Optimization of Total Carotenoid Production by Rhodotorula mucilaginosa from Artichoke Agroindustrial Waste Using Response Surface Methodology

The carotenoids have vast medical, industrial, dietary


Introduction
Carotenoids are natural liposoluble pigments of highly conjugated tetraterpenoid structure and light reflexive properties, showing a diversity of colors such as red, orange, yellow, or purple (Zhao et al., 2019). The total carotenoids are found as carotenes (α, β, γ-carotene and lycopene) and xanthophylls (lutein, zeaxanthin and others) (Maoka, 2020); similarly, they exhibit different bioactive properties: antioxidants, provitamin A precursors, immunomodulators and photo protectors (Amengual, 2019). Carotenoids are mainly extracted from plant sources which require strict and unsustainable chemical processes due to the use of toxic petrochemical solvents (Kultys and Kurek, 2022). Therefore, obtaining carotenoids from microbial sources such as filamentous fungi, bacteria, microalgae and yeast provides an affordable and sustainable alternative, since it allows greater productivity in a short time, absence of seasonal restrictions, and it is non-toxic, facilitated a controlled fermentative process with generation of biodegradable by-products (Ram et al., 2020). The yeasts stand out as producers due to their fast growth and carotenoid production, including genera as Rhodotorula, Sporobolomyces and Xanthophylomyces (Menegazzi et al., 2020).
The genus Rhodotorula genus is the most studied because of its aerobic metabolism and simple nutrition based on carbon and nitrogen, providing its substrate availability, low cost, and elevated cell growth (Liu et al., 2021). To improve the fermentation process, new environment-friendly paths are being acquired with the use of cost/effective substrates including vegetable waste (Kaur et al., 2019). For example, Dias- Moreira et al. (2021) reached an efficiency of 361.3 and 296.6 μg/g of total carotenoids from coffee pulp and skin, respectively. In addition, Sharma and Ghoshal (2020) obtained a maximum content of carotenoids (717.82 mg/g) from an acid extract of onion skin and mung bean at 25° and pH 6 at 110 rpm. In the same way, Cipolatti et al. (2019) achieved to use sugarcane molasses and corn-macerated liquor to obtain an efficiency of 224.8 μg/L volumetric carotenoids using Rhodotorula mucilaginosa.
The agricultural sector is considered an important livelihood activity for the world economy, through the supply of fresh food for consumption and raw materials to the food industry. However, it is estimated that 1300 million metric tons of food are wasted annually, the vegetables, tubers and fruits being the ones with the greatest quantity (Yafetto, 2022). In this way, agro-industrial waste has become an alternative raw material for industrial use due to its wide availability (Ravindran et al., 2018). The agroindustrial waste can be highlighted for the fermentative production of carotenoids, as Otero et al. (2019) got 820 μg/L of carotenoids by Rhodotorula mucilaginosa from parboiled rice water. Likewise, Torres-Alvarez et al. (2019) got 317 μg/g of carotenoids by Rhodotorula mucilaginosa from banana peel extract, and Sinha et al. (2019) achieved to obtain 20 mg/L by Rhodotorula toruloides at 120 h from paddy straw hydrolysate.
Peru is considered the first exporter of fresh vegetables in South America and the eighteenth globally, with 8.34% of artichoke's production generating an export of 10 000 annual tons (Chavez, 2021). In artichoke processing, the resulting waste is 80-85% of the total vegetal biomass, such as bracts which are the external parts of flowers (Monge et al., 2019). The waste of artichoke bracts has a profitable structural composition, with up to 43% of cellulosic content and 10.4 g/100 g reducing sugars, as well as fructose polysaccharides at 15.9 g/100 g of dry weight (Zeaiter et al., 2019). However, little has been used for the production of carotenoids by fermentation.
In this way, the use of artichoke bracts waste represents a sustainable alternative for the collection of carotenoids by biotechnological paths. The main purpose of this investigation was focused on assessing the use of artichoke bracts coming from agroindustrial waste for the production of total carotenoids from a culture of Rhodotorula mucilaginosa. PLUS). Then, it was pulverized and kept in polyethylene bags at room temperature until its use (20 ± 2°C).

Characterization of the agroindustrial waste
The chemical features were determined through infrared spectroscopy from a sample of 0.5 g BABW and ABW, using a Thermo Nicolet IS50 (USA) device through attenuated total reflectance (ATR) at 4000-500 cm -1 (Trilokesh and Uppuluri, 2019). The colors of the reflecting surface of the residue were measured using a spectrophotometer with a color measuring system (Konica Minolta CM-5, Tokio, Japón). The L*, a* and b* values were registered from the waste sample (Aridi et al., 2021). (1)

Fermentative production of carotenoids
The effect of pH (5-8) and stirring speed (80-120 rpm) were assessed in the carotenoids production by a Central Composite Rotational Design (CRD) ( Table 1). The experiment was done in 150 mL flasks with 30 mL of the medium from the hydrolyzed BABW supplied with 0.5% peptone, 0.1% yeast extract, and 0.5% sodium chloride. The initial cell content of the yeast inoculation was 2.6 ± 0.1 at OD 600 , previously cultured. The pilot testing was incubated at 30ºC for 72 hours in agitation according to the experimental design, using a shaking

Isolation and characterization of Rhodotorula mucilaginosa
Different agricultural soil samples were collected (1 kg) from the Center of Plan Genetics at Universidad Nacional de Trujillo, Perú (8°06'46.2"S, 79°02'18.7"W). A suspension of 10 g of soil in 100 mL of sterile saline solution (0.85% NaCl) was prepared, homogenized and left to rest at room temperature for 15 minutes. Serial dilutions were gotten from the supernatant, and aliquots of 100 μL were inoculated over the surface of a Petri dish with agar Sabouraud (glucose 4%, yeast extract 0.1%, peptone 0.5% and sodium chloride 0.5%; pH 6), then it was incubated at 30ºC for 48 h. The signature Rhodotorula-colored colony was selected and its purity was determined by testing its macroscopical and microscopical morphological features (El-Ziney et al., 2018).
The isolated yeast culture was identified through molecular analysis with the use of a Polymerase Chain Reaction (PCR) of the ITS1 e ITS2 (subunit 5.8S rRNA), according to the methodology of Baldera et al. (2022). The DNA was extracted from a highly pure 48 h culture (Nucleic acids extraction and purification kit, innuPREP DNA Kit, Analytik Jena). The amplicon was obtained through the ITS-1 and ITS-2 universal primers, immediately after the amplified rDNA was characterized in a 1.5% agarose gel. The purified PCR product was sequenced (Macrogen, USA) and aligned with sequences through BLAST (National Center for Biotechnology Information, NCBI). A phylogenetic tree was elaborated with the rDNA sequences using MEGA-X software.

Collecting and setting up of agroindustrial waste
The artichoke bracts waste (ABW) was collected from agro-export businesses in La Libertad region, Peru. The ABW were washed and dried at 60°C for 18 h on a stove (MMM, ECOCELL 111). After that, they were ground and sifted at a pore size of 600 μm (Sadh et al., 2018).
The 2% (a/w) of sifted ABW was then bleached (sodium chloride 2%) for 1 h at room temperature (Widiarto et al., 2017). Subsequently, the bleached ABW was vacuum filtered and then washed with distilled water in order to remove chlorine waste. The bleached artichoke bracts waste (BABW) was spread out on a plastic tray and dried at 60°C for 18 h in a stove (MEMMERT, UF260 incubator (LBSI-100A, LABNICS). The yeast biomass was monitored through optical density using a UV-VIS spectrophotometer (UVLINE, 9400).

Total carotenoids extraction and estimate
The carotenoids were extracted with dimethyl sulfoxide (DMSO, Merck) according to the method suggested by Kanzy et al. (2015). The obtained biomass was centrifuged at 6000 rpm for 5 min, the supernatant was removed and washed two times with distilled water. The achieved biomass was resuspended in 20 mL of distilled water and poured into a Petri dish to dry at 45°C for 24 hours on a stove (MMM, ECOCELL 111). 25 mg of dry biomass was weighted and resuspended in 1 mL of DMSO, the mixture was agitated in a vortex for 30 seconds and it was put into Termoblock (VRS, 12621-092) at 55°C for 1 hour, with agitation in a vortex every 15 minutes. Finally, the sample was centrifuged for 5 minutes at 6000 rpm and the pigmented supernatant was recovered for its optical reading at 501 nm through a UV-VIS spectrophotometer (UVILINE, 9400).
The total content of carotenoids was calculated and expressed as Total Fraction of Carotenoids, TFC, (μg/g of dry yeast) and volumetric carotenoids concentration, VCC (μg/L) (Dyaa et al., 2022), according to the next formula: VCC (μg/L) =

Characterization of extracted carotenoids
The maximum absorbance of the extracted carotenoids solution was analyzed using a spectrophotometer (UVILINE, 9400) in the 300-800 nm range. The dry biomass obtained was analyzed through Raman confocal spectroscopy with the use of an Ar laser of 514.5 nm and it was assessed within 800-1800 cm -1 spectra with a 2 cm -1 (Jehlička et al., 2014).

Statistical analysis
The obtained data were subjected to an analysis of variance (ANOVA). They were calculated as media ± standard deviation, and the resulting surfaces were analyzed using Design Expert 13, with statistical significance P < 0.05 (Allahkarami et al., 2021).

Results and Discussion
Isolation and identification of Rhodotorula mucilaginosa strain A total of 10 different morphologies were obtained from microbial isolations, but just a culture showed yeast-like growth with coral pink pigment (designed as TQ21). The isolated TQ21 showed convex colonies of smooth edges with brilliant and coral pink color of mucoid aspect. The cells from TQ21 isolate were verified through microscopy, where blastoconidia and mono-budding ovoid cells (Fig. 1). Additionally, the liquid medium turned yellow to red after 48 hours of growth; these features coincide with Hu et al.'s (2022) reports where they isolated and described different morphological similarities of the Rhodotorula genus. The PCR amplification of regions ITS-1 and ITS-2 of gen 5.8S rRNA produced an amplicon of 159 bp; these sequences were compared with the GenBank database (BLAST). The sequence alignment revealed that the isolated strain showed a 99% similarity with Rhodotorula mucilaginosa PHYTYM9 (MT559975.1); likewise, the phylogenetic analysis based on the sequence of the region ITS 1-2 of gen 5.8S rRNA placed the TQ21 strain near the Rhodotorula mucilaginosa group with a branch length of 0.09 indicating a minor species divergence at the nucleotide level (Fig. 2). The isolated yeast was identified as Rhodotorula mucilaginosa TQ21 and the sequence noted was posted in the NCBI database with access number OM721682.1 (https://www.ncbi. nlm.nih.gov/nuccore/2195005066).

Characterization and conditioning of the waste
The FTIR waste spectrum without treatment showed a peak of over 3290 cm −1 attributed to the O-H stretching vibration contained in its cellulose, hemicellulose, and lignin (Yuli et al., 2020). In the same way, the 2920 cm -1 peak was assigned to the C-H cellulose stretching (Zheng et al., 2019). The 1730 cm −1 band was assigned to the C=O lignin vibration (Falah et al., 2020). The 1610 cm −1 peak corresponded to the O-H vibrations due to the water traces (Mat, 2014). Also, the peaks in 1410-1420 cm −1 were associated with the group CH 2 , related to cellulose crystallinity; while, 1370 cm −1 to the C-H deformation in cellulose and hemicellulose (Akcay and Yalcin, 2021). At the same time, band 1320 cm −1 was related to the vibration of the C-H syringyl and guaiacyl lignin ring (Zheng et al., 2015). According to Bolio-López et al. (2016), the peak within 1240-1250 cm −1 was assigned to aromatic ester groups (syringyl lignin). The peak 1030 cm −1 was attributed to the stretching vibration of CO and OH groups; and 898 cm −1 to β-glucoside links (De-Rosa et al., 2010). The presence of lignin in the artichoke waste linked to its cellulose was corroborated, which represents a tipping point for hydrolysis (Fig. 3) (Ma et al., 2014).
Sodium hypochlorite (NaClO) was used for the selective removal of lignin since this last mentioned behaves as a recalcitrant and fermentation inhibitor (Morone et al., 2017). After treating the residue with NaClO, a variation in the transmittance percentage was observed in the featuring peaks of lignin at 1720-1730, 1320, and 3290-3330 cm −1 (Fig. 3), probably because of the decrease in the lignin content (Aridi et al., 2021). This lignin decrease was contrasted through the whiteness  (Table 2), where the ABW obtained 46.2% showing a significant difference with BABW (72.1%). The color of the vegetal waste reveals natural impurities, pigments, and lignin that stay between the cellulose fibers, this way, an elevated whiteness index would point to an increase of accessible surface to cellulose (Alghooneh et al., 2017). The BABW was submitted to hydrolysis to obtain reducing sugars (RS) for fermentation (Fig. 4). A maximum concentration of 3.1 g/L RS was obtained applying 8% (w/w) of BABW with 2.5% of H 2 SO 4 ; at the same time, according to Tukey test, it significantly differs from the rest of used treatments (P < 0.05). Similarly, Nwogwugwu et al. (2019) reported maximum values of up to 3.1 g/L of RS using waste of calabash pulp (Crescentia cujete). In addition, Christy et al. (2021) obtained within 0.54-36 g/L RS by applying acid hydrolysis palm waste (Borassus flabellifer). Moreover, it is known that a higher concentration of acid and longer treatment time may contribute to the production of furfural compounds from hydrolyzed sugars and may affect fermentation (Deshavath et al., 2017).

Production of total carotenoids
The impact of pH and agitation speed was evaluated on the culture medium of hydrolyzed BABW to identify the maximum conditions in the production of carotenoids.
In Table 3  In Fig. 5, it is observed that the VCC value slightly increases whenever pH decreases (until pH 5), and at the same time, the agitation speed increases. Indeed, Arias (2014) pointed out that at pH 5 higher biomass and carotenoids were obtained, which suggests that Rhodotorula mucilaginosa prefers slightly acid conditions. Besides, during the microbial growth, organic acids are released, and these intermediaries are assimilated again stimulating the carotenogenesis. In the same way, the agitation speed increase reflects good stimuli in VCC as much as in CBC, since it influences the nutrient distribution and the oxygen distribution in the culture medium, stimulating enzymes such as phytoene desaturase, β-carotene hydrolase, and lycopene cyclase (Borba et al., 2018). In FTC, the variation of agitation speed was not significant (P > 0.05); nonetheless, the pH differs positively between 5-6 values showing better pigmentation at treatment 3 with a CBC of 2.4 g/L, in spite the maximum CBC of 3.3 (Mata- Gómez et al., 2014).

Characterization of obtained carotenoids
The preliminary identification of the extracted pigment was achieved through UV-VIS spectrum by a scan within 200-800 nm, and a maximum of 500 nm was reported within regions 440-550 (Fig. 6a). These results are corroborated by Sharma and Ghoshal (2020b) who confirmed carotenoids presence within 400-500

. Colored contour charts showing the interaction between pH and agitation speed for maximum production of VCC (a), FTC (b), and CBC (c)
nm, showing specific peaks of 450, 485 and 490 for beta-carotene, torulene and torularhodin, respectively. In that same way Mussagy et al. (2019) described that DMSO, the solvent used, conducts higher release rates of intracellular carotenoids, where torularhodin, the most polar carotenoid exhibits a maximum absorbance at 500 nm.
With focal RAMAN microscopy (Fig. 6b), the peak 1003 cm −1 was identified as attributable to a combination of δ (C = CH) methyl in the flat balancing and δ (C−H) flexion modalities outside the chart (Marshall and Marshall, 2010); in addition, the peak 1154 cm −1 assigned to v2 stretching (C-C) symmetrical and peak 1509 cm -1 attributable to symmetric stretching v1 (C = C) in phase (Mot et al., 2017). The results are similar to those reported by Pacia et al. (2016), who describe peaks of 1511, 1154 and 1002 cm −1 as features for carotenoids (beta-carotene). Our preliminary results demonstrated for the first time that the Rhodotorula mucilaginosa TQ21 isolate presented potential in the production of total carotenoids from a culture medium based on hydrolyzed artichoke waste, in addition to the advantages that can be obtained through the design of response surface methodology to maximize VCC and FTC through optimization of pH values and stirring speed.

Conclusions
A strain of Rhodotorula mucilaginosa TQ21 was identified and cultivated, showing the ability to grow and produce carotenoids from artichoke hydrolyzed waste as the only carbon source. Response surface methodology was used by a central compound rotational design and optimal values of 2968.95 μg/L VCC (12258.53 μg/g of TFC) were obtained at pH 5 and 120 rpm at 30°C for 72 hours. In addition, TQ21 pigments showed typical carotenoid characteristics by their UV-VIS spectrum at 500 nm, and peaks of 1003, 1154 and 1509 cm −1 were observed by RAMAN analysis. The results suggest that Rhodotorula mucilaginosa TQ21 constitutes a promising microorganism for the commercial production of carotenoids from artichoke waste. This process means an economical way to reduce the requirement of high-cost nutrients and at the same time mitigate environmental pollution, which contributes to subsequent scaling and wide use in the global market of carotenoids for food, cosmetics and animal additives.