Analysis of natural and synthetic folates in pharmaceuticals and foods: a review

Abstract

Folate (vitamin B₉) plays an important role in cell division and tissue growth by participating in the amino acid metabolism and nucleotide metabolism. Folate deficiency among pregnant women and children is common around the world, particularly in developing countries. Accurate folate analysis is critical for quality control in pharmaceutical and food products. This review briefly introduces the physiochemical properties, folate deficiency and food fortification of synthetic and natural folates, and the emphases are focused on the progress of sample pretreatment and analysis methods of these compounds in pharmaceuticals and foods in the last decade.

---

1 Introduction

Folate is a generic term for compounds sharing a common chemical backbone and the same vitamin activity. It acts as a coenzyme in several single-carbon transfer reactions, and thus plays an important part in numerous metabolic processes like the synthesis of DNA and RNA.¹ A few authors have summarized research progress on natural and synthetic folates recently,^2–5 mainly focusing on low folate status and related diseases,² folate bioavailability and biofortification,³ electrochemical determination of folic acid,⁴ and unsubstituted food folates.⁵

In this review, we will systematically summarize the recently reported sample treatment and analysis methods for natural and synthetic folates in pharmaceuticals and foods. The classification, physiochemical properties of folate and its deficiency and food fortification will briefly be discussed.

1.1 Classification

Folates include naturally occurring folates and synthetic folic acid in supplements and fortified foods.⁶ Natural folates exist in different forms that vary in both their oxidation state and the carbon group linked to the N5 and N10 positions of the pteridine ring.⁷ Based on the literature, common natural folates are grouped into 5-methyl-tetrahydrofolate (5-CH₃-THF), formyl folates and unsubstituted folates. According to the oxidation states of the pteridine moiety, unsubstituted folates mainly consist of three types: fully oxidized folic acid (FA), reduced 7,8-dihydrofolate (DHF) and 5,6,7,8-tetrahydrofolate (THF).⁵ Formyl folates include 5-formyl-tetrahydrofolate (5-HCO-THF) and 10-formyl-tetrahydrofolate (10-HCO-THF) as well as their interconversion products such as 5,10-methenyl-tetrahydrofolate (5,10-CH⁺=THF), 5,10-methylene-tetrahydrofolate (5,10-CH₂-THF), and 5-formimino-tetrahydrofolate (5-CHNH-THF).⁸ Natural folates occur widely in green leafy vegetables, citrus fruit, egg yolk, liver, certain berries, nuts, beans, cereals, wheat flour and yeast.

In 1931, Wills found a biological factor from yeast. This substance cured a megaloblastic anemia. In 1940, Snell and Peterson extracted one biological factor from spinach leaf and they named it “folic acid”. This name has been used ever since, but today folic acid (FA) only refers to the synthetic form of folate. FA was synthesized for the first time in 1945 by Shane and Carpenter.⁹ FA is industrially produced through chemical synthesis. China is the main producer of synthetic folic acid with a capacity of ca. 3000 metric tons per year. Nowadays, FA is widely used in various pharmaceuticals and fortified foods, such as flour, bread, fruit juices and infant food.²

1.2 Chemical structures and physiochemical properties

Various folates have a similar chemical structure. FA is composed of a pteridine ring, p-aminobenzoic acid (p-ABA) and L-glutamic acid. The chemical structures of folates are displayed in Fig. 1. Changes of chemical structures cause some differentiation of FA from natural folates. The pteridine moiety of FA exists in a fully oxidized state so that FA is much more stable than natural folates. Furthermore, natural folates often contain additional glutamate residues, which reduce their bioavailability.³

Fig. 1 Chemical structures of folates. A polyglutamyl tail is usually attached via the γ-carboxyl group of the glutamate moiety in natural folates. (A) Folic acid (FA); (B) 7,8-dihydrofolate (DHF); (C) tetrahydrofolate (THF) and various one-carbon substituents of THF.

The water solubility of FA is 1.6 mg L⁻¹ at 25 °C and it is slightly soluble in methanol, less in ethanol and butanol, and insoluble in acetone, chloroform, ether and benzene. FA and natural folates are unstable in acidic solution and relatively stable in a neutral to basic medium. They are sensitive to ultraviolet light, air and heat. Suitable antioxidants, low temperature and a nitrogen atmosphere can slow down their degradation process.

1.3 Folate deficiency

Folate is an essential nutrient for humans. For folate deficiency, the most classical disease is megaloblastic anemia. Insufficient intake of folate in early pregnancy may increase the risks of neural tube defects (NTDs), low birth weight, cleft palate and other birth defects. The symptoms for folate deficiency also include canker sores, sore tongue, headaches, heart palpitations, irritability, and behavioral disorders. Furthermore, folate deficiency could increase plasma homocysteine concentrations, the frequency of chromosomal breaks and causes disruption of DNA synthesis, repair and methylation.¹⁰ Hence, it may increase the incidences of cardiovascular disease, stroke and some cancers.¹¹ However, the evidence is insufficient, and further research is also needed.

1.4 Recommended dietary allowances (RDA)

Folate can't be self-synthesized by the human body and must be supplied from the diet. Humans can obtain folate from natural foods, FA fortified foods and some pharmaceuticals such as FA tablets and multivitamin tablets. One dietary folate equivalent (DFE) is defined as 1 μg food folate. The bioavailability of natural folate is lower than that of FA. Thus, one DFE is equivalent to 0.6 μg FA from fortified foods or dietary supplements consumed with foods. The U.S. National Institutes of Health (NIH) suggested that folate RDA for adults was 400 μg DFE.¹² Pregnant women need more folate because of the higher metabolic rate, greater DNA synthesis and rapid cell division.¹³ Hence, pregnant women should ingest 600–1000 μg DFE per day through folic acid tablets and foods to reduce the risk of birth defects.

In 2015, the World Health Organization (WHO) published a guideline for folate. It says that high folic acid intake has not been reliably shown to be associated with negative health effects.¹⁴ However, a recent study reported that excessive FA intake may be associated with vitamin B₁₂ deficiency, lower natural killer cell activity, higher insulin resistance and higher risk of unilateral retinoblastoma, breast cancer and other cancers.¹⁵ As yet, little research has been done in this area.

1.5 Food fortification

Poor folate intake leading to folate deficiency is common around the world, especially in developing countries. Since the 1990s, the United States and Canada firstly introduced mandatory fortification of cereal-grain products with FA in a concentration of 1.4 and 1.5 parts per million, respectively.¹⁶ Nowadays, the prevalence of folate deficiency is decreased in more than 70 countries by supplying FA fortified foods. Epidemiological studies have led to the identification of FA as a primary prevention strategy for NTDs.¹⁷ Red cell folate levels can reflect folate status perfectly. According to the statistics of the United Nations International Children's Emergency Fund (UNICEF), the serum folate level is on the rise. Today, there is virtually no folate deficiency anaemia in North America. In Canada, the incidence of neuroblastoma has fallen by more than 60% since the country began using FA fortified flour.¹⁸ Ocampo et al.¹⁹ analyzed the data from MotherToBaby cohort studies in the United States and Canada, and found that the use of folic acid-containing supplements might reduce the risk of gestational hypertension (GH) and preeclampsia (PE) for pregnant women. What's more, along with the increase of the duration of FA use came a lower risk for GH and PE. Statistical data are rather sparse about folate status in developing countries.

Along with extensive use of fortified foods, the safety of synthetic FA in fortified foods has become an increasing concern recently. It's reported that FA may have adverse effects including vitamin B₁₂ deficiency, changing DHF reductase enzyme activity and inducing cancer under certain conditions.²⁰ A good alternative would be to increase the natural folate level in food crops. Vitamin bio-fortification strategies using food grade microorganisms have been reported in recent years.²¹ A study by Laiño and colleagues²⁰ demonstrated, for the first time, that milk fermented with selected strains of lactic acid bacteria can improve folate status and prevent folate deficiency. Among lactic acid bacteria, S. thermophilus is regarded as an important industrial dairy starter to synthesize folates. Meucci et al.²² studied 50 strains of S. thermophilus, and they found that only 14 strains were able to grow and accumulate different amounts of FA in a culture medium without folates.

2 Sample pretreatment

The precise analysis of folates is complicated because of their low level in samples, multiplicity of folate forms, their instability to chemical and physical factors and complex matrix interference of food.²³ Different methods used in pretreatment can influence the quantitative accuracy. The choice of the sample pretreatment method depends on sample matrixes and the following detection methods. There is no standard pretreatment method for folate analysis. Generally, sample pretreatment involves several steps: extraction of folates from the matrix; incubation with α-amylase, protease, and folate conjugase; purification and concentration⁶ (Fig. 2).

Fig. 2 Common sample pretreatment procedures for folate analysis in pharmaceuticals and food.

2.1 Sample extraction

Several types of buffers such as phosphate,²⁴ ammonium acetate²⁵ and 4-(2-hydroxyethyl)piperazine-1-ethanesulfonic acid/2-(cyclohexylamino)ethanesulfonic acid (HEPES/CHES)²⁶ at pH 4.5–7.5 have been used in folate extraction. The parameters such as the sample amount, the volume, the concentration and pH of the buffer, heating temperature and time could affect the extraction efficiency. In some food samples, part of folate is often bound to a protein. Heat treatment is usually used to liberate the bound form of folate from the protein.²⁷ However, high temperature may increase folate degradation. As an alternative to heat treatment, high hydrostatic pressure (HHP) is an emerging and non-thermal technique used in the food industry for high-quality and safe food products.²⁸ HHP is also used in sample pretreatment to improve extraction efficiency. As shown by Nassim's group,²⁹ HHP can dissolve non-soluble granules in egg yolk and improve the recovery of 5-CH₃-THF from egg yolk. The parameters of HHP were set to 400 MPa for 5 min. Under these conditions, the folate degradation was negligible.

2.2 Enzyme treatment

The choice of enzymes depends on the sample matrices and the subsequent detection methods. When the sample contains protein and starch, α-amylase and protease are used to release folates that are trapped in carbohydrates and proteins. FA interacts with α-amylase and protease by multi non-covalent interactions including hydrophobic interactions, hydrogen bonding or electrostatic interactions.³⁰ Natural folates are often in polyglutamyl forms, and folate conjugase can break down natural polyglutamyl folates into mono- or di-glutamate forms. Rat serum or plasma, porcine kidney, chicken pancreas and human plasma are the major sources of folate conjugase.²⁷ Nevertheless, when the sample is fortified food, the folate conjugase is not always required.

Tri-enzyme treatment using chicken pancreas conjugase, α-amylase, and protease was first advocated for microbiological assay of total folate in foods in 1990.³¹ This procedure is particularly effective for cereal-based and milk-based foods. Motta et al.³² used a tri-enzymatic extraction technique for extraction of five folates in quinoa, amaranth and buckwheat. LC-MS/MS was applied to the determination of the target folates with satisfactory results. As they said, the detection method is also a key factor that influences the selection of enzymes. Czarnowska-Kujawska et al.²³ used HPLC for folate analysis in vegetables and fruit samples, and compared the effectiveness of two commonly used folate conjugases from porcine kidney and rat plasma. Rat plasma was chosen as the source of folate conjugase in this research. Chandra-Hioe and coworkers²⁷ compared enzymatic treatments to release FA and endogenous 5-CH₃-THF from infant milk with enzyme-free heat extraction. The results indicated that enzyme-free heat extraction of folate was more compatible with UPLC-MS/MS than enzymatic treatment.

2.3 Purification and concentration

The sample matrix interference should be eliminated before determination because the folates are at a very low level in the extraction solution. Many techniques including solid phase extraction (SPE), solid phase microextraction (SPME), dispersive liquid–liquid microextraction (DLLME) and affinity chromatography have been reported for purification and concentration of folates in sample solutions.

2.3.1 Solid phase extraction (SPE). SPE is a popular sample pretreatment technique integrating separation, purification and concentration. The typical steps of SPE are as follows: (1) SPE column equilibration; (2) sample loading; (3) washing; (4) analyte elution. There are some reports regarding the SPE technique, most of which utilize strong anion exchange (SAX)²³ and C18 (ref. 33) as SPE sorbents. SAX^23,24,26,34 is the most frequently used technique with satisfactory recoveries for different folate forms. However, the enriching factor of SPE is usually low in reported literature. Some novel sorbents have been used in SPE techniques in recent years, such as magnetic molecularly imprinted nanoparticles (MMIPs),³⁵ styrene divinylbenzene,³⁶ and weak anion exchange (WAX).³⁷ Hussain et al.³⁵ used MMIPs as a SPE sorbent for extraction of FA in orange and spinach samples. The recoveries in spiked samples at different concentration levels were 95–104% for orange and 99.5–102.5% for spinach. The relative standard deviations (RSDs) of the recoveries were less than 0.5%. The presented MMIP material had the advantages of simplicity and good selectivity compared with other SPE sorbents. On the whole, SPE is a good way to isolate analytes from various matrices with good recovery. However, SPE has some drawbacks such as being time-consuming, a tedious procedure, requirement of large amounts of organic solvents, low selectivity and unsatisfactory enriching factors.

2.3.2 Solid phase microextraction (SPME). SPME is one of the most widely-used microextraction techniques that permit integration of sampling, sample clean-up, and pre-concentration into a single step for chemical analysis.³⁸ It is simple, cost-effective and time-saving. Prasad et al.³⁹ reported a consolidated molecularly imprinted micro-solid phase extraction (MIMSPE) technique for ultratrace analysis of FA. This method is highly sensitive because of a high enrichment factor. The ratio between the volume of the sample and solid phase was approximately 530. The LOD was as low as 0.0036 ng mL⁻¹ and the RSD was 0.13%. But this method has not been applied to FA analysis in pharmaceutical and food samples.

2.3.3 Dispersive liquid–liquid microextraction (DLLME). DLLME was first introduced in 2006 for the determination of polycyclic aromatic hydrocarbons (PAHs) in water samples by using gas chromatography.⁴⁰ DLLME is simple, cheap, environmentally friendly, and could offer high enrichment factors of a wide gap between acceptor and donor phases.⁴¹ Nojavan et al.⁴² developed ion-pair based DLLME for simultaneous extraction of five folate derivatives in wheat flour, egg yolk and orange juice samples. 1-Octanol and ethanol were chosen as the extraction solvent and the disperser solvent, respectively. 0.01% (w/v) cationic surfactant at pH 11 was employed to obtain the maximum response for the extraction and determination of folate derivatives. The sensitivity of this method is satisfactory because of the high enrichment factors of 108–135.

2.3.4 Affinity chromatography. Affinity chromatography is a technique for separating and purifying analytes from sample matrices based on a highly specific and sensitive interaction between an antigen and antibody, an enzyme and a substrate, or a receptor and ligand. Saini et al.⁴³ detected folate in a food plant sample named Moringa oleifera L. using HPLC. The sample was purified by immunoaffinity chromatography using folate binding proteins. However, it needed folate-binding proteins which have not been commercialized. And it is more expensive and time-consuming than other purification techniques.²⁶ Furthermore, folate-binding proteins exhibit low affinity to some forms of folates such as 5-HCO-THF.

Folate and FA and its derivatives are prone to be oxidized during the sample pretreatment. A dark place, lower temperature and neutral to alkaline environment can slow the oxidative process. What's more, antioxidants are also important in resisting oxidation. Ascorbic acid,⁴⁴ sodium ascorbate,⁴⁵ β-mercaptoethanol,⁴⁶ 2,3-dimercapto-1-propanol²⁴ and 1,4-dithiothreitol³⁴ are commonly used antioxidants. Sometimes, two or more antioxidants are used together to have a better antioxidant effect. Nitrogen can also prevent folates from oxidation.

3 Analysis methods for folates

A variety of methods have been developed for the separation and quantification of FA and its analogues in different matrices, including microbiological assay (MA), bio-specific procedures, high-performance liquid chromatography (HPLC), liquid chromatography-tandem mass spectrometry (LC-MS/MS) and capillary electrophoresis (CE). Other techniques such as various sensors, spectrophotometry⁴⁷ and a flow injection chemiluminescence method⁴⁸ have been reported too.

3.1 Microbiological assay (MA)

Some microorganisms require folate as a nutritional factor for their growth and multiplication, which they are unable to synthesize by themselves. These microorganisms are used for quantitative folate analysis. Microbiological assay was firstly used for the determination of folate using Lactobacillus casei in 1968.¹⁴ MA had been a standard method of folate analysis over the past few decades. To date, most of the available data about folate content in foods have been obtained by MA. MA is sensitive to various forms of folate with bioactivity. Along with continuously going deeper into research, some new MA methods were established during recent years. Fajardo et al.⁴⁹ reported a 96-well microtiter plate MA method for the determination of the total folate in 21 fresh-cut vegetables and fruits using chloramphenicol-resistant Lactobacillus casei subspecies rhamnosus (NCIMB 10463). The detection range of FA was 0–30 pg/100 μL and the recoveries were 85–110%. The sensitivity and reproducibility were improved, compared with traditional MA methods. However, MA can't distinguish the different forms of folate, so it can only analyze the total folate in the samples.

3.2 Bio-specific assay

Bio-specific assay can be classified into two types, i.e. the assays based on the use of naturally occurring vitamin binding proteins with either radiolabels or enzyme labels; and those based on the specific interaction of an antibody with its antigen, e.g. enzyme linked immunosorbent assay (ELISA).⁵⁰ As for the present, much of the bio-specific assays are used in clinical diagnosis rather than food analysis. The national health and nutrition examination survey (NHANES) of U.S. monitored the folate status of U.S. populations from 1998–2010. The Bio-Rad radioassay (BR) was used to measure serum and RBC concentrations from 1988–2006, and MA was used from 2007–2010. The latter is considered to be more accurate.⁵¹ Most bio-specific assays are tedious, time-consuming and unstable. Immunoassay has been also applied in some biosensors recently.

3.3 High performance liquid chromatography (HPLC)

The limitation of MA in analysis of folate derivatives has promoted the use of liquid chromatographic techniques.⁵⁰ High performance liquid chromatography (HPLC) has become a popular method for analysis of folates in the past decade. The HPLC methods for analysis of folates in pharmaceuticals and foods are shown in Table 1. Various detectors including ultraviolet detectors (UVDs),^35,42 diode array detectors (DADs),^52,53 electrochemical detectors (ECDs)⁵⁴ and fluorescence detectors (FLDs)^23,26,29 have been used. An UVD responds to all folates, but the UVD lacks the sensitivity to detect folates at the trace levels required. As a result, a FLD is a better choice for analyzing folates in complex matrices. Czarnowska-Kujawska et al.²³ reported an HPLC method with fluorometric detection for analysis of 5-CH₃-THF and THF in fresh fruits and vegetables with the LOQs of 0.4–0.6 ng mL⁻¹. A SPE cartridge using strong anion exchange was used for sample purification and the recoveries were between 92% and 105%. THF and 5-CH₃-THF are natively fluorescent. In contrast, FA is a weakly fluorescent molecule and it needs fluorescence derivatization. Tornero et al.²⁶ developed a HPLC-FLD method for simultaneous determination of FA, THF and 5-CH₃-THF in cereals, spinach and tomatoes. A post-column on-line photoderivatization was used for FA analysis. The LODs of FA, THF and 5-CH₃-THF were 7.47 ng mL⁻¹, 15.3 ng mL⁻¹, 22.9 ng mL⁻¹, respectively. Wang et al.⁵⁵ developed ultra-performance liquid chromatography (UPLC) coupled with ion chromatography (IC) to simultaneously determine folic acid and inorganic anions in folic acid tablets. As for folic acid, the LOD was 0.0061 mg L⁻¹, and the RSDs for peak area and retention time were 0.49% and 0.14%, respectively.

Table 1 The HPLC methods for analysis of folates in pharmaceuticals and foods

Analytes	Sample matrix and preparation	Analytic conditions	Method performances	Ref.
FA, 5-CH3-THF, 5-HCO-THF, THF, 10-HCO-THF, 10-HCO-DHF	Two fruits and six vegetables; extracted, enzyme-treated (α-amylase, rat plasma), and SPE (SAX)	Column: phenomenex synergi C18 (250 × 4.6 mm, 4 μm); mobile phase: acetonitrile (ACN) : 30 mmol L^−1 phosphoric acid buffer (pH 2.3); detector: UVD (290 nm), FLD: ex 290 nm and em 360 nm	LOQ = 0.4 (5-CH3-THF); 0.6 (THF) ng mL^−1; range = 0.3–66.3 (5-CH3-THF), 0.4–55.7 (THF) ng mL^−1; recovery = 109% ± 9 (5-CH3-THF); 93% ± 8 (THF)	23
FA, THF, 5-CH3-THF, 10-HCO-THF, 5-HCO-THF and pteroic acid	New Zealand spinach; extracted, heated, centrifuged, and filtered	Column: C18 (100 × 2.0 mm, 5 μm); mobile phase: 100 mmol L^−1 potassium phosphate buffer (pH 2.0) : methanol (85 : 15, v/v); detector: DAD	LOD = 1.33–36.2 mg/100 g	25
FA, THF, 5-CH3-THF	Cereals, spinach, tomatoes; extracted, enzymatically deconjugated (rat plasma), and SPE (SAX)	Column: C8 (50 × 4.6 mm, 2.7 μm); mobile phase: 2 mmol L^−1 formic acid (pH 4) : ACN (6 : 94, v/v); detector: fast scanning fluorimetric detector (FSFD) (ex 280 nm and em 360 nm for THF and 5-CH3-THF and 440 nm for FA)	LOD = 7.47; 15.3; 22.9 ng mL^−1; range = 50–1000; 100–1000; 100–1000 ng mL^−1; RSD (intra-day) = 5.28%; 1.28%; 1.36% (n = 5); recovery = 92.73–105.0%	26
FA	Spinach, orange; extracted, filtered, and SPE (10 mg MMIPs)	Column: C18 (250 × 4.6 mm); mobile phase: ACN : water (90 : 10, v/v); detector: UVD (280 nm)	Recoveries = 95–104% (orange), 99.5–102.5% (spinach); RSD (for recoveries) < 0.5%	35
5-HCO-THF, 10-HCO-THF, 5-CH3-THF, 5,10-CH2-THF, FA	Wheat-flour, egg yolk, and orange juice; extracted, filtered, and IP-DLLME (extraction solvent: 1-octanol, disperser solvent: ethanol)	Column: ODS (250 × 4 mm, 5 μm); mobile phase: phosphate buffer (pH 3.5) : ACN (90 : 10, v/v); detector: UVD (280 nm)	LOD = 2–4.1 ng g^−1; range = 1–200 ng g^−1; RSD = 5.2–7.4%; recoveries = 98–102%	42
FA, VC and five other water-soluble vitamins B	Vitamins with mineral tablets (VMT); diluted, ultrasound, and filtered	Column: C18 (250 × 4.6 mm, 5 μm); mobile phase: 50 mmol L^−1 ammonium dihydrogen phosphate buffer (pH 3.0) : ACN (85 : 15, v/v for FA); detector: DAD (280 nm for FA)	LOD = 0.010 mg L^−1 (FA); range = 2.952–8.856 mg L^−1 (FA); RSD = 1.37% (intra-day) and 1.75% (inter-day) (FA); recoveries = 97.95–99.51% (FA)	52
FA, VC and four other water-soluble vitamins B	Pharmaceutical preparations; centrifuged	Column: C18 (150 × 3.0 mm, 5 μm); mobile phase: 0.1% formic acid (pH 2.6) : ACN; detector: DAD (285 nm)	Range = 0.6–12 μg mL^−1 (FA); RSD < 2.06% (FA); recoveries = 99.13–100.48% (FA)	53
FA and five inorganic anions	Folic acid tablets; dissolved, ultrasound, diluted, and filtered	Column: C18 (150 × 4.6 mm, 2.6 μm); mobile phase: methanol : deionized water (3 : 97, v/v); detector: UVD (280 nm)	LOD = 0.0061 mg L^−1 (FA); range = 0.1–10.0 mg L^−1 (FA); RSD = 0.14% (retention time) and 0.49% (peak area) (FA); recovery = 96.4–105.8% (FA)	55
FA	Fortified cereals and juices; extracted, centrifuged, filtered	Column: Diamond Hydride™ (75 × 4.6 mm, 4.2 μm); mobile phase: (A) H2O + 10 mmol L^−1 ammonium formate, (B) 90 : 10 acetonitrile : H2O + 10 mmol L^−1 ammonium formate; detector: VWUVD (284 nm)	Range = 0.1–2.0 mg L^−1; RSD = 0.02–0.14% (samples), 0.02–0.08% (standard solutions); recoveries = 90%	82
5-CH3-THF	Nutritional products; centrifuged	Column: C18 analytical column (250 × 4.6 mm, 5 μm); mobile phase: 0.05 mol L^−1 KH2PO4 (pH 2.9): ACN; detector: FLD	LOD = 10 μg kg^−1; RSD = 1.1%; recovery = 92.8%	83

---

As for HPLC and UPLC, some detectors by themselves are not confirmatory such as UVDs. Some auxiliary methods are used for confirmatory analysis. For example, a DAD detector can provide a 3-D scanning spectrum which can be used for qualitative analysis. However, the most efficient method for confirmatory analysis is MS/MS detection.

3.4 LC-MS/MS analysis

Although HPLC holds a high position for analysis of folates in foods and pharmaceuticals, it has some limitations. A strong matrix effect in complex samples influences the precision of analysis. Furthermore, conventional LC detectors are unable to detect the very low level of folates and FA in the samples. LC (UPLC) coupled with MS/MS has very high separation efficiency as well as strong identification ability. Therefore, the application of UPLC-MS/MS in the analysis of FA and folates in foods and pharmaceuticals increased sharply in recent years. In AOAC Method 2011.06 (ref. 37) and AOAC Method 2013.13 (ref. 34), UPLC coupled to a triple quadrupole mass spectrometer is applied to the determination of total folates in infant formulas and adult nutrients. Shohag et al.²⁴ established an UPLC-tandem quadrupole mass spectrometry method for the determination of seven folates in plant leaves within five minutes of running time for the first time. A SPE (SAX) technique was used for sample purification after the tri-enzyme treatment. The range was 0.1–60 μg/100 g. The recoveries were 71.27–99.01%, and the total folate in four plant leaves were 117.45–223.74 μg/100 g fresh weight (FW). Stable isotopes such as ¹³C-labeled (¹³C₅) isotopes^27,34 and deuterium-labeled (²H₄) isotopes⁵⁶ of folates are usually used as internal standards (ISs) to improve the accuracy in MS analysis. Methotrexate has also been used as an IS for quantification of five folates.⁵⁷Table 2 shows the LC-MS/MS methods for the determination of folate in pharmaceuticals and foods.

Table 2 The LC(UPLC)-MS/MS methods for analysis of folates in pharmaceuticals and foods

Analytes	Sample matrix and preparation	Analytic conditions	Method performances	Ref.
THF, 5-CH3-THF, 5-HCO-THF, 5,10-CH^+=THF, FA, FA-glu3 10-HCO-THF	Plant leaf; extracted (tri-enzyme), centrifuged, filtered, and SPE (SAX)	Column: ACQUITY UPLC BEH C18 column (50 mm × 2.1 mm, 1.7 μm); mobile phase: 0.1% formic acid in water and 0.1% formic acid in acetonitrile; detector: tandem quadrupole mass spectrometer	LOD = 0.003–0.021 μg/100 g FW; range = 0.1–60 μg/100 g; RSD = 1.7–7.8%; recoveries = 71.27–99.01%	24
FA, 5-CH3-THF	Infant milk formulae and adult nutritionals; heated, centrifuged, filtered, and SPE (styrene divinylbenzene)	Column: waters Acquity HSS T3 (100 × 1.0 mm, 1.8 μm); mobile phase: 0.1% (v/v) formic acid and ACN; detector: TSQ quantum access tandem quadrupole mass spectrometer (Thermo-Scientific, San Jose, CA) (FA: m/z 442 → 295; 5-CH3-THF: m/z 460 → 313); internal standard: [^13C5]-FA and [^13C5]-5-CH3-THF	LOD = 1.4 (FA) and 7.5 (5-CH3-THF) ng mL^−1; range = 3.1–1400 (FA) and 16.2–1400 (5-CH3-THF) ng mL^−1; RSD = 6% (intra-day), 8% (inter-day); recoveries = 85% (FA) and 95% (5-CH3-THF)	27
FA, 5-CH3-THF, THF, 5-HCO-THF, 10-HCO-THF	Quinoa, amaranth, and buckwheat; tri-enzyme treatment (α-amylase, protease and rat serum), centrifuged, and filtered	Column: HSS T3 (150 × 2.1 mm, 1.8 μm); mobile phase: 0.1% formic acid solution in water and 0.1% formic acid solution in ACN; detector: triple quadrupole mass spectrometer (Waters ACQUITY® TQD, Waters Co., Milford, USA); internal standard: (^13C5) FA and (^13C5) 5-CH3-THF	LOD = 0.56–1.02 μg/100 g; range = 7–100 ng mL^−1; RSD < 10%	32
FA, 5-CH3-THF	Standard reference materials (SRMs): infant/adult nutritional formula, fortified breakfast cereal, whole milk powder, and whole egg powder; tri-enzyme treatment (α-amylase, protease and rat serum), and SPE (C18)	Column: Agilent Zorbax SB-C18 (150 × 2.1 mm, 3.5 μm); mobile phase: 0.1% formic acid in water and 0.1% formic acid in ACN; detector: triple quadrupole mass spectrometer	RSD < 8%	33
FA, 5-CH3-THF	Infant formula and adult/pediatric nutritional formula; enzyme-treatment (α-amylase, protease), and SPE (SAX)	Column: HSS T3 (150 × 2.1 mm, 1.8 μm); mobile phase: 0.5% (v/v) acetic acid in water and ACN; detector: triple quadrupole mass spectrometer (Agilent 6460)	LOD = 0.10 (FA) and 0.05 (5-CH3-THF) μg/100 g; range = 0.33–300 μg/100 g; RSD < 5.4%; recoveries = 89–114%	34
THF, 5-CH3-THF, 10-FFA, 5-HCO-THF, FA, FA-glu3, 10-MFA	Infant formula and adult nutritionals; tri-enzyme treatment (α-amylase, protease and rat plasma), and SPE (WAX)	Column: HSS T3 (150 × 2.1 mm, 1.8 μm); mobile phase: 10 mmol L^−1 ammonium formate buffer (pH 2.8); detector: triple quadrupole mass spectrometer (Waters Quattro Premier)	Range = 10–19 100 μg/100 g; RSD = 4.69–13.89%; recoveries = 94.10–101.34%	37
FA, 5-CH3-THF, THF, 5-HCO-THF, 5,10-CH^+=THF	Tomatoes; homogenized, tri-enzyme treatment (α-amylase, protease and rat serum), centrifuged, and filtered	Column: Luna C18 (250 × 4.60 mm, 5 μm); mobile phase: 0.1% (v/v) formic acid in water and ACN; detector: Exactive™ Plus Orbitrap mass spectrometer (Thermo Fisher Scientific, USA)	LOD = 0.15–0.37 ng mL^−1; range = 1–30 (FA, 5-CH3-THF, 5-HCO-THF), 0.5–30 (5,10-CH^+=THF), 2–30 (THF) ng mL^−1; recoveries = 48–131%	57

---

3.5 Capillary electrophoresis (CE)

In the past decade, CE has been applied to analysis of folic acid and its analogues in pharmaceuticals and foods. Various detectors can be coupled to CE, including UVDs,⁵⁸ DADs,⁵⁹ laser-induced fluorescence (LIF) detectors,⁶⁰ chemiluminescence detectors (CLDs)⁶¹ and MS,⁶² which makes CE a powerful analytical tool. Silva and colleagues⁵⁸ developed a CE-UVD method for FA and nine other water-soluble vitamins in food supplements with a detection limit of 0.28 mg L⁻¹. The sensitivity of CE analysis is unsatisfactory because of its short light path and small sample volume. To overcome this limitation, a laser-induced fluorescence detector has been used for analyzing trace-level folic acid. The sensitivity of the LIF detector is about 1000 times higher than that of the UVD. Zhao et al.⁶⁰ reported a CE-LIF method for the simultaneous determination of three essential amino acids and three B vitamins (FA, niacinamide and riboflavin). The assay had a linear range of 0.005 to 0.80 μmol L⁻¹ for FA, with a detection limit of 5.0 nmol L⁻¹. Chemiluminescence detectors (CLDs) are also regarded as one of the most sensitive detectors and furthermore, it does not require a bulky light source.⁶³ Zhao et al.⁶¹ established a CE-CLD method for the detection of FA in pharmaceutical tablets, apple juice, and human urine. Its LOD was 2.0 × 10⁻⁸ mol L⁻¹, and the recoveries were 94.3–105.8%. In general, CE has the advantages of high resolution, high speed, and low cost. However, its main drawbacks are bad reproducibility and unsatisfactory sensitivity. A variety of CE methods for folate analysis are summarized in Table 3.

Table 3 The CE methods for analysis of folates in pharmaceuticals and foods

Analytes	Sample matrix and preparation	Analytic conditions	Method performances	Ref.
FA and nine other water-soluble vitamins	Food supplements (drinks, tablets, powders, and injection solutions); diluted, and filtered	Column: a fused silica capillary (60 cm × 50 μm) and 49.85 cm of effective length; mobile phase: hydroxide solution and water; detector: UVD (214 nm)	LOD = 0.28 mg L^−1 (FA); range = 6–60 mg L^−1; RSD = 1.9% (intra-day) and 6.7% (inter-day); recoveries = 93.57–100.05% (FA)	58
FA	Green, red and mignon lentils; extracted, centrifuged, and filtered	Column: agilent uncoated fused silica capillary (effective lengths 55 cm, 75 μm I.D.); mobile phase: 10 mM sodium tetraborate in 10% (v/v) methanol at pH 9; detector: DAD (200 nm); internal standard: methylparaben	LOD = 6.12 × 10–7 mol L^−1; range = (1.2–4.8) × 10–5 mol L^−1; RSD = 0.65% (inter-day); 0.48–0.79% (intra-day); recoveries = 112%	59
FA, two other water-soluble vitamin B and three amino acids	Four health drink samples; diluted and derivatized	Column: uncoated fused-silica capillaries (60 cm × 50 μm), with an effective length of 50 cm); mobile phase: running buffer (Na2B4O7), NaOH, water solution; detector: LIF (488 nm excitation, 520 nm emission)	LOD = 5.0 nmol L^−1; range: 0.005–0.80 μmol L^−1 (FA); RSD = 1.09% (migration time) and 3.57% (peak area) (FA); recoveries = 93.32% ± 2.58 (FA)	60
FA	FA tablets, apple juice, and human urine; diluted and heated	Column: a 50 cm × 75 μm I.D. uncoated fused silica capillaries; electrophoresis electrolyte: 35 mmol L^−1 sodium borate (pH 9.4) containing 0.8 mM luminol; detector: chemiluminescence detector (CLD)	LOD = 2.0 × 10^−8 mol L^−1; range = 5.0 × 10^−8 to 1.0 × 10^−5 mol L^−1; RSD = 1.5% (peak area) and 1.1% (migration time); recoveries = 94.3–105.8% (FA tablet)	61
FA and eight other water-soluble vitamin B	Pharmaceuticals (one injection solution and two tablets); diluted and filtered	Column: Agilent CZE column (50 μm I.D. uncoated fused silica capillary tube of an 820 mm total length; mobile phase: 50 mmol L^−1 formic acid (pH 2.05); detector: MS/MS	LOD = 0.1715 μg mL^−1 (FA); range = 3.125–50 μg mL^−1; RSD = 4.44% (FA); recoveries = 93.08% (FA)	62
FA	Fortified instant Asian noodles; extracted, enzyme-treated (α-amylase), centrifuged, and filtered	Column: uncoated fused-silica capillaries (75 μm I.D., 50 cm effective length); mobile phase: 8 mM phosphate-12 mmol L^−1 borate run buffer with 5% MeOH at pH 9.5; detector: DAD (214 nm); internal standard: nicotinic acid	LOD = 5.3 mg L^−1; recoveries = 96–103%	84

---

3.6 Sensors

Sensors have been playing an increasingly important role in analysis of FA and its analogues due to its high sensitivity, small size and negligible environmental impact. In recent years, there have been numerous sensors reported, such as luminescent sensors,^64–67 electrochemical sensors,^68–71 and biosensors,^46,72 for the quantification of folates. Luminescent sensors can be classed into photoluminescent sensors, chemiluminescent sensors and electrochemiluminescent sensors. Many materials are used as fluorescent probes in the photoluminescent sensors such as 8-aminonaphthalene-1,3,6-trisulfonate anchored to Zn–Al–CO₃-layered double hydroxide (ANTS-anchored Zn–Al–CO₃-LDH) particles,⁶⁴ polyethylenimine-capped silver nanoclusters (PEI-AgNCs)⁶⁶ and quantum dots (QDs).⁶⁷ Liu et al.⁶⁴ designed a fluorescent sensor based on ANTS-anchored Zn–Al–CO₃-LDH particles. This method was rapid, inexpensive and label-free, and it had excellent sensitivity (LOD = 0.1 μmol L⁻¹) and repeatability (RSD ≤ 3.5%). However, it was only applicable to the analysis of FA in the simulated phosphate buffer solution (pH = 7.4) instead of real samples. Zhang et al.⁶⁶ detected FA in FA tablets, wheat flour, liquid and powdered milk using PEI-AgNCs as a fluorescence probe in the luminescent sensors (LOD = 0.032 nmol L⁻¹) (Fig. 3A). Electrochemiluminescent sensors have also been applied to the folate analysis. Li et al.⁷³ established a rapid and sensitive electrochemiluminescence sensor based on a 3D graphene/CdSeTe/Ru(bpy)₃²⁺-doped silica nanocomposite modified electrode (Fig. 3B). The linear range of FA detection in FA tablets was 1.0 × 10⁻¹¹ mol L⁻¹ to 1.0 × 10⁻⁶ mol L⁻¹, and the LOD was 3.6 × 10⁻¹² mol L⁻¹.

Fig. 3 (A) Chemical structure of PEI and schematic representation of FA detection (reproduced from ref. 66); (B) the preparation procedures for a 3D graphene/CdSeTe/RuSiNP modified electrode (reproduced from ref. 73).

Electrochemical techniques are much more sensitive and easy to use, and offer low detection limits compared to conventional methods.⁴ Electrochemical impedance spectroscopy (EIS) has been used for folate analysis and electrochemical characterization of electrodes such as square wave voltammetry (SWV),⁶⁸ differential pulse voltammetry (DPV),⁷⁴ cyclic voltammetry (CV),⁷⁵ and linear sweep voltammetry (LSV).⁷⁶ In order to improve selectivity and sensitivity for folate analysis, many chemically modified electrodes (CMEs) have been used in electrochemical sensors. Numerous nanomaterials such as nanowires,⁷¹ nanoparticles,⁷⁷ nanotubes⁷⁶ and nanocomposites⁷⁸ are used as modifiers for electrodes. Recently, room temperature ionic liquids (RTILs)^79,80 have been applied in chemically modified electrodes because of their higher ionic conductivity, wider electrochemical windows and environmental friendliness. Jamali et al.⁷⁹ developed a sensitive nanosensor based nanoalloy (Pt:Co) room temperature ionic liquid (RTIL) modified carbon paste electrode for analysis of FA in fortified juice, FA tablets and mint vegetable. The LOD was 4.0 × 10⁻⁸ mol L⁻¹, and the RSD was 0.94%.

Biosensors also provide a new choice in folate analysis. The biosensor incorporating surface plasmon resonance (SPR) offers an excellent platform for the detection of folate in pharmaceuticals and foods. Indyk⁸¹ developed an optical biosensor assay utilizing folate-binding proteins. In his study, the cross-reactivity of folate-binding proteins for both FA and 5-CH₃-THF facilitated a more reliable estimate of the total folate in samples. It may provide a precise and accurate alternative to conventional methods for folate routine analysis in milk-based products. A pencil graphite electrode modified with salmon sperm ds-DNA was used in an electrochemical biosensor for FA analysis by Mirmoghtadaie and co-workers.⁷² Biosensors have a higher selectivity and specificity for bio-active substances than other sensors in complex matrices. Nevertheless, sensors have some drawbacks such as poor reproducibility, difficulty in simultaneous determination and complex packaging techniques, which limit their wide application. Table 4 shows the various types of sensors for analysis of folates in pharmaceuticals and foods.

Table 4 The sensors for analysis of folates in pharmaceuticals and foods

Analytes	Sample matrix and preparation	Detection techniques and employed materials	Method performances	Ref.
FA	Infant formula and adult/pediatric nutritional formula; enzymatic and heat treatments	SPR optical biosensor assay; exchangeable carboxymethyldextranfunctionalized chip using folate binding protein (FBP)	LOD = 6.5 μg/100 g; RSD = 3.48% (intra-day) and 4.63% (inter-day)	46
FA	Aqueous media	Fluorescence detection (ex 355 nm and em 506 nm); ANTS-anchored Zn–Al–CO3-LDH particles	LOD = 0.1 μmol L^−1; range = 1–200 μmol L^−1; RSD = 0.7–3.4%; recoveries = 97.4–104.0%	64
FA	Pharmaceutical tablets	Fluorescence detection; BSA-stabilized gold nanoclusters (AuNCs)	Range = 120.0 ng mL^−1 to 33.12 μg mL^−1; LOD = 18.3 ng mL^−1; RSD < 4%; recoveries = 93.5–95.7%	65
FA	FA tablets, wheat flour, liquid milk and milk powder, urine; centrifuged	Fluorescence detection (ex 375 nm and em 452 nm) PEI-AgNCs	LOD = 0.032 nmol L^−1; range = 0.1 nmol L^−1 to 2.75 μmol L^−1; RSD < 3.5%; recoveries = 96.0–105.0%	66
FA	Aqueous media	Fluorescence detection; copper- or manganese-doped ZnS QDs	LOD = 11 μmol L^−1	67
FA and vanillin	Chocolate, biscuit and coffee milk; sonicated, centrifuged, and diluted	Electrochemical detection (SWV); cadmium oxide nanoparticle decorated with single wall carbon nanotubes (CdO/SWCNTs) and 1,3-dipropylimidazolium bromide (DPIB) as a binder (CPE/CdO/SWCNTs/DPIB)	LOD = 0.06 ± 0.01 μmol L^−1 (FA); range = 0.1–1200 μmol L^−1 (FA)	68
FA and 6-thioguanine	Folic acid tablets, 6-thioguanine tablets and urine samples	Electrochemical detection (SWV); 2-chlorobenzoylferrocene/ZnO–CuO nanoplate modified carbon paste electrode (2CBFZCCPE)	Range = 15–2000 μmol L^−1; RSD = 1.8–3.4% (FA); recoveries = 97.4–103.3% (FA)	69
FA, ascorbic acid and uric acid	B complex tablets, FA and human urine; centrifuged and filtered	Electrochemical detection; Mn–SnO2 nanoparticle modified glassy carbon electrode	Range = 1–500 μmol L^−1 (FA); LOD = 25 nmol L^−1 (FA); RSD = 3.51% (FA); recoveries = 99.4–102.3%	70
FA	FA tablets; dissolved and filtered	Electrochemical detection (SWV); bismuth nanowires (BiNWs) on glassy carbon (GC) substrates	Range = 10–150 nmol L^−1; LOD = 9.53 nmol L^−1; RSD = 2.5% recovery = 92.33%	71
FA	FA tablets, fortified wheat flour, and spinach; centrifuged and filtered	Electrochemical biosensor (DPV); a pencil graphite electrode modified with salmon sperm ds-DNA	Range = 0.1–10.0 μmol L^−1; LOD = 1.06 × 10^−8 μmol L^−1; RSD < 4.6%; recoveries = 98.4–100.8%	72
FA	FA tablets	Electrochemical detection (DPV); nanosized gold and graphene modified carbon ionic liquid electrode	Range = 0.01–50.0 μmol L^−1; LOD = 2.7 nmol L^−1; RSD = 2.1%; recoveries = 97.2–98.4%	74
FA	Vitamin supplement formulations and human serum; vitamin supplement: filtered and diluted	Electrochemical detection (SWV); polypyrrole-α-polyoxometalate-AuNPs (PPy-α-POM-AuNP modified gold (Au) electrode)	Range = 1.0–44.0 nmol L^−1; LOD = 0.12 nmol L^−1; RSD = 5.3% (intra-day), 5.47% (inter-day)	75
FA	FA tablets, oats; centrifuged	Photoelectrochemical detection; carbon nanohorn supported interwoven titanate nanotubes	Range = 1 × 10^−10 mol L^−1 to 5 × 10^−5 mol L^−1; LOD = (2.5 ± 0.005) × 10^−11 mol L^−1; recoveries = 96.1–101.5%	76
FA, VC	Five pharmaceutical tablets, blood serum, and lemon juice; filtered and diluted	Electrochemical detection (DPV); magnetite (Fe3O4) nanoparticle modified carbon paste electrode (Fe3O4-CME)	Range = 6.50 × 10^−8–9.80 × 10^−5 mol L^−1 (FA); LOD = 2.01 nmol L^−1 (FA); RSD = 2.23% (intra-day), 1.49% (inter-day) (FA); recoveries = 98.9–99.5% (FA)	77
FA, glutathione, nicotinamide adenine dinucleotide;	Pharmaceutical tablets, human blood, and urine; tablets: filtered and diluted	Electrochemical detection (SWV); carbon paste electrode (CPE) modified ZnO/CNT nanocomposite	Range = 3.0–700 μmol L^−1 (FA); LOD = 1.0 μmol L^−1 (FA); RSD = 2.9%	78
FA	Fortified juice, FA tablets and mint; centrifuged and filtered	Electrochemical detection (SWV); a nanoalloy (Pt:Co) room temperature ionic liquid (RTIL) modified carbon paste electrode	Range = 1.0 × 10^−7–5.0 × 10^−4 mol L^−1; LOD = 4.0 × 10^−8 mol L^−1; RSD = 0.94%	79
FA	Tablets, apple juice, and urine; centrifuged and filtered	Electrochemical detection (SWV); ZnO/nanoparticle (NP) modified carbon ionic liquid paste electrode (ZnO/NP/CILPE)	Range = 0.05–1.5 μmol L^−1 (slope = 1.776 μA μM^−1) and 1.5–550 μmol L^−1 (slope = 0.033 μA μM^−1); LOD = 0.01 μmol L^−1	80
FA	Pharmaceutical products (foligen and folvite); dissolved, sonicated, filtered, and diluted	Electrochemical detection (CV, LSV and chronoamperometry); hydroxyapatite (HA) nanoparticle (NP) modified glassy carbon electrode (GCE)	Range = 1.0 × 10^−7–3.5 × 10^−4 mol L^−1; LOD = 75 nmol L^−1; RSD = 5.6%; recoveries = 98.48–103.25%	85
FA	Water	Surface enhanced Raman scattering (SERS); gold nanorod (AuNR) grafting of molecularly imprinted polymers (MIPs)	Range = 1.0 × 10^−8–1.0 × 10^−4 mol L^−1; LOD = 0.1 μmol L^−1	86
FA		Electrochemical detection; on/off-switchable molecularly imprinted polymer (MIP) modified electrode	Range = 1.0–200 μmol L^−1; LOD = 0.9 μmol L^−1	87
FA	Aqueous, serum and pharmaceutical samples	Electrochemical detection (differential pulse and cathodic stripping voltammetric); molecularly imprinted polymer (MIP)-carbon consolidated composite fiber	LOD = 0.191–0.197 ng mL^−1; RSD = 0.9–1.4%; recoveries = 98.5–100.7%	88

---

4 Conclusion and perspectives

Folate deficiency is one of the major global public health issues. Low folate intake and poor folate status are generally present in pregnant women and children, especially in developing countries. More data are needed on various natural folate bioavailabilities and the folate content in foods typically consumed should be established for the different regions of the world. For different forms of folates in pharmaceuticals and foods, it is necessary to develop more easy methods with high sensitivity, specificity and good reproducibility. Compared with traditional MA methods, HPLC and CE can distinguish different folate forms. Various detectors are used in HPLC and CE. A MS/MS detector has higher sensitivity and specificity and it can accurately quantify various forms of folates at trace levels. In recent years, sensors have become a better alternative for folate analysis, especially electrochemical sensors and biosensors. The application of various novel materials as modifiers for electrodes increased the selectivity and decreased the detection limit of the sensors. In a word, the determination of natural and synthetic folates should be developed in a rapid, easy, sensitive, green way. The development of the techniques for rapid pretreatment and simultaneous determination of FA and its derivatives will be still the key research point in the future.

Conflicts of interest

There are no conflicts to declare.

References

1. F. Martin, E. C. Giménez and E. Konings, J. AOAC Int., 2016, 99, 19–25.
2. R. Obeid, K. Oexle, A. Rißmann, K. Pietrzik and B. Koletzko, J. Perinat. Med., 2016, 44, 261–268.
3. R. K. Saini, S. H. Nile and Y. Keum, Food Res. Int., 2016, 89, 1–13.
4. S. Akbar, A. Anwar and Q. Kanwal, Anal. Biochem., 2016, 510, 98–105.
5. H. S. Strandler, J. Patring, M. Jägerstad and J. Jastrebova, J. Agric. Food Chem., 2015, 63, 2367–2377.
6. R. Iyer and S. K. Tomar, J. Food Sci., 2009, 74, R114–R122.
7. C. Serrano-Amatriain, R. Ledesma-Amaro, R. López-Nicolás, G. Ros, A. Jiménez and J. L. Revuelta, Metab. Eng., 2016, 38, 473–482.
8. M. Jägerstad and J. Jastrebova, J. Agric. Food Chem., 2013, 61, 9758–9768.
9. W. G. P. B. Brown, Adequate food for all: culture, science, and technology of food in the 21st century, CRC press, Boca Raton, 2009.
10. B. C. Blount, M. M. Mack, C. M. Wehr, J. T. MacGregor, R. A. Hiatt, G. Wang, S. N. Wickramasinghe, R. B. Everson and B. N. Ames, Proc. Natl. Acad. Sci. U. S. A., 1997, 94, 3290–3295.
11. J. Y. Park, S. E. Vollset, A. Melse-Boonstra, V. Chajès, P. M. Ueland and N. Slimani, Mol. Nutr. Food Res., 2013, 57, 562–581.
12. Folates, https://ods.od.nih.gov/factsheets/Folate-HealthProfessional/#h2, accessed October 2017.
13. C. McStay, S. Prescott, C. Bower and D. Palmer, Nutrients, 2017, 9, 123.
14. WHO (World Health Organization), Guideline: Optimal serum and red blood cell folate concentrations in women of reproductive age for prevention of neural tube defects, WHO, 2015.
15. J. Selhub and I. H. Rosenberg, Biochimie, 2016, 126, 71–78.
16. C. M. Witthöft, Encyclopedia of Dairy Sciences, ed. J. W. Fuquay, P. F. Fox and P. L. H. McSweeney, Elsevier/Academic Press, Amsterdam; Boston, 2nd edn, 2011, Vitamins | Folates, pp. 678–686.
17. J. B. Wallingford, L. A. Niswander, G. M. Shaw and R. H. Finnell, Science, 2013, 339, 1222002.
18. Vitamin & Mineral Deficiency: a Global Progress Report. Part 8 the case of folate, http://https://www.unicef.org/media/files/vmd.pdf, accessed October 2017.
19. M. P. G. De Ocampo, A. Mrg, M. Ca, A. Je and M. Tr, Women Birth, 2017 DOI:10.1016/j.wombi.2017.08.128.
20. J. E. Laiño, H. Zelaya, M. Juárez Del Valle, G. Savoy De Giori and J. G. LeBlanc, J. Funct. Foods, 2015, 17, 22–32.
21. V. Capozzi, P. Russo, M. T. Dueñas, P. López and G. Spano, Appl. Microbiol. Biotechnol., 2012, 96, 1383–1394.
22. A. Meucci, L. Rossetti, M. Zago, L. Monti, G. Giraffa, D. Carminati and F. Tidona, Food Microbiol., 2018, 69, 116–122.
23. M. Czarnowska-Kujawska, E. Gujska and J. Michalak, J. Food Compos. Anal., 2017, 57, 64–72.
24. M. J. I. Shohag, Q. Yang, Y. Wei, J. Zhang, F. Z. Khan, M. Rychlik, Z. He and X. Yang, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci., 2017, 1040, 169–179.
25. E. P. D. P. Azevedo, C. A. de Azevedo Filho, B. S. Santos, L. T. Da Silva, A. E. C. Fai and J. A. Da Paix O, Food Anal. Methods, 2016, 9, 1321–1332.
26. E. Martín Tornero, A. Espinosa-Mansilla and I. Durán Merás, Microchem. J., 2017, 133, 333–345.
27. M. V. Chandra-Hioe, M. P. Bucknall and J. Arcot, Food Chem., 2017, 234, 365–371.
28. C. Y. Wang, H. W. Huang, C. P. Hsu and B. B. Yang, Crit. Rev. Food Sci. Nutr., 2016, 56, 527–540.
29. N. Naderi, Y. Pouliot, J. D. House and A. Doyen, Innovative Food Sci. Emerging Technol., 2017, 43, 191–200.
30. W. Shi, Y. Wang, H. Zhang, Z. Liu and Z. Fei, Food Chem., 2017, 226, 128–134.
31. J. I. Martin, W. O. Landen Jr, A. G. Soliman and R. R. Eitenmiller, J. - Assoc. Off. Anal. Chem., 1990, 73, 805–808.
32. C. Motta, I. Delgado, A. S. Matos, G. B. Gonzales, D. Torres, M. Santos, M. V. Chandra-Hioe, J. Arcot and I. Castanheira, J. Food Compos. Anal., 2017, 64, 181–187.
33. J. E. Camara, M. S. Lowenthal and K. W. Phinney, Anal. Bioanal. Chem., 2013, 405, 4561–4568.
34. K. Meisser-Redeuil, S. Bénet, C. Gimenez, E. Campos-Giménez and M. Nelson, J. AOAC Int., 2014, 97, 1121–1126.
35. S. Hussain, S. Khan, S. Gul, M. I. Pividori and M. Del Pilar Taboada Sotomayor, React. Funct. Polym., 2016, 106, 51–56.
36. M. V. Chandra-Hioe, M. P. Bucknall and J. Arcot, Anal. Bioanal. Chem., 2011, 401, 1035–1042.
37. J. Szpylka, J. DeVries, A. Cheney and S. House, J. AOAC Int., 2012, 95, 1547–1554.
38. S. Ansari and M. Karimi, Talanta, 2017, 164, 612–625.
39. B. B. Prasad, M. P. Tiwari, R. Madhuri and P. S. Sharma, Anal. Chim. Acta, 2010, 662, 14–22.
40. M. Rezaee, Y. Assadi, M. M. Hosseini, E. Aghaee and F. Ahmadi, J. Chromatogr. A, 2006, 1161, 1–9.
41. H. Yan and H. Wang, J. Chromatogr. A, 2013, 1295, 1–15.
42. Y. Nojavan, M. Kamankesh, F. Shahraz, M. Hashemi and A. Mohammadi, Talanta, 2015, 137, 31–37.
43. R. K. Saini, P. Manoj, N. P. Shetty, K. Srinivasan and P. Giridhar, J. Food Sci. Technol., 2016, 53, 511–520.
44. P. Upadhyaya, K. Tyagi, S. Sarma, V. Tamboli, Y. Sreelakshmi and R. Sharma, Food Chem., 2017, 217, 610–619.
45. S. Vishnumohan, J. Arcot and R. Pickford, Food Chem., 2011, 125, 736–742.
46. H. Indyk and D. Dowell, J. AOAC Int., 2012, 95, 298–300.
47. R. Matias, P. R. S. Ribeiro, M. C. Sarraguça and J. A. Lopes, Anal. Methods, 2014, 6, 3065–3071.
48. S. M. Wabaidur, S. M. Alam, S. H. Lee, Z. A. Alothman and G. E. Eldesoky, Spectrochim. Acta, Part A, 2013, 105, 412–417.
49. V. Fajardo, E. Alonso-Aperte and G. Varela-Moreiras, Food Chem., 2015, 169, 283–288.
50. J. Arcot and A. Shrestha, Trends Food Sci. Technol., 2005, 16, 253–266.
51. C. M. Pfeiffer, J. P. Hughes, D. A. Lacher, R. L. Bailey, R. J. Berry, M. Zhang, E. A. Yetley, J. I. Rader, C. T. Sempos and C. L. Johnson, J. Nutr., 2012, 142, 886–893.
52. P. Jin, L. Xia, Z. Li, N. Che, D. Zou and X. Hu, J. Pharm. Biomed. Anal., 2012, 70, 151–157.
53. E. Deconinck, S. Crevits, P. Baten, P. Courselle and J. De Beer, J. Pharm. Biomed. Anal., 2011, 54, 995–1000.
54. P. A. Ramos-Parra, R. Urrea-López and R. I. Díaz De La Garza, Food Res. Int., 2013, 54, 177–185.
55. F. Wang, M. Cao, N. Wang, N. Muhammad, S. Wu and Y. Zhu, Food Chem., 2018, 239, 62–67.
56. A. Freisleben, P. Schieberle and M. Rychlik, Anal. Bioanal. Chem., 2003, 376, 149–156.
57. K. Tyagi, P. Upadhyaya, S. Sarma, V. Tamboli, Y. Sreelakshmi and R. Sharma, Food Chem., 2015, 179, 76–84.
58. D. Silva, J. V. Visentainer, N. Souza and C. C. Oliveira, Food Anal. Methods, 2013, 6, 1592–1606.
59. U. D. Uysal, E. M. Oncu-Kaya and M. T. El, Chromatographia, 2010, 71, 653–658.
60. D. Zhao, M. Lu and Z. Cai, Electrophoresis, 2012, 33, 2424–2432.
61. S. Zhao, H. Yuan, C. Xie and D. Xiao, J. Chromatogr. A, 2006, 1107, 290–293.
62. K. Marakova, J. Piestansky, E. Havranek and P. Mikus, Pharmazie, 2014, 69, 663–668.
63. X. Wang, K. Li, L. Yao, C. Wang and A. Van Schepdael, J. Pharm. Biomed. Anal., 2018, 147, 278–287.
64. P. Liu, D. Liu, Y. Liu and L. Li, J. Solid State Chem., 2016, 241, 164–172.
65. B. Hemmateenejad, F. Shakerizadeh-shirazi and F. Samari, Sens. Actuators, B, 2014, 199, 42–46.
66. J. R. Zhang, Z. L. Wang, F. Qu, H. Q. Luo and N. B. Li, J. Agric. Food Chem., 2014, 62, 6592–6599.
67. M. Geszke-Moritz, G. Clavier, J. Lulek and R. Schneider, J. Lumin., 2012, 132, 987–991.
68. S. Cheraghi, M. A. Taher and H. Karimi-Maleh, J. Food Compos. Anal., 2017, 62, 254–259.
69. H. Beitollahi, S. G. Ivari and M. Torkzadeh-Mahani, Mater. Sci. Eng., C, 2016, 69, 128–133.
70. N. Lavanya, E. Fazio, F. Neri, A. Bonavita, S. G. Leonardi, G. Neri and C. Sekar, J. Electroanal. Chem., 2016, 770, 23–32.
71. A. Ananthi, S. S. Kumar and K. L. Phani, Electrochim. Acta, 2015, 151, 584–590.
72. L. Mirmoghtadaie, A. A. Ensafi, M. Kadivar and P. Norouzi, Mater. Sci. Eng., C, 2013, 33, 1753–1758.
73. X. Li, X. Tan, J. Yan, Q. Hu, J. Wu, H. Zhang and X. Chen, Electrochim. Acta, 2016, 187, 433–441.
74. X. Wang, Z. You, Y. Cheng, H. Sha, G. Li, H. Zhu and W. Sun, J. Mol. Liq., 2015, 204, 112–117.
75. A. Babakhanian, S. Kaki, M. Ahmadi, H. Ehzari and A. Pashabadi, Biosens. Bioelectron., 2014, 60, 185–190.
76. H. Dai, Y. Li, S. Zhang, L. Gong, X. Li and Y. Lin, Sens. Actuators, B, 2016, 222, 120–126.
77. M. P. Kingsley, P. B. Desai and A. K. Srivastava, J. Electroanal. Chem., 2015, 741, 71–79.
78. J. B. Raoof, N. Teymoori, M. A. Khalilzadeh and R. Ojani, Mater. Sci. Eng., C, 2015, 47, 77–84.
79. T. Jamali, H. Karimi-Maleh and M. A. Khalilzadeh, LWT--Food Sci. Technol., 2014, 57, 679–685.
80. A. Taherkhani, T. Jamali, H. Hadadzadeh, H. Karimi-Maleh, H. Beitollahi, M. Taghavi and F. Karimi, Ionics, 2014, 20, 421–429.
81. H. E. Indyk, Int. Dairy J., 2011, 21, 783–789.
82. J. E. Young, M. T. Matyska and J. J. Pesek, J. Chromatogr. A, 2011, 1218, 2121–2126.
83. P. W. Johns, J. H. Baxter and M. C. Terp, Food Anal. Methods, 2017, 10, 3255–3263.
84. R. Hau Fung Cheung, J. G. Hughes, P. J. Marriott and D. M. Small, Food Chem., 2009, 112, 507–514.
85. P. Kanchana and C. Sekar, Spectrochim. Acta, Part A, 2015, 137, 58–65.
86. R. Ahmad, N. Félidj, L. Boubekeurlecaque, S. Lautruong and S. Gamderouich, Chem. Commun., 2015, 51, 9678–9681.
87. N. Karimian, M. H. A. Zavar, M. Chamsaz, A. P. F. Turner and A. Tiwari, Electrochem. Commun., 2013, 36, 92–95.
88. B. B. Prasad, R. Madhuri, M. P. Tiwari and P. S. Sharma, Biosens. Bioelectron., 2010, 25, 2140–2148.

---