Abstract

   Pigmented rice is widely popular because of its abundant phytochemicals
   and various health benefits. This study investigated the variations in
   the carotenoids and flavonoids of Yangxian-pigmented rice (black,
   green, purple, yellow, and red) through LC‒MS/MS targeted metabolomics.
   A total of 20 carotenoids were identified, with 18 xanthophylls (90 %)
   and 2 carotenes (10 %). Principal component analysis (PCA) demonstrated
   that PC1 and PC2 accounted for 55.4 % and 26.4 %, respectively, of the
   variance among pigmented rice samples. PLS-DA revealed 5 significantly
   different carotenoids (VIP >1 and P value < 0.05) across these rice
   varieties. Additionally, 378 metabolites, consisting of 361 flavonoids
   (95.5 %) and 17 tannic acids (4.5 %), were detected. PCA revealed that
   PC1 and PC2 contributed 52.5 % and 24.8 % of the variance,
   respectively, and PLS-DA identified 147 significantly differential
   flavonoids (VIP >1 and P < 0.05) among the samples. Multivariate
   statistical analysis revealed variation within the groups and
   separation among the groups, suggesting relatively significant
   differences in carotenoids and flavonoids among pigmented rice. KEGG
   pathway analysis mapped 5 carotenoid metabolites to three metabolic
   pathways (carotenoid biosynthesis, metabolic pathways, and secondary
   metabolite biosynthesis) and 22 flavonoid metabolites to five pathways
   (secondary metabolite biosynthesis, metabolic pathways, flavonoid
   biosynthesis, flavonoid and flavonol biosynthesis, and anthocyanin
   biosynthesis). These findings can support research on bioactive
   compounds and functional food processing in pigmented rice.

   Keywords: Pigmented rice, Metabolites, Carotenoids, Flavonoids, KEGG

Graphical abstract

   Image 1
   [33]Open in a new tab

Highlights

     * •
       The carotenoids and flavonoids of Yangxian pigmented rice were
       characterized.
     * •
       A total of 20 carotenoids and 378 flavonoid metabolites were
       identified.
     * •
       Xanthophylls, lutein, and α-carotene are the prevailing carotenoid
       components.
     * •
       147 differential flavonoids and 5 differential carotenoids were
       screened.

1. Introduction

   Pigmented rice, a healthy grain, is preferred by consumers because of
   its many phytonutrients, anthocyanins, phenolic compounds, and fatty
   acids, which promote physical health ([34]Qi et al., 2019; [35]Samyor
   et al., 2017; [36]Seo et al., 2011). In recent years, pigmented rice
   has attracted increasing attention because of its abundance of
   bioactive compounds. Compared with white rice, pigmented rice is rich
   in nutritional components and contains a variety of phytochemicals,
   such as carotenoids, flavonoids, and anthocyanins ([37]Zhao et al.,
   2020), which possess anti-inflammatory, antioxidant, anticancer, and
   skin-whitening activities. Yangxian County (Hanzhong, China) is known
   for its locally pigmented rice ([38]Jin et al., 2023).
   Yangxian-pigmented rice refers mostly to rice of five colors: yellow,
   green, purple, black, and red. They are loaded with multivitamins,
   phytochemicals, dietary fiber, and trace elements with functional
   nutrients, such as carotenoids and flavonoids ([39]Jin et al., 2023;
   [40]Qi et al., 2019; [41]Cheng et al., 2025).

   Carotenoids, as members of the tetraterpene family (C40 isoprenoid
   compounds), are among the most important pigments in plants; are
   responsible for the yellow, orange or red color of rice seeds, fruits
   and flowers; and include several oxygen-free carotenes as well as
   oxygen-containing xanthophylls ([42]Zhang et al., 2023b). Flavonoids
   represent a large group of polyphenolic secondary metabolites widely
   distributed in plant tissues and can be classified into six major
   groups: flavones, flavanones, isoflavones, anthocyanins, flavanols, and
   flavonols ([43]Zhang et al., 2021). In Korean pigmented rice,
   carotenoid contents clearly differ with grain color, and the levels of
   quercetin, β-carotene and lutein are significantly greater in black
   rice than in red and white rice ([44]Kim et al., 2010). Moreover,
   previous studies revealed that black rice contains significantly higher
   levels of cyanidin-3-glucoside, peonidin-3-glucoside and quercetin than
   red rice is rich in catechin and epicatechin, with cyanidin-3-glucoside
   being the predominant anthocyanin in red rice, as determined via UPLC
   and metabolome analyses ([45]Chen et al., 2022), and that the ternary
   complex MYB-bHLH-WD40 and its constituent transcription factors are
   involved in regulating the biosynthesis of anthocyanins and
   proanthocyanidins in pigmented rice ([46]Mackon et al., 2021).

   Metabolomics, as an emerging discipline, has developed following omics
   studies on genes, transcription and proteins and has developed into an
   important field of systems biology ([47]Chen et al., 2020).
   Metabolomics can conduct an entire analysis of the changes in
   metabolites contained in a specific component of an organism, aiming to
   reveal the dynamic change patterns of this metabolome under external
   intervention or physiological conditions of diseases. Metabolomics
   research can be divided into two major categories: targeted
   metabolomics and nontargeted metabolomics ([48]Chen et al., 2023).
   Nontargeted metabolomics can be applied to detect different metabolites
   in samples and discover potential biomarkers, thereby enabling the
   study of disease mechanisms or drug efficacy and prediction after
   intervention ([49]Ribbenstedt et al., 2018). Previous research has
   revealed the links between metabolite patterns and seed geographic
   origin via the analysis of 121 metabolites in japonica and indica rice
   seeds ([50]Hu et al., 2014). A study involving nontargeted metabolomic
   analysis revealed that the glycolysis pathway and tricarboxylic acid
   cycle were significantly increased during the yellowing of stored rice
   ([51]Liu et al., 2020). Nevertheless, targeted metabolomics makes use
   of the multiple-reaction monitoring (MRM) method to identify predefined
   metabolites and enables absolute qualitative and quantitative analysis
   of the measured compounds ([52]Cao et al., 2020). In a study by
   [53]Zhang et al. (2023a), 732 metabolites were identified through
   metabolomic profiling of black, red, glutinous, and common white rice,
   and 265 differentially abundant metabolites were screened among red and
   black rice groups through a widely targeted metabolomics-based
   approach.

   Liquid chromatography–mass spectrometry (LC‒MS) is an advanced
   analytical technique that separates analytes and detects them on the
   basis of their mass, offering high sensitivity and a wide detection
   range ([54]De Vos et al., 2007; [55]Chen et al., 2013). It has been
   widely applied in targeted, untargeted, pseudotargeted, and widely
   targeted metabolomics for analyzing complex samples ([56]Yang et al.,
   2024). LC‒MS/MS-based targeted metabolomics has become the most
   commonly used method for the separation and identification of
   metabolites; for example, [57]Mughal et al. (2025) analyzed the
   contents and variation patterns of isoflavones in soybean seeds and
   pods via LC‒MS/MS-based targeted metabolomics. In addition,
   LC‒MS/MS-based targeted metabolomics has also been applied to the
   metabolomic analysis of crops such as avocado seeds ([58]Younis et al.,
   2022) and Kinnow mandarin ([59]Saini et al., 2022). In summary, these
   studies suggest that LC‒MS/MS-targeted metabolomics provides a highly
   sensitive and accurate approach for identifying and quantifying
   specific metabolites.

   As carotenoids and flavonoids are vital phytochemicals in pigmented
   rice, it is possible to characterize them through targeted
   metabolomics. To increase the added value of Yangxian-pigmented rice,
   our previous studies performed quantitative analysis of anthocyanin
   ([60]Han et al., 2021) and cyanidin-3-glucoside ([61]Zheng et al.,
   2020) in black rice and hypoglycemic and antiosteoporosis functions in
   diabetic nephropathy, and we also explored the odor compounds in raw
   and processed pigmented rice ([62]Jin et al., 2022, [63]2023; [64]Cheng
   et al., 2025). Nevertheless, the results of metabolomic profiling of
   carotenoids and flavonoids in Yangxian-pigmented rice are still
   unknown.

   In this study, carotenoids and flavonoids in Yangxian-pigmented rice
   varieties were characterized via LC‒MS/MS targeted metabolomics.
   Furthermore, differentially abundant metabolite and KEGG pathway
   analyses have also been implemented to describe differences in
   carotenoids and flavonoids in various colored rice varieties ([65]Gao
   et al., 2013), with the hope of contributing more insights into the
   nutritional traits of Yangxian colored rice.

2. Materials and methods

2.1. Materials and reagents

   The five pigmented rice samples were purchased from Zhoudahei Food Co.,
   Ltd. (Hanzhong, China) in February 2025, and the milling procedure was
   performed according to [66]Ye et al. (2024). Yangxian-pigmented rice
   varieties were farmed and reaped in Yangxian County (Hanzhong, China)
   ([67]Fig. 1A), and the cultivars were No.1 Shuangya Hei (black), No. 1
   Shuangya Lv (green), No. 1 Shuangya Zi (purple), Shuangya Hong (red),
   and Shuangya Huang (yellow) ([68]Jin et al., 2023). These rice
   varieties are predominantly cultivated in flatlands and hilly regions
   at elevations ranging from 450 to 650 m (with a soil pH of
   approximately 7.2 and are rich in organic matter). The whole
   reproductive and growth period was approximately 130–180 days,
   according to the Rice Variety Planting Label of Zhoudahei Food Co.,
   Ltd. (Hanzhong, China). The differences in major nutrients (including
   crude protein, lipids, ash, moisture, and starch) among these pigmented
   rice varieties were reported in our previous publication ([69]Cheng et
   al., 2025).

Fig. 1.

   [70]Fig. 1
   [71]Open in a new tab

   (A) Photographs displaying Yangxian-colored rice. (B) Histogram of
   carotenoid levels in different pigmented rice samples. (C)
   Classification of 378 metabolites. (D) Pie-class plot of overall
   flavonoids in different pigmented rice samples.

   Acetonitrile (ACN), ethanol (EtOH), and methanol (MeOH) of
   high-performance liquid chromatography (HPLC) grade were purchased from
   Merck (Darmstadt, Germany). Sinopharm supplied the acetone, and Aladdin
   supplied the butylated hydroxytoluene (BHT). CNW and Rhawn provided
   sodium chloride (NaCl) and methyl tert-butyl ether (MTBE),
   respectively. The source of potassium hydroxide (KOH) was Hushi.
   Milli-Q water (Millipore, Bradford, USA) was used throughout every
   experiment. The standards were procured from Sigma‒Aldrich (St. Louis,
   MO, USA) and BOC (NY, USA), with formic acid also sourced from
   Sigma‒Aldrich. The standard stock solutions were mixed with MTBE and
   MeOH (1 mg/mL) and stored at −20 °C.

2.2. Targeted metabolomics analysis of carotenoids

2.2.1. Sample preparation and extraction

   A modified procedure from [72]Zheng et al. (2022) was used to analyze
   the carotenoids in pigmented rice, which were subsequently prepared and
   extracted. After being freeze-dried and ground into a powder (30 Hz,
   1.5 min), the samples were stored at −80 °C. A mixture of n-hexane,
   acetone, and ethanol (1:1:1, v/v/v) was used to extract 50 mg of powder
   after it had been weighed. The extract was vortexed for 20 min at
   −25 °C. After centrifugation at 12,000 rpm (5 min) at 4 °C, the upper
   liquids were collected. The residue was re-extracted following the same
   procedure under the same conditions. The mixture was subsequently
   evaporated to dryness and redissolved in 100 μL of dichloromethane. For
   additional LC‒MS/MS analysis, the solution was passed through a 0.22 μm
   membrane filter.

2.2.2. UPLC parameters

   A UPLC-APCI-MS/MS system (UPLC, ExionLCTM AD,
   [73]https://sciex.com.cn/; MS, Applied Biosystems 6500 Triple
   Quadrupole, [74]https://sciex.com.cn) was used to analyze the sample
   extracts. Chromatographic separation was carried out via a UPLC
   chromatographic column, YMC C30 (3 μm, 2.0 mm ∗ 100 mm), maintained at
   28 °C with a flow velocity of 0.8 mL/min. The mobile phase consisted of
   methyl tert-butyl ether (MTBE) with 0.01 % BHT (B) and
   methanol/acetonitrile (1:3, v/v) with 0.01 % BHT and 0.1 % formic acid
   (A). The gradient program began at 0 % B (0–3 min), increased to 70 % B
   (3–5 min), increased to 95 % B (5–9 min), and then increased back to
   0 % B (10–11 min) at 28 °C, 0.8 mL/min, and 2 μL injection.

2.2.3. APCI-MS/MS conditions

   A QTRAP® 6500+ LC‒MS/MS system with an APCI Heated Nebulizer was used
   for the analyses. Both linear ion trap (LIT) and triple quadrupole
   (QQQ) designs were used for positive ion mode scans. Software called
   Analyst 1.6.3 was in charge of the system. The ion source was in
   APCI + mode at 350 °C, and the curtain gas (CUR) pressure was 25.0 psi.
   These were the APCI source characteristics. The predetermined multiple
   reaction monitoring (MRM) method was employed to assess carotenoids,
   and Analyst 1.6.3 software was used to gather the data. Multiquant
   3.0.3 software (Sciex) was used to quantify each metabolite. For every
   MRM transition, the mass spectrometer parameters, including the
   collision energy (CE) and declustering potential (DP), were tuned.
   Specific MRM transitions were monitored for each period, corresponding
   to the metabolites eluted during that time frame.

2.3. Targeted metabolomics assay of flavonoids

2.3.1. Flavonoid extraction procedure

   Flavonoid extraction was performed with slight modifications to the
   approach described by [75]Kim et al. (2010). The rice samples were
   placed in a lyophilizer (Scientz-100F), freeze-dried and then ground
   down into powder form via a grinder (MM 400, Retsch) at 30 Hz for
   1.5 min. A 50 mg portion of the pulverized sample was weighed via an
   electronic balance (MS105DΜ). The powder was then mixed with 1200 μL of
   70 % methanolic aqueous internal standard extract precooled to −20 °C
   (if less than 50 mg of sample was used, the extractant was added at a
   velocity of 1200 μL/50 mg of sample). Six times in total, the mixture
   was vortexed for 30 s every 30 min. After passing through a 0.22 μm
   microporous membrane, the supernatant was transferred to an injection
   vial for UPLC‒MS/MS analysis following 3 min of centrifugation at
   12,000 rpm.

2.3.2. UPLC conditions

   On the basis of the following analysis criteria: UPLC chromatographic
   column, Agilent SB-C18 (1.8 μm, 2.1 mm ∗ 100 mm), the sample extracts
   were analyzed through a UPLC‒ESI‒MS‒MS‒MS system (UPLC, ExionLC™ AD,
   [76]https://sciex.com.cn/) and a tandem mass spectrometry system
   ([77]https://sciex.com.cn/). Solvent A consisted of clean water with
   0.1 % formic acid, and solvent B consisted of acetonitrile with 0.1 %
   formic acid, making up the mobile phase. A gradient program that used
   95 % A and 5 % B as starting conditions for 9 min and then a linear
   gradient to 5 % A and 95 % B was used to measure the samples over the
   next minute. This composition was maintained for 1 min. Afterwards, the
   ingredients were adjusted to 95 % A and 5.0 % B within 1.1 min and
   maintained at a flow velocity of 0.35 mL per min for an additional
   2.9 min. The column oven temperature was set to be constant at 40 °C,
   with an injection volume of 2 μL. The effluent was alternatively linked
   to an ESI-triple quadrupole-linear ion trap (QTRAP)-MS.

2.3.3. ESI-Q trap-MS/MS

   The ESI source was run with the following settings: 500 °C for the
   source temperature; 5500 V for the positive ion mode; and −4500 V for
   the negative ion mode. Curtain gas (CUR) and ion source gases I (GSI)
   and II (GSII) were set at 50, 60, and 25 psi, respectively. It was set
   to high collision activation dissociation (CAD). While the collision
   gas (nitrogen) was adjusted to medium, triple quadrupole (QQQ) scans
   were acquired in MRM experiments. For every MRM transition, the
   collision energy (CE) and declustering potential (DP) were optimized
   separately. At each interval, a particular set of MRM channels that
   correspond to the metabolites that were eluted at that moment should be
   considered.

2.4. Statistical analyses

   The means ± standard deviations (SDs) of three separate runs were used
   to express all of the data. The significance analysis (Tukey's test)
   and correlation analysis were conducted via the SPSS 27.0 software
   package (SPSS, Inc., Chicago, IL, USA). The relative proportions of
   various metabolites were estimated via the peak volume normalization
   method, and Origin 2021 provided a picture of the linked histogram.
   Using SIMCA 14.1, chemometric statistics and analysis were performed.
   Cluster analysis and heatmapping were performed via MetaboAnalyst 6.0
   (MetaboAnalyst, Qu'ebec, Canada) software to visualize the clustering
   of multivariate data.

3. Results and discussion

3.1. Carotenoids and flavonoids in five different pigmented rice varieties

   Certain pigmented rice varieties contain phytochemicals that determine
   their distinctive coloration. These natural pigments can typically be
   classified into several major categories, including chlorophylls,
   carotenoids, and flavonoids ([78]Kim et al., 2010). Carotenoids and
   flavonoids serve as photoprotective agents and antioxidants, providing
   defense against oxidative injury. The present study characterized
   carotenoids and flavonoids in different colored rice samples through
   UPLC‒MS/MS-targeted metabolomics. On the basis of the metabolite
   retention time (RT) and peak shape, the chromatographic peaks of the
   identified metabolites in the samples were corrected to ensure
   qualitative and quantitative verification. The instrumental conditions,
   TIC chromatogram, TIC chromatogram overlap ([79]Fig. S1 and S2) and
   coefficient of variation ([80]Fig. S3) confirmed the high stability of
   the instruments, providing crucial assurance for the reliability and
   repeatability of the current data results for pigmented rice samples.

   A total of 20 carotenoids were identified in the Yangxian pigmented
   rice varieties ([81]Table 1, [82]Fig. 1B), which were classified into
   two categories: carotene and xanthophyll. This finding is in line with
   the results of [83]Nakano and Wiegertjes (2020). Among these,
   xanthophylls were the predominant carotenoids, accounting for
   65.74 %–83.58 % of the carotenoid content. In contrast, carotenes were
   present at much lower concentrations, with only α-carotene and
   β-carotene detected. The results conformed to [84]Lamberts and Delcour
   (2008). Furthermore, 18 types of xanthophylls were identified,
   including lutein dipalmitate, lutein, lutein oleate,
   violaxanthin-myristate-caprate, violaxanthin, neoxanthin, echinenone,
   lutein stearate, zeaxanthin palmitate, canthaxanthin, neochrome
   palmitate, zeaxanthin, violaxanthin dilaurate, lutein distearate,
   violaxanthin myristate, 5,6 epoxide-lutein-caprate-palmitate, lutein
   palmitate, and antheraxanthin dipalmitate. In general, lutein,
   α-carotene, violaxanthin myristate and 5,6
   epoxide-lutein-caprate-palmitate are present at relatively high
   concentrations in pigmented rice. [85]Fratianni et al. (2005) reported
   similar results. [86]Table 1 shows that the content of carotenoids
   varied with grain color. The content of carotenoids in black- and
   green-colored rice was the highest, whereas that in red varieties was
   the lowest. Additionally, [87]Belefant-Miller and Grunden. (2014)
   reported that purple cultivars contained a lower content of carotenoids
   than red cultivars did. These differences may be caused by genetic
   variations among different rice cultivars ([88]Kim et al., 2010).
   [89]Garcia et al. (2023) revealed that photosynthesis is fundamental to
   rice growth and development and affects dry matter accumulation, yield
   formation, and grain quality. [90]Walter & Strack. (2011) reported that
   carotenoids support the regulation of physiological functions and
   promote health in both plants and animals. These findings provide
   useful data for studying the growth and functional ingredients of
   colored rice in rice cereals.

Table 1.

   Carotenoid levels in Yangxian-colored rice (μg/g).
   Number Name CAS RT Yellow Black Purple Green Red
   1 α-carotene 7488-99-5 5.91 2.313 ± 0.115b 1.245 ± 0.044d
   1.464 ± 0.100c 2.530 ± 0.010a 1.556 ± 0.044c
   2 β-carotene 7235-40-7 6.27 0.034 ± 0.003b 0.047 ± 0.003a
   0.035 ± 0.003b 0.000 ± 0.000c 0.000 ± 0.000c
   3 lutein dipalmitate 547-17-1 7.48 0.082 ± 0.004bc 0.085 ± 0.005bc
   0.066 ± 0.002c 0.258 ± 0.016a 0.087 ± 0.008b
   4 lutein 127-40-2 3.93 1.218 ± 0.098c 2.892 ± 0.065a 1.731 ± 0.090b
   0.586 ± 0.022d 0.164 ± 0.006e
   5 lutein oleate – 6.60 0.159 ± 0.018b 0.173 ± 0.020b 0.144 ± 0.006b
   0.218 ± 0.025a 0.149 ± 0.021b
   6 violaxanthin-myristate-caprate – 6.86 0.176 ± 0.002b 0.195 ± 0.009b
   0.217 ± 0.011a 0.195 ± 0.015b 0.192 ± 0.001b
   7 violaxanthin 126-29-4 1.55 0.025 ± 0.002c 0.041 ± 0.000a
   0.030 ± 0.001b 0.017 ± 0.002d 0.000 ± 0.000e
   8 neoxanthin 14660-91-4 1.90 0.043 ± 0.002c 0.113 ± 0.004a
   0.068 ± 0.001b 0.021 ± 0.000d 0.000 ± 0.000e
   9 echinenone 432-68-8 5.53 0.001 ± 0.000c 0.002 ± 0.000a 0.001 ± 0.000b
   0.001 ± 0.000d 0.000 ± 0.000e
   10 lutein stearate – 7.02 0.072 ± 0.019a 0.063 ± 0.004a 0.065 ± 0.002a
   0.00 ± 0.000b 0.053 ± 0.008a
   11 zeaxanthin palmitate – 6.98 0.162 ± 0.015b 0.182 ± 0.022b
   0.182 ± 0.025b 0.292 ± 0.006a 0.162 ± 0.004b
   12 canthaxanthin 514-78-3 4.70 0.000 ± 0.000a 0.000 ± 0.000a
   0.000 ± 0.000a 0.000 ± 0.000a 0.000 ± 0.000a
   13 neochrome palmitate – 6.58 0.041 ± 0.003b 0.000 ± 0.000d
   0.000 ± 0.000d 0.130 ± 0.007a 0.029 ± 0.003c
   14 zeaxanthin 144-68-3 4.59 0.030 ± 0.004a 0.000 ± 0.000c
   0.000 ± 0.000c 0.023 ± 0.002b 0.000 ± 0.000c
   15 violaxanthin dilaurate – 6.67 0.027 ± 0.002b 0.034 ± 0.001a
   0.030 ± 0.003b 0.000 ± 0.000c 0.000 ± 0.000c
   16 lutein distearate – 7.67 0.000 ± 0.000b 0.000 ± 0.000b
   0.000 ± 0.000b 0.025 ± 0.000a 0.000 ± 0.000b
   17 violaxanthin myristate – 6.11 0.959 ± 0.080 ab 1.061 ± 0.031 ab
   1.080 ± 0.098a 1.062 ± 0.093 ab 0.893 ± 0.028b
   18 5,6epoxy-lutein-caprate-palmitate – 6.98 1.300 ± 0.099b
   1.496 ± 0.035b 1.447 ± 0.163b 2.380 ± 0.074a 1.352 ± 0.050b
   19 lutein palmitate – 6.86 0.130 ± 0.016b 0.128 ± 0.009b 0.082 ± 0.011c
   0.265 ± 0.022a 0.109 ± 0.009bc
   20 antheraxanthin dipalmitate – 7.25 0.078 ± 0.007b 0.107 ± 0.004a
   0.090 ± 0.013b 0.036 ± 0.003c 0.019 ± 0.002d
   [91]Open in a new tab

   The carotenoid with the highest content was lutein in black-colored
   rice, with values of up to 2.892 μg/g ([92]Table 1). Lutein, a natural
   plant pigment with multiple biological activities, has beneficial
   effects on vision problems, inflammatory reactions and neuroprotection
   ([93]Mitra et al., 2021). Furthermore, the content of 5,6
   epoxide-lutein-caprate-palmitate was the second highest, at 1.496 μg/g
   ([94]Table 1), and the contents of α-carotene and violaxanthin
   myristate were slightly lower. Moreover, the content of the vitamin A
   precursor, namely, β-carotene, in black-colored rice was significantly
   greater than that in the other colored rice samples ([95]Melini et al.,
   2019). Although [96]Kim et al. (2010) reported that black-colored rice
   from Korea contained β-carotene contents ranging from 0.026 to
   0.048 μg/g, the content of black-colored rice measured in this study
   ranged from 0.044 to 0.051 μg/g. The differences in content between
   these samples may be due to differences in genetic traits among the
   rice grains or geographical and environmental factors ([97]Zhang et
   al., 2023b). Neochrome palmitate, zeaxanthin and lutein distearate are
   not found in black species. However, [98]Melini & Acquistucci. (2017)
   reported that the zeaxanthin content was 0.02 μg/g in black-colored
   rice. The different experimental results may be due to the lower
   content of zeaxanthin than the detection limit of the instrument or the
   use of different black rice cultivars in the test samples ([99]Chen et
   al., 2022). Accordingly, lutein has considerable potential for
   development as a dietary supplement, contributing to the prevention of
   age-related macular degeneration, ocular inflammation, and other
   vision-related disorders through the attenuation of oxidative stress
   ([100]Feeney et al., 2017). In addition to ocular health, lutein also
   supports cognitive function and may mitigate inflammation-driven
   neurodegenerative processes. Furthermore, lutein and zeaxanthin can be
   utilized in the formulation of beauty-oriented functional foods, as
   oral intake has been reported to improve overall skin tone, exert
   skin-whitening effects, and delay cutaneous aging, which is likely
   attributable to their antioxidant activities. In addition, lutein may
   serve as a functional additive in health products targeted at
   middle-aged and elderly populations to help reduce the risk of
   atherosclerosis ([101]Ahn and Kim, 2021). Therefore, the high lutein
   content of black rice highlights its potential as a valuable resource
   for the development of antioxidant- and anti-inflammatory-oriented
   functional foods, thereby offering promising applications in health
   promotion and disease prevention.

   Among the red-colored rice cultivars, α-carotene had the highest
   carotenoid content, with values of up to 1.556 μg/g. [102]Nagarajan et
   al. (2017) revealed that α-carotene, a provitamin A compound, has
   therapeutic potential and contributes to the inhibition of malignant
   tumor cell proliferation. The content slightly lower is 5,6
   epoxide-lutein-caprate-palmitate and violaxanthin myristate. [103]Radix
   et al. (2014)Radix et al. (2014) reported that red-colored rice samples
   from Indonesia contained β-carotene values of up to 0.0013 μg/g, as did
   [104]Melini & Acquistucci. (2017) reported that red rice cultivars from
   Thailand had zeaxanthin contents of up to 0.01 μg/g and that β-carotene
   and zeaxanthin were not examined in red rice samples in the present
   study. These differences might be caused by differences in red-colored
   rice varieties and genetic factors [105]Radix et al. (2014). Moreover,
   violaxanthin, neoxanthin, echinenone, canthaxanthin, violaxanthin
   dilaurate and lutein distearate were not detected in red cultivar
   samples in this study. The intake of α-carotene can reduce cancer risk
   by scavenging free radicals and inhibiting inflammation, and it can be
   exploited as a dietary supplement and nutraceutical ([106]Nagarajan et
   al., 2017).

   In purple-colored rice cultivars, lutein, α-carotene, and
   5,6-epoxy-lutein-caprate-palmitate were the major carotenoids, and
   their contents were slightly lower than those in black rice. Among
   them, the concentration of lutein was the highest (1.731 μg/g), and
   these findings are in accordance with the results of [107]Pereira-Caro
   et al. (2013). In alignment with the findings of [108]Samyor et al.
   (2017), the kinds and contents of most carotenoids in black rice and
   purple rice were extremely similar in this study. These studies provide
   important evidence for understanding and developing functional
   ingredients from black rice and purple rice.

   In yellow-colored rice cultivars, α-carotene was the predominant
   carotenoid, with the maximum concentration reaching 2.313 μg/g, which
   may be associated with the yellow coloration of the grains
   ([109]Lamberts and Delcour, 2008). The main carotenoids are 5,6
   epoxide-lutein-caprate-palmitate (1.300 μg/g), lutein (1.218 μg/g), and
   violaxanthin myristate (0.960 μg/g). A similar study by [110]Zhu et al.
   (2022) reported that the contents of canthaxanthin and echinenone were
   as low as 0.0001–0.0008 μg/g. This evidence indicates that its
   antioxidant capacity can be harnessed for the development of health
   foods and beverages, thereby providing consumers with additional health
   benefits.

   Among the green-colored rice cultivars, the carotenoid with the highest
   content was α-carotene (2.530 μg/g). As a naturally occurring
   carotenoid pigment, α-carotene has a markedly greater antioxidant
   capacity than β-carotene does and contributes to enhancing immunity and
   preventing age-related ocular degeneration ([111]Ku et al., 2019).
   Moreover, the content of 5,6 epoxide-lutein-caprate-palmitate was
   2.43 μg/g, which was the second highest among the green cultivars.
   Subsequently, they were 5,6-epoxy-lutein-caprate-palmitate (2.380 μg/g)
   and violaxanthin myristate (1.062 μg/g). The color of green rice may be
   associated with the contents of these main lutein compounds ([112]Yang
   et al., 2012). In addition, violaxanthin dilaurate, canthaxanthin,
   lutein stearate and β-carotene were not identified. These findings may
   be attributed to the particularity of genetic factors in green-colored
   rice.

   Overall, there are 16 kinds of carotenoids in black cultivars,
   including 14 xanthophylls and 2 carotenes. The red species contains 12
   kinds of carotenoids, including 11 xanthophylls and 1 carotene. The
   identified types of carotenoids in purple-colored rice are similar to
   those in black-colored rice ([113]Samyor et al., 2017). Moreover,
   yellow-colored rice contains the most carotenoids, with 18 types,
   whereas green-colored rice contains 16 carotenoids ([114]Table 1). In
   conclusion, not all of the substances listed above can be detected in
   the five colored rice varieties, which might be caused by the substance
   content being lower than the instrument detection limit or the absence
   of the substance ([115]Petroni et al., 2017). Additionally, the lutein
   content in colored rice varies according to the color of the grains,
   and genetic differences among different pigmented rice varieties might
   also lead to variations in carotenoid content ([116]Kim et al., 2010).
   Compared with light-colored rice, deep-colored pigmented rice contained
   a relatively high content of lutein, such as the highest lutein content
   (2.892 μg/g) in black-colored rice and the lowest lutein content
   (0.164 μg/g) in red rice, which is consistent with the findings
   reported by [117]Melini and Acquistucci (2017). Similarly, [118]Kim et
   al. (2010) reported that the lutein content in black rice cultivars was
   obviously greater than that in red rice cultivars in several Korean
   pigmented rice samples. Factors such as instrument accuracy and
   experimental error may also lead to slight differences in the amount of
   carotene measured. Nevertheless, other factors, such as the climate
   environment (heat, dryness, precipitation), geographic position of
   growth, reaping and postharvesting handling, and storage conditions,
   might impact the carotenoid content of rice ([119]Del Rio et al.,
   2013). Hence, colored rice represents a promising source of functional
   foods and cosmetic additives. Carotenoids exhibit excellent antioxidant
   properties, which help prevent various chronic diseases, reduce the
   risk of obesity and cardiovascular disorders, and contribute to skin
   whitening and antiaging effects.

   [120]Fig. 1C shows the identification of 378 metabolites, encompassing
   361 flavonoids (95. 5 %) and 17 tannins (4.5 %). Flavonoids can be
   divided into isoflavones, flavonols, flavonoids, anthocyanins,
   dihydroflavonols, dihydroflavonoids, chalcones, and other flavonoids.
   17 kinds of tannic acids have been identified in Yangxian-pigmented
   rice, probably due to proanthocyanidins formed by the polymerization of
   flavanols (such as catechins) in flavonoids, so tannins can be
   classified into the flavonoid metabolome ([121]Goufo and Trindade,
   2014). Furthermore, there were notable differences (P < 0.05) in the
   flavonoid content among the different color rice varieties. The total
   flavonoid contents of black-colored, purple-colored, yellow-colored,
   green-colored, and red-colored rice sequentially decreased in this
   order, with black-colored rice having the highest content,
   approximately eight times greater than that of red-colored rice
   ([122]Fig. 1D). The greater the degree of pigmentation is, the greater
   the flavonoid content, which is consistent with the conclusions of
   [123]Yu et al. (2021). As shown in [124]Fig. 3A, flavones presented the
   highest relative abundance in Yangxian pigmented rice varieties,
   followed by flavonols and anthocyanins, with chalcones being the least
   abundant. Among the flavonoids, 167 flavones, 98 flavonols, 36
   flavanones, 19 flavanols, 16 chalcones, 15 other flavonoids and 10
   flavanonols were present in Yangxian whole grains. Moreover, 14 kinds
   of anthocyanins among the 361 flavonoids were identified in
   Yangxian-colored rice. However, [125]Zhang et al. (2023a) reported that
   50 flavonoids, 22 flavonols, 19 flavonoid carbonosides, 12
   anthocyanins, 5 flavanols and 3 chalcones were present in black and red
   grains. The variations in detection results may be attributable to
   genetic material differences among pigmented rice varieties and
   geographical environmental variations, among other factors ([126]Del
   Rio et al., 2013). The results demonstrated that black-colored and
   purple-colored rice presented a high relative abundance of
   anthocyanins, whereas red-colored and green-colored rice presented a
   very low relative abundance ([127]Fig. 3A). [128]Zheng et al. (2022)
   revealed that black-colored rice has a high content of anthocyanins, a
   class of water-soluble pigments that are part of the flavonoid family,
   which exhibit antioxidant properties and could help decrease the risk
   of cardiovascular diseases and certain cancers to some extent.
   Anthocyanins impart purple to blue pigmentation to these colored
   kernels ([129]Zhang et al., 2023c). [130]Ghasemzadeh et al. (2018)
   reported that anthocyanins are absent in most red and white rice
   grains; however, low levels may be present in several red and brown
   rice grains. These findings are in agreement with those of this study.
   As shown in [131]Fig. 3A, the relative abundance distributions of
   flavonoids (including flavonols, flavanones, flavanonols, flavanols,
   chalcones, anthocyanins and other flavonoids) in purple-colored and
   black-colored rice were clearly similar. Our study revealed that
   carotenoids presented highly similar compositions and concentration
   profiles in both purple-colored and black-colored rice, with slightly
   higher levels observed in black-colored rice. Similarly, an identical
   pattern was demonstrated for flavonoid metabolites between these two
   pigmented rice varieties. [132]Cheng et al. (2025) reported that
   volatile metabolites of different abundances in black and
   purple-colored rice were highly similar. Additionally,
   cyanidin-3-O-rutinoside, cyanidin-3-O-(2″-O-glucosyl) glucoside,
   cyanidin 3-O-glucoside and peonidin 3-O-rutinoside were the main
   anthocyanins in the black-colored rice samples ([133]Fig. 3B). These
   main anthocyanin varieties detected in purple cultivars are similar to
   those detected in purple cultivars, and their contents differ.
   [134]Liang et al. (2023) reported that black-colored rice contains high
   concentrations of cyanidin-3-O-rutinoside, cyanidin-3-O-glucoside, and
   peonidin 3-O-rutinoside. These findings provide a foundation for
   exploring the expression of genetic traits and contribute to the
   development of functional foods in pigmented rice. In particular,
   anthocyanins in purple rice have great potential as health-promoting
   foods for the prevention of heart and cardiovascular diseases, as well
   as for the development of anticancer agents ([135]Zheng et al., 2022).

Fig. 3.

   [136]Fig. 3
   [137]Open in a new tab

   Flavonoid metabolomic profiles of Yangxian-pigmented rice varieties.
   (A) Stacked bar chart of the hierarchical distribution of flavonoid
   subclasses in pigmented rice. (B) Stacked bar chart of the anthocyanins
   in pigmented rice samples. (C) PCA score plot. (D) PLS-DA score plot.
   (E) Cross-validation by a permutation test.

3.2. Multivariate statistical analysis of carotenoids and flavonoids in
colored rice

   Multivariate statistical analysis of five pigmented rice samples was
   carried out to determine the differences in their metabolites. PCA,
   PLS-DA, and clustering models were used to investigate the metabolite
   profiles of pigmented and nonpigmented rice ([138]Zhang et al., 2023b)
   and the differences in volatile flavor compounds among distinct colored
   millet cultivars ([139]Jin et al., 2023).

   In this study, PCA was also used to analyze the detected metabolites
   and assess the general differences among the samples. PCA can reflect
   the degree of overall metabolic differences among groups and the extent
   of variation within group samples. To explore the metabolic variations
   among pigmented rice cultivars, the 20 kinds of carotenoids detected
   were studied through principal component analysis (PCA), with PC1
   accounting for 55.4 % of the variance and PC2 accounting for 26.4 %,
   together totaling 81.8 %. These findings indicate that the two
   principal components revealed key characteristics of the rice grain
   samples ([140]Fig. 2A). As shown in [141]Fig. 2A, the carotenoids
   exhibited closer clustering in pigmented rice of the same color,
   including certain overlaps, and the different colors were separated.
   These findings indicate that the intergroup differences are
   significant, whereas the intragroup similarity is high in pigmented
   rice samples. The heatmap clustering results of 20 carotenoids in five
   different colored rice cultivars are shown in [142]Fig. 2B. As shown,
   significant similarities in both the content and composition of
   carotenoids were observed between black- and purple-colored rice, both
   of which contain high levels of lutein and α-carotene. Similarly,
   black-colored rice contains relatively higher levels of lutein,
   neoxanthin, lutein palmitate, etc., than purple-colored rice. These
   findings may be due to the high levels of carotenoid intermediates
   present in the grains of black- and purple-colored rice ([143]Frei and
   Becker, 2004) and the relatively high expression of genes involved in
   pyruvate metabolism ([144]Chettry et al., 2019). Most of the
   carotenoids in red-colored and green-colored rice differ greatly, and
   green rice contains relatively high contents of carotenoids ([145]Fig.
   2B–[146]Table 1), including lutein dipalmitate, lutein oleate, and
   zeaxanthin palmitate. Moreover, the characteristic carotenoids in
   yellow-colored rice are zeaxanthin and lutein stearate, which are
   remarkably disparate from those in other pigmented rice. Genetic
   variations among different rice cultivars may account for these
   differences in carotenoids ([147]Kim et al., 2010), which influence the
   pathway differences in carotenoid biosynthesis. Moreover, these
   distinctions lead to significant variation in the levels of these
   carotenoid intermediates among cultivars ([148]Chettry et al., 2019).
   [149]Kim et al. (2021) employed an HPLC‒MS approach to analyze
   carotenoids for the classification of red, black and nonpigmented rice
   samples and to explore the correlation of rice varieties and
   metabolites. This study revealed that carotenoids could also be
   utilized for the general classification of metabolic profiles among
   pigmented rice varieties. In general, the five colored rice samples
   could be classified into four types, namely, black and purple-colored
   rice (with relatively high similarity), red-colored rice, green-colored
   rice, and yellow-colored rice, which was in agreement with the PCA
   score graphic ([150]Samyor et al., 2017). This phenomenon may be
   attributed to the high genetic similarity between black rice and purple
   rice ([151]Jin et al., 2022), resulting in only minor differences in
   the carotenoid biosynthetic pathway. Thus, the five varieties of
   colored rice could be well differentiated via PCA and cluster analysis
   of the carotenoids. Similarly, as the Venn diagram shows, there are a
   total of 20 carotenoid metabolites, and 10 of these carotenoid
   compounds, including α-carotene, lutein, and violaxanthin myristate,
   are universally present among the pigmented rice varieties ([152]Fig.
   2D). Moreover, the color of the carotenoids in the cluster heatmap
   ranged from blue to red, which represents the individual material
   composition from low to high and can also better indicate the
   differences in the contents and diversity of the carotenoid compounds
   among the five pigmented rice cultivars. The PLS-DA score plot can
   usually be applied to visualize the classification function of a model
   and establish an interaction simulation between change metrics and
   specimen classification ([153]Wang et al., 2019). A PLS-DA model was
   constructed to accurately identify the distinctive carotenoids among
   Yangxian-colored rice cultivars. The values of R^2Y and Q^2 approaching
   1 (R^2X = 0.994, R^2Y = 0.98, and Q^2 = 0.919) validate the reliability
   and stability of the PLS-DA model in identifying differentially
   abundant metabolites among specimen groups. Clearly, the score graph of
   PLS-DA could markedly separate Yangxian-colored rice, confirming the
   differences in the carotenoid metabolites of pigmented rice varieties
   ([154]Fig. 2C), which is consistent with the results of the PCA profile
   analysis ([155]Fig. 2A). Then, following 200 permutation tests, the
   intercept is negative (−0.557), with the restoration curve of analog
   Q^2 crossing the horizontal axis, and the original value is always
   higher than the R^2 and Q^2 values randomly generated by all
   permutations ([156]Fig. 2E), which can avoid overfitting and ensure the
   reliability of PLS-DA. The permutation test analysis of the PLS-DA
   model demonstrated statistically significant validity (P < 0.05),
   enabling the identification of differentially expressed metabolites on
   the basis of their variable importance in projection (VIP) scores.

Fig. 2.

   [157]Fig. 2
   [158]Open in a new tab

   Carotenoid metabolomic profiles of Yangxian-colored rice varieties. (A)
   PCA score plot, (B) clustering heatmap, (C) PLS-DA score plot, (D) Venn
   diagram of group comparisons, (E) cross-validation by a permutation
   test, (F) VIP value distribution of carotenoids.

   Given that VIP >1.0 and a significance level of P < 0.05, 5 carotenoids
   contribute to the isolation of five pigmented rice cultivars, including
   violaxanthin-myristate-caprate, zeaxanthin, α-carotene, violaxanthin
   myristate, and lutein oleate ([159]Fig. 2F). LC‒MS-targeted amino acid
   profiling was used to explore the differences in nutrient components in
   different colored rice varieties in the study of [160]Zhu et al.
   (2024). Research has indicated that the 20 carotenoids detected can be
   used for sorting and analyzing the metabolic profiles of colored rice
   cultivars.

   With respect to the flavonoid metabolites in pigmented rice,
   significant variations were observed among the different varieties. In
   the PCA score graphic, the two principal components collectively
   explained 76.9 % of the total variance, whereas PC1 and PC2 contributed
   49.2 % and 27.7 % of the variance, respectively ([161]Fig. 3C).
   Moreover, the first principal component (PC1) and the second principal
   component (PC2) demonstrated strong cohesion within the groups and
   clear separation among the five colored rice varieties. The
   distributions of purple rice and black rice are clearly more closely
   distributed than those of other colored rice. This shows that there is
   little difference in flavonoids between black rice and purple rice in
   Yangxian County but that there is an obvious difference among the
   remaining colors. The results demonstrated that genetic traits strongly
   influenced flavonoids in pigmented rice varieties because of the
   significant intergroup variation coupled with marked intragroup
   homogeneity, which was caused mainly by the differences in their
   anthocyanin profiles among different colored rice samples ([162]Yu et
   al., 2021). Previous studies have indicated that different metabolic
   profiles may be related to genetic factors, climatic conditions,
   nutrient availability, and variations in geographical environments
   ([163]Wang et al., 2023; [164]Fu et al., 2019; [165]Kim et al., 2010).
   In the PLS-DA model, the metabolites detected in Yangxian-colored rice
   varieties were fitted and simulated, with R^2X(cum) = 0.988,
   R^2Y(cum) = 0.996, and Q^2 (cum) = 0.962. Analysis and comparisons
   among 378 flavonoids of the five pigmented rice cultivars via the
   PLS-DA model revealed that, with the exception of some overlap between
   Yangxian purple- and black-colored rice, other different colored rice
   cultivars were clearly separated from each other ([166]Fig. 3D), which
   was consistent with the conclusions of the above PCA ([167]Fig. 3C).
   There was a significant similarity of flavonoids in Yangxian black- and
   purple-colored rice. To verify the PLS-DA mode, following 200
   permutation tests, the restoration curve derived from modeling Q^2
   exhibited a statistically significant negative intercept (−0.536),
   whereas the permutation test results demonstrated that all randomly
   generated R^2 and Q^2 values were consistently and significantly lower
   than the original model's predictive performance metrics, thereby
   validating the model's robustness against overfitting ([168]Fig. 3E).
   The results indicated that the PLS-DA model is adequately predictable
   and is not overfit ([169]Zhu et al., 2024). To identify the
   significantly different flavonoid metabolites in the five pigmented
   rice varieties of Yangxian, we used the VIP method, which is based on
   the PLS-DA model, to preliminarily determine the metabolite differences
   between different species and used the P value to screen out the
   different metabolites in greater depth.

   According to the PLS-DA model results, the differential flavonoids
   among the groups were screened and determined via the following
   standards: variable importance in projection (VIP) ≥ 1 and P < 0.05
   ([170]Yu et al., 2021). A total of 147 significantly different
   flavonoids in pigmented rice were screened via this model analysis. The
   top fifty differentially abundant metabolites were recorded and are
   summarized (with decreasing VIP values) in [171]Table 2. As shown,
   delphinidin-3-O-galactoside, eschweilenol c∗,
   apigenin-7-O-(6″-feruloyl)glucuronide,
   5,7,3′,4′-tetrahydroxy-6-methoxyflavone-8-c-[glucosy-(1–2)]-glucoside,
   and 7-methoxy-3-[1-(3-pyridyl)methylidene]-4-chromanone were the five
   substances with the highest VIP values among the 147 flavonoids.
   [172]Tiozon et al. (2023) reported that anthocyanins are susceptible to
   light-induced degradation. For example, delphinidin-3-O-galactoside, a
   water-soluble anthocyanin, was markedly reduced during germination,
   likely due to its diffusion into the soaking water. Research by
   [173]Zhao et al. (2020) revealed that copigmentation with flavonols,
   glycosides, and phenolic acids significantly modified the chromatic
   properties of peonidin-3-O-glucoside, delphinidin-3-O-glucoside, and
   malvidin-3-O-glucoside.

Table 2.

   Several flavonoid profiles with significant differences (among
   Yangxian-colored rice varieties).
   Number CAS Differential Metabolites Class VIP P-value FDR
   01 28500-00-7 Delphinidin-3-O-galactoside Anthocyanidins 1.998
   2.406E-10 6.317E-10
   02 211371-02-7 Eschweilenol C∗ Tannin 1.924 1.521E-07 2.602E-07
   03 – Apigenin-7-O-(6″-feruloyl)glucuronide Flavones 1.884 6.330E-13
   3.988E-12
   04 –
   5,7,3′,4′-Tetrahydroxy-6-methoxyflavone-8-C-[glucosyl-(1–2)]-glucoside
   Flavones 1.850 7.010E-10 1.677E-09
   05 – 7-Methoxy-3-[1-(3-pyridyl)methylidene]-4-chromanone Other
   Flavonoids 1.797 1.581E-10 4.461E-10
   06 – Ginnalin A (2,6-Di-O-Galloyl-1,5-Anhydro-D-Glucitol) Tannin 1.765
   1.227E-08 2.392E-08
   07 – 4′,5-Dihydroxy-3′,5′-dimethoxyflavone Flavones 1.753 8.679E-10
   2.000E-09
   08 1262536-83-3 Fisetinidol-(4α,6)-gallocatechin Flavanols 1.751
   2.984E-11 1.035E-10
   09 176665-78-4 3′-O-methyl-ellagic acid-4-O-beta-D-xylopyranoside∗
   Tannin 1.742 1.544E-06 2.334E-06
   10 20315-25-7 Procyanidin B1 Proanthocyanidins 1.737 1.347E-13
   1.273E-12
   11 970-74-1 Epigallocatechin Flavanols 1.736 3.768E-03 4.342E-03
   12 – Quercetin-3-O-(6″-O-acetyl)glucoside Flavonols 1.724 1.240E-07
   2.150E-07
   13 79763-28-3 Arecatannin B1 Tannin 1.719 1.387E-09 3.030E-09
   14 20310-89-8 Isovitexin-7-O-glucoside(Saponarin) Flavones 1.717
   7.343E-09 1.461E-08
   15 25694-72-8 Luteolin-7-O-neohesperidoside (Lonicerin)∗ Flavones 1.714
   7.475E-06 1.024E-05
   16 13241-32-2 Neoeriocitrin Flavones 1.712 3.916E-07 6.448E-07
   17 32769-01-0 Tricin-7-O-Glucoside∗ Flavones 1.684 4.109E-05 5.265E-05
   18 32602-81-6 Kaempferol-3-O-neohesperidoside∗ Flavonols 1.683
   2.981E-07 4.942E-07
   19 25474-11-7 Jaceosidin-7-O-Glucoside∗ Flavones 1.672 5.809E-07
   9.264E-07
   20 22888-70-6 Silibinin∗ Flavanonols 1.668 1.435E-10 4.141E-10
   21 37064-31-6 Procyanidin C2 Proanthocyanidins 1.667 1.815E-10
   4.917E-10
   22 23567-23-9 Procyanidin B3 Proanthocyanidins 1.648 5.296E-13
   3.452E-12
   23 18003-33-3 6-Hydroxyluteolin Flavones 1.647 4.859E-10 1.193E-09
   24 27215-04-9 Quercetin-3-O-(4″-glucosyl)glucoside; Meratin∗ Flavonols
   1.642 1.519E-06 2.307E-06
   25 29106-49-8 Procyanidin B2 Proanthocyanidins 1.638 6.514E-16
   3.517E-14
   26 72581-71-6 Isosilybin∗ Flavanones 1.637 4.351E-12 2.006E-11
   27 15648-86-9 Myricetin-3-O-galactoside∗ Flavonols 1.632 5.686E-05
   7.164E-05
   28 3563-98-2 Luteolin-7-O-rutinoside∗ Flavones 1.631 1.383E-05
   1.854E-05
   29 552-58-9 Eriodictyol (5,7,3′,4′-Tetrahydroxyflavanone) Flavanones
   1.621 3.125E-13 2.148E-12
   30 – Peonidin-3-O-(6″-O-caffeoyl)glucoside Anthocyanidins 1.619
   3.185E-11 1.092E-10
   31 – 1,3,6-trihydroxy-2,5,7-trimethoxyxanthen-9-one Other Flavonoids
   1.606 2.460E-07 4.114E-07
   32 – Gossypetin-3-O-glucoside∗ Flavonols 1.603 5.536E-03 6.284E-03
   33 527-95-7 Herbacetin Flavonols 1.595 8.199E-10 1.901E-09
   34 28610-31-3 8-Prenylkaempferol Flavonols 1.591 1.218E-17 4.603E-15
   35 15486-33-6 3,5-Dihydroxy-7,4′-dimethoxyflavone Flavonols 1.587
   5.921E-09 1.184E-08
   36 – Isorhamnetin-3-O-(2″-O-xylosyl)glucoside Flavonols 1.586 1.226E-12
   7.240E-12
   37 – Andrographidine D aglycone Flavanones 1.578 5.472E-14 6.895E-13
   38 37744-61-9 Iristectorin A∗ Isoflavones 1.572 2.326E-07 3.925E-07
   39 10236-47-2 Naringenin-7-O-Neohesperidoside(Naringin)∗ Flavanones
   1.566 4.540E-06 6.427E-06
   40 5084-19-5 5-Hydroxy-3,3′,4′,5′,7-Pentamethoxyflavone∗ Flavones 1.558
   1.547E-10 4.397E-10
   41 – 5-Hydroxy-3,7,8,3′,4′-Pentamethoxyflavone∗ Flavones 1.555
   2.589E-11 9.146E-11
   42 489-35-0 Gossypetin(3,3′,4′,5,7,8-Hexahydroxyflavone) Flavonols
   1.553 1.207E-10 3.621E-10
   43 51803-68-0 3,3′-Di-O-Methylellagic acid 4′-glucoside Tannin 1.552
   8.749E-13 5.422E-12
   44 2174-59-6 5-Demethylnobiletin;
   5-Hydroxy-6,7,8,3′,4′-Pentamethoxyflavone∗ Flavones 1.544 1.788E-10
   4.897E-10
   45 – 5,3′-dihydroxy-6,7,4′-trimethoxyflavone-8-O-β-D-glucoside∗
   Flavones 1.542 1.036E-07 1.822E-07
   46 89915-54-8 Tricin-7-O-(2″-O-rhamnosyl)galacturonide Flavones 1.538
   4.686E-10 1.158E-09
   47 – Orientin-6-C-arabinoside Flavones 1.525 1.091E-11 4.225E-11
   48 – Tricin-7-O-rutinoside Flavones 1.525 6.444E-08 1.160E-07
   49 3301-49-3 5,4′-Dihydroxy-3,7-dimethoxyflavone(Kumatakenin)∗
   Flavonols 1.525 1.057E-11 4.207E-11
   50 86579-00-2 Carthamone∗ Chalcones 1.512 5.356E-07 8.579E-07
   [174]Open in a new tab

   The heatmap illustrated the clustering results of 147 flavonoids with
   significantly different abundances across five different pigmented rice
   varieties ([175]Fig. 5), revealing more pronounced differences, and
   four major clusters were identified on the basis of the variations in
   the relative contents of the metabolites. The cluster 1 metabolites
   were the most abundant in yellow-colored rice, whereas the cluster 2
   metabolites were the most abundant in red-colored rice, the cluster 3
   metabolites were the most abundant in green-colored rice, and the
   cluster 4 metabolites were the most abundant in purple-colored and
   black-colored rice varieties. Notably, black rice presented
   significantly higher levels of differentially abundant metabolites than
   purple rice did in cluster 4 ([176]Fig. 5). Moreover, the flavonoid
   profiles of black-colored rice are more similar to those of
   purple-colored rice, and the metabolites with different abundances
   identified in red- and yellow-colored rice are strikingly highly
   similar. These findings indicate that the pigmented rice cultivars
   presented significantly different metabolic profiles. Three parallel
   repeated experiments were conducted in every group, revealing small
   intragroup variability and high repeatability between different colored
   rice varieties in Yangxian and obvious separation between the groups,
   which demonstrated that there were obvious differences in flavonoids in
   different colored rice varieties ([177]Zhang et al., 2023a). The
   content and composition of flavonoids in rice are closely associated
   with coloration, with considerable variation among different cultivars.
   In particular, the darker the color of rice is, the greater the
   flavonoid content. These differentially expressed flavonoids are
   related mainly to anthocyanin biosynthesis ([178]Yu et al., 2021).

Fig. 5.

   [179]Fig. 5
   [180]Open in a new tab

   Clustering heatmap of 147 significantly differential flavonoids in
   Yangxian-pigmented rice varieties.

   [181]Fig. 6C presents the results of the Pearson correlation analysis
   of the top 50 differentially abundant metabolites (based on VIP values)
   among the 147 significantly different flavonoids identified in Yangxian
   five pigmented rice. A positive correlation was detected between the
   contents of flavones and anthocyanins among the flavonoids. [182]Kim et
   al. (2010) reported parallel results. Similarly, the five most
   significant positive correlations (r > 0.999) were observed between
   procyanidin B1 and procyanidin B3 (r = 0.9995), cyanidin-3-O-rutinoside
   and naringenin-4′-O-glucoside∗ (r = 0.9994),
   cyanidin-3-O-(2″-O-glucosyl)glucoside and cyanidin-3-O-rutinoside
   (r = 0.9993), procyanidin B2 and procyanidin B3 (r = 0.9993), and
   procyanidin B1 and procyanidin B2 (r = 0.9993). The findings of
   [183]Deluc et al. (2006) were consistent with those of this study.
   [184]Kim et al. (2010) reported significantly positive correlations
   between cyanidin-3-O-glucoside and peonidin-3-O-glucoside (r = 0.9885)
   or quercetin (r = 0.9400). [185]Liang et al. (2023) reported that
   cyanidin-3-O-glucoside and cyanidin-3-O-rutinoside were present at
   relatively high concentrations in black-colored rice. As shown in
   [186]Fig. 6C, comparative analysis of the yellow, black, green, purple,
   and red rice varieties revealed that most of the 147 significantly
   differentially abundant flavonoids presented strong correlations
   (|r| > 0.9), with the majority showing positive correlations and only a
   minority demonstrating negative correlations. Additionally, the 16 most
   significant negative correlations involved procyanidin C1, with the top
   five being procyanidin C1 and velutin
   (5,4′-dihydroxy-7,3′-dimethoxyflavone)∗ (r = −0.9935), procyanidin C1
   and acacetin-6-C-glucoside (r = −0.9909), procyanidin C1 and
   eriodictyol (r = −0.9867), procyanidin C1 and
   tricin-7-O-(2″-O-rhamnosyl)galacturonide (r = −0.9861), and procyanidin
   C1 and isoscutellarein∗ (r = −0.9839). [187]Fan et al. (2024) reported
   that the contents of procyanidin B1 and procyanidin C1 were greater
   than those of other anthocyanins and that their levels gradually
   increased during the grain germination process in red rice samples. The
   present study revealed similar results: a positive correlation exists
   between the contents of flavonoids and carotenoids in rice varieties,
   and there might be several types of crosstalk reported by [188]Kim et
   al. (2010).

Fig. 6.

   [189]Fig. 6
   [190]Open in a new tab

   (A) KEGG pathway enrichment analysis of carotenoids in colored rice,
   (B) KEGG pathway enrichment analysis of differential flavonoid
   metabolites in pigmented rice varieties, (C) differentially abundant
   metabolite correlation heatmap of flavonoids.

   According to the PLS-DA model, differential flavonoid metabolites
   between groups were identified via the following criteria: VIP ≥1 and
   fold change (FC) ≥ 2 or ≤ 0.5 ([191]Wang et al., 2023). The relevant
   screening results are displayed with volcano plots ([192]Fig. 4A–J).
   There were 227 significantly different metabolites in black-colored vs
   yellow-colored rice (180 upregulated and 47 downregulated), 245 in
   green-colored vs black-colored rice (29 upregulated and 216
   downregulated), 16 in purple-colored vs black-colored rice (7
   upregulated and 9 downregulated), 219 in purple-colored vs
   green-colored rice (188 upregulated and 31 downregulated), 167 in
   green-colored vs yellow-colored rice (34 upregulated and 133
   downregulated), 198 in purple-colored vs yellow-colored rice (143
   upregulated and 55 downregulated), 270 in red-colored vs black-colored
   rice (24 upregulated and 246 downregulated), 171 in red-colored vs
   green-colored rice (63 upregulated and 108 downregulated), 281 in
   red-colored vs purple-colored rice (26 upregulated and 255
   downregulated), and 194 in red-colored vs yellow-colored rice (18
   upregulated and 176 downregulated). A total of 281 differentially
   abundant metabolites were selected between the red-colored rice and
   purple-colored rice samples. There were 26 metabolites whose expression
   was upregulated (flavones, proanthocyanidins, other flavonoids,
   tannins, etc.) and 255 metabolites whose expression was downregulated
   (flavones, other flavonoids, chalcones, flavonols, and anthocyanidins).
   These results demonstrated that the metabolites significantly differed
   between red-colored and purple-colored rice due to the large number of
   significantly differentially abundant metabolites, which may be due to
   the significant genetic differences between red- and purple-colored
   rice, which in turn lead to differential regulatory expression of genes
   involved in flavonoid metabolic pathways ([193]Wan et al., 2024). These
   flavonoids and tannins could explain the different metabolites detected
   in different colored rice ([194]Zhu et al., 2024). Obviously, black-
   and purple-colored rice are quite similar, as only 16 significantly
   differentially abundant metabolites were screened between black- and
   purple-colored rice. The 7 upregulated metabolites were flavones and
   proanthocyanidins, and the 9 downregulated metabolites were flavones,
   flavanols and tannins. [195]Zhang et al. (2023a) revealed that the
   various metabolites found in different colored rice varieties may be
   explained by phenolic acids and flavonoids. The results also
   demonstrated that flavonoids and tannins could be among the reasons for
   the differences in metabolites among rice samples with different
   colors. The discovery of differentially abundant metabolites
   established a foundational framework for advancing functional genomics
   and nutritional biochemistry research on pigmented brown rice.

Fig. 4.

   [196]Fig. 4
   [197]Open in a new tab

   (A–J) Volcano plots showing the differentially abundant metabolite
   expression levels between the black-colored and yellow-colored rice
   groups (A), the green-colored vs black-colored rice group (B), the
   purple-colored vs black-colored rice group (C), the purple-colored vs
   green-colored rice group (D), the green-colored vs yellow-colored rice
   group (E), the purple-colored vs yellow-colored rice group (F), the
   red-colored vs black-colored rice group (G), the red-colored vs
   green-colored rice group (H), the red-colored vs purple-colored rice
   group (I) and the red-colored vs yellow-colored rice group (J).

3.3. KEGG and enrichment analysis of metabolites with significant differences
in carotenoid and flavonoid contents

   KEGG, as a principal database related to public pathways, was used to
   process information and identify the metabolic pathways relevant to the
   significantly enriched compounds. It involves systematic analysis of
   internal cellular metabolic pathways, which aids in the study of
   complex biological behaviors and their different metabolites ([198]Gao
   et al., 2013). The enrichment of metabolic pathways relevant to the
   notably abundant chemicals found was also examined via the KEGG
   database. Twenty carotenoid metabolites were found and analyzed in this
   study. Eight of the 20 carotenoids were successfully annotated in the
   KEGG database and mapped to five known pathways: biosynthesis of
   secondary metabolites, carotenoid biosynthesis, metabolic pathway,
   biosynthesis of cofactors, and biosynthesis of various plant secondary
   metabolites. Notably, 5 of these 8 differentially abundant metabolites
   detected in the comparative analysis of the five colored rice varieties
   were associated with three specific KEGG pathways ([199]Fig. 6A):
   carotenoid biosynthesis, metabolic pathways, and the biosynthesis of
   secondary metabolites. These five differentially abundant metabolites
   were all xanthophylls, namely, violaxanthin, neoxanthin, lutein,
   canthaxanthin, and echinenone. These two metabolic pathways of
   carotenoid biosynthesis and the biosynthesis of secondary metabolites
   were highly enriched. [200]Kim et al. (2021) reported that the
   biosynthesis of secondary metabolites was highly enriched in black rice
   compared with nonpigmented rice, and secondary metabolic pathways were
   closely related to the biosynthesis of terpenoids (including
   carotenoids) in pigmented rice. In this study, we also reached a
   consistent conclusion: in both the biosynthesis of secondary
   metabolites and the carotenoid pathways, the number and types of
   significantly different metabolites annotated were identical through
   KEGG and enrichment analyses, namely, lutein, neoxanthin, violaxanthin,
   canthaxanthin, and echinenone ([201]Fig. 6A). Research has revealed
   that the carotenoid biosynthetic pathway starts with the synthesis of
   geranylgeranyl pyrophosphate via the MEP pathway ([202]Zhang et al.,
   2023b; [203]Chettry and Chrungoo, 2020). Phytoene synthase (PSY) and
   phytoene desaturase (PDS) are the primary rate-limiting enzymes in the
   carotenoid biosynthetic pathway and play crucial roles in carotenoid
   synthesis ([204]Ruiz-Sola and Rodriguez-Concepcion, 2012). Without
   these enzymes, carotenoids, such as β-carotene, lutein, and zeaxanthin,
   cannot be synthesized. In addition, four carotenoid hydroxylases (CHs),
   categorized into β-ring hydroxylases (BCHs) and cytochrome P450-type
   hydroxylases (CYP97A and CYP97C), catalyze the hydroxylation of
   α-carotene and β-carotene, leading to the formation of lutein and
   zeaxanthin ([205]Zhang et al., 2023b). On the basis of the above
   analysis, the pathways of carotenoid biosynthesis and secondary
   metabolite biosynthesis in pigmented rice strongly influence the types
   and contents of certain differentially abundant carotenoid metabolites,
   thereby contributing to the distinctive coloration of Yangxian
   pigmented rice.

   Moreover, 378 metabolites were analyzed and detected in this study, of
   which 55 metabolites could be annotated by KEGG into 6 existing
   pathways in the database: secondary metabolite synthesis, flavonoid
   biosynthesis, metabolic pathways, flavonoid and flavonol biosynthesis,
   isoflavone biosynthesis, and anthocyanin biosynthesis. In addition, 22
   of 147 differentially abundant flavonoids in the comparison of the five
   colored rice varieties could be annotated to the five metabolic
   pathways of the existing KEGG database, including secondary metabolite
   biosynthesis, metabolic pathways, flavonoid biosynthesis, flavonoid and
   flavonol biosynthesis, and anthocyanin biosynthesis. These findings
   suggest that there are various metabolite profiles among the five
   colored rice varieties and that these differences may be associated
   with grain color. This finding is in agreement with the results of
   [206]Zhang et al. (2023a). As shown in [207]Fig. 6B, among the five
   pigmented rice samples, the only one that was noticeably enhanced was
   anthocyanin biosynthesis (P value < 0.05), and the 5 metabolites with
   the greatest differences in abundance were identified in this pathway,
   namely, cyanidin-3-O-glucoside, cyanidin-3-O-(2″-O-glucosyl)glucoside,
   cyanidin-3-O-rutinoside, pelargonidin-3-O-glucoside, and
   peonidin-3-O-glucoside. Previous studies have demonstrated that
   anthocyanin biosynthesis is regulated by the flavonoid biosynthesis
   pathway, among which the upregulation of ANS gene expression and its
   concomitant enzymatic activation drive increased anthocyanin production
   ([208]Wan et al., 2024). Additionally, the biosynthesis of anthocyanins
   is regulated by a ternary transcriptional complex comprising R2R3-MYB,
   bHLH, and WD40 factors, which differentially activate structural genes
   involved in enzyme production ([209]Mackon et al., 2021). Additionally,
   the anthocyanin biosynthesis pathway, as a branch of the general
   flavonoid biosynthesis pathway, starts with phenylalanine as the
   substrate ([210]Koes et al., 2005). The metabolic pathway of
   anthocyanins begins with a major enzyme called dihydroflavonol
   4-reductase (DFR). Without this enzyme, the synthesis of
   proanthocyanidins, anthocyanins, and pigmented rice cannot proceed
   ([211]Dai et al., 2012). The presence of anthocyanins in rice leads to
   distinctive coloration of the caryopsis. The color variation in rice
   husks is caused by differences in the composition and concentration of
   pigments. Anthocyanins are the cause of color differences ([212]Goufo
   and Trindade, 2014). The findings of this study on the KEGG-enriched
   metabolic pathway of flavonoids are consistent with previously reported
   results, thereby providing further validation and support for the
   conclusions of the present investigation. To date, the anthocyanins
   cyanidin-3-glucoside (C3G) and peonidin-3-glucoside (P3G) have been the
   predominant anthocyanins found in rice ([213]Mackon et al., 2021).
   Notably, anthocyanins are secondary metabolites present in pigmented
   cereals and exhibit diverse bioactivities, including a reduction in
   lipid accumulation, suppression of inflammation, and enhancement of
   antioxidant capacity. The results indicated that Yangxian colored rice
   is a highly promising functional food with potential applications in
   the production of health foods and beverages, thereby offering
   additional health benefits to consumers ([214]Yu et al., 2021).

4. Conclusions

   In summary, Yangxian-colored rice contains 20 carotenoids, with
   xanthophylls, lutein, and α-carotene being the predominant carotenoids.
   Five carotenoids (violaxanthin-myristate-caprate, zeaxanthin,
   α-carotene, violaxanthin myristate, and lutein oleate) significantly
   promoted the classification of five pigmented rice varieties on the
   basis of multivariate statistical analysis. Black rice contains more
   lutein, 5,6 epoxide-lutein-caprate-palmitate, α-carotene and
   violaxanthin myristate. KEGG enrichment analysis revealed that these
   differential carotenoids are related to carotenoid biosynthesis,
   metabolic pathways, and the biosynthesis of secondary metabolites. In
   addition, 378 metabolites, including 361 flavonoids and 17 tannic
   acids, were identified in Yangxian-colored rice. A total of 147
   significantly different flavonoids were screened. The black-colored and
   purple-colored rice varieties presented greater similarity. The 22
   significantly different flavonoids were found to be correlated with
   secondary metabolite biosynthesis, metabolic pathways, flavonoid
   biosynthesis, flavonoid and flavonol biosynthesis, and anthocyanin
   biosynthesis via KEGG enrichment analysis. Among these pathways,
   anthocyanin biosynthesis, which is orchestrated by the flavonoid
   biosynthetic pathway, significantly differed (P < 0.05). The
   cyanidin-3-O-glucoside, cyanidin-3-O-(2″-O-glucosyl)glucoside,
   cyanidin-3-O-rutinoside, pelargonidin-3-O-glucoside, and
   peonidin-3-O-glucoside with the greatest differences in abundance were
   associated with anthocyanin biosynthesis. These results provide
   reference information for understanding the variations in carotenoids
   and flavonoids among pigmented rice varieties in the future.

CRediT author statement

   Jingyuan He: Investigation, Methodology, Formal analysis, Writing
   original draft & editing.

   Kaiqi Cheng: Investigation, Visualization, Software, Data curation.

   Zhou Yang: Data curation, Visualization, Validation.

   Jingzhang Geng: Resources, Supervization.

   Wengang Jin and A.M. Abd El-Aty: Conceptualization, Funding
   acquisition, Project administration, Supervization, Writing -review &
   editing.

Declaration of competing interest

   The authors declare that they have no known competing
   financialinterestsor personal relationships that could have appeared to
   influence the work reported in this paper.

Acknowledgements