Chapter 23 summarized appraisal of analytical techniques

The results obtained from a microbiological assay provide an estimate of the total biological activity of a particular vitamin in the sample extract presented for analysis.. The microbio

Summarized Appraisal of

Analytical Techniques

The purpose of this chapter is to make some inferences from the analytical methodology described in previous chapters and to illustrate the application of current analytical techniques

23.1 Microbiological Assays

The turbidimetric microbiological assay is useful for estimating the total vitamin activity of a food or diet because it can be applied to all of the B-group vitamins and it is sensitive enough to measure the levels

of naturally occurring vitamin The assay organisms are selected on the basis of their specific requirement for vitamin forms that are biologically active in humans Microbiological assays using lactic acid bacteria or yeasts require the chemical or enzymatic liberation of bound vitamin The equipment, bacterial cultures, and ready-made media are commer-cially available and the assay procedure follows a standardized protocol Once everything is set up, batches of samples can be analyzed simul-taneously for most of the B vitamins The use of microtiter plates allows automated determination and computer analysis of data The results obtained from a microbiological assay provide an estimate of the total biological activity of a particular vitamin in the sample extract presented for analysis Whether the results reflect bioavailable vitamin depends on the extraction procedure employed

the yeast Saccharomyces cerevisiae underestimates the total vitamin B6 content if the sample contains predominantly pyridoxamine (PM) (e.g.,

a processed meat product), as the microbial growth response to this vitamer is markedly less than that to pyridoxal (PL) or pyridoxine (PN) This unequal response is of little concern in plant-derived foods or foods that are fortified with PN HCl Kloeckera apiculata has been pro-posed as the assay organism in the standard turbidimetric and semiauto-mated radiometric – microbiological assay on the basis of an equivalent

growth response to all three free vitamers However, this proposal has not found acceptance in certain other laboratories, in which K apiculata was found to exhibit an even lower relative response to PM than that obtained with S cerevisiae

The microbiological assay using Lactobacillus plantarum is the standard method for determining biotin in foods, but the nutritional significance

of the results obtained is largely dependent on the extraction procedure employed Maximum liberation of bound biotin from animal tissues

these conditions promote losses of biotin in plant tissues, and a somewhat milder acid hydrolysis must be used for plant foods and food composites Bitsch et al [1] reported that treatment of samples with liquid nitrogen, followed by digestion with papain, was capable of quantitatively releas-ing bound biotin (includreleas-ing biotin from biocytin) from both plant and animal tissues

Total folate is determined using Lactobacillus rhamnosus (casei) after enzymatic deconjugation to monoglutamyl forms The basic premise using this approach is that all active monoglutamyl folates have identical equimolar growth-support activities for L rhamnosus under the conditions

of the assay This assumption is, however, a subject of controversy Careful control of pH and buffering capacity of the medium appear to be essential prerequisites The overall picture reveals that the microbiological assay for folate is fragile, rather than robust A comparison of HPLC and micro-biological results using the same extract of each food for both assays showed significant differences [2,3] For oat flakes, cabbage, orange juice, and particularly whole-wheat flour, the L rhamnosus assay yielded higher folate values, suggesting the influence of nonfolate com-pounds on the bacterial growth response

23.2 High-Performance Liquid Chromatography

23.2.1 Introduction

HPLC is the current method of choice for determining the fat-soluble vitamins and can distinguish between added and indigenous vitamin The ability to quantify individual vitamers is important if the vitamers have different potencies The ability to separate carotenoids provides the opportunity to identify these compounds using sophisticated detec-tors Increases in lower potency cis isomers of carotenoids are an early indication of heat damage in thermally processed foods

The high repeatability of HPLC makes it ideal for determining the added thiamin, riboflavin, nicotinamide, pyridoxine, pantothenic acid, and folic

acid in fortified foods The determination of naturally occurring B vitamins by HPLC is made more complicated by the existence of

need for more complicated extraction procedures to deal with the various bound forms Pantothenic acid and biotin, which lack a strong chromophore, can be determined at naturally occurring levels after con-version to a fluorescent derivative HPLC lacks the required sensitivity

to accurately measure the extremely low indigenous levels of vitamin

despite the poor optical absorptivity of dehydroascorbic acid and its lack of electrochemical activity It cannot be guaranteed that HPLC will account for all of the biologically active forms of a vitamin that would

be measured by microbiological assay and, in this respect, the inherent specificity of HPLC may lead to an underestimation of total vitamin activity Thiamin results in food products analyzed by the microbiological assay were found to be 15 – 39% higher than results obtained by HPLC [4]

23.2.2 Fat-Soluble Vitamins

23.2.2.1 Vitamin A

The AOAC method for determining retinyl palmitate in fortified fluid milk [5] is based on the method of Thompson et al [6], with the addition

of retinyl acetate as internal standard A 2-ml milk sample is treated with ethanol to denature the proteins and fracture the fat globules The total lipid fraction is extracted into hexane, and the proteins are removed by centrifugation An aliquot of the hexane extract is injected onto a silica column for normal-phase chromatography with UV detection

The AOAC method for determining naturally occurring retinol in all food categories [7] involves saponification and neutralization with glacial acetic acid The solution is then diluted with tetrahydrofuran/

acid salts formed during saponification After centrifugation, analysis is performed by reversed-phase HPLC with UV detection

23.2.2.2 Carotenoids

Most published methods for carotenoid analysis employ nonaqueous reversed-phase chromatography with photodiode array detection Silica-based polymeric C18-bonded-phase column packings achieve the

isomers [8] The introduction of a polymeric C30-bonded phase has advanced carotenoid analysis, as it is capable of resolving geometric isomers of asymmetric carotenoids such as a-carotene as well as those

of b-carotene [9] The polymeric C30-bonded phase has been used success-fully to analyze provitamin A carotenoids in foods [10,11]

23.2.2.3 Vitamin D

and enteral nutritional products [12] is based on the method of Sliva

et al [13] and involves saponification, solid-phase cleanup and concen-tration, and quantitative reversed-phase HPLC The AOAC method for

semipreparative normal-phase HPLC, and quantitative reversed-phase

raw meat and liver using two steps of semipreparative HLPC [15]

23.2.2.4 Vitamin E

All of the eight unesterified tocopherols and tocotrienols can be separated isocratically by normal-phase HPLC and detected fluorometrically, making this technique ideal for the analysis of vegetable oils and fats [16] Normal-phase HPLC with UV detection is the current AOAC method for determining supplemental vitamin E in milk-based infant formula [17]

gradient elution is used Isocratic separation of all E vitamers can, however, be achieved using a pentafluorophenylsilica polar reversed-phase column and a methanol/water mobile reversed-phase [18]

23.2.2.5 Vitamin K

The AOAC method for determining vitamin K in milk and infant formulas [19] was developed by Indyk and Woollard [20] The analytical protocol involves lipase digestion, solvent extraction, and nonaqueous

and fluorescence detection Woollard et al [21] extended the enzymatic

trans-phylloquinone as well as menaquinones and

and dihydrophylloquinone in margarines and margarine-like pro-ducts [22] For the determination of phylloquinine and menaquinones

in foods of animal origin, semipreparative normal-phase HPLC was used to isolate a vitamin K fraction, which was then analyzed by

23.2.3 Water-Soluble Vitamins

23.2.3.1 Thiamin and Flavins

HPLC methods are well suited for determining thiamin and riboflavin

in fortified foods and foods containing appreciable amounts of these vita-mins The use of a common extraction procedure allows these vitamins

to be chromatographed either simultaneously or successively with a high degree of precision Recovery data have indicated that riboflavin remains stable during the precolumn oxidation of thiamin to thiochrome [24] Postcolumn chemistry eliminates the problem of reducing sugars (produced during acid hydrolysis) competing with thiamin for the oxidizing agent (ferricyanide) [25] The HPLC procedure proposed by Reyes and Subryan [26] incorporates a number of desirable features designed to provide reliable results with high sensitivity and to prolong the life of the analytical column Simultaneous determination of thiamin and riboflavin is achieved through simple reversed-phase chromato-graphy at neutral pH using a single fluorescence detector Cleanup and concentration of the ferricyanide-treated sample extract is effected by solid-phase extraction, which provides a purified and neutralized sol-ution for injection Thiamin values for raw and processed foods were the same (within experimental error) as those obtained by the 1984 AOAC fluorometric method [27], which uses Bio-Rex 70 in the purifi-cation step The HPLC was further validated by the good agreement obtained between HPLC values for both thiamin and riboflavin with certified values on three dry AACC (American Association of Cereal Chemists) check samples

Abdel-Kader [28] analyzed various enriched and nonenriched foods for thiamin using the AOAC manual fluorometric method and an HPLC method involving postcolumn derivatization and fluorometric detection

of thiochrome There was no statistical difference between the values obtained by the two procedures

Several investigators have compared riboflavin values obtained by HPLC with those obtained by the AOAC manual fluorometric method Good or reasonable agreement has been reported for the determination

of riboflavin in rice and rice products [29]; milk, eggs, and dairy products [30,31]; infant formula products [32], blanched soya beans [33]; and various raw and processed foods [24,34 – 36] Dramatically higher ribofla-vin values using the AOAC method were reported for soy products due to interference from fluorescent impurities that were not removed by oxi-dation with permanganate [37] Evidence for this statement was that treat-ment of the samples with permanganate before HPLC analysis failed to eliminate all the nonvitamin peaks from the chromatogram Higher ribo-flavin values have also been reported for fortified cereal products using a semiautomated modification of the AOAC method [25]

Favorable comparisons between HPLC and microbiological assay (L rhamnosus) for riboflavin determinations have been reported for the analysis of infant formula products [32], blanched soya beans [33], and

a wide range of foods [35,38 – 40] For nonenriched flours and flours with a high rate of extraction, the HPLC values were as much as

25 –50% lower than those found with microbiological assay [39] This discrepancy was considered to be caused by an underestimation of

chroma-tographic separation and nonmeasurement of biologically active isomeric riboflavin monophosphates

23.2.3.2 Niacin

An HPLC method for determining niacin with solid-phase cleanup and UV detection produced results for SRM 1846 milk-based infant formula that were within uncertainty ranges of the certified value [41] Lahe´ly et al [42] revitalized HPLC methodology for niacin by postcolumn conversion of nicotinamide and nicotinic acid to fluorescent derivatives

by UV irradiation, thereby increasing the selectivity and sensitivity of detection Cleanup of acid hydrolyzates was not necessary The quantifi-cation limit was estimated at approximately 0.2 mg/g for a 5-g test sample

of food Rose-Sallin et al [43] tested this method on a range of fortified products and obtained results that were in good accordance with certified values of both SRM 1846 and VMA 195 (a fortified cereal)

23.2.3.3 Vitamin B6

all bound forms by autoclaving samples at 1208C for 30 min in 0.1 N

plant-derived food samples, which contain significant amounts of

could be estimated indirectly by analyzing the sample with and without the double-enzyme treatment The difference in PN content between the two sets of results gives an estimate of PN-glucoside The validated method of Reitzer-Bergaentzle´ et al [45] employs an acid phos-phatase treatment that does not hydrolyze glycosylated PN and therefore yields results that represent bioavailable vitamin B6 The sample prep-aration includes chemical conversion of PM to PL and subsequent reduction of PL to PN, thus only one peak (orginating from phosphorylated and free vitamers) need be measured in the chromatogram

23.2.3.4 Pantothenic Acid

Woollard et al [46] developed an HPLC method for determining the free

pan-tothenate in infant formulas The problem of poor spectral specificity in the low UV was tackled by combining photodiode array detection and online spectral analysis A comparison of results with microbiological assay data showed good agreement

23.2.3.5 Biotin

The low concentrations (5– 50 ng/g) of biotin in most foods and the absence of a strong chromophore in the biotin molecule have precluded direct sensitive detection of this vitamin by HPLC However, postcolumn derivatization with avidin – fluorescein 5-isothiocyanate (FITC) reagent enables biotin to be determined in natural foods by HPLC – fluorescence with a detection limit of 5 ng/g [47]

23.2.3.6 Folate

Doherty and Beecher [48] developed a robust procedure for distinguish-ing between naturally occurrdistinguish-ing 5-methyl-THF and added folic acid

in foods Folate deconjugation was accomplished by a tri-enzyme treat-ment and a large-capacity polystyrene–divinylbenzene-based solid-phase extraction column provided sample cleanup with quantitative recovery

of analytes Folic acid was converted to a fluorescent product by photoly-sis of the appropriate segment of the column effluent and detected by pro-grammed fluorescence The earlier-eluting 5-methyl-THF is naturally fluorescent Results for foods were in close agreement with those from micobiological assay of total folate Ndaw et al [49] determined total folate in foods by chemically converting folates to 5-methyl-THF before sample purification by affinity chromatography

23.2.3.7 Vitamin C

Many HPLC methods for determining vitamin C in foods have been reported, using a variety of separation techniques and detection systems HPLC is accurate and sensitive for the determination of

C The poor optical absorptivity of dehydroascorbic acid and its lack of electrochemical activity necessitate the chemical reduction of this com-pound to ascorbic acid if absorbance or electrochemical detection is to

be used Fluorescence detection requires the chemical oxidation of ascorbic acid to dehydroascorbic acid and subsequent derivatization

If this derivatization is performed postcolumn, and ascorbic acid and dehydroascorbic acid are separated, the oxidation step can be omitted

A simple and convenient protocol is to reduce the dehydroascorbic acid in the sample extract with dithiothreitiol and measure the ascorbic acid (representing total vitamin C) by reversed-phase HPLC and UV detection, using a photodiode array detector [50 –52] A low-pH mobile phase is necessary to achieve ion suppression, but a “shielded” stationary phase will be sterically protected from attack by hydrolyzing protons Reversed-phase HPLC separates ascorbic acid from other organic acids and from erythorbic acid This methodology has been subjected to

an interlaboratory study and found to be suitable for the routine determination of total vitamin C in fruit juices and selected foods at

5 –60 mg/100 g [52]

23.3 Supercritical Fluid Chromatography

Supercritical fluid chromatography (SFC) offers the possibility of a com-pletely automated analysis of fat-soluble vitamins as it can be coupled directly to a supercritical fluid extraction (SFE) module The exclusion

of light, moderate temperature, and an oxygen-free environment backed by the solvating power of supercritical carbon dioxide are condu-cive to the analysis of fat-soluble vitamins and carotenoids The elimin-ation of organic solvents, normally consumed in large volumes, is beneficial with regard to pollution, inhalation, fire risk, and cost Published applications include the separation of cis –trans isomers of

and capillary [55] columns

23.4 Capillary Electrophoresis

Capillary electrophoresis has been used successfully in various appli-cation fields such as biochemistry, biotechnology, pharmaceutical analysis, and clinical chemistry Relatively little impact has been made

in food analysis Advantages of capillary electrophoresis over HPLC are superior resolution of sample components, shorter run times, more robust columns, and cheaper operating costs Disadvantages are a lower sensitivity attributable to the extremely low (nanoliter) injection volumes and small volume of the detector cell (a small section of the capillary column)

Micellar electrokinetic capillary chromatography (MECC) has been applied to the determination of thiamin in meat [56] and milk [57] and found to compare favorably with HPLC Sample cleanup using a

cation-exchange column allowed concentration of the purified sample extract

Cataldi et al [58] established the optimal conditions to quantify ribofla-vin, flavin mononucleotide (FMN), and flavin adenine dinucleotide (FAD) in common food samples by capillary zone electrophoresis (CZE) with laser-induced fluorescence detection The sensitivity of the detector

is two or three orders of magnitude higher than a conventional fluor-escence detector This allows dilution of the sample extract, thereby reducing matrix effects and the interference of other native fluorescent compounds Samples were extracted with a solvent mixture and centri-fuged, with no requirement for additional cleanup

In a CZE method applied to the determination of niacin in fortified and natural foods [59 –61], sample cleanup by solid-phase extraction concen-trated the sample and provided the opportunity to redissolve the purified extract in a solvent that was compatible with the running buffer Analysis

of a standard reference cereal sample (VMA 195) gave an average concen-tration of niacin that compared well with the certified value [59] Niacin values obtained for the various cereals, meats, and other foods compared favorably with results obtained by HPLC [59,61] Analysis of concen-trated yeast spreads gave results that were in good agreement with results obtained by the AOAC colorimetric procedure [60]

Capillary electrophoresis has been applied to the determination of vitamin C in fruits and vegetables [62] and citrus juice [63] using MECC and CZE modes, respectively In both applications, results compared well with those obtained by HPLC Choi and Jo [64] used the technique of sample stacking to concentrate sample extracts

23.5 Flow-Injection Analysis

Flow-injection analysis has been used to automate several AOAC colorimetric and fluorometric methods for water-soluble vitamins, with

a resultant improvement in precision The use of immobilized ascorbate oxidase in flow-injection systems with amperometic detection [65,66] offers the opportunity for a precise, accurate, and automated vitamin C analysis using relatively simple apparatus

23.6 Biospecific Methods

protein-binding assays (EPBAs), and biomolecular interaction analysis

(BIA) use microtitration plates, thereby allowing the batchwise analysis

of large numbers of sample extracts, without the problems of handling and disposal of radioactive material encountered with radio-labeled protein-binding assays (RPBAs) and radioimmunoassays The high tech-nology is built into the reagents, so the assays are simple to perform using automated procedures It is still necessary, however, to chemically

or enzymatically extract the vitamins from the food matrix, as these biospecific techniques depend upon the vitamins being in their free forms

yet been reported Antisera raised against PL respond mainly to PM, although one antiserum exhibited 80% cross-reactivity with PN and could therefore be used in an ELISA for the determination of added PN

in fortified foods [67]

An ELISA developed by Finglas et al [68] can substitute for the microbiological assay in the determination of total pantothenic acid in foods, as shown by the high correlation for the analysis of foods using the same extract The detection limit of the ELISA was at least tenfold lower than that of the microbiological assay

Individual EPBAs have been reported for the determination of

and R-protein as the respective vitamin-specific binding proteins The biotin EPBA [69] has broad specificity and will include analogs

of this vitamin, making it comparable to the microbiological assay The folate EPBA [70] shows similar responses for 5-methyl-THF and 5-formyl-THF, but a different response for folic acid Therefore, it is not possible to assay 5-methyl-THF, 5-formyl-THF, and folic acid simultaneously However, the folate content of nonfortified foods (containing mainly 5-methyl-THF and 5-formyl-THF) and the added folic acid in fortified foods can be determined in separate assays, using 5-formyl-THF and folic acid as the respective calibrants Folate results from the EPBA and the L rhamnosus assay of Phillips and Wright [71] have been compared using regression analysis [72] The data gave a linear relationship over the range 0– 400 mg/100 g with a correlation coefficient, r, of 0.939, which signified good agreement The folate values obtained using both assays were significantly higher than values reported in food composition tables, which were obtained using the

L rhamnosus method of Bell [73] In an interlaboratory study [74], the analysis of a candidate reference material (lyophilized Brussels sprouts) by EPBA yielded results that were highly variable, indicating that further work is necessary to standardize the methodology The

the sensitivity to measure the levels of naturally occurring vitamin B12

in foods