*2.3. Moisture Analysis*

Moisture was determined in triplicate in the oven (SP 200, SP Labor®, São Paulo, Brazil), at 65 ± 1 ◦C, for approximately 72 h [29].

#### *2.4. Extraction of Carotenoids*

The carotenoid pigments were extracted from the lyophilized material (peel + pulp), according to the recommended procedures by Rodrigues-Amaya et al. [30]. Three grams of the edible portion of the samples (peel + pulp) were crushed by the use of a mortar and pestle, and a few drops of distilled water were added and extracted to color exhaustion with 20 mL of acetone (7 times) in an ultrasonic bath (Labsonic LBS 1-H22.5, Treviglio, Bergamo, Italy) for 10 min each time. Then, the extracts were individually centrifuged (Awel MF20-R, Multifunction Refrigerated Centrifuge, Blain, France) at 4000 rpm, 5 ◦C, for 10 min in order to withdraw clear solution on the top. The acetonic extracts were pooled together was concentrated to about 25 mL, in a rotary evaporator (Buchi-heating bath B-491, Buchi, Milan, Italy) at temperature below 35 ◦C. The dry product was diluted with equal volumes (25 mL) of a mixture of ethyl ether and hexane (1:1) and distilled water (50 mL) and worked up with a separating funnel. The lipofilic phase, cleared by the hydrophilic impurities, was evaporated to dryness using a rotary evaporator (Buchi-heating bath B-491 at 35 ◦C, and the residue was dissolved in 2 mL of MeOH/MTBE (1:1) and filtered in filter units (PTFE, 0.45μm, 13mm, Sigma Aldrich, Milan, Italy) prior to HPLC analysis. Samples were stored at −20 ◦C until they were analyzed.

#### *2.5. Analysis of Carotenoids by HPLC-DAD-APCI-MS*

The analysis was performed on an HPLC system (Shimadzu, Kyoto, Japan) equipped with a CBM-20A controller, two LC-20AD pumps, a DGU-20A3R deaerator, a SIL-20AC autosampler, a CTO 20AC column oven and an SPD-M20A photo diode array detector. The data were processed with the Labsolution software. For MS analysis a mass spectrometer detector (LCMS-8040) was used, equipped with an APCI (atmospheric pressure chemical ionization) interface, both in positive and negative ionization mode.

The column used was YMC C30 (250 mm × 4.6 mm × 5 μm); the mobile phases: MeOH/MTBE/H2O (81:15:4, solvent A), MeOH/MTBE/H2O (6:90:4, solvent B); the linear gradient used was: 0–100% B from 0 to 140 min. The column temperature was maintained at 30 ◦C. The flow was 0.8 mL/min and the injection volume was 20 μL.

The UV-Vis spectra were acquired in the range 220–700 nm, while the chromatograms were extracted at 450 nm (sampling frequency: 4.16 Hz, time constant: 0.64 s). The MS was set up as follows: Scan, both APCI positive (+) and negative (-); atomized gas flow (N2): 2.0 L/min; drying gas flow: 5 L/min; Time of the event: 0.06 s; range *m*/*z*: 300–1200; interface temperature: 350 ◦C; Desolvation line (DL) temperature: 300 ◦C; thermal block: 300 ◦C. The samples were analyzed in triplicate.

#### *2.6. Identification and Quantification of Carotenoids*

Carotenoids were identified by their UV-Vis spectra, MS spectra, elution order, comparison with the available standard and literature data.

The kumquat carotenoids quantification was performed from the analytical curves. External standards quantitative determination of each compound was performed using all reference materials listed in Section 2.1, in the concentration range from 5 to 50 μg/mL at six concentration levels. The results were obtained from the average of three determinations and the CV% was below 8% in all the LC measurements. The *R* coefficient for the calibration curves was always above 0.9962, with LOD and LOQ values of 0.07 and 0.22 ppm for β-carotene, 0.1 and 0.33 ppm for β-cryptoxanthin, 0.06 and 0.18 ppm for lutein, 0.08 and 0.3 ppm for zeaxanthin, and 0.12 and 0.24 ppm for physalein, respectively.

#### **3. Results and Discussion**

#### *3.1. Carotenoids Qualitative Profile of Brazilian Kumquat*

Figure 2 shows the chromatographic profile of the carotenoid composition in not saponified kumquat fruits extracts. The identified compounds are shown in Table 1, together with the UV–Vis and MS spectra information. In Figure 3 are reported the UV-Vis (PDA) and mass spectrum of β-citraurin-laurate and β-citraurin-myristate, detected in the kumquat carotenoid extracts. It can be appreciated that the esterification does not affect the PDA spectra of β-citraurin.

**Figure 2.** Chromatographic profile of native carotenoids of kumquat from Brazil: peel + pulp. Identification of the compounds are in Table 1.


**Table 1.** Compounds identification for Figure 2 (kumquat from Brazil: peel + pulp).

Rt: retention time; PDA: photodiode array; λnm: wavelength of maximum absorption; MS: mass spectrometry; APCI: atmospheric pressure chemical ionization.

**Figure 3.** UV-Vis (PDA) and mass spectrum (APCI negative) of β-citraurin-laurate (**A**) and β-citraurinmyristate (**B**) of kumquat from Brazil.

Moreover, both the molecular ions [M]−• at respectively *m*/*z* 614 and *m*/*z* 642 relative to the β-citraurin-laurate and β-citraurin-myristate esters, obtained in the negative APCI mode, are also clearly shown in the same figure. Eleven different carotenoids were identified in kumquat from Brazil; four different β-citraurin esters and three β-cryptoxanthin esters were identified.

The predominant carotenoids were β-citraurin-laurate and β-cryptoxanthin-laurate, both esterified with lauric acid. The chemical structures of these carotenoids are shown in Figure 4. Interestingly, no free β-citraurin was detected in the present study.

**Figure 4.** Chemical structures of β-cryptoxanthin-laurate and β-citraurin-laurate, main carotenoids detected in kumquat.

Shirra et al. [9] detected only 4 carotenoids in kumquat from Italy without saponification (β-carotene, β-cryptoxanthin, lutein and zeaxanthin), which were also detected in the present study. In contrast to our study, Agos et al. [18] reported only β-citraurin and β-cryptoxanthin in the free form in kumquat from Hungary. However, these researchers used saponification after the extraction process and did not quantify the carotenoid esters. Saponification may result in destruction or structural transformation of carotenoids [31]. Huyskens et al. [19] studied the qualitative composition of kumquat carotenoids from Israel by thin-layer chromatography and reported several carotenoids, including β-citraurin, β-cryptoxanthin, lutein, β-carotene and zeaxanthin which were, also determined in the present study. In addition, Huyskens et al. [19] reported that violaxanthin was the predominant component in kumquat from Israel, whereas in our study β-citraurin-laurate was the major component.

Esterification with saturated fatty acids improves the stability of xanthophylls such as β-citraurin and β-cryptoxanthin against heat and UV light, but does not affect their antioxidant activity [32]. During the storage and processing of the fruits the xanthophyll esters were more stable than the free xanthophyll [33]. The pigment β-citraurin is responsible for the citrus reddish color, it derives from of β-cryptoxanthin or zeaxanthin and accumulates in some citrus varieties [34,35]. In these fruits, the accumulation of β-citraurin is not a common event, it is observed only in the flavedos of some varieties during fruit ripening [34]. Frutita, a tropical fruit from Panama, showed a very high content of β-citraurin [36].

Recent studies have shown that β-cryptoxanthin and β-citraurin esterified with lauric acid, myristic acid and palmitic acid are found in the Mandarin Satsuma, [35].

The dietary intake of β-cryptoxanthin has been shown to prevent and reduce some pathologies such as: cancer, diabetes and rheumatism due to its antioxidant activity [37–39]. Breithaupt et al. [40] have verified that in chili, papaya, peach and persimmon, β-cryptoxanthin is mainly esterified with saturated fatty acids. Furthermore, the bioavailability of β-cryptoxanthin esters is comparable to the non-esterified form, since fatty acids can be effectively hydrolyzed by β-cryptoxanthin esters before intestinal absorption in the human body [40]. In several citrus varieties, the esterified form β-citraurin is found [35]. As already observed in another study [41], we have found β-cryptoxanthin and β-citraurin in both free and esterified forms.

#### *3.2. Carotenoids Quantitative Profile of Brazilian Kumquat*

In the present study, β-citraurin-laurate was the carotenoid found in the highest content in kumquat from Brazil (607.33 μg/100 g fresh matter, representing 27.80% of total carotenoids), followed by β-cryptoxanthin-laurate (552.59 μg/100 g fresh matter, representing 25.31% of total carotenoids). Thus, β-citraurin-laurate and β-cryptoxanthin-laurate represented 53.11% of the total carotenoids

found. The different forms of β-citraurin (4 components esterified with fatty acids) and β-cryptoxanthin (1 component in free form and 3 in esterified form) represented 84.34% of the carotenoids found in kumquat from Brazil. Forms of β-citraurin and β-cryptoxanthin esterified with fatty acids accounted for 79.54% of the total carotenoids in kumquat. These results show the importance of the study of the intact carotenoids composition in different matrices (Table 2).


**Table 2.** Carotenoid content in kumquat from Brazil (peel + pulp).

\* Mean of 3 repetitions ± standard deviation (SD). The carotenoid content in the fresh kumquat was calculated based on the average moisture content (*n* = 3) of the lyophilized fruit (14.08%) and the fresh fruit (81.16%).

β-carotene was found in smaller amounts (120.19 μg/100 g fresh matter), representing 5.50% of total carotenoids in kumquat, as well as free xanthophylls (β-cryptoxanthin, lutein and zeaxanthin), which represented 14.96% of total carotenoids (326.96 μg/100 g) (Table 2). β-carotene and β-cryptoxanthin possess provitamin A activity, playing a key role in human health [42]. Differently from our results, Wang et al. [20] reported that β-cryptoxanthin was the major carotenoid present in kumquat cultivated in Taiwan, while Schirra et al. [9] found lutein to be the major carotenoid present in kumquat from Italy.

Schirra et al. [9] found a lower concentration, in fresh matter, of β-carotene (33 μg/100 g), β-cryptoxanthin (26 μg/100 g), lutein (44 μg/100 g) and zeaxanthin (24 μg/100 g) in kumquat from Italy when compared to the concentrations of these carotenoids in kumquat from Brazil, reported here for the first time. Wang et al. [41] found contents in dry matter that were much lower than reported here, for lutein (9.9 μg/100 g), zeaxanthin (10.4 μg/100 g), β-cryptoxanthin (183 μg/100 g) and β-carotene (131 μg/100 g) in kumquat cultivated in Taiwan. Carotenoid composition and contents in fruits can be influenced by genetic factors, geographical regions, fruit processing, storage methods Giuffrida et al. [25] and environmental conditions Lu et al. [26], which can explain the differences found in these studies. Besides the climatic factors, the irrigation, soil conditions, fertilizers and herbicides may also affect the carotenoids accumulation [26]. Regarding the carotenoids content in food, Britton [43] have proposed the following classification ranges: low: 0–0.1 mg/100 g; moderate: 0.1–0.5 mg/100 g; high: 0.5-2 mg/100 g; very high: >2 mg/100 g. Thus, the total carotenoid content of kumquat from Brazil was very high (2185.16 μg/100 g).
