Yeasts Associated with Various Amazonian Native Fruits

Fruits constitute excellent habitats for yeasts, mainly due to their low pH, availability of nutrients, and active fruit-associated vectors. These traits are variable across the type and maturity of the fruit. Changes in the community in response to varying availability of nutrients, production of mycotoxins, and the arrival of new yeast species are evident (Tournas and Katsoudas 2005; Starmer and Lachance 2011).

The majority of research has focused on the diversity of yeasts on grapes and wine-related samples due to their application in the winemaking process (Guillamón et al. 1998; Filho et al. 2017), although some expansions have been made beyond this zone of interest. Koricha et al. (2019) identified yeasts from lemon, mango, and guava fruits, with Candida albicans, Debaryomyces hansenii, Kodamaea ohmeri, Rhodotorula mucilaginosa, among others, found to be present. Vadkertiová et al. (2012) studied the diversity of yeasts and yeast-like microbes associated with fruits and blossoms of apple, plum, and pear orchards in Slovakia. Trindade et al. (2002) investigated yeasts inhabiting the fresh and frozen pulps of Brazilian tropical fruits. Notably, some fruits have been described as sources of new yeasts (Bhadra et al. 2008; Sipiczki 2011).

Yeast diversity on the wide variety of Amazonian native fruits (ANF) has not been widely investigated, with reports focusing mainly on other tropical fruits like passion fruit (Passiflora edulis), mangaba (Hancornia speciosa), umbu (Spondias tuberosa), and acerola (Malpighia glabra) (Trindade et al. 2002; Da Silva et al. 2005; Grondin et al. 2015). The Amazonian rainforest’s environmental characteristics suggest the possibility of finding diverse yeast communities, including new species or strains with new characteristics of biotechnological interest (Morais et al. 1995; Da Silva et al. 2005). Yeasts represent a promising source for obtaining microbial enzymes (Trindade et al. 2002; Da Silva et al. 2005; Raveendran et al. 2018), flavorings (Grondin et al. 2015), and can be used as biocontrol agents for postharvest fruit diseases (Janisiewicz et al. 2010; Ruiz-Moyano et al. 2016). Yeasts or their metabolites isolated from the Amazonian rainforest may display unique characteristics due to the particularities of their habitat. Accordingly, this study aimed to identify and characterize yeasts isolated from Amazonian native fruits using molecular and phenotypic methods to glimpse also their potential biotechnological features.

A total of 81 yeast isolates were obtained from one hundred specimens of 10 different fruit species from the Peruvian Amazonia. Typing at the strain level by rep-PCR discriminated 29 strain profiles (Table I). Our naming system for the strain profiles employed the letter P, followed by a 2-digit number. One representative of each strain profile was chosen for further RFLP analysis of the 5.8S-ITS region. Using this methodology, we were able to distinguish 12 restriction profiles (Table II).

Phylogenetic analysis of the D1/D2 domains of the 26S rRNA gene was used to identify all 29 representative strains, revealing 16 species belonging to 6 genera. Our dataset comprised 45 sequences from the 26S rRNA partial gene (Fig. 1). The final alignment resulted in 578 aligned positions and 288 variable sites. All isolates were identified as ascomycetous and non-Saccharomyces species. Hanseniaspora (40.7%) was the most common genus, followed by Candida (35.6%), and Debaryomyces (17.3%). H. opuntiae (24.7%) was found to be the most prevalent species among all the isolates, followed by C. tropicalis with 16.0%. D. hansenii and H. thailandica were also present at 11.1% of strains obtained (Table II). The highest number of strain profiles was observed in both H. opuntiae and H. thailandica (five strain profiles each), followed by C. tropicalis (three strain profiles). C. carpophila, D. hansenii, and D. nepalensis exhibited two strain profiles; others exhibited one (Table II).

Analysis of yeast presence on Amazonian fruits showed that H. opuntiae was found most frequently across the ANF of our study, with a presence on five of the studied fruits, while C. tropicalis was the second most common, with presence on four of the studied fruits. Yeasts species tended to cluster together with multiple species present on each fruit. The highest number of species of yeasts was found associated with cidra (five species) and ungurahui (four species) fruits, while the lowest was found in camu camu and charichuelo fruits, with only one species being associated with each. Pomarrosa, taperiba, and ubos fruits shared a similar yeast profile, with each harboring H. opuntiae and H. thailandica. Cidra and taperiba showed the highest number of strain profiles (data not shown). Cidra exhibited six strain profiles corresponding to five yeast species; taperiba showed six strain profiles, from three yeast species (Table II).

The biochemical profiles of selected yeasts (one representative per each strain profile in most cases) are presented in Table III. Some phenotypic traits were investigated as hydrolytic capabilities for potential biotechnological applications. Isolates showed a diverse range of phenotypic characteristics, with differentiation evident even between strains belonging to the same species. Fermentation of lactose and growth at 60% glucose was negative in all isolates tested. Hydrolytic capabilities were rarely detected, and lipolytic activity was determined in only one isolate (P11 strain profile). Degradation of Tween 80, carboxymethyl cellulose, casein, gelatin, and starch was not evident in any strain.

Fruits possess essential traits that make them suitable habitats for yeasts. In this study, we isolated yeasts from 10 ripe ANF of the region of Loreto, Peruvian Amazonia, and belonging to the genera Citrus, Garciniama, Mauritia, Myrciaria, Oenocarpus, Poraqueiba, Solanum, Spondias, and Syzygium. Repetitive sequence-based PCR (rep-PCR) yielded 29 strain profiles of yeasts from these fruits. Although this method was initially developed for fingerprinting bacterial genomes (Versalovic et al. 1991), it has also been applied in describing fungal diversity in various samples (Ceugniez et al. 2015; Filho et al. 2017).

For preliminary visualization and identification of the microbial community, we conducted an RFLP analysis of the amplified 5.8S rRNA gene with the two flanking internal transcribed spacers ITS1 and ITS2 (Esteve-Zarzoso et al. 1999), yielding 12 restriction profiles. The majority of RFLP restriction profiles could not be matched with previous reports (Guillamón et al. 1998; Esteve-Zarzoso et al. 1999). As far as we know, restricted profiles belonging to our strains P01, P04, P05, P07, P11, P12, P19, and P23 had no match to any previously published strains. Meanwhile, P13, P14, P16, P17, P18, P24, P25, P27, P28, P29, and P30 were similar to H. opuntiae, H. pseudoguilliermondii, H. thailandica or H. uvarum. This is probably because the methodology in question includes only a limited number of strains currently isolated from other types of fruits and environments. Nonetheless, this approach provided important information about the profiles and, in some cases, the species. We found that several yeast species exhibited the same restriction profile of the ITS region. For example, C. carpophila and Meyerozyma caribbica (restriction profile II, Table II). However, all the species could be differentiated using rRNA gene sequencing (Jindamorakot et al. 2009).

When various typing methods are used together, higher-profile diversity can be observed than when single methods are used (Padilla et al. 2016). Thus, for the purpose of supporting and determining results, we also carried out the sequencing of the D1/D2 domains of the 26S rRNA gene (Kurtzman and Robnett 1998). The identification consensus of all strains was achieved by analyzing the information gathered from these combined techniques. Individual identities were ascribed to each of 81 isolates grouped in 29 strain profiles (Table I).

The distribution of species and strains varied across the ANF in this study. Communities were dominated by the genus Hanseniaspora, followed by Candida and Debaryomyces. More than one yeast species was present on all fruits except camu camu and charichuelo (Table II). In the conditions of this study, we believe that the nature of the fruit peels (chemical composition, thickness, aspect) may be one of the principal reasons for the yeast profiles observed on the ANF, but further research into the nature of these fruits is needed.

The characteristics of fruits strongly influence the diversity of yeasts and other microbes found in their adherent communities. The peels of fruits can contain various proportions of carbohydrates, crude fibers, lipids, crude proteins, minerals, and anti-nutrients. Adherent microbes must develop ways to access such materials (Villachica 1996; Romelle et al. 2016). Additionally, some fruit skins are thinner than others or have indentations that make them more prone to yeast colonization (Tournas and Katsoudas 2005). Interestingly, the highest number of yeast species was found associated with cidra fruit, a citrus species with low pH, and ungurahui, which is considered one of the most useful plants for indigenous people in Amazonia. Also, cidra harbors strains belonging to four genera (the highest number of genera among our ANF), possibly because low pH is a favorable condition for yeast growth. The range of pH of citrus fruits tends to be between 2.3 and 3.6 (Irkin et al. 2015). Ungurahui is employed for medicinal and cosmetic purposes, and to prepare a milk-like alcoholic beverage called chicha (Montúfar et al. 2010). Ungurahui was found colonized by members of the Candida genus, which may explain why the fruit is used to produce fermented alcoholic beverages, as frequently different species of Candida are present in fruits used for alcoholic fermentation (Fleet 2003; Capozzi et al. 2015).

Conditions such as climate, geography, and other factors also interact to determine yeast diversity on the fruit surfaces (Andrews and Harris 2000; Fonseca and Inácio 2006; El Sheikha et al. 2009). Similarly, the stage of fruit maturity also plays an important role in determining the composition of yeast communities (Morais et al. 1995), though in our study, all fruit samples were mature. Hence, we showed the composition of yeast communities at that stage. Thus, nutrient changes and physicochemical characteristics exert an effect on the diversity of yeasts.

The same species identified in this study have been reported in other investigations using samples as diverse as non-Amazonian fruits, other plant surfaces, grape-associated products or even clinical samples (Kurtzman et al. 2011a). The genera Hanseniaspora and Candida have been typically associated with grape juice in the first stages of alcoholic fermentation during winemaking and have been identified as the main genera in some yeast diversity studies on fruits (Trindade et al. 2002; Vadkertiová et al. 2012; Grondin et al. 2015). H. opuntiae have been mainly found in the microbiota of cocoa bean fermentations (Fernández Maura et al. 2016). H. opuntiae have also been identified in the pineapple vinification process in Angola (Dellacassa et al. 2017). H. opuntiae can be referred as a ubiquitous yeast in nature. This fact is corroborated in our study, where H. opuntiae was found among half of the fruits tested and exhibited the high number of strains.

H. thailandica was first reported by Jindamorakot et al. (2009) in samples of insect frass, crabapple mangrove (Sonneratia caseolaris) flowers, lichen, and rotted Psidium guajava fruit from different locations in Thailand. In our study, both H. opuntiae and H. thailandica showed various strain profiles and tended to be present in consortium with other species of the Hanseniaspora genus (Table II). In contrast to the other representatives of the Hanseniaspora genus in our study, H. pseudoguilliermondii and H. uvarum showed low prevalence.

Trindade et al. (2002) isolated yeasts from fresh and frozen pulps of the Brazilian tropical fruits pitanga (Eugenia uniflora), mangaba (Hancornia speciosa), umbu (Spondias tuberosa), and acerola (Malpighia glaba). The authors found 405 different strains belonging to 42 ascomycetous and 28 basidiomycetous species, including various species of Candida, Hanseniaspora, Kluyveromyces, Rhodotorula, and Saccharomyces, among others. Candida showed the highest species richness, as was also the case in our investigation. However, we observed only ascomycetous yeasts, and one reason for this could be the temperature of 30°C we used for the isolation of yeasts. Surprisingly, none of the isolates identified by Trindade et al. (2002) were coincident with our results. This could be due to the nature of the fruits.

The Candida genus is widely found in yeast diversity studies on fruits. Candida tropicalis has been described in various ecological niches (Las Heras-Vazquez et al. 2003; Limtong et al. 2014). C. pseudohaemulonii is ordinarily found in clinical samples at hospitals (Sugita et al. 2006; Oh et al. 2011). However, we found that C. pseudohaemulonii is also associated with citrus fruit in consortium with other yeast genera representing novel information.

D. hansenii was found on pear fruit surfaces by Chand-Goyal and Spotts (1996) from diverse areas in the Pacific Northwest United States. Interestingly, D. hansenii has been described as harboring particular features for biotechnological applications (Prista et al. 2016). In our investigation, D. hansenii appeared to prefer Garciniama crophylla tree fruit as a habitat and was the only yeast species found on this fruit.

K. ohmeri has been mainly reported as a rare human pathogen (Al-Sweih et al. 2011; Fernández-Ruiz et al. 2017). However, it has also been described as being associated with food (Ezeokoli et al. 2016). In our work, we found K. ohmeri associated with cidra. The genus Martiniozyma has recently been described (Kurtzman 2015), and C. asiatica is now recognized as Martiniozyma asiatica. M. asiatica has been previously detected in natural samples from various Asian countries (Limtong et al. 2010). In our study, M. asiatica tended to cohabit with other yeast species associated with umarí fruit.

In order to analyze phenotypic characteristics of the isolates, and possibly find useful traits for biotechnological purposes (Da Silva et al. 2005; Molnárová et al. 2013), some phenotypic tests were carried out. Variation in phenotypic traits of the species compared to previous reports (Kurtzman et al. 2011a) may be due to diverse factors, including the dynamic environmental conditions of Amazonia, which may influence the physiological features. Certain environmental conditions may switch specific genes on or off, causing the broad strain variation. In addition, the patterns we observed could also be ascribed to the effects of fruit species. These factors have been shown to contribute to species variation (Lane et al. 2011; Qvirist et al. 2016).

In terms of hydrolytic capabilities, lipase production was only detected in C. akabanensis, which was isolated exclusively from aguaje, a fruit with high fatty acid content. It is possible that C. akabanensis employs lipase to, in some way, utilize the fatty acids present in the pulp. More generally, however, the rarity of hydrolytic activity detected in our study is not unexpected, since it appears that these yeasts tend to use straightforward sources of carbon such as simple sugars (glucose in most cases, Table III). Ecologically, this is a cost-effective strategy, considering that the surface of fruits in the Amazonia tends to constitute harsh environmental conditions. It is important to consider that strains can be very heterogeneous both genetically and biochemically (Prista et al. 2016; Visintin et al. 2016), and also this variability can be strongly influenced by the nutritional composition of the samples they are obtained from. Furthermore, as far as we know, there are no reports of the same tests for hydrolases for all the species of this study to compare. Thus, the majority of negative hydrolytic profiles appear in agreement with the information described by Kurtzman et al. (2011a). Nevertheless, we recommend investigating hydrolytic capabilities using basal nutrients different from YPD and non-synthetic or residual substrates because non-natural substrates can result in a different biochemical response of the yeasts. More suitable substrates can be starch of potato, olive oil, or beef suet.

Comparing our results with previous works shows that yeasts are ubiquitous on different fruits, and even in different types of samples. The surface of Amazonian fruits, although a hostile environment, can be an interesting source of yeast strains displaying diverse phenotypic traits. Different yeasts found in the ANF studied seem to be influenced mainly by the nature of the fruits and their environment. ANF may constitute a good source of new species or strains of yeasts with particular characteristics for biotechnological purposes. Further investigation is needed in order to explore the potential industrial applications of these yeasts in food, feed ingredients, biocatalysis, or biocontrol.