点击显示 收起
1 From the Unilever Health Institute, Unilever Research Vlaardingen, Vlaardingen, Netherlands.
2 The ethyl-ß-apo-8'-carotenoate used in the study was donated by Hoffmann-La Roche, Basel, Switzerland. 3 Address reprint requests to AJC Roodenburg, Unilever Health Institute, Unilever Research Vlaardingen, Post Office Box 114, 3130 AC Vlaardingen, Netherlands. E-mail: annet.roodenburg{at}unilever.com.
ABSTRACT
Background: Fat-soluble vitamin E and carotenoids are regarded as being protective against chronic diseases. Little is known about the effect of dietary fat on the bioavailability of these compounds.
Objective: The objective of this study was to assess the effect of the amount of dietary fat on plasma concentrations of vitamin E and carotenoids after supplementation with these compounds.
Design: During two 7-d periods, 4 groups of 14–15 volunteers received daily, with a low-fat hot meal, 1 of 4 different supplements: vitamin E (50 mg), - plus ß-carotene (8 mg), lutein esters (8 mg lutein), or placebo. The supplements were provided in a low- or high-fat spread supplied in random sequence during either of the 2 experimental periods.
Results: As anticipated, plasma concentrations of vitamin E, - and ß-carotene, and lutein were significantly higher in the supplemented groups than in the placebo group. The amount of dietary fat consumed with the hot meal (3 or 36 g) did not affect the increases in plasma concentrations of vitamin E (20% increase with the low-fat spread and 23% increase with the high-fat spread) or - and ß-carotene (315% and 139% with the low-fat spread and 226% and 108% with the high-fat spread). The plasma lutein response was higher when lutein esters were consumed with the high-fat spread (207% increase) than with the low-fat spread (88% increase).
Conclusion: Optimal uptake of vitamin E and - and ß-carotene requires a limited amount of fat whereas the amount of fat required for optimal intestinal uptake of lutein esters is higher. 2000;71:–93.
Key Words: Dietary fat • bioavailability • absorption • ß-carotene • -carotene • lutein • carotenoids • -tocopherol • vitamin E • Netherlands
INTRODUCTION
Epidemiologic studies showed that an increased intake of vitamin E or carotenoids is associated with a reduced risk of cardiovascular disease (1–3) and cancer (4–6). Vitamin E and carotenoids have biological activity as antioxidants (7), which may mediate these putative beneficial effects. Important dietary sources of vitamin E are oils and fats (8, 9); the richest dietary sources of carotenoids are fruit and vegetables (10, 11). Vitamin E and carotenoids are fat-soluble compounds and their absorption involves solubilization in bile salts and incorporation into micelles. Actual intestinal uptake is believed to occur by passive diffusion along with uptake of dietary fat. The presence of dietary fat is thought to be important for micelle formation in the small intestine; dietary fat therefore may also be crucial for absorption of vitamin E and carotenoids (12–14).
Little information is available on the influence of dietary fat on vitamin E absorption in humans and results from animal studies are conflicting. Dimitrov et al (15) concluded from a small study in humans that dietary fat enhances the absorption of vitamin E. However, in some studies in rats, the amount of dietary fat present did not influence the apparent absorption of vitamin E (16, 17). In contrast, vitamin E absorption, measured as lymphatic appearance of radiolabeled vitamin E in rats, was increased with higher intakes of saturated fat (18).
Most previous research on the effect of dietary fat on the bioavailability of carotenoids in humans focused on ß-carotene (19–23). The results of these studies indicated that the presence of fat was essential for the intestinal uptake of ß-carotene. However, in some studies, the influence of dietary fat was compared with the effect in the complete absence of fat at the time of ß-carotene ingestion (20, 21). Although this situation may be applicable to the very-low-fat diets of some populations in developing countries, it is not representative of a Western diet. For comparison purposes, an average hot meal in the Netherlands contains as much as 40 g fat (24). The present study was designed to investigate the effect of the amount of dietary fat on the plasma response to supplementation with vitamin E or carotenoids. We used 2 amounts of dietary fat that are achievable in a Western diet.
SUBJECTS AND METHODS
Subjects
Participants were recruited from the local population by means of advertisements in newspapers. Sixty nonsmoking subjects (23 men and 37 women) aged between 18 and 70 y and with a reported body mass index (BMI; in kg/m2) between 19 and 30 were selected. The subjects did not use medication, apart from oral contraceptives, or follow a medically prescribed diet or weight-loss regimen. They did not use any supplements containing vitamin C, vitamin E, carotenoids, calcium, or iron. The subjects' body weights had been stable for 1 mo before the start of the study. None of the women were pregnant or lactating. The selected subjects were apparently healthy as evaluated by a medical history questionnaire.
Two women were withdrawn from the study, one because her fasting plasma cholesterol concentration was >8.5 mmol/L and one because of intercurrent illness not related to the study. Data from a third subject were not used because we suspected prior use of carotenoid supplements (plasma - and ß-carotene concentrations were 285 and 2068 nmol/L, respectively). The protocol for the study, which had been approved by the Medical Ethical Committee Unilever Nederland, was fully explained to the volunteers, who gave their written, informed consent before participation.
Experimental design
A split-plot design was used for this study. Each subject was randomly assigned to 1 of 4 experimental groups after stratification for sex, age, and BMI. One group received a placebo during two 7-d experimental periods. The other 3 groups received supplements of vitamin E, - plus ß-carotene, or lutein esters. The placebo and antioxidant supplements were provided in a low- or high-fat spread. All subjects consumed these low- and high-fat spreads in a crossover design during either of the 2 experimental periods. The subjects were randomly assigned to treatment sequences. The experimental periods were separated by a washout period of 5 wk during which no intervention was carried out. The spreads were consumed daily with a low-fat hot meal in the evening. The amounts of fat, carotenoids, and vitamin E consumed during the remainder of the study were restricted.
The subjects were instructed to maintain their usual lifestyles and patterns of physical activity during the entire study period, including the washout period. In addition, subjects were not allowed to take any supplements containing vitamin C, vitamin E, carotenoids, calcium, or iron. Before and after each experimental period, each subject's body weight was measured and fasting venous blood samples were collected. At the end of the experimental periods, blood sampling occurred 12 h after the last experimental meal. At the end of the first experimental period, body height was also measured.
Dietary restrictions
All subjects were asked to adhere to strict dietary prescriptions to limit their intakes of fat, carotenoids, and vitamin E during the experimental periods. Subjects were instructed to eat breakfast and a low-fat lunch containing negligible amounts of vitamin E and carotenoids, a maximum of only 0.25 g fat after lunch and in the late evening, and no fat for 2 h before and after the supplemented hot meal. From previous work, it was deduced that this range of 2 h would minimize the influence of fat from dietary sources other than the experimental meal (25). After lunch, the subjects were not allowed to consume any products containing vitamin E or carotenoids. Besides the low-fat hot meal and the experimental spread, the subjects received an additional low-fat spread without ß-carotene and vitamin E (Promise Ultra; VandenBergh Foods Co, Lisle, IL) to use with their lunch during the experimental periods. Compliance with the instructions was assessed by asking the subjects about their diets during the experimental periods and their consumption of the experimental spreads and hot meals and by weighing the amount of spread left over in the tubs.
Experimental spreads
The supplementary dosage of vitamin E was 50 mg -D-tocopherol/d (67% purity; Sigma Chemical Co, St Louis), equivalent to 5 times the recommended daily allowance of 10 mg/d for men aged 22–50 y (26, 27). The carotenoid-enriched spreads provided either 15 µmol (8 mg) - and ß-carotene/d (30% suspension in oil; Vegex Carotene; Quest International Ireland Ltd, Cork, Ireland) or 15 µmol (8 mg) lutein/d, primarily as lutein diesters (3.2% suspension in oil; Vegex Lutein OS30; Quest International Ireland Ltd). The daily amount of carotenoids added to the spread was 1–1.3 times the daily carotenoid intake in the United States (6 mg) (10) and the Netherlands (7 mg) (28). The control spread contained no added carotenoids or vitamin E.
The subjects were required to eat 50 g of the experimental spread daily; the low-fat spread contained 3% fat (by wt) and the high-fat spread contained 80% fat (by wt). In this way, a maximum contrast between the low- and high-fat intervention, together with an acceptable meal, could be achieved. The spread was freshly prepared in our laboratory for each experimental period. The fat contents and fatty acid compositions were kept similar in the high- and low-fat control and supplemented spreads. The fatty acid compositions of the low- and high-fat spreads reflected those of commercially available spreads.
Experimental meals
The subjects were instructed to consume the spread together with the low-fat hot meal that was provided for consumption at home. These low-fat hot meals consisted of 70 g prepared lean meat and 160 g low-carotenoid vegetables containing negligible amounts of vitamin E. Consumption of low-fat gravy and potatoes was allowed ad libitum. These low-fat hot meals were designed to contain <5 g fat. The low-fat spread contained 1.5 g fat and the high-fat spread provided an additional amount of 40 g. Thus, total fat intake per meal was calculated to be <6.5 g for the low-fat meal and <45 g for the high-fat meal.
To correct for the effects of meal preparation on the amounts of fat, carotenoids, and vitamin E consumed, duplicate portions were collected. According to the instructions given to the subjects, representative meals were prepared for each study group and each period. The average amount of potatoes (266 g) and low-fat gravy (29 g) used by the subjects was calculated on the basis of the quantities used, as reported by the subjects by questionnaire. The samples were stored under argon at -80°C until analyzed.
Laboratory analyses
Plasma
Fasting venous blood samples were collected before and after each intervention period. Plasma was prepared in tubes containing EDTA and stored at -80°C. Plasma total cholesterol and triacylglycerol concentrations were determined spectrophotometrically by using enzymatic methods with commercially available test kits (Boehringer Mannheim, Mannheim, Germany). Carotenoids in plasma were extracted with heptane:dichloromethane (5:1, by vol) by using ethyl-ß-apo-8'-carotenoate (Hoffmann-La Roche, Basel, Switzerland) as an internal standard. The sample was injected onto a Suplex 25 x 4.6-mm ID, 5-µm column (pkb-100; Supelco Inc, Bellefonte, PA) and eluted with 93.5% methanol:acetonitrile:toluene (A) (55:44:2, by vol) and 6.5% water containing 0.1% ammonium acetate (B). Ten minutes after injection of the samples, a linear gradient started from 93.5% A and 6.5% B to 100% A at 45 min. Between 45 and 47 min, the eluent returned to 93.5% A and 6.5% B, which was maintained until 55 min. The flow rate was 1.0 mL/min. Detection of carotenoids was performed by ultraviolet-visible spectrometry at 450 nm. Plasma -tocopherol concentrations were determined by HPLC on a 5-µm column (Lichrospher RP-18; Merck, Darmstadt, Germany). -Tocopheryl acetate (Merck) was used as an internal standard. The mobile phase consisted of methanol:isopropanol:water (50:50:8, by vol) and the flow rate was 0.6 mL/min. -Tocopherol was detected by ultraviolet-visible spectrometry at 292 nm and -tocopheryl acetate at 284 nm.
Experimental meals
Duplicate samples of the meals were analyzed for contents of fat, fatty acids, -tocopherol, - and ß-carotene, and lutein. The values in Table 1 are based on the chemical analyses and are expressed as daily intakes from an average hot meal. To determine the fat contents, celite (Merck) was added to the duplicate samples, which were then freeze-dried. The dried samples were extracted twice with dichloromethane in a Soxtec system (System HT6; Tecator, Hoganas, Sweden). The recovery rate of the fat extraction was 99–103%, as determined with a standard addition of sunflower oil to the duplicate samples. The CV was 1.4%. The fat contents of the meals containing the low- and high-fat spreads (3 and 36 g, respectively) were lower than expected (6.5 and 45 g, respectively) (Table 1). Fatty acid composition was determined as described previously (17). Extraction of vitamin E and carotenoids from the meals containing the high-fat spread was performed with heptane:diethylether (3:1, by vol). These extractions were repeated 3 times; for the meals containing the low-fat spread, diethylether:methanol (5:2, by vol) was used. These extractions were repeated 2 times with diethylether. -Tocopherol and carotenoids were analyzed by reversed-phase HPLC with a C30 S-5µm, 4.6 x 150-mm column (YMC, Inc, Wilmington, NC). Ethyl-ß-apo-8'-carotenoate and -tocopheryl acetate were used as internal standards. The column was eluted with 95% methanol:tert-butylmethylether:1.5% aqueous ammonium acetate (A) (83:15:2, by vol) and 5% tert-butylmethyl ether:methanol:1.5% aqueous ammonium acetate (B) (90:8:2, by vol). Ten minutes after sample injection, a linear gradient started from 95% A and 5% B to 55% A and 45% B at 22 min. Between 22 and 34 min, a linear gradient was used from 55% A and 45% B to 5% A and 95% B, which remained until 39 min; from 39 to 44 min, the eluents returned to 95% A and 5% B, which was maintained until 50 min. The flow rate was set at 1.0 mL/min. Carotenoids were detected by ultraviolet-visible spectrometry at 450 nm, -tocopherol at 292 nm, and -tocopheryl acetate at 284 nm.
View this table:
TABLE 1.. Average nutrient compositions of low- and high-fat meals consumed daily1
Statistics
Different supplement conditions were tested against the between-subject error and the effect of the amount of fat was tested against the within-subject error. The changes over the experimental periods of the variables measured in the supplemented groups were compared with those in the control group. The following factors were used in the analysis of variance: amount of fat, type of supplement, subject, sex, experimental period, and interaction of all these factors. If a significant interaction between amount of fat and type of supplement was observed, the analysis was executed separately for each amount of fat and type of supplement. Significant deviations from the control group were determined by the Dunnett test. The data were analyzed by using the statistical package SAS (SAS Institute, Cary, NC).
To compare the relative plasma response after - and ß-carotene and lutein supplementation, the changes in each of the plasma carotenoid concentrations were expressed relative to their intakes. For each fat intake, differences in the relative responses of - and ß-carotene in the group that received the - plus ß-carotene supplement were determined with a paired t test. Differences between lutein response in the lutein group and - or ß-carotene response in the group that received the - plus ß-carotene supplement were determined by using an unpaired t test. All comparisons were 2-sided. P values <0.05 were considered significant.
RESULTS
Baseline characteristics and compliance
More women than men participated in the study (Table 2). Mean (±SD) BMI and age of all subjects at baseline were 25.0 ± 3.2 and 46.4 ± 13.4 y, respectively. There were no significant differences among the 4 groups (data not shown). Compliance, as assessed by using compliance forms and weighing spread left over in the tubs, was excellent. All subjects consumed all of the meals during the 2 experimental periods and <3% of the spread was left over. The average calculated energy consumption per meal was 6 MJ, which was high for Dutch standards [4 MJ for men and 3 MJ for women aged 22–50 y (24)] but indicated that there was possibly some compensation for reduced energy intake due to the strict dietary prescriptions.
View this table:
TABLE 2.. Body weight and plasma lipid concentrations at baseline and changes after 7 d consumption of a low- or high-fat meal not supplemented (control) or supplemented with vitamin E or carotenoids1
Body weight, plasma cholesterol, and triacylglycerol
In all groups, body weight and plasma concentrations of cholesterol and triacylglycerol decreased during each of the two 7-d intervention periods (Table 2). However, there were no significant differences between the supplemented groups and the control group. The decrease in body weight was significantly smaller after the high-fat intervention, whereas the decrease in plasma triacylglycerol concentration was significantly smaller after the low-fat intervention.
Plasma -tocopherol and carotenoids
Because plasma -tocopherol and carotenoid concentrations are associated with plasma lipid concentrations (29), they were corrected for plasma cholesterol and triacylglycerol concentrations before statistical analysis. However, because the uncorrected concentrations are more informative, these concentrations and changes in concentrations of plasma -tocopherol and carotenoids after supplementation with the low- or high-fat spreads are shown in Table 3.
View this table:
TABLE 3.. Plasma concentrations of -tocopherol and carotenoids at baseline and changes after 7 d consumption of a low- or high-fat meal not supplemented (control) or supplemented with vitamin E or carotenoids1
Plasma concentrations of -tocopherol, - and ß-carotene, and lutein were significantly higher after consumption of spreads supplemented with vitamin E, - and ß-carotene, or lutein than after consumption of the control spread. There were decreases in plasma -tocopherol and carotenoid concentrations in the control group and in plasma lycopene and ß-cryptoxanthin concentrations in all groups. Changes in plasma concentrations of lycopene and ß-cryptoxanthin in the supplemented groups were not significantly different from those in the control group.
There was no significant effect of the amount of fat on the response in plasma -tocopherol concentration after consumption of the vitamin E–supplemented spreads (Table 3). In the case of the spreads supplemented with - and ß-carotene, the increases in plasma - and ß-carotene concentrations after the low-fat intervention were slightly, but not significantly, larger than after the high-fat intervention (Table 3). This could be explained at least in part by a higher amount of - and ß-carotene in the low-fat spread (Table 1, Figure 1). The increase in plasma lutein concentration was significantly larger after the high-fat intervention than after the low-fat intervention.
FIGURE 1. . Mean (±SEM) plasma carotenoid responses after 7 d consumption of a low-fat () or high-fat () meal supplemented with carotenoids, expressed as the increase (in nmol/L) per µmol carotenoid supplemented (-carotene intake, 4.1 and 3.8 µmol/d; ß-carotene intake, 8.1 and 7.8 µmol/d; lutein intake, 14.0 and 13.3 µmol/d in low- and high-fat meals, respectively). With the low- and high-fat meals, the relative plasma responses were significantly different among carotenoids (P < 0.001 for -carotene compared with ß-carotene, -carotene compared with lutein, and ß-carotene compared with lutein). For lutein only, the relative responses between the low- and the high-fat meals were significantly different (P < 0.001).
Because - and ß-carotene and lutein esters were consumed in about the same amounts (Table 1), we compared the responses in plasma carotenoid concentrations relative to their intakes. For the low- and high-fat interventions, these relative responses were significantly higher for - and ß-carotene than for lutein and the relative response for -carotene was significantly (20%) higher than for ß-carotene (Figure 1). In addition, the relative response for lutein was significantly higher with the high- than with the low-fat intervention (Figure 1).
DISCUSSION
The present study showed that the bioavailability of vitamin E, -carotene, and ß-carotene was similar when these compounds were consumed with a low-fat meal (3 g fat) or a high-fat meal (36 g fat), whereas the bioavailability of lutein esters was significantly lower when lutein esters were consumed as part of a low-fat meal. Bioavailability was assessed in terms of changes in plasma concentrations after 7 d supplementation with vitamin E, - and ß-carotene, or lutein esters.
In the control group, there were reductions in plasma concentrations of carotenoids other than those supplied to the subjects and in plasma -tocopherol concentrations. These reductions may have been due to an interaction between nutrients or may have been the result of good compliance with the dietary restrictions. The subjects were instructed to refrain from foods rich in vitamin E and carotenoids and the experimental meals contained virtually no carotenoids and little vitamin E.
Analysis of the composition of the meals showed a slightly smaller amount of vitamin E and carotenoids in the high-fat meals than in the low-fat meals. This may have been due to the larger amount of unsaturated fatty acids in the high-fat spreads. It is likely that the supplemented antioxidants were used to prevent oxidation of these polyunsaturated fatty acids. The possibility that this also reduced the bioavailability of the antioxidants from the high-fat spreads cannot be excluded (17, 18). In addition, dietary polyunsaturated fatty acids, but not saturated fatty acids, had been shown to increase intestinal 15,15'-dioxygenase activity in rats, suggesting a higher conversion of provitamin A carotenoids to retinol (30). Although it was impossible to quantify these effects in the present study, they may explain the slightly lower plasma response after - and ß-carotene supplementation with the high-fat spread than with the low-fat spread.
Vitamin E
Dietary fat is generally believed to be necessary for the intestinal uptake of vitamin E (12, 14). A previous crossover study with only 6 human volunteers showed a larger increase in plasma -tocopherol concentrations when 5 d supplementation with vitamin E was followed 6–8 h later by a high fat intake (45 g fat) than when followed by a low fat intake (6 g fat) (15). However, that study was rather small and could have been flawed by the fact that the vitamin E dose with the high-fat intake was higher than that with the low-fat intake (15). We conclude from the present results that only a small amount of fat (3 g) is sufficient to ensure the uptake of vitamin E.
-Carotene and ß-carotene
The results of previous studies suggest that the presence of dietary fat is crucial to the bioavailability of ß-carotene. However, in those studies the effect of dietary fat was compared with the effect in the complete absence of fat at the moment of ingestion of ß-carotene (20, 21). Recently, Jalal et al (23) reported a significantly larger increase in serum retinol concentrations in vitamin A–deficient children when a sweet-potato snack (providing 4.5 mg ß-carotene/d) was ingested with 18 g fat than when ingested with 3 g fat. In addition, Jayarajan et al (19) found no difference in improvement in vitamin A status when 5 or 10 g dietary fat was added to spinach (providing 1.2 mg ß-carotene/d), whereas 0 g fat resulted in a smaller increase. Despite differences in food matrices in these studies, the data suggest that a minimum amount of 3–5 g dietary fat may be needed in a meal to ensure intestinal carotene uptake. In the present study, the low-fat meal with which - and ß-carotene were ingested contained 3 g fat. Apparently, this amount of fat was sufficient for the intestinal uptake of - and ß-carotene.
Lutein
The influence of dietary fat on the bioavailability of lutein differed from the influence on vitamin E and - and ß-carotene. About 3 g dietary fat was not sufficient to ensure absorption of the same magnitude as when 36 g fat was present. This result was unexpected because lutein is less lipophilic than are - and ß-carotene and a larger effect of the amount of dietary fat would be expected for the more lipophilic carotenoids. However, the lutein used in this study was esterified, mainly with palmitic acid (31). Although the partition coefficient between fat and water for the lutein esters is unknown (32), it is likely that lutein esters are more lipophilic than are - and ß-carotene. It can thus be speculated that in the presence of only small amounts of fat, the emulsification of lutein esters in the intestine is less than that of - and ß-carotene.
Granado et al (31) detected some lutein monopalmitate in plasma after 4 mo supplementation with lutein esters (15 mg/d). However, it is generally assumed that most lutein esters are hydrolyzed before or during absorption and that free lutein is absorbed, which is the case with cholesteryl esters (33) and ß-cryptoxanthin esters (34). Hydrolysis of esters is mediated by esterases and possibly also by lipases. The excretion of these enzymes by the pancreas is regulated by the presence of fat in the stomach and the duodenum. In addition, the activity of esterases and lipases is substantially enhanced if the amount of fat is sufficient to form lipid-aqueous interfaces in the duodenum (33). Our results suggest that the release of esterases and lipases or the formation of lipid-aqueous interfaces was hampered at a low fat intake, resulting in a reduced uptake of lutein esters.
It was shown previously that the bioavailability of lutein esters is similar to that of lutein when supplemented to a diet (35). Results from the present study suggest, however, that this may depend on the amount of fat consumed with the lutein esters. If lutein is present as free lutein, as is the case in lutein-rich vegetables such as spinach, broccoli, kale, and green peas (36), the amount of fat required for optimal uptake may be smaller than in the case of lutein esters.
Comparison of relative plasma responses of carotenoids
Using similar amounts of - and ß-carotene and lutein allowed us to compare plasma responses. The daily intakes of carotenoids varied only slightly (Table 1). The relative plasma response was calculated by dividing the plasma responses measured by the amount of carotenoids supplied daily, thus assuming a linear dose-response relation (Figure 1). Substantial differences in relative plasma responses among the carotenoids were shown. Previous studies also showed that the relative plasma response of lutein was less than that of ß-carotene after 3–4 wk supplementation (37, 38). In line with the findings of van het Hof et al (39), the relative plasma response of -carotene exceeded that of ß-carotene after supplementation with palm oil carotenoids. These differences may have resulted from differences in absorption, metabolism, or tissue distribution between carotenoids.
The results of the present study illustrate that enrichment of food products with -carotene, ß-carotene, lutein, and vitamin E can be an effective way to enhance the plasma status of these antioxidants and thus possibly reduce the risk of degenerative diseases (1–6). Fat-containing products are suitable carriers of these fat-soluble antioxidants and only a small amount of fat is needed to ensure uptake of these minor components. The minimum amount of fat required depends, however, on the physicochemical characteristics of the fat-soluble compounds.
In conclusion, the present study showed that consumption of a low- or high-fat meal enriched with vitamin E, -carotene and ß-carotene, or lutein esters enhanced the plasma concentrations of these antioxidants. A small amount of fat was sufficient to optimize the uptake of vitamin E and - and ß-carotene, whereas a comparable amount of lutein esters required a larger amount of fat.
ACKNOWLEDGMENTS
We thank Jan-Willem van den Berg, Marieke Beute, Cor Blonk, Jan Don, Willy Dubelaar, Yvonne Gielen, Renate Jacobs, Gerard Kivits, Jolanda Mathot, Wim van Nielen, Wil van Oort, Henk van Oosten, Irene Samwel, Esther Schra, Sjaak Sies, Miranda Slotboom, Saskia van Stroe-Biezen, Henk van Toor, Jan van Toor, Wim Tuitel, Suzanne van der Veen, Tom Wiersma, and Koos van Wijk for their technical assistance. We acknowledge Clive E West, Wageningen Agricultural University, for his helpful comments on the manuscript.
REFERENCES