Folates-food sources , analyses , retention and bioavailability By

Health benefits of folates regarding their prevention of neural tube defects in babies and occlusive vascular diseases caused by elevated plasma homocysteine, their link to mental fitness and possibly certain forms of cancer have already been recognised. However, analyses of food folates is still tedious because of a lack of validated methods for characterisation and quantitation of the great number of native folate forms, but also due to a lack of adequate methods for sample pretreatment. Therefore, the assessment of folate losses through industrial and household food processing is still incomplete, as well as knowledge on folate bioavailability in humans. This paper reviews the state of art for the occurrence and analysis of folates in foods. Furthermore, results are summarised from studies of folate retention during food processing and the assessment of folate bioavailability.


Introduction
Folates represent an important group among the B-vitamins, participating in one-carbon transfer reactions required within the cell, especially for purine and pyrimidine biosynthesis (DNA and RNA) and amino acid interconversions.Optimal folate nutrition and status are linked to diminished risk for neural tube defects, occlusive vascular diseases (1)(2)(3)(4) and possibly some forms of cancer (5).Recently, a relationship to cognitive and mental function has also been discussed (6).
In the latest edition of the Nordic Nutritional Recommendations (1996), the daily intake for adults was increased from 200 pg to 300 pg folate and for pregnant women a daily intake of 400 pg was recommended (7).When publishing the dietary reference intakes (DRI) in 1998, the US Food and Nutrition Board included the concept of possible health-protective effects of folate by increasing recommendations for adults to 400 pg/d from the previous 200 pglday (8).Moreover, the US Food and Nutrition Board recommends women who plan a pregnancy to consume an additional 400 pg synthetic folic acid from fortified foods or supplements.Such dramatic increases of recommended intakes for folate combined with the fact that the average daily intake of folate among Western populations is generally lower than recently set recom-mendations, emphasise the need for a critical evaluation of the dietary sources of folates.Most of the food folate data derive from microbiological analysis with often insufficient methodological control.There is still today little reliable information about which folate forms and concentrations are present in food and what impact food processing techniques have on folate retention.Moreover, knowledge about human folate bioavailability from native food sources or after fortification is still incomplete (9).Possible risks and benefits from food fortification with folic acid are currently a subject of controversy.
This review aims to give brief information on folate contents and analysis in food, folate losses during food processing and the assessment of folate bioavailability (for more detailed information see [10][11][12][13].

Folate chemistry and stability
According to recommendations of the IUNS Committee on Nomenclature (1986), "folate" should be used as the generic term for the class of compounds having similar chemical and nutritional properties to pteroyl-L-glutamic acid (folic acid) (14).While the pteridine ring of folic acid exists in oxidised form, native folates have either two or four additional hydrogens in their pteridine ring forming dihydro-or tetrahydrofolates.Thus, dietary folates exist primarily as reduced, one-carbon-substituted forms of pteroylglutamates, with up to seven glutamyl residues attached to the p-aminobenzoic group by y-peptide linkage.Five different one-carbon units are known to be linked at N5-and/or NIo-position of the pteroyl group: methyl (5-CH,), formyl (5-or 10-HCO), formimino (5-CHNH), methylene (5,lO-CH,) and methenyl (5,lO-CH) (Figure 1).Altogether the theoretical number of all native folate vitamers reaches several hundred (1 I).
All folates are in danger of oxidative degradation enhanced by oxygen, light and heat, resulting in a splitting of the molecule into biologically inactive forms, of which p-aminobenzoylglutamate is one major form.There are considerable differences in stability between various reduced folate forms; the order of stability is: 5-HCO-H4folate > 5-CH,-H4folate > 1 0-HCO-H4folate > H,folate.
Moreover, the stability is pH-dependent.Folic acid exhibits substantially greater stability than the reduced folate forms.The chemistry of folates makes the vitamin one of the most vulnerable to losses during food processing.If present in adequate amounts, antioxidants, e.g.ascorbic acid and thiols, protect folates.The rate of reaction for folate breakdown in the presence of oxygen depends on the type of folate derivative and the nature of the food matrix, in particular with respect to pH, buffer composition, catalytic trace elements and antioxidants (1 1 -13).

Folate content in foods
A brief look into various food tables from different countries (Table 1) shows that many vegetables and pulses are rich sources of folate, with folate concentrations up to 600 pg/100 g in some beans and chick peas and around 200 pgI100 g in leafy vegetables.A sort of general rule is that the lower the water content in the vegetable the higher the folate concentration, and moreover that leafy vegetables are good folate sources (folium means leaf).Folate concentrations in fruits and berries are usually one-tenth those of vegetables, ranging from a few pg to approx.50 pg/ 100 g.The highest concentrations are to be found in frozen concentrated orange and grapefruit juice, strawberries and several nuts with folate concentrations of about 50-1 00 pg/ 100 g or more.
Meat and meat products, except liver which is the storage organ, contain little folate, while chicken and fish are moderate sources (1 5).Folate concentrations in milk are only around 5 pg1100 ml, but milk is however of interest due to its content of a specific high-affinity folate-binding protein.This may play an important role in the regulation of absorption and bioavailability of dietary folates from the gastrointestinal tract (16).An increase of the folate content usually results from fermentation of milk and whey products.In hard cheeses, up to 40 pg/ 100 g and more were quantified.Soft cheeses like Camembert and Brie contain 60-100 pg/ lOOg (17).High in folate is egg yolk, with up to 90 pg/100 g (15).
Cereals are also an important dietary source for folate.Some cereal fractions like bran and germs contain a few hundred pg folate per 100 g, while bread from wholemeal flour contains 50-1 00 pg/100 g (Table 1).Baker's yeast with its extremely high folate concentration (approx .1000-4000 pg/100 g) contributes to the folate content in soft bread.Commonly, foods are ordered into groups being "rich", "good" and "moderate" sources, with folate concentrations of > 100 pg ,SO-100 pg and 15-50 pg per serving, respectively (1 8).
Folate analysis by HPLC can provide information on individual folate forms present in food, but currently only few data are available; some examples from vegetables, fruits and dairy products are given in Table 2.The sum of concentrations from individual folate forms analysed by HPLC cannot be directly compared with total folate concentrations assessed by microbiological methods, as the latter response to nearly all tetrahydro-, dihydro-and fully oxidised folate forms in the food sample.It is still unclear, however, which concentration best reflect the amount of folate bioavailable for humans.Both tables show notable discrepancies between folate data for some foods, which cannot only be explained by differences with folate content but rather by methodological differences in respect to folate quantification and sample pretreatment, emphasising the need for re-assessment using better controlled methods.

Folate analysis
Traditionally, microbiological assay procedures with Lactobacillus casei (ATCC 7469), which responds to most folate derivatives with up to three polyglutamate residues, are used for folate quantification (19).Most folate values published in food tables today were established by microbiological methods.
The use of (radio-) protein binding procedures is common for clinical diagnostics in plasma, serum or whole blood mainly containing 5-CH3-H4folate (20).These tests are based on non-specific competitive binding of folates from the test sample and radiolabelled folic ,acid to a folate-binding protein.Accurate control of the assay pH is required, and formylated folates are only bound to a small percentage (2 1,22).Protein-binding procedures have not often been used successfully for food folate analyses and their application might be restricted to food matrixes which contain mainly 5-CH3-H4folate (23)(24)(25)(26)(27)(28)(29)(30).Most foods, however, contain a variety of folate forms (3 1-36).
Folate stabilisation before and during analysis is necessary as folates are sensitive to oxygen, heat, light, pro-oxidants and extreme shifts of pH-value (10)(11)(12)(13).Optimisation of stabilisation procedures is hampered, because individual folates forms possess different pH-optima for maximum stability.With the use of antioxidants throughout sample preparation, folates are successfully protected from interconversion (44) and from oxidative degradation (45)(46)(47)(48).Additional exclusion of oxygen by overlay with nitrogen (49-5 1 ), as well as use of low temperatures and shelter from light, should already be applied when homogenising the samples.
Commonly, folates are extracted in slightly acidic to slightly alkaline conditions by heat, using autoclaving with microbiological folate assay ( 19,25,52) or a boiling water bath with HPLC procedures (40,4 1).
Prior to quantification with both microbiological and HPLC procedures, deconjugation of folate polyglutamates to monoglutamates is required (9,lO).As deconjugase preparations are not commercially available, they have to be prepared by the investigator.Partially purified suspensions of y-glutamyl-hydrolases (EC 3.4.22.12) from hog kidney (37), chicken pancreas (53) and plasma from humans or other species (25,54,55) are used; to a lesser extent enzyme preparations from intestines (56)(57)(58).For HPLC determination, the use of human plasma or hog kidney deconjugation is advisable, because these exopeptidases produce folate monoglutamates (37,59).Procedures for folate deconjugation have to be optimised in respect to time, temperature, incubation milieu, folate stabilisation and substrate-enzyme ratio depending on the characteristics of the sample   Recently, increased total folate concentrations were reported after additional use of proteolytic and amylolytic enzymes in cereal-based foods, some vegetable and dairy products or composite food samples (33,52,(60)(61)(62).Procedures for this so called tri-enzyme extraction are yet not widely established.Further investigations are needed for accurate food folate quantification in order to revise current folate data in food tables (62).
Purification of food extracts is necessary when using HPLC determination.Several solid phase extraction procedures with commercial disposable cartridges have been reported (37,40,41,5 1,63,64).Less common is the use of affinity chromatography with folate-binding protein attached to agarose gel (33,6567).The affinity columns have to be prepared by the investigator, they are comparatively expensive, show low affinity to 5-HCO-H,folate (65) and possess a short life-time (3 1).
The need for validated methods for folate determination becomes obvious.Attention should not only be focused on the method of determination, but also on techniques for extraction, stabilisation and deconjugation of folates including quality control of methods by e.g.recovery studies, use of certified reference material and adequate calibration procedures.These tasks were recently tackled in the studies of the "EU SMT project on improvements in the determination of folates in food" (25,(68)(69)(70), which have resulted in the preparation of four certified reference materials (CRM) from milk powder, wholemeal wheat flour, lyophilised pig's liver and mixed vegetables.

Folate losses or retention during storage and processing
Most foods in modern nutrition are consumed after being processed by household or industrial procedures.This aims for microbiological safety, convenience regarding storage and distribution and optimal nutritional value and organoleptic appeal (1 3).Several methods of industrial and household processing were investigated regarding their impact on folate retention, as summarised in Table 3.Most studies reflect negative effects from processing, causing increasing losses with increasing severity of processing conditions in terms of heating temperature and time.Although quite a few studies could be found focusing on the influence of processing on folate concentration in vegetables, the picture is still fragmentary.For fruits and berries there is practically no information at all on processing and storage stability of folates .
Leakage and oxidative degradation are the major reasons for folate losses during processing and storage.After thermal processing of vegetables, as displayed in Table 3a, both leakage and oxidative degradation can cause folate losses up to 70-80%.In high-temperature short-time processing such as pasteurisation or UHTtreatment of milk, no leakage but oxidative degradation of folate occurs resulting in rather modest losses between 0 and 20% (7 1).Oxidation is also the major cause of folate losses during steam-flaking , spraydrying and extrusion cooking of cereals when producing pre-cooked cereal products for gruel, porridge, biscuits etc. (72).
Oxidation of tetrahydrofolates , e .g .5-CH3-H4folate to 5-CH3-dihydrofolate, can occur, which as a very unstable compound is prone to further degradation.However, Lucock et al. (73) showed that 5-CH3-dihydrofolate can be reduced to 5-CH3-H4folate in the presence of ascorbic acid and can subsequently be salvaged under physiological conditions.Many food sources contain endogenous ascorbic acid which can prevent folate oxidation (74).From 10-HCO-H4folate the formation of oxidation products like 10-HCOdihydrofolate or 10-HCO-folic acid is possible.
Heat treatment can also cause interconversion of several folate forms.Gregory et al. (37) studied interconversion of the very labile 10-HCO-H4folate to the more energetically stable 5-HCO-H4folate.Heat-induced conversion is rapidly completed, so that formulated folates can only be analysed as one 5-HCO-H4folate fraction by HPLC.
It is still questionable whether certain forms of processing, which include rise of temperature, may have an impact on folate pattern and concentrations due to heat destruction of endogenous food enzymes.An experiment by Ristow et al. (75) indicated lower folate bioavailability from cooked compared to raw cabbage in a bioassay using chicken.When endogenous folate deconjugase was not inactivated by heat treatment, folate polyglutamates from raw cabbage were presumably deconjugated to a greater extent prior to feeding.These "endogenously" deconjugated folate monoglutamates were according to the authors more bioavailable.
On the other hand, folate monoglutamates might be more susceptible to oxidative or enzymatic degradation after endogenous deconjugation.This idea could be supported by the findings of Malin et al. (74), who determined folate losses of 84% after 188 days of storage at -2 1 "C in non-blanched Brussels sprouts samples.In blanched samples, no losses occurred during storage.
Food processing can lead to positive effect when applying fermentation procedures, as compiled and studied for milk products, e.g.yoghurt, buttermilk and cheeses by Wigertz (17), and as traditionally practised with some vegetables and during bread-making with yeast and sour dough.
Some food processing including fiactionation results in fractions with either increased or decreased folate concentrations.A typical example of this is milling.When whole grains of cereals like wheat, rye and oats are milled, the grains become separated into bran, germs, wholemeal flours etc. resulting in markedly increased concentrations of folates in the bran parts and decreased folate content in the branfree flour fractions (Table 1).
There is a need for intensified studies, especially with a view to developing new techniques enhancing folate content and bioavailability in products.Together with gentle techniques of processing, this can maximise folate retention.which is desirable in the development of functional foods.

Human folate bioavailability
Bioavailability is commonly defined as the absorption and metabolic utilisation of a nutrient, involving processes of intestinal absorption, transport, metabolism and excretion (76,77).Conjugated dietary folates consumed by man require deconjugation to their monoglutamate forms prior to absorption (78).Human intestinal brush border deconjugase is a zinc-dependent exopeptidase with optimum activity at pH 6.5 catalysing stepwise hydrolysis of polyglutamyl folates (56,79).Once deconjugated, the folate monoglutamates are absorbed by an active energy-dependent carrier mediated process at physiological concentrations, and by passive diffusion at higher concentrations (80).Absorption takes place mainly in the jejunum and is markedly influenced by pH, with a maximum at pH 6.3.Oxidised folates in physiological amounts are reduced and methylated when passing through the intestinal mucosa or during passage of the liver, the storage organ of folates.All intracellular folates exist in polyglutamyl form, whereas transport in plasma occurs in the methylated monglutamyl form (80).
Folate bioavailability can be influenced by several intrinsic or extrinsic factors.Directly linked to the organism are folate status, health and gastro-intestinal function, age, sex and antagonistic effects from use of drugs and alcohol.Extrinsic factors are many depending on the ingested folate (e.g .oxidation status and substituent, length of polyglutamate chain), the food matrix (e .g .folate interaction with substances from the food matrix, substances with antifolate activity), or processing and storage.
Initially bioassays were performed to determine folate bioavailability using rats 8 1,82), pigs (83,84), chickens (85) or even monkeys (86).Bioassays make it possible to determine folate concentrations in several response tissues such as liver, kidneys, serum and whole blood and allow the assessment of growth or reproduction activity as parameters for folate bioavailability.Often bioassays are performed according to a so-called depletion-repletion protocol, whereby animals are fed a folate-free diet during the depletion period and then receive various doses of folate during repletion.Accumulation of folate until steady-state is determined in the response tissues of each group to create a dose response curve (87).
Cliflord et al. (88,89), using a rat bioassay model, reported folate bioavailability between 70 and 120% from food items like lyophilised orange juice, cooked lima beans, cooked leafy vegetables, wheat germ, eggs and mushrooms.These figures are higher than usually found in human studies based on similar food items.One reason could be a more efficient folate deconjugation in rats, since pancreas juice of rats, in contrast to man, is a very rich source of deconjugases (90).Another factor is that the rat model is performed over a period of 1-2 weeks and folate intake might include a contribution Scand J NutdNiiringsforskning 4/99 Prinz-Langenohl et al 1999 (98) aPteGlux = pteroylglutamic acid (suffix 1 -7 for number of glutamate rests attached), no suffix = pteroylmonoglutamic acid (folic acid); hS-HCO-H4folate = 5formyltetrahydrofolate; 'AUC = area under the plasma concentration curve; d5-CH3-H,folate = 5-methyl-tetrahydrofolate from folates produced and absorbed in the colon or simply due to coprophagy.Data from human studies originate mainly from single dose protocols.There is a growing discussion about the appropriateness of bioassays in order to predict folate metabolism in humans, due to quantitative and qualitative physiological differences between species.
In vitro methods have been used to investigate properties of endogenous deconjugases from various species (1 2,9 1,92) or folate binding to other food components, e.g.dietary fibre (85).Wigertz (17) studied folate retention in milk products using an in vitro model simulating the human gastro-intestinal tract.However, in vitro methods can reflect the complexity of in vivo folate absorption, deconjugation and metabolism only to a limited extent.
Studies involving human volunteers provide the best means to assess folate bioavailability by the use of long-term or short-term protocols.Long-term folate kinetics usually investigate the impact of repeated folate supplementation on folate status, e.g.feeding fortified vs. natural foods vs. folic acid supplements (93,94).Subsequently, folate concentrations in whole blood or erythrocytes are quantified (93,95), which reflects the folate status of the past three months.These studies are intensive as regards costs and time.Short-term folate absorption and elimination kinetics are often studied after application of a single oral folate dose in the form of a pharmaceutical preparation or a fortified or natural food sample.Plasma and urinary folate concentrations can be determined by protein-binding assay (96)(97)(98), microbiological (99)(100)(101)(102)(103)(104) and HPLC methods (42,(105)(106)(107).Comparison of the area under the plasma response curve or urinary folate excretion allow determination of relative bioavailability of different folate forms, from various doses or application techniques (98,108,109).
In recent years, the use of stable isotope protocols was introduced by Gregory et al. (95,110,111).The incorporation of isotopically labelled folates (from fortified food samples) into the diet provides a tool to estimate folate turnover by urinary and fecal excretion of the labelled compounds and their metabolites (108).Advantage of the use of stable isotopes is that isotopically labelled folates from the dose and endogenous folates from body stores can be differentiated.Due to lack of commercial availability of labelled compounds, intensive analytical procedures and the difficulty of mass spectrometric determination, stable isotope procedures are not commonly applicable.
A drawback of some short-term protocols for the determination of folate kinetics is the fact that the application of a single folate dose of physiological magnitude only results in a small plasma and urine folate response.Therefore often pharmacological doses or loading of body stores are used (97,109), which do not reflect physiological conditions.Table 4 summarises in chronological order sample results from human studies of folate bioavailability and absorption kinetics, even showing contradictory results.Whichever protocol for the determination of folate bioavailability is applied, the limitations of each approach have to be taken into consideration.
Summing up the state of the art concerning folate bioavailability in man, synthetic pteroylmonoglutamic acid (folic acid) is absorbed almost completely when ingested as a physiological dose without simultaneous consumption of food, while synthetic pteroylpolyglutamates seem to be less available (50-80%) even when given in the absence of food (1 12).
Figures describing absorption of native dietary folates, which exist as a mixture of reduced mono-and polyglutamate forms, have to be interpreted with care, but are generally in the magnitude 40-70%.
Matrix effects from foods, e.g.interaction with deconjugase inhibitors, binding or encapsulation to different food constituents, pH effects or losses during gastro-intestinal digestion, owing to oxidative degradation of native folates are some factors that might contribute to reduced bioavailabili ty.
A food constituent of particular interest is folate-binding protein (FBP) from milk.In cow's milk most of the folate is bound to FBP.Folate bound to FBP is mainly absorbed in the ileum at a slower rate than the free folate which is absorbed in the jejunum.FBP has been suggested to play a role in the absorption of folate by preventing its uptake by intestinal bacteria, and it may also directly promote the transport of folate across the intestinal mucosa.Most of these studies investigating the role of FBP on folate bioavailability are performed using animals (16,(113)(114)(115)(116).
Studies performed on infants by Ek and Magnus (1 17,118) showed that breast-fed babies had a better folate status than bottle-fed babies.Breast-fed babies sustained their folate status on an intake of 55 pg folate per day while bottle-fed babies needed 78 pg folate per day.
It is suggested that these discrepancies are due to the presence of FBP in human milk which is destroyed in heat-processed milk formula.

Outlookneed for future investigations
The importance of an adequate folate intake to maintain health has already been recognised.Despite that, there is still uncertainty about folate requirements during e.g.infancy, growth, pregnancy or when suffering from chronically or inflammatory diseases.Quantitative aspects of human folate absorption, bioavailability and metabolism are still poorly understood.
Several methods for folate determination are established today, but the need for strict methodological validation has to be stressed, including international method calibration and harmonisation to revise current food data tables.
Another issue is the evaluation of current and development of new techniques of food processing with respect to beneficial or adverse effects on folate content, retention and bioavailability.Taken together, these aspects all point to a need to re-evaluate national health policies, and to participate in the discussion of (mandatory) food fortification and functional foods.