**3. Results and discussion**

#### **3.1. Evolutionary relationships of palm wine yeasts and their relatives**

Yeasts facilitate several industrial food fermentation processes, which often consist of a desired specific strain [19]. This may be why domestication is believed to be the main driver of specific yeast prevalence in a geographical location. The understanding of the ecological basis of yeast diversity in nature remains fragmented and cross-kingdom competition has been proposed as a method to generate industrially useful yeast strains with new metabolic traits [20]. Palm wine yeasts are yet to enjoy significant diversity study hence a look at their relatives will enable more information to be generated.

In the last decade, there has been increase in submissions of palm wine yeast sequences based on 26S rRNA genes mainly due to quality checks by academicjournals. The identification of new strains is accompanied by performing a search with the basic local alignment search tool [21] followed by submission of DNA sequences to the GenBank. According to Benson et al. [22], GenBank is a comprehensive database that contains publicly available nucleotide sequences for up to 370,000 formally described species. It is common knowledge that these submissions which contain a lot of information are generated mainly through submissions from investigators around the world. Each sequence data received is curated by the GenBank annotation staff to ensure that it is free from errors after which accession numbers are assigned.-

All the sequences used in this study were the first versions submitted by investigators. The maximum likelihood method was preferred for the trees constructed because it is computationally intense and all possible trees are considered. Also the method can be useful for widely divergent groups or other difficult situations [23].

#### **3.2.** *Candida ethanolica*

The yeast *C. ethanolica* is not widely reported in palm wine. It has been reported as a nonconventional yeast which may present massive resource of yeast biodiversity for industrial applications because it has been found to be adapted to some of the stress factors present in harsh environmental [24]. In that report, it was found that *C. ethanolica* tolerated up to 7% v/v ethanol. This could be useful information for new palm wine drink development especially now that there is increasing interest in non-*Saccharomyces* yeasts with peculiar features able to replace or accompany *S. cerevisiae* during specific industrial fermentations [25].

The *C. ethanolica* strain from *Raphia* sp. (**Figure 1**) and *Elaeis* sp. (**Figure 2**) palm wine showed close relationships with other *Candida* species*.* The relatives of *Raphia* sp. palm wine that emanated from the same node (**Figure 1**) came from diverse sources. The flanking close relatives- (KY283163 and DQ466540) of *C. ethanolica* (HG425332) were isolated from composite microbial powders for aquaculture in China [26] and composite cocoa fermentation in Ghana [27]. Other close relatives included species from the genus *Pichia.* The *P. deserticola* strain (KM005182) from the same node as the reference strain was from aerobic deterioration of total mixed ration silage in China [28]. For *Elaeis* sp. (**Figure 2**) palm wine, close relatives to *C. ethanolica* (HG425336) strain were from a laboratory culture collection with unidentified source [29] and a tannin tolerant yeasts associated with naturally fermented *Miang* leaves in Thailand [30]. A close *P. deserticola* strain of unstated source in GenBank was from a large characterization study [31].

In both *Elaeis* sp. and *Raphia* sp. palm wine, several monophyletic groups were formed with other *Pichia* species namely *P. deserticola*, *P. Manshurica* and *P. galeiformis* which indicate polyphyletic relationships. The polyphyletic nature of *Pichia* has been demonstrated by Kurtzman and Robnett [29] in the analysis of gene sequences that included all known ascomycetous yeasts. Apart from possible similar conserved regions, previous nomenclature at the time of submission of the sequences may also be the reason why *Pichia* species of different genus were observed as close relatives of *C. ethanolica* from *Elaeis* sp. and *Raphia* sp. palm trees.

It has been reported that ascomycetic fungi submitted to the database previously have been- assigned names based on their life stages [32, 33]. For example, it was shown that the name for the fungi *Candida krusei* is based on the anamorphic stage whereas its telemorph stage

**Figure 1.** Phylogenetic analysis of *Candida ethanolica* (HG425332-underlined) from *Raphia* sp. palm wine. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site.

name is *Pichia kudriavzevii*. It also has an older name *Issatchenkia orientalis*. The whole *Candida*  species consists of up to 850 organisms, which can be distantly related [34]. Hence in order to avoid the confusion, the International Botanical Congress in Melbourne in July 2011, made a change in the international code of nomenclature for fungi and adopted the principle of one fungus can only have one name and ended the system of permitting separate names to- be used for anamorphs [35]. The report emphasized that this validated all legitimate names proposed for a species, regardless of what stage they were typed and can serve as the correct name for that species.

#### **3.3.** *Sachharomyces cerevisiae*

The yeast *S. cerevisiae* is generally known to be the most used microorganism in the food and drink manufacturing sector. The organism is the dominant yeast species isolated from many studies on palm wine. However, it is unclear whether *S. cerevisiae* as a species occurs naturally or exists solely as a domesticated species [36]. *S. cerevisiae* strains are genetically diverse, largely as a result of human efforts to develop strains specifically adapted to various fermentation processes. These adaptive pressures from various ecological niches may generate behavioral differences among these strains [37]. In a review [8], it was suggested that domestication in *Saccharomyces*, is most pronounced in beer strains, because they live in their industrial niche always and allow only limited genetic admixture with wild stocks and minimal contact with natural environments. Due to this restriction, it was pointed out that

**Figure 2.** Phylogenetic analysis of *Candida ethanolica* (HG425336-underlined) from *Elaeis* sp. palm wine. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site.

beer yeast genomes show complex patterns of domestication and divergence, making both ale (*S. cerevisiae*) and lager (*S. pastorianus*) strains ideal models to study domestication.

The relatives of palm wine *S. cerevisiae* was not distributed among many species or different genus observed for *Candida species*. Two nodes were observed for the *S. cerevisiae* trees constructed for *Elaeis* sp. (**Figure 3**) and *Raphia* sp. (**Figure 4**). The yeast strain isolated from *Elaeis* sp. (**Figure 3**) was in a different branch from most of its relative whereas it was vice versa for the palm wine yeast from *Raphia* sp. (**Figure 4**) palm wine. As observed for *Candida species*, isolation of *S. cerevisiae* species was from different sources. The close relatives flanking the palm wine strain from *Elaeis* sp. palm wine (HG425328, **Figure 3**) with accession numbers KU862639 and MF966566 were isolated from grape surface [38] and pear sough dough [39] whereas the close relatives of *Raphia* sp. palm wine (HG425338, **Figure 4**) with accession numbers GU080046 and HM191669 were isolated from must of spontaneous fermentation [40] and grape juice used to brew *Musalais,* a beverage made from compressed grapes [41]*.* 

It is believed that 99% of yeasts is still unknown [42], and *S. cerevisiae* fermentation could be specific to a particular substrate, hence more studies of *S. cerevisiae* from different palm trees will be beneficial. The genus *Saccharomyces* was previously divided into two groups namely


**Figure 3.** Phylogenetic analysis of *S. cerevisiae* (HG425328-underlined) from *Elaeis* sp. palm wine. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site.

*Saccharomyces sensu stricto* and *Saccharomyces sensu lato* and the sensu stricto strains are mostly associated with the fermentation industry [43]. The *S. cerevisiae* in this study are *sensu stricto*. Comparative genomics analysis of *S. cerevisiae* and closely related species has contributed to our understanding of how new species emerge and has shed light on various mechanisms that contribute to reproductive isolation [44]. This knowledge can be applied to palm wine yeasts to ascertain how they differ from well characterized yeasts.-

#### **3.4.** *Pichia kudriavzevii*

From recent molecular studies of yeasts present in palm wine, the yeast species *Pichia kudriavzevii*  has emerged as a prevalent non-*Saccharomyces* yeast species in the drink. The genus has shown probiotic potentials [45] and multistress-tolerance [46]. It is worth looking closely at this genus because it has been shown that some *P. kudriavzevii* strains can produce higher quantities of ethanol from lignocellulosic biomass than conventional cells of *S. cerevisiae* at 45°C [47].

The tree constructed for *P. kudriavzevii* showed the least divergence when compared to *S. cerevisiae* or *Candida* palm wine yeast relatives. All the relatives and the *Elaeis* sp. palm wine strain (HG425333) originated from one node and formed separate taxonomic units


**Figure 4.** Phylogenetic analysis of *S. cerevisiae* (HG425338-underlined) from *Raphia* sp. palm wine. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site.

(**Figure 5**). In contrast, the *P. kudriavzevii* (HG425335) from *Raphia* sp. palm wine formed a separate clade and did not lie on the same branch with the relatives (**Figure 6**). This indicates intraspecies diversity and confirms findings reported previously [11]. In that study, intraspecies diversity was suggested because *P. kudriavzevii* (HG425335) from *Raphia* sp. palm wine formed a separate clade with palm wine isolates from Mexico instead of isolates from the same geographical location.

The information contained in the sequence submission of close relatives of *P. kudriavzevii*  strains also shows different sources of isolation. The strains close to the yeast from *Elaeis* sp. palm wine (HG425333, **Figure 5**) with accession numbers KY283159 and KM234455 show- that isolation was from composite microbial powders for aquaculture [21] and naturally fermented cashew apple juice [48] whereas a close relative of *Raphia* sp. palm wine (HG425335, **Figure 6**) with accession number KU167717 was isolated from activated sludge from textile dyeing [49].

#### **3.5. Geographical origin and sources of palm wine yeast relatives**

After ascertaining the sources of very close relatives from the phylogenetic trees constructed, the shortlisted 600 sequences from the aforementioned yeast genera were further examined and the information found was used to group the isolates according to country of isolation, food, beverage, and non-edible source.

**Figure 5.** Phylogenetic analysis of *P. kudriavzevii* (HG425333-underlined) from *Elaeis* sp. palm wine. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site.

#### *3.5.1. Isolates submitted by country of origin and source-*

Overall, sequences examined for the aforementioned yeasts genera were submitted from 38- countries [18] and the top 6 countries is presented in this report. Sequence data for both *Elaeis*  sp. (**Figure 7**) and *Raphia* sp. (**Figure 8**) palm trees show that highest number of submissions to the Genbank database was from China. The top three countries from which palm wine yeast relatives originated were the same for both palm tree species. This suggests that a large number of palm wine yeasts may have common ancestors with yeasts found in China. The origins or sources of palm wine yeasts relatives were spread across beverages, food, and non-food sources. The prevalence of *S. cerevisiae*, *P. kudriavzevii*, and *C. ethanolica* from these sources is shown for *Elaeis* sp. palm tree (**Figure 9**) and *Raphia* sp. palm tree (**Figure 10**). In both palm wine from *Elaeis*  and *Raphia* palm trees, yeasts relatives of *S. cerevisiae* and *P. kudriavzevii* species were isolated mainly from beverage sources whereas relatives representing *C. ethanolica* species were isolated from non-food sources. The sources of isolation revealed that the closest relatives of palm wine yeasts were from various sources and not specific to any particular food or fermentation mix.-

**Figure 6.** Phylogenetic analysis of *P. kudriavzevii* (HG425335-underlined) from *Raphia* sp. palm wine. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site.

**Figure 7.** Top six countries from which sequences of palm wine yeast relatives of *Elaeis* sp. palm tree were submitted to the GenBank.

**Figure 8.** Top six countries from which sequences of palm wine yeast relatives of *Raphia* sp. palm tree were submitted to the GenBank.

**Figure 9.** Distribution of palm wine yeast relatives with reference to yeasts from *Elaeis* sp. palm wine according to beverage (), food (∎), and non-food (⊞) sources.

A report [50] found that laboratory estimates of optimum growth temperature could be used to predict global distributions of free-living microbes. Also, it was pointed out that population genetic analyses show that the genetic diversity of *S. cerevisiae* is high in the tropics and subtropics of China [51, 52]. It was suggested that without further sampling in tropical and subtropical regions, it is not possible to differentiate whether the higher diversity of *S. cerevisiae*  in Asia reflects a greater habitat area or an Asian origin for *S. cerevisiae*. It would be beneficial- to carry out further studies in order to establish if palm wine yeasts were taken from Africa to Asia or vice versa. The diversity could also be high in temperate regions because a study examined *S. cerevisiae* and *S. paradoxus* in northeast America and uncovered a large diversity of yeasts [53]. Up to 24 yeast isolates could not be assigned to any known species and it was suggested that the yeasts identified may be of taxonomic, medical, or biotechnological importance.-

**Figure 10.** Distribution of palm wine yeast relatives with reference to yeasts from *Raphia* sp. palm wine according to beverage (), food (∎), and non-food (⊞) sources.

#### **3.6. G+C composition of palm wine yeast relatives**

The G+C composition is a well known evolutionary property of eukaryotes, archaea, and bacteria. There are suggestions by Chen et al. [54], that concordance between proteomic architecture and the genetic code is related closely to genomic G+C content and phylogeny. It has been suggested that yeasts with higher G+C content have a higher recombination rate [55] and recombination is believed to be suppressed around centromeres [56]. The data in **Table 1** present the average nucleotide composition and G+C content of partial sequences of 26S rRNA genes analyzed. It shows concentration of arginine, guanine and thiamine, and cytosine concentration in *S. cerevisiae, P. kudriavzevii*, *or C. ethanolica* obtained from the aforementioned palm trees. Data were obtained after measuring nucleotide frequencies (%)


Nucleotide concentration was obtained after analysis with MEGA 7.0 software. T/U, thiamine/uracil; C, cytosine; A, arginine; G, guanine.

**Table 1.** Average nucleotide composition and G+C content obtained from yeasts from *Raphia* sp. (R) or *Elaeis* sp. (E) palm wine and their relatives after measuring nucleotide frequencies (%) in 100 sequences relative to each yeast species shown.

in 100 sequences of strains relative to each palm wine yeast species listed. It was observed that the G+C content in *P. kudriavzevii* and *C. ethanolica* was higher than that of *S. cerevisiae*. This suggests that the *P. kudriavzevii* and *C. ethanolica* have a higher recombination rate than *S. cerevisiae* strains analyzed in this report. The G+C range observed is within the reported average genomic G+C-content range (13–75%) among species [57]. It was also found to be within range of G+C content (38.3–52.9%) of the *MAT* locus reported [58] in different- *Saccharomycetaceae* species.

Further studies are required because G+C-content is associated with multiple biases of different nature during down stream operations and these biases may include sequencing technologies, biological, and methodological reasons [57]. Another factor that could affect- the G+C content is that some yeasts like *Lachancea kluyveri* show an intriguing compositional heterogeneity in that a region of the chromosome has an average G+C content of 52.9% which is significantly higher than the 40.4% global G+C content of the rest of the- genome [58].
