Chemical Composition of Coffea arabica Beans and Grafts at Different Roasting Levels

Abstract

The main commercial species of coffee is Coffea arabica; however, it is highly susceptible to biotic and abiotic factors. For this reason, scions of Coffea arabica species are used on robusta (Coffea canephora) rootstocks since grafting can modify the organoleptic, nutraceutical, and nutritional characteristics of crops. The objective of this study was to evaluate the effects of five roasting degrees (light, medium light, medium, medium dark, and dark) in two varieties of Coffea arabica species (Colombia Rojo and Costa Rica 95) and their respective grafts on C. canephora variety Romex. The concentrations of acrylamide, caffeine, and chlorogenic acids were analyzed using HPLC, and the nutrient content was analyzed using plasma induction atomic emission spectrophotometry. The graft of the Colombia Rojo variety with the medium dark roast had the highest concentration of caffeine (1.523%). The same variety with the light roast had the highest concentration of chlorogenic acid (3.088 mg g⁻¹), while the highest acrylamide content (336.70 ng g⁻¹) was found in the same variety with the medium light roast. The roasting degree, variety, and grafting of coffee influence the organoleptic, nutraceutical, and nutritional contents of coffee infusions.

1. Introduction

Coffee (Coffea arabica) cultivation represents a multi-billion-dollar sector that encompasses a long value chain from the farmer to the consumer. Coffee is the second most consumed beverage worldwide, with an expected production of 10.68 billion kg of green beans from 2023 to 2024 [1].

Coffee flavor is a quality determinant and considered the main attribute of sensory quality. Some non-volatile chemicals are present in green coffee, which are subject to drastic changes during the roasting process and are responsible for the distinctive flavor by while acting as precursors, i.e., free sugars, carbohydrates, chlorogenic acids, trigonelline, proteins, and lipids [2].

It is very important to monitor the identification of contaminants that are generated in the processing of coffee, especially chemical substances that represent a health risk, such as acrylamide. Acrylamide is classified as carcinogenic and is an amide formed through the Maillard reaction when foods rich in carbohydrates are heated to temperatures above 120 °C [3].

Roasted coffee exhibits antioxidant activity as it is rich in melanoidins that are generated during the roasting process [4]. In addition to the antioxidant activity of chlorogenic acids, they have been associated with the prevention of cancer and diabetes [5].

According to the authors of [1], the two most important coffee species are C. arabica (arabica coffee), which represents more than 60% of global production, and C. canephora, better known as robusta coffee.

Of the coffee produced in Mexico, approximately 96% is arabica and 4% is robusta. However, arabica varieties are more susceptible to coffee rust, a pest that has affected the coffee sector in the country since 2012, which has caused a reduction in production of more than 50% [6].

For this reason, in Mexico, the cultivation of the coffee variety Colombia Rojo (C. arabica) has been initiated. This variety exhibits resistance to coffee rust and originates from the C. arabica variety Caturra and the Timor hybrid [7]. Similarly, the Costa Rica 95 coffee variety, which is the result of hybridization between Timor and Caturra, is mainly characterized by resistance to coffee rust, high yields, and an ability to adapt to hot areas and acidic soils [8].

To manage nematodes, grafts of arabica varieties (susceptible) have been used on less susceptible robusta coffee rootstocks, which has proven to be a successful strategy for the control of this pest [9].

This study analyzed the contents of acrylamide, caffeine, chlorogenic acids, and minerals in coffee beans of Colombia Red, Colombia Red Graft, Costa Rica 95, and Costa Rica Graft with five different roasting levels.

2. Materials and Methods

2.1. Plant Material and Experimental Conditions

The present study was carried out with coffee beans from five-year-old plants of the Colombia Rojo and Costa Rica 95 varieties and their grafts (Colombia Rojo graft and Costa Rica 95 graft), planted in the congregation of La Laja, municipality of Tlaltetela, Veracruz, Mexico (19°17′10.8″ N, 96°56′51.9″ W, and altitude of 1250 m). The plants were grown under managed shade with approximately 70% solar radiation using the traditional selective planting method (planting, fertilization, and pruning), with 2000 to 2500 shrubs ha⁻¹ in cambisol soils. For the experiment, 150 trees of each variety were randomly selected from one hectare. Coffee cherries were harvested in the second week of February, and uniform red cherries were selected. Coffee cherry samples (20 kg) from a single harvest of each variety were immediately treated with wet milling until green coffee beans (gold) with 12% moisture content were obtained. The beans were left to rest for seven days, and defective beans were eliminated. Subsequently, manual quartering was carried out to obtain 5 kg samples for roasting.

2.2. Roasting Procedure

The coffee beans were roasted in a MT-1 fluidized bed roaster (Trejo; Zacatlan, Puebla, Mexico), with a capacity of 5 to 7 kg. The roast levels were light, medium light, medium, medium dark, and dark. The variables measured included initial humidity (%), initial and final temperatures (°C), initial weight (g), first and second crack (roasting curve), and temperature increase (5 °C every 30 s). The roasting times were between 8 and 13 min, and the final roasting time was defined by the color of the bean according to each level on the AGTRON/SCAA roasting scale, following the protocol established by the author of [10]. Three replicates were performed, and the roasting process was halted via ventilation once the desired roasting degree was reached (

Table 1. Conditions of coffee beans with five roast levels.

Conditions	Light	Medium Light	Medium	Medium Dark	Dark
	Roasting Degree
Initial temperature (°C)	150	150	160	190	220
Final temperature (°C)	174.81	180–190	190–210	200–220	230–240
Initial weight (g)	1000.00	1000.00	1000.00	1000.00	1000.00
Agtron color (Å)	75–70	70–65	65–58	58–45	45–35
Initial humidity (%)	12	12	12	12	12
Final humidity (%)	3.3	2.5	1.9	1.5	1.3

).

2.3. Quantification of Chlorogenic Acids (5-CQA) and Caffeine Concentrations

Analysis of chlorogenic acids (5-CQA) was performed on an HPLC system (Dionex, ICS 3000; San Jose, CA, USA) equipped with a UV/VIS detector set at a 325 nm wavelength. An isocratic mobile phase consisting of 75% HPLC-grade water and 25% H₃PO₄-acidified methanol was applied at a flow rate of 1.2 mL min⁻¹, using a 5 mm, 4.6 × 150 mm Acclaim™ 120 C18 reverse phase column (ThermoFisher^®, Sunnyvale, CA, USA) and a 5 μm, 10 × 2 mm ZORBAX C18 column (Agilent Technologies, Santa Clara, CA, USA).

The caffeine concentration was determined using a variation in NMX-F-182-SCFI-2011 [11]. Results were expressed as a percentage of dry matter weight (% DMW). Caffeine was extracted using a thermoreactor (Hach, DRB200; Loveland, CO, USA) at 135 °C for 5 min. Quantification was performed on a Dionex ICS 3000 HPLC (San Jose, CA, USA) equipped with a UV/VIS detector at a wavelength of 272 nm. An isocratic mobile phase consisting of 75% HPLC-grade water and 25% H₃PO₄-acidified methanol was applied at a flow rate of 1.5 mL min⁻¹ on a 120 C18 5 mm 4.6 × 150 mm column (Acclaim™ Model, ThermoFisher, Sunnyvale, CA, USA) with H₃PO₄. Samples were run in triplicate. The results were expressed as a percentage of dry matter weight (% DMW). Calibration curves for chlorogenic acids and caffeine were prepared using eight concentration points (0, 0.01, 0.008, 0.0065, 0.005, 0.004, 0.002, and 0.0005%) and analyzed in duplicate to verify the linearity of the methods. For these calibration curves, chlorogenic acids 95% CAS 327-97-9 and ≥95% CAS 58-08-2 C1778 (Sigma-Aldrich Saint Louis, Saint Louis, MO, USA) and caffeine ≥ 99% CAS 58-08-2 C1778 (Sigma-Aldrich, Saint Louis, MO, USA) were used.

2.4. Quantification of Acrylamide

Acrylamide extraction was performed by first adding 50 mL of water to a 5 g coffee sample and heating it at 80 °C for 5 min. Then, 1 mL of water was extracted and strained through an SPE cartridge (Thermo Scientific™ HyperSep™ Hypercarb™, Waltham, MA, USA) to separate the acrylamide. Acrylamide quantification was performed using a DAD UV/VIS HPLC (Agilent, 1200, Waldbronn, Karlsruhe, Germany) at a wavelength of 195 nm. An isocratic mobile phase consisting of 85% HPLC-grade water and 15% H₃PO₄-acidified methanol was applied at a flow rate of 0.4 mL min⁻¹, using an Acclaim™ 120 C18 5 µm 4.6 × 150 mm reverse phase column (ThermoFisher^® Sunnyvale, CA, USA) and a ZORBAX C18 5 μm, 10 × 2 mm column protector (Agilent Technologies, Santa Clara, CA, USA). The calibration curve was obtained using acrylamide ≥ 99.9% 79-06-1 (Sigma-Aldrich; Saint Louis, MO, USA) at eight concentration points (0, 0.01, 0.008, 0.0065, 0.005, 0.004, 0.002, and 0.0005 mg mL⁻¹), performed in duplicate to verify the linearity of the methods.

2.5. Nutritional Analysis

The concentrations of the nutrients P, K, Ca, Fe, Cu, Zn, and Mn were determined following the methodology described by the author of [12] with coupled plasma induction atomic emission spectrophotometry (Agilent ICP-AES, model 725-ES; Victoria, Australia). The N content was quantified using the micro-Kjeldahl method. Results were expressed in g kg⁻¹ for macronutrients and mg kg⁻¹ for micronutrients.

2.6. Experimental Design

A completely randomized design with a factorial arrangement was used. The study factors included the genotypes (from four estates) of C. arabica L. var. Colombia Rojo and Costa Rica 95, along with their respective grafts onto C. canephora P. var. Romex, and five roasting levels (light, medium light, medium, medium dark, and dark). This resulted in 20 treatments, each with four replicates (Appendix A and Appendix B).

Statistical Analysis

With the resulting data, ANDEVA with Tukey’s comparison test (p ≤ 0.05) was performed using the SAS statistical package version 11 (SAS Institute, 2011).

3. Results

3.1. Caffeine, Chlorogenic Acid, and Acrylamide Contents in Coffee Beans

Analysis of the variety factor revealed significant statistical differences between treatments. The highest concentration of caffeine (Figure 1a,b) was found in the beans of the Colombia Rojo graft, along with the highest concentration of acrylamide (Figure 1c,d). Meanwhile, the Costa Rica 95 coffee variety presented the highest content of chlorogenic acid (Figure 1e,f).

3.2. Caffeine, Chlorogenic Acid, and Acrylamide Contents in Coffee Beans at Five Roasting Levels

The highest level of caffeine was found in the medium dark roast. The medium light roast level presented the highest acrylamide content, while the highest concentration of chlorogenic acid (2.87 mg g⁻¹) was found in coffee beans with a light roast (Figure 1).

3.3. Concentrations of Caffeine, Chlorogenic Acid, and Acrylamide Based on the Interaction of the Variety and Its Respective Grafts

When analyzing the interaction factor between variety and roasting level, the grafted Colombia Rojo beans with a medium roast exhibited the highest concentration of caffeine (Figure 2a). The highest concentration of chlorogenic acid (Figure 2b) occurred in the grafted Colombia Rojo beans with a light roast. The highest amount of acrylamide (Figure 2c) was observed in the grafted Colombia Rojo beans with a medium light roast. When analyzing the effect of roasting on acrylamide content, it was observed that the coffee beans with a medium light roast had the highest acrylamide content, while the lowest contents were found in the beans with a dark roast.

3.4. Macro- and Micronutrient Content in Coffee Beans

When analyzing the effect of the interaction factor variety by roasting level, no significant statistical differences were found in the nitrogen contents for Colombia Rojo and its graft (Colombia Rojo graft) and Costa Rica 95 at the five roasting levels; however, the Costa Rica 95 graft deviated from this tendency by presenting the highest nitrogen (Figure 3a) contents in the coffee beans with a dark roast level and the lowest with a medium roast. Regarding the contents of phosphorus (Figure 3b), potassium (Figure 3c), and calcium (Figure 3d), statistical differences were found between the treatments in the four varieties at the five roasting levels. Regarding calcium, the highest contents of this nutrient were found in the coffee beans of the Colombia Rojo variety with a medium dark roasting level, and the lowest contents were found in the Colombia Rojo graft and the Costa Rica 95 variety with medium and dark roasting levels, respectively.

The content of micronutrients, influenced by the interaction factor between coffee variety and roasting level, presented significant statistical differences between treatments. The highest concentrations were observed as follows: Fe (Figure 4a) in the Colombia Rojo variety with the light roasting level; Cu (Figure 4b) in the Costa Rica 95 variety with medium roasting; Zn (Figure 4c) in the Colombia Rojo variety with medium roasting; and Mn (Figure 4d) in the Colombia Rojo variety with dark roasting. Bo did not follow this trend since there were no significant statistical differences between the treatments.

4. Discussion

The differences in the contents of caffeine and chlorogenic acid between the pure varieties and the grafts may be due to the fact that the gene expression of the rootstock influences the gene expression of the scion. The author of [13] observed in orange tree plants that the molecular mechanism underlying dwarfism was mainly reflected in genes responsible for phytohormone metabolism, nitrogen (N) metabolism, carbon (C) metabolism, and photosynthesis in scions and rootstocks. They also observed that grafting affected the difference in the expression of secondary metabolism genes, such as some genes related to phenylpropanoid and flavonoid biosynthesis.

The content of metabolites in the fruit can also be affected by this factor. Mehic et al. [14] studied the influence of rootstock on sweet cherry fruit quality and found that rootstock genotype and cultivar/rootstock interaction influenced the phenolic profile of sweet cherry fruit. They also reported that there were significant differences between rootstocks regarding their individual polyphenol content.

The highest concentration of acrylamide was found in the Colombia Rojo graft variety, according to the author of [15]. The differences in acrylamide concentration observed between arabica and canephora coffee varieties could be because robusta coffee varieties contain higher concentrations of amino acids and sugars, and the decomposition of sucrose in coffee during heating and oxidation of lipids are the main sources of carbonyl compound formation, which is essential for the conversion of asparagine into acrylamide [16]. Although no pure strains of C. canephora were studied, it is important to emphasize that in the observed cases, the differences between pure strains and grafts could be due to plastid genome transfer at the graft junctions. Nucleic acids, such as microRNAs, are known to be systematically propagated through the phloem via plasmodesmata. While these intercellular plastid channels normally only allow the transit of small molecules, plasmodesmata are likely responsible for lateral gene transport in grafts [17].

The increase in caffeine content in darker roast beans could be mainly attributed to their caffeine thermostability and mass loss of thermolabile compounds during the roasting process [18]. Consistent with this work, caffeine levels were higher in the early roasting degrees but decreased significantly from light, medium light, and medium to medium dark roast coffees. As the temperature increased to higher levels (dark roast), the caffeine concentration decreased, which could be a result of caffeine sublimation at higher temperatures during roasting [19].

Among the chlorogenic acids, 5-caffeoylquinic acid (5-CQA) was found in greater proportion in the grafted coffee beans of the Colombia Rojo variety (3.09 mg g⁻¹) and the Costa Rica variety (3.04 mg g⁻¹). During the roasting process of coffee beans, reactions occur between their components. Notable changes include decreases in polysaccharides, proteins, chlorogenic acids, and trigonelline; degradation of sucrose; increase in ash; and the formation of melanoidins [20].

These differences between varieties are probably because some edaphoclimatic growing conditions, such as altitude, modify the characteristics of coffees, such as their composition [21,22]. The degree of roasting modifies the chlorogenic acid content. It has been observed that the higher the degree of roasting, the lower the chlorogenic acid content because these compounds undergo thermal degradation as the roasting temperature increases. This is because high temperatures during the roasting process can lead to the breakdown of carbon–carbon bonds in the structure of chlorogenic acids, leading to thermal degradation and isomerization of the structures of chlorogenic acids [23].

This may occur during the roasting of coffee beans because acrylamide forms rapidly in the early stages of coffee roasting. This coincides with the depletion of free asparagine, causing the concentration of acrylamide to peak before the first crack, potentially exceeding 1000 µg/kg. As roasting continues and asparagine is depleted, the acrylamide concentration starts to decrease rapidly because it contains an electrophilic vinyl group, making it highly reactive [16].

When analyzing the effect of the interaction factor between coffee variety and roasting degree on the micronutrient content, the highest concentrations were observed as follows: Fe in the Colombia Rojo variety with a light roasting degree; Cu in Costa Rica 95 with a medium roast; Zn in the Colombia Rojo variety with a medium roast; and Mn in the Colombia Rojo variety with a dark roast. Fe was the fourth most abundant mineral found in the coffee samples, followed by Mn, Cu, and Zn, as described by other authors [24].

The differences between the pure and grafted varieties could be because, in grafted plants, RNA molecules are highly transmissible between adjacent cells and throughout the plant, transported by the plant plasmodesmata [17]. The author of [25] reported that in grafting, the long-distance movement of mRNA from the rootstock to the shoots of the graft scions (grafted variety) can cause visible changes in the morphology of the leaves of the latter. This indicates that translocated RNAs are functional; not only can they be transported over long distances via phloem binding in the graft, but they also act as signal molecules that regulate growth and development, generating resistance to stress in remote cells or tissues and resulting in changes in plant traits, rootstocks, and scions [26].

It was observed that N and K were the main minerals found in coffee beans, followed by Mg and Ca. According to the author of [27], after nitrogen, potassium (K) is the most accumulated nutrient in the fruits of the coffee plant, because K is related to the enzymatic activation of several metabolic processes, such as photosynthesis, protein and carbohydrate synthesis, and the maintenance of cell turgor. It is also directly related to the transport of sugars from the source to the fruits.

According to the findings of the author of [24], the mineral content found in coffee is not affected by the degree of roasting, which is why the greatest variation occurs in macro- and micronutrients. For example, arabica coffees present high concentrations of K, Na, Ca, Mg, Mn, and Fe [28]; however, the contents of P and Cu in robusta coffee are higher than in arabica coffees [29]. Therefore, the quality of coffee is directly related to the chemical composition of roasted beans, which is influenced by variety, origin, and climate [30].

5. Limitations

The implementation of the method for acrylamide quantification was developed and validated by the authors; however, the method has not yet been published for reproducibility.

The reproducibility of the culture is another limitation, since the edaphoclimatic conditions of the area are difficult to replicate.

6. Conclusions

This study observed a relationship between coffee varieties and roasting degree of coffee beans when macro- and micronutrients, caffeine, and chlorogenic acid levels were assessed. The highest concentrations of these compounds were found in the Colombia Rojo coffee variety. The use of the Colombia Rojo graft coffee variety, along with medium and dark roasted coffee beans, resulted in the highest concentrations of caffeine. Light roasting of coffee beans resulted in the highest chlorogenic acid content, while medium dark roasting resulted in the highest concentrations of caffeine. Medium dark and dark roasting of coffee beans produced the highest concentrations of acrylamide, so this type of roasting is not recommended, confirming the relationship between coffee varieties and the degree of roasting of their beans.