Purification of bioactive peptides from spent yeast autolysates

.


Introduction
As one of the main brewing by-products (2-4 kg of spent yeast per 100 L of beer) (Fȃrcas ¸ et al., 2017) and designation of human consumption-generally recognized as safe (GRAS), spent brewer's yeast have been commercialized in the food market for years.Despite thousands of described yeast species, Saccharomyces cerevisiae (S. cerevisiae) still dominates this market (Ma et al., 2023;Oliveira et al., 2022a) due to their high protein (40-74% of dry weight) and essential amino acids (70.2-124 g/100 g protein) contents that allow the development not only of dietary supplements, but also of multiple functional foods, such as cooked ham, low-calorie mayonnaise, meat substitutes, yogurt, vegetable juices, toast spreads, soups and ready-to-eat meal (Zeko-Pivač et al., 2023).
In the last decades, the interest for other yeast ingredients has been growing in several economic sectors, due to their established biologic relevance properties, leading to the development of several extraction and purification processes for their production over the years (Marson et al., 2020a).Bioactive peptides are one example of these molecules demonstrating, among other, antimicrobial, antihypertensive and antioxidant properties, mainly in agri-food and cosmetic sectors (Oliveira et al., 2022a).However, aiming their production, yeast extraction processes often resort to physical, chemical or enzymatic extraction methodologies, where proteins are cleaved and freed from cell envelope, thus compromising the obtention of specific bioactive peptide fractions.For that reason, several isolation and purification steps have been added to the extraction process in order to decrease yeast cell debris and other contaminants that can influence the final biological properties.One must, however, bear in mind that the final process cost and selectively are inversely related, meaning that a value-added ingredient needs to be produced towards an overall cost-effective process (Oliveira et al., 2022a).
In addition to its application in food industry, S. cerevisiae strains have been engineered in large scale production of commercial target ingredients through their fermentation processes, such as farnesene, propanol and butanol, among others (Oliveira et al., 2022a).Similarly to what happens in brewing industry, these large-scale fermentation processes generate significant amounts of spent yeast which must be valued in sustainable and circular economy approaches.
Considering the bioactive potential of spent yeast peptides and their significant amount generated as by-products in industrial waste, the aim of the present study was the application of purification processes to a peptide-rich extract from engineered spent yeast autolysate previously produced as detailed in Oliveira, Odila Pereira et al (Oliveira et al., 2022a)., for the obtention of bioactive peptides-rich fractions with targeted biological properties for dietary supplementation.After literature review (Oliveira et al., 2022a), two purification methodologies were chosen: ultrafiltration and gel filtration, using size-exclusion chromatography (SEC).The purified peptide-rich fractions were then characterized and assessed for their potential bioactive properties, namely antimicrobial, anticholesterolemic, antihypertensive and antioxidant properties.

Peptide-rich extract production
The peptide-rich extract used in the present study was chosen from a previous work (Oliveira et al., 2022a) using protein content and molecular weight (MW) profile of extract as criteria, together with protein recovery and sustainability of the extraction processes evaluated.Briefly, a S. cerevisiae strain from Amyris, Inc. (Emeryville, California, USA) engineered to produce β-farnesene, identified as "ESY1″, was autolysed for 16 h, at 50 • C and 120 rpm in a New Brunswick™ Innova® 40/40 R Benchtop Orbital Shaker from (Eppendorf, Hamburg, Germany).After yeast treatment, the intrinsic enzymes were inactivated (95 • C, 5 min) and the cell suspension centrifuged (4696g, 10 min, 4 • C).The obtained supernatant, rich in protein and peptides, was collected and freeze-dried (Freeze-dryer Alpha 2-4 LSCbasic, Martin Christ, Osterode am Harz, Germany).

Size-exclusion chromatography (SEC)
Size exclusion chromatography, using a Hiprep™ 26/10 Desalting column (GE Healthcare) was performed in an ÄKTA Pure™ 25 M (Cytiva, USA) purification system, and monitored by ultra-violet absorption at 280 nm using a UV detector.Lyophilized peptide-rich extract was resuspended in ultrapure water to a final concentration of 9 g/L and filtered through a 0.45 µm filter (Chromafil® PET-45/25, Macherey-Nagel, Germany) before being loaded into a pre-equilibrated column (injection volumes of 2 mL).Ultrapure water, at a flow rate of 5 mL/ min, was used as elution buffer and fractions of 5 mL were collected.
BCA and dry weight.Protein concentration of SEC fractions was determined using the Pierce™ BCA protein assay kit (Thermo Fisher Scientific Inc., Massachusetts, USA), according to the manufacturer's instructions.Bicinchoninic acid (BCA) interacts with Cu 1+ resulting from the reduction of Cu 2+ ions from the copper sulphate by peptide bonds in alkaline pH, forming a purple-coloured complex with a strong linear absorbance at a wavelength of 562 nm with increasing protein concentrations (Smith et al., 1985).Protein content of the fractions was calculated after dry weight determination, performed at 105 • C for 24 h according to standard procedures of the Association of Official Analytical Chemists (Association, 2005).
Peptide sequences were obtained using the Proteome Discoverer 2.5.0.400 software (Thermo Scientific, Bremen, Germany), as described by Osório et al (Osório et al., 2021)., using S. cerevisiae protein sequence database instead.After identifying the peptide sequences, we conducted an in silico search of these sequences against databases that contain information on known peptide activities.The databases used in our analysis included the Database of Antimicrobial Activity and Structure of Peptides (DBAASP) (Pirtskhalava et al., 2021), the Data repository of antimicrobial peptides (DRAMP) (Shi et al., 2022), the Collection of Anti-Microbial Peptides (CAMP) (Gawde et al., 2023), the Antimicrobial Peptide Database (APD) (Wang et al., 2016) and the BIOPEP-UWM (MinkiewiczIwaniakDarewicz, 2019).Unique matches and partial matches were both considered.

Peptides bioactivity 2.4.1. Antimicrobial Activity
Time-growth inhibition curves of Escherichia coli (DSM1576) and Staphylococcus aureus (DSM799) (Deutsche Sammlung von Mikroorganismen und Zellkulturen, Braunschweig, Germany) were performed to assess the potential antimicrobial activity.According to the reported values of antimicrobial peptides in literature (Oliveira et al., 2022a), all samples were prepared at 0.5% (w/v) in Mueller Hinton (MH) growth medium (Biokar Diagnostic, Beauvais, France) and sterilized using 0.22 µm filters (Millipore, Billerica, MA, USA).Both bacterial strains were used as monocultures and grown in Tryptic Soy Agar (TSA; Biokar Diagnostic, Beauvais, France) before the assay, with 3 g/L of yeast extract (Sigma-Aldrich, Munich, Germany) at 37 • C for 24 h under aerobic conditions.One colony was then picked, transferred to a tube with 10 mL of MH and grown at the same abovementioned conditions.On the day of experiment, bacteria inoculums were adjusted to an optical density (OD) at 625 nm of 0.08-1 (corresponding to a cell density of 1 × 10 8 cells/mL), followed by a 10-times dilution in MH to obtain the working inoculums.For the bacterial growth inhibition assay, 980 µL of each sample were inoculated with 20 µL of the working bacterial inoculums.After mixing, 200 µL of the suspensions were pipetted to a 96-well microtiter plate (Nunc, Darmstadt, Germany), and the OD at 625 nm was assessed during 24 h at 37 • C, with 1 h intervals, using a microplate reader (Epoch, Vermont, USA).Inoculated MH medium was used as a negative control and inoculated MH medium with 0.5% (w/v) ampicillin, a concentration known to inhibit E. coli and S. aureus growth (Proma et al., 2020) was used as positive control.Blanks of the samples were performed to correct sample colour OD interference.

Angiotensin converting enzyme (ACE) inhibition assay
Amorim et al (Amorim et al., 2019a).methodology was exploited for ACE inhibition assay using o-Abs-Gly-p-nitro-Phe-OH trifluoroacetate salt (Bachem, Bubendorf, Switzerland) as substrate and 42 mU/mL of ACE (peptidyl-dipeptidase A from rabbit lung) (Sigma-Aldrich, St Louis, USA).ACE inhibition (%) was quantified using Eq.1: where F control and F sample are the fluorescence values obtained for the control (maximum ACE activity) and the sample, respectively.The calculation of IC50 values (concentration needed to inhibit 50% of ACE activity) for all samples was performed by non-linear fitting of the data, using a four-parameter logistic regression model.The assay was performed in triplicate.
Oxygen radical absorbance capacity (ORAC).ORAC was assessed according to Coscueta et al (Coscueta et al., 2019).Trolox standard curve (with a concentration range of 10-80 µM) allowed the quantification of TE antioxidant activity of the samples (μmol/g sample), after assessing the area under curve (AUC) for each sample.The assay was performed in triplicate.

HMG-CoA (3-hydroxy-3-methylglutaryl coenzyme A) reductase inhibition assay
HMG-CoA reductase inhibition was assessed by the colorimetric HMG-CoA Reductase Activity Assay Kit (Abcam, Cambridge, United Kingdom), according to the manufacturer's guidelines.The lower biocompatible Caco-2 concentration of peptide-rich extracts from spent yeast, previously assessed (Oliveira et al., 2022b), was used in this assay.This method is based on consumption of NADPH by the enzyme which is evaluated by the decrease of absorbance at 340 nm during 15 min.Quantification of the HMG-CoA reductase activity (% inhibition) is performed using absorbance values of two time points within the linear range, and with a minimum of 2 min apart, as recommended by the manufacturer.Atorvastatin was used as a reference drug inhibitor and Pravastatin as a commercial inhibitor (control).The assay was performed in triplicate.

Statistical analysis
Statistical analysis was performed using the Real Statistics Resource Pack software (Release 7.2) and results were expressed as average ± standard deviation.Outliers were excluded using the interquartile range with multiplier of 2.2 and data normality was checked using the Shapiro-Wilk test.Protein and bioactivities results were subjected to the analysis of variance (ANOVA) followed by Tukey's post hoc test.

Protein content (% w/w)
Purification of peptide-rich extract by SEC allowed the separation of four peptidic fraction as observed in the chromatogram (Fig. 2), that were collected and assessed for their protein content.Differently, only two purified fractions, "ESY1 > 1 kDa" and "ESY1 < 1 kDa", were obtained from ultrafiltration purification and quantified for their protein content by Dumas (Dumas, 1831).Protein content in a variety of samples can be rapidly determine by this method, an analytical method for the quantification of nitrogen, that is in accordance with international standards such as AOAC (Dumas, 1831;Association, 2005).However, due to quantity restrictions, protein content in SEC fractions was determined by the combination of BCA and dry weight determinations instead.
Protein content of the different fractions obtained by ultrafiltration A.S. Oliveira et al. and SEC, as well as protein content of the protein-rich extract obtained after autolysis of spent yeast (16 h, 50 ºC) (Oliveira et al., 2022b) can be seen in Table 1.Results show that protein content significant increased (p < 0.05) in ESY1 < 1 kDa (61.7 ± 0.6%) when compared to ESY1 > 1 kDa (22.5 ± 0.1%).The protein content here reported for ESY1 < 1 kDa is in accordance to the previous results of peptide rich-extract obtained from waste streams of mannan extraction from engineered spent yeast (S. cerevisiae) purified by ultrafiltration using a 1 kDa cut-off membrane (Oliveira et al., 2022b).Protein content also significantly increases (p < 0.05) in Section 4 (89.0 ± 0.8%) when compared to Section 1 (20.2 ± 3.4%), Section 2 (13.1 ± 4.3%) and Section 3 (43.6 ± 3.1%).These results are somewhat expected, since during both filtration processes, molecules are separated based on their sizes.Thus, in the ultrafiltration process, smaller molecules like minor proteins, peptides and amino acids, will pass through the 1 kDa cut-off membrane towards fraction, producing the cleaner ESY1 < 1 kDa fraction, while other larger molecules than proteins that are present in the extract will be retained in fraction ESY1 > 1 kDa, thus decreasing protein content in this fraction.Similarly, during gel filtration by SEC, these molecules with higher MW will be eluted first, while molecules with smaller MW will be trapped within the porous matrix, being eluted later, in fractions where the amount.Furthermore, the protein content determined for the Section 2 fraction, reveals that the higher number of other molecules are in that fraction.
When comparing both purification approaches, it is possible to conclude that SEC produced the fraction with the highest (p < 0.05) protein amount (Section 4; 89.0 ± 0.8%), supporting reports of being one of the widely techniques for isolation of antimicrobial, antihypertensive and antidementia peptides from S. cerevisiae (Oliveira et al., 2022a).However, costs associated to chromatography equipment are substantially higher than those associated with the ultrafiltration methodology, indicating that SEC needs to produce a value-added ingredient in order to justify their use.Indeed, this is the reason why chromatographic techniques are more frequently applied in purification of pharmaceutical proteins and peptides (Pratap et al., 2017).Nevertheless, some authors have already reported their use to produce yeast peptides with important biological properties, such as antihypertensive (Kim et al., 2004;Kohama et al., 1990;Kanauchi et al., 2005;Mirzaei et al., 2015;Ni et al., 2012a, 2012b), antimicrobial (Gddoa Al-sahlany et al., 2020;Branco et al., 2014Branco et al., , 2017;;Caldeira et al., 2019) and antioxidant (Mirzaei et al., 2015).In contrast, the use of ultrafiltration with filter membranes offers several advantages for concentrating proteins and peptides.One significant advantage is its low energy consumption.Additionally, it provides a gentle treatment of molecules, preserving their integrity and functionality.These reasons contribute to the widespread adoption of ultrafiltration in food processing industry (Zeko-Pivač et al., 2023;Vollet Marson et al., 2020;Mohammad et al., 2012) or to the production of peptides with important bioactivities for nutraceutical supplementation, including antihypertensive (Oliveira et al., 2022b;Amorim et al., 2019a;Kim et al., 2004;Mirzaei et al., 2015;Ni et al., 2012a;Hu et al., 2014), antimicrobial (Gddoa Al-sahlany et al., 2020;Branco et al., 2014;Caldeira et al., 2019;Comitini et al., 2005;Albergaria et al., 2010), antioxidant (Oliveira et al., 2022b;Amorim et al., 2019a;Mirzaei et al., 2015;Hu et al., 2014;Guo et al., 2020;Jung et al., 2011) and, more recently, anticholesterolemic (Oliveira et al., 2022b).Furthermore, in this work, the ultrafiltration fractions obtained can be classified as "rich in protein" according to the European Commission (European Commission, Nutrition claims, 2012).This designation highlights their potential as ingredients for nutraceutical market, as previously reported (Oliveira et al., 2022a).
All things considered, both techniques appear to be interesting processes to be included in protein-rich extracts production, but their choice requires a cost-effectiveness evaluation directly related to the economic sector to be applied.
(23.9%), and small percentages of 3-5 kDa (1.72%) and 5-10 kDa (0.15%).For this reason, a 1 kDa cut-off membrane filtration was chosen for peptides separation.However, these results are very different from other works reporting MW of 5-35 kDa in spent yeast autolysates.Indeed, peptides under 3 kDa were only obtained after including enzymatic hydrolysis in process (Amorim et al., 2019b;Marson et al., 2022;Xie et al., 2017).Protein and peptides MW distribution of the ultrafiltration and SEC fractions was assessed by HPLC-SEC.Unfortunately, the MW of ESY1, and respective ultrafiltration fractions, could not be determined by HPLC-SEC, since a good chromatographic resolution was not obtained with our column, probably related with the complexity of the matrix.For that reason, it was assumed that ESY1 ultrafiltration fractions own MW associated with cut-off membrane separation, that is, peptides above 1 kDa in ESY1 > 1 kDa and peptides under 1 kDa in ESY1 < 1 kDa.Results regarding the protein and peptides MW distribution in SEC fractions (Table 2) show that the most representative population in Section 1 had a MW around 3 kDa (75.4%), while the most representative in fraction Section 2 had approximately 1 kDa (45.9%).Peptides below 300 Da characterized fraction Section 3 (91.0%)and amino acids composed the majority of fraction 4 (92.6% under 200 Da) (Table 2).Comparing the MW distribution of all samples, Section 2 and Section 3 seemed to be similar with the original ESY1, being mostly constituted by ≤ 1 kDa peptides, and promising bioactive fractions.
After analysis of protein content and MW determinations, it possible to hypothesise that fractions ESY1 < 1 kDa and Section 4 are the ones with the most promising for bioactive properties, as it has been reported that peptides containing 3-20 amino acids are highly specific in their biological functions and in choosing biological and metabolic targets (Oliveira et al., 2022a).

Proteomics analysis
After proteomics analysis, several peptide sequences were identified in ESY1 and respective fractions obtained by ultrafiltration and SEC, from approximately 600 to 5000 Da (Supplementary material 2).As expected, the original extract "ESY1″ had the highest number of peptide sequences identified, along with fraction ESY1 > 1 kDa (6343).These were followed by Section 1 (1803) and Section 4 (599), while both less than 300 peptides were identified in Section 2 and Section 3 (Table 3).The main purpose of this analysis was to identify potential bioactivities related to the peptide sequences found in the samples by matching those sequences with bioactive peptides reported in different peptide bioactivity databases (PBD, see section 2.1.3for details).After conducting a thorough search in the data found in each of the PBD, sequences of our peptides present in all our fractions were found within the sequences of peptides listed in the PBD, with antibacterial, antimicrobial and antifungal activities reported (Table 3).However, it is important to note that only partial matches were identified during our analysis.Specifically, our peptide sequence was found as a component within a larger peptide sequence in the PBD, and an exact match was not found.
* PBD: BAASP, DRAMP, CAMP, APD and BIOPEP-UWM.Several authors already described antimicrobial activities of yeast peptides (Gddoa Al-sahlany et al., 2020;Branco et al., 2014Branco et al., , 2017;;Caldeira et al., 2019), and therefore these results are not surprising.However, considering the partial matches observed, it is plausible to argue that there is untapped potential to enhance the antimicrobial activity of these extracts.One potential approach is to reduce autolysis time, thereby minimizing peptide breakdown and achieving a closer match to those peptides described in the PBD.Nonetheless, the antimicrobial potential is still present and, for that reason, the effect upon the growth curves for two important food-related pathogen agents, exposed to our samples, was addressed and detailed in the next section.

Antimicrobial
To evaluate the potential antimicrobial activity of the different peptide fractions collected after ultrafiltration and SEC, a growth inhibition assay was carried out, and results of the fractions impact on E. coli and S. aureus growth can be seen in Fig. 3. Contrary to what was expected from the PBD results, none of the peptide fractions inhibited bacterial growth.As a matter of fact, all fractions (except Section 1 against S. aureus) seem to be promoting bacterial growth.
Antimicrobial peptides are usually amphipathic and cationic short peptides, whose activity strongly depends on their amino acid sequence, size and charge (Powers and Hancock, 2003;Jenssen et al., 2006).Although their mechanism is not clear yet, the best described mode of action for antimicrobial peptides regards the electrostatic interaction between the cationic peptides and the negatively charged components present on the pathogen membrane bilayers, allowing them to accumulate on the membranes and form pores by 'barrel-stave', 'carpet' or 'toroidal pore' mechanisms (Oliveira et al., 2022a;Mookherjee et al., 2020).For these reasons, the lack of antimicrobial activity, when compared with the peptide sequences in the PBD, might be due to the fact that the peptides identified in the different fractions have an inferior number of amino acids with a positive net charge, such as lysine (K), arginine (R) or tryptophan (W), know to play an important role in the antimicrobial activity of these peptides (Powers and Hancock, 2003;Chan et al., 2006).On the other hand, it is possible that the promotion of bacterial growth in these fractions is attributed to residual sugars.As previously reported, peptide-rich fractions derived from spent yeast contain certain amounts of sugar, potentially including glucose, which could contribute to bacterial proliferation (Oliveira et al., 2022b).Nevertheless, it would be interesting to investigate the potential antimicrobial properties of these peptide fractions against other pathogenic agents in the future.Since different microorganisms have distinct membrane components, their interactions with the peptides in each fraction could vary, leading to diverse outcomes.Exploring such variations could provide valuable insights into the broader antimicrobial potential of these peptide fractions.

Anticholesterolemic, antihypertensive and antioxidant activities
Even though the peptide fractions did not display any antimicrobial activity, other important bioactivities for nutraceutical market, such as antihypertensive, antioxidant and anticholesterolemic activities, were assessed.
3-Hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase is a catalyst in the conversion of HMG-CoA to mevalonate, an important intermediate in the cholesterol biosynthesis (DeBose-Boyd, 2008).HMG-CoA reductase inhibitors such as statins have been, thus, widely used in the treatment of hypercholesterolemia.In this work, the original extract and the different peptide fraction collected after both purification steps were assessed for their anticholesterolemic properties.As expected, after the addition of the ultrafiltration and SEC for purification of the extract autolysate, a significant higher HMG-CoA reductase inhibition (p < 0.05) was observed for all the fractions obtained, when compared with the original ESY1, thus increasing from 40.3 to 70.8-77.7%,depending on the fractions (Table 4, Supplementary material 3).In fact, these findings align with our previous study, which documents that spent yeast peptide fractions, both below and above 1 kDa, exhibited a HMG-CoA reductase inhibition range of 62.0-71.3%(Oliveira et al., 2022b).However, these results do not seem to be related with protein content or predominant MW of fractions, since no tendency was observed in this sense.ACE is a zinc metallopeptidase that catalyses the conversion of angiotensin I to angiotensin II, a strong vasoconstrictor with direct effect on the raising of blood pressure (Zhao, 2008).For this reason, molecules with ACE inhibition activity can help in the management of hypertension.The peptide-rich extract and the different fractions were evaluated for their antihypertensive activity using an ACE inhibition assay and by calculating their IC50 (Table 4, Supplementary material 3).Original ESY1 and fractions Section 4 and ESY1 < 1 kDa, showed the highest activity, corresponding to the lowest IC50 values, followed by Section 2, Section 3 and ESY1 > 1 kDa, and, finally, Section 1 (p < 0.05).While the higher ACE inhibitions of fractions ESY1 < 1 kDa and Section 4 are in accordance with their low MW distribution and protein content/purity, the high ACE inhibition value obtained for original ESY1 indicates the contribution not only of all proteins and peptides, but also of some other components of the extract.Section 1 was characterized with the highest MW distribution (~3 kDa) which can justify its lowest bioactivity.Indeed, small peptides (~ 2-20 amino acids) have been hypothesised have ACE inhibition activity by binding to the hydrophobic C-terminal tripeptide of the enzyme substrate, preventing  ACE-substrate binding, and, consequently, the conversion of angiotensin I to the angiotensin II (Kim et al., 2004).Overall results, mainly those obtained for ESY1, ESY1 < 1 kDa and Section 4, are in accordance with our previous work (Oliveira et al., 2022b) where IC50 values from 0.99 to 1.72 mg/mL were reported.Oxidative stress, caused by high intracellular levels of reactive oxygen species (ROS) has been frequently associated with aging and agingrelated diseases and antioxidant molecules can delay the oxidation by scavenging ROS through different reactions, thus reducing ROS to less reactive forms (Shields et al., 2021).Aromatic amino acids present in the peptide sequences have been presumed to be capable to quench the free radicals through a direct electron transfer mechanism, while the pyrrolidine ring of proline residues may interact with secondary structure of peptides, thus increasing their flexibility and ability to quench singlet oxygen (Amorim et al., 2019a).The potential antioxidant activity of the original extract and the peptide fractions were determined using two scavenging-based assays: ABTS •+ and ORAC.While the ABTS •þ assay measures the relative ability of antioxidants to reduction of cation radical ABTS •+ into ABTS • as compared with a Trolox, a water-soluble analogous of vitamin E and a known antioxidant, the ORAC assay is based on generation of the radical peroxyl (ROO•), a radical with biological importance (Prior et al., 2003).
In the ABTS •+ assay, fractions Section 4 and ESY1 < 1 kDa showed the highest activity, followed by ESY1, ESY1 > 1 kDa and Section 3. Section 1 and Section 2 demonstrated to have the lowest activity (p < 0.05) (Table 4, Supplementary material 3).Overall, the antioxidant activities are lower than those previously obtained for other peptide-rich extracts (Oliveira et al., 2022b) which might be related with differences in protein content of fractions.On the other hand, the SEC fractions demonstrated the highest ORAC vales, especially Section 3 and Section 4, followed by original ESY1 and ultrafiltration fractions (p < 0.05) (Table 4, Supplementary material 3).Similar antioxidant activity results to the ones obtained for the original ESY1, and respective ultrafiltration fractions, were reported by Costa et al (Costa et al., 2023).for spent yeast waste streams fractions (654-1041 μmol TE/g extract).However, SEC fractions showed a substantial increase in the ORAC values to ~ 2000 μmol TE/g extract for Section 1 and Section 2, and ~ 40000 μmol TE/g extract for SEC fractions with low MW peptides and high protein content/purity (Section 3 and SEC4) thus demonstrating their application for antioxidant value-added products on nutraceutical market, or even in the cosmetic sector, as previously suggested by Costa et al (Costa et al., 2023).
When comparing the results obtains by the two antioxidant assays, it is possible no see some differences between them.When working with peptide-rich extract, or even peptide fraction, data interpretation can be tricky due to their complex nature and composition.Moreover, different assays have been used to evaluate antioxidant properties, and the results obtained in those assays are frequently different among them, as all methodologies present advantages and limitations (Schaich et al., 2015).Nevertheless, it is clear that both ultrafiltration and SEC originate some fraction with higher biological activities than the original peptide-rich extract.

Conclusion
The present work reports the production of peptide fractions derived from a spent yeast extract autolysate.The fractions were obtained using 1 kDa ultrafiltration or gel filtration, using SEC.These processes allowed to generation of fractions with distinct characteristics, mainly in terms of protein content, MW distribution and peptide sequences.While the original extract autolysate and respective ultrafiltration and SEC fractions did not show any antimicrobial effect, interesting results regarding their antihypertensive, anticholesterolemic and antioxidant properties were observed.These results suggest that these fractions hold promise as potential candidates for dietary supplementation in the form of a single product.All purified peptide fractions showed higher HMG-CoA reductase inhibition values than the original extract.These results demonstrate these processes as crucial methodologies for the development of new anticholesterolemic ingredients using yeast autolysates.The peptide fractions with the highest protein content/purity (ESY1 <1 kDa and Section 4) revealed potential for the formulation of antihypertensive ingredients, as these were the ones with the higher ACE inhibition values.Regarding the antioxidant activity, significantly superior ORAC values were observed in fractions Section 4 and Section 3.This suggests their advantageous use in other economic sectors such as cosmetics, where antioxidant molecules have a large anti-aging impact.
In addition, results presented in this study suggests that the addition of a purification step in the production of bioactive peptides from spent yeast autolysates can be advantageous for generating ingredients with specific targeted bioactivities for the nutraceutical market.These purification processes enable the isolation and concentration of bioactive peptides, enhancing their potential application in the development of functional food and dietary supplements.These findings emphasize the importance of implementing purification strategies in order to optimize the production of bioactive peptide ingredients for the nutraceutical industry.Results are expressed as average ± standard deviation (n = 3).*For Caco-2 biocompatible concentrations (Oliveira et al., 2022b).

Fig. 2 .
Fig. 2. Elution profile of the SEC purification of peptide-rich extract "ESY1″.Four different peptide fractions, corresponding to the four peaks observed, were collected (Sections 1 to 4).

Fig. 3 .
Fig. 3. Growth curves for E. coli (A) and S. aureus (B) exposed to original ESY1 extract and fractions obtained from ultrafiltration (ESY1 >1 and ESY1 <1; on the left) and Sections (1 to 4; on the right).

Table 2
Protein and peptides MW distribution in SEC fractions represented as percentage of chromatogram total area (% total area).

Table 3
Peptide sequences identified in original ESY1 extract and fractions obtained from ultrafiltration (ESY1 >1 and ESY1 <1) and Sections (1 to 4), together with their bioactivities reported in PBD.