DEVRIES ET AL.:JOURNAL OF AOAC INTERNATIONAL VOL. 88, NO. 1, 2005 5

DIETARY SUPPLEMENTS

Microbiological Assay-Trienzyme Procedure for Total Folates in Cereals and Cereal Foods: Collaborative Study

JONATHAN W. D EVRIES Medallion Laboratories, General Mills Inc., 9000 Plymouth Ave North, Minneapolis, MN 55427 JEANNE I. RADER U.S. Food and Drug Administration, Center for Food Safety and Nutrition, 5100 Paint Branch Pkwy, College Park, MD 20740-3835 PAMELA M. KEAGY and CAROL A. HUDSON U.S. Department of Agriculture, Western Regional Research Center, 800 Buchanan St, Albany, CA 94710

Collaborators: G. Angyal; J. Arcot; M. Castelli; N. Doreanu; C. Hudson; P. Lawrence; J. Martin; R. Peace; L. Rosner; H.S. Strandler; J. Szpylka; H. van den Berg; C. Wo; C. Wurz

In 1996, U.S. Food and Drug Administration the standard test tube method; others used the regulations mandated the fortification of enriched microtiter plate modification for endpoint cereal-grain products with folic acid, thereby quantitation with equal success. For the required emphasizing the need for validated methods for products, the relative standard deviation between total folates in foods, particularly cereal products. laboratories (RSDR) ranged from 7.4 to 21.6% for The AOAC Official Methods (944.12, 960.46) 8 fortified (or enriched) products compared with currently used for the analysis of folate in foods for expected (Horwitz equation-based) values of compliance purposes are microbiological methods. 11–20%. RSDR values were higher (22.7–52.9%) for When the fortification regulations were finalized, no 2 unfortified cereal-grain products. For the optional Official AOAC or Approved AACC methods for products, the RSDR ranged from 1.8 to 11.2% for folate in cereal-grain products were in place. The 8 fortified products. RSDR values were higher AOAC Official Method (992.05) for folic acid in infant (27.9–28.7%) for 2 unfortified cereal-grain products. formula does not incorporate important Based on the results of the collaborative study, the improvements in the extraction procedure and was microbiological assay with trienzyme extraction is not considered suitable for the analysis of folates in recommended for adoption as Official First Action. foods in general. A microbiological assay protocol using a trienzyme extraction procedure was prepared and submitted for comments to 40 laboratories with recognized experience in folate olate and its polyglutamyl homologs (Figure 1) are analysis. On the basis of comments, the method Fessential vitamins for humans. Naturally occurring was revised to have the conjugase folates are reduced derivatives; fully oxidized folic acid is (gamma-glutamyl-carboxy-peptidase) treatment only found in the diet when foodstuffs are fortified or enriched follow a protease treatment, to include the use of (i.e., supplemented with folic acid) or when naturally occurring cryoprotected inoculum, and to include the folates are oxidized. Deficiency of folate leads to megaloblastic spectroscopic standardization of the standard and anemia and a general impairment of cell division that is more optional use of microtiter plates. Thirteen apparent in tissues that turn over rapidly (e.g., the laboratories participated in a collaborative study of hematopoietic system and cells lining the digestive tract). 10 required and 10 optional cereal-grain products, Weakness, tiredness, diarrhea, and anorexia are associated with including flour, bread, cookies, baking mixes, and folate deficiency. Recent studies have provided evidence that ready-to-eat breakfast cereals. The majority of the folate supplementation reduces the incidence of neural tube participating laboratories performed the assay by defects in certain high-risk populations (1–4). Studies are currently in progress to determine whether folate deficiency per se may contribute to the development of atherosclerosis (5, 6). Submitted for publication August 2004. In 1996, the U.S. Food and Drug Administration (FDA) The recommendation was approved by the Methods Committee on Commodity Foods and Commodity Products as First Action. See “Official finalized regulations mandating the fortification of enriched Methods Program Actions,” (2004) Inside Laboratory Management, cereal-grain products with folic acid. Accurate measurement of July/August issue. folate content of foods is essential for purposes of food labeling Corresponding author’s e-mail: [email protected]. and for the development of databases of folate content of foods. 6 DEVRIES ET AL.:JOURNAL OF AOAC INTERNATIONAL VOL. 88, NO. 1, 2005

Figure 1. Folic acid structure (various forms available in the diet).

For food labeling and database purposes, folates have Collaborative Study traditionally been analyzed with microbiological assays, such as AOAC Official Methods 944.12, 992.05,and960.46. For the collaborative study, cereal-based products with a AOAC Official Method 992.05 uses a conjugase enzyme range of protein, fat, and carbohydrate (fiber and starch) were treatment during extraction for conversion of the polyglutamyl selected (Table 2). Products were selected to cover a wide range vitamers to folic acid. Advances in folate methodology were of folate levels, including both fortified and unfortified made by a number of workers (7–13). In 1990, DeSouza and products. Results from individual test samples were used to Eitenmiller (14) showed that the use of protease and amylase in determine overall variability (SDR,R,andRSDR). Twenty addition to the conjugase significantly increased the measured different test samples were prepared by grinding, level of folate in foods. Also in 1990, Martin et al. (15) homogenizing, sealing in glass jars, and encoding. demonstrated the broad applicability of the trienzyme All food products for the collaborative study were (conjugase, protease, and amylase) methodology to a wide purchased at local supermarkets in quantities adequate to variety of foods. Other researchers (16–19) further ensure that homogeneous test samples would be supplied to all demonstrated the application of the trienzyme system. Finally, collaborators. After grinding (if necessary), test samples were in 1998, Rader et al. (20) found that digesting the test portion mixed to ensure homogeneity, sealed in glass jars, encoded, and with protease followed by a protease deactivation step before placed in the dark for storage until shipment. Collaborators treatment with the other enzymes resulted in the highest were provided with instructions to analyze individual test samples and report the results. Collaborators were provided measured folate level (Table 1). Laboratories performing the with 2 sets of test samples: one set of 10 was labeled “required” basic AOAC Official Methods were applying the trienzyme and were to be analyzed first. The second set was labeled procedure in a variety of ways, including the use of microtiter “optional” to allow the laboratory to analyze them as time and plates instead of test tubes for the determinative Lactobacillus casei growth and turbidity measurement steps.

Although the microbiological assay for folate was used for Table 1. Effect of sequence of enzyme pretreatments, several decades, significant differences in scientific opinion mg/100 g test sample found existed regarding the optimum procedures. To ascertain that the optimum method would be collaboratively studied, a complete Trienzyme, Trienzyme, Product Conjugase conjugase first protease first review of the literature was undertaken, a method based on the review was drafted in AOAC format, and input from analysts active in folate analysis was sought. The method was submitted All-purpose flour 26.1 ± 1.25 33.1 ± 0.3 41.3 ± 1.4 to 40 laboratories active in folate analysis for comment and Rotelle 31.4 ± 1.65 39.6 ± 1.3 49.6 ± 2.2 feedback. The written method was revised and the microtiter Long-grain and 50.7 ± 8.3 59.4 ± 1.3 71.1 ± 3.5 plate end point measurement option was included. Based on the wild rice comments of the laboratories, the final version of the method Bran cereal 629.3 ± 48.05 677.8 ± 52.2 665.8 ± 29.6 intended for collaborative study was prepared. DEVRIES ET AL.:JOURNAL OF AOAC INTERNATIONAL VOL. 88, NO. 1, 2005 7

Table 2. Characteristics of samples tested, percentage makeup

Product Solids Protein Fat Carbohydrate Fiber

Wheat flour, fortified white 88.1 10.3 1.0 76.3 2.7 Wheat flour, unfortified white 88.1 10.3 1.0 76.3 2.7 Wheat flour, whole wheat 89.7 13.7 1.9 72.6 12.6 White bread from wheat flour 63.3 8.2 3.6 49.5 2.3 Macaroni, dry 89.7 12.8 1.6 74.7 2.4 Cookies, low-fat, gingersnap 94.7 5.6 9.8 76.9 Cookies, fat-free oatmeal 87.5 5.9 1.5 78.6 7.3 Cookies, high-fat shortbread 96.3 6.1 24.4 64.5 1.6 Cookies, oatmeal, regular 94.3 6.2 18.1 68.7 3.1 a RTE cereal, corn flakes 97.4 8.1 0.3 86.1 6.0 RTE cereal, extruded oat 95.0 15.1 6.4 69.1 12.0 RTE cereal, wheat flakes 95.4 9.5 1.7 9.5 10.0 RTE cereal, crispy rice 97.5 6.4 0.4 88.6 1.4 RTE cereal, extruded bran 97.0 14.3 1.8 74.4 29.9 Pancake mix, low-fat 90.9 10.0 1.7 73.6 2.7 Biscuit mix, high-fat 90.8 8.0 15.4 63.3 2.3 Tortillas 55.9 5.7 2.5 46.6 5.2 Corn meal 88.4 8.5 1.7 77.7 5.2 a RTE = Ready-to-eat.

resources allowed. Laboratories were given the option of using behind safety barrier with hot water, steam, or the test tube method, the microtiter plate method, or both. electric mantle heating. Use effective fume Because laboratories conducting this study were generally removal device to remove flammable vapors as familiar with microbiological assays, no practice samples were produced. Leave ample headroom in flask and add supplied. Laboratories were encouraged, but not required, to boiling chips before heating is begun. All controls, purchase and include NIST SRM 1846 as a reference material unless vapor sealed, should be located outside of for the study. SRM 1846, available from the National Institutes vapor area. (5) Toxic liquids: Use effective fume of Standards and Technology (NIST), is a spray-dried, removal device to remove vapors as produced. milk-based infant formula. The mass fraction value for folic Avoid contact with skin. acid in this material is 1.29 ± 0.28 mg/kg. SRM 1846 was also included as an unknown test sample among those sent as part of Method uses the organism Lactobacillus casei subspecies the required examination set. rhamnosus-ATCC 7469 to determine the amount of folate present in foods and vitamin concentrates. Method can also be AOAC Official Method 2004.05 semiautomated through the use of automated dilution and Total Folates in Cereals and Cereal Foods turbidity reading instruments or the 96 well microtiter plate and Microbiological Assay-Trienzyme Procedure reader system. Note: Folates are light and oxygen sensitive. Use First Action 2004 of yellow lighting and low actinic glassware is recommended. (Applicable to cereal grains and cereal grain foods Preparation and storage of samples under subdued lighting is containing added folate (folic acid) or natural occurring folates essential. with levels from 7.6 mg/100 g to 100% folate.) A. Principle See Table 2004.05A for the results of the interlaboratory study supporting acceptance of the method. Foods, with water added, are autoclaved to break up particles, gelatinize starch, destroy extraneous microorganisms, Caution: See Appendix B for laboratory safety. and denature proteins to enhance enzymatic attack and make (1) Ammonium hydroxide. (2) Hydrochloric acid. folate more extractable. Folate (pteroylglutamic acid in various (3) Potassium or sodium hydroxide: Extremely forms) occurs naturally in foods bound to glutamic acid caustic; can cause severe burns; protect skin and residues of varying chain lengths. Most naturally occurring eyes. (4) Flammable liquids: Perform operations folates cannot be used by the assay organism. Folic acid 8 DEVRIES ET AL.:JOURNAL OF AOAC INTERNATIONAL VOL. 88, NO. 1, 2005

Table 2004.05A. Interlaboratory study results of trienzyme method for determination of folate in cereal grains and grain products

Single Double No. Mean, Recovery, Grubbs Grubbs a Material of labs mg/100 g sR RSDR,% R HORRAT % (Lab ID) (Lab ID)

Flour, fortified 13(1) 134.1 15.9 11.83 44.4 0.77 15 Flour, unfortified 12(2) 59.8 13.5 22.67 37.9 1.31 2, 4 Bread, enriched 12(2) 114.3 12.9 11.31 36.2 0.72 2, 10 Pasta, enriched macaroni 13(1) 223.5 28.8 12.88 80.6 0.91 2 Cereal oats 14(0) 666.8 144.1 21.61 403.5 1.80 Cereal wheat 12(2) 1365.6 101.4 7.42 283.8 0.69 4, 7 Cereal rice 14(0) 684.9 147.3 21.51 412.6 1.80 Tortilla 13(0) 34.8 18.4 52.94 51.5 2.82 Baking mix 14(0) 173.3 37.4 21.59 104.8 1.47 Standard Reference Material 12(0) 145.6 21.4 14.71 60.0 1.07 113 a Laboratories retained after outliers removed (number of outlier laboratories removed in parentheses). Each laboratory ran one test per material.

(pteroylglutamic acid) is extracted from the matrix by a triple (k) Test tube rack holder.—Large, with cover. To hold 4 or enzyme system. A protease and an amylase are used to digest more test tube racks. the food matrix and aid in the release of folates. Desiccated (l) Filter paper.—No. 2V, 12.5 cm. chicken pancreas conjugase is used to hydrolyze (m) Balances.—Analytical, readability to at least 4 places; folylpolyglutamates to folyldiglutamates which, along with and top loading, 3 places. folic acid, can be used by the assay organism. The freed folates (n) Hot plate stirrer. are extracted and diluted with basal medium containing all (o) Syringe.—Fitted with long needles, sterilizable, capable required growth nutrients except folate, and the turbidity of the of delivering 50 mL. Lactobacillus casei subspecies rhamnosus growth response for the test samples is compared quantitatively to that of known (p) Desiccator. standard solutions. The method allows the optional use of (q) Optional for tube assay.—Automated assay tube 96 well microtiter plates and a microplate reader in place of the reading; Gilson (Middleton, WI) escargot fractionator Model standard test tube system for semiautomated turbidity SC-30 or Model 222 sample changer modified with air determination. agitation system and connected to spectrophotometer with flow cell and either printer or computer, or Autoturb II diluter and B. Apparatus and Materials reader (Mitchum-Schaefer Inc., Indianapolis, IN). (r) Optional microtiter plate system.—96 Well microtiter (a) Spectrophotometer.—To read 20 ´ 150 mm (or 18 ´ plate reader and 96 well microtiter plates. A reader with 150 mm) test tubes (Agilent, Palo Alto, CA). appropriate filter(s) for 570–630 nm and efficient software for (b) Water bath.—Covered, 37° ±0.2°C, with rotary shaker. calculation is suitable. Note: For some plate reader systems, (c) Test tubes.—Disposable, glass, Borosilicate, 20 ´ 150 or ´ blackwall microtiter plates may be necessary to prevent 18 150 mm. deviations in readings between edge row wells and center row (d) pH meter.—With long combination electrode. wells. Use of microtiter plate systems requires (e) Vortex mixer. filter-sterilization of solutions using 0.22 mm sterilization filter (f) Volumetric flasks.—Class A, low actinic and clear; 25, units (15 and 500 mL, and 1 L), available as Nalge Disposable 50, 100, 250, 500, and 1000 mL. Filterware (Nalge Nunc International, Rochester, NY) or (g) Repetitive pipetters.—50 mL capacity with 1.25 mL Corning Disposable Filterware (Corning, NY). disposable tips; and adjustable digital 100–1000 mL. (h) Pipeting machine(s).—To deliver 1 and 5 mL aliquots. C. Reagents Optional pipetter.—12 channel to use with 96 well microtiter plates. Use distilled or double-distilled water throughout. (i) Volumetric pipets.—ClassA;1,2,3,4,5,10,20,25,and (a) Adenine sulfate. 50 mL. (b) Agar.—Bacto, Difco No. 0140-01 (BD, Franklin (j) Centrifuge tubes.—Oak Ridge type 28.6 ´ 106.1 mm, Lakes, NJ). polypropylene, reusable; and disposable, 50 mL. (c) Agar.—Bacto Lactobacilli agar (Difco No. 0900-15). DEVRIES ET AL.:JOURNAL OF AOAC INTERNATIONAL VOL. 88, NO. 1, 2005 9

(d) Alpha-amylase.—From Aspergillus oryzae, phosphate monobasic in water and dilute to 1 L with water. 150–250 U/mg protein (burette), Sigma Chemical Co. (St. Adjust pH to 7.0 with 4M potassium hydroxide. Louis, MO) No. A-6211 is suitable. (jj) Potassium hydroxide.—4M. Dissolve 224 g potassium (e) p-Aminobenzoic acid (PABA). hydroxide in ca 500 mL water. Caution: Cool; dilute to 1 L with (f) Antifoam.—Mix 1.5 mL Dow Corning (Midland, MI) water. Antifoam AF or B with 100 mL water, and mix. Shake before (kk) Sodium hydroxide.—4M. Dissolve 160 g sodium using. hydroxide in ca 500 mL water. Caution: Cool; dilute to 1 L with (g) Ascorbic acid. water. (h) L-Asparagine monohydrate. (ll) Sodium hydroxide.—0.01M. Pipet 2.5 mL 4M sodium (i) Biotin. hydroxide into 1 L volumetric flask. Dilute to volume with (j) Calcium pantothenate. water. (k) Casein hydrolysate.—Vitamin-free acid hydrolyzed (mm) Adenine–guanine–uracil solution.—Dissolve 1.0 g (Humko Sheffield Chemical, Lyndhurst, NJ). Hy-Case amino, each of adenine sulfate, C(a), guanine hydrochloride, C(r), and ICN Cat. No. 104864, is suitable. uracil, C(bb), in 50 mL warm HCl (1 + 1), C(ff), cool, and (l) Celite.—Optional as filter aid. dilute with water to 1 L. Store at 10°C. Prepare fresh every (m) Conjugase source.—Chicken pancreas, desiccated, 3 months. Difco Bacto Cat. No. 0459-12. No substitutions. (nn) Xanthine solution.—Suspend 1.0 g xanthine, C(cc), in ° (n) L-Cysteine×HCl. 150–200 mL water, heat to ca 70 C,add30mLNH4OH (2 + 3), (o) Ferric sulfate heptahydrate. C(ee), and stir until solid dissolves. Cool and dilute to 1 L with ° (p) Folic acid.—Reference standard USP Cat. No. 28600. water. Store at 10 C. Prepare fresh every 3 months. (q) Glutathione. (oo) Asparagine solution.—Dissolve 10 g L-asparagine monohydrate, C(h), in water and dilute to 1 L. Store at 10°C. (r) Guanine hydrochloride. (pp) Vitamin solution for folate.—Dissolve 10 mg PABA, (s) Liver.—Difco No. 0133-01. C(e), 40 mg pyridoxine hydrochloride, C(w),4mgthiamin (t) Niacin. hydrochloride, C(z), 8 mg calcium pantothenate, C(j), 8 mg (u) Octanol. niacin, C(t), and 0.2 mg biotin, C(i), in ca 300 mL water. Add (v) Protease.—From Streptomyces griseus, powder Type 10 mg riboflavin, C(x), dissolved in ca 200 mL 0.02M acetic XIV ca 4 U/mg solid, Pronase E from Sigma Chemical Co. No. acid, C(dd). Add solution containing 1.9 g anhydrous sodium P-5147. acetate and 1.6 mL acetic acid in ca 40 mL water. Dilute to 2 L (w) Pyridoxine hydrochloride. with water. Prepare fresh every 2 months. (Not necessary to (x) Riboflavin. prepare if using commercial basal medium preparation in (y) Sodium ascorbate. reagent, C(uu)). (z) Thiamin hydrochloride. (qq) Saline.—Sterile. Dissolve 9 g NaCl in 1 L water. (aa) D,L-Tryptophan. Dispense 10 mL portions to 20 ´ 150 mm test tubes capped (bb) Uracil. with plastic top. Sterilize 15 min at 121°–123°C and store in (cc) Xanthine. refrigerator. Prepare fresh weekly. × (dd) Acetic acid.—0.02M. Dissolve 1.2 mL glacial acetic (rr) Mixed salts solution.—Dissolve 20 g MgSO4 7H2O, × × acid to ca 500 mL water. Dilute to 1 L with water. 1gNaCl,1gFeSO4 7H2O,and1gMnSO4 H2O in water. (ee) Ammonium hydroxide (NH4OH; 2 + 3).—Carefully Dilute to 1 L. Add 1 mL concentrated HCl. Prepare fresh every mix 2 volumes concentrated ammonium hydroxide with 3 months. 3 volumes water. (ss) PABA–vitamin B6 solution.—Dissolve 50 mg PABA, (ff) Hydrochloric acid (HCl; 1 + 1).—Carefully mix equal C(e), and 120 mg pyridoxine hydrochloride, C(w), in 200 mL volumes concentrated HCl and water (see Caution). water. Add 0.95 g sodium acetate and 0.8 mL acetic acid to ca (gg) Phosphate buffer (extract buffer).—pH 7.8. Dissolve 40 mL water. Combine the 2 solutions and dilute to 500 mL 1.42 g sodium phosphate dibasic and 1.0 g ascorbic acid in with water. Store at 10°C. water and dilute to 100 mL. Adjust pH to 7.8 with 4M NaOH. (tt) Agar maintenance medium.—Into1Lhotwater Prepare fresh on day used. Assay requires ca 35 mL buffer for containing 10 mL 100 ng/mL vitamin B12, dissolve 48 g each test sample and standard. Lactobacilli agar, C(c),and3gBactoagar,C(b). After agars (hh) Phosphate buffer (assay buffer).—pH 6.8. Dissolve dissolve, dispense 10 mL portions to 20 ´ 150 mm test tubes, 1.42 g sodium phosphate dibasic and 1.0 g ascorbic acid in ca and cap. Cover tubes to prevent contamination, autoclave 85 mL water and dilute to 100 mL. Adjust pH to 6.8 ± 0.02 with 15 min at 121°–123°C, and store in refrigerator. Will keep 4M NaOH. Prepare fresh on day used. Assay requires indefinitely. 12–15 mL buffer for each test sample and standard tube for test (uu) Folic acid-free, double strength basal tube assay; ca 2.5 mL buffer for each test sample and standard medium.—Prepare as in Table 2004.05B. [Alternatively, for microtiter plate assay option. commercially available Difco Folic Acid Casei Medium (No. (ii) Phosphate buffer (for making standard 0822-15) can be used. Prepare as directed: Suspend 9.4 g Difco solutions).—0.1M, pH 7.0. Dissolve 13.61 g potassium Casei Medium and 50 mg ascorbic acid in 100 mL water, boil 10 DEVRIES ET AL.:JOURNAL OF AOAC INTERNATIONAL VOL. 88, NO. 1, 2005

Table 2004.05B. Preparation of basal medium

Amount of basal medium to prepare, mL

250 500 1000

Add reagents in order listed, mL

Adenine–guanine–uracil solution, C(mm)2.5510 Xanthine solution, C(nn) 5 10 20 Asparagine solution, C(oo)153060 Vitamin solution for folate, C(pp) 50 100 200 Mixed salts solution, C(rr) 5 10 20

PABA–vitamin B6 solution, C(ss)2.5510 Add ca 100 mL water and the following solids, g

Vitamin-free casein, hydrolyzed 2.5 5 10 Dextrose 10 20 40 Potassium phosphate, dibasic 0.25 0.5 1 Potassium phosphate, monobasic 0.25 0.5 1

Sodium acetate–3H2O 16.6 33.2 66.4 Glutathione, C(q) 0.00125 0.0025 0.005 Dissolve the following solids, g, in dilute HCl and add to above solution

L-Cysteine×HCl, C(n) 0.125 0.25 0.5 D,L-Tryptophan, C(aa) 0.05 0.1 0.2 Mixwell,adjusttopH6.8withNaOH,anddiluteto1Lwithwater

for 1–2 min, and cool.] Dispense 200 mL portions into 16 oz (3) Diluted standard solution.—100 ng/mL. For use in polyethylene bottles, cap, and freeze (use within 2 months). working inoculum. Dilute 10 mL working standard solution, (vv) Standard solutions.—(Use low actinic glassware.) C(vv)(2), to ca 90 mL with water and adjust pH to ca 7.5 with (1) Stock solution.—100 mg/mL. Accurately weigh 50 mg USP HCl. Dilute to 100 mL with water. Top with enough toluene to folic acid, C(p), that has been dried to constant weight in vacuo keep surface covered (usually 3–5 mL). Store in refrigerator. ° at 70 C and stored in dark over P2O5 in desiccator and dissolve Prepare fresh on day of use. in 0.1M, pH 7.0, phosphate buffer, C(ii), in 500 mL volumetric (ww) Liquid culture medium.—50% Basal medium with flask. Dilute to volume with 0.1M phosphate buffer. Top with 0.4 ng/mL folic acid and 10 mg/mL solubilized liver. Suspend enough toluene to keep surface covered (usually 3–5 mL). 0.1 g liver, C(s), in 100 mL water. Hold mixture for 1 h at 50°C, Store in refrigerator. Check purity of standard and verify and filter. Top with enough toluene to keep surface covered concentration of stock solution by pipetting 10 mL stock (usually 3–5 mL). Store in refrigerator. Pipet 20 mL to 1 L solution to 100 mL volumetric flask and diluting to volume volumetric flask. Add (via pipet) 8 mL diluted folic stock with 0.1M, pH 7.0, phosphate buffer. Measure absorbance of solution (100 ng/mL), C(vv)(3). Dilute to volume with water. solution at 282 and 346 nm in AUs using 0.1M, pH 7, Mix equal volumes of solution with basal medium solution, phosphate buffer as a blank. Folic acid concentration in the C(uu). Dispense 10 mL portions diluted medium to 20 ´ stock solution: 150 mm screw-cap test tubes, autoclave 15 min at 121°–123°C, and cool tubes rapidly. Store in refrigerator. Difco Bacto micro Folic acid, mg/mL = inoculum broth No. 0320-02 is also satisfactory as liquid absorbance/absorptivity ´ dilution factor ´ 1000 ´ MW culture medium. Will keep indefinitely. (xx) Conjugase solution.—Chicken pancreas solution where absorptivity = 27 600 at 282 nm and 7200 at 346 nm, (5 mg/mL). Weigh 0.5 g chicken pancreas, C(m), and add MW = 441.4. 100 mL buffer, pH 7.8, C(gg). Stir vigorously for 10 min. (2) Working standard solution.—1 mg/mL. Dilute 5 mL Transfer to 20 ´ 150 mm test tubes and centrifuge for 10 min at stock solution to ca 475 mL with water and adjust pH to ca 7.5 2000 rpm. Decant supernatant through glass wool pledget into with HCl. Dilute to 500 mL with water. Prepare fresh on day of beaker, cover with parafilm, and store in refrigerator. Prepare use. This will be diluted further when standards are run in fresh on day of use. Each test solution and standard requires parallel with samples. 4–5 mL conjugase solution. DEVRIES ET AL.:JOURNAL OF AOAC INTERNATIONAL VOL. 88, NO. 1, 2005 11

(yy) Alpha-amylase solution.—20 mg/mL. Dissolve 0.5 g standards to pH 4.5 with HCl (1 + 1) and dilute to 100 mL with alpha-amylase, C(d), in 25 mL water. Store in refrigerator. water. Filter ca 20 mL through 2V filter paper. Use of Celite is Prepare fresh on day of use. Each test solution and standard acceptable to obtain clear filtrate. (Note: Filtered solutions can requires 1 mL alpha-amylase solution. be set aside in the dark at 4°C overnight.) (zz) Protease solution.—2mg/mL.Dissolve0.05g (3) Options.—For standard tube/spectrometer assay, protease, C(v), in 25 mL water. Filter through glass wool proceed to next step, E(c). If using 96 well microtiter plate, pledget if necessary and store in refrigerator. Prepare fresh on filter-sterilize ca 5 mL filtered solution, using 0.22 mm filter day of use. Each test solution and standard requires 1 mL system, into sterile screw-cap vials and store in dark at 4°C until protease solution. ready to assay. Proceed directly to F(b). D. Inoculum (c) Diluted standard and sample extracts (test tube method only).—(1) Standard.—0.3 ng/mL. Pipet 15 mL standard (a) Test organism.—Lyophilized L. casei subspecies extract into 500 mL volumetric flask. rhamnosis, ATCC 7469. (2) Test samples.—Pipet 10 mL of each test sample extract (b) Cryprotected inoculum.—Dissolve 4.7 g Difco Folic into respective 500 mL volumetric flasks. Add equal volume Acid Casei Medium in 50 mL distilled water. Heat to boiling, phosphate buffer, pH 6.8, to each standard and test flask. (If cool in ice, then add 50 mL water and 25 mg sodium ascorbate. additional serial dilutions are necessary, add equal volume of Add 0.5 mL diluted folic standard solution, C(vv)(3), mix, and buffer to volume pipetted to flask.) Dilute to volume with sterile-filter. Suspend one vial lyophilized L. casei in 1.0 mL of water. If concentration of folate is low, use larger aliquot; if above medium using sterile techniques. Transfer 0.15 mL of too high, use smaller aliquot or larger volumetric flask. this suspension to culture medium. Incubate at 37°C for 18 h. Mix 120 mL glycerol and 30 mL water, autoclave, and cool in F. Determination ice bath. Cool incubated bacterial culture in ice bath and add 100 mL of the cold 80% glycerol. Mix gently but well. Aliquot (a) Assay using test tubes.––(Use of Brewer automatic 2.0 mL aliquots into sterile tubes (stir suspension occasionally pipetting machine is recommended.) to maintain even suspension). Can be stored at –20°C for up to Prepare twenty-one 20 ´ 150 mm tubes to contain 0, 0, 3 months and at –70°C for up to 6 months. 0.02, 0.04, 0.06, 0.08, and 0.10 ng/mL folic acid by adding in (c) Working inoculum.—For tube cultures, dilute 2.0 mL triplicate 0, 0, 1, 2, 3, 4, and 5 mL diluted (0.3 ng/mL) standard frozen suspension, (b), to 50 mL with sterile saline. Mix on a extract, E(c), into the tubes. Three 0 mL tubes will be used as Vortex mixer. Use this suspension as working inoculum. uninoculated blanks. For 96 well plates, add 5 mL sterile saline to 2.0 mL Prepare twelve 20 ´ 150 mm tubes for each test sample solution direct from freezer. Mix on a Vortex mixer. Use this solution. In triplicate, place 1, 2, 3, and 4 mL diluted test suspension as working inoculum. sample extracts into the tubes. Add water to each to a total volume of 5 mL. Add 5.0 mL folic acid-free double strength E. Preparation of Test Solutions basal medium, C(uu), to each tube. Cover tubes to prevent ° ° (a) Product preparation.—Grindsolidstopass40mesh bacterial contamination. Autoclave 6 min at 121 –123 C. ° sieve and store in air-tight containers, free from light exposure. Cool tubes as rapidly as possible to <40 C to minimize (b) Hydrolysis and extraction.—Use distilled water unless browning reactions between amino acids and sugars in the otherwise indicated. basal medium which darken the color and reduce the (1) Standard.—10 ng/mL. Pipet 1.0 mL working standard availability of essential amino acids, and to prevent injury to solution (1 mg/mL), C(vv)(2), to 125 mL Erlenmeyer flask the inoculum. Treat all blanks, standards, and test solutions containing 20 mL phosphate buffer, pH 7.8, C(gg), mix, and identically. add 30 mL water. Aseptically inoculate each tube, except for one set of (2) Test samples.—Accurately weigh test portion equal to triplicate blank tubes (containing 0 mL standard extract) with 0.25–1.0 g dry basis solids containing ca 1 mgfolicacidinto a50mL drop working inoculum delivered from a sterile respective 125 mL Erlenmeyer flask. Add 20 mL phosphate syringe fitted with a long needle. To ensure that drop falls onto buffer, pH 7.8, and mix thoroughly. (If product is low in folate, surface of tube contents, angle syringe slightly. Incubate at do not take >1.0 g. To compensate for lower levels, use a larger 37°C for 22 h. aliquot in the second serial dilution below.) Dilute with water to (b) Assay using optional 96 well microtiter plate.—(Note: 50 mL. Add 0.1–1.0 mL octanol (antifoam) to all flasks. Cover To prevent contamination of sterile microtiter plates and flasks with 50 mL beaker, autoclave 15 min at 121°–123°C, and solutions, be sure that bench area is clean and not susceptible to cool. Add additional 10 mL buffer to each flask. Add 1 mL airborne contaminants. A biosafety hood will reduce the risk of protease solution and incubate 3 h at 37°C. Inactivate enzyme contamination.) by autoclaving (or placing in boiling water bath) 3 min at Using 0.22 mm filter system, filter-sterilize phosphate buffer, 100°C,andcool.Add1mLalpha-amylase solution to each pH 6.8, C(hh), and basal medium, C(uu), 2.5 mL of each flask, cover, and incubate 2 h at 37°C. Add 4 mL conjugase solution per test sample plus ca 70 mL extra. Using 12 channel solution, cover, and incubate 16 h at 37°C. Inactivate enzymes pipetter, pipet 150 mL buffer, pH 6.8, C(hh), to wells A1–H12 by autoclaving 3 min at 100°C, and cool. Adjust tests and of sterile microtiter plate. 12 DEVRIES ET AL.:JOURNAL OF AOAC INTERNATIONAL VOL. 88, NO. 1, 2005

Pipet 150 mL standard extract (from F above) to wells Table 2004.05C. Potential sources of error m G1–G2. For each sample extract, pipet 150 L in duplicate into Cause Type of error in result high/low wells in row G, i.e., 150 mL sample one extract to each well G3 and G4, 150 mL sample 2 extract to G5 and G6, etc., up to G12. Incomplete wetting of Low random (Therefore, 5 sample extracts and one standard extract per sample at extraction microtiter plate can be run.) Mix contents of each well to Sample splattered during mixing Low random homogeneity by using a 12 channel pipetter and repeating aspiration and delivery steps. Make serial dilutions (x2) of Incorrect preparation of extraction Low bias buffers standard and samples by transferring 150 mL from wells G1–G12 to F1–F12, and mix; then from F1–F12 to E1–E12, Weighing error High/low random etc. For samples A1–A12, withdraw 150 mL from each well Balance out of calibration High/low bias after mixing, and discard. Air bubbles in diluter Low random Add 1 mL working inoculum per 1 mL folic acid-free basal Diluter not in calibration Low bias medium. Mix well. Note: A minimum of 15 mL inoculated Pailing power board on High/low random media (15 mL inoculum in 15 mL medium) is needed per plate spectrophotometer (one standard and 5 samples). Prepare amount needed plus ca pH electrode calibation error Low bias 70 mL extra. Add 150 mL inoculated basal medium to wells Sample spilled Low random A1–H12. (H is the inoculated blank row.) Put plate into plastic bag, and seal. Incubate 22 h at 37°C. Place pan of water in Stress on assay organism Low bias incubator to ensure adequate humidity to inhibit evaporation of Incorrect assay medium Low bias to no water from outer wells. Remove plate and let stand ca 30 min to growth in assay tubes equilibrate to room temperature. Mix contents of each well using 12 channel pipetter by repeating aspiration and delivery steps until bacterial A random error can be high or low, but of indeterminate suspension becomes homogeneous. Read optical density of all magnitude. A high or low bias affects every individual analysis wells on microtiter plate reader at 595 or 600 nm using in the same way, at the same magnitude. inoculated blanks (H1–H12) as reference blank. Ref.: J. AOAC Int. 88, 7–12(2005)

G. Calculations Results and Discussion Prepare standard concentration response curve by plotting average % T reading for each level of standard solution used Eighteen sets of data were received from 14 laboratories, against the amount of standard folate contained in the including 5 sets for the microtiter plate turbidity end point respective tubes. Determine amount of folate per mL for each measurement option. One laboratory submitted 2 sets of data sample tube by interpolation from standard curve. Discard any for each of the method options, i.e., test tube and microtiter tubes with % T value for samples equivalent to <0.5 mL or plate. At that laboratory, one analyst performed the test tube >4.5 mL standard solution. Calculate concentration of folate in assay, while another analyst independently performed the extract solution for the sample using the remaining tubes. microtiter plate assay. Therefore, the test tube and microtiter Average the values obtained and calculate ±10% of the average. plate data sets were independent of each other, but the 2 sets of Accept tubes within the ±10% range, and if the number of tubes test tube data were not independent, and the 2 sets of microtiter ³ with acceptable values is 2/3 of the original number of tubes plate data were not independent from each other. Therefore, one used in the 4 levels of sample assay solution, calculate folate set of data of each type (test tube and microtiter plate) were content in original sample from average of acceptable tubes. If selected by flipping a coin. Thus, data from only one tube set £ 2/3 of tubes have acceptable values, then the assay must be for Laboratory 1 and from one microtiter plate set from repeated for the sample. Alternatively, computer programs Laboratory 1 were used. designed and validated for the calculation can be used. Data from one other laboratory (both test tube data and Optional microtiter plate reader.—Reduce data by plotting microtiter plate data) were at odds with the data from all the results and determining folate concentration in wells as above, other laboratories, with some discrepancies exceeding the data or use appropriate data analysis software (Microplate Manager, from the other laboratories by more than an order of magnitude. BioRad, Richmond, CA, is suitable). Average the results of the That laboratory indicated they observed precipitates in their duplicates. standard solution(s). Further attempts to resolve the reasons for H. Quality Assurance the significant discrepancies proved futile; therefore, all the data from the laboratory were removed from statistical Glassware must be low actinic and must be cleaned consideration. Laboratory 7 failed to provide data for meticulously and heated 1–2 h at 250°C to destroy any folic 2 samples, indicating they misunderstood the labeling between acid residues. Folic acid should be stored in a desiccator prior to the required and optional samples and ran 2 optional samples stock standard preparation. See Table 2004.05C. instead. Laboratory 4 provided no explanation for its missing DEVRIES ET AL.:JOURNAL OF AOAC INTERNATIONAL VOL. 88, NO. 1, 2005 13 135 130 109 119

b Figure 2. Comparison of test tube vs microtiter plate end points.

data point. The results submitted from the laboratories are shown in Table 3.

100 97 109 121 127 91 111 Statistical analysis was performed using standard statistical protocol, with outliers determined by the use of the 2-step Lab No. b Grubbs test. The results of the statistical treatment are shown in Table 2004.05A. Grubbs outliers are noted in Tables 3 and 2004.05A. The results of the collaborative study were excellent. For the required samples, the RSDR ranged from 7.4 Tube Microtiter plate

1470 1315 1450 1467 1410 1240to 1265 21.6% 1210 1323 for 8 fortified (or enriched) samples compared with expected (Horwitz equation-based) values of 11–20%. RSDR b values were higher (22.7–52.9%) for 2 unfortified cereal-grain 74 52 72 44 53 50 38 73 56 79

1875 samples. For optional samples, the RSDR ranged from 1.8 to 11.2% for 8 fortified samples. RSDR values were higher b b (27.9–28.7%) for 2 unfortified cereal-grain samples. The variability for the method is well within that expected for analytes at this level, as evidenced by the Horwitz ratio

54 127 (HORRAT column) which is within the expected range of 0–2 120220 115 288 128 266 131 229 56 225 175 230 223 203 225 193 212 except for one sample, namely, the white corn tortilla chip a (Figure 2). The reason for the higher than expected variation for b b b the white corn tortilla chips is unknown. g/100 g

m NIST SRM 1846 was included as part of the required sample set as a means of assessing accuracy of the methodology. The 1234781011121517156 NIST mass fraction value for folic acid in this material is 1.29 ± 0.28 mg/kg. The mean value obtained by the collaborators was 1.45 ± .23, well within the expected range. The value obtained in the collaborative study is slightly higher than the NIST value. This may be due to the fact that only conjugase enzyme was used for the original establishment of the NIST value. Finally, there was excellent agreement between the results of the method using test tubes versus microtiter plates for the determination step (Figure 2); therefore, both can be used interchangeably and effectively. Laboratories participating in the collaborative study were required to analyze the samples listed in Table 3. In addition, each laboratory was provided with 10 extra optional samples (Table 4) to analyze if resources permitted, to provide a broader picture of method capability and performance. Seven sets of data were submitted to the Study Directors (Table 4). Between-laboratory cereal: oats 615 972 730 793 671 763 715 633 609 743 546 684 377 484 c statistical analysis of the optional samples is shown in Table 5. The Each laboratory ran one test perGrubbs material. outlier. RTE = Ready-to-eat. Table 3. Interlaboratory results for determination of total folate, a b c RTE cereal: wheat flakes (highly fortified) 1332 1375 1530 1060 Bread: white, enriched 121 212 RTE Pasta: enriched macaroni 216 372 Required samples Flour: fortified, white 137 148 119 138 158 164 135 132 119 223 Flour: unfortified, whole wheat 72 110 RTE cereal: crispy riceBaking mix: all-purposeTortilla: white corn tortilla chipsStandard reference material: milk-based infant formula 169 190data for 139 the 590 35optional 154 993 252 66 samples 770 180 20 722 are 145 229 in 48 704 excellent 141 206 873 112 147 agreement 732 159 166 11 620 128 111 with 30 631 those 181 118 13 776 164 150 431 21 178 132 24 621 164 143 468 60 133 658 182 45 23 56 14 DEVRIES ET AL.:JOURNAL OF AOAC INTERNATIONAL VOL. 88, NO. 1, 2005

Table 4. Interlaboratory study results for determination of total folate in optional samples mg/100 g sample

Tube Microtiter plate

Lab No.

Sample 1-a 1-b 3 10 1-a 1-b 6

Flour: unfortified, white 41 32 20 29 39 40 43 Bread: buns (hamburger/hot dog) 128 137 115 118 98 119 136 Corn meal: yellow, enriched 190 215 182 190 174 200 193 a RTE cereal: corn flakes 634 644 710 639 492 568 596 RTE cereal: wheat bran flakes 360 356 455 380 307 346 402 Baking mix: pancake mix 126 141a 110 118 109 112 113 Cookies: gingersnaps 98 106 82 94 87 85 98 Cookies: animal crackers 29 29 14 18 33 25 28 Cookies: oatmeal (high fiber) 113 64 52 49 70 54 59 Muffin mix: plain, without fruit or topping 121 115a 104 104 142 100 103 a RTE = Ready-to-eat. obtained for the other samples of the study, further indicating the J. Langemeier, American Institute of Baking, Manhattan, KS breadth of applicability of the method to cereal matrixes. N. Rawley, American Institute of Baking, Manhattan, KS C. Pfeiffer, Centers for Disease Control and Prevention, Recommendations Atlanta, GA D. Horne, Department of Veterans Affairs, Nashville, TN In light of the excellent collaborative study results obtained T. Tamura, University of Alabama, Birmingham, AL with regard to between- and within-laboratory data, the Study J.F. Gregory, III, University of Florida, Gainesville, FL Directors recommend adoption of the method as First Action by We also thank the following collaborators for their AOAC INTERNATIONAL, and as a First Approval Method of participation in this study: Analysis of the American Association of Cereal Chemists. G. Angyal, U.S. Food and Drug Administration, Washington, DC Acknowledgements J. Arcot, University of New South Wales, Sydney, Australia M. Castelli, Ralston Analytical Labs, St. Louis, MO We thank the following scientists who participated in the N. Doreanu, The Long Co., Chicago, IL organization of the study or reviewed and commented on the C. Hudson, U.S. Department of Agriculture, Western methodology: Regional Research Center, Albany, CA

Table 5. Statistical results of trienzyme method for determination of folate in optional cereal grains and grain products

Optional samples Labs Mean, mg/100 g SD RSDR,% R HORRAT

a RTE cereal: corn flakes 5 631.40 54.01 8.55 151.23 0.71 RTE cereal: wheat bran flakes 5 387.80 43.41 11.19 121.55 0.86 Corn meal: yellow, enriched 5 196.00 12.43 6.34 34.80 0.44 Bread: buns (hamburger/hot dog) 5 128.33 12.52 9.75 35.06 0.63 Baking mix: pancake mix 4 113.25 3.40 3.01 9.52 0.19 Muffin mix: plain, without fruit or topping 4 102.75 1.89 1.84 5.29 0.12 Cookies: gingersnaps 5 93.00 9.75 10.48 27.30 0.65 Cookies: oatmeal (high fiber) 5 55.60 5.94 10.69 16.63 0.61 Flour: unfortified, white 5 32.80 9.15 27.89 25.62 1.47 Cookies: animal crackers 5 22.80 6.53 28.66 18.28 1.43 a RTE = Ready-to-eat. DEVRIES ET AL.:JOURNAL OF AOAC INTERNATIONAL VOL. 88, NO. 1, 2005 15

P. Lawrence, Laboratory of the Government Chemist, (7) Toepfer, E.W., Zook, E.G., Orr, M.L., & Richardson, L.R. Teddington, Middlesex, UK (1951) U.S. Department of Agriculture Handbook No. 29, J. Martin, U.S. Food and Drug Administration, Atlanta, GA USDA, Washington, DC R. Peace, Health Canada, Ottawa, Ontario, Canada (8) Rabinowitz, J.C. (1960) in The Enzymes,Vol.2,2ndEd.,P.D. L. Rosner, TestAmerica, Inc., Chicago, IL Boyer, H. Lardy, & K. Myrback (Eds), Academic Press, New H.S. Strandler, National Food Administration, Uppsala, York, NY, p. 185 Sweden (9) Flynn, L.M. (1965) J. Assoc. Off. Anal. Chem. 48, 1230–1235 J. Szpylka, General Mills, Mineapolis, MN (10) Ford, J.E., Salter, D.N., & Scott, K.J. (1969) J. Dairy Res. 36, H. van den Berg, TNO Nutrition and Food Research 435–446 Institute, Zeist, The Netherlands (11) Kirsch, A.J., & Chen, T.S. (1984) J. Food Sci. 49, 94–98 C. Wo, ENI Laboratories, Eurofins Scientific, Dayton, NJ (12) Tamura, T. (1990) in Folic Acid Metabolism in Health and C. Wurz, Hoffman-LaRoche, Basel, Switzerland Disease: Contemporary Issues in Clinical Nutrition,M.F. Picciano, E.L.R. Stockard, & J.F. Gregory (Eds), Wiley-Liss, New York, NY, pp 121–137 References (13) Angyal, G. (1996) in U.S. Food and Drug Administration Methods for the Microbiological Analysis of Selected (1) http://www.cc.nih.gov/ccc/supplements/folateref.html Nutrients, AOAC INTERNATIONAL, Gaithersburg, MD, pp (2) Daily, S., Mills, J.L., Molloy, A.M., Conley, M., Lee, Y.J., 1–8, 21–28 Kirke, P.N., Weir, D.G., & Scott, J.M. (1997) Lancet 350, (14) DeSouza, S., & Eitenmiller, R. (1990) J. Micronutr. Anal. 7, 1666–1669 37–57 (3) Crandall, B.F., Corson, V.L., Evans, M.I., Goldberg, J.D., (15) Martin, J.I., Landen, W.I., Jr, Soliman, A-G.M., & Eitenmiller, Knight, G., & Salafsky, I.S. (1998) Am.J.Med.Genet.78, 381 R.R. (1990) J. Assoc. Off. Anal. Chem. 73, 805–808 (4) Institute of Medicine, Food and Nutrition Board (1998) Dietary Reference Intakes: Thiamin, Riboflavin, Niacin, (16) Ryu, K.S., Eitenmiller, R.R., & Pesti, G.M. (1994) J. Sci. Food Agric. 64, 389–392 Vitamin B6, Folate, Vitamin B12, Pantothenic Acid, Biotin, and Choline, National Academy Press, Washington, DC (17) Tamura, T., Mizuno, Y., Johnston, K.E., & Jacob, R.A. (1997) (5) Malinow, M.R., Duell, P.B., Hess, D.L., Anderson, P.H., J. Agric. Food Chem. 45, 135–139 Kruger, W.D., Phillipson, B.E., Gluckman, R.A., & Upson, (18) Pfeiffer, C.M., Rogers, L.M., & Gregory, J.F. (1997) J. Agric. B.M. (1998) N.Engl.J.Med.338, 1009–1015 Food Chem. 45, 407–413 (6) Herbert, V. (1999) in Nutrition in Health and Disease,M. (19) Tamura, T. (1998) J. Nutr. Biochem. 9, 285–293 Shils, J. Olson, M. Shike, & A.C. Ross (Eds), Williams & (20) Rader, J.I., Weaver, C.M., & Angyal, G. (1998) Food Chem. Wilkins, Baltimore, MD 62, 451–465