Yeasts from rumen of adult Zebu cows and fermentations of lignocellulosic materials

Commercial yeast strains are a promising alternative to chemical additives in ruminants; but are not self to the rumen ecosystem. In this study, the aim was to identify and evaluate the growth, mass production, and enzymatic activity of yeast strains naturally isolated from bovine rumen in submerged fermentations containing roughages with high concentration of lignocellulose. An experiment was initially conducted to compare the degradation of Urochloa decumbens (UD) hay or sugarcane bagasse (SB) by eight yeast isolates and a control without microorganisms. Subsequently, further experiments were performed to compare the degradation of the two substrates by two yeast strains (V5 and V62) selected (from the eight used in the initial screen) and a control without microorganisms for four different fermentation periods.

increased the pectinase activities at 96 h of fermentation of U. decumbens showing potential of industrial application and for animal diets. Additionally, V5 strain would be selected and represent potential probiotic for cattle raised in tropical pasture Background The complex ruminal microbioma may supply proteins, energy, and vitamins to the ruminants, which contribute to their growth and production. This ecosystem comprises known species of bacteria, filamentous fungi, and protozoa [1,2]. Nevertheless, little is known about the yeast species in the rumen, that constitute up to 6.76 log of colony forming unit (CFU) per ml of ruminal fluid in healthy cattle of different ages [3][4][5].
Research evaluating the incorporation of Saccharomyces cerevisiae yeast from exogenous sources in the ruminant diet has shown a reduction in the overall ruminal oxygen content and increased ruminal bacterial populations and microbial protein production [6]. This yeast may reduce lactic acid levels, to control the ruminal pH, and to favour cellulose digestion. This eukaryote also prevented diseases and reduced faecal excretion of pathogenic bacteria [7,8]. However, some of the commercial strains of S. cerevisiae have exhibited limited growth in the rumen [9,10].
The natural occurrence of yeast in the ruminal fluid has been reported for cattle fed with different sources of tropical roughage [3][4][11][12][13]. Surprisingly, we detected a significantly higher population of yeast in adult cows raised on pastures of Uroclhoa spp.
during the dry season than young cattle [4]. The characterisation of yeasts from the rumen of cattle that grazed on lignified pastures and having low nutritional value would support the selection of isolates that may facilitate the degradation of cell wall in roughages with high lignocellulose concentrations during dry seasons. In this study, our aims were to identify and evaluate the growth, biomass production, and enzymatic activity of yeast strains isolated from bovine rumen in submerged fermentations containing roughages with high concentrations of lignocellulose.

Identification and molecular characterization of yeasts from bovine rumen
In this study were identified three yeast species from the rumen of the cow fed on lignified UD pasture, considering morphology and physiological tests and molecular analyses. Among the identified species, Pichia kudriavzevii was the most common, followed by Rhodotorula mucilaginosa and Candida tropicalis ( Table 1). The phylogenetic analyzes of D1-D2 domain sequences of the 26S rRNA showed theses isolates clustered with reference strains of the their respective species and intraspecific variations were not detected ( Figure. 1).

Degradation of lignocellulosic substrates
In fermentations containing SB, no significant differences were detected for cell mass production between the evaluated yeast strains (Table 2). However, for fermentations with UD, the strains of R. mucilaginosa (V16 and V10) had lower mass production than other strains (p < 0.05). Yeast mass productions were significantly higher in fermentations using UD than SB (p < 0.001, Table 2).
The V5 and V10 and V62 (Pichia kudriavzevii) showed a higher rate of dry matter reduction of UD than other strains (p < 0.05); however, the strain V5 was the most efficient in dry matter reduction of SB ( Table 2). As expected, we detected a higher dry matter reduction of UD than SB by the yeast strains (p < 0.05).
The strain V5 (R. mucilaginosa) showed higher biomass production in the medium containing UD and together with V62 promoted more reduction of final pH of fermentation medium in comparison to other strains (p < 0.05, Table 2). The strain V12 (C. tropicalis) increased the final pH of medium containing UD, possibly through its proteolytic action.
We detected a low degradation rate of the evaluated substrates (< 3.99%).
In the second experiment, comparing the two selected strains, there were no significant variations of pH, xylanase and pectinase activities and in the total protein from the supernatants of the fermentations between these strains ( Figures. 2, 3 and 4). The isolates did not exhibit expressive xylanase activity under the evaluated fermentation conditions. However, the regression analysis showed that V5 reduced ph at 96 h of fermentation ( Figure. 3) and V62 and V5 increased the pectinase activities at 96 hs of fermentation ( Figure. Table 2 Mean of final pH, yeast biomass (mg), dry matter (dm) reduction and degradation rate (%) of sugarcane bagasse (SB) and Urochloa decumbens (UD) in submerged fermentation containing yeast from rumen fluid of Nelore cows Reduction of dm = 1-(final dm / initial dm) **Degradation rate = (dm reduction with yeast -dm reduction of control) x 100

Discussion
These three yeast species identified in this study have also been reported in the ruminal micobiota analysis of cattle fed with tropical forages [3][4][5]. Pichia kudriavzevii was also the most frequent yeast in the rumen of cattle from South of Minas Gerais, Brazil, which may be due to its ability to better adapt to ruminal conditions. All strains of this species showed growth in anaerobic conditions and in temperatures that predominate in the rumen [5].
The yeast R. mucilaginosa has also been detected in ruminal fluid samples from three fistulated cows and in the hay that was used to feed these animals [13]. In other study involving three fistulated Holstein cows, the researchers isolated yeast colonies and identified the Levica 18 strain, which was 98% similar to R. mucilaginosa [ 14].The genus Rhodotorula is common in the environment and can be frequent in soil, water, milk, fruit juices and air samples [15]. This species has the capacity to assimilate glucose, sucrose and galactose [16].
Strains of Candida tropicalis were also identified in a study of fistulated cows fed with Trijolium pratense L. [13]. In a analyse conducted in northern Minas Gerais, Brazil, evaluating the rumen microbiota of goats fed with tropical pasture, 90% of the total yeast isolated corresponded to the species Pichia membranifaciens and 10% to C. tropicalis [ 17].
A strain of C. tropicalis (BPU1) was also isolated from the rumen of the Malabari goat, showing dual production of biosurfactant and polyhydroxybutyrate in a simple mineral salt medium, using vegetable oil as the sole carbon source [18].
The occurrence of yeast in the rumen environment is still poorly supported in the scientific literature despite their high population, especially in adult cows raised on low quality pastures [4]. Studies associated with the rumen microbiota have frequently ignored the presence of these yeast fungi in the rumen ecosystem [5]. Although the three yeast species identified in this study have also been reported in other studies of this ruminant site, the role of these microorganisms should be better elucidated in the ruminal ecosystem of animals of different categories fed with different diets.
In a review, we verified that these yeast species are potential producers of enzymes involved in the oxidative degradation of lignocellulosic biomass such as superoxide dismutase and peroxidase based on its information available from the UniProt p database (Table 3) [19][20][21]. Genes coding for enzymes that act on the degradation of hemicellulose and lignin as 1,3-β-glycosidases, mannosidases, trehalases, esterases and deacetylases were reported in P. kudriavzevii, [19,20]. Sequencing of the complete genome of this species revealed that yeast has the potential for fermentation of xylose, xylitol dehydrogenase and xylulokinase, enzymes that are considered important in the production of second-generation ethanol, as reported by Chan et al., [19].
The cultivation conditions employed in this study were not favourable to a substantial production of the fibrolite enzymes, possibly due to the low nutrient contents of the lignocellulolytic substrates and of the minimum culture medium. Hence, future studies should focus on adding specific nutrients and promoting other cultivation conditions that would favour the better expression of enzymes and metabolites that would improve the degradation of these lignocellulose-containing materials. In this study, the strain V5 (R. Mucilaginosa, a non-pathogenic species) might be an ideal strain to be selected and evaluated with further tests to improve the digestibility of UD and to represent a natural yeast probiotic for cattle raised in tropical pasture.

Isolates of yeast from rumen
Eight yeast strains isolated from the rumen of Nellore cows reared in U. decumbens lignified pasture in northern Minas Gerais were evaluated for degradation of two roughages. These microorganisms were collected by puncture of the ventral rumen sac and isolated in our previous study [4] and were selected because of their high populations in the rumen environment of adult beef cows (> 6 log CFU per ml of ruminal fluid). These strains were grown in Sabouraud broth and stored in an ultra-freezer at -80º C and deposited and available in the Yeast Culture Collection of the Institute of Agricultural Sciences of the Universidade Federal de Minas Gerais.

Characterization and identification
Preliminarily we grouped the yeast strains according to colony morphology and physiological tests performed according to the procedures described by Kurtzman et al. [22]. For molecular analyses, the yeast isolates of rumen were grown on Sabouraud agar for seven days, and DNA was extracted according to a method previously described by Rosa et al., [23]. The D1/D2 region of rDNA was amplified by polymerase chain reaction (PCR) using primers NL1 (5'-GCA TAT CAA AAG GAA GAG TAA GCC-3 ') and NL4 (5'-GGT AAG CTT CGC TGT CCG G-3'), according to a method previously described by White et al., [24].The amplified product was quantified with a NanoDrop 1000ND (NanoDrop Technologies) and the concentration was adjusted to 100 ng µL − 1 for use in sequencing reactions.
Sequencing was performed with DYEnamic (Amersham Biosciences, USA) in a Mega-BACE 1000 automated sequencing system at the Genome Analysis Center and Gene Expression.
The rDNA sequences were analysed using BLASTn (v.2.215) of BLAST 2.0 at the National Center for Biotechnology Information (NCBI) website [25]. Conspecific strains differed by no more than three among the 500-600 nucleotides of the D1/D2 domains and the isolates with 99% sequence similarity to deposited sequences were considered as the same species [22].

Phylogenetic analyzes
The D1-D2 variable domain sequences of the 26S rRNA gene from the yeast isolates were used to reconstruct their phylogenies using the MEGA X version 10.1 (BETA) program [26].
Analyzes were performed individually with the sequences belonging to the Basidiomycota and Ascomycota, which were aligned using the Clustal W method [27]. For each genus, the alignments were performed by including sequences of other yeast strains deposited in GenBank. Akaike's information criterion was used to identify the most appropriate evolution model for the dataset of each phylum. These datasets were estimated by the Maximum Likelihood Method, based on the Tamura-Nei model [28] and discrete Gamma distribution was used to model the evolutionary rate differences between sites [5 categories (+ G, parameter = 0.4430)] for Basidiomycota and [5 categories (+ G, parameter = 0.3043)] for Ascomycota. Tree robustness was estimated by bootstrap analysis with 1000 replicates [29] and all nucleotide sequences were submitted to GenBank, assigned MN380262 to MN380269 accession numbers.

Lignocellulosic materials
The material of Urochloa decumbens (UD) hay was collected during the dry season (March to June) on a farm located in the Montes Claros region, Northern Minas Gerais, Brazil and the sugarcane bagasse (SB) was provided by alcohol and sugar company São Judas (SADAbio), located in the Jaíba city at same region. The species of theses forages were identified according to their morphological characteristics expressed in the flowering periods [30].
These fibrous materials were dried and grounded in a Willey knife mill to 1-3 mm particle size. Subsequently, subsamples were subjected to chemical composition analysis [30] as described in Table 4.  The inoculums of each yeast strain were prepared after incubation in Sabouraud agar medium (Acumedia, Lansing, Michigan) plus chloramphenicol (150 mg / l) for 48 hours.
Subsequently, yeast colony masses were added and incubated in Sabouraud broth at 37ºC for 48 hours. After this period, 3 mL of inoculum containing approximately 3.2 × 10 4 CFU / mL was added to tubes containing 30 mL of C medium that contained the lignocellulosic substrates SB or UD. In the control tubes, same volume of the culture medium with no inoculum was added.
The tubes were incubated in a shaker (NT 715, Novatécnica, São Paulo, Brazil) at 37 ºC and 150 RPM for seven days. The pH of the medium was measured before and after incubation, using a digital potentiometer (pH 1800, Waterproof Pen pH Tester, Instrutherm, São Paulo, Brazil).
After this period, the tubes were centrifuged to separate the yeast mass, which, together with the non-woven textile (NWT − 100 g/m2) bags., were kept in a circulating oven (TE-394/3, Tecnal, São Paulo, Brazil) for four days at 40 ºC until they presented constant weight. Subsequently, the dry yeast mass and dry matters of SB or UD were obtained using a moisture determinant (MOC63u Shimadzu, Kyoto, Japan).

Production of enzyme extracts
In the second experiment, the strains V5 (Rhodotorula mucilaginosa) and V62 (Pichia kudriavzevii) were selected, considering their higher degradation rates of UD and reduction of pH by them in the firth experiment. Submerged fermentation occurred as previously described, comparing the two sectioned yeast strains and the two carbon sources (SB or UD) at 0, 48, 96, 144 and192 hours of incubation.
After each fermentation period, the tubes containing the yeast and their controls were

Statistical analysis
This first experiment was conducted in a 2 × 9 factorial design with four replications, comparing the eight yeast isolates and the control (without microorganisms), and the degradation of the two substrates (UD or SB). The variables observed were pH, dry yeast mass, dry mater (dm) of lignocellulosic material residues and their degradation rate.
Reduction of dm = 1-(final dm / initial dm) Degradation rate = (dm reduction with yeast -dm reduction of control) x 100 In the second experiment, a 2 × 3 × 4 factorial design was performed to compare the two carbon sources, the two yeast strains selected from the first experiment and the control without yeast, in four periods, with four replications. The variables analysed were pH, pectinase and xylanase activity and total protein production. After normality and homogeneity test, the data were subjected to analysis of variance test and the means were compared to a 5% of significance. Data were analysed also by regression analyses using the SAEG statistical package,

Availability of data and materials
The datasets used and/or analyzed during the current study and yeast strains are available from the corresponding author on reasonable request

Authors' contributions
ERD conceived, designed and coordinated the study. SML, JCN, HARM, CES, ADLJ participated in data collection, analyses and drafting the manuscript. ERD, JCS and VLS analyzed, wrote and submitted the manuscript for publication. All the authors revised and approved the final manuscript.

Ethical Approval
The yeast isolates were collected in previous experiments submitted and approved by the

Consent for publication
Not applicable.