Comparative Analysis of the Alkaline Proteolytic Enzymes of Yarrowia Clade Species and Their Putative Applications

Proteolytic enzymes are commercially valuable and have multiple applications in various industrial sectors. The most studied proteolytic enzymes produced by Yarrowia lipolytica, extracellular alkaline protease (Aep) and extracellular acid protease (Axp), were shown to be good candidates for different biotechnological applications. In this study, we performed a comprehensive analysis of the alkaline proteolytic enzymes of Yarrowia clade species, including phylogenetic studies, synteny analysis, and protease production and application. Using a combination of comparative genomics approaches based on sequence similarity, synteny conservation, and phylogeny, we reconstructed the evolutionary scenario of the XPR2 gene for species of the Yarrowia clade. Furthermore, except for the proteolytic activity of the analyzed Yarrowia clade strains, the brewers’ spent grain (BSG) was used as a substrate to obtain protein hydrolysates with antioxidant activity. For each culture, the degree of hydrolysis was calculated. The most efficient protein hydrolysis was observed in the cultures of Y. lipolytica, Y. galli, and Y. alimentaria. In contrast, the best results obtained using the 2,2-azinobis (3-ethyl-benzothiazoline-6-sulfonic acid (ABTS) method were observed for the culture medium after the growth of Y. divulgata, Y. galli, and Y. lipolytica on BSG.


Introduction
Nearly two decades ago, the genus Yarrowia was still limited to Yarrowia lipolytica species [1,2]. In 2004, Péter et al. [3] described Candida galli, a sister species of Y. lipolytica, which was later renamed Y. galli. In 2005, Cletus Kurtzman [4] defined the Yarrowia clade as a taxonomic group, including Y. lipolytica and C. galli, as well as Candida hispaniensis and Candida bentonensis. Since then, numerous species have been assigned to the Yarrowia clade primarily based on marker sequencing, which has become a common tool in taxonomic studies. Today, the Yarrowia clade comprises Y. lipolytica, Yarrowia alimentaria, Yarrowia bubula, Yarrowia brassicae, Yarrowia deformans, Yarrowia divulgata, Y. galli, Yarrowia hollandica, Yarrowia keelungensis, Yarrowia osloensis, Yarrowia parophonii, Yarrowia phangngaensis, Yarrowia porcina, Yarrowia yakushimensis, and C. hispaniensis [5]. The species listed above share many morphological and physiological characteristics of Y. lipolytica. One is the ability to exist as yeast cells, pseudomycelium, or true septate hyphae. However, the best known feature of all the species is their high capacity for lipid accumulation, which are the carbon storage compounds accumulated during nitrogen-limiting conditions [6,7]. Yarrowia clade species vary in terms of the maximum growth temperature, ranging from 21 • C for Y. alimentaria to 37 • C for Y. phangngaensis [7].
Most of the species belonging to the Yarrowia clade originate from European, American, or Asian regions, being isolated from various natural habitats, which are also common for Y. lipolytica, such as dairy and meat products (Y. alimentaria, Y. osloensis, Y. divulgata, Y. galli, Y. bubula, and Y. porcina), mangrove or marine waters (Y. keelungensis, Y. phangngaensis and Our research provides a comprehensive analysis of the alkaline proteolytic enzymes of Yarrowia clade species, including phylogenetic studies, synteny analysis, and protease production and application. Furthermore, in order to address the economic aspects of protease production, a protein-rich spent grains medium was selected. Finally, considering the potential applications of Yarrowia proteases, we established a link between the high degree of protein hydrolysis and antioxidant activity of the obtained hydrolysates.

Comparative Analysis of Yarrowia Alkaline Proteases from a Genomic Point of View
The search for alkaline proteases encoding genes in the genome of the different Yarrowia species showed the presence of 5 (Y. phangngaensis) to 21 (Y. galli) genes (Table 1). Five genes were also found in C. hispaniensis, which has a basal position in the clade. Due to the lack of access, genomes of Y. brassicae and Y. parophonii were not examined in terms of alkaline protease genes. Instead, three strains of Y. lipolytica were analyzed: the Polish strain A101, the German strain H222, and the laboratory strain E150 (CLIB 122) issued from a crossing between the French W29 and the American ATCC 18942 followed by many back-crosses [8]. The genome of E150 was the first fully sequenced genome and is still the reference for genome assembly and annotation of new Y. lipolytica genomes.  16 3 Y. lipolytica H222 16 4 Y.alimentaria CBS 10151 19 5 Y. bubula CBS 12934 16 6 Y. deformans CBS 2071 15 7 Y. divulgata CBS 11013 15 8 Y. galli CBS 9722 21 9 Y. hollandica CBS 4855 15 10 Y. keelungensis CBS 11062 17 11 Y. osloensis CBS 10146 19 12 Y. phangngaensis CBS 10407 5 13 Y. porcina CBS 12935 19 14 Y. yakushimensis CBS 10254 15 15 C. hispaniensis CBS 9996 5 Based on sequence similarity, we defined 13 different groups of putative alkaline proteases as shown on the phylogenetic tree ( Figure 1).
Most groups contained one Y. lipolytica representative except for three groups with two proteins with multiple counterparts that could not be separated based on sequence similarity. One of these groups corresponds to the XPR2 group, i.e., the Aep encoded by YALI0F31889g, part of which (highlighted by a blue curve in Figure 1) gathered one protein per species except for Y. phangngaensis and C. hispaniensis for which no ortholog was found. The genetic distances between these genes are small, as shown by the short branches in the tree. The other proteins in the XPR2 group exhibited very long branches representing proteins emanating from successive duplication events, which occurred in Y. osloensis, Y. porcina, Y. galli, and Y. yakushimensis. Synteny analysis of these genes showed different insertion sites suggesting that these duplications occurred independently ( Figure S1). One exception to this rule is the conservation of synteny in the upstream region of YAOS0S01e24608g and YAPO0S01e09780g genes, which are probably derived from the same event ( Figure S1). In Y. yakushimensis, YAYA0S05e13234g, the syntenic copy of XPR2, was successively duplicated three times on scaffold YAYA0S6. These duplication events involved not only XPR2 but also the homolog of YALI0F31845g (F31845 in Figure S1).
In contrast, there is extensive synteny conservation in the region of XPR2 close homologs (part of the XPR2 group with short branches) with some insertions of tRNA genes, transposable elements, and intervening genes ( Figure 2). Nevertheless, we observed a rearrangement occurring one gene upstream of the XPR2 homolog (upstream of YALI0F31867g) in Y. bubula. An additional rearrangement generated a loss of synteny one gene downstream of XPR2 (downstream from YALI0F31911g). As synteny is conserved in Y. bubula, Y. hollandica, and Y. alimentaria, which are basal species of the genus Yarrowia, the most parsimonious explanation is that the rearrangement occurred in the other species (grey arrow in Figure 2), leading to the actual order of genes in Y. lipolytica and closely related species.
in Y. bubula, Y. hollandica, and Y. alimentaria, which are basal species of the genus Yarrowia, the most parsimonious explanation is that the rearrangement occurred in the other species (grey arrow in Figure 2), leading to the actual order of genes in Y. lipolytica and closely related species. The phylogenetic tree was constructed from an alignment of 446 amino acids with phyML with a LG evolutionary model. XPR2 orthologs are in red and flanking genes are in different blue or yellow to brown colors reflecting orthologs in the different species. Transposable elements (MITE, LTR, LINE) and tRNA genes are represented as described in the legend. The grey arrow close to the phylogenetic tree indicates the rearrangement event leading to the actual gene order in Y. lipolytica and closely related species.

Expression Analysis of Y. lipolytica W29 Alkaline Proteases in Various Culture Conditions and Signal Peptide Prediction
To verify which of the 16 putative proteases may play an important role in protein hydrolysis by the analyzed strains, we took advantage of previously published RNA-seq data of Y. lipolytica W29 [26] and unpublished data performed on a derivative of E150 (CLIB 122) (genotype: E150, ku70::URA3,LEU2). Twelve putative protease encoding genes showed low or no expression ( Table 2). The XPR2 gene is highly expressed in minimal medium with glucose but poorly in oleic acid and tributyrin. YALI0A06435g and YALI0A08360g were moderately expressed on the three tested carbon sources in yeast nitrogen base (YNB) medium. Surprisingly, YALI0B16500g, a homolog of Saccharomyces cerevisiae PRB1-protease B, is highly expressed in YPD medium at high and low pH. Previous studies showed that XPR2is induced by pH above 6 and/or by the presence of peptone [27]. Thus, we checked the expression of all putative protease encoding genes in these conditions and, as expected, XPR2was expressed on the YPD medium, although moderately, probably due to the genotype of E150 (CLIB 122) derivative (his-1, ura3-302, leu2-270, xpr2-322,ku70::URA3, LEU2). In turn, low expression of XPR2 was observed on YPD glucose at pH 4 ( Table 2). The highest expression was observed for YALI0B16500g on both YPD and YPD at pH 4. Furthermore, the functional annotation of YALI0B16500g predicts The phylogenetic tree was constructed from an alignment of 446 amino acids with phyML with a LG evolutionary model. XPR2 orthologs are in red and flanking genes are in different blue or yellow to brown colors reflecting orthologs in the different species. Transposable elements (MITE, LTR, LINE) and tRNA genes are represented as described in the legend. The grey arrow close to the phylogenetic tree indicates the rearrangement event leading to the actual gene order in Y. lipolytica and closely related species.

Expression Analysis of Y. lipolytica W29 Alkaline Proteases in Various Culture Conditions and Signal Peptide Prediction
To verify which of the 16 putative proteases may play an important role in protein hydrolysis by the analyzed strains, we took advantage of previously published RNA-seq data of Y. lipolytica W29 [26] and unpublished data performed on a derivative of E150 (CLIB 122) (genotype: E150, ku70::URA3, LEU2). Twelve putative protease encoding genes showed low or no expression ( Table 2). The XPR2 gene is highly expressed in minimal medium with glucose but poorly in oleic acid and tributyrin. YALI0A06435g and YALI0A08360g were moderately expressed on the three tested carbon sources in yeast nitrogen base (YNB) medium. Surprisingly, YALI0B16500g, a homolog of Saccharomyces cerevisiae PRB1protease B, is highly expressed in YPD medium at high and low pH. Previous studies showed that XPR2 is induced by pH above 6 and/or by the presence of peptone [27]. Thus, we checked the expression of all putative protease encoding genes in these conditions and, as expected, XPR2 was expressed on the YPD medium, although moderately, probably due to the genotype of E150 (CLIB 122) derivative (his-1, ura3-302, leu2-270, xpr2-322, ku70::URA3, LEU2). In turn, low expression of XPR2 was observed on YPD glucose at pH 4 ( Table 2). The highest expression was observed for YALI0B16500g on both YPD and YPD at pH 4. Furthermore, the functional annotation of YALI0B16500g predicts it to be localized in the vacuole. Thus, we checked all proteases for the presence of a putative signal peptide. It appeared that all of the putative proteases, including YALI0B16500p, could potentially be secreted ( Table 2).  Expression on YNB medium with OA, GLU, and TB was performed with W29 [26], whereas expression on YPD and YPD at pH 4 was performed with a derivative of E150 (CLIB 122) of genotype E150, ku70::URA3, LEU2 (Neuvéglise, unpublished). OA-oleic acid; GLU-glucose; TB-tributyrin.

Analysis of Expression of XPR2Alkaline Protease in the BSG Medium by Y. lipolytica
Expression analysis of all the genes encoding putative proteases in the BSG-based medium was performed. Due to the large number of genes to be tested, we analyzed the expression only in the Y. lipolytica W29 strain. Out of the 16 genes encoding putative proteases, 7 were expressed under the analyzed conditions. The highest expression was observed for XPR2, followed by the YALI0B16500 gene ( Figure 3). These data are, however, in contrast with the proteomic studies, in which the analyzed bands on SDS-PAGE were covered mostly with Aep protein (details see Section 2.7). These results were further verified by overexpression of the YALI0B16500 gene in the PO1d strain (deleted for the XPR2 gene), where no protease activity was detected on skimmed milk agar medium at pH 6.8.

Investigation of the Proteolytic Activity of Yarrowia Clade Species
The proteolytic activity of the Yarrowia clade was analyzed using skimmed milk agar plates. Results showed huge diversity of the clade members in terms of their abilities to secrete proteases (Table 3). Y. phangngaensis, Y. deformans, and C. hispaniensis showed no proteolytic activity, indicated by the lack of proteolysis zones on the skimmed milk agar ( Table 3). The highest proteolytic activity was observed for Y. yakushimensis (15.05 ± 0.52), followed by Y. lipolytica (14.80 ± 0.40), Y. keelungensis (14.02 ± 0.04), and Y. alimentaria (13.51 ± 1.01) ( Figure 4). The proteolytic activity of fungal, yeast, and bacterial strains derived from an artisanal raw ewe's milk cheese was investigated by Ozturkoglu-Budak and colleagues [28]. Among the five tested yeast species, only Y. lipolytica AF_335977 and Debaryomyces hansenii AF_210327 were able to secrete proteases. However, the proteolytic activity of AF_335977 was described as medium. At the same time, it was the best result among tested yeast species. In the studies by Gardini and colleagues [29], several yeast species isolated at different stages of Pecorino Crotonese production and ripening were characterized by their technological properties. However, only five strains, identified as Y. lipolytica and D. hansenii, were able to hydrolyze skimmed milk proteins.
Furthermore, Y. lipolytica and numerous other yeast species isolated from starter-free raw cow's milk cheeses were investigated for extracellular proteolytic activity. The highest activity of proteolytic enzymes was observed for Y. lipolytica and K. lactis, while the highest caseinolytic activity was presented by Y. lipolytica and P. guilliermondii [30]. Similarly, the proteolytic activity of yeast species isolated at different stages during the processing of pastirma (the most popular dry-cured meat product in Turkey) was investigated [23]. The yeast isolates, classified as Candida zeylanoides (the predominant species), C. deformans, C. galli as well as C. alimentaria (currently classified as Yarrowia genus), showed clear zones surrounding the colonies. Our results clearly confirm the potential of some Yarrowia species to secrete high amounts of proteases into the growth medium.

Characterization of the Proteolytic Activity of Yarrowia Clade Growing on BSG
The Yarrowia clade members differed significantly in terms of their proteolytic abilities. Y. phangngaensis, Y. deformans, and C. hispaniensis showed no proteolytic activity under the analyzed conditions ( Figure 5). In Y. lipolytica W29, XPR2 is the highest expressed gene in the BSG medium ( Figure 3). The absence of proteolytic activity in Y. phangngaensis and C. hispaniensis could be explained by the loss of the XPR2 homolog (Figure 1), which is, however, not the case in Y. deformans. The latter species was analyzed for the expression of XPR2 ortholog (YADE0S25e00298g) in BSG medium. However, no transcript was detected.
The proteolytic activity of Y. hollandica, Y. porcina, Y. yakushimensis, Y. bubula, Y. osloensis, Y. divulgata, and Y. brassicae did not exceed 0.5 U/mL ( Figure 5). The highest proteolytic activity was detected for Y. lipolytica W29 with 2.47 U/mL/min ( Figure 5). Other Yarrowia species showed strong proteolytic activities: Y. galli (1.96 U/mL/min), Y. alimentaria (1.38 U/mL/min), Y. keelungensis (1.07 U/mL/min), and Y. parophonii (0.80 U/mL/min). These results allow us to consider the above-mentioned strains as interesting producers of proteolytic enzymes with potential industrial applications in the future. In the cultures of Y. lipolytica, Y. alimentaria, and Y. parophonii, the proteolytic activity remained stable throughout the process, indicating continuous enzyme secretion and protein hydrolysis. In turn, in cultures of Y. galli and Y. keelungensis, a significant decrease in protease activity after reaching the maximum point was observed. It is possible that Y. galli and Y. keelungensis secreted a significant amount of enzymes at the beginning of the culture, allowing rapid protein hydrolysis and providing sufficient nutrients to decrease protease secretion.
The proteolytic activity of Y. lipolytica W29 and Y. lipolytica IMUFRJ 50862, isolated from an estuary near Rio de Janeiro, were analyzed in the research of Braga and colleagues [31], who investigated the effect of lipase secretion on the γ-decalactone production. In a medium optimized for lipase production, the proteolytic activity reached 0.0034 U/mL/min and 0.00151 U/mL/min for W29 and IMUFRJ 50862, respectively. The difference in results obtained in our work, caused by the medium composition, clearly indicates the potential for using BSG as a substrate.
In the present work, high proteolytic activity was associated with significant hydrolysis of BSG protein. Furthermore, it caused the accumulation of hydrolysis products, expressed as free α-amino groups. In the culture of Y. lipolytica, Y. galli, Y. alimentaria, and Y. keelungensis, characterized already as leading proteolytic strains, the concentration of α-amino groups reached 4367µg/mL, 4245 µg/mL, 3939 µg/mL, and 3885 µg/mL, respectively ( Figure 6). Interestingly, in our previous work using proteolytic bacteria, the concentration of free α-amino groups in the same BSG-based medium never exceeded 2600 µg/mL [32]. This result clearly shows the benefits of using Yarrowia clade yeast when applying BSG.
Eight strains of Y. lipolytica, isolated from goat/water buffalo mozzarella cheese, were previously studied by Suzzi and colleagues [33] for their ability to breakdown milk proteins. Y. lipolytica isolates showed weak proteolytic activity (expressed as leucine concentration [mg] per 100 mL of milk) after 8 days of the process. However, between the 8th to the 14th day, a huge increase in proteolytic activity was observed. After 14 days, the concentration of leucine reached 300-600 mg/100 mL, while in our research, the concentration of free α-amino groups in the BSG medium reached a maximum of 350-450 mg/100 mL for the most active strains ( Figure 6). It is worth pointing out that the time of culture in our research was only 5 days. This result further confirms the proteolytic abilities of Yarrowia clade members. For six out of fifteen tested species, the degree of BSG protein hydrolysis (DH%) exceeded 50% (Figure 7). The best results were observed for Y. lipolytica with 72.78% hydrolyzed proteins. Therefore, we conclude that the hydrolysis performed by Yarrowia species enabled a major protein breakdown. Furthermore, in similar studies with Rhizopus sp., DH% of the BSG protein reached 59.0 ± 5.4% [34]. This further confirms that BSG proteins are strong inducers of the Yarrowia clade proteolytic enzymes and, thus, that it is an appropriate substrate for optimization studies. Moreover, given that filamentous fungi are common producers of hydrolytic enzymes, the results obtained by Y. lipolytica, Y. galli, Y. alimentaria, and Y. keelungensis are more than satisfactory.
The protein hydrolysis in the cultures of Yarrowia clade members was associated with environmental alkalization (Figure 8). As a result, the final pH fluctuated between 8.5 and 9. Only in the Y. phangngaensis, Y. deformans, and C. hispaniensis culture, where a lack of proteolytic activity and no hydrolysis of the BSG proteins were observed, did the pH not exceed 8. Medium alkalization in the cultures of Y. lipolytica confirmed the previous studies in which K. lactis, D. hansenii, Pichia guilliermondii, Pichia fermentans, S. cerevisiae, Candida spp., and Rhodotorula spp., as well as 21 strains of Y. lipolytica, were analyzed for their properties influencing cheese production [30]. As a result, 16/21 and 10/21 of Y. lipolytica strains showed alkalizing properties on citrate agar and litmus milk, respectively.
To summarize the current information, species such as Y. lipolytica, Y. alimentaria, Y. galli, and Y. keelungensis are the most efficient and universal (in terms of medium composition) proteolytic enzyme producers.

Profile of Yarrowia Clade Proteolytic Enzymes Secreted in BSG-Based Medium
The gelatin zymography allowed for the detection of several proteolytic enzymes produced extracellularly during the growth of Yarrowia clade species in BSG medium ( Figure S2). Molecular masses of proteolysis bands observed on zymography gel are summarized in Table 4. It is important to underline that some proteolysis bands represent protease complexes formed as a result of interactions between enzyme monomers. This can be confirmed in the Y. lipolytica example. The molecular weight of Y. lipolytica proteases is 40-90 kDa. Therefore, producing an enzyme with a molecular weight of 110 or 130 kDa is impossible. Nevertheless, based on obtained results, it is possible to define the enzymes/enzymes complex involved in BSG protein hydrolysis.
As shown, Y. lipolytica and Y. alimentaria were the most active species in proteolytic activity and the most efficient in BSG protein hydrolysis. The protease profile of Y. lipolytica and Y. alimentaria was represented by two main proteolytic bands at 110 and 130 kDa (Table 4). However, a proteolytic band at 130 kDa was also detected in Y. brassicae and Y. divulgata culture, where low proteolytic activity and DH% < 30 were estimated. Therefore, we suggest a key role of the 110 kDa protein complex in BSG protein hydrolysis.
Furthermore, strong proteolytic activity in BSG medium as well as significant DH% of BSG protein, was demonstrated for Y. galli. The profile of proteolytic enzymes of Y. galli was represented by two bands with a molecular mass of 37-39 kDa and 40 kDa (Table 4). Nevertheless, the production of 40 kDa protease of low proteolytic activity (<0.5 U/mL) was determined in the Y. osloensis culture medium. Following this, protease of 37-39 kDa seems significant for BSG protein hydrolysis. It is important to underline that the same enzyme profile as Y. galli was observed for Y. parophonii, in which proteolytic activity (0.80 U/mL/min) was almost twice as low as Y. galli (1.96 U/mL/min). An explanation for this is the difference in the expression level of the 37-39 kDa enzyme.
Finally, a protease of molecular mass 30 kDa, secreted in Y. keelungensis (Table 4), is a crucial factor affecting BSG protein hydrolysis as indicated by high proteolytic activity and significant DH%. Production of proteolytic enzymes of 30 kDa was also observed in the Y. porcina culture medium. However, based on the results of the proteomic analysis presented in Section 2.7, enzymes of Y. keelungensis and Y. porcina are, with a high probability, different compounds of similar molecular weight.
Production of proteolytic enzymes in Y. lipolytica was already reported. An alkaline extracellular protease of 35 kDa was produced by Y. lipolyticaYITun15 [35]. This enzyme showed significant activity in a wide spectrum of factors tested. A Y. lipolytica coconut copra isolate produced protease of 61.5 kDa. The fish flower was applied as a substrate and enzyme production inductor [36]. In other studies, a 28 kDa protease was secreted by an auxotrophic recombinant strain of Y. lipolytica Po1d (JMY 1852). The enzyme was produced as an undesirable factor affecting human interferon alpha 2b production [37]. Another example involved the production of the intercellular protease of 97 kDa by Y. lipolytica strain IPN-MI-1 of the yogurt origin. The enzyme showed stable activity in the pH range of 5.0-9.0, and 37 • C was the optimal temperature [38].
In order to conduct proteomic analysis of Yarrowia clade enzymes, protein electrophoreses were performed. Unfortunately, despite a significant increase in protease concentration obtained using acetone precipitation, some of the proteases detected by gelatin zymography did not appear on the electrophoretic gel. Therefore, only the samples presented in Table S1 were subjected to proteomic investigation.

Protein Sequence Coverage of Yarrowia Clade Proteases to Other Fungal Enzymes
Proteomic analysis of the enzymes expressed in the BSG medium culture of Y. lipolytica, Y. keelungensis, and Y. alimentaria showed high protein sequence similarity to Aep of Y. lipolytica E150 (CLIB122) encoded by XPR2 (Table S1). The amino acid sequence of the Y. lipolytica enzyme complex of 130 kDa and 110 kDa was 40% and 44% aligned to the sequence of Aep, respectively. Moreover, the protein sequence coverage of the Y. alimentaria enzyme complex (130 kDa) and Y. keelungensis enzyme (30 kDa) to Aep reached 42% and 35%, respectively.
In the analysis of these results, two factors were important. First, we used public databases of fungal proteins to analyze the proteomic data, but the proteomes of Yarrowia species are not available, except those of Y. lipolytica strains. Despite this, the results of the proteomic analysis confirmed close affinity between Yarrowia species. Moreover, matched peptide sequences were found in the position range of 158-454, corresponding to the proper chain of Aep, which is converted into the active form of the enzyme upon signal peptide and propeptide cleavage.
The proteolytic enzyme of Y. osloensis (65 kDa) and Y. parophonii (65 kDa) were, to a great extent, 32% and 26%, respectively, aligned to alkaline extracellular protease of Yarrowia sp. E02 and Yarrowia sp. B02, while sequence coverage of Y. porcina (30 kDa) protease to the Carboxypeptidase of Y. lipolytica YB392 coded by B0I71DRAFT_162965 gene reached 25%. The full set of Yarrowia clade enzymes proteomic analysis results is available in the Supplementary Materials (Table S1).

Antioxidant Activity of BSG Protein Hydrolysates
Three methods were applied in order to estimate the antioxidant activity of BSG protein hydrolysates obtained in the cultures of Yarrowia clade species. Significant antioxidant activity was detected using the ABTS method. The highest activity, reaching 2606.13 µM TEAC/g, was detected in the Y. divulgata culture medium (Figure 9). This result was followed by the activity of Y. galli, Y. keelungensis, and Y. lipolytica samples, reaching 1771.87 µM TEAC/g, 1730.73 µM TEAC/g, and 1722.63 µM TEAC/g, respectively. It was more than twice higher than the activity of the control sample (603.73 µM TEAC/g). According to our previous research, the antioxidant activity of the control samples is caused by the partial hydrolysis of BSG protein occurring during media sterilization [32]. High antioxidant activity of post-culture media of Y. galli, Y. keelungensis, and Y. lipolytica strains resulted mainly from the efficient hydrolysis of spent grain protein (DH% > 60%), which is not the case of Y. divulgata, where DH% did not exceed 30%. In studies where protein hydrolysis is carried out directly in microbial cultures, the antioxidant activity of hydrolysates may be affected by the products of microbial metabolism. BSG is a waste material containing, in addition to protein, cellulose (17-25%), non-cellulosic carbohydrates (25-35%), lignin (8-28%), lipids (10%), and ash (5%) [39]. Therefore, due to the hydrolysis, other compounds, such as phenolics, may be exposed or released to the culture medium affecting the antioxidant activity of hydrolysate [40]. Y. divulgata is the object of extensive studies concerning erythritol production [22,41]. According to den Hartog et al. (2010), erythritol was recognized to possess strong antioxidant activity [42]. Therefore, simultaneous production of erythritol as a process accompanying BSG protein hydrolysis should be considered possible.
In our previous work, the antioxidant activity of BSG protein hydrolysates, obtained using the ABTS method directly in proteolytic bacteria cultures, ranged from 345.18 to 1621.31 µM TEAC/g peptides [32]. In this work, most of the analyzed samples showed a total activity of about 1500 µM TEAC/g. Therefore, using some Yarrowia clade yeast proteolytic enzymes allowed us to obtain hydrolysates with higher antioxidant activity compared to bacterial enzymes. It is likely related to DH%, reaching higher values in the yeast cultures, showing that high antioxidant activity is strongly associated with a high degree of protein breakdown.
The other two methods did not demonstrate the significant antioxidant activity of hydrolysates compared to the control. The best result obtained using the DPPH method is 144.90 µM TEAC/g in the culture medium of Y. hollandica, i.e., almost 10 times lower than the activity detected using the ABTS method ( Figure 9). The antioxidant activity of Yarrowia clade culture samples did not reach the control activity (147.80 µM TEAC/g). However, these results align with the work of Wołosiak and colleagues [43], in which the limitation in the solubility of peptides and amino acids in the system of DPPH assay (free radical resuspended in ethanol solution) is underlined. In addition, numerous problems related to the formation of iridescence in the presence of amide compounds were described. Therefore, the authors recommended the ABTS method as appropriate for the research on the samples containing significant quantities of peptides. On the other hand, due to large differences in the DPPH and ABTS systems, when possible, it is still advisable to apply both methods.
The antioxidant activity of BSG protein hydrolysates was evaluated using a ferric reducing antioxidant power (FRAP) assay in the work of Vieira and colleagues [44]. Activities of hydrolysates reached 0.062 mg TE/mg dw (247.71 84 µM TEAC/g), 0.098 mg TE/mg dw (391.55 µM TEAC/g), and 0.101 mg TE/mg dw (403.53 µM TEAC/g) when using brewery spent yeast extract, neutrase, and alcalase, respectively, for hydrolysis purposes. According to Vieira and colleagues, FRAP of protein hydrolysates depends not only on the molecular size of peptides but on the presence of specific peptide sequences and the amino acid composition of the hydrolysis products. Antioxidant activities measured in Yarrowia clade cultures ranged between 140-240 µM TEAC/g using the FRAP method, with the highest result obtained for the hydrolysate of Y. divulgata (233.13 µM TEAC/g) (Figure 9). In addition, the obtained activities were not significantly different from the control sample (181.46 µM TEAC/g). Presumably, due to their chemical properties, the peptides in our samples did not precipitate in the reduction reaction of Fe 3+ into Fe 2+ as electron donors, which is the basis of the FRAP method.
Nowadays, antioxidant peptides are increasingly applied in the production of nutraceuticals and functional food ingredients [45], pharmaceuticals [46], and in the field of dermatology [47]. As an aging population, we search for ingredients of increased potency. Therefore, peptides of the activity that result in diminished lines and wrinkles, smoother skin texture, and reduced redness and skin discoloration are highly sought after in the skincare industry [47]. Moreover, the unstable and reactive nature of free radicals results in cell damage leading to diseases such as hypertension, cardiovascular disease, cancer, diabetes mellitus, and neurodegenerative and inflammatory diseases. Therefore, antioxidant peptides protect the body by scavenging free radicals and inhibiting lipid peroxidation reactions, thus preventing oxidative damage [46]. In addition, antioxidant peptides are also applied in food systems to preserve food products against lipid oxidation leading to the development of undesirable off-flavors, odors, and potentially toxic reaction products [45]. Therefore, we believe that the peptides obtained as a result of BSG protein hydrolysis can be used in the food, pharmaceutical, and medical industries.

Bioinformatic Analysis of Yarrowia Clade Proteolytic Enzymes
Protease sequences were extracted from the Yarrowia genomes previously published byČervenák and colleagues [48] by sequence homology to alkaline extracellular protease XPR2 (YALIA101S03e00364g) from Y. lipolytica A101. The blastp threshold for E-value was set to 1 × 10 −20 . Protein sequences were aligned using Clustal Omega [49] in the multi-platform graphical user interface Seaview version 4.6.2 [50]. Gblocks [51] was used to select blocks of conserved sites useful for phylogenetic analysis. Trees were reconstructed by maximum likelihood using PhyML 3.0 [52] with an LG evolutionary model. Robustness of the trees was assessed by the approximate likelihood ratio test approach (aLRT). Synteny analyses were done manually using Artemis genome browser [53].

Analysis of Alkaline Proteases Expression in Y. lipolytica Grown in Various Culture Conditions and Prediction of Putative Signal Peptide
Previously published data [26] were used to estimate the expression level of each Y. lipolytica W29 alkaline protease on yeast nitrogen base (YNB) with oleic acid, glucose, and tributyrin. Additional unpublished data from cDNA sequence analysis (Neuvéglise and colleagues) were used for analyzing the expression of the same genes in YPD and YPD at pH 4.0. Signal peptide predictions were performed with SignalP-5.0 [54] and DeepLoc-1.0 [55].

Analysis of Alkaline Proteases Expression in Y. lipolytica Grown on Spent Grains
The analysis of gene expression encoding verified and putative alkaline protease in Y. lipolytica was performed in the yeast cells growing on BSG according to the methodology described in Section 3.5. After 72 h of culture, the biomass was filtered through Whatman filters and collected into 50 mL Falcon tubes. The biomass was divided into 2 mL samples, centrifuged at 15,000 rpm for 1 min, washed 3 times with sterile distilled water, and frozen in liquid nitrogen. RNA was extracted from three biological replicates using RNeasy Micro Kit (Qiagen; Santa Clarita, CA, USA) following the manufacturer's instructions. The obtained RNA was DNase digested (Thermo Fisher, Waltham, MI, USA) for reverse transcription reaction using Maxima First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, Inc., Waltham, MA, USA). The obtained cDNA was a template for PCR reaction using Phire Green Hot Start II PCR Master Mix (Thermo Fisher Scientific, Inc., Waltham, MA, USA) and primers listed in Table 5. Similarly, expression of Y. deformans YADE0S25e00298g gene was analyzed following the above-described methodology. Actin gene (ACT1, YALI0D08272g) was chosen as a housekeeping gene in this analysis.

Cultivation of Yarrowia Clade Species in the BSG Medium
The analyzed species belonging to the Yarrowia clade were cultured in a minimal medium supplemented with BSG as carbon and nitrogen sources to stimulate proteolytic enzyme secretion. Yeast cultures were performed in 250 mL Erlenmeyer flasks containing 50 mL of the medium composed of BSG (50 g/L), MgSO 4 (1 g/L), KH 2 PO 4 (0.1 g/L), K 2 HPO 4 (0.13 g/L), CaCl 2 (0.5 g/L), and FeSO 4 × H 2 O (0.01 g/L). After sterilization, the pH of the medium was adjusted to 7.0 by adding sterile 5 M NaOH (addition of approx. 320 µL per flask). Before inoculating the main culture, yeast was grown in YPD medium for 24 h under180 rpm agitation at 25 • C, centrifuged for 10 min at 7500 rpm, and washed with distilled water. The optical density was standardized to reach OD 600 = 0.25 in the BSG medium. Non-inoculated medium was cultured under the same conditions and used as a control. Each culture was cultivated for 168 h at 25 • C on a rotary agitation (180 rpm). One milliliter samples were taken every 48 h and analyzed for proteolytic activity, and the concentration of free α-amino groups was estimated. As pH is an important factor affecting the type of proteases produced by Yarrowia clade species, appropriate pH measurements were performed.

Determination of Hydrolysis Efficiency
In order to define involvement of proteases secreted by Yarrowia clade species on the availability and solubilization of BSG proteins, %DH was estimated. For this purpose, the concentration of free α-amino groups was determined using ninhydrin method [56]. The maximum concentration of α-amino groups detected for each culture was compared to the data obtained for the protein hydrolysis using hydrochloric acid (total hydrolysis), performed as described previously [32]. The %DH was calculated according to the following equation: %DH = (BH × 100)/TH (1) where TH (µg/mL) and BH (µg/mL) are the average concentration of α-amino groups obtained, respectively, as a result of the total hydrolysis (TH) or in the yeast culture (BH) calculated per 1 g of BSG sample.

Analysis of the Proteolytic Activity
In order to evaluate the proteolytic activity of the yeast species, the method of Anson was applied [57]. Shortly, proteolytic activity was estimated by mixing 0.45 mL of casein (1.1% in 50 mM Tris-HCl buffer, pH 7.5) with 0.5 mL of 50 mM Tris-HCl buffer (pH 7.5) and 0.05 mL of the analyzed sample. Reaction mixtures were incubated at 55 • C for 10 min, followed by adding 0.5 mL 8% (w/v) TCA to terminate the reaction. For each sample, appropriate negative controls were performed (no enzyme addition). As a positive control, protease from Bacillus licheniformis (Sigma-Aldrich, Germany) was applied. Finally, samples were centrifuged at 12,000 rpm for 15 min, and the absorbance of supernatants was determined at 280 nm. Tyrosine was used as a standard in the concentration range of 0-2 mM/L. One unit of enzyme activity was defined as the amount of enzyme required to release 1 µM of tyrosine per 1 mL of culture medium within 1 min under the experimental conditions.
The proteolytic activity of the Yarrowia clade species was also detected using skimmed milk agar plates. For this purpose, the standardized biomass solutions (OD 600 = 1.0) were spotted (5 µL) on the agar plates containing 2% (w/v) of agar and 10% of sterile skimmed milk (0.5%) (w/v). Inoculated plates were air dried under sterile conditions followed by 24 h incubation at 25 • C. Finally, the proteolysis zones were measured. As a control, non-inoculated medium was spotted on the agar plates and incubated under the same conditions. Both types of proteolytic activity measurements were performed in triplicate. Gelatin zymography was carried out according to the methodology described previously [32]. The samples, collected after 72 h of Yarrowia clade cultivation in BSG medium, were filtered through 0.45 µm membrane and concentrated using Amicon Centrifugal Filters in order to apply approximately 5µg of proteins on a gel well. Finally, samples were mixed 1:1 with a Sample buffer (62.5 mM Tris-HCl (pH 6.8), 0.01% bromophenol blue, 2% SDS, and 25% glycerol), cooled down, and subjected to SDS-PAGE with a 10% gel containing 0.1% gelatin. Thermo Scientific™ PageRuler™ Prestained Protein Ladder (Thermo Fisher Scientific, Cleveland, OH, USA), 10 to 180 kDa was used as a molecular weight standard. The non-inoculated BSG medium was used as a control.

Estimation of Proteolytic
In order to conduct proteomic analysis of secreted proteases, SDS-PAGE electrophoresis was performed. Samples were collected at 72 h, and proteins were precipitated using acetone. Each sample, 400 µL, was cooled down on ice and mixed with 1600 µL of acetone, previously cooled to −80 • C. Mixtures were shaken vigorously and incubated at −80 • C for 1 h. Samples were centrifuged at 13,000 rpm for 15 min. Supernatants were discarded, and final protein precipitates were evaporated using vacuum evaporator to remove residual acetone. Protein precipitates were dissolved in the 50 µL of 50 mM Tris-HCl buffer, pH 7.4, in order to obtain 8 times concentrated samples. For strains Y. divulgata, Y. hollandica, Y. bubula, Y. brassicae, and Y. yakushimensis, the described procedure was not enough to obtain positive result of electrophoresis. The process was thus repeated with 1 mL of the medium concentrated 20-fold.
In order to run the SDS-PAGE, 8-and 20-times concentrated samples were mixed with 4×Laemmli Sample Buffer and loaded on 10% gel. Gels were run at 150 V followed by staining with Coomassie Brilliant Blue R250 (40% methanol, 10% acetic acid, 0.1% Coomassie blue) for 1 h and destained overnight (40% methanol, 10% acetic acid). Thermo Scientific™ PageRuler™ Prestained Protein Ladder, 10 to 180 kDa, was applied as a molecular weight standard. Uninoculated BSG medium was used as a control.

Comparative Proteomic Analysis of the Yarrowia Clade Proteolytic Enzymes
Protein bands cut out from the SDS-PAGE gels were analyzed in the Laboratory of Mass Spectrometry, Institute of Biochemistry and Biophysics of Polish Academy of Sciences (Warsaw, Poland). Gel pieces were dried with ACN and reduced (treated) with 10 mM DTT in 100 mM NH 4 HCO 3 for 30 min at 57 • C. Cysteines were then alkylated with 0.5 M iodoacetamide in 100 mM NH 4 HCO 3 (45 min in dark at room temperature). Proteins were digested overnight at 37 • C with 10 ng/µL of trypsin in 25 mM NH 4 HCO 3 (Promega). Afterward, samples were concentrated and desalted on an RP-C18 pre-column (Waters), and further peptide separation was achieved on a nano-Ultra Performance Liquid Chromatography (UPLC) RP-C18 column (Waters, BEH130 C18 column, 75 µm i.d., 250 mm long). NanoACQUITY UPLC system, using a 45 min linear acetonitrile gradient, was applied. Column outlet was directly coupled to the Electrospray ionization (ESI) ion source of the Orbitrap Elite mass spectrometer (Thermo), working in the regime of data-dependent MS to MS/MS switch with HCD peptide fragmentation. The raw data files were pre-processed with Mascot Distiller software (version 2.7, MatrixScience). The obtained peptide masses and fragmentation spectra were compared to libraries using the National Center Biotechnology Information (NCBI) non-redundant database nr 20210206 (349124773 sequences; 126215745164 residues), with "other Fungi" filter (16225947 sequences) using the Mascot search engine (Mascot Server v. 2.7, MatrixScience). The following search parameters were applied: enzyme specificity was set to trypsin, fixed modifications: Carbamidomethyl (C), variable modification: Oxidation (M), and one missed cleavage allowed. Finally, offline mass recalibration of peptide and fragment masses in individual LC-MS/MS runs was performed to reduce mass errors, with resulting parent mass window of around 5 ppm and fragment mass window of around 0.01 Da. Decoy Mascot option was activated, but the FDR computations were unsuccessful due to vastness of database and low number of spectra originating from samples of limited complexity. Therefore Mascot-provided score threshold of 45 was used for peptide/protein identification.

Estimation of Antioxidant Activity of BSG Protein Hydrolysates
The antioxidant activity of the hydrolysates was investigated using two methods based on the depletion of the radical cation, ABTS and DPPH assay, and the FRAP method involving reduction of colorless, oxidized Fe 3+ form of iron into a blue-colored Fe 2+ form. All determinations were performed according to Ciurko et al. [58]. Despite a similar mechanism of action, there are significant differences between ABTS and 2,2-diphenyl-1picrylhydrazyl (DPPH) method, mostly related to the solubility of free radicals. In this research, ABTS radical was dissolved in water. Therefore, it was appropriate to estimate activity of water-soluble antioxidants, while DPPH radical, dissolved in ethanol solution, is more suitable to estimate activity of hydrophobic compounds. According to the work of Wołosiak and colleagues [43], deactivation of DPPH radical occurs predominantly via single electron transfer (SET) and uncommonly by hydrogen atom transfer (HAT), which has a dominant in the neutralization of ABTS radical. Differences in the described inactivation system cause recommendations for using both methods to determine the antioxidant activity of tested compounds. The antioxidant activity was estimated in samples after 120 h of the culture when DH% reached maximum value. Some modifications of DPPH assay were made: 10 µL samples were mixed with 190 µL of 0.1 mM DPPH radical solution in Eppendorf tubes (1.5 mL). Samples were incubated for 30 min in dark conditions, followed by centrifugation at 12,000 rpm for 6 min. After centrifugation, 100 µL of each sample was transferred to the microplate wells, and the absorbance at 517 nm was measured using Tecan instrument (Tecan Group Ltd., Männedorf, Switzerland). As a control, uninoculated medium incubated in the same conditions was used. Trolox (a synthetic analog of alphatocopherol) was used to estimate Trolox equivalent antioxidant capacity. Standard curves were prepared for 0-2000 µM, 0-1000 µM, and 0-400 µM for ABTS, DPPH, and FRAP methods, respectively. The obtained results were calculated per gram of protein hydrolysis products. For this purpose, peptide concentration in every sample was analyzed according to the Lowry method [59], with some modifications. A sample with a value of 10 µL was mixed with 50 µL AB solution and incubated for 10 min at room temperature. AB solution was prepared by mixing reagent A, composed of 0.1 M NaOH and 0.2 M Na 2 CO 3 , with reagent B containing 1% sodium citrate in a ratio of 50:1. After 10 min, 5 µL of Folin-Ciocalteau reagent was added to the mixture and incubated for additional 30 min at room temperature. Finally, the absorbance of each sample was measured at 750 nm using Tecan microplate reader. Peptide concentration was measured using peptide digest standard (Thermo Fisher Scientific, Cleveland, OH, USA) applied in a concentration range of 0-1000 µg/mL. The results were expressed as mean ± SD. Statistical significance was determined using Student's t-test. The significance level was set at p < 0.05.

Conclusions
The major conclusion of this work is that species of the Yarrowia clade showed an important diversity of alkaline proteases with different secretion abilities. In addition, the obtained data showed that the BSG could be used as a substrate for Yarrowia clade yeasts to obtain protein hydrolysates with antioxidant activity. Results presented in our work clearly confirm the potential of Yarrowia clade species to secrete high amounts of proteases into the growth medium.
The presented research is the basis for scientists who would like to develop knowledge about the proteolytic enzymes of the Yarrowia clade leading to their widespread commercial use. Based on the data presented here, further research can be performed on optimizing protease or bioreactor scale production. Another direction of further research may be the development of protease purification technology. In this study, proteolytic enzymes were used to produce bioactive peptides as a consequence of BSG protein breakdown, and hydrolysates with significant antioxidant activity were obtained. We believe that the presented research can contribute to the development of a system of biopeptide production, applied as a nutraceutical and functional food ingredient or as pharmaceutical and active cosmetic ingredients for the skin care industry.