SILVIA JULIANA MARTINEZ GELVEZ 

 

 

 

 

 

 

IMPROVEMENT CHARACTERISTICS OF SEMI-DRY COFFEE FERMENTATION 

WITH STARTER YEASTS  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

LAVRAS – MG 

2017 



 

SILVIA JULIANA MARTINEZ GELVEZ 

 

 

 

 

IMPROVEMENT CHARACTERISTICS OF SEMI-DRY COFFEE FERMENTATION 

WITH STARTER YEASTS  

 

 

 

 

 

Dissertação apresentada à Universidade Federal de 

Lavras, como parte das exigências do Programa de Pós-

Graduação em Microbiologia Agrícola, área de 

concentração em Microbiologia Agrícola, para a 

obtenção do título de Mestre. 

 

 

Orientadora 

Dra. Rosane Freitas Schwan 

 

Coorientador 

Dr. Disney Ribeiro Dias 

 

Coorientadora  

Dra. Maria Gabriela da Cruz Pedrozo Miguel 

 

 

 

 

 

 

 

 

 

 

LAVRAS – MG 

2017 



 

 

 

  

 

Gelvez, Silvia Juliana Martinez. 

       Improvement characteristics of semi-dry coffee 

fermentation with starter yeasts / Silvia Juliana Martinez 

Gelvez. - 2017. 

       56 p. 

 

       Orientador(a): Rosane Freitas Schwan. 

       . 

       Dissertação (mestrado acadêmico) - Universidade Federal 

de Lavras, 2017. 

       Bibliografia. 

 

       1. Qualidade de café. 2. qPCR. 3. Compostos voláteis. I. 

Schwan, Rosane Freitas. . II. Título. 

 

 

Ficha catalográfica elaborada pelo Sistema de Geração de Ficha 

Catalográfica da Biblioteca Universitária da UFLA, com dados 

informados pelo(a) próprio(a) autor(a). 



 

SILVIA JULIANA MARTINEZ GELVEZ 

 

 

IMPROVEMENT CHARACTERISTICS OF SEMI-DRY COFFEE FERMENTATION 

WITH STARTER YEASTS 

 

 

MELHORAMENTO DAS CARACTERÍSTICAS NA FERMENTAÇAO DO CAFÉ 

SEMI-SECO COM LEVEDURAS INICIADORAS 

 

 

Dissertação apresentada à Universidade 

Federal de Lavras, como parte das exigências 

do Programa de Pós-Graduação em 

Microbiologia Agrícola, área de concentração 

em Microbiologia Agrícola, para a obtenção 

do título de Mestre. 

 

APROVADA em 14 de fevereiro de 2017. 

 

Dra. Jussara Coelho   UFES 

 

Dra. Patricia Bernardes   UFES 

 

Dr. Disney Ribeiro Dias   UFLA 

 

 

Dra. Rosane Freitas Schwan 

Orientadora 

 

Dr. Disney Ribeiro Dias 

Coorientador 

 

Dra. Maria Gabriela da Cruz Pedrozo Miguel 

Coorientadora  

 

 

 

 

 

 

LAVRAS – MG 

2017 



 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

To my parents Nubia and Jaime, for their love, advices, mentoring and support 

To my sister’s, Monica and Maria Paula, for their care and love 

Dedico  



 

AGRADECIMENTOS 

 

To God, which guide me through this adventurous road and health. 

To my best and dearest advisor Dr. Rosane Freitas Schwan, for her orientation, 

teaching, trust and opportunity. 

To the Universidade Federal de Lavras (UFLA) and Programa de Microbiologia 

Agricola, for the opportunity to use their instalations during my master project. 

To the Brazilian agencies Conselho Nacional de Desenvolvimento Científico e 

Tecnológico of Brasil (CNPQ), Fundação de Amparo a Pesquisa do Estado de Minas Gerais 

(FAPEMIG), and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES).  

To Gabi, for her guidance and help during the execution of my project and the writing 

of the article. 

To all technicians and lab friends, for their help, knowledge and laughs, especially to 

Ana Paula, Cidinha and Rose. 

To all the people I left without mentioning but where a part of my life and in which 

during this phase made things easy and possible. 

 

Many thanks! 

 

 

 

 

 

 

 

 

 

 

  



RESUMO 

 

Brasil é conhecido mundialmente por ser um dos maiores produtores e exportadores de café. 

Antes de produzir a bebida, os grãos de café devem passar por método de processamento, que 

pode ser o seco, semisseco e o úmido. A escolha do tipo de processamento vai depender do 

produtor ou do agricultor. Durante o processo, os frutos ou grãos são fermentados 

espontaneamente durante a secagem até atingir a umidade de, aproximadamente, 11%. 

Algumas vezes, as culturas iniciadoras são utilizadas para ajudar na fermentação, resultando 

em cafés especiais. Os microrganismos naturais do café envolvem leveduras, fungos 

filamentosos, bactérias aeróbicas e facultativas que, ao longo da fermentação, consomem 

açúcares, produzindo ácidos e, consequentemente, baixando o pH. Cafés de alta qualidade 

têm sabor e aroma característicos e a maioria tem compostos benéficos para a saúde. No 

presente estudo, o objetivo foi avaliar o comportamento de três leveduras iniciadoras 

(Saccharomyces cerevisiae CCMA 0543, Candida parapsilosis CCMA 0544 e Torulaspora 

delbrueckii CCMA 0684) inoculadas em café processado via semisseca, implementando dois 

métodos de inoculação, inoculação direta e inoculação em balde. A população total de 

leveduras, bactérias lácticas e mesófilas foi avaliada por plaqueamento. Além disso, a 

população das culturas iniciadoras foi monitorada por PCR em tempo real (qPCR). Os 

metabólitos produzidos e consumidos nos grãos verdes e torrados durante a fermentação 

foram avaliados por cromatografia líquida (HPLC) e cromatografia gasosa (GC-MS). 

Finalmente, foi feito o teste de xícara para avaliação da bebida final. O resultado de contagem 

em placa mostrou que o método em balde manteve alta população de leveduras, bactérias 

lácticas e mesófilas no final da fermentação (secagem). A sacarose foi consumida em todos os 

tratamentos avaliados. Os ácidos cítrico e succínico foram detectados durante todos os tempos 

de fermentação. Após torra, a média de ácido clorogênico foi maior no método em balde, e o 

mesmo foi observado para as concentrações de trigonelina e cafeína, exceto no tratamento 

com T. delbrueckii CCMA0684. Nos grãos verdes, álcoois e ácidos, e nos grãos torrados, 

pirazinas e piridinas foram os principais compostos voláteis detectados por GC-MS. As notas 

do teste de xícara foram acima de 80 para os dois métodos de inoculação, demonstrando que a 

inoculação teve efeito positivo e produziu bebidas de boa qualidade.  

 

Palavras-chave:  Qualidade de café. qPCR.  Compostos voláteis.  Fermentação de café. 

 

 

 

 

 

 

 

 

 

  



 

ABSTRACT 

 

Brazil is known worldwide for being one of the highest producers and exporters of coffee. 

Before beverage production, beans go through different processing methods: dry, semi-dry 

and wet, choosing one will depend on the producer or farmer. During the process, cherries or 

beans are fermented spontaneously while drying until reaching approximately 11% of 

moisture. Sometimes starter cultures are used to help fermentation that results in special 

coffees. Natural organisms of coffee involve yeasts, filamentous fungi, aerobic bacteria and 

facultative bacteria that consume sugars, produce acids that subsequently low pH. Coffees of 

high qualities have characteristic aromas and flavors and most have beneficial health 

compounds. This work aimed at evaluating the behavior of three-yeast previously tested 

starters (Saccharomyces cerevisiae CCMA 0543, Candida parapsilosis CCMA 0544 and 

Torulaspora delbrueckii CCMA 0684) inoculated in coffee process via semi-dry method, 

implementing two inoculation methods: direct inoculation and bucket inoculation. Total 

population of yeast, lactic acid and mesophilic bacteria was evaluated by plating. Posteriorly, 

the population of starters was monitored by real-time polymerase chain reaction (qPCR). 

Metabolites consumed and produced during fermentation in both inoculation methods were 

evaluated using liquid chromatography (HPLC) and gas chromatography (GC-MS) for green 

and roasted beans. Finally, a cup test was carried out as sensorial analysis. As a result, plate 

counting showed that bucket method maintained a higher population of yeast, lactic acid and 

mesophilic bacteria at the end of fermentation (drying). Sucrose was consumed in all tested 

treatments. Citric and succinic acid were detected during all fermentation processes. After 

roasting, the average levels of chlorogenic acid were higher for the bucket method and the 

same for trigonelline and caffeine concentration, except for the T. delbrueckii CCMA0684 

assay. Group of acids and alcohols in green beans and pyrazines and pyridines in roasted 

beans were the main volatile compounds. Scores from both inoculation methods in the cup 

test were above 80, proving that the inoculation had a positive effect and produced beverages 

of good quality.  

 

Keywords:  Coffee quality.  qPCR.  Volatile compounds.  Coffee fermentation. 

 

 

 

 

 

 

 

 

 

 

 

 

 

  



RESUMEN 

 

Mundialmente Brasil es conocido por ser uno de los mayores productores y exportadores de 

café. Para producir la bebida, sus granos pasan por diferentes métodos de procesamiento, 

pudiendo ser estos seco, semi-seco y húmedo. Su selección dependerá del productor o 

agricultor. Durante el proceso, los frutos o granos son fermentados espontáneamente mientras 

son secados hasta alcanzar aproximadamente un 11% de humedad. Algunas veces se utilizan 

culturas iniciadoras para ayudar en la fermentación, resultando en cafés especiales.  Los 

organismos naturales del café comprenden levaduras, hongos filamentosos, bacterias 

aeróbicas y facultativas que a lo largo de la fermentación consumen azucares y producen 

ácidos que consecuentemente bajan el pH. Los cafés de alta calidad tienen sabores y aromas 

característicos, y la mayoría contienen compuestos benéficos para la salud. En este estudio, el 

objetivo fue evaluar el comportamiento de tres levaduras iniciadoras previamente ensayadas 

(Saccharomyces cerevisiae CCMA 0543, Candida parapsilosis CCMA 0544 y Torulaspora 

delbrueckii CCMA 0684) e inoculadas en un café procesado por el método semi-seco, 

implementando dos métodos de inoculación: inoculación directa e inoculación en balde. La 

población total de levaduras, bacterias lácticas y mesófilas fue evaluada por cultivo en placa. 

Luego, la población de las levaduras iniciadoras fue monitoreada por PCR en tiempo-real 

(qPCR). Los metabolitos consumidos y producidos durante la fermentación en los granos 

verdes y tostados para ambos métodos de inoculación, fueron evaluados utilizando 

cromatografía liquida (HPLC) y cromatografía gaseosa (GC-MS). Finalmente, para el análisis 

sensorial se realizó una prueba de taza. Como resultado del recuento en placa, se observó que 

al final de la fermentación (secado) el método en balde mantuvo una mayor población de 

levaduras, bacterias lácticas y mesófilas. En todos los tratamientos evaluados la sacarosa fue 

consumida. Los ácidos cítrico y succínico fueron detectados en todos los tiempos de 

fermentación. Después de la torrefacción, el promedio de ácido clorogênico fue mayor en el 

método en balde, siendo observado de igual manera para las concentraciones de trigonelina y 

cafeína, excepto en el tratamiento con T. delbrueckii CCMA0684. En los granos verdes los 

ácidos y alcoholes, y en los granos tostados las pirazinas y piridinas, fueron los principales 

compuestos volátiles. Los valores para ambos métodos en la prueba de taza fueron mayores a 

80, demostrando que la inoculación posiblemente tuvo un efecto positivo y se produjeron 

bebidas de buena calidad.  

 

Palabras-clave: Calidad de café. qPCR. Compuestos volátiles. Fermentación de café. 

 

  



 

SUMÁRIO 

 

  PRIMEIRA PARTE ................................................................................................... 10 
1   INTRODUCTION ...................................................................................................... 10 
2   LITERATURE REVIEW .......................................................................................... 12 
2.1   General characteristics of coffee ............................................................................... 12 
2.2   Coffee production in Brazil ....................................................................................... 12 
2.3   Coffee processes .......................................................................................................... 14 
2.3.1  Pulp and mucilage ...................................................................................................... 16 
2.3.2  Beans composition ...................................................................................................... 17 
2.4   Fermentation of coffee ............................................................................................... 19 
2.4.1  Diversity and roles of microorganisms ..................................................................... 19 
2.5   Coffee flavor and volatile compounds ...................................................................... 21 
  REFERENCES ........................................................................................................... 24 
  SEGUNDA PARTE – ARTIGO ............................................................................... 27 

 ARTIGO 1 - Different inoculation methods for semi-dry processed coffee 

using yeasts as started cultures .................................................................................. 27 
 
 

 

 

 

 

 

 

 

 

 

 

 



10 

 

PRIMEIRA PARTE 

 

1 INTRODUCTION 

 

Coffee is a dark brown-slightly bitter beverage made from ground and roasted beans 

that is served hot or iced. It is also considered one of the most popular beverages around the 

world because of its valuable properties and beans composition (BALLESTEROS; 

TEIXEIRA; MUSSATTO, 2014). The main production is located in South and Central 

America, The Caribbean, Africa and Asia (RESTUCCIA et al., 2015), being Brazil, Paraguay, 

Venezuela, Colombia, Indonesia, Ethiopia, India and Mexico the main coffee producers 

(LEONG et al., 2014). Brazil, apart from being one of the largest producers in the world, is 

also the largest exporter of coffee. Total production for crop in 2016 was around 151,624 

bags; for the commercial Arabica specie, the production was of approximately 95,204 bags 

compared to the Robusta specie with 56.419 thousand bags (INTERNATIONAL COFFEE 

ORGANIZATION - ICO, 2017). 

After harvesting, coffee can be processed by different methods:: dry, wet and semi-dry 

process (SILVA et al., 2013). In the dry method, the whole cherry is dried under the sun or in 

a mechanical dryer after being washed. During this process, the most important step is the 

drying operation because it determines the quality of coffee. In the wet method, which 

requires a big amount of water, the pulp is eliminated by a pulper and beans are placed in 

fermentation tanks and submerged under water for 48 h to remove mucilage. The semi-dry or 

pulped natural method is an intermediate of both dry and wet; at the beginning, cherries are 

depulped by a wet mechanical process and beans are dried with its mucilage in patios 

(DUARTE; PEREIRA; FARAH, 2010; SILVA et al., 2013). 

Therefore, the type of process should be carefully chosen because sometimes it 

changes concentrations of several biochemical compounds such as acetic acid, lactic acid, 

caffeine, chlorogenic acid and others. Furthermore, it also contributes to the formation of new 

compounds, either give coffee its flavor. In addition, a few of non-volatile or volatile 

compounds have health benefits such as, antioxidant activity, antidiabetic activity and reduce 

cholesterol levels (BELGUIDOUM et al., 2014). In this sense, determination of these 

compounds before and after processing becomes important because they are influenced by 

coffee variety, geographical origin, roasting conditions and microbiota during fermentation. 

Fermentation of coffee fruits occurs naturally regardless the processing method. The 

objective of fermentation is to remove the mucilaginous layer, which is rich in 



11 

 

polysaccharides (pectins), and to reduce the water content (SILVA et al., 2013). Most 

microorganisms are responsible for fermentation, and these are indigenous species that 

originate as natural microbiota of coffee, such as bacteria (positive and negative), yeast and 

filamentous fungi. Population of each microbial group varies and depends on the processing 

method and extent of water loss (SILVA et al., 2000). Species of yeasts most commonly 

found during coffee fermentation are Pichia guilliermondi, Pichia anomala, Hanseniaspora 

uvarum, Saccharomyces cerevisiae, Debaryomyces hansenii and Torulaspora delbrueckii. 

Bacteria of the genera Erwinia, Klebsiella, Aerobacter, Escherichia and Bacillus can also be 

found (AVALLONE et al., 2001; SILVA et al., 2008, 2013). It has been described that 

fermented coffee has a better quality and its control minimizes the production of poor 

beverages. The incomplete fermentation results in a second fermentation during drying and 

storage, and over-fermentation or bad fermentation produces butyric and propionic acids 

(AVALLONE et al., 2002). 

One option to optimize and enhance coffee fermentation is the usage of starter 

cultures. Besides this advantage, they can also prevent the growth of filamentous fungi that 

produces ochratoxins (EVANGELISTA et al., 2014; SILVA et al., 2013). The 

implementation of initial cultures can increase beverage value without increasing the cost. 

Moreover, the selection of microorganisms should be based on pectinase production, acidic 

and other compounds since they interfere in the final product (SILVA et al., 2013). Yeasts are 

a good example of having antagonistic effects against fungi and a strong pectinolytic activity, 

such as S. cerevisiae (MASOUD et al., 2004); in other words, they are potential tools. 

The behavior of such starter cultures during fermentation and other present microbiota 

can be trace with culture dependent and independent methods since both can give a complete 

scene of the microbial diversity.This means that in some cases the low population of species 

considered viable but non-culturable (VBNC) does not always growth in agar and are 

identified by molecular methods (VILELA et al., 2010). The qPCR methods have showed to 

be useful for quantitative analysis of microorganisms in food fermentations (BRINKMAN et 

al., 2003), for example in the recent research of semi-dry coffee by Ribeiro et al. (2017) and 

other foods such as a cocoa fermentation in Batista et al. (2015). 

In summary, this study aims to evaluate the dynamic behavior of inoculated yeasts 

implemented in semi-dry coffee fermentation by qPCR, metabolites production by HPLC and 

GC-MS and cupping test analysis of the beverage.  

 

  



12 

 

2 LITERATURE REVIEW  

 

2.1 General characteristics of coffee  

 

The botanist Linnaeus first described the coffee specie Coffea arabica L. in 1753.  The 

genus Coffea L. belongs to the family Rubiaceae subfamily Ixoroideae, tribe Coffeeae DC. 

The genus Coffea L. currently comprises 104 species classified into the two subgenera: Coffea 

subgen. Coffea with 95 species and Coffea subgen. Baracoffea with nine species (DAVIS et 

al., 2006). 

Three species used for beverage coffee production belong to Coffea subgen. Coffea: C. 

Arabica L. (arabica coffee), C. canephora Pierre ex A. Froehner (robusta coffee) and C. 

liberica Bull ex Hiern (liberica or excels coffee). C. arabica L. and C. canephora account for 

almost all the coffee produced and consumed worldwide (DAVIS et al., 2006, 2007). 

In terms of natural origin, genus Coffea L. species was exclusively found in Africa, 

Madagascar and the Madagascar Islands. The cultivated species C. arabica L. probably 

originated from the west side of the Great Rift Valley in southern Ethiopia, where 

subspontaneous population still grows. Coffee was first explored by the Arabians who may 

have introduced plants from Ethiopia to Yemen as early as 575. Two varieties of arabica 

coffee, called Typica and Bourbon, spread from Yemen (SAKIYAMA; GAVA, 2015). 

 

2.2 Coffee production in Brazil 

 

Brazil is so far the most important arabica coffee producer in the world and grows 

cultivars essentially derived from Typica and Bourbon. Cultivation of arabica coffee in Brazil 

began with the introduction of the first seeds and seedlings from Guyana in 1727, being first 

introduced to the Typica cultivar. The second arabica coffee introduced into Brazil was the 

Bourbon Vermelho cultivar from the Reunion Island in 1852, since this cultivar had greater 

productivity than the original Typica cultivar. Natural mutations in the Typica cultivar 

produced the Amarelo de Botucatu cultivar, in 1871. The third arabica coffee introduced was 

the Sumatra cultivar (cultivar of Typica variety) from the Sumatra Island in 1896, which 

presented good vegetative vigor, beverage quality and its productivity was below 

expectations. Selection of natural hybrid between the cultivars Bourbon Vermelho and 

Amarelo de Botucatu originated the Bourbon Amarelo cultivar, in 1930. Dwarf cultivars 

Caturra Vermelho and Caturra Amarelo were selected from Bourbon Vermelho, in 1937 

(SAKIYAMA; GAVA, 2015). 



13 

 

Nowadays, the two most important cultivars grown in Brazil are Mundo Novo and 

Catuaí (Catuaí Vermelho and Catuaí Amarelo). Natural hybridization between Sumatra with 

Bourbon Vermelho in 1943 generated the Mundo Novo cultivar. The hybridization between 

Mundo Novo and Caturra Amarelo in 1949 originatedthe Catuaí Amarelo and Catuaí 

Vermelho cultivars. The Catuaí Amarelo cultivar was backcrossed with Mundo Novo in order 

to improve the vegetative vigor and gave the following new dwarf cultivars: Rubi, Topázio, 

Ouro verde, Ouro Amarelo, Ouro Bronze and Travessia (SAKIYAMA; GAVA, 2015). 

In Brazil, the main states producing coffee are Minas Gerais, Espírito Santo, São 

Paulo, Bahia, Paraná, Rondônia and Goiás (Figure 1), with 98.6 % of total production. 

Arabica coffee is the most produced in all the country (75.1% out of the total). The Minas 

Gerais state cultivates approximately 98.87% of specie arabica in their growing areas 

(COMPANHIA NACIONAL DE ABASTECIMENTO - CONAB, 2013). 

 

Figure 1 - Map of main states producing coffee in Brazil.  

 
Source: CONAB (2013). 



14 

 

2.3 Coffee processes  

 

The processing of coffee initiates after harvesting the coffee cherries, followed by the 

removal of both the pulp and hulk, which can be by either dry, wet or semi-dry methods 

(MUSSATTO et al., 2011) (Figure 2). 

Coffee that is processed by the wet method is called washed or parchment coffee; it 

requires reliable pulping equipment and adequate supply of clean water (MURTHY; NAIDU, 

2012). In this process, the objective is, before drying, to remove both the pulp and the 

mucilage, covering the seeds in an environmental friendly way. For initiation, only ripe 

cherries will be harvest and used, and depending on the product harvested, separation may 

vary (mechanically or not); when mechanically, cherries may be pulped in a water flow. Then, 

mucilage is removed by fermentation (can be placed in tanks), followed by washing or 

machines processes (BRANDO; BRANDO, 2015). 

 

  



15 

 

Figure 2 - Coffee processing methods. 

 
Source: Dias et al. (2015), modified. 

 

The dry method, also known as natural processing, is implemented in countries such 

as Brazil and Ethiopia that have extended periods of sunshine (LEE et al., 2015). For this 

process, coffee cherries are either handpicked or machine-harvested when most of them are 

matured. Consequently, the levels of maturity are not consistent among the harvested coffee 

cherries. Following harvesting, coffee cherries are then put on patios and left to dry under the 

sun in layers of approximately 5-8 or 10 cm for 10–25 days, where they are constantly heaped 

and re-spread, until reaching a moisture of 11-12%. Along drying, natural microbial 



16 

 

fermentation takes place and enzymatic hydrolysis leads to a breakdown of the pulp and 

mucilage within the coffee cherry, leaving it intact (BRANDO; BRANDO, 2015). 

On the other hand, the semi-dry or pulped natural method is a variation of wet and dry 

processes, which started to be used in Brazil in the early 1990s. This process aims separation 

of ripe and unripe cherries by flotation using water. Following, ripe cherries are pulped and 

seeds are dried with the mucilage that was not previously removed surrounding the 

parchment, fermentation process occurs directly under the sun. Coffee seeds resulting from 

this method are called pulped natural coffees (BRANDO; BRANDO, 2015). 

 

2.3.1 Pulp and mucilage 

   

Before describing two important parts of coffee cherries, the pulp and mucilage, 

Figure 3 shows their position on a sectioned cherry. Coffee pulp is the first by-product 

obtained during processing and represents 29% dry-weight of the whole cherry. Coffee pulp is 

rich in carbohydrates, proteins and minerals (especially potassium) and it also contains 

appreciable amounts of tannins, polyphenols and caffeine. The organic components present in 

coffee pulp (dry weight) includes tannins 1.80 to 8.56%, total pectic substances 6.5%, 

reducing sugars 12.4%, non-reducing sugars 2.0%, caffeine 1.3%, chlorogenic acid 2.6%, and 

total caffeic acid 1.6% (MURTHY; NAIDU, 2012). 

 

Figure 3 - Section of a coffee cherry. 

 

Source: Murthy and Naidu (2012) and Narita and Inouye (2014). 

 

The quantity of mucilage (Pectic adhesive layer) in coffee depends on the ripening 

stage, moisture and size of fruits. A green bean has near 1.3% of mucilage, least ripe over 

8.4% and overripe between 1 to 23%. Specie Arabica has around 44% of pulp and 11% of 



17 

 

mucilage (QUINTERO, 2012). The mucilage is a gelatinous, translucent and sweet substance, 

which is richer in sugar in arabica than in robusta coffee (SAKIYAMA; GAVA, 2015); it is a 

source of fermentable carbohydrates for coffee fermentation, minerals, such as Ca, K and P, 

and amino acids (FOOD AND AGRICULTURE ORGANIZATION OF THE UNITED 

NATIONS - FAO, 2006). In addition, it is a layer that contains 84.2% of water, 8.9% 

proteins, 4.1% sugar, 0.91% pectic substances and 0.7% ash (LEE et al., 2015). 

 

2.3.2 Beans composition 

 

Chemical composition of green coffee from two species is shown in Table 1. Caffeine 

is the most known component of coffee beans (also thermostable). In raw Arabica coffee, 

caffeine can be found in values varying between 0.8% and 1.4% (w/w), while for the Robusta 

variety these values vary between 1.7% and 4.0% (w/w). However, coffee bean is constituted 

by several other components, including cellulose, minerals, sugars, lipids, tannin, and 

polyphenols. Minerals include potassium, magnesium, calcium, sodium, iron, manganese, 

rubidium, zinc, copper, strontium, chromium, vanadium, barium, nickel, cobalt, lead, 

molybdenum, titanium, and cadmium. Among the sugars, sucrose, glucose, fructose, 

arabinose, galactose, and mannose are present. Several amino acids such as alanine, arginine, 

asparagine, cysteine, glutamic acid, glycine, histidine, isoleucine, leucine, lysine, methionine, 

phenylalanine, proline, serine, threonine, tyrosine, and valine can also be found in these. 

Additionally, coffee beans contain vitamin of complex B, the niacin (vitamin B3 and PP) and 

chlorogenic acid (MUSSATTO et al., 2011). 

The coffee bean is rich in polysaccharides mainly consisting of mannans or 

galactomannans, type II arabinogalactans, and cellulose (CAMPOS-VEGA et al., 2015). 

Furthermore, amino acids are considered the main components, contributing directly to the 

typical flavor and aroma left in roasted beans. Hence, levels of amino acids vary according to 

the processing method and can accumulate under stress conditions (such as osmotic pressure) 

(DIAS et al., 2012). 



 

 

1
8
 

Table 1 - Chemical composition of green coffee. 

Components Arabicaa Robustaa Constituents 

Soluble carbohydrates 9 - 12.5 6 - 11.5  

Monosaccharides 0.2 - 0.5 Fructose, glucose, galactose, arabinose (traces) 

Oligosaccharides 6 - 9 3 - 7 Sucrose (>90%), raffinose (0–0.9%), stachyose (0–0.13%) 

Polysaccharides 3 - 4 

Polymers of galactose (55–65%), mannose (10–20%), arabinose 

(20–35%), glucose (0–2%) 

Insoluble polysaccharides 46 - 53 34 - 44  

Hemicelluloses 5 - 10 3 - 4  

Cellulose,  β(1-4) mannan 41 - 43 32 - 40  

Acids and phenols    

Volatile acids 0.1  

Nonvolatile aliphatic acids 2 - 2.9 1.3 - 2.2 Citric acid, malic acid, quinic acid 

Chlorogenic acid 6.7 - 9.2 7.1 - 12.1 Mono-, dicaffeoyl-, and feruloylquinic acid 

Lignin 1 - 3  

Lipids 15 - 18 8 - 12  

Wax 0.2 - 0.3  

Oil 7.7 - 17.7 Main fatty acids: 16:0 and 18:2 (9,12) 

N compounds 11 - 15  

Free amino acids 0.2 - 0.8 Main amino acids: Glu, Asp, Asp-NH2 

Proteins 8.5 - 12  

Caffeine 0.8 - 1.4 1.7 - 4.0 Traces of theobromine and theophylline 

Trigonelline 0.6 - 1.2 0.3 - 0.9  

Minerals 3 - 5.4  
a Value % dry-weight basis. Source: Mussatto et al. (2011). 

 



19 

 

2.4 Fermentation of coffee 

 

In all processing methods, the objective of fermentation is to remove the layer of 

mucilage from the seed to which adheres, while its degraded fruits are simultaneously dried to 

11-12% of moisture. During fermentation, physicochemical changes occur in grains, such as 

reduction in water content, simple sugars are consumed, and aroma and flavor precursors are 

formed. Fermentation is a step that occurs naturally regardless of the processing method 

(SILVA et al., 2013). 

The optimum temperature for fermentation is 30–35 ºC. The coffee masses are stirred 

2–3 times or more during fermentation period, depending on the method of processing. The 

degradation of mucilage takes approximately between 6 to 72 h, either by drying or 

submerged for fermentation depending on the specie of coffee, the inherent concentration of 

pectinolytic enzymes, ambient temperature and elevation (BRANDO; BRANDO, 2015; 

MURTHY; NAIDU, 2012). 

Yeast, aerobic bacteria and facultative bacteria population increase in the first hours of 

fermentation in open systems and without water, as the process move forward pH decreases 

and inhibits growth of certain organisms in the mucilage (QUINTERO, 2012). 

A poor control of fermentation would be reflected in coffee aroma. Over-fermentation 

results in the production of black or ‘‘stinker’’ beans with poor visual and aroma 

characteristics. These beans are commonly associated with fruity, flora, sour and alcoholic 

attributes. Therefore, this shows that there is only a fine margin between the fermentation 

process and the quality of coffee aroma (LEE et al., 2015). 

Optimization of coffee fermentations and production of special and good quality 

beverages can be achieved by using starter cultures. Upon selecting starters certain criteria 

must be used, just as, depending on their capacity to degrade mucilage, maintenance of 

population, metabolite production and others during fermentation drying. However, after 

selecting the appropriate strain, producers/farmers must carefully choose the inoculation 

method they will be using, since the effect can either be positive or turn into something 

expected. 

 

2.4.1 Diversity and roles of microorganisms 

 

Three important roles that microorganism execute during coffee fermentation are the 

production of pectinolytic enzymes to degrade the mucilage and pulp, spoilage of food, and 



20 

 

production of mycotoxins due to poor drying and storage. The microbiota present in coffee 

fruits is complex and diverse and include yeasts, filamentous fungi, and bacteria (SILVA et 

al., 2000). 

In past studies, several species of microorganisms have been isolated from the 

fermentation phase of wet-processing such as aerobic bacteria - Klebsiella ozaenae, K. 

oxytoca, Erwinia herbicola and E. dissolvens and lactic acid bacteria such as - Leuconostoc 

mesenteroides, Lactobacillus brevis are the bacterial species that were isolated from the 

fermentation process. Yeast species such as Kloeckera apis apicualata, Candida 

guilliermondii, Cryptococcus albidus, C. laurentii, Pichia kluyveri, P. anomala, 

Hanseniaspora uvarum, Saccharomyces cerevisiae, Debaryomyces hansenii and Torulaspora 

delbrueckii have also been identified (AGATE; BHAT, 1966; AVALLONE et al., 2001; 

EVANGELISTA et al., 2015; MASOUD et al., 2004). 

In Arabica and Robusta coffees processed by the semi-dry process, genera of Bacillus, 

Flavobacterium, Serratia, Pseudomonas and Lactic acid bacteria were found. Yeasts is the 

second group of microorganism in depulped fruit coffee in this process and some identified 

species are S. cerevisiae, Axula spp., Candida ernobii, C. membranifaciens, Kluyveromyces 

spp., H. uvarum and Torulaspora delbrueckii (SILVA, 2015). 

Previous studies of Silva et al. (2013) showed that indigenous and nonindigenous 

bacteria and yeast exhibited pectinolytic activity. Part of the potential starter cultures that 

were identified includes yeast cultures such as Saccharomyces, Pichia and Candida, which 

showed higher pectinolytic enzyme activity for efficient mucilage degradation during 

fermentation. Strains were also evaluated for their ability to enhance the quality of coffee 

fermentation in wet, dry and semi-dry processing and produced coffees with distinctive 

flavor. It was found that the employment of a selected culture for fermentation during coffee 

processing enhanced the quality of coffee aroma compared to coffees produced from 

fermentation involving indigenous microbiota (EVANGELISTA et al., 2013, 2014). 

A culture-independent method such as quantitative real-time PCR (QPCR) is known 

as a fast, advantageous and sensible technique that is used for quantifying culture starters 

(BATISTA et al., 2015). Several recent studies in food fermentation, such as in Batista et al. 

(2015) with cocoa fermentations and Ribeiro et al. (2017) with coffee fermentation used this 

technique to monitor their yeast starters and showed promising and reliable results. 

 

 



21 

 

2.5 Coffee flavor and volatile compounds  

 

Flavor of coffee is complex and it develops in various stages of processing and cup 

preparation as shown in Figure 4. In dry processing, a ‘hard’ coffee with a medicinal flavor is 

produced while wet processing yields a better quality coffee with less body, higher acidity and 

more aroma than the dry processing. The semi-dry (semi-washed or pulped natural) offers a 

coffee with intermediate body. Furthermore, roasting has the most influence on coffee flavor 

(Figure 5), with temperatures varying between 180 ºC to 240 ºC for 8 to 15 min. During 

roasting, endothermic and exothermic processes begin from heat being transferred to the bean 

through hot gases or contact with the metal surface of the coffee roaster, which reduces water 

content of the coffee beans and causes puffing and cooling to produce desirable 

characteristics. The impact of roasting on flavor comes from the degradation and formation or 

release of numerous chemical compounds through Maillard reactions, Strecker degradation, 

break down of amino acids, degradation of trigonelline, quinic acid, pigments, lipids and 

interaction between intermediate products (SUNARHARUM; WILLIAMS; SMYTH, 2014). 

 

Figure 4 - Factors that affect coffee flavor. 

 

Source: Sunarharum, Williams and Smyth (2014). 

 

 



22 

 

Tannins are the main phenolic compounds in coffee pulp and in the seed, phenolic 

compounds are present predominantly as a family of esters formed between certain 

hydroxycinnamic acids and quinic acid, collectively known as chlorogenic acids (CGA). 

Other phenolic compounds, such as tannins, lignans and anthocyanins are also present in 

coffee seeds in smaller amounts. CGA, which are present in high concentrations in green 

coffee seeds (up to 14 %), have a marked influence in determining coffee quality and play an 

important role in the formation of coffee flavor (FARAH; DONANGELO, 2006). 

A variety of other volatile compounds comprise several chemical classes including 

alcohols, aldehydes, ketones, carboxyclic acids, pyrazines, pyrroles, pyridines, sulfur 

compounds, furans, furanones, phenols, oxazoles among others. These compounds vary 

significantly in concentration and sensory potency which makes coffee flavor extremely 

complex, and explains why different coffee types may exhibit such diverse, unique and 

specific flavors (SUNARHARUM; WILLIAMS; SMYTH, 2014). 

 

Figure 5 - Main classes of volatile compounds formed during roasting from non-volatile 

precursors in the green beans. 

 

Source: Yeretzian et al. (2002). 

 



23 

 

Studies have shown that gas chromatography–mass spectrometry analysis (GC–MS) 

of volatile components in coffee together with principal component analysis (PCA) had 

allowed discrimination of Arabica/Robusta blends (AGRESTI et al., 2008) and analysis of 

compounds produced in cup (BRESSANELLO et al., 2017). 

 



24 

 

REFERENCES 

 

AGATE, A. D.; BHAT, J. V. Role of pectinolytic yeasts in the degradation of mucilage layer 

of Coffea robusta cherries. Applied Microbiology, Baltimore, v. 14, n. 2, p. 256-260, Mar. 

1966. 

 

AGRESTI, P. D. C. M. et al. Discrimination between defective and non-defective Brazilian 

coffee beans by their volatile profile. Food Chemistry, London, v. 106, n. 2, p. 787-796, 

2008. 

 

AVALLONE, S. et al. Involvement of pectolytic micro-organisms in coffee fermentation. 

International Journal of Food Science and Technology, New Jersey, v. 37, n. 2, p. 191-

198, Feb. 2002. 

 

AVALLONE, S. et al. Microbiological and Biochemical study of coffee fermentation. 

Current Microbiology, New York, v. 42, p. 252-256, Apr. 2001. 

 

BALLESTEROS, L. F.; TEIXEIRA, J. A.; MUSSATTO, S. I. Chemical, functional, and 

structural properties of spent coffee grounds and coffee Silverskin. Food Bioprocess 

Technology, New York, v. 7, p. 3493-3503, Dec. 2014. 

 

BATISTA, N. N. et al. Dynamic behavior of Saccharomyces cerevisiae, Pichia kluyveri and 

Hanseniaspora uvarum during spontaneous and inoculated cocoa fermentations and their 

effect on sensory characteristics of chocolate. Food Science and Technology, Amsterdam, v. 

63, p. 221-227, Sept. 2015. 

 

BELGUIDOUM, K. et al. HPLC coupled to UV - vis detection for quantitative determination 

of phenolic compounds and caffeine in different brands of coffee in the Algerian market. 

Journal of the Taiwan Institute of Chemical Engineers, Taiwan, v. 45, p. 1314-1320, July 

2014. 

 

BRANDO, C. H. J.; BRANDO, M. F. Methods of coffee fermentation and drying. In: 

SCHWAN, R.; FLEET, G. (Ed.). Cocoa and coffee fermentations. Boca Raton: CRC, 2015. 

chap. 10. 

 

BRESSANELLO, D. et al. Coffee aroma: chemometric comparison of the chemical 

information provided by three different samplings combined with GC-MS to describe the 

sensory properties in cup. Food Chemistry, London, v. 214, p. 218-226, Jan. 2017. 

 

BRINKMAN, N. E. et al. Evaluation of a rapid, quantitative real-time PCR method for 

enumeration of pathogenic Candida cells in water. Applied and Environmental 

Microbiology, Washington, v. 69, n. 3, p. 1775-1782, 2003. 

 

CAMPOS-VEGA, R. et al. Spent coffee grounds: a review on current research and future 

prospects. Trends in Food Science & Technology, Amsterdam, v. 45, p. 24-36, Sept. 2015. 

 

COMPANHIA NACIONAL DE ABASTECIMENTO. Acompanhamento da safra 

brasileira: café. Brasília, DF, 2013. Available from: <http://www.conab.gov.br/>. Access in: 

10 Oct. 2015. 

 



25 

 

DAVIS, A. P. et al. An annotated taxonomic conspectus of the genus Coffea (Rubiaceae). 

Botanical Journal of the Linnean Society, London, v. 152, p. 465-512, 2006. 

 

DAVIS, A. P. et al. Searching for the relatives of Coffea (Rubiaceae, Ixoroideae): the 

circumscription and phylogeny of Coffeeae based on plastid sequence data and morphology. 

American Journal of Botany, Saint Louis, v. 94, p. 313-329, Mar. 2007. 

 

DIAS, D. et al. Management and utilization of wastes from coffee processing. In: SCHWAN, 

R.; FLEET, G. (Ed.). Cocoa and coffee fermentations. Boca Raton: CRC, 2015. chap. 15. 

 

DIAS, E. C. et al. Amino acid profiles in unripe Arabica coffee fruits processed using wet and 

dry methods. European Food Research and Technology, Berlin, v. 234, n. 1, p. 25-32, Jan. 

2012. 

 

DUARTE, G. S.; PEREIRA, A. A.; FARAH, A. Chlorogenic acids and other relevant 

compounds in Brazilian coffees processed by semi-dry and wet post-harvesting methods. 

Food Chemistry, London, v. 118, n. 3, p. 851-855, 2010. 

 

EVANGELISTA, S. R. et al. Improvement of coffee beverage quality by using selected yeasts 

strains during the fermentation in dry process. Food Research International, Burlington, v. 

61, p. 183-195, July 2013. 

 

EVANGELISTA, S. R. et al. Inoculation of starter cultures in a semi-dry coffee (Coffea 

arabica) fermentation process. Food Microbiology, Amsterdam, v. 44, p. 87-95, Dec. 2014. 

 

EVANGELISTA, S. R. et al. Microbiological diversity associated with the spontaneous wet 

method of coffee fermentation. International Journal of Food Microbiology, Rome, v. 210, 

p. 102-112, Oct. 2015. 

 

FARAH, A.; DONANGELO, C. M. Phenolic compounds in coffee. Brazilian Journal of 

Plant Physiology, Campos dos Goytacazes, v. 18, n. 1, p. 23-36, Jan./Mar. 2006. 

 

FOOD AND AGRICULTURE ORGANIZATION OF THE UNITED NATIONS. Good 

hygiene practices along the coffee chain. Rome, 2006. Available from: 

<http://www.fao.org/home/en/>. Access in: 10 Oct. 2015. 

 

INTERNATIONAL COFFEE ORGANIZATION. Trade statistics tables. Available from: 

<http://www.ico.org/trade_statistics.asp?section=Statistics>. Access in: 10 Feb. 2017. 

 

LEE, L. W. et al. Coffee fermentation and flavor: an intricate and delicate relationship. Food 

Chemistry, London, v. 185, p. 182-191, Oct. 2015. 

 

LEONG, K. H. et al. Diversity of lactic acid bacteria associated with fresh coffee cherries in 

Taiwan. Current Microbiology, New York, v. 68, n. 5, p. 440-447, 2014. 

 

MASOUD, W. et al. Yeast involved in fermentation of Coffea arabica in East Africa 

determined by genotyping and by direct denaturating gradient gel electrophoresis. Yeast, 

Malden, v. 21, n. 7, p. 549-556, Apr. 2004. 

 



26 

 

MURTHY, P. S.; NAIDU, M. M. Sustainable management of coffee industry by-products and 

value addition: a review. Resources, Conservation and Recycling, Amsterdam, v. 66, p. 45-

58, Sept. 2012. 

 

MUSSATTO, S. I. et al. Production, composition, and application of coffee and its industrial 

residues. Food and Bioprocess Technology, New York, v. 4, n. 5, p. 661-672, July 2011. 

 

NARITA, Y.; INOUYE, K. Review on utilization and composition of coffee silverskin. Food 

Research International, Burlington, v. 61, p. 16-22, July 2014. 

 

QUINTERO, G. I. P. Factores, procesos y controles en la fermentación del café. 

Manizales: Avances Técnicos Cenicafé, 2012. 12 p. 

 

RESTUCCIA, D. et al. Brewing effect on levels of biogenic amines in different coffee 

samples as determined by LC-UV. Food Chemistry, London, v. 175, p. 143-150, May 2015. 

 

RIBEIRO, L. S. et al. Behavior of yeast inoculated during semi-dry coffee fermentation and 

the effect on chemical and sensorial properties of the final beverage. Food Research 

International, Burlington, v. 92, p. 26-32, Feb. 2017. 

 

SAKIYAMA, N.; GAVA, M. Botany and production of coffee. In: SCHWAN, R.; FLEET, G. 

(Ed.). Cocoa and coffee fermentations. Boca Raton: CRC, 2015. chap. 9. 

 

SILVA, C. F. Microbial activity during coffee fermentation. In: SCHWAN, R.; FLEET, G. 

(Ed.). Cocoa and coffee fermentations. Boca Raton: CRC, 2015. chap. 11. 

 

SILVA, C. F. et al. Evaluation of a potential starter culture for enhance quality of coffee 

fermentation. World Journal of Microbiology and Biotechnology, Dordrecht, v. 29, n. 2, p. 

235-247, Feb. 2013. 

 

SILVA, C. F. et al. Microbial diversity during maturation and natural processing of coffee 

cherries of Coffea arabica in Brazil. International Journal of Food Microbiology, Rome, v. 

60, n. 2/3, p. 251-260, Sept. 2000. 

 

SILVA, C. F. et al. Succession of bacterial and fungal communities during natural coffee 

(Coffea arabica) fermentation. Food Microbiology, Amsterdam, v. 25, n. 8, p. 951-957, Dec. 

2008. 

 

SUNARHARUM, W. B.; WILLIAMS, D. J.; SMYTH, H. E. Complexity of coffee flavor: a 

compositional and sensory perspective. Food Research International, Burlington, v. 62, p. 

315-325, Aug. 2014. 

 

VILELA, D. M. et al. Molecular ecology and polyphasic characterization of the microbiota 

associated with semi-dry processed coffee (Coffea arabica L.). Food Microbiology, 

Amsterdam, v. 27, n. 8, p. 1128-1135, Dec. 2010. 

 

YERETZIAN, C. et al. From the green bean to the cup of coffee: investigating coffee roasting 

by on-line monitoring of volatiles. European Food Research and Technology, Berlin, v. 

214, n. 2, p. 92-104, Feb. 2002. 

 



27 

 

SEGUNDA PARTE – ARTIGO 

 

ARTIGO 1 - Different inoculation methods for semi-dry processed coffee using yeasts as 

started cultures 

 

Artigo redigido conforme normas da revista International Journal of Food Microbiology 

 

 

 

 

 

 

  



28 

 

ABSTRACT 

Coffee is a bitter beverage consumed worldwide due to its pleasant and stimulating 

properties. This study evaluates the behavior of Saccharomyces (S.) cerevisiae (CCMA 0543), 

Candida (C.) parapsilosis (CCMA 0544), and Torulaspora (T.) delbrueckii (CCMA 0684) as 

starter cultures for semi-dry processed coffee using two inoculation methods: (1) direct 

inoculation of the yeast starter culture solution by spraying the coffee beans (the direct 

method), and (2) inoculation in which coffee beans are fermented with a yeast starter solution 

in polystyrene buckets for 16 hours (the bucket method). The microbial population was 

evaluated by plating and real-time or quantitative polymerase chain reaction (qPCR). The 

microbial metabolic response of both the bucket and direct inoculation methods during 

fermentation was evaluated using high performance liquid chromatography (HPLC) and gas 

chromatography–mass spectrometry (GC-MS). Quantification of sugars, acids, caffeine, 5-

caffeoylquinic acid (5-CGA), and trigonelline were performed in HPLC. For volatile 

compounds, GC-MS and a sensorial test were also carried out. Citric and succinic acids were 

detected throughout the fermentation period. Chlorogenic acid concentration levels were 

higher for the bucket method after roasting. Roasted coffee beans also had a higher caffeine 

concentration, with the exception of the T. delbrueckii (CCMA0684) assay. Acids, pyrazines 

and pyridines were the main volatile compounds in both green and roasted coffee beans. 

Coffee cupping results proved that both inoculation methods scored well in terms of coffee 

quality. Inoculating beans with yeast using the bucket method before sun drying encouraged 

microbial growth and metabolite production, generating a better quality coffee. 

 

Keywords: coffee quality; qPCR, volatile compounds; coffee fermentation 

 

  



29 

 

1. Introduction 

Coffee is a dark-brown, slightly bitter beverage made from ground and roast coffee 

beans, which is widely consumed and considered popular due to its stimulant properties and 

composition (Ballesteros et al., 2014; Contreras-Calderón et al., 2016). South and Central 

America, the Caribbean, Africa and Asia are the main coffee-producing areas (Restuccia et 

al., 2015). Brazil is the major global coffee producer and exporter, with 55.000 thousand bags 

produced in 2016 (ICO, 2017). Within Brazil, Minas Gerais, Espírito Santo and São Paulo are 

the main coffee producing states (CONAB, 2017). 

Coffee beverages are produced mainly by the processing of two species of the Coffea 

(C.) genera: C. canephora (robusta coffee) and C. arabica (Arabica coffee). Arabica coffee is 

recognized to have better quality and flavor and accounts for approximately 65% of the world 

coffee trade and 75% of the coffee trade in Brazil. The largest numbers of Arabica coffee 

cultivars are found in plantations in Brazil’s Mundo Novo and Catuaí (Amarelo and 

Vermelho) (Sakiyama et al., 2015). 

Coffee fermentation occurs naturally, regardless of the process, in order to remove the 

mucilage from seeds and reduce water content, with beans being dried simultaneously until 

moisture content is reduced to between 11 and 12% (Silva et al., 2013). Most microorganisms 

responsible for fermentation are autochthonous (bacteria, yeast, and filamentous fungi). The 

population of each microbial group may vary depending on the processing method and the 

extent of water loss (Silva et al., 2000). During certain processing stages, the contribution of 

acetic acid, lactic acid, caffeine, chlorogenic acids and other compounds improves coffee 

flavor and gives health benefits. Coffee beans possess antioxidant and antidiabetic properties 

that can reduce cholesterol levels (Belguidoum et al., 2014). Measuring these compounds 

before and after processing is important because they are influenced by coffee variety, 

geographical origin, and microbiota during fermentation and roasting. 

The role of yeasts is essential for coffee fermentation, preventing toxigenic 

filamentous fungi growth and boosting the production of pectinolytic enzymes, which help the 

degradation of the coffee mucilage and pulp (Ramos et al., 2010; Silva et al., 2013). Using 

yeasts as starter cultures during coffee fermentation can improve the quality of the end 

product. Silva et al. (2013) and Evangelista et al. (2014b) have reported promising results 

after the direct inoculation of Saccharomyces, Pichia and Candida yeast strains over coffee 

beans. 

The use of molecular and biochemical techniques to evaluate starter cultures has been 

applied to different foods in previous research (Batista et al., 2015; Bressanello et al., 2017; 



30 

 

Evangelista et al., 2014b; Menezes et al., 2016). Starter culture populations have also been 

monitored by qPCR (Gil-Serna et al., 2009; Postollec et al., 2011; Michel et al., 2016). In 

combination, the methods used in GC-MS are also relevant because the process separates 

volatile compounds according to the stated conditions and generates data about unknown 

compounds from fermentation, which may derive from coffee or other samples (Batista et al., 

2015; Bressanello et al., 2017; Evangelista et al., 2014b; Ziółkowskaet al., 2016). 

The aim of this study was to use qPCR to evaluate the dynamic behavior of three 

inoculated yeasts: S. cerevisiae (CCMA 0543), C. parapsilosis (CCMA 0544), and T. 

delbrueckii (CCMA 0684). The study took place in Brazil and used direct and bucket 

inoculation methods to inoculate Catuaí Amarelo depulped cherries during a semi-dry 

process. The effect of the inoculation on the chemical composition of the bean (sugars, acids, 

and volatiles) was analyzed using gas and liquid chromatography. Sensorial analysis (coffee 

cupping) of the resulting beverage was carried out to confirm whether inoculation contributed 

to coffee quality. 

 

2. Material and methods 

2.1. Coffee cherries 

Cherries from the C. arabica variant Catuai Amarelo were obtained from a producing 

farm near Lavras, Minas Gerais, Brazil. The cherries were processed using a semi-dry 

method, which consisted of depulping and washing, leaving only the mucilage and parchment 

(Brando and Brando, 2015). 

 

2.2. Starter cultures 

Three yeast strains—S. cerevisiae (CCMA 0543), C. parapsilosis (CCMA 0544) and 

T. delbrueckii (CCMA 0684)—from the Culture Collection of Agricultural Microbiology 

(CCMA) in Lavras were used as starter cultures (Evangelista et al., 2014b). Each yeast was 

grown in one liter of YEPG medium (20 g/L glucose, 10 g/L yeast extract, 10 g/L peptone 

soy, and pH 3.5) at 28ºC for 24 hours or until they reached a concentration of 109 cells/ml. 

After, cells were washed and diluted in 500 ml of water for inoculation. 

 

2.3. Fermentation and drying 

Two inoculation methods were tested: (1) direct inoculation by spraying the yeast 

solution on depulped cherries in wooden frames (1m x 1m) with plastic nets, and (2) 

inoculation of yeast solution on depulped cherries in polystyrene buckets for 16 hours, with 



31 

 

the fermented beans then being transferred to wooden frames for drying. A total of 8 

treatments containing 10 kg of depulped cherries were carried out: three used direct 

inoculation; three used the bucket method; and two were used as controls (one sample was not 

inoculated but piled in a bucket and the other was not inoculated) (Figure 1). All treatments 

were fermented and sun dried until a moisture content of 10 to 11% was reached (measured 

using a water activity meter, provided by Pawkit). During the fermentation and drying 

process, samples of approximately 100 g were collected in sterile plastic bags until the 

appropriate moisture content was reached (after 0, 16, 64, 112, 256, and 352 hours, 

respectively). 

 

Figure 1 – Flow diagram of the inoculation process and treatment. 
 

 

 

2.4. Microbiological analysis 

2.4.1. Microbial count 

Initial and final samples (10 g) of fermented beans were homogenized in 90 mL 

saline-peptone water (0.1% (v/v) bacteriological peptone (Himedia) and 0.8% (v/v) NaCl 

(Merck, Whitehouse Station, NJ)) in a stomacher at normal speed for 5 minutes, serially 

diluted, and then plated in triplicate. YEPG (g/L glucose 20 (Merck), yeast extract 10 

(Merck), peptone soy 10 (Himedia) and agar 20 (Merck), pH 3.5) with 0.01% (w/v) 

chloramphenicol, MRS agar containing 0.1% (w/v) nystatin, and Nutrient agar were used to 



32 

 

count the total yeast, lactic acid bacteria (LAB), and mesophilic bacteria, respectively. Plates 

were incubated at 30°C for 48 hours. 

 

2.4.2. Quantitative polymerase chain reaction (qPCR) 

Total DNA was extracted from samples at six different fermentation times (after 0, 16, 

64, 112, 256, and 352 hours of drying, respectively)] using the “DNA Purification from 

Tissues” protocol (QIAamp DNA Mini Kit (Qiagen, Hilden, Germany)) in accordance with 

the manufacturer's instructions.  

Specific primers for S. cerevisiae (Díaz et al., 2013), C. parapsilosis (Hays et al., 

2011) and T. delbrueckii (Zott et al., 2010) were used (Table 1). The specificity of each 

primer pair was confirmed by searching in GenBank using BLAST 

(http://www.ncbi.nlm.nih.gov/BLAST/). A qPCR analysis was used to quantify S. cerevisiae, 

C. parapsilosis and T. delbrueckii, as described by Batista et al. (2015). Three independent 

qPCR assays were performed for each treatment. For standard curves, all yeast species were 

cultivated in YEPG broth at 28ºC for 24 hours. The yeast populations were then counted 

using a Neubauer chamber. DNA was extracted using the QIAamp DNA Mini Kit (Qiagen, 

Hilden, Germany) and serially diluted (1:10) from 1010 to 103 cell/mL. Each point of the 

calibration curve was measured in triplicate. 

 

Table 1 – Specific primer used for qPCR analysis. 

Specie 
Primers 

Name Sequence Source 

S. cerevisiae 
SC-5fw 5’AGGAGTGCGGTTCTTTGTAAAG3’ 

Díaz et al., 2013 
SC-3bw 5’ TGAAATGCGAGATTCCCCT3’ 

T. delbrueckii 

PRIMER 1 

Tods L2 
5’ CAAAGTCATCCAAGCCAGC 3’ 

Zott et al., 2010 
PRIMER 2 

Tods R2 
5’TTCTCAAACAATCATGTTTGGTAG3’ 

C. parapsilosis 
SADH-F 5’ GCTGCGGCTTCAACTGATGC 3’ 

Hays et al., 2011 
SADH-R 5’ CTTGGTCACGAGCCTCC3’ 

 

2.5. Chemical analysis 

2.5.1. Carbohydrates, organic acids, glycerol and ethanol by high performance 

liquid chromatography (HPLC) 

The samples were evaluated after 0, 16, 112, 256, and 352 hours of fermentation. For 

extraction, 3 g of the sample was homogenized with 20 ml of Milli-Q water by vortexing for 

10 minutes at room temperature. Then fluids were centrifuged two times at 100 x g for 10 

minutes at 4ºC and were filtered through a 0.2 mm cellulose acetate filter. Later, a 



33 

 

chromatographic analysis was performed, according to Evangelista et al. (2014a). Calibration 

curves were constructed with different concentrations of standard compounds. 

 

2.5.2. Caffeine, chlorogenic acid and trigonelline 

Measurement of the non-volatile compounds (caffeine, chlorogenic acid, and 

trigonelline) in times 0, 112, 352 hours of fermentation and roasting was carried out by 

HPLC, according to Malta and Chagas (2009). Identification and quantitative analysis were 

performed using calibration curves of caffeine, trigonelline and 5-CGA (Sigma-Aldrich, Saint 

Luis, EUA). 

  

2.5.3. Volatile compounds 

Samples of green and roasted beans at both the start and end of the fermentation 

process were used for GC-MS analyses. Coffee samples were macerated with liquid nitrogen 

to extract manual headspace, according to Evangelista et al. (2014a). Volatile compounds 

were identified by comparing their mass spectra to those in the NIST11 Library. In addition, 

an alkane series (C10–C40) was used to calculate the retention index (RI) for each compound 

and compared with RI values found in the literature data. 

 

2.6. Sensorial analysis—coffee cupping 

Green coffee samples were prepared according to the Specialty Coffee Association of 

America (SCAA, 2013). The coffee was roasted in a laboratory roaster (Probatino, Leogap 

model, Brazil) with a capacity of 150 g and was then ground in an electric mill (Pinhalense 

ML-1, Brazil). A panel of three trained coffee tasters with Q-Grader Coffee Certificates 

evaluated the samples. The methodology applied to evaluate coffee was conducted according 

to SCAA standards (SCAA, 2013), which assess ten sensorial attributes: fragrance, flavor, 

aftertaste, acidity, body, uniformity, balance, sweetness, cleanliness, and score. 

 

3. Results 

3.1. Microbiological analyses  

3.1.1. Microbial count by plating 

Total yeast count at the beginning of fermentation varied slightly according to the 

treatment (which strains were used as the starter culture) and inoculation method (direct or 

bucket) (Figure 2). With the direct method, the yeast count was 6.40 log CFU/g for the 

control, 7.25 log CFU/g for the sample inoculated with S. cerevisiae (CCMA 0543), 7.85 log 

https://www.google.com.br/search?biw=1517&bih=735&q=St.+Louis&stick=H4sIAAAAAAAAAOPgE-LUz9U3sLC0SK5U4gAxzcoryrW0spOt9POL0hPzMqsSSzLz81A4VhmpiSmFpYlFJalFxQDMHhGVQwAAAA&sa=X&sqi=2&ved=0ahUKEwjVuu3Zmc_PAhXBHJAKHeysA-QQmxMIggEoATAO
https://www.google.com.br/search?biw=1517&bih=735&q=St.+Louis&stick=H4sIAAAAAAAAAOPgE-LUz9U3sLC0SK5U4gAxzcoryrW0spOt9POL0hPzMqsSSzLz81A4VhmpiSmFpYlFJalFxQDMHhGVQwAAAA&sa=X&sqi=2&ved=0ahUKEwjVuu3Zmc_PAhXBHJAKHeysA-QQmxMIggEoATAO


34 

 

CFU/g for C. parapsilosis (CCMA 0544), and 7.15 log CFU/g for T. delbrueckii (CCMA 

0684). With the bucket method, the yeast count was 6.90 log CFU/g for the control, 7.11 log 

CFU/g for the sample inoculated with S. cerevisiae (CCMA 0543), 7.87 log CFU/g for C. 

parapsilosis (CCMA 0544), and 7.63 log CFU/g for T. delbrueckii (CCMA 0684). However, 

at 352 hours of the fermentation process, there was a greater difference between treatments in 

relation to the yeast count. Treatments inoculated with the yeast S. cerevisiae (CMMA 0543) 

(6.10 log CFU/g) using the bucket method and treatments inoculated with C. parapsilosis 

(CCMA 0544) (5.30 log CFU/g) and T. delbrueckii (CCMA 0684) (5.34 log CFU/g) using the 

direct method had higher yeast counts than other treatments. The final yeast population using 

the direct method differed between the control (3.6 log CFU/g), S. cerevisiae (CCMA 0543) 

(4.0 log CFU/g), C. parapsilosis (CCMA 0544) (5.3 log CFU/g), and T. delbrueckii (CCMA 

0684) (5.3 log CFU/g) treatments. 

 

  



35 

 

Figure 2 - Total population of yeast (A), lactic acid bacteria (B) and mesophilic bacteria (C) 

in coffee beans inoculated with the following yeasts: S. cerevisiae (CCMA 0543) 

( ), T. delbrueckii (CCMA 0684) ( ), C. parapsilosis (CCMA 0544) ( ), and 

Control ( ). 
 

 

 

The bacteria population remained constant at the beginning of fermentation for both 

methods and all treatments (Figures 2B and 2C). However, at 352 hours, LAB and mesophilic 

populations were higher with the bucket method. The LAB population (4.88 log CFU/g) and 

the mesophilic bacteria count (4.87 log CFU/g) were higher with the bucket method when 

inoculated with S. cerevisiae (CCMA 0543). In addition, higher values were found at final 

LAB times (4.30 log CFU/g) and with mesophilic bacteria (4.77 log CFU/g) in direct 

treatments inoculated with C. parapsilosis (CCMA 0544).  



36 

 

3.1.2 Monitoring of starter strains by qPCR 

This study used qPCR to evaluate both control coffee fermentations that were not 

inoculated and fermentations that were inoculated with S. cerevisiae (CCMA 0543), C. 

parapsilosis (CCMA 0544), and T. delbrueckii (CCMA 0684) (Figure 3). The specificity of 

primers was tested in a conventional PCR assay before running. Resulting standard curves 

parameters were observed, with fluctuating values for R2 (from 0.992 to 0.998), slope (from -

3.112 to -3.864), and efficiency (81% to 110%). 

Yeasts used as starter cultures were quantified using inoculated coffee samples and 

control samples (Figure 3). Population of inoculated samples with S. cerevisiae (CCMA 

0543) ranged from 8.0 to 7.6 log cell/g with the bucket method, and from 6.6 to 7.5 log cell/g 

with direct inoculation. However, in control samples, S. cerevisiae (CCMA 0543) behaved 

differently for both methods. With the bucket method, its population was higher for the first 

112 hours and then declined, reaching 4.0 log cell/g at the end of processing. With the direct 

method, the highest S. cerevisiae (CCMA 0543) population was also observed at the start of 

fermentation (5.24 log cell/g), reaching 3.5 log cell/g at the end of processing. 

 

  



37 

 

Figure 3 - Dynamic behavior of S. cerevisiae (A), T. delbrueckii (B) and C. parapsilosis (C) 

during 352 hours of fermentation and drying in Catuaí semi-dry coffee beans with 

inoculation (full line) and control (dashed line), using direct (▲) and bucket (●) 

inoculation methods, measured by qPCR. (Continue) 

   
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

2

3

4

5

6

7

8

9

10

0 60 120 180 240 300 360S
. 
ce

re
vi

si
a

e 
p

o
p

u
la

ti
o

n
 (

L
o

g
 c

el
l/

g
)

Time (h)

2

3

4

5

6

7

8

9

10

0 60 120 180 240 300 360T
. 

d
el

b
ru

ec
ki

i
p

o
p

u
la

ti
o

n
 (

L
o

g
 c

el
l/

g
)

Time (h)



38 

 

Figure 3 - Dynamic behavior of S. cerevisiae (A), T. delbrueckii (B) and C. parapsilosis (C) 

during 352 hours of fermentation and drying in Catuaí semi-dry coffee beans with 

inoculation (full line) and control (dashed line), using direct (▲) and bucket (●) 

inoculation methods, measured by qPCR. (Conclusion) 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

The T. delbrueckii (CCMA 0684) population was similar for the treatments performed 

in the present study, independent of inoculation method (Figure 3B). The yeast population 

remained higher with the bucket than with the direct method, ranging in inoculated assays 

from 8.7 to 8.6 log cell/g and from 8.4 to 8.8 log cell/g for the control at the end of the drying 

process. In relation to C. parapsilosis (CCMA 0544), the inoculation method contributed to 

the persistence of the species during coffee processing. Inoculated treatments had a larger 

population than the control (for both bucket and direct inoculation) (Figure 3C). The 

population of this yeast varied from 10 to 9 log cell/g from bucket inoculation, and from 8.9 

to 8.1 log cell/g for direct inoculation. In the control, the average of the final population was 

5.6 log cell/g for the bucket method and 5.4 log cell/g for the direct method. 

 

3.2 Target metabolites 

Organic compounds such as citric, malic, succinic, lactic, acetic, propionic, isobutyric, 

and chlorogenic acid, and glycerol and ethanol from the inside (green bean) and outside of the 

fermented beans were detected and quantified by HPLC (Table 2). The concentrations of 

citric acid were lowered during fermentation-drying. Lactic and acetic acids presented inverse 

behavior in relation to the inoculation method used. With the bucket method, there were 

higher concentrations of lactic acid (an average of 1.21 g/L on the inside and 2.25 g/L on the 

outside) than with acetic acid (0.20 g/L on the inside and 0.19 g/L on the outside). With the 

2

3

4

5

6

7

8

9

10

0 60 120 180 240 300 360C
. 

p
a

ra
p

si
lo

si
s 

p
o

p
u
la

ti
o

n
 (

L
o

g
 c

el
l/

g
)

Time (h)



39 

 

direct method, there were higher concentrations of acetic acid (1.10 g/L on the inside and 2.07 

g/L on the outside) than with lactic acid (0.13 g/L on the inside and 0.22 g/L on the outside).  

Although lactic acid was produced throughout 352 hours of fermentation with both 

inoculation methods, when C. parapsilosis (CCMA 0544) and T. delbrueckii (CCMA 0684) 

were inoculated, there was a decrease in lactic acid production with direct inoculation (data 

not shown). Lactic and malic acid were produced, but only on the outside of the coffee bean 

after 112 hours of fermentation (from 0.07 to 0.13 g/Kg with the bucket method and from 

0.26 to 0.30 g/Kg with the direct method). Malic concentrations were reduced for the same 

period on the inside of the grain. 

At the start of the fermentation process, no acetic acid was detected on the outside of 

the grain with either inoculation method. However, at the end of the fermentation, higher 

concentrations were observed with the direct method (from 0.82 to 1.14 g/Kg on the inside 

and from 0.15 to 2.12g/Kg on the outside) when compared with the bucket method (from 0.20 

to 1.61 g/Kg on the inside and from 0.10 to 0.30 g/Kg on the outside) (Table 2). Propionic 

acid was presented in both inoculation methods. Isobutyric acid had the highest values on the 

inside of the bean (ranging from 1.35 to 4.73 g/Kg with the bucket method and from 2.24 to 

6.04 g/Kg with the direct method). Neither isovaleric nor butyric acid was detected during 

fermentation in any of the treatments studied. Chlorogenic acid increased in all fermentation 

samples, but decreased after roasting (18.39 g/Kg in grains inoculated using the bucket 

method and 14.91 g/Kg in grains inoculated using the direct method).  

The inoculation methods affected ethanol production, with higher concentrations when 

the bucket inoculation method was used. On the outside of the grain, the concentration ranged 

from 0.88 to 0.16 g/Kg, and from 1.04 to 0.17 g/Kg on the inside. With the direct inoculation 

method, the ethanol concentration was similar both on the inside and the outside of the coffee 

grains (Table 2). Where bucket inoculation was used, coffee beans had a higher glycerol 

concentration (ranging from 0.17 to 0.88 g/Kg on the inside and from 0.36 to 2.25 g/Kg on the 

outside). With the direct method of inoculation, the concentration ranged from 0.12 to 0.59 

g/Kg on the inside and from 0.19 to 0.96 g/Kg on the outside. 

 

  

    



 

 

4
0
 

Table 2 - Average values of organic compounds for all yeast treatments and control from inside and outside the 

bean. 

 

T= Roasted beans;      ND= Not detected



41 

 

3.3 Carbohydrates, trigonelline and caffeine 

Sucrose, fructose, glucose, trigonelline and caffeine concentrations were analyzed by 

HPLC (Figure 4). All treatments showed sucrose consumption after 112 hours of fermentation 

(with the bucket method, sucrose concentration ranged from 6.7 to 18.2 g/Kg for the control, 

from 4.5 to 18.6 g/Kg when inoculated with S. cerevisiae (CCMA 0543), from 5.2 to 17.8 

g/Kg when inoculated with C. parapsilosis (CCMA 0544), and from 7 to 14.5 g/Kg when 

inoculated with T. delbrueckii (CCMA 0684). With the direct method, the concentration 

ranged from 0.9 to 23.4 g/Kg for the control, from 3.8 to 19.9 g/Kg when inoculated with S. 

cerevisiae (CCMA 0543), from 4.5 to 22.1 g/Kg when inoculated with C. parapsilosis 

(CCMA 0544), and from 5.5 to 21.2 g/Kg when inoculated with T. delbrueckii (CCMA 

0684)). 

 

  



42 

 

Figure 4 - Carbohydrates detected in coffee beans during fermentation and drying: glucose 

(A) and fructose (B), exposed to treatments with inoculation—by S. cerevisiae 

(CCMA 0543) (▲), T. delbrueckii (CCMA 0684) ( ), and C. parapsilosis 

(CCMA 0544) (●)—and without inoculation—the control (♦)—using the bucket 

(full line) and the direct inoculation methods (dashed line). 
 

 

 

 

In contrast to fructose, glucose concentrations were lower for all fermentation times as 

shown in Figure 4: from 7.2 to 4.3 g/Kg with the bucket method and from 10.1 to 2.3 g/Kg 

with the direct method for the control; from 4.8 to 0.3 g/Kg for the bucket method and from 

8.5 to 2.8 g/Kg for the direct method for samples inoculated with S. cerevisiae (CCMA 0543); 

from 7.0 to 2.2 g/Kg for the bucket method and from 8.7 to 3.3 g/Kg with the direct method 

for samples inoculated with C. parapsilosis (CCMA 0544); and from 5.2 to 1.5 g/Kg with the 

bucket method and from 9.7 to 3.6 g/Kg with the direct method for samples inoculated with T. 

delbrueckii (CCMA 0684).  

Trigonelline concentrations in roasted beans were higher when the bucket inoculation 

method was used for all treatments studied, with C. parapsilosis (CCMA 0544) having the 

highest concentration (11.68 g/Kg). In green coffee, trigonelline concentrations were higher 

0
2
4
6
8

10
12
14

0 60 120 180 240 300 360

C
o

n
ce

n
tr

at
io

n
 g

/K
g

Time (h)

A

0
5

10
15
20
25
30
35
40
45
50

0 60 120 180 240 300 360

C
o
n
ce

n
tr

at
io

n
 g

/K
g

Time (h)

B



43 

 

only in the control treatment. In all other treatments via bucket inoculation, concentrations 

were lower than with direct inoculation. Caffeine concentration in roasted beans was also 

higher when the bucket method was used with S. cerevisiae (CCMA 0543) (24.76 g/Kg), C. 

parapsilosis (CCMA 0544) (21.94 g/Kg) and the control (18.79 g/Kg) assays. This was not 

the case for T. delbrueckii (CCMA 0684) (15.40 g/Kg). In green beans, caffeine was more 

stable for all yeasts when direct inoculation was used, apart from small variations in 

concentration levels. 

 

3.4 Volatiles compounds 

A total of 112 compounds were detected using the headspace-MS method and their 

IDs are displayed in the supplementary material. The compound groups detected in green and 

roasted coffee beans include aldehydes, ketones, alcohols, acids, esters, lactones, furans, 

pyrroles, pyrazines, aromatics, and others (Figure 5). Some of the detected compounds 

include important sensory values involved in flavor and aroma, including fruit-like, sweet, 

caramel, nutty, and floral tastes and smells, among others. 

 

Figure 5 – Profile of volatile compounds identified by HS-SPME GC-MS during coffee 

fermentation, with bucket inoculation of green (A) and roasted beans (C), and 

with direct inoculation of green (B) and roasted beans (D). Exposed treatments: 

(1) control (without inoculation); (2) inoculation with S. cerevisiae (CCMA 

0543); (3) inoculation with C. parapsilosis (CCMA 0544); and (4) inoculation 

with T. delbrueckii (CCMA 0684). Aldehydes ( ), Ketones ( ), Alcohols (

), Acids ( ), Esters ( ), Lactones ( ), Furans ( ), Pyrroles ( ), Pyrazines 

and pyridines ( ), Aromatics ( ), and others ( ). 

 

 



44 

 

In the green beans, alcohols and acids were the groups with more volatile compounds, 

regardless of the inoculation method (Figure 5). As can be seen in Figures 5A and 5B, the 

bucket and direct inoculation methods resulted in a small difference between the compound 

groups identified. However, the treatments used with bucket inoculation (Figure 5A) had 

more compounds than with direct inoculation (Figure 5B). Vanillin was only found in the 

direct treatment inoculated with S. cerevisiae and control, resulting in a vanilla-like flavor 

(Supplemental material). 

After roasting, there was a greater difference between the bucket inoculation method 

(Figure 5C) and the direct method (Figure 5D). Treatment with S. cerevisiae (CCMA 0543) 

and T. delbrueckii (CCMA 0684) using the bucket method showed higher numbers of the 

identified compounds. With direct inoculation, treatment with C. parapsilosis (CCMA 0544) 

showed a greater number of volatile compounds. 

 

3.5 Sensorial analyses  

With regard to sensorial analysis, coffee from all the treatments achieved scores above 

80, which indicates good quality. The sensorial results showed that there were no differences 

between treatments and inoculation methods. Coffee beverages produced using direct 

inoculation methods achieved scores of 81.5, 80.8, 81.0, and 81.3 for the control treatment, 

and treatment with T. delbrueckii (CCMA 0684), C. parapsilosis (CCMA 0544) and S. 

cerevisiae (CCMA 0543), respectively. Coffee beverages produced using the bucket method 

achieved scores of 81.4, 81.0, 81.3, and 81.4 for the control treatment, and treatment with T. 

delbrueckii (CCMA 0684), C. parapsilosis (CCMA 0544) and S. cerevisiae (CCMA 0543), 

respectively.   

 

4 Discussion 

First, this study measured plate counts of LAB, mesophilic bacteria and yeast (Fig. 2). 

Our main objective was to evaluate whether the bucket inoculation method would contribute 

to the persistence of the starter cultures during the coffee fermentation process. In general, the 

bucket method maintained a higher population of yeast, lactic acid and mesophilic bacteria at 

the end of fermentation (Figures 2A, 2B and 2C). Probably due to the favorable 

environmental conditions this method provided, such as availability of oxygen, temperature, 

and pH which lead to their survival and growth. Factors such as moisture and the temperature 

of the coffee beans can affect the degree of colonization and the colonization species (Silva et 

al., 2008), this may offer an explanation for a yeast and bacteria population above 6 log 



45 

 

CFU/g and for the similarity between treatments at the beginning of fermentation for both the 

bucket and direct methods. It can be assumed that inoculation with the yeast C. parapsilosis 

(CCMA 0544) using the direct method favors the growth of yeast and bacteria, resulting in 

slightly higher counts than with other treatments. In a previous study using the same process 

(semi-dry), including the direct inoculation method but with different coffee varieties of 

coffee, Evangelista et al. (2014b) observed that C. parapsilosis stood out from the other 

treatments with a higher population (approximately 7.5 log CFU/g) after 30 days of 

fermentation and drying. In our study, C. parapsilosis (CCMA 0544) prevailed in all 

treatments when using the bucket inoculation method and it can therefore be assumed that this 

method favors the growth of the inoculant and other coffee microbiota. 

In this study, three potential yeasts—S. cerevisiae (CCMA 0543), C. parapsilosis 

(CMMA 0544) and T. delbrueckii (CCMA 0684)—previously selected and tested based on 

their pectinolytic activity and organic acid production by Silva et al. (2013), were monitored 

using two inoculation methods (the bucket and the direct method), which had not previously 

been evaluated. The study used qPCR to evaluate the Catuai Amarelo variety (Figure 3). Our 

results demonstrate that the starter cultures gave positive results for Catuai Amarelo. 

Evangelista et al. (2014b) tested the same cultures using the direct method but with another 

variety of coffee. In that study, the direct inoculation method had positive results when using 

starter cultures. 

The bucket method showed a higher population at 352 hours with samples inoculated 

with S. cerevisiae than with non-inoculated coffee beans, indicating that this method helps the 

multiplication of cells and probably of specie S. cerevisiae, which is considered to be an 

epiphytic strain of coffee (Evangelista et al., 2014a; Masoud et al., 2004; Silva et al., 2000; 

Silva et al., 2013). Although the concentration of cells at the beginning of fermentation was 

supposed to be the same for all the yeast treatments whatever method was used, cell 

concentration was higher with the bucket method. This proves that some cell concentration is 

lost when spraying by air flow. The different inoculation methods did not affect the T. 

delbrueckii population. The bucket method favored the growth of the non-saccharomyces 

yeast Candida parapsilosis.  

As the concentration of acids may vary depending on the process used, the variety of 

coffee, the inoculation method and other factors, their presence may act as good quality 

indicators of the final product (Silva et al., 2008; Sunarharum et al., 2014; Vilela et al., 2010). 

Each yeast behaved differently with varying metabolite concentration levels. Some desirable 

dominant acids, such as citric and succinic acid, were detected throughout fermentation 



46 

 

(Table 2). In contrast to the study by Evangelista et al. (2014b), which tested the same 

inoculants, citric and succinic acid concentrations on the outside i8-of the bean were higher at 

the beginning of direct fermentation. There were lower acid concentrations at the end of 

fermentation, although a more stable population was observed with the bucket method. The 

bucket inoculation method maintained microbial population levels during the fermentation 

process (Figures 2 and 3, respectively) and generated a good quality final beverage. 

Chlorogenic acids are a major family of phenolic compounds and are responsible for 

coffee pigmentation and astringency (Duarte et al., 2010). The bucket method showed a 

constant increase in their levels before roasting (Table 2). This can be explained in accordance 

with Fagan et al. (2011) who have argued that chlorogenic acid is transported from the inside 

to the outside surface of the fruits to protect against attack from microorganisms when fruits 

are mature. After roasting, lower concentrations of these compounds were detected (with the 

bucket method, the concentration was 18.39 g/Kg and with the direct method 14.91 g/Kg). 

With its antioxidant properties, this result has potential for human health (Farah et al., 2006). 

As for trigonelline, after roasting, beans inoculated using the bucket method maintained 

higher concentrations (data described only), leading to desirable aroma compounds. 

Carbohydrates, glucose, fructose and sucrose were found in the mucilage and may be 

degraded by enzymatic reactions, with a positive impact on coffee aroma (Lee et al., 2015). 

All sucrose was degraded as expected (data described only). Lower glucose concentrations 

were observed for both inoculation methods when compared with fructose (Figure 4A). The 

level of maturation and the processing method used can influence sugar concentration (Knopp 

et al., 2006; Quintero, 2012). A possible explanation for lower glucose concentration is the 

microorganism’s presence as it is a primary energy source in metabolic processes. As with 

glucose, fructose was also consumed (Figure 4B). However, at the end of fermentation, both 

sugars increased for most of the treatments. There no clear explanation for this but it was 

probably due to enzymatic reactions inside the coffee bean (Figure 5), which may be 

responsible for fresh and floral notes (Toledo et al., 2016). In roasted beans, pyrazines and 

pyridines were the most abundant group, in accordance with Sunarharum et al. (2014), who 

argue that pyrazines are part of the top two main classes of compounds. Arising after roasting, 

they add significantly to coffee flavor by exhibiting nutty, earthy, roasted, and green aromas. 

Both main groups (acids and pyrazines) can be correlated with attributes perceived by the 

certified tasters in this study, including yellow fruits, caramel and almonds. This study 

identified important volatile markers of raw and roasted coffee such as, 2, 3-butanediol, 2, 5-

dimethylpyrazine, and 2, 3-dimethylpyrazine (complementary material). 



47 

 

Conclusions 

The bucket inoculation method favored microorganisms and metabolic response. The 

starter yeasts—S. cerevisiae (CCMA 0543), C. parapsilosis (CCMA 0544), and T. delbrueckii 

(CCMA 0684)—worked well with the Catuaí Amarelo coffee variety because they changed 

the behavior of the microbiota and the chemical composition during the process. In addition, 

all the coffees produced in this study resulted in scores over 80, which indicates good quality. 

 

Acknowledgments 

The authors thank the Brazilian agencies Conselho Nacional de Desenvolvimento 

Científico e Tecnológico of Brasil (CNPQ), Fundação de Amparo a Pesquisa do Estado de 

Minas Gerais (FAPEMIG), and Coordenação de Aperfeiçoamento de Pessoal de Nível 

Superior (CAPES). We also thank Dra. Suzana Evangelista Reis and Resfriado farm, from 

Lavras in the state of Minas Gerais, for the collected samples. 

 

REFERENCES 

 
Ballesteros, L.F., Teixeira, J.A., Mussatto, S.I., 2014. Chemical, Functional and 

Structural Properties of Spent Coffee Grounds and Coffee Silverskin. Food Bioprocess Tech. 

7, 3493–3503. 

Batista, N.N., Ramos, C.L., Ribeiro, D.D., Pinheiro, A.C.M., Schwan, R.F., 2015. 

Dynamic behavior of Saccharomyces cerevisiae, Pichia kluyveri and Hanseniaspora uvarum 

during spontaneous and inoculated cocoa fermentations and their effect on sensory 

characteristics of chocolate. LWT- Food Sci. and Technol. 63, 221-227. 

Belguidoum, K., Amira-Guebailia, H., Boulmokh, Y., Houache, O, 2014. HPLC 

coupled to UV – Vis detection for quantitative determination of phenolic compounds and 

caffeine in different brands of coffee in the Algerian market. J. Taiwan Inst. Chem. E. 45, 

1314–1320. 

Brando, C.H.J., Brando, M.F., 2015. Methods of coffee fermentation and drying. In: 

Schwan,R.F., Fleet, G.H. (Eds.), Cocoa and Coffee Fermentation. CRC Taylor & Francis, 

BocaRaton, FL,pp. 367–396. 

Bressanello, D., Liberto, E., Cordero, C., Rubiolo, P., Pellegrino, G., Ruosi, M.R., 

Bicchi, C., 2017. Coffee aroma: Chemometric comparison of thechemical information 

provided by three different samplings combined with GC–MS to describe the sensory 

properties in cup. Food Chem. 214, 218-226. 



48 

 

Czerny, M. and Grosch, W., 2000. Potent Odorants of Raw Arabica Coffee. Their 

Changes during Roasting. J. Agric. Food Chem. 48, 868-872. 

Contreras-Calderón, J., Mejía-Díaz, D., Martínez-Castaño, M., Bedoya-Ramírez, D., 

López-Rojas, N., Gómez-Narváez, F., Medina-Pineda, Y., Vega-Castro, O., 2016. Evaluation 

of antioxidant capacity in coffees marketed in Colombia: Relationship with the extent of non-

enzymatic browning. Food Chem. 209, 162-170. 

Companhia Nacional de abastecimento (CONAB), 2017. Available in: 

http://www.conab.gov.br/OlalaCMS/uploads/arquivos/17_01_17_14_51_54_boletim_cafe_-

_janeiro_de_2017.pdf. (Accessed at: 2017) 

Díaz, C., Molina, A.M., Nahring, J., Fischer, R., 2013. Characterization and dynamic 

behavior of wild yeast during spontaneous wine fermentation in steel tanks and amphorae. 

BioMed Res. Int., 1-13. 

Duarte, G.S., Pereira, A.A., Farah, A., 2010. Chlorogenic acids and other relevant 

compounds in Brazilian coffees processed by semi-dry and wet post-harvesting methods. 

Food Chem. 118, 851–855. 

Evangelista, S.R., Miguel, M.G.C.P., Cordeiro, C.S., Silva, C.F., Pinheiro, A.C.M., 

Schwan, R.F., 2014b. Inoculation of starter cultures in a semi-dry coffee (Coffea arabica) 

fermentation process. Food Microbiol. 44, 87–95. 

Evangelista, S.R., Silva, C.F., Miguel, M.G.C.P., Cordeiro, C.S., Pinheiro, A.C.M., 

Duarte, W.F., Schwan, R.F., 2014a. Improvement of coffee beverage quality by using selected 

yeasts strains during the fermentation in dry process. Food Res. Int. 61, 183–195. 

Fagan, E.B., Souza, C.H.E., Pereira, N.M.B., Machado, V.J., 2011. Effect of time on 

coffee bean (Coffea sp) grouth in cup quality. Biosci. J. 27, 729-738. 

Farah, A., Donangelo, C.M., 2006. Phenolic compounds in coffee. Braz. J. Plant 

Physiol. 18, 23-36. 

Flament, I. and Bessière-Thomas, Y., 2001. Coffee Flavor Chemistry. John Wiley & 

Sons, LTD., England. 

Gil-Serna, J., González-Salgadob, A., González-Jaén, M.T., Vázqueza, C., Patiño, B., 

2009. ITS-based detection and quantification of Aspergillus ochraceus and Aspergillus 

westerdijkiae in grapes and green coffee beans by real-time quantitative PCR. Int. J. Food 

Microbiol. 131, 162-167.  

Hays, C., Duhamel, C., Cattoir, V., Bonhomme, J., 2011. Rapid and accurate 

identification of species belonging to the Candida parapsilosis complex by real-time PCR and 

melting curve analysis. J. Med. Microbiol. 60, 477-480. 



49 

 

ICO. International Coffee Organization. 2017. Available in: 

http://www.ico.org/trade_statistics.asp?section=Statistics. (Accessed at: 2017). 

Knopp, S., Bytof, G., Selmar, D., 2006. Influence of processing on the content of 

sugars in green Arabica coffee beans. Eur. Food Res. Technol. 223, 195-201. 

Lee, L.W., Cheong, M.W., Curran, P., Yu, B., Liu S.Q., 2015. Coffee fermentation 

and flavor – An intricate and delicate relationship. Food Chem. 185, 182–191. 

Malta, M.R., Chagas, S.J.R., 2009. Avaliação de compostos não-voláteis em diferentes 

cultivares de cafeeiro produzidas na região sul de Minas Gerais. Acta Sci-Agron. 31, 57-61. 

Masoud, W., Bjorg, L.B., Jespersen, L., Jakobsen, M., 2004. Yeast involved in 

fermentation of Coffea arabica in East Africa determined by genotyping and by direct 

denaturating gradient gel electrophoresis. Yeast 21, 549–556. 

Menezes, A.G.T., Batista, N.N., Ramos, C.L, Reis, A.A.S, Efraim, P., Pinhero, 

A.C.M., Schwan, R.F., 2016. Investigation of chocolate produced from four different 

Brazilian varieties of cocoa (Theobroma cacao L.) inoculated with Saccharomyces cerevisiae. 

Food Res. Int. 81, 83-90. 

Michel, E., Monfort, C., Deffrasnes, M., Guezenec, S., Lhomme, E., Barret, M., 

Sicard, D., Dousset, X., Onno, B., 2016. Characterization of relative abundance of lactic acid 

bacteria species in French organic sourdough by cultural, qPCR and MiSeq high-throughput 

sequencing methods. Int. J. Food Microbiol. 238, 35-43. 

Postollec, F., Falentin, H., Pavan, S., Combrisson, J., Sohier, D., 2011. Recent 

advances in quantitative PCR (qPCR) applications in food microbiology.Food Microbiol. 28, 

848-861. 

Quintero, G.I.P., 2012. Factores, procesos y controles en la fermentación del café. 

Avances Técnicos Cenicafé, Manizales, Colombia. 

Ramos, D.M.B., Silva, C.F., Batista, L.R., Schwan, R.F., 2010. In vitro inhibition of 

toxigenic filamentous fungi by Pichia sp.and Debaryomyces sp. isolates from coffee (Coffea 

arabica) fruits.Acta Sci-Agron. 32, 397-402. 

Restuccia, D., Spizzirri, G., Parisi, O., Cirillo, G., Picci, N., 2015. Brewing effect on 

levels of biogenic amines in different coffee samples as determined by LC-UV. Food Chem. 

175, 143–150. 

Sakiyama, N. and Gava, M., 2015. Botany and Production of Coffee. In: Schwan,R.F., 

Fleet, G.H. (Eds.), Cocoa and Coffee Fermentation. CRC Taylor & Francis, Boca Raton, 

FL,pp. 341-365. 



50 

 

SCAA— Specialty Coffee Association of America. Available in: http://www.scaa.org, 

2013. 

Silva, C.F., Schwan, R.F., Dias, E.S., Wheals, A.E., 2000. Microbial diversity during 

maturation and natural processing of coffee cherries of Coffea arabica in Brazil. Int. J. Food 

Microbiol. 60, 251–260. 

Silva, C.F., Batista, L.R., Abreu, L.M., Dias, E.S., Schwan, R.F., 2008. Succession of 

bacterial and fungal communities during natural coffee (Coffea arabica) fermentation. Food 

Microbiol. 25, 951-957. 

Silva, C.F., Vilela, D.M., Cordeiro, C.S., Duarte, W.F., Dias, D.R., Schwan, R.F., 

2013. Evaluation of a potential starter culture for enhance quality of coffee fermentation. 

World J. Microb. Biot. 29, 235–247. 

Sunarharum, W.B., Williams, D.J., Smyth, H.E., 2014. Complexity of coffee flavor: A 

compositional and sensory perspective. Food Res. Int. 62, 315–325. 

Toledo, P.R.A.B, Pezza, L., Pezza, H.R., Toci, A.T., 2016. Relationship Between the 

Different AspectsRelated to Coffee Quality and Their VolatileCompounds.Compr. Rev. Food 

Sci. F. 15, 705-719. 

Vilela, D.M., Pereira, G.V.M., Silva, C.F., Batista, L.R., Schwan, R.F., 2010. 

Molecular ecology and polyphasic characterization of the microbiota associated with semi-dry 

processed coffee (Coffea arabica L.). Food Microbiol. 27, 1128-1135. 

Ziółkowska, A., Wąsowicz, E., Jeleń, H., 2016. Differentiation of wines according to 

grape variety and geographical origin based on volatiles profiling using SPME-MS and 

SPME-GC/MS methods. Food Chem. 213, 714-720. 

Zott, K., Claisse, O., Lucas, P., Coulon, J., Lonvaud-Funel, A., Pomarede-Masneuf, I., 

2010. Characterization of the yeast ecosystem in grape must and wine using real-time PCR. 

Food Microbiol. 27, 559-567. 

 

 

 

 

 

 

 

 

 

 

 



 

 

5
1
 

Supplemental material 

Volatile compounds found in treatments inoculated with yeasts and without inoculation (continue) 

Group Compound Sample of treatment 
Sensory attributes* 

(Odor or flavor) 

Aldehydes Hexanal 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 Fruity 

2-Butenal, 3-methyl- 1, 2, 3, 4, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 Almond 

2-Octenal, (E)- 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16  

2-Nonenal, (E)- 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 Fatty, cardboard-like 

Benzaldehyde 3, 5, 7, 8, 9, 11, 14, 15, 16 Almond, sweet 

2-Pentenal, (E)- 1, 2, 3, 4, 9, 11, 12   

Pentadecanal- 1, 4, 5, 6, 8, 9, 10, 11, 12, 13, 14, 16   

2-Heptenal, (Z)- 1 Apple-like 

Heptanal 5, 6, 7, 15, 16 Fruity-like 

Nonanal 5, 6, 7, 8, 13, 14, 16   

Decanal 7, 8   

2,4-Nonadienal, (E,E)- 9, 10, 12  Oily, Green 

2-Hexenal 15  

2,6-Nonadienal, (E,Z)- 13, 14, 15, 16   

2-Furancarboxaldehyde 17, 18, 19, 20, 21, 22, 23, 24  Sweet, bread-like 

5-Hydroxymethylfurfural 17, 18, 19, 20, 21, 23, 24   

Vanillin 9, 10  Vanilla-like 

Ketones 2-Heptanone 1, 2, 3, 4, 9, 10, 11, 12, 13, 14, 16  

5,9-Undecadien-2-one, 6,10-dimethyl-, (E)- 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16  

2-Butanone, 3-hydroxy- 3  

 2-Pentadecanone, 6,10,14-trimethyl- 1, 2, 4, 5, 6, 8, 9, 10, 11, 12, 13, 14, 15, 16, 18, 20, 23  

3-Octen-2-one, (E)- 5, 7, 8, 15 Earthy 

2,3-Octanedione 10, 12, 14, 15, 16  Cheese 

2-Undecanone, 6,10-dimethyl- 10  Citronella 

Acetoin 17, 18, 19, 20, 21, 22, 23, 24   



 

 

5
2
 

Supplemental material 

Volatile compounds found in treatments inoculated with yeasts and without inoculation (continue) 

Group Compound Sample of treatment 
Sensory attributes* 

(Odor or flavor) 

 2-Propanone, 1-hydroxy- 17, 18, 19, 20, 21, 22, 23, 24  Sweet, caramel 

2-Butanone, 1-hydroxy- 17, 18, 19, 20, 21, 22, 23, 24   

Ethanone, 1-(1H-pyrrol-2-yl)- 17, 18, 20, 21, 22, 23, 24   

Alcohols (S)-(+)-2-Pentanol 3  

1-Butanol, 3-methyl- 1, 2, 3, 4, 5, 6, 7, 8, 12, 16 Fruity 

1-Pentanol 2, 3, 4, 5, 6, 7, 10, 11, 12, 13, 14, 15, 16  

2-Buten-1-ol, 3-methyl- 2, 3, 4, 5, 6, 7, 8, 9, 10, 12, 13, 14, 16  

2-Heptanol 1, 2, 3, 4, 5, 6, 7, 9, 10, 11, 12, 13, 14, 16 Nutty, sweet 

1-Hexanol 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 Fruity 

1-Hexanol, 2-ethyl- 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 Fermented-yeast 

Linalyl alcohol 2, 3, 4, 7  Floral 

1-Octanol 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 12, 13, 14, 15, 16 Coconut 

 2,3-Butanediol, [R-(R*,R*)]- 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16  

2-Octen-1-ol, (E)- 1, 3, 4, 5, 6, 7, 8, 13, 14, 15, 16 Fruity 

Benzyl alcohol 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16  

Phenylethyl Alcohol 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 Floral 

 1-Pentanol, 4-methyl 2, 3   

1-Penten-3-ol 1, 2, 3, 4  

3-Buten-1-ol, 3-methyl- 1, 3  

3-Hexen-1-ol, (Z)- 2, 3, 4, 9, 11, 12  

1-Heptanol 1, 2, 3, 4, 5, 6, 7, 13, 14, 15, 16 Dairy, lactonic 

1-Propanol, 2-methyl- 3, 10, 11, 12  

5-Hepten-2-ol, 6-methyl- 1, 4, 9, 10, 11, 12  

1,6-Octadien-3-ol, 3,7-dimethyl- 1, 2, 5, 6, 8, 9, 10, 11, 12, 13, 14, 15, 16 Floral 

2-Hexen-1-ol, (E)- 1, 2, 4, 9, 10, 11, 12   

Cyclohexanol, 5-methyl-2-(1-methylethyl) 5, 7 Mint-like 



 

 

5
3
 

Supplemental material 

Volatile compounds found in treatments inoculated with yeasts and without inoculation (continue) 

Group Compound Sample of treatment 
Sensory attributes* 

(Odor or flavor) 

 1-Nonanol 5, 6, 7, 8, 14, 15, 16   

3-Octanol, 2-methyl- 6   

1-Butanol 6, 8, 16  

1-Octen-3-ol 13, 14, 15, 16  

Acids Butanoic acid, 3-methyl- 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 17, 18, 19, 20, 21, 22, 23, 24  

Pentanoic acid 1, 2, 3, 4, 5, 6, 8, 13, 14, 15, 16 Fruity, sweaty 

2-Butenoic acid, 3-methyl- 1, 2, 3, 4, 5, 6, 7, 8, 9, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 

21, 22, 23, 24 

 

Hexanoic acid 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 16  

Hexanoic acid, 2-ethyl- 1, 2, 3, 4, 5, 6, 14, 20, 21, 22, 23, 24 Herbal 

 Octanoic acid 1, 2, 3, 4, 5, 6, 7, 9, 10, 11, 12, 16  

Nonanoic acid 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16  

Tetradecanoic acid 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 14, 15, 16  

Hexadecanoic acid 1, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 

21, 22, 23, 24 

 

Hexadecanoic acid, ethyl ester 3, 4, 9, 10, 11, 12, 13  

Pentadecanoic acid 1, 2, 3, 4, 5, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 21, 24  

Hexadecanoic acid, methyl ester 1, 2, 3, 4, 9, 13, 14, 16  

Hexanoic acid, ethyl ester 2  

Hexanoic acid, methyl ester 1  Pineapple-apricot 

Dodecanoic acid 2, 9, 10, 11, 12, 13, 15, 16  

Benzoic acid 1, 4, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16  

Decanoic acid 1, 5, 6, 8, 9, 10, 11, 12, 13, 14, 15, 16  

Butanoic acid 7   

 



 

 

5
4
 

Supplemental material 

Volatile compounds found in treatments inoculated with yeasts and without inoculation (continue) 

Group Compound Sample of treatment 
Sensory attributes* 

(Odor or flavor) 

Esters 1-Butanol, 3-methyl-, acetate 3, 6, 9, 10, 12, 15 Fruity, earthy 

Methyl salicylate 1, 2, 3, 4, 5, 6, 7, 9, 10, 11, 12, 13, 14, 15, 16  

Butanedioic acid, diethyl ester 2, 4   

Ethyl 9-hexadecenoate 2   

Furfuryl formate 17, 18, 19, 20, 21, 22, 23, 24 Floral 

2-Furanmethanol, acetate 17, 18, 19, 20, 21, 22, 23, 24 Floral 

Benzeneacetic acid, ethyl ester 2, 3, 4   

Lactones Butyrolactone 5, 6, 7, 17, 18, 19, 20, 21, 22, 23, 24  

2(3H)-Furanone, dihydro-5-pentyl- 10, 11, 12, 13, 14, 15, 16 Fruity 

Furans Furan, 2-pentyl- 2, 3, 4, 9, 11, 12, 13, 14, 15, 16  Earthy 

Ethanone, 1-(2-furanyl)- 17, 18, 19, 20, 21, 22, 23, 24  

2-Furancarboxaldehyde, 5-methyl- 17, 18, 19, 20, 21, 22, 23, 24 Sweet, caramel 

2-Furanmethanol 17, 18, 19, 20, 21, 22, 23, 24  

Pyrroles Pyrrole 17, 18, 19, 20, 21, 22, 23, 24  

Indole 23  

Pyrazines 

and 

pyridines 

Pyrazine, 2-methoxy-3-(2-methylpropyl)- 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16  

Pyrazine, methyl- 17, 18, 19, 20, 21, 22, 23, 24 Sweet. Nutty 

Pyrazine, 2,5-dimethyl- 17, 18, 19, 20, 21, 22, 23, 24  

Pyrazine, 2,6-dimethyl- 17, 18, 19, 20, 21, 22, 23, 24 Sweet 

 Pyrazine, 2,3-dimethyl- 17, 18, 19, 20, 21, 22, 23, 24 Nutty 

Pyrazine, 2-ethyl-6-methyl- 17, 18, 19, 20, 21, 22, 23, 24  

Pyrazine, 2-ethyl-3-methyl- 17, 18, 19, 20, 21, 22, 23, 24 Nutty 

Pyrazine, trimethyl- 17, 18, 19, 20, 21, 22, 23, 24  

Acetylpyrazine 17, 18, 19, 20, 21, 22, 23, 24  

Pyrazine, ethenyl- 18, 19  

Pyrazine 17, 18, 20, 21, 23, 24  



 

 

5
5
 

Supplemental material 

Volatile compounds found in treatments inoculated with yeasts and without inoculation (conclusion) 

Group Compound Sample of treatment 
Sensory attributes* 

(Odor or flavor) 

 Pyridine, 2,6-dimethyl- 6, 7, 13, 14, 15, 16  

Pyridine 17, 18, 19, 20, 21, 22, 23, 24  

Aromatics Mesitylene 3, 4, 5, 6, 7, 11  

o-Xylene 1, 2, 3, 5, 7, 9, 11  

Styrene 1, 2, 4, 5, 6, 9, 10, 11, 12, 13, 14, 15, 16  

p-Xylene 1, 5, 9  

Benzene, 1-ethyl-4-methyl- 1, 4, 7, 8, 11, 13, 15, 16 Earthy 

Benzene, 1,2,4-trimethyl- 1, 9, 10, 11, 13, 14, 15, 16  

p-Cymene 5, 6, 7, 8, 13  

Benzene, 1,2-dimethyl- 7  

Maltol 17, 18, 20, 21, 23, 24 Fruity, caramel 

Others 2-Buten-1-ol, 3-methyl-, acetate 1, 2, 3, 4, 9, 10, 11, 12  Banana-like 

2-Butanone, 1-(acetyloxy) 17, 18, 19, 20, 21, 22, 23, 24 Coffee 

Green beans without yeast inoculation (control): Bucket inoculation 1= initial time (0 h) and 5= final time (352 h) of fermentation; Direct inoculation 9= 

initial time (0 h) and 13= final time (352 h) of fermentation 

Green beans with yeast Saccharomyces cerevisiae (CCMA 0543): Bucket inoculation 2 =initial time (0 h) and 6 =final time (352 h) of fermentation; Direct 

inoculation 10 =initial time (0 h) and 14 =final time (352 h) of fermentation 

Green beans with yeast Candida parapsilosis (CCMA 0544): Bucket inoculation 3 =initial time (0 h) and 7 =final time (352 h) of fermentation; Direct 

inoculation 11 =initial time (0 h) and 15 =final time (352 h) of fermentation 

Green beans with yeast Torulaspora delbrueckii (CCMA 0684): Bucket inoculation 4 =initial time (0 h) and 8 =final time (352 h) of fermentation; Direct 

inoculation 12 =initial time (0 h) and 16 =final time (352 h) of fermentation 

RB = Roasted beans: Bucket inoculation 17 =control, 18 =Saccharomyces cerevisiae (CCMA 0543), 19 =Candida parapsilosis (CCMA 0544) and 20 = 

Torulaspora delbrueckii (CCMA 0684) 

RD =Roasted beans: Direct inoculation 21 = control, 22 =Saccharomyces cerevisiae (CCMA 0543), 23 = Candida parapsilosis (CCMA 0544) and 24 

=Torulaspora delbrueckii (CCMA 0684) 

* Sensory attributes are taken from: Flament, I. and Bessière-Thomas, Y.(2001) and Czerny and Grosch (2000). 



56 

 

5
6
 

FINAL CONSIDERATIONS 

 

Information regarding culture starters and optimization of coffee fermentations with 

different kind of varieties tested with all coffee process and other strains is still lacking for 

literature review since the topic is recent and is still been under research. Based on the review, 

final coffee beverages are affected by a diversity of factors such as environmental conditions, 

inoculation strains, quality of grains, concentration of starters, natural microbiota and more. 

As a result, the biochemical compounds are always changing and coffee fermentations are 

considered complex. 

In this work, the objectives stated were reached, such as evaluating the effect the 

inoculation method has on Catuai Amarelo coffee using previously selected strains as starters. 

The results obtained showed that bucket fermentation is a promising method for producers or 

farmers to do when inoculating, since generated special coffees with different attributes. In 

addition, this kind of inoculation (bucket) becomes an advantage for farmers, since materials 

are available and present low cost. 

Due to coffee fermentations, special coffees with a variety of sensorial differences can 

be obtained. Therefore, the steps to make fermentation successfully must be taught in farmers 

because it can raise the value of their crop and income. In future studies, it would be 

significant to correlate the selected strains used for inoculation with the metabolites they 

actually produce