2.1 Chemicals and Reagents

Solvents used for sample dilution and chromatography were purchased from Carl Roth (Karlsruhe, Germany), Sigma-Aldrich (Steinheim, Germany), and VWR (Darmstadt, Germany) in gradient or analytical grade. Formic acid (FA) was obtained from Grüssing (Filsum, Germany). Water for dilution and HPLC analysis was purified using a Milli-Q Gradient A10 system (Millipore, Schwalbach, Germany). β-Glucuronidase from Helix pomatia was purchased from Sigma-Aldrich.

Proline betaine hydrochloride (manufacturer's designation: stachydrine hydrochloride) was purchased from Extrasynthese (Genay, France). Hesperetin and phloretin were obtained from Sigma-Aldrich and naringenin was obtained from Carl Roth. Taxifolin was obtained from Leblang GmbH (formerly Foxtank GmbH; Berlin, Germany).

All analytes were dissolved in acetonitrile (ACN) or ACN/H₂O and the multi analyte stock solutions were prepared at a concentration of 50 µg mL⁻¹ in ACN/H₂O and stored at −80 °C. In addition, working solutions containing 1.0 and 0.1 µg mL⁻¹ in ACN/H₂O (10/90) were used for matrix matched calibration.

2.2 Intervention Study

A short-term dietary intervention study was conducted with 32 subjects. All subjects were informed about the scope and the aim of the study and gave written consent about their participation. The study was approved by the research ethical committee of the University Hospital Münster, Germany (File reference: 2014-632-f-S) as a follow-up study of ref. [1].

2.2.3 Study Design

The dietary intervention study was conducted over the course of 8 days. During the study, the subjects were not allowed to consume any orange juice, citrus fruits, juices or any products containing citrus fruits. The only exception was a specific amount of orange juice in the morning of day 4 of the study (see Figure 1). For the whole period of the study, the participants were requested to document their daily diet using food diaries. Except for the mentioned avoidance of citrus fruits, juices or products containing citrus fruits, no further diet restrictions were made.

All subjects were asked to collect one spot urine sample in the morning before the orange juice consumption on day 4, which was used as a blank sample. Before the orange juice intake, each participant was either assigned to the low, medium or high consumption group. Each subject then received a portion of orange juice which was consumed within one hour; subjects of the low consumption group received 250 mL, subjects of the medium consumption group 500 mL, and subjects of the high consumption group ingested 1 L of orange juice. In the first 48 h following orange juice intake, the participants collected a spot urine sample every time they urinated, as well as an aliquot of the morning urine sample on the 6th day of the study; all points in time of urination had to be recorded. Furthermore, they collected an aliquot of the morning urine sample approximately 72 h (7th day) and 96 h (8th day) after orange juice consumption (see Figure 1). All urine samples were pre-frozen at −20 °C and subsequently stored at −65 °C until analysis.

2.3 Analysis of Urine Samples

After thawing the urine samples to room temperature, 100 µL of urine were mixed with 10 µL 1.5 m sodium acetate solution (pH 4.8) and 25 µL of the injection standard taxifolin (1 µg mL⁻¹). Subsequently, 30 µL of β-glucuronidase solution (1500 U in 150 mm sodium acetate solution pH 4.8) were added and the samples were incubated at 37 °C for 1.5 h, followed by addition of 10 µL ACN (+ 0.1 % FA). The solutions were vortexed before adding 825 µL water (+ 0.1% FA), resulting in a dilution of the urine of 1:10 (v/v) over all steps. After centrifugation at 15 000 × g for 5 min to remove solid residues, the supernatant was directly used for HPLC-MS/MS analysis. All samples were prepared in duplicate.

The fluid intake of the participants was not monitored. In order to take dilution of urine into account the urinary creatinine levels were determined for each sample and biomarker concentrations normalized to µg mg⁻¹ creatinine. About half of the samples were measured by the central laboratory of the University Hospital Münster using an ADVIA 1800 clinical chemistry analyzer by application of the Jaffé method (Siemens Healthcare Diagnostics, Eschborn, Germany). The other half was analyzed using an in-house method described by Sueck et al.^[23]

2.4 HPLC-MS/MS Conditions

The HPLC-MS/MS analysis was carried out by using an Agilent 1260 series system (Agilent Technologies, Santa Clara, USA) with a G1312B Binary Pump, a G1367E autosampler with a G1330B thermostat, a G4225A degasser, a G1316A column oven, and a G4208A instant pilot module coupled to an SCIEX QTRAP 5500 mass spectrometer (SCIEX, Darmstadt, Germany).

2.5 Validation of the Analytical Method

The validation of the analytical method for this follow-up study was carried out as previously described^[1] by using external matrix matched calibration. The spiking levels for matrix matched calibration equal concentration levels of 0.001, 0.0025, 0.005, 0.0075, 0.01, 0.025 0.05, 0.075, 0.1, 0.25, 0.5, 0.75, 1.0, 2.5, and 5.0 µg mL⁻¹ for each analyte in the samples measured by HPLC-MS/MS and tenfold higher concentrations in the undiluted urine samples.

2.6 Statistical Analysis and Data Presentation

Statistical analysis and data presentation were performed as described by Saenger et al.^[1] For statistical comparison of the groups, the urinary concentration of each potential biomarker for orange juice consumption in the period 0–24 h, 0–48 h, and 24–48 h after consumption was summed up for each subject. Statistical comparison within one group was based on excretion kinetics. The excretion at each point in time was compared to the concentration of the biomarker of interest in the blank urine sample. Changes were interpreted as significant at a p-value of 0.05 or lower. The most significant changes are discussed in the RESULTS section below; a detailed summary of the results of all executed statistical tests can be found in the Supporting Information (Tables S2–S6, Supporting Information). The data regarding the excretion kinetics are presented as box-whiskers-plots according to Tukey. Outliers are indicated in the corresponding figures as ●, however all outliers were included in the statistical evaluation.

2.7 Determination of the Biomarker Limit Concentration

Since the maximum biomarker concentration of all blank samples (x_max) was used, the calculated limit concentration is higher as if the mean value of all blank samples would have been used. Therefore, the result of the calculation is more reliable and the likelihood of orange juice intake is very high in case the limit concentration of the biomarker in urine is exceeded by the respective concentration in the analyzed sample.

3.1 Dietary Intervention Study Design

A dietary intervention study was performed with 32 volunteers to identify urinary biomarkers characteristic for the consumption of orange juice and to differentiate between low, medium, and high intake. The design of the study is summarized in Figure 1 and a detailed description can be found in the Experimental Section. Briefly, the participants of the study were requested to avoid any orange juice and related products for the first three days of the study (wash-out phase) and throughout the study, except for one orange juice portion on day 4. The volunteers were divided into three intake groups. On the morning of day 4, the subjects assigned to the low, medium, and high consumption group consumed 250 mL, 500 mL, and 1 L orange juice, respectively. Subsequently, an aliquot of every given urine sample was collected in the first 48 h following orange juice intake. All participants were requested to furthermore collect an aliquot of the morning urine sample on the 6th as well as 7th day of the study, equaling approximately 72 and 96 h after orange juice consumption. The morning urine sample collected prior to the orange juice intake on day 4 was used as blank urine sample. All urine samples obtained over the course of the study were analyzed by HPLC-MS/MS regarding their contents of proline betaine, hesperetin, naringenin, and phloretin, which represent potential biomarkers for orange juice intake. Besides investigating the biomarkers’ excretion kinetics, the sum of the urinary biomarker concentrations determined for the consumption groups were compared for the periods of 0–24, 24–48, and 0–48 h after orange juice intake to examine if the consumption of different amounts is distinguishable with statistical significance.

3.2 Analytical Method Validation

An in-house validation of the HPLC-MS/MS method was performed and the parameters apparent recovery (R_A), limit of detection (LOD), and limit of quantification (LOQ) were determined by application of a matrix-matched calibration. Blank urine samples were spiked with proline betaine, hesperetin, naringenin, and phloretin. The limits of detection were 0.0056 µg mL⁻¹ for hesperetin, 0.0065 µg mL⁻¹ for phloretin, 0.0071 µg mL⁻¹ for naringenin, and 0.057 µg mL⁻¹ for proline betaine. The recovery rates ranged from 40.6 ± 1.2% (hesperetin) to 67.2 ± 5.4% (phloretin). The coefficients of variation of the analysis were lower than 5% for all analytes, evidencing a high precision of the analytical method. A summary of the validation parameters can be found in Table S7, Supporting Information.

Besides the quantitative analyses of the four analytes, synephrine sulfate was detected in some of the urine samples obtained after orange juice intake. The identity of this compound was confirmed by using accurate mass spectrometry measurements and recording fragmentation spectra using HPLC coupled to high-resolution mass spectrometry (HPLC-HRMS) (For details see Method S1, Supporting Information). MRM transitions for synephrine sulfate were derived from its structure. All samples were qualitatively analyzed for synephrine sulfate by using the calculated MRM transitions.

3.3 Biomarker Excretion after Orange Juice Intake

3.3.1 Proline Betaine

Within the first two hours after consumption, the proline betaine concentration in urine rises significantly. The main amount is excreted after 24 h, but the proline betaine concentration in urine is still significantly higher after 24 h compared to the one of the blank urine sample (see Figure 2). Comparing some specific points in time (see Figure 3), it can be concluded that the proline betaine concentration is still significantly higher up to 72 h (low consumption) and 96 h (medium and high consumption) after orange juice consumption compared to the blank urine sample collected before intake (Figure 3).

The sum of the urinary proline betaine concentrations for the periods of 0–24, 0–48, and 24–48 h after orange juice consumption are shown in Figure 4. All three groups can be distinguished based on the sum of the urinary proline betaine concentrations with statistical significance for these periods, even for the period of 24–48 h after consumption (p ≤ 0.05, Tables S2, S5, and S6, Supporting Information).

3.3.2 Naringenin and Hesperetin

The naringenin and hesperetin concentration in urine did not rise as fast as the proline betaine concentration, with the maximum excretion reached 4–10 h after consumption. In contrast to proline betaine, the urinary concentrations of naringenin and hesperetin decreased below the LOQ in most samples within the first 24 h after intake (see Figures 5 and 6). Regarding the above-mentioned specific points in time after orange juice intake, only the hesperetin concentration in the urine samples representing the last excretion on the day of consumption is significantly higher compared to the one in blank urine samples. In case of naringenin, such a significant difference is only observable for the high consumption group (see Figures S1 and S2, Supporting Information).

The different intake groups were compared based on the sum of the biomarker concentrations for the periods of 0–24, 0–48, and 24–48 h after orange juice consumption as described for proline betaine. Similar to the urinary proline betaine concentration, the median of the sum of the concentrations of naringenin and hesperetin in urine increased with the consumed amount of orange juice (see Figures 7 and 8, median equals horizontal bar in the boxes). Due to high variations within the groups (evidenced by the wide boxes and whiskers) however, a significant difference was not apparent between the sum of the concentrations of the three groups. The only exception is the comparison between the sum of the urinary concentrations of naringenin for the low and the high consumption group within the first 48 h after orange juice intake (p-value = 0.0279; for details on significance levels see Table S6, Supporting Information).

In addition to the analysis of all urine samples of the human study after enzymatic treatment with β-glucuronidase to cleave potential glucuronides, an intact glucuronide of hesperetin was isolated from human urine as one of its main metabolites in this study. The linkage of the respective glucuronic acid moiety to the hesperetin backbone was determined to be at position 3′ based on nuclear magnetic resonance (NMR) spectroscopy (for experimental details and NMR data see Method S2, Table S8, and Figure S4, Supporting Information). This is in accordance with data presented by Brand et al. describing hesperetin-3′-O-glucuronide as the main metabolite of hesperetin by employing human liver microsomes.^[25] As shown with this example, isolation and structure elucidation of other metabolites hardly or not at all commercially available might be feasible, thus enabling the access to reference compounds and allowing for the direct analysis of these metabolites without having to rely on enzymatic hydrolysis during sample preparation in future experiments.

3.3.5 Phloretin

Phloretin was also analyzed in urine samples as it was postulated as a potential metabolite of naringenin.^[8] Even though phloretin was detectable in the urine samples collected after orange juice intake from 29 out of the 32 participants, there was no correlation between the ingested amount of orange juice and the sum of the concentrations of phloretin regarding the first 24 h after intake (Figure 9). However, the detailed excretion kinetics indicate an increase of the urinary phloretin excretion after orange juice intake (see Figure S3, Supporting Information).

Based on a student t-test a statistical significance (p ≤ 0.05) was not evident regarding the urinary phloretin excretion (Figure 9) nor the individual excretion kinetics (Figure S3, Supporting Information). By taking the entries in the food diaries into account, the observed phloretin excretion can be explained by previous consumption of apples or apple products in the hours before the orange juice intake in some of the cases. However, other participants excreted higher amounts of phloretin within the first 24 h after orange juice intake without having consumed apples or apple products during this time based on their entries in the food diaries.

4.1 Proline Betaine

In the present study, it was demonstrated that it is possible to distinguish between low (250 mL), medium (500 mL), and high (1 L) orange juice consumption based on the sum of the urinary proline betaine excretions within the first 24 and 48 h after orange juice consumption. This supports a study presented by Lloyd et al. who showed that two groups ingesting different amounts of citrus products can be distinguished using the proline betaine excretion in urine.^[12] Gibbons et al. further evidenced that it is possible to estimate the consumed amount of orange juice by analyzing urine samples derived from another study regarding their proline betaine content based on an external calibration derived from data of a consumption study where different portions of orange juice were consumed.^[18] Despite the observed fast increase and decrease of the urinary proline betaine concentration after orange juice intake, a blank urine sample and samples collected for up to 72 h after intake can still be distinguished by using the biomarker limit concentration proposed in this study. In addition, such a biomarker limit concentrations (Table 1) might generally prove helpful to verify whether the consumption of a certain food can be assumed, especially when the data are derived from a large number of subjects. The proline betaine concentration is significantly higher even in urine samples collected up to 96 h after orange juice consumption compared to the blank urine samples. In contrast to other studies^{[11, 16]} the subjects in the presented study not only drank 250 mL of orange juice but were split in three groups with low (250 mL), medium (500 mL), and high (1 L) consumption. Furthermore, we analyzed the biomarker excretion for up to 96 h. In agreement with studies reported in literature independent of the consumed volume of orange juice, the main excretion of proline betaine takes place in the first 24 h after orange juice consumption, with a maximum urinary concentration detectable in the samples collected 2 to 6 h after intake (see Figure 2); but as shown in the present study, the excretion is not finished within the first 24 to 48 h. Even 72 h after consumption the proline betaine concentration in urine, in all groups, is significantly higher than in the blank urine sample. The low consumption group drank 250 mL of orange juice, which is similar to the setting by Lang et al.^[16]

4.2 Naringenin and Hesperetin

In contrast to proline betaine, the three consumption groups cannot be distinguished based on the urinary naringenin or hesperetin concentration, although naringenin and hesperetin were reported to be potential qualitative biomarkers for orange juice consumption. Manach et al. showed that both are only detectable in urine or blood after intake of orange juice and reported a maximum plasma concentration 5.4–5.8 and 2.0–4.6 h after intake for hesperetin and naringenin, respectively.^[7] In urine, both flavanones were detected first in samples collected 6–11 h after orange juice consumption. In these specific samples they also found the highest concentrations of naringenin and hesperetin. Neither naringenin nor hesperetin were detectable in plasma or urine 24 h after orange juice intake.^[7] Krogholm et al. found a maximum hesperetin concentration in plasma about 4.9 h after consumption and a maximum naringenin concentration about 3.6 h after consumption.^[26] In agreement with the results presented by Manach et al.,^[7] they did not detect any hesperetin or naringenin in the urine samples collected about 24 h after consumption.^[26] These data from literature correlate with the results of the present study. Herein, the maximum hesperetin and naringenin concentration in urine is reached approximately 4–10 and 2–12 h after consumption. In general, this is the first study reporting the extensive urinary excretion kinetics of hesperetin and naringenin after orange juice intake. A statistically significant differentiation between the blank urine sample and samples collected after orange juice consumption based on the urinary hesperetin or naringenin concentration is possible for 4–14 h after the consumption (see Table S5, Supporting Information). This illustrates the fast urinary clearance of the flavanones from the human body. The hesperetin levels in urine seem to rise slightly overnight. While there is no statistically significant difference regarding these levels between the urine samples obtained about 14 h after consumption and the blank urine sample, a difference was evident about 24 h after orange juice intake if blank samples are compared to morning urine samples collected the day after consumption (see Table S4, Supporting Information). Using the sum of the concentrations for either naringenin or hesperetin, the three consumption groups are not distinguishable with statistical significance, except in case of the comparison of the low and high consumption group regarding the sum of the naringenin concentration 0–48 h after intake (p = 0.0279) (Figure 7). This might be due to high interindividual variations regarding the naringenin and hesperetin concentration in urine within these groups. Nevertheless, the medians of the sum of the urinary concentrations of naringenin and hesperetin of the three intake groups indicate a trend, as they rise with the amount of orange juice consumed. In total, the results obtained in the present study regarding the applicability of hesperetin and naringenin as biomarkers for orange juice consumption confirm the statement of Lloyd et al. who found that these two compounds represent rather limited biomarkers for citrus consumption, especially compared to proline betaine.^[12]

4.3 Biomarker Limit Concentration

Besides investigating the urinary excretion kinetics after orange juice intake and comparing the different intake groups by means of the sum of the urinary concentrations of naringenin or hesperetin, the biomarker limit concentration was calculated (see Table 1). The biomarker limit concentration of hesperetin indicates its suitability as a qualitative short-term biomarker for orange juice consumption, as this concentration is still exceeded in 80% of the urine samples collected up to 24 h after consumption in case of the high consumption group.

In preliminary experiments, we found that proline betaine, hesperetin, and naringenin are potential biomarkers for the intake of citrus fruits in general and related products, not specifically for orange juice consumption (Method S3, Table S9, and Figure S5, Supporting Information). This supports Andersen et al. who stated that some dietary biomarkers are rather biomarkers for foods derived from a whole botanical family instead specific biomarkers for a single fruit or vegetable.^[27]

4.4 Synephrine Sulfate

In 2017, Bader et al. were the first to describe synephrine as a potential biomarker for orange juice and tangerine juice consumption. They showed a fast excretion: the maximum synephrine concentration in urine was reached approximately 4 h after intake, followed by a rapid decrease. About 12 h after the consumption of 250 mL orange juice, the urinary synephrine excretion was completed.^[17] Kusu et al. previously reported a relation between consumption of satsuma pulp and synephrine excretion. In this study, a rather fast urinary excretion of synephrine was described as well, with the maximum concentration in urine reached approximately 2–3 h after consumption and a complete excretion observed 16 h after intake.^[28] As both groups subjected the urine samples to hydrolysis or enzymatic treatment with glucuronidase and sulfatase prior to the analysis, no metabolites of synephrine were directly analyzed. Ibrahim et al. however described unconjugated hydroxymandelic acid, hydroxyphenylglycol sulfate, synephrine sulfate (main metabolite with 47%), and synephrine glucuronide as metabolites of synephrine.^[21] In the present study, the urine samples were not treated with sulfatase and synephrine sulfate was therefore directly analyzed. As no reference substance was available for synephrine sulfate and a sulfatase treatment led to a major decrease of the flavonone concentrations in the urine samples, only a semi-quantitative approach was applicable. In good agreement with studies reported by Bader et al. and Kusu et al.,^{[17, 28]} a fast increase and subsequent decrease of the urinary synephrine excretion was observed after orange juice intake for all subjects in the present study. In contrast to the two previous studies, synephrine sulfate was still detectable up to 34 h after consumption. Furthermore, the intake of a higher orange juice amount resulted in a longer detection period regarding synephrine sulfate in the urine samples collected after consumption.

4.5 Phloretin

Phloretin is described as a specific biomarker for apple consumption.^[1] Nevertheless, Ito et al. found a potential link between the consumption of citrus juice and the excretion of phloretin as well, and proposed phloretin to be a metabolite of naringenin.^[8] In 2017, Liu et al. detected phloretin after incubation of dried young fruits of Citrus aurantium L. (Fructus aurantii immaturus) with a human fecal microbiota. Based on these results, they postulated a reductive cleavage of the C-ring of naringenin during its metabolism yielding phloretin as metabolite.^[29] In the present study, a link between urinary phloretin excretion and prior orange juice consumption was evident for some subjects. This supports the assumption that phloretin represents a metabolite of naringenin, as described by Liu et al.^[29] However, no statistically significant correlation between the consumption of orange juice and the urinary phloretin excretion was given, and not all subjects excreted phloretin after orange juice intake. The latter might be due to the fact that the intestinal metabolism varies widely between individuals, leading to high individual differences in the metabolic profiles.^[30]

4.6 General Remarks

As there are several influencing factors, such as diet, absorption, excretion, and metabolism of humans, that can be the cause of high interindividual variations, one single biomarker might not be sufficient to draw conclusions concerning the recent food intake of an individual. In contrast, a multiple-biomarker approach as described by Lundblad and by Garcia-Aloy et al. can be a useful alternative for future studies.^{[31, 32]} Such an approach relies on the analysis of several biomarkers instead of one, and the intake of the respective food item is assessed by taking the excretion of all these biomarkers into account. In this manner, individual variations as well as variations caused by the composition of food are reduced and more reliable results are thus obtained.^{[31, 32]} For future studies we recommend the collection of spot urine samples for a period of at least 24 h or morning urine samples for several days. Especially the time-course gives much more information compared to 24 h urine samples and allows an estimation of the kinetic of excretion. Collecting morning urine samples for several days could also provide information on the suitability of analytes as long-term biomarkers.