Yellow maize with high β-carotene is an effective source of vitamin A in healthy Zimbabwean men^12,34

Production of intrinsically labeled yellow maize

Seeds of a high–β-carotene maize line (DE_exp × CI7 − F₁) developed by Torbert Rocheford (Purdue University, West Lafayette, IN) were kindly provided for these studies (16). Seeds were imbibed (soaked in water) and germinated on filter paper for 4 d, until primary roots were ≥3 cm in length. Seedlings were then planted in plastic canisters (5-cm diameter), which had their bottoms removed and replaced with a black plastic mesh (4 mm × 4 mm openings); each seedling's root was inserted through the mesh. Seedling canisters were positioned in lids that were suspended over 20-L tubs of continuously aerated nutrient solution that contained the following micronutrients in mmol/L: KNO₃, 1; KH₂PO₄, 1; Ca(NO₃)₂, 1; MgSO₄, 1; and K₂SiO₄, 0.1; and the following micronutrients in nmol/L: CaCl₂, 25; H₃BO₃, 25; MnSO₄, 2; ZnSO₄, 2; CuSO₄, 0.5; H₂MoO₄, 0.5; and NiSO₄, 0.1; but no deuterium oxide (²H₂O) was added. Iron was added in chelated form as N-hydroxyethyl-ethylenediamine-triacetic acid [Fe(III)HEDTA] at 20 nmol/L. MES buffer (adjusted with KOH) was added at 2 mmol/L to maintain the nutrient solution pH at between 5.4 and 5.8. Plants were maintained on this solution until flowering (≈3.5 mo after planting); the nutrient solution was topped off as needed with a refill solution that contained the same nutrients as above, but without additional MES buffer or K₂SiO₄. A complete change-out of the solution (with the use of the starting solution) was performed at 6-wk intervals. The plants were grown in an environmental growth chamber (Conviron Model PGW36; Winnipeg, Canada) at the US Department of Agriculture (USDA)–Agricultural Research Service Children's Nutrition Research Center in Houston, Texas; they were maintained on a 16-h 20°C/8-h 15°C day/night regimen and 70% relative humidity. Light was provided during the day from a combination of fluorescent and incandescent lamps; the intensity of photosynthetically active radiation was 600 nmol photons m⁻²s⁻¹ at 0.5 m above the tubs.

At the time of flowering, pollen being shed from male tassel flowers at the tops of plants was manually shaken and brushed onto the exposed silks of female flowers on the ear shoots, to ensure a high degree of pollination success. Approximately 10 d after pollination, the tassels were excised from the plants (to shorten the stalks), and the plants were placed in a clear plastic-walled chamber situated within the same environmental growth chamber. At this time the plants were transferred to a new nutrient solution that contained 23 atom% ²H₂O and the following macronutrients in mmol/L: KNO₃, 5; KH₂PO₄, 2; Ca(NO₃)₂, 2; MgSO₄, 1; and K₂SiO₄, 0.1; the following micronutrients in nmol/L: CaCl₂, 25; H₃BO₃, 25; MnSO₄, 2; ZnSO₄, 2; CuSO₄, 0.5; H₂MoO₄, 0.5; and NiSO₄, 0.1; and Fe(III)HEDTA at 20 nmol/L and MES buffer at 2 mmol/L. The 23 atom% ²H₂O allowed us to achieve a target peak enrichment of M + 9 [original mass of β-carotene (M) plus 9 atoms of ²H] for the maize β-carotene (determined empirically in pilot studies). Plants were maintained on this media, with the solution topped off as needed (with the same ²H₂O solution), until the ears had matured (≈45 d later). The ²H₂O nutrient solution was aerated with an air stream that was purged of water vapor. During this labeling period, the temperature within the labeling system was maintained at 25°C (day) and 15°C (night); the relative humidity was maintained at between 35% and 75%. The clear plastic-walled labeling system was used to maintain an elevated ²H₂O concentration in the gas atmosphere surrounding the plants and developing ears, thereby achieving better enrichment. This closed system also required that carbon dioxide was added to the gas atmosphere within the chamber during the daylight hours to support plant growth; the carbon dioxide concentration within the chamber was monitored with an infrared carbon dioxide gas analyzer (model 225-MK3; Analytic Development Co, Hertfordshire, United Kingdom) and was maintained at 400 ± 50 ppm. At maturity, the ears were collected and dried at room temperature (in low light) for 10 d. Seeds were then removed from the ears and stored in sealed plastic bags at −20°C until shipment to Boston for further processing. The seeds used in this study were collected from 2 separate harvests, in June and October 2006.

Production of intrinsically labeled yellow maize dose

The first batch of 612 g of deuterium-labeled yellow maize was received in June 2006 from the USDA–Agricultural Research Service Children's Nutrition Research Center, Houston, TX, soon after harvest. The second batch of 262 g of deuterium-labeled yellow maize was received from the same laboratory in October 2006. On receipt from Houston, the yellow maize was analyzed by HPLC and liquid chromatography (LC)–mass spectronomy (MS)in the Carotenoids and Health Laboratory, Jean Mayer USDA Human Nutrition Research Center on Aging at Tufts University. This was done to determine whether the levels of provitamin A carotenoids and deuterium enrichment were sufficient for a human study (Table 1). The yellow maize kernels were vacuum-packed and stored at −80°C until further analysis. In January 2008, the yellow maize kernels from both the June 2006 and the October 2006 harvests were pooled, homogenized, and ground into flour with a grinder (Braun KSM 4B; Braun Inc, Woburn, MA). On 16 January 2008, the yellow maize thick porridge (sadza, the traditional staple food in Zimbabwe) was prepared by cooks and dietitians in the kitchen of the Metabolic Research Unit at the Jean Mayer USDA Human Nutrition Research Center on Aging at Tufts University according to Zimbabwean recipes (17). The sadza was prepared as follows: 4 cups of water were boiled in a pot; one-half cup of yellow maize flour was mixed with one cup of cold water in a separate bowl, and the mixture was added to the boiling water in a pot and stirred for ≈2 min with a wooden spoon; the mixture was left to boil into a paste/gel for ≈10 min, and more yellow maize flour was then added slowly while the mixture was stirred continuously, until a thick gel was formed (sadza). The sadza was left to simmer for 5 min and then allowed to cool. In total, 729 g of yellow maize flour was used to make the sadza. The total weight of the yellow maize sadza was 2401 g. (Traditionally, only yellow maize flour and water are used to make sadza and in this study nothing else was added to the sadza during cooking.) (17). The sadza was divided into 8 doses, each of which weighed 300 g. After cooking, the all-trans-β-carotene equivalence of the yellow maize sadza was 4.1 μg/g wet weight (Table 1). Each dose contained 1.2 mg all-trans-β-carotene equivalents and was individually vacuum packed and then stored at −80°C until transportation under ice packs for 20 h to Bulawayo, Zimbabwe. The yellow maize doses with ice packs were still frozen on arrival in the city of Bulawayo, Zimbabwe, on 20 January 2008. On arrival at the study site at the Department of Applied Biology and Biochemistry, National University of Science and Technology in the city of Bulawayo, Zimbabwe, the yellow maize doses were stored at −80°C until consumption on 4 February 2008.

Volunteers and study design

To recruit 8 volunteers for the study, 12 healthy men (aged >40 y and <70 y), nonsmoking, and not having taken vitamin A or β-carotene supplements within the past month, were screened as volunteers from the city of Bulawayo. Eight subjects were finally enrolled and completed the study. No specific racial or ethnic background was required. Of the 8 volunteers, one was white, one was Indian (south Asian origin), and 6 were black Zimbabwean men. Volunteers attended a screening session during which they were instructed on how to follow a diet containing low amounts of vitamin A and carotenoid-containing foods. They were provided with lists of foods to select and to avoid while living at home. The subjects were provided with food record books and instructed to list all food and drink consumed. No alcohol was allowed during the study. The following situations excluded potential volunteers from the study: severe and symptomatic cardiac disease or hypertension; history of bleeding disorders; chronic history of gastric, intestinal, liver, pancreatic, or renal disease; any portion of the stomach or the intestine removed (aside from the appendix); history of intestinal obstruction or malabsorption; active smoking (smoking was stopped ≥1 mo before the beginning of the study and during the study); history of chronic alcoholism; a convulsive disorder; or an abnormality in screening blood or urine samples. HIV status was not used in the inclusion or exclusion criteria and was not asked or assessed. Informed written consent was obtained from all volunteers under the guidelines established by the Institutional Review Board of Tufts University, the Tufts Medical Center, and the Medical Research Council of Zimbabwe.

Study design and procedures

For the 2 wk before the first dose, and on days 16–21, days 23–28, and days 30–35, subjects consumed their normal diets, but without vitamin supplements or foods that contained large amounts of β-carotene or vitamin A. A diagrammatic summary of the study design and procedures is shown in Figure 1. Participants received all their meals at the study site from day 1 to day 15, and on day 22, day 29, and day 36.

The subjects were admitted at 0700 on day 1 of the study after an overnight fast. At 0730, a fasting 10 mL blood (time = 0 h) was withdrawn into a no-additive evacuated tube from a forearm vein by registered nurses. At that time the yellow maize sadza dose was taken out of the freezer and left to stand until equilibrated to room temperature. Subjects then spread 20 g butter on their yellow maize dose and consumed it, together with a capsule containing 0.5 g corn oil, under the supervision of coinvestigators, nurses, and research assistants to ensure compliance. In this study the 20 g butter was added to the sadza only to ensure intestinal absorption of maize β-carotene; this is not how sadza is traditionally consumed in Zimbabwe (17). Note that maize seeds also contain endogenous oil. Based on previous analyses (18), we estimate that the sadza made from this hybrid corn would have introduced ≈1.2 g of endogenous fat per 100 g sadza, or 3.6 g fat per our 300-g serving size. Thus, each serving of sadza, with butter and corn oil added, would have contained ≈24.1 g fat. Ten milliliters of blood were drawn at 3, 6, 9, 11, and 13 h after the dose (an intravenous line was inserted for drawing these samples). A low–vitamin A, low–β-carotene breakfast, lunch, and dinner were offered at 0900, 1200, and 1900, respectively. Fasting blood samples were collected at 0730 from day 2 to day 7, and a low–vitamin A, low–β-carotene breakfast, lunch, and dinner were provided at the same times as on day 1.

On day 8 the subjects were admitted at 0730; a fasting 10 mL of blood (time = 0 h) was withdrawn from a forearm vein into a no-additive evacuated tube. The subjects took 1 mg [¹³C₁₀]retinyl acetate (synthesized by the Cambridge Isotope Laboratory, Andover, MA) in a 0.5-g corn oil capsule, with 300 g white maize sadza and 20 g butter. The [¹³C₁₀]retinyl acetate, which converts to [¹³C₁₀]retinol once in circulation, was used as a reference dose to assess the conversion efficiency of the maize β-carotene to retinol. Ten milliliters of blood were withdrawn at 3, 6, 9, 11, and 13 h after the dose (an intravenous line was inserted for drawing these samples). A low–vitamin A, low–β-carotene breakfast, lunch and, dinner were provided at the same times as on day 1. White maize sadza was prepared at the National University of Science and Technology from locally grown noncommercial white maize. White maize is white because it has little or no carotenoid content (19).

Fasting blood samples were collected on days 9, 10, 11, 13, 15, 19, 22, 29, and 36, after the subjects had fasted for 12 h overnight. A low–vitamin A, low–β-carotene breakfast, lunch, and dinner were provided at the same times as on day 1. Throughout the 36-d study period, all the blood samples were kept at room temperature for 0.5 h after drawing and were then centrifuged at 4°C and 800 × g for 15 min. Serum was processed and stored at −80°C until shipment to the United States. After the study was completed, ice packs frozen at −80°C were packed with serum vials in an insulated box and shipped to the United States (an 18-h trip). The serum vials were still frozen on arrival in Boston in March 2008 and were stored at −80°C in the Carotenoids and Health Laboratory at Tufts University until processing, which started in April 2008.

HPLC analysis of serum samples

Under red light, serum samples stored at −80°C were taken out and thawed until equilibration at room temperature before aliquots were prepared for analysis. Three milliliters of chloroform: methanol (2:1, vol:vol) was added to a 100-μL serum sample. The mixture was vortexed and then centrifuged for 10 min at 4°C and at 3000 rpm. The chloroform layer was collected. Hexane (2 mL) was added to the aqueous layer to reextract the fat-soluble compounds. The hexane layer was combined with the chloroform layer and evaporated under N₂ gas on an N-EVAP (Organomation Associates Inc, South Berlin, MA). The residue was dissolved in 100 μL ethanol, which formed a clear solution, and 20 μL was injected into an HPLC system. Concentrations of carotenoids and retinol in a 100-μL aliquot of serum were measured with an HPLC system equipped with a C₃₀ column (SC-150; Bischoff Chromatography, Leonberg, Germany) and a Waters 2996 programmable photodiode array detector (Waters Corp, Milford, MA), and data were retrieved with the wavelength set at 450 nm for carotenoids and at 340 nm for retinoids for quantification (20). The total amounts of retinoids and carotenoids in serum (endogenous and labeled) were determined with the use of external calibration curves obtained from pure standards (from Sigma-Aldrich, St Louis, MO). Serum samples from each subject were analyzed for concentration of serum carotenoids, vitamin A, and vitamin E. Pooled spare-serum from previous human studies were used as quality control samples and were included in each batch of analysis to determine within-batch, interbatch, and day-to-day variations. Internal standards of retinyl acetate and echinenone were used to calculate recoveries of retinoids and carotenoids, respectively. Together with the percentage enrichment (see below), the amount of labeled retinol formed from the labeled β-carotene dose was determined.

Gas chromatography–electron capture negative chemical ionization–MS and LC–atmospheric pressure chemical ionization–MS analysis

To determine the percentage enrichment of labeled retinol, 400 μL of the serum sample was extracted by following the procedure described in the section “HPLC analysis of serum samples” (21). The extract was injected into an HPLC apparatus equipped with a C₁₈ column (Perkin-Elmer Inc, Norwalk, CT) (21). The retinol collected from the HPLC system was dried under nitrogen gas, and the residue was derivatized with ≈10 μL N,O-bis(trimethylsilyl) trifluoroacetamide that contained 10% trimethylchlorosilane (Pierce, Rockford, IL) for 30 min at 70°C to form retinyl trimethylsilyl ether (22). After cooling, 2–3 μL was injected into the gas chromatography (GC)–MS system. The GC-MS instrument was an Agilent 6890 with a 5973 Network Mass Selective Detector, and the GC column was a Zebron ZB-1MS capillary (Phenomenex Inc, Torrance, CA). Helium and methane were used as the carrier gas and reaction gas, respectively. The GC oven temperature was increased from 50°C to 220°C at a speed of 15°C/min, to 230°C at 5°C/min, and to 310°C at 20°C/min; then it remained at 310°C for 5 min before it was cooled down to 50°C . This method produced a trans retinol peak at 12 min. The mass spectrometer was set at a mass-to-charge ratio (m/z) of 260–280. The linearity of the GC-MS response and the detection limit of the GC–electron capture negative chemical ionization–MS system have been addressed previously (22).

To determine the percentage enrichment of labeled β-carotene in yellow maize and in human serum, an LC–atmospheric pressure chemical ionization–MS analysis of m/z at (M+H⁺)of 537, 538 (¹³C), 539 (¹³C₂), 540 (²H₃) −554 (²H₁₇), 555 (²H₁₇-¹³C), and 556 (²H₁₇-¹³C₂) was conducted (23). The molecular mass of unlabeled β-carotene in positive-mode LC-MS analysis is 537 (M+H⁺), 538 (M+H⁺, ¹³C β-C), and 539 (M+H⁺, ¹³C₂ β-C). In this study the yellow maize β-carotene had the highest enrichment abundance of ²H₉–β-carotene at 546 (M+ H⁺+9), and the deuterium enrichment was randomly distributed to all possible positions of the β-carotene molecule as determined by Nuclear Magnetic Resonance (24). This intrinsic labeling (partial replacement of protons with deuterium with a random distribution in carotenoids) enables an easy differentiation between isotopomers from yellow maize plant food and endogenous, unlabeled carotenoids and their cleavage products. The labeled β-carotene isotopomers had a mass range of from 540 (M+H⁺+3) to 553 (M+H⁺+16). However, to ensure optimal detection of labeled β-carotene and labeled retinol in serum, we selectively monitored only the labeled β-carotene isotopomers 544 (M+ H⁺ +7), 545 (M+ H⁺+8), 546 (M+ H⁺+9), 547 (M+ H⁺+10), and 548 (M+ H⁺+11), along with labeled M_retinol + 4 and M_retinol + 5 (²H₄- and ²H₅-retinol), the latter being the predominant retinol isotopomers cleaved from these labeled β-carotene isotopomers. Because the m/z 544–548 isotopomers of labeled β-carotene represented 60% of the entire yellow maize β-carotene [ie, mass 540 (M+H⁺+3) to mass 553 (M+H⁺+16)], an adjustment of 0.6 was used for the total enrichment of retinol formed from the labeled maize β-carotene (Equation 1). The percentage enrichments measured by GC-MS and the concentration of retinol in serum were used to calculate the concentration of labeled retinol in the circulation. The m/z of M_retinol +4 and M_retinol +5 = m/z 272 and 273 could be clearly detected and integrated to represent the formation of retinol from the labeled maize β-carotene.

Enrichment of labeled retinol from yellow maize β-carotene (25):

GC–electron capture negative chemical ionization–MS was used to measure retinol enrichment from the reference dose (M_retinol +10 m/z 278) and yellow maize (M_retinol +4 and M_retinol +5 = m/z 272 and 273, respectively).

Because the in vivo conversion of β-carotene to vitamin A is a dynamic process that combines the absorption, conversion, and metabolism of β-carotene, an appropriate way to determine the amount of labeled retinol (vitamin A) formed from the labeled yellow maize β-carotene is to use a reference dose with a known amount of vitamin A that is differently labeled. In this study, 1 mg of [¹³C₁₀]retinyl acetate was used as a reference dose and it was given a week after the yellow maize dose. [Here, M_retinol − H₂O equals m/z 268 and M [¹³C₁₀]_retinol equals m/z 268 + 10 = m/z 278 (22).] The total enrichment of labeled retinol was determined by evaluation of the negative ions at M_retinol [m/z 268–270 (¹³C₀-¹³C₂)], M_retinol + 4 and M_retinol +5 [m/z 272–273 (²H₄-²H₅)] and M_retinol +10 [m/z 278–280 (¹³C₀-¹³C₂)]. The retinol m/z values are reduced by the mass of H₂O, because a water molecule equivalent is removed from retinol during ionization in the mass spectrometer (22). The percentage enrichment of labeled retinol derived from [¹³C₁₀]retinyl acetate was calculated by integrating the peak area under the reconstructed mass chromatogram of the negative ions at m/z 278, 279, and 280, divided by the total area response of labeled and unlabeled retinol fragment ions, as shown by Equation 2:

Enrichment of ¹³C retinol from [¹³C₁₀] retinyl acetate (25):

Areas under the curve of labeled retinol [2H] and [13C10]retinol in the serum

Total serum responses (nmol) to the [²H]β-carotene dose and the [¹³C₁₀]retinyl acetate dose were determined by multiplying the total serum volume (estimated at 0.0435 L per kilogram body weight) by the concentration of [²H]β-carotene and [²H]retinol and [¹³C₁₀] retinol in the circulation (nmol/L, determined for each time point of serum sampling by adding all of the enrichment masses). Areas under the curve (AUC) in nmol · d for the serum labeled retinol responses after the ingestion of [²H] β-carotene dose and the [¹³C₁₀]retinyl acetate dose were calculated with the use of the curves of total serum responses (expressed as nmol on the y axis) compared with time (expressed as d on the x axis) via Integral-Curve of Kaleidagraph (Synergy Software, Reading, PA). The conversion of the AUC unit from “nmol · d” to “μg · d” was done with the use of M_retinol = 291 for [²H]retinol and M_retinol = 296 for [¹³C₁₀]retinol. Because of the 7-d delay in the administration of the retinyl acetate dose, the AUCs were calculated for 21 d after each labeled tracer.

Retinol equivalence calculations

The retinol equivalence was calculated by comparison of the AUC of serum [²H]retinol response from labeled yellow maize β-carotene (nmol) with the AUC of 1 mg [¹³C₁₀]serum response, as shown by Equation 3.

[²H] retinol from [²H] yellow maize β-carotene (nmol) (25):

Conversion factor calculations

The conversion factor of β-carotene to retinol by weight (25)

Statistical analysis

Descriptive statistics on age, body mass index (BMI; in kg/m²), serum carotenoids, retinoids, and tocopherols were performed with the use of SPSS, version 15.0.1 (SPSS Inc, Chicago, IL).