Effects of Lyophilized Dietary Yeast Rhodotorula mucilaginosa on Skin and Fillet Pigmentation of Gilthead Seabream (Sparus aurata): A Computer-Based Image Analysis Assessment

Abstract

Skin pigmentation is a crucial factor influencing the market value of gilthead seabream. A three-month feeding trial evaluated the effects of dietary inclusion of Rhodotorula mucilaginosa on skin and fillet pigmentation of gilthead seabream (Sparus aurata). Four diets containing yeast were tested in triplicate tanks using 120 fish in total. Skin and fillet colours were assessed via computer-based image analysis in CIELAB, RGB and HSB spaces. Analysis of total carotenoids was also performed. Yeast inclusion increased L* and Whiteness values in the operculum and enhanced lightness in ventral skin regions. In the abdominal area, RGB values decreased, particularly in the 3% diet. Fillet responses were limited to the red muscle, where the 3% diet significantly increased a* and Chroma values. Overall, inclusion of R. mucilaginosa exceeding 2% influenced seabream skin brightness and total carotenoid content, while 3% inclusion enhanced red muscle pigmentation, suggesting potential as a natural pigment source in seabream feeds.

1. Introduction

The market value of gilthead seabream is based to some extent on its skin pigmentation, with the golden-yellow in the forehead area between the eyes being of great importance [1,2]. Nonetheless, colourimeters cannot accurately estimate the non-homogeneous fish pigmentation [3,4]. This could be attributed to its circular aperture with a diameter of 8 mm, which could lead to the omission of colours and subtones during measurement [5]. Computer-based image (CBI) analysis can accurately measure colour, providing homogeneous conditions without the need for direct contact with tissue under examination [6].

In addition to methodological advances in colour assessment, there is increasing interest in sustainable pigment sources to enhance external appearance and product quality. Among many fish, gilthead seabream cannot synthesize carotenoids de novo and thus relies solely on its diet [7]. Seabream skin pigmentation is under multiparametric regulation, with numerous internal factors like hormones and external factors like nutrition and the environment influencing the chromatic state of the fish [8]. Consequently, carotenoids derived either from natural or synthetic sources have been playing an important role in fish nutrition [9]. Synthetic astaxanthin has been widely used as a feed additive to enhance seabream pigmentation [10]. Nevertheless, there remains considerable scope for further research on natural pigment sources to support the transition towards more sustainable aquaculture practices.

Yeasts could be used as feed additives belonging to the sustainable feed ingredients category [11]. Moreover, they can produce metabolites like carotenoids, which could potentially enhance the health of reared fish [12]. Rhodotorula mucilaginosa has been identified as one of the permanent strains in the gut microflora of certain fish. It is known for producing pigments, the main of those being β-carotene, torulene and torularhodin [13,14]. Although synthetic astaxanthin is extensively used, evidence on yeast-derived pigments in gilthead seabream remains scarce. Thus, this yeast could be potentially used as a natural source of pigments in aquafeeds. Based upon this background, the aim of this study was to determine the potential of the pigmented yeast R. mucilaginosa as a pigmentation source for the gilthead seabream S. aurata using computer-based image analysis as a novel approach to quantify pigmentation.

2. Materials and Methods

Four diets containing 0% (C), 1% (RM1), 2% (RM2) and 3% (RM3) lyophilized Rhodotorula mucilaginosa strain ACA-DC 5340 were prepared and fed to gilthead seabream over a three-month period. Hereafter, “RM treatments” refers collectively to the yeast-supplemented diets (RM1–RM3). Fish were hand-fed twice daily at a 1.2% BW. A detailed description of the diet formulation is provided in Zantioti et al. 2026 [15]. R. mucilaginosa carotenoids were estimated at 207. 74 ± 8.23 μg/g [15]. A total of 120 fish were randomly allocated to 12 tanks (n = 10 per tank, three replicates per diet). Fish were fed twice daily, and environmental conditions were kept constant throughout the trial.

Six fish per treatment were randomly sampled for skin and fillet colour analysis. Fish and fillets were photographed (4032 × 3024 pixels) under standardized lighting conditions (Puluz Photobox 60 × 60 × 60 cm, PULUZ, Shenzhen, China), and five skin regions (Figure 1) and three fillet regions (Figure 2) were analyzed. Colour parameters (CIELAB’s L*- a*- b*, RGB and HSB) were extracted using computer-based image analysis (Adobe Photoshop 2022 Version 23.5.1, Adobe Inc., San Jose, CA, USA), and Chroma and Whiteness were calculated according to the following formulas. CIELAB parameters (L*, a*, b*) and derived indices represent conventional chromaticity standards for objective colour evaluation, while the additional colour spaces were chosen to provide a more holistic characterization of pigmentation. Full definitions and calculations of the colour parameters are provided in the Supplementary Materials (Table S1). Tables containing all measured parameters and their values for each anatomical region are provided in the Supplementary Materials (Table S2).

$$\mathrm{Chroma}=\sqrt{{\left({\mathit{a}}^{\mathbf{\ast }}\right)}^{\mathbf{2}}\mathbf{+}{\left({\mathit{b}}^{\mathbf{\ast }}\right)}^{\mathbf{2}}}$$

$$\mathrm{Whiteness}=100-\sqrt{{\mathbf{(}\mathbf{100}\mathbf{-}{\mathit{L}}^{\mathbf{\ast }}\mathbf{)}}^{\mathbf{2}}\mathbf{+}{\mathbf{\left(}{\mathit{a}}^{\mathbf{\ast }}\mathbf{\right)}}^{\mathbf{2}}\mathbf{+}{\mathbf{\left(}{\mathit{b}}^{\mathbf{\ast }}\mathbf{\right)}}^{\mathbf{2}}}$$

Colour assessment was performed by pixel-based analysis in Adobe Photoshop. For skin regions 1 and 2, a 3 × 3-pixel sampling area was used, whereas for skin regions 3, 4 and 5, an 11 × 11-pixel sampling area was applied. For fillet colour analysis, a 3 × 3-pixel sampling area was used. For each sampling point, three replicate measurements were performed and mean values were used for statistical analysis. Total carotenoid content in skin and fillet samples was determined according to the method described by Waseef et al. 2010, with slight modifications [16]. To ensure representative sampling, skin was collected from various locations on each fish and combined into a single pooled sample per individual before analysis. Data were analysed by one-way ANOVA followed by Tukey’s test. Non-normal datasets were evaluated using Kruskal–Wallis or Welch ANOVA. Statistical significance was set at p < 0.05 [17].

3. Results

Dietary R. mucilaginosa produced several significant effects on skin pigmentation (

Table 1. Summary of significant pigmentation responses in gilthead seabream fed diets containing different inclusion levels of the yeast Rhodotorula mucilaginosa. Only parameters exhibiting statistically significant differences compared to the control group are presented. All colour parameters for all sampling points are provided in the Supplementary Materials. (L*): denotes lightness in the CIE colour space, (a*): denotes the red–green axis in the CIE colour space, (b*): denotes the yellow–blue axis in the CIE colour space, (R): denotes red in the RGB colour space, (G): denotes green in the RGB colour space, (B): denotes blue in the RGB colour space, arrows (↑, ↓) indicate an increase or decrease in the respective values.

	Parameters Affected	Effect	Treatments Affected	Explanation
Skin
Forehead (spot 1)	b*, Chroma	↓	RM1, RM2, RM3	Reduced yellow pigmentation in the forefront band
Operculum (spot 2)	L*, Whiteness, R, G, B	↑	RM1, RM2, RM3	Increased brightness and lightness
Below lateral line (spot 4)	L*, Whiteness, R, G, B	↑	RM2, RM3	Enhanced lightness and RGB intensity in ventral skin
Abdominal area (spot 5)	R, G	↓	RM1, RM3	Lower red and green values in the ventral skin
Fillet
White muscle (spot 1)	b*	↑	RM1, RM3	Slight increase in yellowness, inconsistent across spots
Red muscle (spot 2)	a*, Chroma ↑ Whiteness ↓	↑/↓	RM3	Enhancement of red muscle pigmentation at 3% inclusion

). In the forehead region, fish fed RM diets showed significantly lower b*, Saturation and Chroma values compared to the control (p < 0.001). In the operculum, all yeast-supplemented groups exhibited significantly higher L* and Whiteness values as well as higher RGB intensities (p < 0.01) (Table 1). This could indicate increased brightness in this region. Below the lateral line, fish fed RM2 and RM3 diets displayed significantly higher L*, Whiteness and B values (p < 0.05) consistent with a lighter appearance while RM2 also showed significantly increased R and G values compared to the control (Table 1). In the abdominal region, all RM diets resulted in significantly lower R and G values, with RM3 showing the strongest reduction (p < 0.05) (Table 1).

Although three fillet spots were measured, only the two most responsive spots are shown in Table 1 for conciseness, while the full results for all spots are provided in Supplementary Table S1. Fillet pigmentation differences were limited to specific regions. In white muscle, only b* differed significantly, with RM1 and RM3 showing higher values than the control (p < 0.05), indicating increased yellowness. In contrast, the red muscle in group RM3 exhibited significantly higher a* and Chroma values and lower Whiteness (p < 0.01), indicating an enhancement of red colouration at the highest inclusion levels (Table 1). Total carotenoid concentration in the skin differed among treatments (p < 0.05), with RM2 and RM3 showing significantly higher values (

Table 2. Total carotenoid content (mg/kg dry weight) in skin and fillet samples. Data are presented as mean ± standard error. Different superscripts within a row indicate significant differences. NS: not significant, (**): p-value < 0.01.

Total Carotenoids (mg/kg)
	C	RM1	RM2	RM3	p-Value
Skin	5.12 ± 0.65 b	4.64 ± 0.76 b	7.45 ± 0.55 a	8.27 ± 0.80 a	**
Fillet	1.38 ± 0.26	1.46 ± 0.15	2.32 ± 0.31	1.87 ± 0.29	NS

). Fillet carotenoid content was not affected by the dietary treatment.

4. Discussion

The colouration of Sparidae family is under multifactorial control, being influenced not only by endogenous but also exogenous factors such as nutrition [18]. Skin and fillet pigmentation of gilthead seabream responded selectively to dietary Rhodotorula mucilaginosa. Similar yeast-based pigmentation effects have been reported in other species. Rhodotorula paludigena was shown to influence scale colouration in ornamental koi carp [19]. In the forehead zone, all yeast diets reduced b* and Chroma values, contrasting with previous studies where microalgal carotenoids increased yellowness [1]. This reduction may reflect the carotenoid profile of R. mucilaginosa, which produces mainly β-carotene, torulene and torularhodin rather than the astaxanthin-type pigments naturally associated with yellow hues in Sparidae [20,21]. It is important to note that carotenoid bioavailability is a complex process that involves various steps and phases like liberation, absorption, transport and distribution [22]. Minor methodological effects linked to lateral photography may also have contributed to lower measured yellowness. Pulcini et al. mentioned that lateral shots of the forehead zone between the eyes lead to the recognition of a reduced number of yellow pixels by CBI software compared to frontal shots (2020). CBI appears to provide colour estimates more consistent with visual perception. In Atlantic salmon, CBI-derived colour values corresponded to the expected orange fillet colour. In contrast, values obtained with a colourimeter were associated with a purplish hue [23,24]. Generally, CBI analysis has been increasingly used as a non-invasive and practical alternative for pigmentation analysis; nevertheless, method validation against reference instrumentation remains an important next step.

In contrast, brightness-related parameters increased in several skin regions. Higher L* and Whiteness values in the operculum and ventral body areas suggest enhanced lightness, a pattern similarly reported when seabream was fed diets with inclusion of Phaeodactylum tricornutum [25]. These changes indicate that yeast pigments or associated compounds can alter skin appearance even if they do not intensify yellow coloration. From an applied perspective, increased brightness may be commercially relevant, as discolouration or colour fading is often associated with reduced perceived freshness and lower consumer acceptance [8]. Interestingly, the biological role of LDL may also be relevant to the transport of yeast-derived carotenoids, which are absorbed with dietary lipids and circulate via lipoproteins. In our previous trial with dietary lyophilized R. mucilaginosa, the observed changes in LDL support the hypothesis that the yeast may influence lipid transport and carotenoid trafficking, thereby contributing to pigmentation-related outcomes [15]. The changes observed in CBI are further supported by the analysis of total carotenoids in the skin, which revealed increased carotenoid content in fish fed yeast-containing diets at inclusion levels above 2%. Previous studies have shown that natural carotenoids can be efficiently bioabsorbed and deposited in gilthead seabream. However, this deposition does not necessarily translate into enhanced muscle pigmentation, a pattern also confirmed in the present study [16]. Further targeted analyses are required to identify which specific carotenoid compounds contribute to the observed increase in total carotenoid content.

Effects on fillet pigmentation were limited to specific tissues. White muscle showed only minor inconsistencies in b*, whereas red muscle exhibited a more pronounced response at the highest inclusion level, with increased a* and Chroma and reduced Whiteness. Red muscle, compared to white muscle, contains higher myoglobin levels and supports sustained aerobic swimming [26]. Given the higher vascularisation and lipid content of red muscle [27,28], preferential accumulation of yeast-derived carotenoids such as torularhodin may be plausible. Similar results were observed in trials with arctic, Salvelinus alpinus (L.), which exhibited increased fillet redness when fed diets containing astaxanthin [29]. Another natural source of natural pigment, the red macroalga Gracilariopsis persica, seemed to enhance the colouration of Persian sturgeon fillets [30]. Previous studies also suggest that the oxidation of myoglobin to metmyoglobin may cause a decrease in the a* value [31]. Myoglobin is also responsible for the pigmentation of the flesh, but its concentration in muscle tissue is related to various factors [32]. It is important to note that myoglobin redox status was not assessed in the present study and should be investigated in future work.

Dietary pigmentation strategies using both synthetic and natural sources like microalgae or microbial pigments have been widely evaluated across aquaculture species such as Salmo salar, Oncorhynchus mykiss and Pagrus pagrus, with dose-dependent and tissue-specific responses commonly reported [33,34]. Overall, R. mucilaginosa influenced pigmentation mainly through enhanced brightness and carotenoid content in the skin and increased redness in red muscle, with effects becoming more evident at inclusion levels above 2%. These results highlight the potential of pigmented yeasts as natural additives for modifying seabream appearance, although their carotenoid profile appears less effective in targeting yellow facial pigmentation. Although synthetic astaxanthin remains an efficacious additive to enhance the pigmentation of aquatic animals [35], yeast-derived pigments from R. mucilaginosa represent a promising natural alternative, but require further evaluation and dose optimization for practical application.

CBI has been shown before to correlate strongly with fish colour attributes linked to freshness and other quality indicators, with high correlation coefficients, highlighting its reliability as an objective tool for colour assessment [6]. Although in this study, CBI detected pigmentation differences, no direct comparison with a conventional colourimeter took place. Therefore, full methodological equivalence between CBI and chromatometry cannot be conclusively established within this study. Future studies should be performed using both techniques to directly evaluate their relative accuracy and precision.

5. Conclusions

In conclusion, the inclusion of the red yeast Rhodotorula mucilaginosa affected the skin pigmentation of gilthead seabream. Results become more noticeable at inclusion rates above 2%. Changes in the flesh pigmentation were noticeable only on the red muscle part of the fillet and at an inclusion level of 3%. From an applied perspective, pigmentation is a key quality trait influencing consumer perception and market value in gilthead seabream, and the observed colour modulation supports the interest in sustainable pigment strategies for aquafeeds. These findings support the potential of R. mucilaginosa as a dietary source of natural pigments. However, the underlying mechanisms were not directly assessed and remain to be validated. Future work involving larger sample sizes should examine pigment deposition and antioxidant activity to clarify the mechanisms involved. Also, histological examination and gene expression analyses of pigment-related pathways could help elucidate the mechanisms underlying colour changes. Lastly, although CBI was able to detect pigmentation alterations at different yeast inclusion levels, further work directly comparing CBI with conventional colourimeters is required to fully validate its accuracy and precision.