**Microbial Colorants Production in Stirred-Tank Bioreactor and Their Incorporation in an Alternative Food Packaging Biomaterial**

**Fernanda de Oliveira <sup>1</sup> , Caio de Azevedo Lima <sup>1</sup> , André Moreni Lopes <sup>2</sup> , Daniela de Araújo Viana Marques <sup>3</sup> , Janice Izabel Druzian <sup>4</sup> , Adalberto Pessoa Júnior <sup>5</sup> and Valéria Carvalho Santos-Ebinuma 1,\***


Received: 28 August 2020; Accepted: 15 October 2020; Published: 2 November 2020

**Abstract:** Natural colorants from microbial fermentation have gained significant attention in the market to replace the synthetic ones. *Talaromyces* spp. produce yellow-orange-red colorants, appearing as a potential microorganism to be used for this purpose. In this work, the production of natural colorants by *T. amestolkiae* in a stirred-tank bioreactor is studied, followed by its application as additives in bio-based films. The effect of the pH-shift control strategy from 4.5 to 8.0 after 96 h of cultivation is evaluated at 500 rpm, resulting in an improvement of natural colorant production, with this increase being more significant for the orange and red ones, both close to 4-fold. Next, the fermented broth containing the colorants is applied to the preparation of cassava starch-based films in order to incorporate functional activity in biodegradable films for food packaging. The presence of fermented broth did not affect the water activity and total solids of biodegradable films as compared with the standard one. In the end, the films are used to pack butter samples (for 45 days) showing excellent results regarding antioxidant activity. It is demonstrated that the presence of natural colorants is obtained by a biotechnology process, which can provide protection against oxidative action, as well as be a functional food additive in food packing biomaterials.

**Keywords:** natural colorants; filamentous fungi; stirred-tank bioreactor; biodegradable films; antioxidant; food package

#### **1. Introduction**

Over the years, humans have used colorants to enhance or restore the appearance of products, especially in the food industry [1,2]. Nevertheless, most of the synthetic colorants used in food products are derived from petroleum and may have harmful effects on human health and the environment [3,4]. Therefore, there is a growing preference for natural additives in foods. Regarding the natural color market, a survey by the Grand View Research, Inc. estimated the world color market for 2025 to be

USD 37.49 billion. Specifically, for the food market, the natural colorant consumption will register a 5.9% compound annual growth rate (CAGR) in terms of revenue, in relation to the period from 2018 to 2025 [5].

There are several biological sources that produce natural colorants, such as insects, vegetables, and microorganisms [2,6]. Colorants from microbial origin, especially fungal ones, are of great interest mainly because of the recent improvements in biotechnology and bioprocessing technologies as well as a full control of the production conditions [2,7,8]. Fungus of the genus *Monascus* stands out for presenting good levels of colorant production. However, *Monascus* species co-produce citrinin, a mycotoxin [9]. Thus, other strains such as *Aspergillus*, *Trichoderma*, *Fusarium*, and *Talaromyces* (formerly *Penicillium* spp.) have been the subject of studies in order to produce colorants [8,10–12]. In this context, *Talaromyces amestolkiae* (former *Penicillium purpurogenum*) has aroused interest in the production potential of stable and non-toxic yellow, orange, and red colorants [10,13–15].

One of the main challenges of using filamentous fungi, such as *Talaromyces,* for metabolite production is the control of microorganism growth and form besides the mass transfer inside the bioreactor [16]. Hyphal concentration greatly influences the fermentation broth viscosity [15]. Meanwhile, the oxygen transfer in submerged cultures is an important factor for efficient fungal growth and an important parameter for a high yield of fungi colorants [17]. At a high cell concentration, the culture medium viscosity increases and it limits the oxygen supply for the cells [4,18]. Increasing the stirring speed can maintain dissolved oxygen concentration at high levels, however, potential damage to the fungal hyphae can limit the impeller speed and, consequently, the oxygen and nutrient transfer capability of a bioreactor [19,20]. In order to control cell growth and colorant biosynthesis, the well-directed process parameter shift represents a valuable control strategy [9,21]. In the production of *Monascus* colorants, the pH-shift strategy has been proven to be efficient in colorants biosynthesis [17,22]. Hence, great efforts to improve the bioreactor cultivation conditions are necessary for enhancing the large-scale production of *Talaromyces* colorants.

Biosynthetically, many fungi colorants are polyketides. Polyketide colorants range in structure from tetraketides to octaketides, which have four or eight C<sup>2</sup> units that contribute to the polyketide chain. The structure of the polyketides does not exhibit localized negatively-charged ions. These molecules often have a polyunsaturated function, i.e., a ring system, one or more carbonyl groups, carbolic acid, and functional ester or amide groups that absorb in the UV-visible spectrum [23]. One class of polyketide colorants are azaphilones, which are hexaketide colorants with pyrone-quinone structures and a chiral quaternary center [24]. Investigations focus on six major azaphilone colorants: Rubropunctamine and Monascorubramine (red); Rubropunctatin and Monascorubrin (orange); Monascin and Ankaflavin (yellow) [25]. One of the most interesting characteristics of azaphilone colorants is their potential use as a functional food colorant, because they may exert antioxidant and antimicrobial activities, relatively low cytotoxicity, as well as immunosuppressive, antiviral, anticancer, and cholesterol-reducing properties [26]. Moreover, our research group previously reported that *T. amestolkiae*-fermented broth presented low cytotoxicity against fibroblast cells and effective antimicrobial activity against *Staphylococcus aureus*, a representative food contaminant [13]. In this sense, bio-based materials with antioxidant properties are gaining attention in the food sector [27], mainly because non-ecofriendly packages are considered an environmental problem [28].

The application of natural compounds in active packaging has been increasingly applied in order to produce materials that interact with packaged foods, positively modifying their sensory and nutritional properties, preventing the deterioration of food, mainly caused by lipid oxidation and microbial growth [29]. Particularly, foods with high-fat content and especially those with a high degree of unsaturation are susceptible to deterioration by oxidation. In this way, alternative packaging technologies based on the inclusion of antioxidants compounds can create a barrier, improving the stability of packaged products sensitive to oxidation [30]. Therefore, the use of the *T. amestolkiae* natural colorants as an added component in bio-based films represents an alternative for increasing the properties of food packaging.

This work addresses the production of natural colorants by *T. amestolkiae*in a stirred-tank bioreactor. The effects of stirring speed and pH-shift as a control strategy on colorant accumulation are evaluated. After the selection of the best operational condition, the fermented broth containing the colorant is incorporated into cassava starch biodegradable films. The bio-based films prepared are characterized with respect to thickness, water activity, and total solid content. In the end, the cassava starch-based films are used to pack butter, supporting the applicability of natural colorants as additives in bio-based films. The effectiveness of the *T. amestolkiae* colorants protection against oxidative action is addressed on the basis of the peroxide content under accelerated oxidative conditions.

#### **2. Materials and Methods**

#### *2.1. Materials*

Sucrose and yeast extract were purchased from Synth (São Paulo, Brazil) and Acumedia (Lansing, MI, USA), respectively. Cassava starch (composed by 23.5% amylose and 64.2% amylopectin) was donated by Cargill Agrícola S.A. (Porto Ferreira, SP, Brazil). Commercial butter was obtained from Imperial (Bahia, Brazil). Low-density polyethylene (LDPE) film (0.020 mm thickness and 15.86 × 10−<sup>8</sup> g H2O·mm/m<sup>2</sup> ·h·kPa water vapor permeability) was purchased from local markets (Salvador, BA, Brazil). All of the other reagents were of analytical grade.

#### *2.2. Microorganism Maintenance and Colorant Production*

*Talaromyces amestolkiae* DPUA 1275 was generously provided by the Culture Collection of the Federal University of Amazonas (DPUA, Manaus, AM, Brazil). The cultures preserved in distilled water were reactivated in Czapeck Yeast Extract Agar (CYA) and maintained at 30 ◦C for 7 days. The CYA medium had the following composition (g/L in deionized water): K2HPO<sup>4</sup> (1.0), yeast extract (5.0), sucrose (30.0), agar (15.0), and 10 mL/L of concentrated Czapeck [(g/100 mL of deionized water): NaNO<sup>3</sup> (30.0), KCl (5.0), MgSO4·7H2O (5.0), and FeSO4·7H2O (0.1)] [10].

The production process was composed of three phases: pre-inoculum, inoculum, and submerged culture. For the pre-inoculum preparation, a loop of fungus from stock culture was inoculated on a CYA plate and maintained in the same reactivation conditions. For inoculum, five mycelial agar discs (8 mm diameter) of *T. amestolkiae* were punched out from the pre-inoculum with a self-designed cutter and transferred to 50 mL of submerged culture medium in 250-mL Erlenmeyer flasks incubated at 30 ◦C and 150 rpm for 72 h. The inoculum medium was CYA liquid (without the addition of agar). Then, the entire volume obtained (0.2 L) was aseptically transferred to a single flask and then transferred to the bioreactor. The composition of the submerged culture medium was similar to that used for inoculum preparation, except for the concentration of sucrose and yeast extract, which were 48.50 and 11.80 g/L, respectively [10]. Both the inoculum medium and the culture broth had their pH adjusted to 4.5 with HCl (5 M) and were autoclaved at 121 ◦C for 15 min. The submerged culture was performed in a stirred-tank bioreactor Bioflo® 115 (New Brunswick, Edison, NJ, USA) with 3 L of working volume equipped with two Rushton impellers submersed into the bulk liquid. After sterilization, the bioreactor containing 1.8 L of culture medium was supplied with 0.2 L of the inoculum. The following operational conditions were kept constant by the bioreactor controllers: temperature at 30 ◦C and aeration rate at 2.0 vvm. The experiments were carried out for 240 h, and every 24 h, an aliquot was withdrawn. At the end of the bioprocess, the fermented broth was filtered first using filter paper Whatman #1 (Whatman, Marlborough, England) and later using a 0.45 µm filter acquired from Millipore. The filtrated samples were used to determine the production of yellow, orange, and red colorants as well as sucrose consumption. In fact, during the cultivation of *Talaromyces* spp., the three natural colorants (yellow, orange, and red) were produced at the same time. In order to guarantee that the colorants were produced, we analyzed the specific wavelength for each one of them, according to Section 2.3.

In the first set of experiments, the effects of the stirring speed from 100 to 600 rpm were evaluated. Next, pH-shift experiments were carried out at 500 rpm, adjusting the pH to 8.0 with NaOH (1 M) after the sucrose depletion (96 h of cultivation).

#### *2.3. Analytical Methods*

The sucrose concentration was determined according to the methodology described by Dubois et al. [31] and the pH was measured using a pH-meter (model MPC 227, Mettler Toledo, Columbus, OH, USA). The production of extracellular colorants was estimated by spectrophotometric analysis (model DU 640, Beckman, Irving, TX, USA) by reading the absorbance of supernatant at 410, 470, and 490 nm, which corresponds to the maximum absorbance for yellow, orange, and red colorants, respectively. The results were expressed in terms of absorbance units (AU) [10].

#### *2.4. Kinetic Parameters*

To calculate the substrate conversion factor (sucrose) in product (*YP*/*S*) and productivity (*P*), we employed Equations (1) and (2), respectively:

$$\chi\_{P/S} = \frac{Abs\_f - Abs\_0}{S\_0 - S\_f} \tag{1}$$

$$P = \frac{Abs\_f}{t} \tag{2}$$

where: *S*<sup>0</sup> and *S<sup>f</sup>* are the initial and final sucrose concentrations; *Abs<sup>f</sup>* and *Abs*<sup>0</sup> are the final and initial colorant absorbance, respectively, and *t* is the time (h).

#### *2.5. Preparation of Cassava Starch-Based Films and Incorporation of Natural Colorants*

The cassava starch-based films were prepared according to the methodology described by Silva et al. [32]. In this way, the bio-based films were prepared by the casting technique, mixing cassava starch (4.0 wt%), plasticizers (0.7 wt% sucrose and 1.4 wt% inverted sugar), and the fermented broth containing the colorants (5.0 wt%). The fermented broth was used because of the synergy between the colorants molecules and the different metabolites present on fermented broth that increase the antioxidant power. Dispersions were heated (70 ± 2 ◦C), degassed for 30 min in ultrasonic bath to remove the bubbles, placed in polystyrene Petri dishes (150 × 15 mm), and dehydrated in an oven with airflow and circulation at 35 ± 2 ◦C for 24 h. A cassava starch-based film without addition of the fermented broth was prepared and used as control. Resulting bio-based films were stored in desiccators (at 23 ± 2 ◦C with 60 ± 2% relativity humidity) with a supersaturated solution of magnesium nitrate for 48 h before the characterization studies.

#### *2.6. Cassava Starch-Based Films Characterization*

The thickness was evaluated using a flat parallel surface micrometer (Mitutoyo model 103–137, precision 0.002 mm). Six measurements were taken at random positions around the film sample. The results were expressed as the mean ± respective standard deviations and used for the calculation of the contact area (mm).

The water activity (aw) was measured with a Decagon, Aqualab Lite, as calibration standards used pure water (aw of 1.000 ± 0.001%) and LiCl (aw of 0.500 ± 0.015%). Preconditioned samples (4 cm<sup>2</sup> ) were cut from the center of the films and evaluated in triplicate at room temperature (25 ± 2 ◦C).

The total solid content was determined by measuring the weight loss of films upon drying (105 ◦C) until constant weight. This determination was performed in triplicate. The results were expressed as the mean ± respective standard deviations.

#### *2.7. Application of Cassava Starch-Based Films for Packaging Butter*

Butter was packed in the cassava starch-based films with colorant additives. Square-shaped films (5 × 2 cm) of 0.164 and 0.212 mm in thickness were molded (Sealer Sulpack SM 400 TE, Caxias do Sul, Rio Grande do Sul, Brazil). Butter samples were homogenized and frozen in small pieces (with 3 × 2 cm and 10.00 ± 0.54 g), after they were involved with the film, the bubbles of oxygen were removed, and the film was sealed.

The samples of packaged butter were stored for different periods of time (0, 7, 15, 30, and 45 days) and under accelerated oxidative conditions (64% of relative humidity at 25 ± 2 ◦C). These analyses were carried out in a dark room to avoid the effects of light interference (i.e., color degradation) in the samples. For this, three types of packages were prepared: cassava starch-based films with colorant additives (CFC), cassava starch-based films without colorant additives (CF), and conventional plastic (CP). In addition, unpackaged butter (U) was used as control. To evaluate the antioxidant action of the films formulation, the oxidative stability of the butter stored under accelerated oxidative conditions was monitored for 45 days and the peroxide content (PC) was determined by titration method according to the methodology described by the Association of Official Analytical Chemists–AOAC (2000) [33].

#### *2.8. Statistical Analysis*

The data were analyzed by ANOVA using a StatSoft v.7 program (StatSoft, Inc., Tulsa, OK, USA). The Tukey test was used to evaluate the mean difference in results at the 95% level of significance.

#### **3. Results and Discussion**

#### *3.1. Production of Natural Colorants in Bioreactor: E*ff*ect of Stirring Speed*

The high viscosity of the medium can result in oxygen diffusion limitation, due to the combination of the high biomass concentration and the fungal morphology. Hence, efforts have been made in order to overcome mass transfer problems [16,34]. Therefore, the effects of the stirring speed and pH were investigated by our group. In this sense, the production of natural colorants by *T. amestolkiae* in a stirred-tank bioreactor was evaluated, we maintained the aeration rate constant at 2.0 vvm and varied the stirring speed from 100 to 600 rpm in order to achieve proper oxygen diffusion in our cultivations. The results are depicted in Figure 1.

As can be seen in Figure 1, the stirring speed influenced the colorants' yield. At 100 rpm, the colorants' production stopped after 120 h, probably because under this condition the microorganism distribution inside the bioreactor was not homogeneous. It is usually observed that submerged cultures of filamentous fungi present a pseudo plastic behavior (i.e., a non-Newtonian characteristic) [35], and low speed, associated with high medium viscosity, does not allow for system homogeneity. The main role of stirring in a bioreactor is to improve heat and mass transfer to achieve homogeneity in the system. A proper mixing of components in the culture medium is necessary to ensure an adequate flow in submerged aerobic culture [36]. The stirring of the culture medium may also cause different effects in filamentous microorganisms. Among them may be included: cell wall disruption, changes in filamentous morphology, variation in growth efficiency and growth rate, and the variation in the formation rate of the desired bioproduct [18,36]. The stirring process may damage both mycelial pellets and hyphae structures. The fragmentation of hyphae can result in the creation of small parts of these structures, promoting a new center for biomass growth, or may cause damage to the hypha allowing release of the cytoplasm. On the other hand, total pellet breakdown may occur, and, probably in this case, total autolysis or aggregates can occur [37].

**Figure 1.** Production of yellow (yellow bars), orange (orange bars), and red colorants (red bars) by submerged culture of *T. amestolkiae* in stirred-tank bioreactor at 2.0 vvm and 30 ◦C by varying the stirring speed (100 to 600 rpm) as a function of time (0 to 240 h). The error bars represent 95% confidence levels for the mean of three independent assays.

Increasing the stirring speed could maintain dissolved oxygen concentration at high levels being a key parameter for high yield of *T. amestolkiae* colorants in the stirred-tank bioreactor cultivation. As the stirring speed increased, the homogeneity of the culture medium could be achieved, and it was possible to determine the natural colorant production. Since the oxygen transfer is directly proportional to the shear force and it is strictly related to the morphology of the filamentous fungus, stirring speeds above 600 rpm were not evaluated in this study. Considering the production of orange and red colorants, there was no statistical difference between the results achieved at 500 rpm and 600 rpm (statistical test: one-way ANOVA, *p* = 0.0994). However, as 500 rpm promote a lower shear stress, this stirring speed was chosen for the next experiments. In this condition, the absorbance of the yellow, orange, and red colorants was 1.71 AU400nm, 0.71 AU470nm, and 0.69 AU490nm, respectively. Comparing the colorant production achieved with results previously reported by our research group [10], using the same culture media but in orbital shaker incubator, there was a decrease of 5-fold for the red colorant production in the bioreactor. This result confirms the statement that the scale-up from orbital shaker incubator to large fermenters is a difficult task. In fact, several conditions of cultivation can change in the bioreactor, mainly the oxygen transfer rate in the medium.

Figure 2 shows the results of sucrose consumption and pH over time of submerged culture of *T. amestolkiae* in bioreactor varying the stirring speed. All values are detailed in Table S1 in the supplementary material.

**Figure 2.** Consumption of sucrose (**A**) and pH (**B**) over time (0 to 240 h) of submerged culture of *T. amestolkiae* in stirred-tank bioreactor varying the stirring speed (100 to 600 rpm) at 2.0 vvm and 30 ◦C. The error bars represent 95% confidence levels for the measurements.

Regardless of the stirring speed evaluated, the sucrose consumption showed a similar profile (Figure 2A). In general, the concentration of sucrose decreased during the cultivation and a pronounced consumption of sucrose at 96 h occurred. At the end of cultivations, the final sucrose concentration was close to 2 g/L in all conditions studied.

With respect to pH variation, regardless of the stirring speed evaluated, there was an increase in its value throughout the cultivation, with the pH value of 4.5 reaching values close to 7.5 after 240 h of cultivation (Figure 2B). This profile was similar to what occurred in the submerged culture of *T. amestolkiae* (former *Penicillium purpurogenum*) in the orbital shaker incubator studied by Santos-Ebinuma et al. [10]. Therefore, pH 4.5 is suitable to provide good cell growth.

#### *3.2. Production of Natural Colorants by Submerged Culture in Bioreactor: pH-Shift Strategy*

The stirring speed of 500 rpm was selected to evaluate the pH-shift control strategy on *T. amestolkiae* colorant accumulation. At 500 rpm, the initial pH of 4.5 changed to approximately 8.0 at the end of cultivation (Figure 2B). According to Orozco and Kilikian (2008) [38], the production of red colorants by the submerged culture of *Monascus purpureus* in a bioreactor can be favored by a pH change. These authors selected a pH of 5.5 for the growth step and pH 8.0 for the production step. Thus,

a similar strategy was used in this work in order to increase the colorant production in the bioreactor cultivation. In this way, at 500 rpm of stirring speed, a cultivation was carried out and, after 96 h, the pH was changed to 8.0 by the addition of 5 M NaOH. The time of 96 h was selected according to the experiments carried out that varied the stirring speed, because it was observed that, at this point, there was a marked decrease in the concentration of the primary carbon source (Figure 2B). Figure 3 depicts the production of yellow, orange, and red colorants, as well as the pH for the assays performed at 2.0 vvm, 30 ◦C, and 500 rpm.

**Figure 3.** Production of yellow (yellow bars), orange (orange bars), and red colorants (red bars), pH (circle) and sucrose (square) as a function of time (0 to 240 h) during the submerged culture of *T. amestolkiae* in a stirred-tank bioreactor at 2.0 vvm, 30 ◦C, and 500 rpm. The black arrow indicates the pH-shift from 4.5 to 8.0 at 96 h for the production step. The error bars represent 95% confidence levels for the measurements.

After the pH-shift, it is possible to observe a significant increase in the production of all the colorants, with this increase being more significant for the orange and red ones, both close to 4-fold. Furthermore, the influence of the pH-shift on colorant production was evident. The maximum yield of yellow, orange, and red colorants was 3.20 AU400nm, 2.56 AU470nm, and 2.45 AU490nm, respectively. These results corroborate those found by Orozco and Kilikian (2008) [38], who mention that different pH levels during the growth and production steps may lead to an improvement in the colorant production. Although the colorant production with the pH-shift strategy was greater than for the first set of experiments, the product yield was still lower than for the best condition in the orbital shaker incubator, which shows that further studies are necessary to improve the production of *T. amestolkiae* natural colorants in a stirred-tank bioreactor.

Compared with no pH change condition, all kinetic parameters evaluated improved with the pH-shift control strategy (Table 1). The rate of the substrate conversion to product increased from 0.037 to 0.067 AU·L/g for the yellow colorants and from 0.015 to 0.054 and 0.051 AU·L/g for the orange and red colorants, respectively. Similarly, the productivity of all colorants increased in this second experimental condition.

Generally, in bioreactor cultivation, reproducible kinetic growth of filamentous organisms is difficult to obtain. This phenomenon occurs frequently, due to mycelial aggregation in fermentation broths with long hairy mycelial morphologies that can adhere to surfaces and form a growing biofilm [39]. Many fungal strains grow preferentially on surfaces and will develop thick layers on walls and surfaces in a submerged fermentation process [40]. In the same way, reproducible samples to measure apparent viscosity are difficult to obtain, hence, it was also deemed appropriate to omit the time course for biomass and viscosity.

**Table 1.** Kinetic parameters calculated for substrate conversion factor (sucrose) in product (YP/S) and productivity during the submerged cultivation of *T. amestolkiae* in a stirred-tank bioreactor for the production of yellow (Y, AU400nm), orange (O, AU470nm), and red colorants (R, AU490nm) without a pH change and after a pH-shift from 4.5 to 8.0 at 96 h of cultivation.


AbsO/AbsY: relationship between the production of orange and yellow colorants; AbsO/AbsR: relationship between the production of orange and red colorants.

Even though it is desirable to produce fungal colorants comprised of only one single color component [22], it is known that the fermented broth is composed of yellow, orange, and red colorants [13–15]. However, the careful selection of pH and nitrogen sources is a valid approach to produce extracellular extracts comprised of one predominant color component [14]. It has been widely accepted by the scientific community that orange colorants are the first biosynthetic product and the other colorants derive from the orange ones [41]. In this way, the extracellular colorant composition and concentration is dependent of the orange colorants produced that are transformed into red and yellow colorants under some specific fermentation conditions. For the *Monascus* genus, a low pH (pH 2.5 and 4.0) provides intracellular extracts composed mainly of orange colorants, independently of the nitrogen source employed. At acidic conditions, the secretion of orange colorants into the broth and the reaction with any amino unit for red colorants formation is limited [22]. Meanwhile, a change of pH to levels closer to neutral modify the extent of transformation of orange colorants into red ones [22]. Additionally, for *T. amestolkiae*, the synergistic effect of a low pH and nitrogen source is mainly important in the production and excretion of red colorants. A previous study of our research group presented the production of a glutamic acid-red colorant complex by *T. amestolkiae* in a chemically defined medium with monosodium glutamate (MSG) as the nitrogen source [14]. In the presence of MSG, deep yellow colorants were derived from neutral and basic pH, while deep red colors were derived from acidic pH. The glutamic acid–colorant complex seems to be more water-soluble than those produced in nitrogen complex media (meat extract and meat peptone) by the same strain [13].

In this work, yeast extract was used as complex nitrogen source and the synergistic effect of this nitrogen source with pH favored the increase of the relation between orange colorants and yellow/red ones. This is a reflection of the overall increased rate of substrate to product and productivity. It seems that a close to neutral pH favored the reduction of orange colorants to yellow ones (relation AbsO/Abs<sup>Y</sup> < 1.0 duplicated with pH-shifting), while the specific red conversion was limited (relation AbsO/Abs<sup>R</sup> > 1.0). Consequently, the extracellular extracts contained mainly yellow colorants. Therefore, extracellular red colorants derivatives were not mainly produced at a high pH. It can be considered that natural colorants can be obtained by the submerged culture of *T. amestolkiae* in a bioreactor, however, the culture conditions can be improved and the production increased by varying mainly the pH during cultivation. In order to study a possible application of the colorants produced, the incorporation of the fermented broth into an alternative food-packaging material based on its oxidative protection effect was evaluated.

#### *3.3. Preparation and Characterization of Cassava Starch-Based Films*

Biodegradable films were produced by a formulation of cassava starch, plasticizers, and the fermented broth containing the *T. amestolkiae* colorants produced in stirred-tank bioreactor. The incorporation of natural additives from plants for active packaging has been increasingly applied, such as: coffee-cocoa [42,43], carotenoid and yerba mate extract [44], green tea and palm oil extract [45], among others. However, natural additives produced by biotechnology is an innovative approach. The bio-based films produced were characterized in terms of thickness, water activity, and total solids content. The results are depicted in Table 2.

**Table 2.** Characterization of biodegradable films in the presence or absence of natural colorants in terms of thickness, water activity (aw), and total solids.


\* Statistical test: one-way ANOVA (with Welch correction for unequal variances). The results were expressed as the mean ± respective standard deviations. Equal letters in the same line represent equal values to a significance level of 5%.

The physical properties of the films were determined to evaluate the influence of the colorant incorporation, since they could modify the structure of the polymer matrix, e.g., by weakening the inter-chain bonds. The incorporation of natural colorants in the film formulation caused a significant reduction of film thickness, around 10%, in relation to films with no additives. Other works also reported a reduction in thickness of films formulated with cassava starch and natural additives [43,46]. The control of the thickness in the biodegradable films is an important step, since variations in this parameter can affect some films properties, such as mechanical and barrier properties, which can compromise the package performance [45]. Thickness was also the only parameter addressed that was affected significantly.

The variations in the total solid content in the films after colorant incorporation were not significant, with values of 88.96 and 89.80%, respectively. In general, the incorporation of natural additives in cassava starch-based films does not affect significantly the solid content [43].

Additionally, no significant difference (<0.05) for water activity (aw) between the films produced was observed. These results demonstrate that the presence of natural colorants did not change the water activity of the films. Water activity brings information about the amount of free (also referred as unbound or active) water present in a sample [47] being a parameter used in food preservation. Low water activity reduces the availability of water to microorganisms, avoiding undesirable chemical changes for the storage of products [32]. As water migrates from areas of high aw to areas of low aw, it is important to have low values of aw (<0.600 according to Mathlouthi, 2001 [48]) for food product design. As the aw was around 0.6 for both films, they can be considered as food product packages. In this way, the cassava starch-based films were used to pack butter samples.

#### *3.4. Oxidative Stability of the Packaged Butter during Storage*

The oxidative stability of the butter packaged in different formulations, namely: cassava starch-based films with colorant additives (CFC), cassava starch-based films without colorant additives (CF), and conventional plastic (CP) was evaluated. Unpacked butter (U) was used as the control. To this purpose, the packed or unpackaged butter was stored for 45 days under accelerated oxidative conditions (64% of relative humidity and 25 ◦C) and monitored through peroxide content (PC) at 7, 15, 30, and 45 days of storage. Figure S1 from Supplementary Material shows an image of the butter packaged in each formulation evaluated. Figure 4 shows the results of PC for each condition over the 45 days of storage. All values are detailed in Table S2 in Supplementary Material.

**Figure 4.** Peroxide content (PC, meq/kg) in the butter as a function of the packaging film used, such as: cassava starch-based films with colorant additives (CFC), cassava starch-based films without colorant additives (CF), conventional plastic (CP), and unpackaged butter (U), during different periods of time (0, 7, 15, 30, and 45 days) in storage. The error bars represent 95% confidence levels for the measurements.

From Figure 4 can be seen that, as expected, the PC of the U was higher than of the packaged one. The CF and CP showed higher peroxide indices than CFC. This demonstrates that natural colorants act as both a photooxidative protector and antioxidant agent, which allow a higher oxidative protection of the packaged product when compared to the other conditions evaluated.

After 45 days of storage, the PC increased 329.0% for the exposed butter, 239.7% for the CP, and 231.1% for the CF. While for the CFC, the PC increased only 118.5%. Therefore, even without antioxidant additives (CF), cassava starch-based films have a greater oxidative protection effect when compared to conventional package films (CP). However, the formulation with colorants incorporated in cassava starch-based films (CFC) enhanced the oxidative protection of butter, with a reduction in peroxide value. This shows a synergistic action of the bio-based film and the colorants, improving the shelf life of fatty products. In this way, *T. amestolkiae* colorants can be considered as potential natural antioxidants for the stabilization of lipid-containing foods for active packing formulation. However, additional studies on the concentration of colorants in the formulation should be performed in order to avoid a possible pro-oxidant effect. Since a high content of antioxidant additives can act as a pro-oxidant agent [44,45].

#### **4. Conclusions**

This work evaluated the influence of the stirring speed and pH on the production of natural colorants by *T. amestolkiae* in a stirred-tank bioreactor. The greatest colorant production occurred at 500 rpm, under the pH-shift strategy from 4.5 to 8.0 during the production phase. Although the regulatory mechanisms of the biosynthesis of *T. amestolkiae* colorants are not yet clear, the relationship between the culture conditions and the colorant formation it is of great importance. Despite the difficulties in terms of viscosity of the medium, during the submerged cultivation of filamentous fungi in a bioreactor, there is the possibility of using strategies to control growth and metabolite production, especially in terms of stirring and pH. Moreover, this work demonstrated the potential of natural colorants to be included in biodegradable film formulation. The incorporation of colorants in cassava starch-based films provides oxidative protection in packaged butter, by the decrease of peroxide index. Hence, *T. amestolkiae* colorants have great potential for use as functional food colorants. As future perspectives, our group intends to evaluate the surface morphologies of the films with or

without colorants by scanning electron microscope in order to gather supplementary data about the characteristics and stability of our formulations extending, for example, its shelf-life application.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2309-608X/6/4/0264/s1, Figure S1: Butter packaged in cassava starch-based films with colorant additives (CFC), cassava starch-based films without colorant additives (CF), conventional plastic (CP), and unpackaged butter (U), Table S1: Consumption of sucrose and pH over time (0 to 240 h) of submerged culture of *T. amestolkiae* in stirred-tank bioreactor varying the stirring speed (100 to 600 rpm) at 2.0 vvm and 30 ◦C, Table S2: Peroxide content in the butter as a function of the packaging film used, such as: cassava starch-based films with colorant additives (CFC), cassava starch-based films without colorant additives (CF), conventional plastic (CP), and unpackaged butter (U), during different periods of time (0, 7, 15, 30, and 45 days) in storage.

**Author Contributions:** Conceptualization, F.d.O., C.d.A.L., V.C.S.-E.; methodology, F.d.O., J.I.D., A.P.J., V.C.S.-E.; software, F.d.O., V.C.S.-E.; validation, D.d.A.V.M., A.P.J., V.C.S.-E.; formal analysis, V.C.S.-E.; investigation, F.d.O., V.C.S.-E.; resources, A.P.J., V.C.S.-E.; data curation, A.P.J., V.C.S.-E.; writing—original draft preparation, F.d.O., C.d.A.L., A.M.L., V.C.S.-E.; writing—review and editing, F.d.O., A.M.L., D.d.A.V.M., V.C.S.-E.; visualization, A.P.J., V.C.S.-E.; supervision, J.I.D., A.P.J., V.C.S.-E.; project administration, V.C.S.-E.; funding acquisition, A.P.J., V.C.S.-E. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was developed with the financial support from FAPESP (São Paulo Research Foundation, Brazil) through the projects 2018/06908-8, 2014/01580-3, 2013/05039-2, 2008/58280-0. A.M. Lopes acknowledges the financial support from FAPESP through the projects 2017/10789-1 and 2018/10799-0. This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) - Finance Code 001. The authors also acknowledge the support from the CNPq (National Council for Scientific and Technological Development, Brazil).

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Oil-Based Fungal Pigment from** *Scytalidium cuboideum* **as a Textile Dye**

**Mardonio E. Palomino Agurto <sup>1</sup> , Sarath M. Vega Gutierrez <sup>1</sup> , R. C. Van Court <sup>1</sup> , Hsiou-Lien Chen <sup>2</sup> and Seri C. Robinson 1,\***


Received: 31 March 2020; Accepted: 20 April 2020; Published: 22 April 2020

**Abstract:** Identification of effective natural dyes with the potential for low environmental impact has been a recent focus of the textile industry. Pigments derived from spalting fungi have previously shown promise as textile dyes; however, their use has required numerous organic solvents with human health implications. This research explored the possibility of using linseed oil as a carrier for the pigment from *Scytalidium cuboideum* as a textile dye. Colored linseed oil effectively dyed a range of fabrics, with natural fibers showing better coloration. Scanning electron microscopy (SEM) revealed a pigment film over the fabric surface. While mechanical testing showed no strength loss in treated fabric, colorfastness tests showed significant changes in color in response to laundering and bleach exposure with variable effects across fabric varieties. SEM investigation confirmed differences in pigmented oil layer loss and showed variation in pigment crystal formation between fabric varieties. Heating of the pigmented oil layer was found to result in a bright, shiny fabric surface, which may have potential for naturally weatherproof garments.

**Keywords:** fungal pigment; natural dye; spalting; *Scytalidium cuboideum*; dramada; sustainable clothing

#### **1. Introduction**

Natural dyes and pigments have a long history of use for coloring textiles, from ancient Egypt to the oldest South American cultures [1–5]. Artificial dyes currently dominate the market due to their ease of mass production, low price, and color variety [5]. However, most modern textile colorants are produced using hazardous chemicals and many contribute to water pollution through the production of effluents [6,7]. In recent decades, the low environmental impact and sustainability of natural dyes have become more desired by consumers, driving a shift in the market back to natural sources of coloration [8,9].

There are presently a number of natural colorant alternatives available, such as those derived from barks, fungal pigments, insects, and minerals [3]. Most have never been commercialized due to concerns about their sustainability, cost effectiveness, and profitability [10]. In addition, natural dyes often have problems with colorfastness, and, while methods have been developed to improve dye uptake [11–13], they are energy intensive and increase the production price [14–16].

One source of natural coloration that has been suggested as a replacement for synthetic dyes in fabric dyeing are pigments extracted from wood-rotting fungi. Research by Weber et al. [17] has shown that fungal pigments, when carried in dichloromethane (DCM), with and without mordants, showed promise for dyeing fabrics, using a dripping process. The pigments from four fungi were investigated: *Chlorociboria aeruginosa* (Oeder), Seaver, and *C. aeruginascens* (Nyl.) Kanouse, which produces a blue-green pigment known as xylindein, *Scytalidium cuboideum* (Sacc. & Ellis) Sigler & Kang, which produces a red pigment known as draconin red (Figure 1) that exists as both orange and red crystals, and *Scytalidium ganodermophthorum* Kang, Sigler, Y.W. Lee & S.H. Yun, which produces an unnamed yellow pigment. Further studies showed that polyester absorbed more pigment than other synthetic fabrics and that these colors were stable over time [17,18]. Hinsch [19] later confirmed that *S. cuboideum* and *C. aeruginosa* could be used to dye textiles with no mordant and attributed this to the carbon–carbon interaction of the main quinone structures.

**Figure 1.** The structure of dramada, the orange and red crystal pigment from *Scytalidium cuboideum.*

Despite the success of early testing, the fungal pigments mentioned above never gained commercial traction. This is likely due to the need for the pigments to be carried in DCM, which is a potential human carcinogen and a known greenhouse gas [20–22]. Other solvents were tested for their ability to extract the pigments, and while tetrahydrofuran (THF), acetonitrile (ACN), and acetone were all moderately capable of extraction, all three interacted with the pigments and caused color change [23]. Due to these issues, an entirely new type of solvent was explored: natural oils. These oils proved to be highly successful at carrying the fungal pigments, though not at extracting them, with raw linseed oil allowing for the highest pigment stability of those tested [24]. However, of the tested pigments, only the red pigment from *S. cuboideum* was found to effectively color fabric, although it showed variation in coloration based on the material to which it was applied [25].

The purpose of study was to compare the dyeing capability, heat stability (in different washing temperatures), and pH stability (washing in different detergents), of three soft-rot spalting fungi, carried in raw linseed oil, on various fabrics. While it is already known that the pigments from these fungi are reasonable textile dyes when carried in DCM [17,18], this carrier is not practical for commercial scale use due to toxicity issues. An alternative carrier proposed is raw linseed oil, which is known to carry the pigments at high concentrations but has never been tested in textile applications. Using the oil carrier would result in a dyeing process with fewer environmental issues than the DCM process. Analysis via scanning electron microscopy (SEM) will further the understanding of the interaction between the fungal pigments and the fabrics. The research will assess if the oil changes the structure of the pigments, how the pigments bind to textiles, and if the reaction of the pigments changes due to temperature and pH changes based upon the oil carrier.

This work prepares the fungal pigments to fully enter into the commercial dye industry by assessing the viability of a more eco-friendly carrier and removing the largest hurdle currently facing the soft-rot fungal pigments, their DCM requirement for application.

#### **2. Materials and Methods**

All testing methodology followed either accepted the American Society of Testing Material (ASTM) standards (such as for fabric testing) or were based on testing used in previous work with these spalting pigments on textiles (see citations below).

#### *2.1. Fungal Growth and Pigment Extraction*

Cultures of *S. cuboideum* (UAMH 4802, isolated from oak lumber, location unknown) were used to inoculate malt agar plates amended with white rotted wood chips, following the procedure outlined in Robinson et al. [23]. Plates were allowed to grow for 2–3 weeks before they were dried for 24 h in a fume hood, extracted into DCM, and standardized to CIE *L*\**a*\**b*\* values of *L* = 82.32, *a* = 26.84, and *b* = 13.19, as described in Robinson et al. [23]. After the standardization, raw linseed oil (Sunnyside) was mixed with the pigment carried in DCM and placed on a stir plate for 48 h in order to evaporate the solvent and leave the pigment suspended in the oil, following the methodology by Palomino Agurto et al. [25]. This pigmented oil was used for further testing.

#### *2.2. Fabrics Tested*

Unfinished 100% fabrics of polyester, nylon, cotton, and wool were used (see Table S1). Per each type of fabric, two different densities were used: low-density and high-density. To establish a difference between low-density and high-density, a minimum difference of 40 yarns per inch in the count number was established. Low-density nylon was discarded because its finishing process did not allow dying by the pigmented oil.

#### *2.3. Mechanical Testing*

Mechanical testing was done to evaluate the impact of the oil-pigments on tensile and tear strength of the fabrics, with samples for mechanical testing prepared according to the American Society of Testing Material (ASTM 2013). Submersion dying was carried out using borosilicate glass beakers (250 mL) (brand VWR, Randor, PA, USA), filled with 100 mL of pigmented oil from *S. cuboideum*. Samples from tested fabrics were separately submerged in the pigment solution in the uncovered beakers for 1 h. Once the time of exposure was achieved, samples were removed from the borosilicate beakers and placed over a metal mesh to dry for 48 h.

#### 2.3.1. Tear Strength Test

The tear strength test was performed according to the ASTM D1424 using a falling-pendulum digital Elmendorf-type (produced by SDL ATLAS, Rock Hills, SC, USA) to measure the force required to propagate a single-rip tear. This test was performed on the samples of the four fabrics in filling directions. Twelve repetitions were performed on both dyed and undyed control samples (treated with only raw linseed oil), making a total of 36 samples per fiber fabric and 144 samples in total.

Estimated least squares means tests with Tukey adjustment was performed using SAS 9.8 (SAS Institute, Cary, NC, USA) to compare the effect of treatment and fabric type on stress (MPa) or force (N). The least squares means for main effects were reported if the main effect was statistically significant and not part of any statistically significant interaction or when the main effect was statistically significant and the effect of that factor was always in the same direction for the interacting variable(s). Selected samples were later evaluated using SEM analysis (below) to assess the behavior of pigment crystals inside the fabric structures.

#### 2.3.2. Tensile Strength Test

The breaking (tensile strength) test was performed according to the ASTM D5034-09 (2013), using the Universal Testing Machine Instron Model 5582 (produced by Instron Company, Norwood, MA, USA). Samples were loaded into the tool and then exposed to tensile forces until breaking occurred. The test was repeated for 30 samples for each fabric/density type, with 10 controls, 10 samples dyed with only raw linseed oil, and 10 samples dyed with pigmented oil. After the tensile strength test, affected samples were analyzed using SEM to evaluate any physical change in the attachment of the pigment to the fibers.

Estimated least squares means tests with Tukey adjustment were performed using SAS 9.8 (SAS Institute, Cary) to compare the effect of treatment and fabric type on f (N), and least squares means for main effects were reported if the main effect was statistically significant and not part of any statistically significant interaction or when the main effect was statistically significant and the effect of that factor was always in the same direction for the interacting variable(s). Selected samples were later evaluated using SEM analysis (below) to assess the behavior of pigment crystals inside the fabric structures.

#### *2.4. Analysis of Colorfastness*

#### 2.4.1. Color Variation Across Fabrics

Difference in pigment uptake and coloration were compared between fabrics. Fabric samples were cut into 5.08 × 5.08 cm, and dripping was used to apply fungal pigment. This consisted of applying pigment suspended in raw linseed oil onto the surface of each fabric sample using a disposable pipette. Fifteen drops of the oil-solubilized pigments were applied onto each fabric sample using a 1 mL pipette (brand Gilson, Lewis Center, OH, USA), with drops weighing an average 0.02443 g.

Control samples were compared to samples colored using the dripping methodology, and color analysis was performed on the samples using a Konica Minolta Chroma Meter CR-5 colorimeter, using the CIE *L*\**a*\**b*\* color space, with the ∆E calculation 2000. A one-way ANOVA was performed to test for the effect of fabric type on the response of color variation (∆E) using SAS 9.8 (SAS Institute, Cary).

#### 2.4.2. Laundry Test

Colorfastness to laundering with and without bleach was performed on the fabric samples dyed with the dripping method. The washing test was performed according to the AATCC method 61-2013 (Colorfastness to Laundering: Accelerated Using a Launder-O-Meter) in a Launderometer model LEF (produced by SDL ATLAS). Twelve 5.08 × 5.08 cm samples of each tested fabric type were prepared using dripping methodology (above), with six replicated exposed to bleach in testing and six without exposure. Samples were placed in a canister containing 50 stainless steel balls and AATCC standard detergent solution (with or without bleach). The sealed canisters were placed in the launder-O-meter with hot water (49 ◦C) for 45 min. When the process finished, samples inside the canister were rinsed and air-dried for 48 h. The color was measured in the Konica Minolta Chroma Meter CR-5, utilizing the CIE *L*\**a*\**b*\* color space. A general linear model was used to analyze the interactions between the independent variables of laundry, fabric type, heat, and time of heating with a four-way ANOVA and Tukey–Kramer test (*p* < 0.05) to determine significant differences between groups. One sample for each fabric/treatment was randomly selected and was taken to be analyzed with SEM (below) to visualize any change from the control samples. The response of prepared pigmented fabrics to heating was then compared, with samples previously subjected to heat testing.

#### 2.4.3. Heat Testing

Testing samples were prepared following methods described for the laundry test (using the same number of samples), then heated to one of three common dryer temperatures (low = 50 ◦C, medium = 65 ◦C, and high = 80 ◦C) in a forced air oven to simulate a home dryer. Initial color readings were taken after samples dried using a Konica Minolta Chroma Meter CR-5 colorimeter using the CIE *L*\**a*\**b*\* color space, using the ∆E calculation 2000. Color change was then again measured after 30 and 60 min exposure to heat, and samples were removed with tweezers and transported in baggies to prevent dirtying. Once the color was measured, two samples were randomly selected from each

fabric/treatment and analyzed via SEM (below) to visualize any change that occurred due to the heat. A repeated measurement ANOVA with a mixed model was performed to determine the difference in ∆E and any interactions between the independent variables fabric, time, heat, and treatment, and a Tukey Kramer test was performed to classify possible interactions.

#### 2.4.4. Qualitative Analysis: SEM

The purpose of SEM was to visualize the physical interaction between the pigment when carried in raw linseed oil and the fabrics and how this varied between treatment groups and the control. All tested samples (noted above) were dried for 48 h at room temperature, then were mounted on aluminum studs 5 mm in diameter using carbon adhesive tape. Samples were coated following Vega Gutierrez (2016), with gold-palladium in a Cressington Sputter Coater 108 Auto (Cressington Scientific Instruments, Inc, Cranberry Twp, PA, USA) for 35 s to allow the samples to generate a coating that decreased the electron charging and increased the contrast on the sample images. Samples were analyzed with a FEI QUANTA 600F environmental SEM (FEI Co., Hillsboro, OR, USA). The electron spot used was 2 A and 2 kV (high voltage).

#### **3. Results**

#### *3.1. Influence of Fabric Type on Coloration of Samples*

The color intensity of dyed fabrics was found to vary significantly (df = 619, F = 150.15, *p* < 0.001) depending on the type of fabric, with Tukey groupings showing that low-density cotton had the most pigmentation (∆E = 35.58 ± 1.73), followed by high-density cotton (33.21 ± 3.54), and low-density polyester (27.61 ± 4.22). Nylon, high-density polyester, high-density wool, and low-density wool were not found to differ significantly and showed lower coloration overall.

#### *3.2. Mechanical Testing*

#### 3.2.1. Tensile Strength

No significant difference was seen in tensile strength between untreated fabric samples, fabric samples treated with linseed oil alone, and samples treated with pigmented oil (df = 2, F = 1.75, *p* = 0.1806). The fabric type was found to significantly affect the sample tensile strength (df = 6, F = 114.55, *p* < 0.0001), though no interaction was found between the interaction of fabric type and treatment, which would have indicated differences in strength changes for one or more of the fabrics due to the pigment (df = 12, F = 0.25, *p* = 0.9944).

SEM analysis showed differences in fiber morphology after tensile testing between tested fabric varieties. High-density cotton showed flattened or curvy twisted fibers close to the break area, with hook-like ends at the ends of fibers (Figure S1). Low density cotton also showed curvy, twisted fibers close to break point, with split fiber ends (Figure S2). Nylon showed flattening at the area of breakage and wide separation of fibers, with oil accumulation visible on their surface (Figure S3). High-density polyester showed uneven broken fiber ends and minor flattening of broken ends (Figure S4). Low-density polyester in contrast showed an irregular twisted surface and wavy fibers in the failure area, with oil accumulation evident (Figure S5). High-density wool showed flattened and elongated fibers with a sharp angle on failure sites, showcasing the separation of the cuticle from the fiber core (Figure S6). Low-density wool showed similar characteristics to high-density wool, with clear damage from the test (Figure S7).

#### 3.2.2. Tear Strength

The tearing strength test showed a significant interaction between fabric type and treatment (df = 12, F = 9.76, *p* < 0.0001), with low density cotton showing a significant difference between untreated samples and samples treated with pigmented or non-pigmented linseed oil. Control samples tore at a significantly lower strength of 912.58 N compared to pigmented oil samples. Tukey groupings showed that the low-density wool with oil had the highest tear strength performance (5068.36 N), while high-density cotton treated with oil had the lowest performance (649.67 N) in this test.

High-density cotton showed clear flattened, broken ends with accumulated oil at the ends (Figure S8), while low-density cotton showed twisted sharp ends and wrinkled fibril bundles (Figure S9). High-density wool showed an even shape in the failure area with no loss of cuticle integrity, though a flattened core was present (Figure S10). Low-density wool showed cuticle loss and irregular surfaces, with breaking areas showing visible structural damage (Figure S11). Nylon showed un-flattened fibers with occasional bracket-like structures with accumulation of oil close to the break area on fiber ends, though most sections did not show this oil clumping (Figure S12). High-density polyester showed angled edges and flattened fibers, with oil accumulation and waviness seen in broken fiber ends (Figure S13). These was also seen in low-density polyester, which also showed twisting in the fibers (Figure S14).

#### *3.3. Laundry and Heat Testing*

Color change in response to laundry and heat processes was found to vary in response to testing conditions. The four-way ANOVA interaction between fabric type, laundry type (bleach/no bleach), temperature, and time was not found to be significant, nor were any three-way interactions. Significant two-way interactions included those between heating time and temperature (df = 4, f = 9.07, *p* = 0.0001), laundry type and temperature (df = 4, f = 7.66, *p* = 0.0005), heating time and temperature (df = 2, f = 9.07, *p* = 0.001), fabric type and temperature (df = 9, f = 1.88, *p* = 0.0349), and fabric type and laundry treatment (df = 8, f = 24.67, *p* <0.0001). Overall, these significant interactions suggested that fabrics responded differently to pigmentation, with bleach leading to discoloration and increased exposure to heat (either over time or at higher temperatures) leading to more intense color. Drying temperature seemed especially important, as it was present in all significant interactions, although its effects were mediated by other variables.

In general, the addition of bleach resulted in less color difference between pigmented samples and controls, although for low-density cotton, the addition of bleach resulted in a significantly higher ∆E (more color change). Polyester and nylon showed significant degradation of the pigments after laundry treatment, both with and without bleaching, as can be seen in Figure 1. However, in this case, the ∆E value does not tell the whole story. ∆E measures the total change in color, and, while in most cases in this study this change reflects the intensity of the red coloration, when bleach was applied on a number of fabrics, the hue of the pigmentation changed. In wool of both densities, the color changed from red to blue, and in cotton, the color changed from red to dark purple (Figure 2). Overall, bleach caused a more visible color change on natural fabrics over synthetic fibers. Synthetic fabrics followed a different pattern, presenting less color change after washing and drying.

In addition to the changes in color, as measured by ∆E, and the variation of this effect upon different fabrics, bleach seemed to have a profound effect on the applied pigment in oil, as was qualitatively seen in SEM. While fabrics did not appear degraded in SEM, pigments on samples showed evidence of breakdown. Figure 1 shows pigments looking different in the bleached sample compared to the unbleached sample on the same fabrics at the same drying temperature and time. The analysis of SEM images identified the loss of the pigment layer produced by the interaction of the fabric with the oil. As can be seen in Figure 3, the degradation of oil layers occurred regardless of whether bleach was used during laundry tests. When bleach was used, it severely altered the dye color. All cases showed a strong contrast in comparison with the initial coloration and/or a dramatic loss of the dyed area.

∆ ∆

∆

**Figure 2.** Differences in color and saturation between various fibers treated with 15 drops of dramada-pigmented oil across laundry without bleach, laundry with bleach, and the control (unwashed samples).

∆

∆

**Figure 3.** SEM images of a bleach/no bleach sample at the same temperature, showing degradation. (**a**) Low density polyester (control 15 drops); (**b**) low density polyester laundry without bleach; (**c**) low density polyester laundry with bleach.

No dramada crystals were seen in cotton fabrics, with high density cotton showing a smooth covering of oil with no presence of crystals, which became porous after laundry treatment without bleach, and, in the presence of bleach, no oil was found on the samples (Figure S15). Low density cotton showed a less smooth surface of applied pigment with fissures on the outer layer, likely due to retention in the most superficial layer (Figure S16b). After washing with and without bleach, low density cotton showed retention of a thick layer of pigmented oil with an irregular porous shape (Figure S16). Nylon also showed no crystals in the pigmented oil layer, with an irregular surface that was actually smoother after laundry without bleach. In contrast, laundry with bleach resulted in stripping of much of the oil layer remaining only in spaces between yarns (Figure S17).

High density wool also did not form visible crystals within the pigment layer, which formed a thick, even layer of oil that remained after laundry without bleach. However, laundry with bleach showed visible interstices and a porous layer (Figure S18). The low-density wool allowed for a higher degree of infiltration of the oil into the lower layer of the fabric than the high-density wool. Low density wool also showed a thick pigment layer covering the fibers, though interactions with inner layers of the wood fibers made it difficult to distinguish if a granular surface seen where pigment was directly applied was due to pigment/wool interaction or not. After laundry without bleach, the oil absorbance of the cotton became higher than wool or polyester, as seen through the interstices in the oil layer (indicating a dryer oil coat) (Figure S19c), with samples exposed to bleach also showing a reduction in oil and flaking along a porous surface.

In contrast to other fabrics, polyester showed high concentrations of dramada crystals in below oil layers. High-density polyester showed a somewhat more even oil distribution, with a smoother surface seen after laundry with bleach and a lack of visible coating after laundry with bleach (Figure 4). Low density polyester also showed a reduced oil layer after laundry, along with several facture points. Oil also seemed to be degraded by detergent and bleach, forming cube-like structures between fibers and an overall porous surface (Figure 5). Low density polyester also showed a reduced oil layer after laundry, along with several facture points (Figure S20).

**Figure 4.** SEM images of high-density polyester samples: (**a**) control; (**b**) sample treated with 15 drops of pigmented oil, showing presence of dramada crystals; (**c**) pigmented sample after laundry treatment with no bleach, showing a somewhat smoothed surface; (**d**) pigmented sample after laundry treatment with bleach showing no visible oil layer.

**Figure 5.** Distinctive features seen in polyester. (**a**) High-density polyester sample with a dramada crystal; (**b**) SEM image showing low-density polyester sample after the laundry test with bleach.

∆

∆

### *3.4. Heat Testing*

Significant variations of ∆E were seen between pigment treated fabrics exposed to drying conditions at either 50, 65, or 80 ◦C conditions for either 30 or 60 min. The four-way interaction between all these variables was not statistically significant (df = 12, F = 1.47, *p* = 0.1366), though two three-way interactions were found to be statistically significant. Fabric type and heating were found to influence color change after heating (df = 2, F = 7.41, *p* = 0.0008), with all fabrics having higher ∆E value after the heat treatment. In addition, the interaction between time, temperature, and treatment was significant (df = 12, F = 3.21, *p* = 0.0003), with an increase in ∆E after the heat process; a difference that can be seen in Figure 6. Tukey groups showed that high-density cotton and low-density polyester showed the highest (∆E), while nylon showed the lowest (∆E). ∆ ∆

#### **4. Discussion**

The red pigment produced by *Scytalidium cuboideum*, carried in oil, was found to successfully dye a variety of fabric types without impact on tested mechanical properties, though wide variation in response to colorfastness treatments was seen. The color of the oil pigment varied based on fabric. These differences could be seen in the presence and morphology of the pigmented oil layer under SEM.

The presence of the pigment in oil layer did not affect material strength or tensile strength. As natural dyes have historically been associated with weakening of fabric and decreases in tear strength due their need to use mordants for effective dying [26–28], this is a significant benefit for use of spalting fungal pigments.

Pigmented low density cotton showed a significant difference in tear strength in comparison to untreated samples, but no significance was seen between pigmented oil and non-pigmented oil. This suggests that any differences might be due to the absorption of oil, which could have influenced slippage, leading to tearing. It is also possible that cellulose functional groups may have interacted with pigmented oil and formed a stronger bond; however, this would require further investigation. The fact that significance was only seen in this fabric type further suggests that this effect was not due to the pigment presence and instead suggests that factors, such as fabric thickness and diffusion, may influence dye absorption. Additionally, SEM showed that the pigmented oil did not stay in the internal structures of the high-density cotton but it did stay in the internal structures of low-density cotton, which had more space to absorb it. Both tensile strength and tear strength testing results mirrored those found by Hinsch [19], who reported no significant differences in strength values after application of *S. cuboideum* pigment in the DCM carrier.

While pigment presence was not seen to influence mechanical testing, clear differences in coloration of applied pigment were seen between fabric varieties with both cotton varieties and low-density polyester, showing significantly higher coloration than other fabric types (*p* < 0.001). It is likely that structural differences between fabric types are responsible for these differences in coloration. The more complex shapes created by natural fibers presented a different surface, leading to different interactions with the pigmented oil. A deep view with SEM showed that polyester fibers seemed to have the dye attached mostly in the outer areas of the sample. Polyester fabrics are normally hydrophobic materials; therefore, their affinity for oil-based compounds was supposed to be higher than that of cotton and wool (which are normally lipophobic materials). Previous research [1,19] did not show any relationship between this feature of the fabrics and their ability to be dyed; however, both studies found that polyester had a general affinity for natural dyes. This relationship still needs to be explored with other types of polyester and fabric blends.

While differences were seen between fabric varieties, density differences did not result in consistent patterns of pigment behavior. While it was speculated that the crystal structure contained in pigment from *S. cuboideum* could be entrapped differentially in fabrics of varying densities, SEM investigations showed that fabric density was also not related to crystal presence, with pigmented oil staying in the fibers without forming crystals. In most cases, a thin layer of oil formed on the surface and a marked accumulation was found in the intersections of each yarn. More crystals may have formed with time, as crystals from *S. cuboideum* have been shown to appear in oil after 60 days, potentially due to oxidation processes [29].

The natural fibers retained coloration better than the synthetics, with wool fabrics, in particular, showing stable color after laundering. This has been seen with other fungal dyes (although not from spalting fungi), specifically with *Talaromyces australis* (C.M. Visagie, N. Yilmaz & J.C. Frisvad) and *Penicillium murcianum* (C. Ramírez & A.T. Martínez) [30]. In most cases, laundry with detergent alone was sufficient to remove the oil layer from the applied pigment. It is likely that the detergent used in the laundry test affected the triester groups of the linseed oil through a complex formation with available -OH<sup>−</sup> groups. However, complete removal was not seen on wool, and cotton samples though the coloration persisted, suggesting that the oil may influence pigment fixation, creating a mordant-like effect that has been described in other studies [31].

The pigmented oil layer degraded on exposure to bleach, likely through a saponification reaction. The strong alkalinity of bleach is also likely responsible for the color change seen in laundry samples. The pigment from *S. cuboideum* is known to turn blue under basic conditions, instead of its normal red coloration [32,33]. The mechanism behind this color change is under investigation but has not yet been described. It is possible that the variation in response to bleach seen may have been due to

physical protection of pigment-functional groups pigment by the oil layer or by pigment binding to fabric surfaces.

While bonding mechanisms of fungal pigment attachment to different fabrics were not tested, these likely impacted the differential responses seen in colorfastness. It is possible that hydroxy groups in cotton, and potentially amino acids in wool, may form hydrogen bonds with the methoxy and hydroxy groups present in the fungal pigment dramada. For the polyester fabric, it is likely that the aromatic π-electron system of the phthalic unit may be interacting with the π-electron system of dramada, providing a strong bond between the fabric and the pigmented oil. This interaction has already been reported with an anthraquinone present in natural dyes [1]. It is also interesting that. in previous research studies, polyester fabrics were successfully dyed with *S. cuboideum* when using DCM as a carrier [18] and later with raw linseed oil [25]. However, there is a need to further investigate which interaction causes the bonding between the fabric and the dye.

Heating resulted in changes of fabric color, as it was present in all statistically significant interactions identified, with all fabrics brightening except nylon, which may have been due to low initial pigment uptake. Under SEM, a superficial layer could be seen on the dyed areas after heating. This layer was also seen in samples subjected to laundry testing and it appeared to be made of oil: it presented as a thin, compacted, smooth surface, and, in some cases, it showed interstices or pores. This layer was best seen in high-density wool and low-density cotton, which showed the best performance, as measured by ∆E after the heat test. The layer may have been formed due to evaporation, oxidation, or polymerization of the oil under the effect of the temperatures. The polymerization process of linseed oil is well known [34] and this process may have allowed for molecular stacking of the pigment into its crystal formation. Using high temperatures to accelerate the evaporation of the raw linseed oil has been previously reported in the preparation of oilcloths as a way to reduce drying time [35], although no oil layer, such as the one found in this research, has ever been reported. It is also notable that previous research has shown that temperatures higher than 70 ◦C have been shown to degrade the red color of dramada [29], although this effect was not seen during this research. It is possible that the oil provided some protection for the pigment.

While this research showed that a range of fabrics could be effectively dyed using fungal pigment from *Scytalidium cuboideum*, this method of dying presented colorfastness challenges, especially for synthetic fibers. In addition, the texture of completed cloth may be an obstacle, as the oil does not polymerize to a dry layer quickly, making samples oily. However, treatment of cotton with linseed oil to create a waterproof layer, known as an oilcloth, was commonly done in the past [36] and may be a sustainable waterproofing option. Testing other cellulosic fibers, such as bamboo, rayon, or hemp, should also be investigated, as should the water repellence of the pigmented oil layer. In addition, new technology would likely need to be developed to allow for industrial processing of quantities of this pigmented oil. The economics of this product would require further research. In addition, natural oils have been researched for protection of exterior wood in service [37,38], which may be an additional use.

The challenges seen using pigment from *S. cuboideum* in linseed oil may limit the ability of this technology to be adopted in the market. However, the use of spalting fungal pigments like that from *S. cuboideum* still have great potential for use in the textile industry, especially as they can be used without the need for mordants [18]. Other natural pigments, such as those from plant sources, have relied on mordants for their color stability [39–42], and the lack of mordants needed for fungal pigments represents a significant advantage. While it was hoped that linseed oil would provide a less toxic alternative to DCM, future work into identifying new solvents appropriate for use with dramada would be helpful, especially for use with synthetic fabrics. Recent work using pigment from *S. cuboideum* in inkjet printers may be a particularly effective method of textile dying with fungal pigments [43].

### **5. Conclusions**

Red pigment from *Scytalidium cuboideum* carried in linseed oil was found to effectively dye a variety of fabric varieties, with natural fibers showing better coloration and colorfastness overall. Mechanical properties were found to be essentially unaffected by the application of pigment. Colorfastness was shown to vary among fabrics, with application of bleach associated with a color shift from red to blue of the applied pigment. Heating was found to result in the production of a bright red fabric, which may have potential as a sustainable waterproofing treatment.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2309-608X/6/2/53/s1, Figure S1: SEM images of high-density cotton exposed to tensile strength test; Figure S2: SEM images of low-density cotton exposed to tensile strength test; Figure S3: SEM images of nylon exposed to tensile strength test; Figure S4: SEM images of high-density polyester exposed to tensile strength test; Figure S5: SEM images of low-density polyester exposed to tensile strength test; Figure S6: SEM images of high-density wool exposed to tensile strength test; Figure S7: SEM images of low-density wool exposed to tensile strength test; Figure S8: SEM images of high-density cotton for tear strength test; Figure S9: SEM images of low-density cotton for tear strength test; Figure S10: SEM images of high-density wool for tear strength test; Figure S11: SEM images of low-density wool for tear strength test; Figure S12: SEM images of nylon for tear strength test; Figure S13: SEM images of high-density polyester for tear strength test; Figure S14: SEM images of low-density polyester for tear strength test; Figure S15: SEM images of high-density cotton samples from laundry test; Figure S16: SEM images of low-density cotton samples from laundry test; Figure S17: SEM images of nylon samples from laundry test; Figure S18: SEM images of high-density wool samples from laundry test; Figure S19: SEM images of low-density wool samples from laundry test; Figure S20: SEM images of low-density polyester samples.

**Author Contributions:** Conceptualization, S.C.R. and H.-L.C.; methodology, S.C.R. and H.-L.C.; validation, S.C.R. and H.-L.C.; formal analysis, M.E.P.A., S.M.V.G., and H.-L.C.; investigation, M.E.P.A.; resources, S.C.R. and H.-L.C.; data curation, M.E.P.A. and S.M.V.G.; writing—original draft preparation, M.E.P.A. and S.V.M.G.; writing—review and editing, R.C.V.C., S.M.V.G., H.-L.C., and S.C.R.; supervision, S.C.R. and H.-L.C.; project administration, S.C.R. and H.-L.C.; funding acquisition, S.C.R. and H.-L.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Walmart Foundation Manufacturing Fund.

**Acknowledgments:** Authors gratefully acknowledge Oregon State University Electron Microscopy Facilities, Peter Eschbach, and Teresa Sawyer and would like to thank Ariel Muldoon for help with statistical analysis.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Review*
