Phytochemical Research Using Accelerator Mass Spectrometry

Vegetables and fruits provide an array of microchemicals in the form of vitamins and secondary metabolites (phytochemicals) that may lower the risk of chronic disease. Tracing these phytochemicals at physiologic concentrations has been hindered by a lack of quantitative sensitivity for chemically equivalent tracers that could be used safely in healthy people. Accelerator mass spectrometry is a relatively new technique that provides the necessary sensitivity (in attomoles) and measurement precision (<3%) towards ^sup 14^C-labeled phytochemicals for detailed kinetic studies in humans at dietary levels.

Key words: accelerator mass spectrometry, AMS, carbon-14, kinetics, phytochemical

2004 International Life Sciences Institute

doi: 10.1301/nr.2004.oct.375-388

Introduction

Plant-based diets provide an array of biologically active phytochemicals that are thought to confer specific health benefits, including the prevention and mitigation of chronic disease.1 Recent attention has largely focused on minor chemical constituents such as vitamins and secondary products such as polyphenols, flavonoids, isoflavones, terpenes, and glucosinolates. The prospect of marketing “designer” foods that can claim discrete health benefits has created competition among food and agricultural industries to provide phytochemically dense products to the health-conscious consumer. Strategies that combine conventional breeding and genetic engineering to enhance key phytonutrients in crops are being actively pursued.2-4 However, while many compounds have shown promise in bioassays, these indications have not been convincingly confirmed in human subjects.5 Moreover, although potentially beneficial to human health in small doses, many secondary products are toxic and can cause harm when consumed at excessive levels.6,7 Therefore, efforts to change the functionality of foods with regard to phytochemical content should be guided by sound scientific evidence in human subjects.

Information on human absorption and metabolism is a crucial first step toward defining the relationship between trace phytochemicals and human health, and in this regard, ^sup 14^C-accelerator mass spectrometry (AMS) is expected to play a leading role.8 The focus of this review is centered upon recent developments in the area of accelerator-based techniques, substantiated largely by research conducted at the University of California, Davis, and the Lawrence Livermore National Laboratory (LLNL). AMS and in vivo phytochemical research serves as an introduction to AMS and its developing role in phytochemical research. For further knowledge on the topic, several excellent reviews on AMS in biology and nutrition are available.8- 11

Overview of AMS

In the 1940s, Willard Libby at the Institute for Nuclear Studies developed the method of carbon dating using a highly sensitive Geiger counter. In this form of radiometric dating, the concentration of ^sup 14^C, the longest-lived radioisotope of carbon, records the time since the last carbon exchange with the atmosphere.12 The specific activity (^sup 14^C/C)of the material thus serves as a reverse chronometer, accurate-but not highly so-to about 60,000 years into the past. Upon its development, radiocarbon dating had immediate and widespread consequences for many fields. However, the method of decay counting described by Libby is an inefficient means of quantifying the ^sup 14^C content of a sample because ^sup 14^C decays slowly. To illustrate this, measuring 0.1% of the ^sup 14^C in a sample requires uninterrupted counting for 8.3 years (0.1% 5730 years/ln 2; by contrast, 0.1% of the population of ^sup 32^P is measured in 29.7 minutes). To overcome this limitation, large quantities of sample, anywhere from grams to hundreds of grams, are required. However, more often than not, such quantities of material are either not available or cannot be sacrificed to the measurement. In the late 1970s and throughout the 1980s, a mass- spectrometric method for direct detection of ^sup 14^C and other long-lived isotopes was developed in low-energy nuclear physics laboratories. The technique, AMS, was originally developed for the difficult task of radiocarbon dating. In the 1990s, however, AMS emerged as a useful bio-analytical tool for the quantification of ^sup 14^C and (other long-lived isotopes) in biochemical labeling and tracing.8

AMS uses a type of tandem isotope ratio mass spectrometer that measures the ratios of ^sup 14^C/C to parts per quadrillion or down to as few as 10^sup 5^ atoms of ^sup 14^C. In a typical AMS measurement, samples containing approximately 1 mg of carbon are combusted to CO2, which is reduced to graphite off-line prior to analysis. AMS counts electrostatically accelerated nuclei in a simple particle detector. Molecular isobars are completely disassociated in the charge-changing process, and any atomic isobars are discriminated in the detector. A schematic of the compact bio- AMS instrument at LLNL is shown in Figure 1; the figure legend provides a discussion of its mode of operation.

Figure 1. Schematic of compact bio-AMS system at LLNL. Biologic samples for ^sup 14^C-AMS are typically combusted to CO2 and then reduced to graphite. These samples are then bombarded by 3 to 10 keV Cs ions in a cesium-sputter ion source. This process physically knocks atoms and molecules out of the sample and contributes an electron to a fraction of the ejected particles, forming negative elemental or molecular ions. The production of negative ions removes the primary isobaric interference for radiocarbon, ^sup 14^N, since nitrogen does not form a stable negative ion. Singly ionized negative ions are accelerated to 40 keV and then filtered by a low- energy mass spectrometer that alternately switches between ions of mass around 14, i.e., ^sup 14^C, ^sup 13^CH, and ^sup 12^CH^sub 2^, and mass 13, for the stable isotope measurements, into the accelerator. The field in the injection magnet is held constant and the isotopes are changed by quickly switching the energy (voltage) of the incoming ion beam. Negative ions are accelerated through 520 kV in the first stage of a tandem accelerator. At the end of this first acceleration stage, these ions pass through an electron stripper, a small volume of approximately 45 mT gas constrained by differential pumping. Here, particles undergo collisions, losing valence electrons. Depending on the accelerator potential, charge states of +1 to +4 are created. These positive ions then accelerate away from the positive potential to ground potential in the second half of the accelerator. The loss of electrons in the collision cells destroys all molecules, leaving only nuclear ions at relatively high energies (1.04 MeV for C). Positive ions exiting the accelerator are analyzed with a magnet and an electrostatic analyzer. An off-axis Faraday cup measures the ^sup 13^C^sup 1+^ current after the analyzing magnet. A solid-state particle detector counts individual ^sup 14^C^sup 1+^ ions. The footprint of the compact bio-AMS system is approximately 50 m^sup 2^.

Among the unique characteristics of AMS is its capability to provide the lowest detection limits of any isotopic measurement technique for ^sup 14^C, as well as for other long-lived radionuclides, including ^sup 3^H, ^sup 10^Be, ^sup 26^Al, ^sup 32^Si, ^sup 36^Cl, ^sup 41^Ca, ^sup 59^Ni, ^sup 99^Tc, ^sup 129^I, and several actinide isotopes (^sup 41^Ca is an isotope of nutritional importance that allows lifelong studies; variations in the ^sup 41^Ca/Ca ratio reflect dietary changes and bone resorption caused by hormonal changes or osteoporosis8,13). Radioactivity is not a requirement for AMS; rather, AMS profits from the extremely low natural abundance of radioisotopes. For example, ^sup 14^C is present at 100 amol/mg of carbon in contemporary materials; the stable isotope of ^sup 13^C is present at levels of approximately 1 mol in the same amount of material. Quantifying on ^sup 14^C, therefore, extends the theoretical dynamic range of the measurement by a factor of 10^sup 10^, providing detection sensitivities of attomoles for ^sup 14^C in milligram-sized samples.

^sup 14^C-AMS was originally developed for, and remains most applied to, the area of carbon dating at levels of ratios of ^sup 14^C/C at or below the contemporary levels of ^sup 14^C. For all intents and purposes, AMS did not establish a foothold in bioanalytical tracing until AMS operations at LLNL began in late 1988.8 In the first study of its type, AMS was used to quantify the amount of meat carcinogen MeIQX covalently bound to mouse liver DNA (DNA adduct) following very low-level exposure to a ^sup 14^C- labeled carcinogen.14-16 The first nutrient studies followed, with the long-term biologic tracing of folic acid17 and β- carotene18 at the University of California, Davis, in collaboration with the LLNL in 1997 and 2000, respectively. An important demonstration of AMS in human breath analysis came from Lund University (Sweden) in 1996.19 The first use of AMS analysis in human mass balance and high-performance liquid chromatography (HPLC) metabolite profiling of a pharmaceutical appeared in 2002.20

Currently, there are approximately 100 AMS facilitiesscattered throughout the world. There are seven facilities in the United States, four of which engage in biomedical tracing to some degree. Numerous laboratories also exist in Europe, Asia, and Australia/New Zealand.21 The instruments are still expensive but have dropped to below $1 million for the lower-voltage compact spectrometers. This cost is less than high-range, research-grade nuclear magnetic resonance (NMR) systems, which are relatively abundant. The measurement costs vary with facility access and required sample preparation, but, in general, measurement costs are comparable to other high-resolution mass spectrometry techniques.

Sample turnaround time depends upon the facility and number of samples submitted. The facilities with active biologic measurement programs generally have faster turnaround because there is staff committed to sample preparation and measurement. At LLNL, a single technician prepares 300 samples per week, and the spectrometer is capable of measuring 300 samples per day.22 On average, turnaround at LLNL is 7 to 10 days, and for high-priority samples can be as fast as 2 days.

Human Kinetics

The approach of phytochemical research is comparable to that of pharmacokinetic studies. Pharmacokinetic data-absorption, distribution, metabolism, and excretion-is central to drug development. These data are attainable from the ability to measure drugs at the concentration at which they are found in the human body during the course of therapy. This same statement holds true for phytochemical research. A bioactive phytochemical, in order to be an effective public health prophylactic, nutraceutical, or functional food component, must be absorbed intact or as some active metabolite by most individuals, and must attain effective concentrations at the receptors or tissues. It merits noting, however, that phytochemical kinetic studies are technologically more challenging than similar drug studies in several ways. First, drugs are most often tested using large, milligram- to gram-sized doses, whereas phytochemicals are generally consumed in microgram to milligram quantities. Second, pre-existing concentrations in tissues and fluids limit the ability to follow the fate of dietary constituents when given as a “dose,” either with food or as an isolate. These challenges can be surmounted in ^sup 14^C-AMS experiments that facilitate discrimination of exogenous intake (tracer) from endogenous pools (tracee).

Low Background Tracing of Natural and Synthetic α- Tocopherols

In the following sections, a presentation of recently obtained data from in vivo tocopherol tracing is presented. The data were selected because they provide a full description of the disposition of the dose in plasma, urine, and feces. Furthermore, the study addresses the interesting topic of stereoisomerism and metabolism.

Natural and synthetic forms of α-tocopherol are available for use as vitamin E supplements and fortified foods. Ultimately, the biopotency of either form must be determined on the basis of clinical and biochemical endpoints.23 However, in lieu of these data, relative availabilities and retention are assessed using tracer studies that can record the biologic fate of the compound. Natural α-tocopherol, RRR-α-tocopherol, is a single stereoisomer. The conventional synthetic form, all-rac-α- tocopherol (synthetic), however, consists of equal amounts of eight stereoisomers, with only 1/8 the quantity represented as the natural RRR form. The long-accepted ratio of biopotency for natural to synthetic α-tocopherol is 1.36 based on resorption gestation assays in the rat, an assay that cannot be replicated in humans. Several studies have recorded the cumulative plasma exposures of natural and synthetic α-tocopherol as a surrogate measure for biopotency using deuterium labels and traditional mass spectrometry.24-26 These data suggested a biopotency ratio closer to 2 than 1.36. A primary methodological concern points to the use of pharmacologie doses (15-150 mg) that can overwhelm the physiologic transport systems.27 AMS sensitivity, however, allows tracing nutrients at normal intake levels (or less) that do not saturate or perturb metabolic pathways. This model is perfectly suited, therefore, for the study of natural and synthetic doses of α- tocopherol, as described in the following study conducted in our laboratories.

^sup 14^C-labeled tocopherols (natural and synthetic) were individually tested in a longitudinal design in a single male participant. The first dose consisted of natural α-tocopherol. Three months later, a similarly sized dose of the synthetic mixture was administered. Both doses contained 100 nanoCuries (nCi) of radioactivity and were <1 g in mass (Figure 2). Frequent blood samples (5 mL) were taken via an indwelling venous catheter over the first 2 days after dosing, followed by less frequent samples for 2 months. Twenty-five-microliter aliquots of plasma were graphitized, and carbon isotope ratios were measured to 3% precision by AMS. Data are expressed as femtomoles of ^sup 14^C per milliliter of plasma; these values can be expressed as labeled α-tocopherol equivalents using the specific activity of the labeled compounds.

The doses were equal in terms of their ^sup 14^C content, and therefore similar metabolism would predict equivalent kinetic patterns. The kinetic description of data for both forms is presented in Figure 2. A 2-hour delay in the appearance of activity is consistent with lymphatic absorption. Both forms displayed a complex kinetic behavior of a retained and bound molecule. Visual inspection of the concentration profiles reveals a muted plasma response for the synthetic relative to the natural α- tocopherol experiment, which is confirmed in Figure 3 by cumulative, integrated area-under-the-concentration (AUC) time-course plots of the total plasma exposure. This plot clearly suggests that the total plasma exposure of the synthetic dose is 40% of the natural dose after 10 days post-dose. Based only on this information, a biopotency of approximately 2 would be supported with tracer-sized doses.

Figure 2. Plasma time courses of natural (RRR) and synthetic (all- rac) vitamin E. Two ^sup 14^C doses were given in a longitudinal design in a single male volunteer: the first dose consisted of the single natural isomer (3.7 kBq, 100 nCi, 0.85 g). Three months later, a similarly sized dose of the synthetic mixture was administered. Both doses were given with a cup of whole milk (10 g fat) to promote absorption. Dose masses were 10,000 to 125,000 times less than similar studies using deuterium-labeled isotopes with gas- chromatography or liquid-chromatography mass spectrometry detection.

The small sample requirements for AMS measurement (minimum requirements of 1 mg or less of carbon or about 25 L of plasma) provided for frequent blood sampling and a subsequent detailed temporal record of the kinetic behavior of the labeled tocopherols. The plasma taken in the presented tocopherol study was obtained via catheterization. Such results might well have been generated using microliter-sized capillary finger sticks (similar to home health care testing kits used for the monitoring of blood glucose levels in diabetics), because the first 24 hours of samples consumed less than 1 mL of plasma. This technique would have the dual advantage of obviating the need for costly clinical support while lessening participant resistance associated with discomfort related to venipuncture or catheterization.

Figure 3. Total plasma exposure to ^sup 14^C following administration of an equivalent amount of natural (RRR) and synthetic (all-rac) α-tocopherol. A value of 1 is assigned to the RRR form. After 250 hours, the relative exposure to the synthetic form was about 40% that of the natural α-tocopherol.

Mass Balance

AMS sensitivity can reinvigorate balance studies at doses that will not saturate the binding capacity of the gastrointestinal system. Because AMS is a combustive process and since specific compound identification is not needed, performing balance measurements is relatively straightforward.

In the described tocopherol study, cumulative urine and feces were collected for 8 days post-dose. The loss of label in the feces was biphasic, with the transition between the phases occurring between 2 and 2.5 days. Early stool losses (<3 days) are attributed to unabsorbed compound, whereas later losses are attributed to gastrointestinal tract epithelial cell loss and inefficient recovery of tocopherols in the entero-hepatic pool with bile loss. Losses were virtually equivalent for both compounds and accounted for 25% of the dose in this first phase, suggesting that at sub-physiologic doses the gut does not discriminate between the forms25 and that the dose is 75% bioavailable. This approximation, however, does not take into account potential losses in the gut and other contribution of the rapidly re-excreted tocopherol, and therefore may underestimate the true absorption.

Figure 4. Recovery of ^sup 14^C in stool and urine following the two oral doses of labeled tocopherol. Intestinal absorption was equivalent for the two forms. Post-absorption, the synthetic all- rac form underwent more rapid elimination via the urinary route. After 8 days, approximately 50% of the all-rac and 65% of the RRR was still present in the body.

Kinetic analysis of the ^sup 14^C contents in the urine revealed that the synthetic excreted at approximately 2.5 times the rate relative to the natural α-tocopherol (Figure 4). This result is consistent with the observation that the synthetic stereoisomers are less retained than the natural form, perhaps due to differential affinities for specific binding proteins. Cumulative urine and fecal ^sup 14^C losses indicated that after 8 days, approximately 50% of the synthetic and 65% of the natural α-tocopher\ol were retained. By this estimate, a biopotency closer to 1.3 would be obtained. The discrepant data observed between the plasma and balance estimates of biopotency highlight the importance of complete kinetic information.

Urinary Metabolite Profiling

For establishing metabolite profiles, small (microliter) aliquots of minimally processed plasma or unprocessed urine are separated by liquid chromatographic systems prior to analysis. These radiochromatograms can serve as detectors for assessing the range of biotransformations and identifying target analytes. The low background of ^sup 14^C in the environment provides an unambiguous baseline for clearly identifying ^sup 14^C-enriched fractions collected from chromatographic systems.28,29 Furthermore, since quantitation is based upon an intrinsic property of the molecule, i.e., the label, values are more quantitative than from methods whose accuracy is tied to molecular structure or matrix.

The first described urinary metabolites were identified by Simon et al.30 as α-tocopheronic acid and its lactone in 1956. The eponymous “Simon” metabolites had an open chroman structure, consistent with α-tocopherol that had reacted as an antioxidant. In 1995, however, Schultz et al.31 found the Simon metabolites to be mainly products of oxidative modification of a precursor, 2,5,7,8-tetramethyl-2-(beta-carboxyethyl)-6- hydroxychroman (α-CEHC), during the analytical workup (Figure 5). Subsequently, interest has focused on the dynamics of α- CEHE excretion and a growing list of other minor metabolites. Many questions, however, remain to be answered. For example, evidence from limited studies have reported that only a small fraction-as low as 5% of administered α-tocopherol-is recovered as α-CEHC in the urine.31,33 Therefore, methods are needed that can account for the full complement of biologic metabolites.

Research at our laboratories addressed aspects of this issue by separating urine by reversed-phase HPLC. Samples were treated by two methods. One procedure followed that of Lodge et al.,33 and included deconjugation by glucuronidase treatment and partitioning of the metabolites into ether. The second entailed glucuronidase treatment with a direct injection of the unpartitioned urine. Both urine samples were “cold-spiked” with the α-CEHC and the Simon metabolite at concentrations that would be easily viewed in the column effluent by UV/Vis absorbance. Compound confirmation is derived from co-chromotography with these authentic standards. In the present example, unlabeled α-CEHC and the Simon metabolites were added to the urine prior to processing to track the quality of the chromatographic separation.

Figure 5. Pathways of tocopherol metabolism. Current understanding states that participation as an antioxidant leads to the opening of the chroman ring and lactone products (left side), whereas an intact chroman ring (right side) is hypothesized to represent excretion of excess tocopherol. The Simon metabolite, α-tocopheronolactone, appears to be produced by artifactual oxidation of α-CEHC (2,5,7,8-tetramethyl-2-(beta-carboxyethyl)- 6-hydroxychroman) during the extraction procedure.

The metabolite profiles (radiochromatograms) from the two urine treatments are shown in Figure 6. For the ether extract, peak concentrations of ^sup 14^C coincided with the α-CEHC standard. In contrast, little or no ^sup 14^C appeared at the characteristic elution time for the Simon metabolites. This latter result affirmed the integrity of at least some of the metabolites, as α-CEHC was preserved in the extraction procedure. The activity of ^sup 14^C was not exclusive to the α-CEHC metabolites. From visual examination of “peak tops” in the radiochromatograms, a minimum of six chemically distinct metabolites that carry the ^sup 14^C tag in both the natural and synthetic urine samples were found. The natural and synthetic profiles present some observable differences. The peak after 8 minutes in the synthetic profile was absent in the natural profile, indicating differential pathways of catabolism for the natural and synthetic forms, although fuller kinetic profiles will need to be generated to confirm this observation.

Figure 6. ^sup 14^C contents in HPLC eluents from human urine after consumption of RRR (top panel) and all-rac (bottom panel) α-tocopherol. Baseline is 9.0 amol 0.55 using a blank injection run. The limit of quantification is 1.6 amol per eluent fraction. α-CEHC is the single major metabolite in both tests, although at least nine metabolites are indicated in the combined radiochromatograms.

Processing can alter the recovery and character of the metabolite profiles, including, in the case of reactive metabolites, the formation of artifacts that can lead to spurious theories of metabolism similar to what occurred with the Simon metabolite. Therefore, we analyzed the same urine sample without ether partitioning. Thirty microliters of raw urine treated with glucuronidase was directly injected onto the HPLC system. The metabolite, α-CEHC, was again the main metabolite. The bulk of the radioactivity, however, appeared with the larger, unretained compounds at the early region of the chromatogram. These discrepancies highlight the importance of minimal processing for generating accurate metabolite profiles. Indeed, the ideal approach to sample preparation is to exclude the step altogether or “dilute and shoot.”34 Metabolite profiling by direct injection of unprocessed urine has been demonstrated in other applications.20,35

The limit of quantification for HPLC-AMS is determined by the average of the baseline fractions plus two standard deviations, and is typically in the vicinity of 2 to 20 amol ^sup 14^C.28 Even lower levels of detection can be achieved through application of cleanup methods for removing other sources of biogenic carbon (lipid, proteins, etc.) and higher-specific activity compounds (but not higher radioactive doses).

β-Carotene: Nutrient-Nutrient Interactions

Although marginally tested, it is popularly suspected that vitamin A status inversely relates to the utilization of β- carotene as a vitamin A source in humans.36 To test this hypothesis, populations at risk of marginal vitamin A status, typically children, are fed β-carotene.37 Individuals with low vitamin A status will supposedly obtain greater vitamin A benefit from the administered β-carotene than those with adequate vitamin A status.38 These studies, however, require the participation of experimental subjects of marginal nutrient status, and such subjects are difficult to locate in industrialized countries. Moreover, to the extent that phytonutrition is often focused on optimal health in nutritionally replete populations (rather than correcting deficiencies), it is of interest to have methods that are sensitive to subtle physiologic changes in replete populations.

A recent study highlights the unique opportunities that AMS brings to the study of replete populations, in which other indices might not respond to subtle changes in dietary intake.39 The effect of vitamin A supplements on the absorption and metabolic behavior of a physiologic dose of ^sup 14^C-β-carotene was investigated in a repeat test format. Specific details of the dose timeline are shown in Figure 7. Two normal females ingested 1 nmol ^sup 14^C- β-carotene (100 nCi) in olive oil. Concentrations of labeled β-carotene, retinyl esters, and retinol were determined in plasma for 46 days. Complete fecal and urine collections were also obtained for 14 and 30 days, respectively. After the first test, both subjects consumed daily supplements of 10,000 IU retinyl palmitate for 21 days to raise their vitamin A nutritional status. Both subjects then took a second and similar dose of ^sup 14^C- β-carotene.

In a vitamin A-replete individual, the effect of supplemental dose in moderate excess of normal dietary levels (2 RDA) might not be expected to substantially alter the disposition of β- carotene. The results of this study suggest otherwise. Figure 8 shows that the dose disposition 72 hours post-dose for both the test and re-test periods was exhibited in one subject (this pattern was displayed in both participants). Absorption, as assessed by the cumulative recovery of the ^sup 14^C label in the stools, was raised by about 50% in both subjects. Conversely, the recovery of ^sup 14^C in the urine was reduced by tenfold in both subjects. As a result, 72 hours after dosing, about 80% of the consumed vitamin A was retained in the body of the supplemented individual, whereas only about 35% was retained in the untreated individual. These data suggest that raising the vitamin A status in already replete subjects can markedly raise absorption and subsequent retention of a physiologic dose of β-carotene. Such data have important implications for recommendations regarding the consumption of β- carotene, either in food-based or supplemental forms. Minor changes in the consumption of one nutrient can have additive, or even synergistic, effects on the utilization of a second nutrient, even in nutritionally replete populations that by all definitions are not lacking in either nutrient.

Figure 7. Experimental design and time line. The experiment was designed to incorporate a test and re-test period. In the test period, subjects began complete fecal and urine collection 24 hours in advance of the dose, and continued complete 24-hour collections until day 16 and day 30, respectively. The subjects were then given a 1-nmol dose of [^sup 14^C]β-carotene in an emulsified drink. Cumulative urine was collected for 30 days and cumulative stool for 17 days. This was followed by a 7-week wash-out period. After that time, the administration of a second [^sup 14^C]β-carotene dose marked the beginning of the retestperiod. Three weeks prior to the start of the re-test, subjects began consuming 10,000 IU (3000 mg retinol equivalents [RE]) of vitamin A supplement daily, and continued at that level until 2 weeks after the re-test dose administration. The supplement was then continued at 5000 IU (1500 RE) until the completion of the testing.

Additionally, these data illustrate the value of non-invasive collections, i.e., urine and stool, for the assessment of phytonutrient metabolism. Because the dose input is known, the total body burden of the labeled compound is obtained without the collection of blood specimens. Such experimental models represent an attractive method for assessing the biologic variance in the absorption and retention characteristics of compounds without specific analyte identification methods in the blood. This non- invasive feature, causing minimal discomfort in participants, would greatly facilitate population surveys into the utilization of selected phytonutrients.

Figure 8. Effects of vitamin A supplementation upon the disposition of a tracer dose of ^sup 14^C-β-carotene. The chart illustrates the dose disposition in excreta and subsequent body burden prior to (left) and following (right) supplementation with vitamin A. Supplemental vitamin A resulted in increased absorption (stool) and retention (urine and body burden) of a tracer dose of β-carotene in vitamin A-replete participants.

Cell Population Kinetics Using ^sup 14^C-Folate

Retained nutrients have been especially difficult to study in humans prior to the use of AMS. Figure 9 shows the ^sup 14^C content of human red blood cells over a 7-month period after a single 35-g, 100-nCi dose of ^sup 14^C-folate, a nutrient required in the maturation of blood cells.40 Erythrocyte ^sup 14^C levels displayed a broad maximum from 8 to 103 days post-dose, as the ^sup 14^C- labeled erythrocytes aged and were eliminated. The erythrocyte lifetime was approximately 125 days, similar to the accepted lifetime of 120 days derived previously by older methods. Of the erythrocytes that accumulated ^sup 14^C-folate during maturation, it can be calculated that a single cell contained approximately 130 ^sup 14^C atoms (from ^sup 14^C-folate) above background. In contrast, previous research into cell life cycles used constant infusion of deuterated compound for 48 hours to obtain less than 2 weeks of data.41

Figure 9. ^sup 14^C content of labeled folate in erythrocytes, which require 4.5 days to mature in the bone marrow after accumulating ^sup 14^C-folate before reaching circulation. Their elimination occurs after 130 days in circulation.

The above data are taken from a larger study that described the disposition of an oral dose of ^sup 14^C-folate into plasma, erythrocytes, urine, and stool.17 Using empirical modeling, the data indicated that the mean sojourn time for folate is in the range of 93 to 120 days, and that it took more than 350 days for the absorbed portion of small bolus dose of ^sup 14^C-folic acid to be eliminated completely from the body. The dosages described in this study were sufficiently small that the total radiation exposure was only a fraction of the natural annual background radiation exposure received by Americans, and the generated laboratory waste may legally be classified non-radioactive in many cases.42

Radiation Risk

When discussing the role of radiation in AMS, it is important to first discuss the role of radiation in the natural world. Our world is radioactive, and when assessing the radiation risk in experimental subjects, one must incorporate voluntary and involuntary exposures presented by our environment, diet, and lifestyle choices. Natural radiation from cosmic rays and from naturally occurring radioisotopes accounts for the vast majority of the total average annual effective dose. Specifically, 81% is from natural sources of radiation, and of that, most is from radon. Every food has radionuclides that lead to the accumulation of radioisotopes within the body. The most common radionuclides in food are: potassium 40 (^sup 40^K), radium (^sup 226^Ra), and uranium (^sup 238^U), their associated progeny, and of course radiocarbon (^sup 14^C). Because of these radionuclides (and others), an average human body experiences about 500,000 radioactive disintegrations per minute. The integrated average exposure is about 3600 Sv (microsieverts; a sievert is the deposited energy equivalent to one joule per kilogram).

It is important to consider the amount of ^sup 14^C that naturally occurs in the body. The human body is about 23% carbon. The natural background levels of ^sup 14^C due to formation in the upper atmosphere (cosmogenic production), as well as contributions from anthropomorphic sources (the burning of fossil fuels and atmospheric testing of nuclear weapons) is 6.11 pCi/g carbon (0.23 Bq). Accordingly, approximately 100 nCi (37 kBq) is present at all times in a 70-kg person. An average adult might consume 200 g of carbon in a 24-hour period, which would contain 2.1 nCi of ^sup 14^C. In contrast, a “large” administered radio-dose in AMS experiments is 100 nCi. In the case of vitamins, this quantity was determined to obtain long-term kinetic information associated with the longer turnover times inherent to fat-soluble vitamins (β- carotene T^sub 1/2^ = ~40 days) and even retained water-soluble forms (tblate T^sub 1/2^ = -100 days). Experiments in β- carotene and folic acid17 were conducted for greater than 200 days.18 Many studies will be primarily interested in those events occurring within the period of absorption and distribution of a dose, and therefore studies on a limited duration of several days can be conducted. Where long-term tracing is not necessary, doses of 10 nCi or less suffice. In such cases, the amount of administered radiocarbon would be equal to that consumed in several meals. By this analogy, it is hard to define a quantifiable risk with AMS studies.

Labeling Strategies

AMS is poised to assume a leading role in the study of plant constituents. For effective expansion of AMS into phytochemical investigations, suitably labeled substrates must be available. Many natural products found in foods, however, are difficult to synthesize due to complex stereochemistries and a multiplicity of polymeric forms. However, there are several options to overcome the aforementioned predicament. Labels can be introduced into intact plants, as well as excised tissues, cultured plant cells, algae, or isolated enzymatic homogenates. For intact plants, if the label is efficiently incorporated into the target, the food can conceivably be ingested intact to examine the effects of the plant matrix on the nutrient digestibility (intrinsic bioavailability). Alternatively, plants can be grown in ^sup 14^CO^sub 2^-enriched environments using atmospherically isolated plant growth chambers. Photosynthetic labeling in the presence of ^sup 14^CO^sub 2^ randomly labels the entire phytochemical constitution of the plant, thus creating a radiopharmacy of labeled substrates. When following the appropriate fractionation, this yields a spectrum of compounds that can be used individually, or possibly in synergistic combinations, for in vivo and in vitro studies of human metabolism and catabolism.

The value of isotopes (stable or radioactive) as molecular tracers depends on the detection methods: isotope ratio mass spectometry, NMR, decay production counting, or AMS. Furthermore, the final detection method will determine the level of sophistication used in the labeling protocol. It is established that methods developed around the detection of stable isotopes generally require much higher levels of isotope incorporation into the molecule-to facilitate detection-than corresponding radiometric methods (particularly AMS). One demonstration of biolabeling that contrasts stable and ^sup 14^C protocols is that of β-carotene in the photosynthetic alga Dunaliella salina. This organism accumulates large amounts of β-carotene in response to the stress of nutrient depletion and increased light intensity. Because it is photosynthetic, it offers the opportunity of incorporating labeled CO2. Wilson et al.43 described in detail the construction of a 4.5-L photosynthetic bioreactor for the production of ^sup 13^C- labeled carotenes using a ^sup 13^CO^sub 2^ precursor. The design of the reactor was optimized for the conservation of the label using a gas-recycling protocol during the feeding of the label into the system and a high level of incorporation of label into the β- carotene target. Between 0.7 and 1.4 g/day of NaH^sup 13^CO^sub 3^ were fed for approximately 25 days in several experiments. Mass spectral data demonstrated high levels of incorporation of ^sup 13^C into the β-carotene. In one experiment, the average molecular formula of the β-carotene showed that 30 of the possible 40 carbons in the molecule possessed the ^sup 13^C label. These high levels of label were deemed necessary for the efficient detection of β-carotene in human tracer studies using traditional mass spectrometry. Parker44 demonstrated the value of highly labeled β-carotene in human kinetic studies using gas-chromatography/ infrared mass-spectometry detection.

In contrast to the examples mentioned above, AMS sensitivity obviates the need for high levels of isotopic enrichment or sophisticated labeled chambers. In one experiment conducted in our laboratory (unpublished results, University of California, Davis, 2002), 3 mCi of NaH ^sup 14^CO^sub 3^ was added to a 50-mL culture of D. salina grown in a 50-mL screw-cap bottle. A light source was provided, and the contents of the closed vessel were allowed to grow for 1 week. The final specific activity of the β-carotene measured 30 Ci/mol, which corresponded to a single ^sup 14^C atom per two molecules of β-carotene. Nonetheless, even at these low leve\ls of incorporation, a nanocurie-sized dose of even lower specific activity was traced in human plasma for over 200 days using AMS detection.18

Photosynthetic labeling can also be applied to intact plants using atmospherically sealed plant chambers. An example of a chamber appropriate for radiolabeling with ^sup 14^CO^sub 2^ is illustrated in Figure 10. This chamber ensures controlled cycles of irradiance, temperature, and relative humidity, while quantitatively containing the administered radiolabel. The system is primarily comprised of a stainless steel chamber, a refrigerated water chiller, a water reservoir with pump, a 1000-W metal halide lamp suspended from an adjustable support arm, a control panel, a pressure relief system, and an on-time totalizer for administering CO2.

The value of such a chamber is illustrated in a biolabeling experiment of common spinach (Spinacia oleracea), also conducted in our laboratories.18,45 Mature plants (30 days of age) were placed into the chamber and allowed to acclimate for 2 days. Exposure was initiated by adding 10 mCi per day as a solution of NaH^sup 14^CO^sub 3^ to excess acid, which was repeated for an additional 4 days. Following the final exposure, the plants were maintained for 72 hours in the chamber prior to harvesting to remove residual ^sup 14^CO^sub 2^. The aerial parts of the plant were harvested and were extracted for β-carotene and lutein: specific activities were 1.45 and 0.35 Ci/mol respectively, corresponding to 0.023 and 0.0056 ^sup 14^C atoms per molecule.

Limitations of AMS: Intrinsic Labeling

There is a growing need to assess the availability of nutrients from foods (as opposed to chemically isolated concentrates). Photosynthetic radiolabeling is not practical for this purpose, since the vast majority of the radiolabel resides in non-target plant components that, upon consumption, would expose the participant to potentially large radiative doses. Similar concerns do not pertain to ^sup 13^C, but the costs associated with it dampen the enthusiasm for this approach. Nonetheless, as the value of an intrinsically labeled plant is recognized, this approach will meet important needs. Recent demonstrations of ^sup 13^C photosynthetic labeling were presented for kale46 (Brassica oleracea var. acephala), spinach, and couard greens.47 In the kale study, lutein, phylloquinone, β-carotene, and its biologic metabolite retinol were traced using liquid chromatography-atmospheric pressure chemical ionization mass spectrometry, while lutein isotopomers were the target of the spinach experiment. Either experiment validates the use of stable isotopes for assessment of instrinsic bioavailability. Moreover, the ability to analyze multiple components is clearly a strength of the traditional mass spectrometry approaches.

Figure 10. Diagram of a photosynthetic labeling chamber.

A second important example of the intrinsic labeling concept was demonstrated in the biologic tracing of deuterium-labeled vitamin K (phylloquinone).48 In this example, broccoli grown hydroponically using 31 atom % deuterium oxide was fed to a single participant. The plasma concentration of the labeled vitamin K in the plasma could be traced for approximately 12 hours using gas chromatography/mass spectrometry detection operated in the negative chemical ionization mode.

The concept of intrinsic labeling has been applied to the radiolabeling of folates in the pea plant.49,50 In these experiments, pea seeds were allowed to imbibe ^sup 14^C-para- aminobenzoic acid (pABA) prior to germination and vegetative growth (Figure 11). Significant portions of the administered label appeared in the plant folate pools. While human testing was not the aim of this investigation, it does validate the general concept. The concept was also applied in our laboratories for the specific purpose of assessing the availability of folates when consumed by an intact plant. In this study, ^sup 14^C-pABA was administered to a mature kale plant using the cottonwick method.51 After uptake of the label, the plant was then allowed to grow for 1 week to promote translocation of the dose to the actively growing leaf tips, the area of most active respiration. This experiment showed minimal incorporation. However, delivery of the same label via the root system in young spinach provided substantial (>60%) assimilation of the label into folate pools.

Upon surveying various labeling strategies, it is clear that mass- spectometry techniques will be enlisted to address biolabeling. While AMS excels in terms of sensitivity and quantitation, it also adds complexity, since the labeling scheme should be specific to the desired phytochemical when intrinsic bioavailability is sought. This interesting area of isotopes, mass spectrometry, and intrinsic bioavailability is expected to grow in the coming years through strategic, interdisciplinary efforts.

Figure 11. Stylized representation of folic acid. The gray area represents moeity derived from p-aminobenzoic acid.

Experimental Model Design

Questions continue to be raised concerning the relationship of phytonutrients and diet (fruits and vegetables, meat, fiber, fat) to health and disease. Moreover, they have generated many large clinical trials specifically aimed at testing the effects of select nutrients in reducing the risk of chronic diseases. Most of these trials have produced inconclusive results. The most controversial trials involving the lack of effect of β-carotene in decreasing cardiovascular disease and cancer risk were: the Carotene and Retinol Efficacy Trial (CARET), the Alpha Tocopherol Beta Carotene (ATBC) Trial, and the Physician’s Health Study (PHS). Many explanations have been put forth, including differences between the natural forms of β-carotene and synthetic β-carotene supplements (synthetic all-trans versus “natural” cis-trans isomeric mixtures)52; the lack of experimental data in human systems on pro- oxidant activity of β-carotene under certain conditions53; and the fact that these studies, being at phase II or III, did not go through rigorous development to identify safe dosage, efficacy, distribution, and metabolism of synthetic β-carotene, α- tocopherol, retinols, or the combination of these compounds.54

Despite unfavorable results, the above β-carotene trials prompted reevaluation of nutrition research, as well as the need for methodological, pre-clinical experiments analogous to the testing of a new drug treatment. Epidemiologic data, although helpful in establishing associations between nutrients and disease-risk reduction, cannot replace basic scientific studies to understand absorption, metabolism, catabolism, pharmacokinetics, toxicity, and mechanism of activities of the investigated compounds. If phytochemicals are going to be used to improve health or as treatment for diseases, steps similar to those used in the development of new drugs should be applied. Gescher55 suggested “a structured design, incorporating parallel preclinical studies of the food source and the isolated agent in terms of efficacy, toxicity, biological mechanisms and pharmacokinetics.” Farnsworth, a pioneer in the study of natural drug development, emphasized the two important steps preceding clinical trials and after the active compounds have been isolated and characterized chemically: (1) biologic and chemical standardization and (2) in vitro studies of metabolism, absorption, and toxicity. The common thread in these proposed models is the pre-clinical path, which was lacking in earlier investigations. In this path, complementary roles for both AMS and traditional mass spectrometry can be envisioned. AMS has the sensitivity and specificity towards a ^sup 14^C label to establish metabolite profiles and toxicology data from in vitro and animal models at doses that mimic concentrations relevant to dietary intake levels.

Conclusion

As public interest in phytonutrition continues to increase, there will be an augmented demand for extensive phytochemical research. The fact that foods are inherently phytochemically complex dictates the need to apply scientific techniques to them. However, while these techniques detect synergistic interaction among the many active principals and adjuvant substances in the plant, at the same time they modify the activities of these components. As illustrated by the experiments discussed in this manuscript, the advantages of AMS are unique and extensive. These advantages are best summarized by Dr. John Vogel, an originator of biologic AMS experimentation: “AMS brings (at least) three advantages to biochemical tracing: high sensitivity for finding low probability events or for use of physiologic-sized doses; small sample sizes for painless biopsies or highly specific biochemical separations; and reduction of overall radioisotope exposures, inventories, and waste streams.”

AMS opens the door to increased phytochemical tracing in humans to obtain biochemical data concerning human health at dietary levels of exposure, obviating the need for uncertain extrapolations from animal models, which express marginal relevance to human metabolism. The unparalleled capabilities and benefits of AMS will undoubtedly establish this technique as an invaluable analytical tool in phytochemical research.

Acknowledgements

Andrew Clifford, Yumei Lin, Jennifer Follett, Colleen Carkeet, Charlene Ho, and John Vogel all provided original data of published work or data in preparation for the work presented here. Much gratitude is also extended to Dr. William Cohn of DSM (formerly Roche division Vitamins and Fine Chemicals) and Dr. Robert Parker (Cornell University). Part of this work was performed under the auspices of the US Department of Energy by the University of California, Lawrence Livermore National Laboratory under contract no. W-7405-Eng-48, NIH NCRR P41 RR 13461, CSREES#99-35200-7584, and NIDDK DK48307.\1. Becher GR. Phytonutrients’ role in metabolism: effects on resistance to degenerative processes. Nutr Rev. 1999;57:S3-S6.

2. Grusak MA. Phytochemicals in plants: genomics-assisted plant improvement for nutritional and health benefits. Curr Opin Biotechnol. 2002;13:508-511.

3. Potrykus I. Golden rice and beyond. Plant Physiol. 2001;125:1157-1161.

4. Mullineaux PM, Creissen GP. Opportunities for the genetic manipulation of antioxidants in plant foods. Biochem Soc Trans. 1996;24:829-835.

5. Farnham MW, Simon PW, Stommel JR. Improved phytonutrient content through plant genetic improvement. Nutr Rev. 1999;57:S19- S26.

6. Ames BN, Profet M, Gold LS. Dietary pesticides (99.99% all natural). Proc Natl Acad Sci U S A. 1990;87:7777-7781.

7. Halsted CH. Dietary supplements and functional foods: 2 sides of a coin? Am J Clin Nutr. 2003;77(suppl 4):1001S-1007S.

8. Vogel JS, Turteltaub KW, Finkel R, Nelson DE. Accelerator mass spectrometry. Anal Chem. 1995;67:353A-359A.

9. Vogel JS, Turteltaub KW. Accelerator mass spectrometry as a bioanalytical tool for nutritional research. Adv Exp Med Biol. 1998;445:397-410.

10. Vogel JS. Low background isotope tracing in biochemistry. Nucl Phys News. 2000;March:1-8.

11. Jackson GS, Weaver C, Elmore D. Use of accelerator mass spectrometry for studies in nutrition. Nutrition Research Reviews. 2001;14:317-334.

12. Libby WF. Radiocarbon dating. Science. 1961;133:621-629.

13. Weaver CM, Liebman M. Biomarkers of bone health appropriate for evaluating functional foods designed to reduce risk of osteoporosis. Br J Nutr. 2002;88(suppl 2):S225-S232.

14. Turteltaub KW, Felton JS, Gledhill BL, et al. Accelerator mass spectrometry in biomedical dosimetry: relationship between low- level exposure and covalent binding of heterocyclic amine carcinogens to DNA. Proc Natl Acad Sci USA. 1990;87:5288-5292.

15. Turteltaub KW, Vogel JS, Frantz C, Buonarati MH, Felton JS. Low-level biological dosimetry of heterocyclic amine carcinogens isolated from cooked food. Environ Health Perspect. 1993;99:183- 186.

16. Turteltaub KW, Vogel JS, Frantz CE , Frantz CE, Fultz E. Studies on DNA adduction with heterocyclic amines by accelerator mass spectrometry: a new technique for tracing isotope-labelled DNA adduction. IARC Sci Publ. 1993;293-301.

17. Clifford AJ, Arjomand A, Dueker SR , Schneider PD, Buchholz BA, Vogel JS. The dynamics of folic acid metabolism in an adult given a small tracer dose of 14C-folic acid. Adv Exp Med Biol. 1998;445:239-251.

18. Dueker SR, Lin Y, Buchholz BA, et al. Long-term kinetic study of beta-carotene, using accelerator mass spectrometry in an adult volunteer. J Lipid Res. 2000;41:1790-1800.

19. Stenstrom K, Leide-Svegborn S, Erlandsson B, et al. Application of accelerator mass spectrometry (AMS) for high- sensitivity measurements of 14CO2 in long-term studies of fat metabolism. Appl Radiat Isot. 1996;47:417-422.

20. Garner RC, Goris I, Laenen AA, et al. Evaluation of accelerator mass spectrometry in a human mass balance and pharmacokinetic study-experience with 14C-labeled (R)-6-[amino(4- chlorophenyl)(1-methyl-1H-imidazol-5-yl)methyl]-4-(3-chlorophenyl)- 1-methyl-2(1H)-quinolinone (R115777), a farnesyl transferase inhibitor. Drug Metab Dispos. 2002;30:823-830.

21. Vogel JS, Ognibene T, and Buchholz BA. MS of “semi-” stable isotopes for absolute attomole quantitation. American Biotechnology Laboratory. 2003; 21:20-23. 22. Ognibene TJ, Bench G, Brown TA, Peaslee GF, Vogel JS. A new accelerator mass spectrometry system for 14C-quantification of biochemical samples. International Journal of Mass Spectrometry. 2002;218:255-264.

23. Hoppe PP, Krennrich G. Bioavailability and potency of natural- source and all-racemic alpha-tocopherol in the human: a dispute. Eur J Nutr. 2000;39:183-193.

24. Burton GW, Traber MG, Acuff RV, et al. Human plasma and tissue alpha-tocopherol concentrations in response to supplementation with deuterated natural and synthetic vitamin E. Am J Clin Nutr. 1998;67:669-684.

25. Traber MG, Burton GW, Ingold KU , Kayden HJ. RRR- and SRR- alpha-tocopherols are secreted without discrimination in human chylomicrons, but RRR-alpha-tocopherol is preferentially secreted in very low density lipoproteins. J Lipid Res. 1990;31:675-685.

26. Lauridsen C, Leonard SW, Griffin DA , Liebler DC, McClure TD, Traber MG. Quantitative analysis by liquid chromatography-tandem mass spectrometry of deuterium-labeled and unlabeled vitamin E in biological samples. Anal Biochem. 2001;289:89-95.

27. Cohn W. Bioavailability of vitamin E. Eur J Clin Nutr. 1997;51(suppl 1):S80-S85.

28. Buchholz BA, Dueker SR, Lin Y, Clifford AJ, Vogel JS. Methods and applications of HPLC-AMS. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms. 2001;172:910-914.

29. Gilman SD, Gee SJ, Hammock BD, et al. Analytical performance of accelerator mass spectrometry and liquid scintillation counting for detection of 14C-labeled atrazine metabolites in human urine. Anal Chem. 1998;70:3463-3469.

30. Eisengart A, Milhorat AT, Simon EJ, Sundheim L. The metabolism of vitamin E. II. Purification and characterization of urinary metabolites of alpha-to-copherol. J Biol Chem. 1956;221:807- 817.

31. Schultz M, Leist M, Petrzika M , Gassmann B, Brigelius-Flohe R. Novel urinary metabolite of alpha-tocopherol, 2,5,7,8- tetramethyl-2(2′-carboxy-ethyl)-6-hydroxychroman, as an indicator of an adequate vitamin E supply? Am J Clin Nutr. 1995;62(suppl 6):1527S- 1534S.

32. Schultz M, Leist M, Elsner A , Brigelius-Flohe R. alpha- Carboxyethyl-6-hydroxychroman as urinary metabolite of vitamin E. Methods Enzymol. 1997;282:297-310.

33. Lodge JK, Traber MG, Elsner A , Brigelius-Flohe R. A rapid method for the extraction and determination of vitamin E metabolites in human urine. J Lipid Res. 2000;41:148-154.

34. Henion J, Brewer E, Rule G. Sample preparation for LC/MS/MS: analyzing biological and environmental samples. Anal Chem. 1998;70:650A-656A.

35. Buchholz BA, Fultz E, Haack KW, et al. HPLC-accelerator MS measurement of atrazine metabolites in human urine after dermal exposure. Anal Chem. 1999;71:3519-3525.

36. Solomons NW, Bulux J. Effects of nutritional status on carotene uptake and bioconversion. Ann N Y Acad Sci. 1993;691:96- 109.

37. Vuong Ie T, Dueker SR, Murphy SP. Plasma beta-carotene and retinol concentrations of children increase after a 30-d supplementation with the fruit Momordica cochinchinensis (gac). Am J Clin Nutr. 2002;75:872-879.

38. van Lieshout M, West CE, van Breemen RB. Isotopic tracer techniques for studying the bioavailability and bioefficacy of dietary carotenoids, particularly beta-carotene, in humans: a review. Am J Clin Nutr. 2003;77:12-28.

39. Lemke SL, Dueker SR, Follett JR, et al. Absorption and retinol equivalence of beta-carotene in humans is influenced by dietary vitamin A intake. J Lipid Res. 2003;44:1591-1600.

40. Buchholz BA, Arjomand A, Dueker SR , Schneider PD, Buchholz BA, Vogel JS. Intrinsic erythrocyte labeling and attomole pharmacokinetic tracing of 14C-labeled folic acid with accelerator mass spectrometry. Anal Biochem. 1999;269:348-352.

41. Macallan DC, Fullerton CA, Neese RA, Haddock K, Park SS, Hellerstein MK. Measurement of cell proliferation by labeling of DNA with stable isotope-labeled glucose: studies in vitro, in animals, and in humans. Proc Natl Acad Sci USA. 1998;95:708-713.

42. U.S. Nuclear Regulatory Commission. Part 20: Standards for protection against radiation, Subpart K: Waste Disposal, 20.2005 Disposal of Scientific Wastes. In: Title 10: Code of Federal Regulations. Washington, DC: US Nuclear Regulatory Commission; 2004.

43. Wilson PDG, Hilton MG, Waspe CR , DC Steer, DR Wilson. Production of 13C-labelled b-carotene from Dunaliella salina. Biotechnol Lett. 1997;19:401-405.

44. Parker RS, Brenna JT, Swanson JE , Goodman KJ, Marmor B. Assessing metabolism of beta-[13C]carotene using high-precision isotope ratio mass spectrometry. Methods Enzymol. 1997;282:130-140.

45. Dueker SR, Lame MW, Segall HJ, et al. Production of carbon- 14 labeled phytonutrients in spinach for human nutrition studies employing accelerator mass spectrometry detection. FASEB J. 1998;12:A210.

46. Kurilich AC, Britz SJ, Clevidence BA , Novotny JA. Isotopic labeling and LC-APCI-MS quantification for investigating absorption of carotenoids and phylloquinone from kale (Brassica oleracea). J Agric Food Chem. 2003;51:4877-4883.

47. Lienau A, Glaser T, Tang G , Dolnikowski GG, Grusak MA, Albert K. Bioavailability of lutein in humans from intrinsically labeled vegetables determined by LC-APCI-MS. J Nutr Biochem. 2003;14:663-670.

48. Dolnikowski GG, Sun Z, Grusak MA , Peterson JW, Booth SL. HPLC and GC/MS determination of deuterated vitamin K (phylloquinone) in human serum after ingestion of deuterium-labeled broccoli. J Nutr Biochem. 2002;13:168-174.

49. Chen L, Chan SY, Cossins EA. Distribution of folate derivatives and enzymes for synthesis of 10-formyltetrahydrofolate in cytosolic and mitochondrial fractions of pea leaves. Plant Physiol, 1997;115:299-309.

50. Cossins EA, Chen L. Folates and one-carbon metabolism in plants and fungi. Phytochemistry. 1997;45:437-452.

51. Clifford AJ, Dueker SR, Fabbro EE, et al. Intrinsic radiolabeling of nutrients for human nutrition studies using accelerator mass spectrometry. In: Stevenson NR, ed. Isotope Production and Applications in the 21st Century. Proceedings of the 3rd International Conference on Isotopes Vancouver, Canada, 6-10 September 1999. Vancouver, British Columbia: World Scientific; 2000:377-379.

52. Patrick L. Beta-carotene: the controversy continues. Altern Med Rev. 2000;5:530-545.

53. Zhang P, Omaye ST. Antioxidant and prooxidant roles for beta- carotene, alpha-tocopherol and ascorbic acid in human lung cells. Toxicol In Vitro. 2001;15:13-24.

54. Meyskens FL Jr. Criteria for implementation of large and multiagent clinical chemoprevention trials. J Cell Biochem. 2000;77:115-120.

55. Gescher AJ\, Sharma RA, Steward WP. Cancer chemoprevention by dietary constituents: a tale of failure and promise. Lancet Oncol. 2001;2:371-379.

Le T. Vuong, PhD, Bruce A. Buchholz, PhD, Michael W. Lam, PhD, and Stephen R. Dueker, PhD

Dr. Vuong is with Vitalea Science, Inc., Davis, California; Dr. Buchholz is with the Center for Accelerator Mass Spectrometry, Lawrence Livermore National Laboratory, Livermore, California; Dr. Lam is with the Department of Molecular Biosciences, University of California, Davis; Dr. Dueker is with Vitalea Science, Inc., and the Department of Nutrition, University of California, Davis.

Please address correspondence to: Stephen R. Dueker, Department of Nutrition, University of California, 3135 Meyer Hall, One Shields Ave., Davis, CA 95616; Phone: 530-752-4630; Fax: 530-752-8966; e- mail: srdueker@ucdavis.edu.

Copyright International Life Sciences Institute and Nutrition Foundation Oct 2004