Introduction

Tea and coffee, the most popular beverages in the world, have been consumed for thousands of years. Tea is produced by brewing the dried leaves and buds of Camellia sinensis. Annual production of about 1.8 million tons of dried leaf provides about 40 L of beverage per capita worldwide (Wang and Ho 2009). Coffee is the third most frequently consumed beverage in the world, after water and tea (Wang and Ho 2009). Once ripe, coffee berries are dried, roasted at various temperatures, and then ground and brewed. Many studies have shown the relation between the consumption of tea and coffee and their potential disease prevention properties, which may be due to their polyphenol content (Wang and Ho 2009).

There is a growing interest in disease prevention using natural products, especially polyphenolic compounds, as consumer awareness of their possible beneficial health effects increases. Mangiferin, as an example of the many polyphenolic compounds, appears to have considerable promise as a natural disease chemopreventant.

Mangiferin (1,3,6,7-tetrahydroxyxanthone-C2-ß-glucoside) (Fig. 1) is a C-glycosylated xanthone which is highly resistant to hydrolysis and is widely distributed in the plant kingdom as reviewed by Vyas et al. (2012). The following species are reported to contain mangiferin namely Anemmarhena asphodeloides, Aphloia theiformis, Arrabidaea patellifera, Arrabidaea samydoides, Bersama abyssinica, Bombax ceiba, Bombax malabaricum, Cratoxylum cochinchinens, Cyclopia genistoides, Cyclopia subternata, Folium mangiferae, Folium pyrrosiae, Gentiana lutea, Gentianella nitida, Gnidia involucrata, Hypericum perforatum, Mahkota dewa, Mangifera indica, Mangifera odorata, Mangifera persiciformis, Phaleria cumingii, Phaleria macrocarpa, Mangifera persiciformis, Polygala hongkongensis, Pyrrosia gralla, Rhizoma anemarrhenae, Rhizoma belamcandae, Salacia hainanensis, Salacia oblonga, Salacia reticulate, Swertia macrosperma, Swertia chirata, Swertia mussotii, Swertia punicea, Trichomanes reniforme, and Zizyphus cambodiana (Vyas et al. 2012).

Fig. 1
figure 1

Mangiferin (1) and isomangiferin (2) structures

Full size image

However these plants apart from the Cyclopia family and Mangifera indica (Barreto et al. 2008) are not a viable source as a convenient provider of mangiferin for oral consumption. The Cyclopia family is indigenous to South Africa, and is popular for the preparation of a herbal tea termed “green honeybush.” This represents a viable natural source of mangiferin, in which amounts as high as 11.8 g/100 g extract (n.b. not g/100 g plant material) have been reported (De Beer and Joubert 2010), but the availability of the plant is extremely limited. In addition, Campa et al. (2012) have recently demonstrated mangiferin accumulation in the leaves of seven African Coffea taxa namely C. anthonyi, C. arabica, C. eugenioides, C. heterocalyx cf., C. pseudozanguebariae, C. salvatrix, and C. Sessiliflora which ostensibly represent an excellent natural source of this potentially disease-preventive agent, for which the following properties have been described.

In addition to inhibition of bowel carcinogenesis (Yoshimi et al. 2001), mangiferin is reported to have potent cytoprotective and antigenotoxic effects against CdCl2-induced toxicity in HepG2 cells, on the basis of absorption spectra, which were attributed to a decrease in the levels of reactive oxygen species (Satish Rao et al. 2009), anti-allergic properties (Rivera et al. 2006), anthelminthic properties (Garcia et al. 2003), gastro-protective effects (Carvalho et al. 2007; Morais et al. 2012), potent hepatoprotective effects on CCl4-induced liver damage in mice (Rasool et al. 2012), chemopreventive and chemotherapeutic effects (Rajendran et al. 2008), and antiviral properties (Yoosook et al. 2000). Mangiferin has also been used in foodstuffs for treating diabetes (Wada 2007) and the treatment and prevention of neurodegenerative diseases and ageing symptoms (Matute Almau et al. 2007; Rao et al. 2012). Mangiferin has also been reported to have strong scavenging effects against reactive oxygen species (Barreto et al. 2008) which attack molecules including DNA, proteins, and lipids, contributing to the development of several chronic diseases including cancer (Tang et al. 2004). Mangiferin is considered non-toxic, because its reported oral LD50 value in mice is 400 mg/kg (Jagetia and Baliga 2005).

The aim of this work was the quantitation of mangiferin and isomangiferin (Fig. 1) in methanol extracts of Coffea arabica leaves from Brazil and Costa Rica grown under various environmental conditions and infusion of a commercial source to establish a readily available natural source for future human intervention trials.

Materials and Methods

Chemicals

Acetic acid, dimethyl sulfoxide, n-hexane and methanol were purchased from Merck (Darmstadt, Germany); and acetonitrile from Fluka/Riedel de Haen (Seelze, Germany) while mangiferin was obtained from Extrasynthese (Lyon Nord, Genay, France). All solutions were made up in either double distilled water or methanol.

Plant Material

The coffee leaves of nine cultivars (C. arabica) were obtained from Brazil (Table 1) and Costa Rica (Table 2) as listed. The material was obtained from several different sources in Brazil: Ceará (Guaramiranga), Minas Gerais (Bom Sucesso and Manhumirim), and from Costa Rica (Herredura and Turrialba): the Turrialba cultivars were from CATIE (Solutions for Environment and Development). In addition, one commercial source of coffee tea leaves was used for comparison.

Table 1 Content of mangiferin and isomangiferin in coffee leaves from Brazil
Full size table
Table 2 Content of mangiferin and isomangiferin in coffee leaves from Costa Rica
Full size table

Preparation of Plant Material

Coffee leaves were freeze-dried (Christ, Gefriertrocknungsanlangen, Osterode, Germany) to constant weight. Dried samples were pulverized by blending to a fine homogeneous powder in a Moulinex coffee grinder prior to extraction.

Coffee Leaf Extraction

The various samples of C. arabica leaves (10 g of freeze-dried powder) were extracted in duplicate for 3 h with n-hexane in a soxhlet apparatus to remove lipids. After drying, the material was further extracted, three times for 3 h with methanol, the solutions pooled and evaporated to dryness at 40 °C by rotary evaporation under reduced pressure. The extracts were dissolved in methanol and analyzed by analytical reverse-phase high-performance liquid chromatography electrospray ionization mass spectrometry (HPLC-ESI-MS).

Coffee Leaf Infusion

In order to assess the efficiency of mangiferin extraction by infusion, varying amounts (100–1000 mg) of coffee leaf powder from the commercial source (3.05 ± 0.16 g/kg mangiferins) were transferred to 15 mL Falcon tubes in duplicate, and 10.0 mL of double distilled boiling water were added and maintained in a water bath at 100 °C for 20 min. The solutions were filtered, and 1.0 mL was set aside for HPLC-ESI-MS. To assess the rate of extraction, duplicate 100 mg samples of coffee leaf powder were transferred to 15 mL Falcon tubes and 10.0 mL of double distilled boiling water was added and maintained in a water bath at 100 °C for 1 to 30 min. The solutions were filtered and 1.0 mL was set aside for HPLC-ESI-MS. Finally, to assess the effect of temperature on efficiency of extraction, duplicate 100 mg samples of coffee leaf powder were transferred to 15 mL Falcon tubes and 5.0 mL of double distilled water at temperatures ranging from 35 to 100 °C was added and maintained in a heating block for 5 min. The solutions were filtered and 1.0 mL was set aside for HPLC-ESI-MS.

High-Performance Liquid Chromatography—UV

HPLC-UV analysis was conducted on an Agilent 1100 HPLC, coupled to an Agilent single quadrupole mass-selective detector (HP 1101; Agilent Technologies, Waldbronn, Germany). Chromatographic separation of methanol extracts (10 μL) and aqueous infusion (20 μL) was conducted using a C18, reverse-phase (5 μ), Gemini column (250 × 4 mm I.D.; Phenomenex, Aschaffenburg, Germany). The mobile phase consisted of 2 % acetic acid in water (solvent A) and acetonitrile (solvent B) with the following gradient: 95 % A for 10 min, to 90 % A in 1 min, to 60 % A in 9 min, to 80 % A in 10 min, to 60 % A in 10 min, to 0 % A in 5 min, and continuing at 0 % A until completion of the run. Detection of phenolic compounds was by means of UV absorbance (A) at 257, 278, 320, and 340 nm at room temperature.

High-Performance Liquid Chromatography Electrospray Ionization Mass Spectrometry

HPLC-ESI-MS analysis was conducted on an Agilent 1100 HPLC, coupled to an Agilent single quadrupole, mass-selective detector (HP 1101; Agilent Technologies, Waldbronn, Germany). Chromatographic separation of methanol extracts was conducted using a column of the same type, and dimensions as for analytical HPLC (Latek, Eppelheim, Germany). The mobile phase consisted of 2 % acetic acid in double distilled water (solvent A) and acetonitrile (solvent B) with the following gradient profile: initially 95 % A for 10 min; to 90 % A over 1 min; to 60 % A over 9 min; to 80 % A over 10 min; to 60 % A over 10 min, to 0 % A over 5 min and continuing at 0 % A until completion of the run. Detection of phenolic compounds was by means of UV absorbance (A) at 257, 278, 320, and 340 nm at room temperature. Mass spectra in negative-ion mode were generated under the following conditions: fragmentor voltage = 100 V, capillary voltage = 2500 V, nebulizer pressure = 30 psi, drying gas temperature = 350 °C, and mass range = 100–1500 D. Instrument control and data handling were performed with the HP Chemstation software on a PC. The amount of mangiferin and isomangiferin in the extracts was calculated from a standard curve of an authentic mangiferin standard in the range 0–422 μg/mL (0–1000 μM) at 257 nm (y = 24.084×).

Limits of Detection, Limits of Quantitation, and Repeatability

The estimated of limits of detection (LOD) and limits of quantitation (LOQ) were made using the method based on the regression parameters of the calibration curve (Huber 2003; Basilio et al. 2014). Intra- and inter RSD were made in five different concentrations (range, 25, 50, 125, 250, and 500 μM), with three replicates of each.

Statistics

Correlations between release of mangiferins by infusion, with regard to amount extracted and at various temperatures, were evaluated by linear regression using the Origin program (version 7.5).

Results and Discussion

Limits of Detection, Limits of Quantification, and Repeatability

The sensitivity of the developed method was assessed by determining the LOD and LOQ. The LOD for mangiferin and isomangiferin was 0.164 g/kg while the LOQ was 0.492 g/kg. These results are in agreement with the LOD and LOQ of mangiferin as reported in the literature (Jounert et al. 2003). Intra- and inter-RSD of five different concentrations (25, 50, 125, 250, and 500 μM) were <2 % for three replicates of each.

Identification and Quantitation of Mangiferins in Methanol Extracts of Coffee Leaves

The identification of mangiferin and isomangiferin in the methanol extracts (Fig. 2) and infusions (Fig. 3), which displayed typical UV spectra of xanthones, were determined from the mass spectra of the compounds in negative-ion mode. A pseudo-mole peak was detected at [M–H]− at m/z = 421 (exact mass = 422.085) in negative-ion mode, which was the base peak for both mangiferin and isomangiferin. Bombardment of the mole peaks at a fragmentor voltage of 100 V gave significant fragment ions at m/z 403, 331, 301, and 271 typical of C-glucosides. The proposed fragmentation is shown in Fig. 4, which is in agreement with the literature (Sun et al. 2009). An additional fragment peak at m/z 259 was indicative of the norathyriol xanthone core (Souza JRR et al. 2012). Quantitation was achieved against a standard curve of mangiferin (0–1000 μM) from the reverse-phase analytical chromatograms at 257 nm.

Fig. 2
figure 2

HPLC–ESI-MS chromatograms (negative-ion mode) of a methanol extract of a commercial cultivar of Coffea arabica. Above: UV chromatogram (257 nm) under the conditions used for HPLC–ESI-MS. Below: The individual extracted ions at m/z = 421 [M–H]− of mangiferin  (Peak 1) and isomangiferin (Peak 2)

Full size image
Fig. 3
figure 3

HPLC–ESI-MS chromatograms (negative-ion mode) of an aqueous infusion (100°C) of a commercial cultivar of Coffea arabica. Above: UV chromatogram (257 nm) under the conditions used for HPLC–ESI-MS. Below: The individual extracted ions at m/z = 421 [M–H]− of mangiferin (Peak 1) and isomangiferin (Peak 2)

Full size image
Fig. 4
figure 4

Proposed fragmentation mechanism of the [M–H]− ion for mangiferin and isomangiferin

Full size image

Overall, the concentration (Tables 1 and 2) of mangiferin detected in the methanol extracts was in the range 0.48–4.17 g/kg whereas isomangiferin was in the range 0.16–0.80 g/kg representing 20 % or less in 8 (89 %) of the nine samples. Total mangiferins were in the range 0.67–4.97 g/kg. The range of total mangiferins, from both Brazil (0.67–4.97) and Costa Rica (0.85–4.01), was similar across the samples. Total mangiferin content tended to be considerably higher in Coffee leaves, obtained from trees growing in plantations under natural full-sun conditions, compared to other types of management such as organic treatment. Additionally, in one comparison with a species from Minas Gerais (Bom Sucesso), young leaves contained much higher concentrations (4.97 ± 0.11 vs 1.49 ± 0.19 g/kg) than old leaves.

Quantitation of Mangiferins in Coffee Leaf Tea Infusions

Infusion of varying amounts (100–1000 mg) of coffee leaf powder (Commercial variety) with boiling distilled water and maintenance at this temperature for 20 min revealed (Fig. 5) that release of mangiferins showed a linear dose response (r = 1.00; P = 0.0001). The amount released from 1 g was only 47 % lower than that after prolonged extraction with methanol.

Fig. 5
figure 5

Linear release of total mangiferins from coffee leaf powder (100–1000 mg/10 mL) by infusion at 100 °C of a commercial source of Coffea arabica

Full size image

The release of mangiferins by infusion with boiling water was instantaneous with 53 % extraction (compared to methanol extraction) after only 1 min, with no increase after 30 min (data not shown). Release of mangiferins from coffee leaf powder was strongly temperature dependent. At 35 °C, release of mangiferins was 59 % efficient compared to boiling water, rising to 88 % at 85 °C and to 100 % at 100 °C (Fig. 6). Therefore, infusion of Coffee tea leaves with boiling water (1.6 ± 0.06 g/kg) releases greater than 50 % of the available mangiferins present as compared to methanol (3.05 ± 0.16 g/kg) extraction.

Fig. 6
figure 6

Temperature dependence of total mangiferins release from coffee leaf powder (100 mg/5.0 mL) by infusion of a commercial source of Coffea arabica

Full size image

The data shows that coffee leaves obtained from all nine cultivars of C. arabica from both Brazil (n = 4) and Costa Rica (n = 5) contain appreciable concentrations of mangiferins dominated by mangiferin at 80 % or over with regard to isomangiferin. The ranges, in both the Brazilian and Costa Rican cultivars at 0.67–4.97 and 0.85–4.01 g/kg are in only moderate agreement with the data reported for 7 out of 23 species from Africa as reported by Campa et al. (2012) which is the only other publication on the content of mangiferins in Coffee leaves.

Campa et al. (2012) presented their data in % dry weight, and the range reported in the positive species was 0.026–16.359 % equivalent to 0.26–163.6 g/kg. The higher values are in the range for those reported in young mango leaves (Barreto et al. 2008), which is surprising. Study of the data in Campa et al. (2012) reveals that the standard curve of mangiferin (in an unstated solvent) was generated at 320 nm in the range 10–100 mg/mL representing 23.70–236.97 mM which appears inconceivable, because at these concentrations in any solvent, mangiferin is not totally soluble and would saturate the UV source.

However, it cannot be discounted that the coffee leaves studied by Campa et al. (2012) represent wild species of the old World and that the major cultivated species used in the commercial production of coffee, namely C. arabica, in general, contain less mangiferin in their leaves. Nevertheless, the extremely limited availability of wild type Coffea and Cyclopia species renders their suitability as providers of a natural source of mangiferin, peripheral, when compared to the global availability of Coffea arabica leaves.

Conclusion

In conclusion, our data showing total mangiferins at concentrations in the range 0.67–4.97 g/kg (which are 53 % soluble by infusion) indicate that the leaves of cultivated Coffea species are a useful natural source of mangiferins and consumption of coffee leaf tea brews may contribute significantly to elevated intake of these potentially health-promoting phenolic compounds, the exact relevance of which needs to be determined by future bioavailability and epidemiologic studies.