Skip to main content
NCBI home page
Springer logoLink to Springer
. 2025 May 27;21(3):71. doi: 10.1007/s11306-025-02269-5

Comprehensive profiling of folates across polyglutamylation and one-carbon states

Sevcan Erşan 1, Yu Chen 2, Junyoung O Park 1,✉
PMCID: PMC12116830  PMID: 40423736

Abstract

Introduction

One-carbon metabolism is central to carbon fixation, methylation, and biosynthesis of amino acids, lipids, and nucleotides. Folates are organic cofactors that harbor one-carbon units and shunt them across these metabolic pathways. Despite its essentiality to all life forms, the diverse nature of folate species with various polyglutamylation and one-carbon states makes their measurement challenging.

Objectives

We aim to illuminate one-carbon metabolism by streamlining comprehensive profiling of folate polyglutamates.

Methods

We analyze folate standards and cellular extracts containing diverse folates species by liquid chromatography-mass spectrometry (LC-MS).

Results

We observe that Escherichia coli cells possess diverse folate polyglutamates with one to ten terminal glutamates. Interestingly, most folate polyglutamates form doubly charged ions as well as singly charged ions in LC-MS. Folates also undergo in-source fragmentation. The disparate fates of folates in MS make their quantitation prone to underestimation. Fragmentation by in-source collision-induced dissociation (CID) and LC separation circumvent this issue and facilitate robust and sensitive quantification of folates. In-source CID of folates generates reporter fragment ions that yield higher signals in the mass-to-charge ratio (m/z) range near the maximal mass resolution of Orbitrap MS. Our LC methods complement MS by effectively separating folates based on their polyglutamylation and one-carbon states.

Conclusion

Our metabolomics approach tailored to folate polyglutamates reveals multiple layers of one-carbon metabolism organized by the lengths of polyglutamate tails in folates. Our analytical workflow is broadly applicable to folate profiling across various cell types to advance our knowledge of one-carbon metabolism as well as biotechnology and medicine.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11306-025-02269-5.

Keywords: Folate, One-carbon metabolism, Polyglutamylation, LC-MS, Metabolomics

Introduction

Folates encompass a broad spectrum of one-carbon carriers involved in amino acid and lipid metabolism as well as nucleotide synthesis and DNA methylation [1]. In acetogens, folates facilitate serial reduction of CO2 into a methyl group through the Wood-Ljungdahl pathway, also known as the reductive acetyl-coenzyme A pathway [2]. Each folate molecule consists of pterin, p-aminobenzoate, (poly)glutamate, and optionally a one-carbon unit (Fig. 1a) [3]. Folates are classified by the oxidation states of the pterin and the one-carbon unit as well as the degree of polyglutamylation.

Fig. 1.

Fig. 1

Open in a new tab

Structural diversity of folate polyglutamates in one-carbon metabolism. a Folates (PteGlu) are tripartite molecules composed of a pterin ring and a p-amino-benzoyl group (pAB), which are linked by a methylene bridge to form a pteroyl group (Pte), and glutamate (Glu). Pterin ring can be oxidized at R1 and N8 (PteGlu), reduced at R1 (H2PteGlu), and reduced at both R1 and N8 (H4PteGlu). The one-carbon units of various oxidation states (CH, CH2, CH3 or CHO) can attach at N5 or N10 positions (5-CH3-H4PteGlu, 5-CHO-H4PteGlu, and 10-CHO-H4PteGlu) or bridged in between (5,10-CH = H4PteGlu, and 5,10-CH2-H4PteGlu). Each folate species carries one or more terminal glutamates that constitute a polyglutamate tail. b In one-carbon metabolism, folates carry one-carbon units through redox reactions and interfaces with amino acid metabolism, nucleotide metabolism, and lipid metabolism. Folylpolyglutamate synthase catalyzes the ATP-dependent addition of glutamate to H4PteGlun, 5/10-CHO-H4PteGlun, and 5,10-CH2-H4PteGlun

Most folates inside the cell exist as polyglutamates containing as many as 14 glutamates and participate as coenzymes in one-carbon metabolism (Fig. 1b) [4, 5]. However, past research focused on folate monoglutamates due to their relevance to nutrition [6–9] and their presence in circulation [10]. An ability to reliably measure intracellular folate polyglutamates would advance our understanding of regulation of one-carbon metabolism in health and disease and promote metabolic engineering strategies for one-carbon utilization in biotechnology.

Various liquid chromatography-mass spectrometry (LC-MS) methods have been employed for measurement of folates (Table S1) [11–18]. However, comprehensive analysis of folate polyglutamates remains challenging due to a lack of their reference mass spectra, low intracellular concentrations [5, 13, 17], and instability [19, 20]. Folate polyglutamates can form diverse ion patterns in MS akin to peptides [21]. Adding to the challenge, ionization process in LC-MS often generates in-source fragments and adducts to obfuscate analyte’s identity and quantity [22].

We present LC-MS methods and a predictive LC retention model [23] for comprehensive profiling of folate polyglutamates. We used commercially available folate polyglutamates standards [24] to probe the effect of the length of polyglutamate tails on their retention in hydrophilic interaction chromatography (HILIC) and C18 reversed-phase chromatography. We generated their fragments using in-source collision-induced dissociation (CID) to yield mass spectra carrying structural information and common moieties useful for quantifying diverse folate species [25–29]. We demonstrate comprehensive folate quantification in Escherichia coli. Our analytical approach can be applied to other microorganisms as well as plant and mammalian systems to unveil the diversity and the unknown roles of folate polyglutamates.

Materials and methods

Folate standards and solution preparation

Authenticated standards of select folate species were obtained from Schircks Laboratories (Switzerland) (Table S2). p-Aminobenzoyl-glutamate (pABGlu) and folic acid (PteGlu) were obtained from Sigma-Aldrich (St. Louis, MO). The standards represented all relevant species in one-carbon metabolism but 10-CHO-H4PteGlun due to their unavailability (Fig. 1). The standards were stored in a vacuum desiccator in a freezer until analysis. Stock standard solutions were prepared at 1 mM concentration in a solution of 2.5 mM sodium ascorbate, and 25 mM ammonium acetate at pH 7 [18]. Due to solubility limitations, Tetrahydrofolate (THF or H4PteGlu) was prepared at a concentration of 0.3 mM. Dihydrofolate (DHF or H2PteGlu) was dissolved in 1 mM ammonium acetate and 1 mM ammonium hydroxide at pH 9.4 following vendor recommendation. Brief ultrasonication was applied when necessary to ensure complete solubilization. For direct MS injection, standard solutions were prepared using the same solvent system but without sodium ascorbate.

Microbial culture preparation and folate extraction

Escherichia coli strain K-12 strain MG1655 was cultured in Gutnick minimal medium [30] at 37 °C in a shaker incubator. A 4-mL sample of overnight culture was vacuum-filtered on a nylon membrane (25 mm diameter, 0.2 μm pore size, MilliporeSigma, Burlington, MA). The membrane was immediately transferred to a six-well plate containing 1 mL of precooled (4 °C) extraction solution (80:20 HPLC-grade acetonitrile/water with 2.5 mM sodium ascorbate and 25 mM ammonium acetate at pH 7) to quench metabolism and extract metabolites. Extraction continued for 20 min at 4 °C. The extract was transferred to a microcentrifuge tube and centrifuged at 17,000 x g for 10 min at 4 °C. The supernatant was transferred to a new tube, heated to 60 °C for 5 min to inactivate residual enzymes, and cooled on ice immediately after [18, 31]. The sample was dried under nitrogen gas flow, reconstituted in 40 µL of HPLC-grade water containing 25 mM sodium ascorbate and 25 mM ammonium acetate at pH 7, and transferred to an HPLC vial for LC-MS analysis.

Direct infusion mass spectrometry

A hybrid quadrupole-Orbitrap mass spectrometer (Q Exactive Plus, Thermo Fisher Scientific) with heated electrospray ionization (HESI-II) probe was used for direct MS analysis of folate standards. MS was operated under the following conditions: spray voltage of 3.5 kV, a spray current of 1 µA, a capillary temperature set at 320 °C, the sheath gas with a flow rate of 28 au (arbitrary units), a 350 °C auxiliary gas with a flow rate of 10 au, S-lens RF level of 70, and a sweep gas flow rate of 0.98 au. Mass spectra were collected in both positive and negative ion modes over mass-to-charge ratios (m/z) between 100 and 1500, with a resolution of 140,000 at m/z 200, an automatic gain control (AGC) target at 3e6, and maximum injection time (IT) of 500 ms.

In-source CID activation energy ranging from 0 eV to 100 eV was applied in 10-eV increments. Higher energy collision dissociation (HCD) for MS2 was conducted using all ion fragmentation (AIF) with normalized collisional energy (NCE) from 10 to 50 in 10-unit increments. MS2 spectra were recorded at a resolution of 35,000 at m/z 200 following a full MS1 scan at a resolution of 70,000 at m/z 200.

Stock standard solutions of folate polyglutamates (prepared without sodium ascorbate) were diluted to 400 µM or 100 µM in 95:5 water/acetonitrile buffer at pH 9.4 (20 mM ammonium acetate and 20 mM ammonium hydroxide) or pH 4.0 (10 mM ammonium acetate with 0.1% acetic acid). These solutions were directly injected into the ESI source of MS at a flow rate of 50 µl/min.

Data processing involved analyzing the full scan MS1 and AIF MS2 spectra using Metabolomic Analysis and Visualization Engine (MAVEN, version 26.4) [32] and Xcalibur (version 4.2.47; Thermo Fisher Scientific), respectively. The top 10 most abundant peaks in MS1 spectra were selected after the removal of background noise and isotopic peaks. Additional expected (de)protonated and adduct ions were taken for further analysis.

Liquid chromatography-mass spectrometry

LC-MS analysis was performed using a Vanquish Duo UHPLC system (Thermo Fisher Scientific) coupled to Q Exactive Plus MS. For LC, hydrophilic interaction liquid chromatography (HILIC) or a reversed-phase liquid chromatography (RPLC) were employed. Full MS scan over m/z 100–1500 in both positive and negative ion modes followed with a resolution of 140,000 at m/z 200, AGC target at 3e6, and max IT of 500 ms.

HILIC was adopted from an untargeted metabolomics method [33]. RPLC was adopted from Schober et al. [34]. For HILIC, an Xbridge BEH Amide XP column (130Å, 2.5 μm, 2.1 mm × 150 mm; Waters Corporation, Milford, MA) was used. For RPLC, a superficially porous amide-embedded C18 column (InfinityLab Poroshell 120 Bonus, 120Å, 2.7 μm, 2.1 mm × 150 mm; Agilent Technologies) was used. Both columns were maintained at 25 °C.

Two types of eluent A were used for HILIC. Basic eluent A (pH 9.4) consisted of 20 mM ammonium acetate and 20 mM ammonium hydroxide in 95:5 water/acetonitrile. Acidic eluent A (pH 4.0) consisted of 10 mM ammonium acetate with 0.1% acetic acid in 95:5 water/acetonitrile. Eluent B was acetonitrile. For RPLC, only the acidic eluent A (pH 4.0) was used, and eluent B was acetonitrile. Two segmented gradients were used for each LC type.

HILIC method I. acidic/basic: Both acidic (pH 4.0) and basic (pH 9.4) eluents were used. Chromatographic conditions were as follows: 90% B at 0 min, 90% B at 2 min, 75% B at 3 min, 75% B at 7 min, 70% B at 8 min, 70% B at 9 min, 50% B at 10 min, 50% B at 12 min, 25% B at 13 min, 25% B at 14 min, 0% B at 16 min, 0% B at 20 min, 90% B at 21 min, 90% B at 25 min. The total run time was 25 min with a flow rate of 150 µL/min. The sample injection volume of 5 µL.

HILIC method II. acidic/basic: Both acidic (pH 4.0) and basic (pH 9.4) eluents were used. Chromatographic conditions, modified after “HILIC method I” gradient, were as follows: 90% B at 0 min, 90% B at 2 min, 40% B at 22 min, 0% B at 25 min, 0% B at 28 min, 90% B at 30 min, 90% B at 40 min. The total run time was 40 min with a flow rate of 150 µL/min. The sample injection volume of 5 µL.

RPLC method I: Only acidic eluent (pH 4.0) was tested. Chromatographic conditions were as follows: 2% B at 0 min, 98% B at 20 min, 98% B at 25 min, 2% B at 26, 2% B at 35 min. The total run time was 35 min with a flow rate of 200 µL/min. The sample injection volume was 20 µL.

RPLC method II: Only acidic eluent (pH 4.0) was tested. Chromatographic conditions were as follows: 2% B at 0 min, 2% B at 2 min, 29% B at 8 min, 95% B at 10 min, 95%B at 12 min, 2% B at 14 min, and 2% at 20. The total run time was 20 min with a flow rate of 200 µL/min. The sample injection volume was 20 µL.

Chromatographic resolution (Inline graphic) as an indication of LC separation of two analyte peaks was calculated using the following equation:

graphic file with name d33e586.gif

where Inline graphic and Inline graphic are retention times of respective peak maxima, and Inline graphic and Inline graphic are the individual peaks’ full widths at half maxima (FWHM) in minutes [35].

Identification and quantification of metabolites were performed by processing LC-MS data using MAVEN [32] and comparing retention time and m/z to those of authenticated standards (Table S2) and an in-house database (Table S3). Extracted ion chromatograms (EIC) corresponding to individual m/z values of interest were exported.

Regression analysis for LC retention

The relationship between LC retention times and the degrees of folate polyglutamylation was determined by log-linear regression analysis using retention time measurements from 5-methyltetrahydrofolate (CH3-THF or CH3-H4PteGlu) standards with one to four terminal glutamates. Retention times of CH3-THF with five and six glutamates were predicted from experimental values of 5-formyltetrahydrofolate (5-CHO-THF or 5-CHO-H4PteGlu) standards with five and six terminal glutamates by assuming a constant retention time offset between 5-CHO-THF and CH3-THF across all polyglutamate forms. The model of LC retention times for folate polyglutamates was formulated as follows:

graphic file with name d33e642.gif

where Inline graphic represents the retention time in minutes, Inline graphic is the number of glutamates in the folate species. Inline graphic and Inline graphic are constants representing the intercept and slope of the regression line, respectively. Inline graphic represents a retention time shift specific to the monoglutamate form of each folate species of interest in reference to CH3-THF, for which Inline graphic.

Results and discussion

Adducts and doubly charged ions of folate species are prevalent in direct infusion MS

We sought to identify ions that offer sensitive detection and accurate identification of folate polyglutamates. We characterized the mass spectra resulting from the direct injection of folate standards, which eliminated potential artifacts and interferences that may result from folate instability during liquid chromatography [19, 20]. Full scan MS data were collected to detect comprehensive ion features within a m/z range from 50 to 1500. For each standard injection, we examined the total ion current (TIC) chromatogram and focused on the top 10 peaks with the highest signal intensities (Fig. 2).

Fig. 2.

Fig. 2

Open in a new tab

The top 10 most abundant ions in direct infusion mass spectrometry of folate polyglutamates. Folate polyglutamates with three or more terminal glutamates were doubly charged [M + 2 H]2+. Cationic adducts of sodium, potassium, calcium, and lithium were observed. Mass spectra were collected in full scan positive ion mode. Precursor ion peaks are denoted in red, cationic adducts in blue, interconversion and degradation products in purple, dimer ions in pink, fragment ions in green, and unidentified peaks in grey

Electrospray ionization of folate polyglutamates generated diverse ion products, including singly and doubly charged monomeric and dimeric forms in both positive and negative mode (Fig. 2 and Figs. S1-3). Signal intensities were comparable between positive and negative mode (Table S4). Adduct formation with alkali metals was observed. The type and relative abundance of these adducts depended on folate species, length of polyglutamate tail, and matrix composition. Adduct formation was particularly prevalent in CHO-THF with four to six terminal glutamates with singly and doubly charged lithium adducts surpassing the abundance of [M + H]+, [M + 2 H]2+, [M-H]−, and [M-2 H]2− ions (Fig. 2 and Figs. S1-3). While the high abundance of lithium adduct was due to folate polyglutamate standards being prepared as lithium salts, other cation adducts such as sodium, potassium, and calcium were also observed.

Folate polyglutamates with three or more terminal glutamates displayed doubly charged parent ions [M + 2 H]2+ (Fig. 2), consistent with a previous observation [18]. Folate polyglutamates even with long polyglutamate tails were only singly or doubly charged unlike peptides which may have higher charge states under ESI [36]. The abundance of doubly charged ions became more prominent with longer polyglutamate tails, surpassing those of singly charged ions upon reaching six terminal glutamates.

Cell lysates abound in doubly charged ions and sodium adducts of folate polyglutamates

We profiled folates in E. coli by LC-MS. In positive ion mode, we detected 38 tetrahydrofolate species with one to 10 terminal glutamates (Fig. 3). Tetrahydrofolates with one to four terminal glutamates were mainly observed as singly charged ions [M + H]+. Folate polyglutamates with seven to 10 terminal glutamates were detected almost exclusively as doubly charged ions [M + 2 H]2+. Sodium adducts [M + Na]+ were prevalent for CHO-THF polyglutamates but not for other types of folates. Sodium adducts were abundant despite Gutnick minimal medium not containing sodium likely due to the quenching and extraction solution containing sodium ascorbate. The formyl group of CHO-THF in conjunction with the carboxyl-rich polyglutamate tail, which can readily interact with cations [37], may be responsible for stabilizing the sodium adduct. Thus, detection and quantification of folate polyglutamates in biological samples warrant examining doubly charged ions and monovalent and divalent cation adducts.

Fig. 3.

Fig. 3

Open in a new tab

LC-MS analysis of folate polyglutamates in E. coli. Metabolite extract from E. coli was analyzed on LC-MS set up with HILIC method I, basic eluent, and positive ion mode. CHO-H4PteGlun represented isomers 5-CHO-H4PteGlun and 10-CHO-H4PteGlun. Doubly charged ions [M + 2 H]2+ became more abundant as the number of terminal glutamates (n) in folate polyglutamates increased. The abundance of doubly charged ions overtook that of singly charged ions with n ≥ 6. The plots show LC-MS analysis of folate species from one E. coli culture. The folate profiles including doubly charged ions were reproducible in a biological replicate

Based on ion counts, CH3-THF was the most abundant, especially with four to eight terminal glutamates. For other folate species THF, CHO-THF, and methylenetetrahydrofolate (CH2-THF or CH2-H4PteGlu), polyglutamates with three terminal glutamates were most abundant, consistent with prior observations [17, 18]. Folate species with more than 10 terminal glutamates were not detected in E. coli lysate. In erythrocytes and plants, folates with as many as 11 and 14 terminal glutamates, respectively, have been observed [4, 5]. These variations in the lengths of glutamate tails across different cell types could be attributed to biological differences such as cell type and physiological state [38, 39]. Improved folate extraction and instrument sensitivity may lead to robust quantification of even larger folate polyglutamates.

Folate species undergo in-source fragmentation and interconversion

We detected several prominent peaks besides adducts and doubly charged ions in direct infusion mass spectrometry of folate standards (Fig. 2 and Figs. S1–3). Given that the standard solutions had been freshly prepared just before MS analysis, these peaks suggested in-source modifications. To distinguish between in-source fragmentation and in-sample degradation, we put the standards through LC-MS. If the fragment ions had the same retention times as the parent ions, we determined that the fragments were formed in the source. Otherwise, the “fragment” peaks would have had to come from in-sample (or on-column) degradation.

LC-MS analysis of folate polyglutamate standards revealed their in-source fragments formed by neutral loss of glutamate residues. In positive ion mode, folate polyglutamates produced a common fragment ion F1 at m/z 313.1408 and m/z 327.1200 for CH3-THF and CHO-THF polyglutamates, respectively (Fig. 4a and Table S5). The consistent formation of F1 fragments across all folate standards suggested their potential use as reporter ions to identify and quantify folate species of varying one-carbon oxidation state (Fig. 4b) regardless of their polyglutamylation state (Fig. 4c). Targeting the F1 fragments would improve quantification accuracy and spectral accuracy by allowing a narrow scan range that targets only these fragment ions rather than the broad mass range of folate polyglutamates [40].

Fig. 4.

Fig. 4

Open in a new tab

In-source fragmentation of folate polyglutamates in LC-MS. (a) LC-MS analysis of folate polyglutamates revealed various fragments including those shown as F1 through F5. (b) In-source fragmentation of various folate monoglutamate species produced F1 fragments, whose LC retention times were the same as those of the parent ions. (c) In-source fragmentation of various folate polyglutamates produced F1 fragments. (d) LC-MS of CH2-H4PteGlu standard also produced two CH = H4PteGlu peaks. The left peak shared the same retention time with CH2-H4PteGlu, suggesting its in-source oxidation. The retention time of the right peak aligned with that of CH = H4PteGlu standard (dashed gray line), indicating the in-sample or on-column conversion of CH2-H4PteGlu

Interestingly, with the CH2-THF standard, we observed both in-source and in-sample conversion of its one-carbon unit to form methenyltetrahydrofolate (CH = THF or CH = H4PteGlu) (Fig. 4d). Of the two distinct peaks detected in CH = THF mass channel obtained upon injecting the CH2-THF standard to LC-MS, the earlier peak shared the same retention time as the CH2-THF standard, indicating in-source oxidation of CH-H4PteGlu. The latter peak corresponded to in-sample or on-column oxidation of the CH2-THF standard into CH = THF. Folates undergo interconversion and degradation in solution without proper precautions [19, 20, 41]. While chemical derivatization and elimination of residual enzyme activity mitigate these pitfalls [41], in-source and in-sample conversion may confound accurate identification and quantification of folates in biological samples (Supplementary Note 1).

In-source CID and MS2 generate reporter ions for sensitive folate quantification

While the F1 in-source fragments showed promise for identifying and quantifying folates across polyglutamylation and one-carbon state, they were not as abundant as the parent ions [M + H]+, thus posing a shortcoming in the lower limit of detection (Fig. 4b, c). To improve F1 fragment formation, we tested in-source CID with direct infusion MS. With ion kinetic energies ranging from 30 eV to 50 eV, we observed strong peaks for the F1 fragment ions of all folate species except for CH = THF (Fig. 5a). For CHO-THF, CH3-THF, and THF, we observed F2, and for CH3-THF, and THF, we also observed F4, and F5 with their strongest signals coming at or above 50 eV. Glutamate was also observed as a fragmentation product for all THF species but CH2-THF (Fig. S4a). As ion kinetic energies increased, the signals for parent ions and sodium adducts decreased. For polyglutamated forms of CHO-THF and CH3-THF, the F1 fragment ions displayed the strongest signals across all tested lengths of polyglutamate tails (Figs. 5b and S4b).

Fig. 5.

Fig. 5

Open in a new tab

Fragmentation products from in-source CID of folate polyglutamates. a Upon direct infusion mass spectrometry and in-source CID of folate monoglutamates, the parent ions and cation adducts as well as F1, F2, F4, and F5 fragments were observed under ion kinetic energies ranging from 0 to 100 eV. b The same in-source CID of folate polyglutamates generated singly and doubly charged ions as well as F1 fragments

Fragmentation by AIF on MS2 was performed for comparison (Figs. S5 and S6). On MS2, the F1 ions, resulting from the loss of Glun from folates, were consistently among the most abundant across all folate polyglutamates. Unique to CH2-THF, fragment ion F3 was observed. For CH = THF, we did not observe much discernible fragmentation across the tested NCE range. These observations suggested that the one-carbon bridges between N5 and N10 in CH2-THF and CH = THF lead to fragmentation patterns different from other folate species. The strong signals for F1 fragment ions in both in-source CID and AIF MS2 indicated their utility as sensitive quantification and accurate identification of various folate species regardless of polyglutamate tail lengths. An added benefit of F1 fragment ions was their small sizes amenable to near-maximal mass resolution in Orbitrap-based MS.

HILIC separates folate polyglutamates with different numbers of terminal glutamates

The F1 fragments are useful reporter ions because their structures are distinct for DHF, THF, and those THF molecules that harbor different one-carbon units. Folate species that are different only in the lengths of their polyglutamate tails may be distinguished by the parent ions in MS1 or in MS/MS but not by in-source CID because their F1 fragments are identical. Thus, we needed to separate folate polyglutamates by LC.

We measured the retention times of folate standards on HILIC and RPLC followed by MS (Fig. S7a, b and Table S6). For HILIC eluent A buffers were at pH 4.0 and 9.4, and for RPLC, eluent A buffer was at pH 4.0. We tested different solvent gradients to separate in-source modification from in-sample degradation (Supplementary Note 2). Overall, HILIC retained and separated folate species better than RPLC (Fig. 6a and Table S7). Peak widths in RPLC, indicated by the FWHM increased from 0.14 min to 0.60 min as the number of terminal glutamates increased from one to three (Fig. 6b). In both HILIC and RPLC, increasing lengths of polyglutamate tails increased the retention times of folates.

Fig. 6.

Fig. 6

Open in a new tab

LC retention behavior of folate polyglutamates. a HILIC method II effectively separated folate polyglutamates based on the numbers of their terminal glutamates. b While RPLC method separated folate polyglutamates, peak broadening occurred as the number of terminal glutamates increased. c Both HILIC and RPLC methods separate folate monoglutamates. Although folate monoglutamates eluted earlier in RPLC, the retention time differences between them were larger. d All tested folate standards eluted between 40% and 52% aqueous mobile phase in HILIC. A higher percent of aqueous mobile phase is expected to elute folates with longer polyglutamate tails. e The retention times of folate polyglutamates followed a log-linear relationship with respect to the number of terminal glutamates. f Taking into consideration the baseline differences in folate monoglutamates, the retention times of folates species with various one-carbon units and polyglutamylation in HILIC were predicted. g For the polyglutamate forms of four folate species THF, CHO-THF, CH2-THF, and CH3-THF, the observed retention times matched with the predicted retention times

In addition to polyglutamylation, the pterin ring and the one-carbon unit of folates affected LC retention. Reduction of pterin ring, from folate (PteGlu) to DHF (H2PteGlu) to THF (H4PteGlu), led to longer retention in HILIC but shorter retention in RPLC (Fig. 6c). In HILIC, the tested folate standards eluted between 40% and 52% aqueous mobile phase in the solvent gradient (Fig. 6d). Since more terminal glutamates increased the retention of folate species, > 52% aqueous mobile phase may be required to elute longer polyglutamates. Thus, combined LC-MS (including in-source CID) would lead to robust and accurate measurement of folate polyglutamates.

Folate polyglutamate LC retention prediction and validation in vivo

Since not all folate polyglutamate standards are readily available, we developed an LC retention model to predict the retention times of any folate polyglutamates. To this end, we performed regression analysis using available CH3-THF and CHO-THF polyglutamate standards (Fig. 6e). We observed a log-linear relationship between the number of terminal glutamates and the retention time of folate polyglutamates in HILIC with a high coefficient of determination (R2 = 0.96). To account for the baseline retention time differences between folate monoglutamates, we added a retention time shift (ε) of each folate species relative to CH3-THF, for which ε = 0. The revised model was then employed to predict the LC retentions of six folate species with as many as 10 terminal glutamates (Fig. 6f). For polyglutamated forms of THF, CHO-THF, CH2-THF, and CH3-THF, the predicted retention times in HILIC were consistent with measurements from E. coli lysate (R2 = 0.96) (Figs. 6g and S8a). On the other hand, we observed discrepancies between observed and expected retention times of CH = H4PteGlun and H2PteGlun (Fig. S8b). The inaccurate prediction for the CH = THF and DHF species was due to the subdued effect of additional terminal glutamates on folate retention that results in insufficient separation. In the E. coli lysate samples, all detected DHF polyglutamates but one were observed within 0.4 min of each other, and all detected CH = THF polyglutamates were observed within 0.1 min of each other.

Conclusion

In the present study, we comprehensively characterized the chromatographic behavior, ionization, and in-source conversion of folates with varying polyglutamylation and one-carbon states in LC-MS. The presented LC-MS strategies enabled sensitive and robust quantification of folate polyglutamates across ranges of one-carbon units and terminal glutamates. As most cellular folates exist in polyglutamate forms, we expect our methods to advance our knowledge of one-carbon metabolism. Specifically, we can now start thinking about folate species with the same polyglutamate tail length occupying a layer of one-carbon metabolism and probe if carbon fluxes differ across layers, which may serve different purposes. For example, reductive one-carbon metabolism may utilize different layers (i.e., different pools of folate polyglutamates) than oxidative one-carbon metabolism as carbon flux may be more streamlined within folate pools that have the same or similar numbers of terminal glutamates. Thus, profiling cellular folate polyglutamates will help elucidate multilayer one-carbon metabolism in carbon-fixing organisms and in human disease models, advancing sustainable biotechnology and medicine.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1 (2.4MB, docx)

Acknowledgements

The authors would like to thank the members of the Park laboratory, the UCLA Molecular Instrumentation Center, and the UCLA Metabolomics Center for helpful discussion.

Abbreviations

LC-MS

Liquid chromatography-mass spectrometry

ESI

Electrospray ionization

pAB

p-aminobenzoate

Pte

Pteroyl group or pteroic acid

Glu

L-glutamate

pABGlu

p-aminobenzoyl-glutamate

PteGlu

Pteroylglutamate (folate)

PteGlun

Pteroylpolyglutamate (folate polyglutamate)

H2PteGlu

Dihydropteroylglutamate (dihydrofolate)

DHF

Dihydrofolate

H4PteGlu

Tetrahydropteroylglutamate (tetrahydrofolate)

THF

Tetrahydrofolate

(5-/10-)CHO-H4PteGlu or (5-/10-)CHO-THF

Formyltetrahydrofolate

(5

10-)CH = H4PteGlu or (5,10-)CH = THF: Methenyltetrahydrofolate

(5

10-)CH2-H4PteGlu or (5,10-)CH2-THF: Methylenetetrahydrofolate

(5-)CH3-H4PteGlu or (5-)CH3-THF

Methyltetrahydrofolate

Author contributions

S.E. and J.O.P. conceived the study, designed the experiments, analyzed and interpreted the data, and wrote the manuscript. S.E. conducted the experiments with assistance from Y.C.

Funding

This work was supported by the National Institute of General Medical Sciences of the National Institutes of Health under Award Number R35GM143127 and the National Institutes of Health Instrumentation Grant under Award Number 1S10OD016387-01. The content is solely the authors’ responsibility and does not necessarily represent the official views of the National Institutes of Health.

Data availability

Data is provided within the manuscript or supplementary information files. Raw data files are available from the corresponding author on reasonable request.

Declarations

Conflict of interest

The authors have no relevant financial or non-financial interests to disclose.

Footnotes

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  • 1.Ducker, G. S., & Rabinowitz, J. D. (2017). One-Carbon metabolism in health and disease. Cell Metab, 25, 27–42. 10.1016/j.cmet.2016.08.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Park, J. O., Liu, N., Holinski, K. M., Emerson, D. F., Qiao, K., Woolston, B. M., Xu, J., Lazar, Z., Islam, M. A., Vidoudez, C., Girguis, P. R., & Stephanopoulos, G. (2019). Synergistic substrate cofeeding stimulates reductive metabolism. Nat Metab. 10.1038/s42255-019-0077-0 [DOI] [PubMed] [Google Scholar]
  • 3.Saini, R. K., Nile, S. H., & Keum, Y. S. (2016). Folates: Chemistry, analysis, occurrence, biofortification and bioavailability. Food Research International, 89, 1–13. 10.1016/j.foodres.2016.07.013 [DOI] [PubMed] [Google Scholar]
  • 4.Garratt, L. C., Ortori, C. A., Tucker, G. A., Sablitzky, F., Bennett, M. J., & Barrett, D. A. (2005). Comprehensive metabolic profiling of mono- and polyglutamated folates and their precursors in plant and animal tissue using liquid chromatography/negative ion electrospray ionisation tandem mass spectrometry. Rapid Communications in Mass Spectrometry, 19, 2390–2398. 10.1002/rcm.2074 [DOI] [PubMed] [Google Scholar]
  • 5.Van Haandel, L., Becker, M., Williams, T., Stobaugh, J., & Leeder, J. S. (2012). Comprehensive quantitative measurement of folate polyglutamates in human erythrocytes by ion pairing ultra-performance liquid chromatography/tandem mass spectrometry. Rapid Communications in Mass Spectrometry, 26, 1617–1630. 10.1002/rcm.6268 [DOI] [PubMed] [Google Scholar]
  • 6.Zhang, H., Jha, A. B., Warkentin, T. D., Vandenberg, A., & Purves, R. W. (2018). Folate stability and method optimization for folate extraction from seeds of pulse crops using LC-SRM MS. Journal of Food Composition and Analysis, 71, 44–55. 10.1016/j.jfca.2018.04.008 [Google Scholar]
  • 7.Freisleben, A., Schieberle, P., & Rychlik, M. (2003). Specific and sensitive quantification of folate vitamers in foods by stable isotope Dilution assays using high-performance liquid chromatography-tandem mass spectrometry. Analytical and Bioanalytical Chemistry, 376, 149–156. 10.1007/s00216-003-1844-y [DOI] [PubMed] [Google Scholar]
  • 8.Czarnowska, M., & Gujska, E. (2012). Effect of freezing technology and storage conditions on folate content in selected vegetables. Plant Foods for Human Nutrition, 67, 401–406. 10.1007/s11130-012-0312-2 [DOI] [PubMed] [Google Scholar]
  • 9.Czarnowska-Kujawska, M., Gujska, E., & Michalak, J. (2017). Testing of different extraction procedures for folate HPLC determination in fresh fruits and vegetables. Journal of Food Composition and Analysis, 57, 64–72. 10.1016/j.jfca.2016.12.019 [Google Scholar]
  • 10.Zheng, X. H., Jiang, L. Y., Zhao, L. T., Zhang, Q. Y., & Ding, L. (2015). Simultaneous quantitation of folic acid and 5-methyltetrahydrofolic acid in human plasma by HPLC-MS/MS and its application to a Pharmacokinetic study. J Pharm Anal, 5, 269–275. 10.1016/j.jpha.2015.05.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Schittmayer, M., Birner-Gruenberger, R., & Zamboni, N. (2018). Quantification of cellular folate species by LC-MS after stabilization by derivatization. Analytical Chemistry, 90, 7349–7356. 10.1021/acs.analchem.8b00650 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Luo, S., Duan, H., Zou, Y., Qiu, R., & Wang, C. (2017). Quantification of total folate, folate species and polyglutamyl folate distribution in winged beans (Psophocarus tetragonolobus (L) DC) from different cultivars and growth stages by Ultra-High performance liquid chromatography tandem mass spectrometry. J Nutr Sci Vitaminol, 63, 69–80. [DOI] [PubMed] [Google Scholar]
  • 13.Kopp, M., Dürr, K., Steigleder, M., Clavel, T., & Rychlik, M. (2016). Development of stable isotope Dilution assays for the quantitation of intra- and extracellular folate patterns of Bifidobacterium adolescentis. Journal of Chromatography A, 1469, 48–59. 10.1016/j.chroma.2016.09.048 [DOI] [PubMed] [Google Scholar]
  • 14.Wang, C., Riedl, K. M., & Schwartz, S. J. (2010). A liquid chromatography-tandem mass spectrometric method for quantitative determination of native 5-methyltetrahydrofolate and its polyglutamyl derivatives in Raw vegetables. Journal of Chromatography. B, Analytical Technologies in the Biomedical and Life Sciences, 878, 2949–2958. 10.1016/j.jchromb.2010.08.043 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Leung, K. Y., De Castro, S. C. P., Cabreiro, F., Gustavsson, P., Copp, A. J., & Greene, N. D. E. (2013). Folate metabolite profiling of different cell types and embryos suggests variation in folate one-carbon metabolism, including developmental changes in human embryonic brain. Molecular and Cellular Biochemistry, 378, 229–236. 10.1007/s11010-013-1613-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Gmelch, L., Wirtz, D., Witting, M., Weber, N., Striegel, L., Schmitt-Kopplin, P., & Rychlik, M. (2020). Comprehensive vitamer profiling of folate monoand polyglutamates in Baker’s yeast (Saccharomyces cerevisiae) as a function of different sample Preparation procedures. Metabolites, 10, 1–19. 10.3390/metabo10080301 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Kwon, Y. K., Lu, W., Melamud, E., Khanam, N., Bognar, A., & Rabinowitz, J. D. (2008). A domino effect in antifolate drug action in Escherichia coli. Nature Chemical Biology, 4, 602–608. 10.1038/nchembio.108 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Lu, W., Kwon, Y. K., & Rabinowitz, J. D. (2007). Isotope Ratio-Based profiling of microbial folates. Journal of the American Society for Mass Spectrometry, 18, 898–909. 10.1016/j.jasms.2007.01.017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Gazzali, A. M., Lobry, M., Colombeau, L., Acherar, S., Azaïs, H., Mordon, S., Arnoux, P., Baros, F., Vanderesse, R., & Frochot, C. (2016). Stability of folic acid under several parameters. European Journal of Pharmaceutical Sciences, 93, 419–430. 10.1016/j.ejps.2016.08.045 [DOI] [PubMed] [Google Scholar]
  • 20.Patring, J. D. M., Johansson, M. S., Yazynina, E., & Jastrebova, J. A. (2005). Evaluation of impact of different antioxidants on stability of dietary folates during food sample Preparation and storage of extracts prior to analysis. Analytica Chimica Acta, 553, 36–42. 10.1016/j.aca.2005.07.070 [Google Scholar]
  • 21.Bielow, C., Ruzek, S., Huber, C. G., & Reinert, K. (2010). Optimal decharging and clustering of charge ladders generated in ESI-MS. Journal of Proteome Research, 9, 2688–2695. 10.1021/pr100177k [DOI] [PubMed] [Google Scholar]
  • 22.Jia, Z., Qiu, Q., He, R., Zhou, T., & Chen, L. (2023). Identification of metabolite interference is necessary for accurate LC-MS targeted metabolomics analysis. Analytical Chemistry, 95, 7985–7992. 10.1021/acs.analchem.3c00804 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Gritti, F. (2021). Perspective on the future approaches to predict retention in liquid chromatography. Analytical Chemistry, 93, 5653–5664. 10.1021/acs.analchem.0c05078 [DOI] [PubMed] [Google Scholar]
  • 24.Stoffel, R., Quilliam, M. A., Hardt, N., Fridstrom, A., & Witting, M. (2021). N-Alkylpyridinium sulfonates for retention time indexing in reversed-phase-liquid chromatography-mass spectrometry–based metabolomics. Analytical and Bioanalytical Chemistry. 10.1007/s00216-021-03828-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Domingo-Almenara, X., Montenegro-Burke, J. R., Benton, H. P., & Siuzdak, G. (2018). Annotation: A computational solution for streamlining metabolomics analysis. Analytical Chemistry, 90, 480–489. 10.1021/acs.analchem.7b03929 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Xue, J., Domingo-Almenara, X., Guijas, C., Palermo, A., Rinschen, M. M., Isbell, J., Benton, H. P., & Siuzdak, G. (2020). Enhanced in-Source fragmentation annotation enables novel data independent acquisition and autonomous METLIN molecular identification. Analytical Chemistry, 92, 6051–6059. 10.1021/acs.analchem.0c00409 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Seitzer, P. M., & Searle, B. C. (2019). Incorporating In-Source fragment information improves metabolite identification accuracy in untargeted LC-MS data sets. Journal of Proteome Research, 18, 791–796. 10.1021/acs.jproteome.8b00601 [DOI] [PubMed] [Google Scholar]
  • 28.Xu, Y. F., Lu, W., & Rabinowitz, J. D. (2015). Avoiding misannotation of in-source fragmentation products as cellular metabolites in liquid chromatography-mass spectrometry-based metabolomics. Analytical Chemistry, 87, 2273–2281. 10.1021/ac504118y [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Su, X., Chiles, E., Maimouni, S., Wondisford, F. E., Zong, W. X., & Song, C. (2020). In-Source CID ramping and covariant ion analysis of hydrophilic interaction chromatography metabolomics. Analytical Chemistry, 92, 4829–4837. 10.1021/acs.analchem.9b04181 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Gutnick, D., Calvo, J. M., Klopotowski, T., & Ames, B. N. (1969). Compounds which serve as the sole source of carbon or nitrogen for Salmonella typhimurium LT-2. Journal of Bacteriology, 100, 215–219. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.House, R. (Rae), Soper-Hopper, J., Vincent, M. T., Ellis, M. P., Capan, A. E., Madaj, C. D., Wolfrum, Z. B., Isaguirre, E., Castello, C. N., Johnson, C. D., Escobar Galvis, A. B., Williams, M. L., Lee, K. S., & Sheldon, H. (Eds.). (2024). RD A diverse proteome is present and enzymatically active in metabolite extracts. Nat Commun 15:. 10.1038/s41467-024-50128-z [DOI] [PMC free article] [PubMed]
  • 32.Seitzer, P., Bennett, B., & Melamud, E. (2022). MAVEN2: An updated Open-Source mass spectrometry exploration platform. Metabolites, 12. 10.3390/metabo12080684 [DOI] [PMC free article] [PubMed]
  • 33.Wang, L., Xing, X., Chen, L., Yang, L., Su, X., Rabitz, H., Lu, W., & Rabinowitz, J. D. (2019). Peak annotation and verification engine for untargeted LC-MS metabolomics. Analytical Chemistry, 91, 1838–1846. 10.1021/acs.analchem.8b03132 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Schober, A. F., Mathis, A. D., Ingle, C., Park, J. O., Chen, L., Rabinowitz, J. D., Junier, I., Rivoire, O., & Reynolds, K. A. (2019). A Two-Enzyme adaptive unit within bacterial folate metabolism. Cell Rep, 27, 3359–3370e7. 10.1016/j.celrep.2019.05.030 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Snyder, L. R., Kirkland, J. J., & Glajch, J. L. (1997). Basics of separation. Practical HPLC method development (2nd ed., pp. 21–58). Wiley-Interscience.
  • 36.Iavarone, A. T., Jurchen, J. C., & Williams, E. R. (2001). Supercharged protein and peptide ions formed by electrospray ionization. Analytical Chemistry, 73, 1455–1460. 10.1021/ac001251t [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Wang, K. J., Lai, P., & Li, S. S. (2004). High resolution high-performance liquid chromatography separation of polyglutamic acids. Analytical Biochemistry, 332, 199–201. 10.1016/j.ab.2004.05.042 [DOI] [PubMed] [Google Scholar]
  • 38.Sybesma, W., Starrenburg, M., Kleerebezem, M., Mierau, I., De Vos, W. M., & Hugenholtz, J. (2003). Increased production of folate by metabolic engineering of Lactococcus lactis. Applied and Environment Microbiology, 69, 3069–3076. 10.1128/AEM.69.6.3069-3076.2003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Hjortmo, S., Patring, J., & Andlid, T. (2008). Growth rate and medium composition strongly affect folate content in Saccharomyces cerevisiae. International Journal of Food Microbiology, 123, 93–100. 10.1016/j.ijfoodmicro.2007.12.004 [DOI] [PubMed] [Google Scholar]
  • 40.Su, X., Lu, W., & Rabinowitz, J. D. (2017). Metabolite spectral accuracy on orbitraps. Analytical Chemistry, 89, 5940–5948. 10.1021/acs.analchem.7b00396 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Chen, L., Ducker, G. S., Lu, W., Teng, X., & Rabinowitz, J. D. (2017). An LC-MS chemical derivatization method for the measurement of five different one-carbon States of cellular tetrahydrofolate. Analytical and Bioanalytical Chemistry, 409, 5955–5964. 10.1007/s00216-017-0514-4 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material 1 (2.4MB, docx)

Data Availability Statement

Data is provided within the manuscript or supplementary information files. Raw data files are available from the corresponding author on reasonable request.

Articles from Metabolomics are provided here courtesy of Springer

RESOURCES