-lossy samples. This is significantly compromised in lossy (ionic) samples, a problem which is exacerbated at higher magnetic field strength. Using smaller diameter (e.g. 1–1.7 mm) probes will reduce such losses, while improving intrinsic performance, particularly for mass-limited samples [ 76 , 363 ]. Shaped tubes have also been developed to reduce the influence of the E-field in salty samples [ 364 ].

To greatly expand 1D NMR resolution, and to gain much more structural insights, further development in multidimensional NMR is indispensible, particularly in terms of the speed of data acquisition. There have been multiple developments that improve this aspect, including non-uniform sampling [ 154 , 365 ] and projection–reconstruction [ 366 , 367 ]. There are also hardware approaches, including multiple receivers enabling 2 or more 2D experiments to be acquired simultaneously [ 368 , 369 ], as well as one-shot 2/3D methods which acquire different “increments” at different slices of the sample [ 370 – 376 ]. However, the latter put greater demands on sensitivity [ 377 ].

The number of increments needed for nD NMR can also be reduced using selective excitation methods, or pre-choosing which metabolites present in a mixture are to be sampled, using the Hadamard approach [ 366 ] for targeted analyses. The latter approach will be limited for metabolic profiling and/or SIRM studies, where the goal is to obtain as much coverage as possible, including isotopomer distributions.

Slow relaxation or long T 1 is a significant limitation on the speed of NMR acquisition. Very fast pulsing results in compromised sensitivity and highly distorted intensities, requiring large corrections for quantitative analysis. The SOFAST HMQC experiment that works well for 15 N-enriched proteins [ 378 ] is inappropriate for metabolite analysis, for lack of fast relaxing proton reservoirs and due to the need for sampling the entire spectral region. Fast HMQC using Hartmann-Hahn polarization sharing [ 379 ] has been described, but may be of less value for 13 C isotopomer analysis. A third alternative is to reduce the T 1 of protons by paramagnetic relaxation enhancement, allowing shorter recycle times [ 380 ]. A combination of these solutions is most likely to improve the throughput of 2D NMR methods to make them more attractive for routine NMR analysis, thereby attaining greater reliability for metabolite assignments and isotopomer distribution analysis as well as resolving capacity.

Furthermore, there is a resurging interest in in vivo NMR applications, as magnetic resonance imaging (MRI) and image-guided spectroscopy (MRS) have greatly improved in terms of ease of use, sensitivity, and circumvention of motion-mediated artifacts. Together with further developments in hyperpolarization technologies, particularly in terms of the choice of reagents [ 381 ] and repetitive polarization, MRS will find many more powerful applications not only in basic and preclinical research but also in the clinical arena [ 92 , 298 , 382 , 383 ]. The metabolic insights to be gained from NMR analyses will be a key to our quest for achieving systems biochemical understanding in living organisms.

  • Here we review recent developments in NMR applications to metabolic studies using stable isotope tracers.
  • Recent developments in sample handling, structure determination and quantification are described.
  • The unique strengths of NMR in isotopomer analysis and in vivo investigations are highlighted.
  • New hardware and techniques are illustrated with a wide range of applications in biological research.
  • We also place NMR investigations of metabolic networks in the context of integration with other ‘omics technologies.


This work was supported in part by NIH grants: 1R01CA118434-01A2, 1U24DK097215-01A1, P01 CA163223-01A1 and NIH 5R01ES022191-04. NIH had no role in the preparation of the article.

We thank Drs. P. Dobrowolski and P. Lin for careful reading of the manuscript.

Glossary of abbreviations

8. Uncited reference


  1. (PDF) Research Paper Biochemistry

    research paper of biochemistry

  2. Past Papers:Biochemistry

    research paper of biochemistry

  3. Buy Annals of Clinical Biochemistry Journal Subscription

    research paper of biochemistry

  4. Journal of Biological Chemistry Template

    research paper of biochemistry

  5. Journal of Biochemistry and Molecular Medicine Template

    research paper of biochemistry

  6. (PDF) some papers in biochemistry

    research paper of biochemistry



  2. Msc zoology second paper biochemistry and immunology question 2022-2023✌ #mgsuniversity #viralvideo

  3. introduction to biochemistry

  4. Biochemistry Paper l and ll March 2022

  5. Dmlt Biochemistry Question paper

  6. MSc zoology 1st semester paper biochemistry and cell biology 2023 Lucknow University #shorts


  1. Biochemistry

    Biochemistry is the study of the structure and function of biological molecules such as proteins, nucleic acids, carbohydrates and lipids. Biochemistry is also used to describe techniques suited ...

  2. Biochemistry

    Biochemistry is a Transformative Journal. Biochemistry has been certified as a transformative journal by cOAlition S, committing to a transition to 100% open access in the future. If your research funder has signed Plan S, your open access charges may be covered by your funder through December 31, 2024. Please visit the ACS Open Science website ...

  3. The Journal of Biochemistry

    The Journal of Biochemistry Award Papers. The JB Award is an annual prize which recognized the best papers publishing the journal. The award is presented by the Japanese Biochemical Society at the society's annual meeting. Explore the prize winning articles. Special & Virtual Issues from JB.

  4. Articles

    Citation: BMC Biochemistry 2016 17:10 Content type: Research article Published on: 20 May 2016. View Full Text View PDF Tubulin is a molecular target of the Wnt-activating chemical probe. In drug discovery research, cell-based phenotypic screening is an essential method for obtaining potential drug candidates. ...

  5. Home Page: Journal of Biological Chemistry

    A journal of the American Society for Biochemistry and Molecular Biology. The Journal of Biological Chemistry (JBC) welcomes high-quality science that seeks to elucidate the molecular and cellular basis of biological processes. Papers published in JBC can therefore fall under the umbrellas of not only biological chemistry, chemical biology, or biochemistry, but also allied disciplines such as ...

  6. Biochemistry

    The antihyperglycemic potential of pyrazolobenzothiazine 1, 1-dioxide novel derivative in mice using integrated molecular pharmacological approach. Saman Taj. , Usman Ali Ashfaq. & Anwarul Hasan ...

  7. Biochemistry

    Snapshots of genetic copy-and-paste machinery in action. LINE-1 DNA elements self-duplicate, inserting the copy into new regions of the genome — a key process in chromosome evolution. Structures ...

  8. Biochemistry Current Issue

    Characterization of a Structurally Distinct ATP-Dependent Reactivating Factor of Adenosylcobalamin-Dependent Lysine 5,6-Aminomutase. Amanda L. Darbyshire. and. Kirsten R. Wolthers *. Biochemistry 2024, 63, 7, 913-925 (Article) Publication Date (Web): March 12, 2024. Abstract.

  9. High-Impact Research

    Explore a collection of the most read and most cited articles making an impact in the Journal of Biochemistry published within the past two years.This collection will be continuously updated with the journal's leading articles so be sure to revisit periodically to see what is being read and cited.

  10. Journal of Biological Chemistry

    A journal of the American Society for Biochemistry and Molecular Biology. The Journal of Biological Chemistry (JBC) welcomes high-quality science that seeks to elucidate the molecular and cellular basis of biological processes. Papers published in JBC can therefore fall under the umbrellas of not only biological chemistry, chemical biology, or biochemistry, but also allied disciplines such as ...

  11. Publications

    Publications. Publishing world-leading research and reviews across seven journals, we work in partnership with researchers, institutions, and funders to share knowledge and advance the molecular biosciences. As a self-publishing learned society, all our profits are returned to the life science community in support of our charitable activities.

  12. About Biochemistry

    Biochemistry provides an international forum for publishing exceptional, rigorous, high-impact research in all aspects of biological chemistry. Topics may include, but are not limited to: The structure, function, regulation, and/or dynamics of natural, designed, or evolved biopolymers (e.g. proteins, DNA, RNA, oligosaccharides) The chemical ...

  13. 312985 PDFs

    Biochemistry - Science topic. This group is designed for the exchange of thoughts on Biochemistry | Explore the latest full-text research PDFs, articles, conference papers, preprints and more on ...

  14. Home

    Biochemistry (Moscow) is a peer-reviewed journal that publishes comprehensive research in all fields of biochemistry and related fields. Coverage extends to biochemical aspects of molecular biology, bioorganic chemistry, microbiology, immunology, physiology, and biomedical sciences. Encompasses experimental methods in biochemistry and ...

  15. Biochemistry Research International

    Biochemistry Research International publishes original research articles as well as review articles covering all areas of biological chemistry. About this journal. Editor spotlight. Chief Editor, Professor Andrei Surguchov, is based at the University of Kansas Medical Center, USA. His current research focuses on the structure-function ...

  16. A revolution in biochemistry and molecular biology education informed

    The biochemistry students at UNL have been highly successful as evidenced by co-authorship on research papers, presentations, and awards. Over the past five years, biochemistry students have presented their research at the ASBMB annual meeting, where they have had opportunities to talk with the leaders in the field.

  17. Top 10 in Biochemistry and Molecular Biology

    Explore our most highly accessed biochemistry and molecular biology articles in the first quarter of 2018. Featuring authors from around the World, these papers highlight valuable research from an ...

  18. Process Biochemistry

    Process Biochemistry is an application-orientated research journal devoted to reporting advances with originality and novelty, in the science and technology of the processes involving bioactive molecules and living organisms.These processes concern the production of useful metabolites or materials, or the removal of toxic compounds using tools and methods of current biology and engineering.

  19. Biomolecules

    This Topical Collection, "Feature Papers in Biochemistry", aims to collect high-quality research articles, review articles, and communications on all aspects of biochemistry. It is dedicated to recent advances in the research area of biochemistry, and comprises a selection of exclusive papers from the Editorial Board Members (EBMs) of the ...

  20. Biochemistry Ahead of Print

    Poacic Acid, a Plant-Derived Stilbenoid, Augments Cell Wall Chitin Production, but Its Antifungal Activity Is Hindered by This Polysaccharide and by Fungal Essential Metals. Adi Yona. and. Micha Fridman *. Biochemistry, Articles ASAP (Article) Publication Date (Web): March 27, 2024. Abstract.

  21. Teaching and Learning Medical Biochemistry: Perspectives from a Student

    To do so, medical biochemistry educators need first to define core concepts and biochemical principles that teachers and students could review (or relearn) in a short period. The educator will then teach new concepts and medical applications of biochemistry, in an in-depth fashion, via a series of carefully designed PBL (problem-based learning ...

  22. Applications of NMR spectroscopy to systems biochemistry

    1. Introduction. Since the discovery of nuclear magnetic resonance phenomenon in solids and liquids by Bloch and Purcell in 1945 [ 1, 2 ], NMR spectroscopy has proven to be a powerful and versatile tool for structure elucidation for organic chemists, followed by structural and dynamic determination of macromolecules for the structural ...

  23. IJMS

    This Special Issue aims to collect high quality review papers in all the fields of Biochemistry. We encourage researchers from related fields to contribute review papers highlighting the latest developments in Biochemistry, or to invite relevant experts and colleagues to do so. Full length comprehensive reviews will be preferred.