# Fermentative Production of Erythritol from Cane Molasses Using Candida magnoliae: Media Optimization, Purification, and Characterization

^{*}

## Abstract

**:**

_{2}PO

_{4}out of 12 contributing factors. Further, the concentration of molasses (200–300 g/L), yeast extract (9–12 g/L), and KH

_{2}PO

_{4}(2–5 g/L) were optimized using response surface methodology coupled with numerical optimization. The optimized erythritol yield (99.54 g·L

^{−1}) was obtained when the media consisted of 273.96 g·L

^{−1}molasses, 10.25 g·L

^{−1}yeast extract, and 3.28 g·L

^{−1}KH

_{2}PO

_{4}in the medium. After purification, the liquid chromatography–mass spectrometry (LC-MS) analysis of erythritol crystals from this optimized fermentation condition showed 94% purity. Glycerol was produced as the side product (5.4%) followed by a trace amount of sucrose and mannitol. The molecular masses of the erythritol were determined through mass spectrometry by comparing [M + Na] + ions. Analysis in electrospray (ES) positive mode gave (m/z) of 145.12 [M + 23]. This study has reported a higher erythritol yield from molasses and used osmotolerant yeast Candida magnoliae to assimilate the sucrose from molasses.

## 1. Introduction

## 2. Materials and Methods

#### 2.1. Materials

#### 2.2. Culture Conditions for Measuring the Growth Curve

^{−1}), yeast extract (10 g·L

^{−1}), KH

_{2}PO

_{4}(5 g·L

^{−1}), and MgSO

_{4}·7H

_{2}O (0.25 g·L

^{−1}). A loop-full of culture from freshly prepared slants of Candida magnoliae 3470 was inoculated into 10 mL LFM in test tubes and incubated at 28 °C, 210 rpm, for 48 h. Two milliliters of this seed culture was aseptically transferred to 23 mL LFM, and the fermentation broth was incubated at 28 °C, 210 rpm for 168 h. The broth samples were withdrawn at successive intervals and analyzed for optical density (OD) at 600 nm. The dry cell weight (DCW), plate count, and change in pH of C. magnoliae were estimated from a corresponding standard curve. The erythritol yield was estimated using the erythrose reductase enzyme assay.

#### 2.3. Preparation of Cell Extracts

_{2}) for 30 min. The cell suspension was homogenized by an ultrasonicator at 180 W, 70% duty cycle for 30 min with intermittent cooling to prevent denaturation of proteins. Enzyme extract was centrifuged at 10,000× g for 30 min at 4 °C. The supernatant was analyzed for protein content by Lowry’s test using BSA (bovine serum albumin) as a standard to obtain 74.11 mg/mL of protein [28].

#### 2.4. Erythrose Reductase Assay

#### 2.5. Fermentative Production of Erythritol

_{2}PO

_{4}(phosphorus source), MgSO

_{4}(cofactor), and CaCO

_{3}(for pH stabilization), pH of the media, the volume of inoculum, temperature, time, agitation, and media volume. The Plackett–Burman design was employed to screen out the most influential factors among those 11 variables. In addition to erythritol yield, the dry cell weight and change in pH during fermentation were measured as two other responses to the Plackett–Burman design (Table 1).

_{i}) were converted into corresponding dimensionless coded values (x

_{i}) using Equation (1), where X

_{max}and X

_{min}are the maximum and minimum values of X

_{i}, respectively.

_{i}) on the response (Y

_{i}) was calculated using Equation (2).

_{i+}and Y

_{i−}are the response values obtained for the factor x

_{i}varied at high (+1) and low (−1) levels, respectively, and n represents the number of trials. From the relative contribution (%) values in the Pareto chart, a set of top three factors influencing the erythritol yield (Y

_{1}, g·L

^{−1}) was screened out (Table 1). These three variables were taken forward for optimization using response surface methodology (RSM).

#### 2.6. Response Surface Methodology

_{1}g·L

^{−1}), yeast extract (X

_{2}g·L

^{−1}), and KH

_{2}PO

_{4}(X

_{3}g·L

^{−1}). The domain of the variables was selected from the literary information. The RCCD matrix resulted in 20 experimental runs with 2

^{3}factorial runs (±1 level in coded values), 2 × 3 axial runs (±α level in coded values), and 6 repeated runs at the center point (0 levels in coded values) (Table 2).

_{1}g·L

^{−1}) in the medium after fermentation. A quadratic polynomial model was developed for erythritol yield (Y

_{1}g·L

^{−1}) as a function of three independent variables, viz., x

_{1}, x

_{2}, and x

_{3}, as described in Equation (3).

_{0}to β

_{9}are the regression coefficients generated from minimizing the sum of the square of errors using Equation (4).

_{10×1}is the matrix consisting of β

_{0}to β

_{9}; [D]

_{20×10}is the matrix composed of 10 parameters from Equation (3) and the corresponding values for 20 experimental runs. The matrix [y

_{a}]

_{20×1}is the actual set of response values for 20 experimental runs. The adequacy of model fitting was checked by analysis of variance (ANOVA) data. The R

^{2}value was used to know the overall predictive capability of the model. The statistical significance of the fit of the polynomial model equation was checked by the variance test (F-test) with a confidence interval of 95% of the mean. The significance of the regression coefficient was tested by p-value (probability of accepting null hypothesis). The response surface model and contour plots were developed for the erythritol yield as a function of the two variables between x

_{1}, x

_{2}, and x

_{3}.

#### 2.7. Numerical Optimization

_{1}, x

_{2}, and x

_{3}targeting a maximum erythritol yield (Y

_{1}g·L

^{−1}). It was targeted to maximize the value of the overall desirability function (d

_{1}) according to Equation (5).

_{1}is the desirability index for the response (erythritol yield, Y

_{1}); the value of d

_{1}ranges between 0 and 1, and a magnitude nearer to 1 represents the most desirable ones. The weightage (w) represents the nature of the curve followed by d

_{1}between 0 to 1. It is taken as a linear pathway (w = 1). L

_{1}and U

_{1}are the lower and upper limits of the Y

_{1}.

#### 2.8. Analysis of Fermentation Broth

^{−1}with a total run time of 6 min. The quantification was performed by a standard external technique using the peak area of reference compounds. The standard erythritol of 1 µg·mL

^{−1}solution showed a 1,347,578 nm

^{2}area under the curve in LC-MS.

#### 2.9. Purification of Fermentation Broth

^{−}

^{1}. The concentrated solution was then allowed to cool to 20 °C under gentle agitation and seeded with a trace amount of erythritol to initiate crystallization. The solution was then incubated at 4 °C overnight. Brittle white erythritol crystals formed were collected by filtration, washed twice with cold distilled water, and dried at 50 °C for two hours in a hot air oven. Further, the purified erythritol crystals were characterized by LC-MS by comparison with the mass spectra of standard or reference erythritol. The injection volume of the molasses sample was 10 µL comprising 5% broth, and it was compared with pure erythritol solution (1 g/L). The concentration of erythritol present in the molasses was calculated by comparing the ratio between the area under the curve of standard erythritol and molasses broth, respectively.

#### 2.10. Statistical Analysis

## 3. Results and Discussion

#### 3.1. Growth Curve of Candida magnoliae

_{600nm}) and the dry cell weight (DCW, mg) (Equation (6)). An exponential relationship connected the cell count (N cfu/mL) to OD by an exponential relationship (Equation (7)).

^{2}for Equations (6) and (7) were 0.92 and 0.99, respectively. This indicated the adequacy of model fitting.

#### 3.2. Screening of Significant Factors Using Plackett–Burman Design

_{1}) varied between 22.5 between 125.9 g·L

^{−1}. The DCW ranged from 170 to 930 mg, whereas ΔpH varied between 3.0 and 5.3. A set of linear equations has evolved to describe the changes in three responses as a function of independent variables (x

_{1}to x

_{11}). The equation for the responses in terms of coded value is presented in Equation (8).

_{1}–x

_{11}represent the coded values (−1 to +1) for the independent variables X

_{1}–X

_{2}in Table 1. Here, −1 corresponds to the lower limit and +1 corresponds to the upper limit of the domain. The adjusted R

^{2}for the linear equations obtained for Y

_{1}, Y

_{2}, and Y

_{3}were 0.85, 0.75, and 0.91, respectively; the respective F-values were 15.54, 7.72, and 28.4. This reflects that the changes in the response values are not due to noise; rather, these are influenced by the factors. The significant factors (p-value < 0.05) were screened out for each response. Interestingly, for each equation, only three independent variables were significant. For instance, x

_{1}, x

_{2}, and x

_{3}influenced Y

_{1}significantly; Y

_{2}was affected most by x

_{1}, x

_{3}, and x

_{10}; and the contribution of x

_{2}, x

_{5}, and x

_{6}was greatest for Y

_{3}. The coefficient in the coded form equation (Equation (8)) shows that molasses (x

_{1}) was the most significant factor influencing the erythritol formation (Y

_{1}) positively. The positive influence signifies that with an increase in molasses concentration (factor), the erythritol yield (response) is improved or increased. In addition, increasing the concentration of KH

_{2}PO

_{4}(x

_{3}) increased the Y

_{1}significantly, whereas yeast extract (x

_{2}) negatively influenced erythritol production. On the contrary, it was found that molasses concentration (x

_{1}) negatively contributed to dry cell weight (Y

_{2}), whereas the more KH

_{2}PO

_{4}(x

_{3}) and media volume (x

_{10}) were used, the greater the dry cell weight obtained. It was also found that yeast extract (x

_{2}) and inoculum volume (x

_{6}) significantly influenced the change in pH as per Equation (8). Therefore, the concentrations of molasses (x

_{1}), yeast extract (x

_{2}), and KH

_{2}PO

_{4}(x

_{3}) were taken forward in the next step for the optimization of erythritol yield (Y

_{1}).

#### 3.3. Response Surface Modeling of Erythritol Yield

_{1}), nitrogen (yeast extract, x

_{2}), and phosphate (KH

_{2}PO

_{4}, x

_{3}) were the critical medium components for erythritol production (Y

_{1}). These three factors significantly influenced erythritol yield (Y

_{1}) (Table 2). The yield ranged between 28.2 and 105.6 g·L

^{−}

^{1}. Increasing molasses (carbon source) led to a higher product yield. For instance, at 9 g·L

^{−}

^{1}yeast extract and 2 g·L

^{−}

^{1}KH

_{2}PO

_{4}, the erythritol yield increased from 35.1 to 59.2 g·L

^{−}

^{1}when the molasses concentration was increased from 200 to 300 g·L

^{−}

^{1}(Table 2). Molasses contains different carbon and nitrogen sources such as sucrose, thiamine, and so on, which are easily assimilated, supporting the cell growth of osmotolerant microbes. The components of molasses vary greatly. Molasses is complex, and it contains mainly sucrose besides other components. The proximate composition of the molasses used in this study is summarized in Table 3. The molasses’ reducing and non-reducing sugar contents are 181 and 352 g·L

^{−1}, respectively. The cumulative concentration of nitrogenous compounds (in terms of crude protein) was 3.5 g·L

^{−1}, whereas fat content was 4.2 g·L

^{−1}. The molasses was black and contained 22.5% moisture with 11.5% ash.

^{−1}yeast extract and 5 g·L

^{−1}KH

_{2}PO

_{4}, the yield reduced from 96.6 to 89.4 g·L

^{−1}when the molasses concentration was increased from 200 to 300 g·L

^{−1}(Table 2). The osmotic stress might be responsible for this trend in erythritol yield [26]. The osmotic pressure of the medium increased from 0.988 to1.482 kPa when the molasses concentration changed from 200 to 300 g·L

^{−1}. The dry cell weight of the biomass recovered from 1 L medium decreased from 20.23 to 18.36 g. The reduction in cell biomass might be due to the unavailability of soluble oxygen in the medium. In addition, the C:N ratio might play a role in diverting the balance between biomass production and polyol formation [29].

^{−1}molasses and 2 g·L

^{−1}KH

_{2}PO

_{4}mixed in the fermentation medium, the erythritol yield increased from 35.1 to 63.1 g·L

^{−1}when the yeast extract concentration increased from 9 to 12 g·L

^{−1}, respectively. At 250 g·L

^{−1}molasses and 3.5 g·L

^{−1}KH

_{2}PO

_{4,}the yield was elevated from 57.8 to 65.8 g·L

^{−1}for 7.98 to 13.02 g·L

^{−1}yeast extract, respectively. Yeast extract is commonly used as the nitrogen source for the fermentative production of sugars such as erythritol. It is a source of thiamine required to produce erythritol [7]. The yeast powder contained (w/w) protein (30%), fat (0.42%), sodium chloride (0.67%), ash (12.18%), and total volatile nitrogen (9.2%) with a moisture of 4.72% and pH of 6.29. Molasses contained a certain amount of nitrogen compounds (3.5 g·L

^{−1}), as indicated in Table 3. Both yeast extract and molasses provide organic nitrogen; thus, an increase in either component led to a higher erythritol yield in the medium [17].

_{2}PO

_{4}concentration in the medium significantly enhanced the erythritol yield. Keeping the molasses (200 g·L

^{−1}) and yeast extract (9 g·L

^{−1}) fixed, the product yield showed an elevation from 35.1 to 96.6 g·L

^{−1}when the phosphate concentration increased from 2 to 5 g·L

^{−1}. However, after reaching an optimum, a reverse trend was found. For instance, Y

_{1}was 80.8 and 78.6 g·L

^{−1}when the KH

_{2}PO

_{4}concentration increased from 0.98 to 6.02 g·L

^{−1}at 250 g·L

^{−1}molasses and 10.5 g·L

^{−1}yeast extract. For the growth of Candida species, the role of inorganic phosphate is indispensable. In addition, a higher concentration of KH2PO4 led to a higher osmotic stress surrounding the Candida magnoliae. This eventually might promote the cellular erythritol production so that osmotic stress inside the cell is balanced [10,30].

_{1}) as a function of molasses (x

_{1}), yeast extract (x

_{2}), and KH

_{2}PO

_{4}concentration (x

_{3}). The F-value for the quadratic model was 10.62 compared to the same 1.67 and 0.67 for the linear and factorial interaction model, respectively. The cubic model showed an F-value of 0.59. Therefore, the quadratic polynomial model was taken forward to visualize the square and interaction effect between x

_{1}, x

_{2}, and x

_{3}on the response. The corresponding lack of fit p-value was 0.4539 (insignificant). Lack of fit signifies that the influence of noise variables on the response. Statistically, an insignificant lack of fit is expected for a model to be fit. Additionally, an insignificant lack of fit refers that any change in response is well connected to the variation in the controllable factors such as x

_{1}, x

_{2}, and x

_{3}(Table 4).

^{2}= 0.92) and adjusted R

^{2}(0.84) depicts the model adequacy. The predicted response values are also reflected in Table 3. The percentage error between the actual and predicted values is below 10%, except for the outlier at one condition (run: 5, showing an error of 12.7%). This signifies that except that outlier, the deviation in the responses from the fitted line is within 10%. A low value of the coefficient of variation of 12.40% indicates a very high degree of precision and good reliability of the experimental values. The coefficient of variation dictates the relative size of the standard deviation in comparison to the fitted value including all the outliers. A coefficient of variation of 12.40% is well accepted from a statistical point of view. The adequacy precision, which measures the signal-to-noise ratio, shows a value of 9.9, whereas it is desirable to have this value > 4. This quadratic polynomial model navigates the design space when presented graphically. The points accommodate themselves along the diagonal, indicating the high level of statistical significance of the model.

_{1}, x

_{2}, and x

_{3}in the polynomial equation positively influence the response, which implies that on increasing the concentration of either of these three components, the yield of erythritol also increases significantly. The major contributor to increasing the erythritol yield is molasses concentration, followed by KH

_{2}PO

_{4}and yeast extract. The interaction between molasses and yeast extract concentration (x

_{1}x

_{2}) is significant at p < 0.05, whereas other interaction terms, such as x

_{1}x

_{3}and x

_{2}x

_{3}, do not statistically (p > 0.1) influence the response (Y

_{1}). All of the three-square terms (x

_{1}x

_{1}, x

_{2}x

_{2}, and x

_{3}x

_{3}) showed significant (p < 0.05) negative coefficients in the polynomial model. The negative square terms support the parabolic nature of the curve, corroborating the notion that, initially, the yield will increase with an increase in either x

_{1}, x

_{2}, or x

_{3}. However, the yield will be compromised beyond a particular concentration, and the shape of the curve will be reversed. Figure 1 depicts the changes in erythritol yield with respect to a simultaneous change in molasses and yeast extract concentration. This contour plots are drawn at a fixed phosphate concertation of 3.5 g/L. The contour plot showing the interaction between x

_{1}and x

_{2}on the response follows this parabolic trend. When the concentration of molasses and yeast extract was 200 and 9 g·L

^{−1}, the erythritol yield was 55.6 g·L

^{−1}, whereas when the concentration was increased to 250 and 11 g·L

^{−1}, the erythritol yield increased to 90 g·L

^{−1}. However, the yield of erythritol declined to 81.9 g·L

^{−1}when the concentration was further increased to 280 and 11.3 g·L

^{−1}. A similar influence of glucose concentration on erythritol yield was observed by Park et al. [31]. In another study, it was seen that glycerol started to appear in the broth as a by-product instead of erythritol production. This decline can be attributed to an increase in the osmolarity of the fermentation media, which might reduce the oxygen transfer rate towards the cells, which is crucial for erythritol formation [26].

#### 3.4. Numerical Optimization of Fermentation Conditions

_{1}, x

_{2}, and x

_{3}. The desirability value (D) is connected to Y, which is again the function of the independent variables. The desirability value (D) was maximized, leading to a maximum erythritol yield. The relative importance of the response was 5 out of 5. The maximum erythritol yield predicted was 99.54 g·L

^{−1}under the optimized condition of molasses, yeast extract, and KH

_{2}PO

_{4}concentrations of 273.96, 10.25, and 3.28 g·L

^{−1}, respectively. A high desirability value of 0.88 supported this optimized condition. The actual experimental yield of erythritol obtained was 98.89 g·L

^{−1}. The error during validation was less than 1%, corroborating the model fitting accuracy. Similar results of erythritol yield have been reported in the literature. For instance, Hijosa-Valsera et al. [32] reported an erythritol yield of 106.4 g·L

^{−1}from 300 g·L

^{−1}molasses using the yeast Moniliella pollinis. In another study, a three-level factorial design with glycerol, urea, and NaCl as the three most contributing factors for erythritol yield was used. The predicted erythritol yield was 100.6 g·L

^{−1}using 214.5 g·L

^{−1}glycerol with 1.69 g·L

^{−1}urea and 37.2 g·L

^{−1}NaCl [33]. Savergave et al. [26] used a four-level factorial design for the optimization of erythritol, which included glucose 238 g·L

^{−1}, yeast extract 9.2 g·L

^{−1}, KH

_{2}PO

_{4}5.16 g·L

^{−1}, and MgSO

_{4}0.23 g·L

^{−1}as factors. They predicted the optimized erythritol yield of 87.8 g·L

^{−1}. At the optimized condition, the erythritol yield was 0.625 g/g of dry cell weight. Deshpande et al. [34] obtained an erythritol yield of 0.38 g/g of dry cell weight when molasses was used as a feedstock by Moniliella pollinis. Rakicka et al. [35] employed a two-stage chemostat process with glycerol and reported an erythritol yield of 0.66 g/g of dry cell weight.

#### 3.5. Purification and Characterization of Erythritol

^{2}, respectively (Figure 2).

_{3}

^{−}is 61.987, while the m/z of [erythritol + NO

_{3}]

^{−}is 184.046. The additional peaks in the mass spectra denote the presence of minor compounds associated with the erythritol crystals involved in the purification and crystallization steps.

^{−1}was optimized using response surface methodology using 273.96 g·L

^{−1}molasses, 10.25 g·L

^{−1}yeast extract, and 3.28 g·L

^{−1}KH

_{2}PO

_{4}in the medium, which was comparable to or higher than the erythritol yield obtained using glucose fermentation by Candida magnoliae. Moreover, the study uses molasses, a by-product of the sugarcane industry, for the production of erythritol, bringing the process a step closer toward attaining sustainability as opposed to using commonly used substrates such as pure glucose/fructose, which increases the production cost. The erythritol yield was estimated by erythrose reductase enzyme assay, and the results obtained were validated using LC-MS. The erythritol crystals purified from the fermentation broth showed 94% purity. Further study should focus on identifying the crystal structure of the obtained erythritol. Improving the yield of erythritol by gene modification through site-specific mutagenesis of Candida magnoliae can be studied further. The influence of flocculants and the yeast grown in various molasses on the erythritol yield may be explored. Besides, scaling-up studies of fermentation media inside a bioreactor would be of great interest.

## Author Contributions

## Funding

## Institutional Review Board Statement

## Informed Consent Statement

## Data Availability Statement

## Conflicts of Interest

## References

- Lee, K.J.; Lim, J.Y. Optimized conditions for high erythritol production by Penicillium sp. KJ-UV29, mutant of Penicillium sp. KJ81. Biotechnol. Bioprocess Eng.
**2003**, 8, 173–178. [Google Scholar] [CrossRef] - Martău, G.A.; Coman, V.; Vodnar, D.C. Recent advances in the biotechnological production of erythritol and mannitol. Crit. Rev. Biotechnol.
**2020**, 40, 608–622. [Google Scholar] [CrossRef] [PubMed] - Grembecka, M.; Lebiedzińska, A.; Szefer, P. Simultaneous separation and determination of erythritol, xylitol, sorbitol, mannitol, maltitol, fructose, glucose, sucrose, and maltose in food products by high performance liquid chromatography coupled to charged aerosol detector. Microchem. J.
**2014**, 117, 77–82. [Google Scholar] [CrossRef] - Ahuja, K.; Rawat, A. Erythritol Market Trends 2020–2026: Global Industry Report; Global Market Insights Inc.: Selbyville, DE, USA, 2020; Available online: www.gminsights.com/industry-analysis/erythritol-market (accessed on 16 March 2020).
- Daza-Serna, L.; Serna-Loaiza, S.; Masi, A.; Mach, R.L.; Mach-Aigner, A.R.; Friedl, A. From the culture broth to the erythritol crystals: An opportunity for circular economy. Appl. Microbiol. Biotechnol.
**2021**, 105, 4467–4486. [Google Scholar] [CrossRef] - Tomaszewska-Hetman, L.; Rymowicz, W.; Rywińska, A. Waste conversion into a sweetener—development of an innovative strategy for erythritol production by Yarrowia lipolytica. Sustainability
**2020**, 12, 7122. [Google Scholar] [CrossRef] - Tomaszewska, L.; Rymowicz, W.; Rywińska, A. Mineral supplementation increases erythrose reductase activity in erythritol biosynthesis from glycerol by Yarrowia lipolytica. Appl. Biochem. Biotechnol.
**2014**, 172, 3069–3078. [Google Scholar] [CrossRef] - Tomaszewska, L.; Rywińska, A.; Rymowicz, W. High selectivity of erythritol production from glycerol by Yarrowia lipolytica. Biomass Bioenergy
**2014**, 64, 309–320. [Google Scholar] [CrossRef] - Yang, L.B.; Zhan, X.B.; Zheng, Z.Y.; Wu, J.R.; Gao, M.J.; Lin, C.C. A novel osmotic pressure control fed-batch fermentation strategy for improvement of erythritol production by Yarrowia lipolytica from glycerol. Bioresour. Technol.
**2014**, 151, 120–127. [Google Scholar] [CrossRef] - Yu, J.H.; Lee, D.H.; Oh, Y.J.; Han, K.C.; Ryu, Y.W.; Seo, J.H. Selective utilization of fructose to glucose by Candida magnoliae, an erythritol producer. Appl. Biochem. Biotechnol.
**2006**, 131, 870–879. [Google Scholar] [CrossRef] - Guo, J.; Li, J.; Chen, Y.; Guo, X.; Xiao, D. Improving erythritol production of Aureobasidium pullulans from xylose by mutagenesis and medium optimization. Appl. Biochem. Biotechnol.
**2016**, 180, 717–727. [Google Scholar] [CrossRef] - Mirończuk, A.M.; Rakicka, M.; Biegalska, A.; Rymowicz, W.; Dobrowolski, A. A two-stage fermentation process of erythritol production by yeast Y. lipolytica from molasses and glycerol. Bioresour. Technol.
**2015**, 198, 445–455. [Google Scholar] [CrossRef] [PubMed] - Rzechonek, D.A.; Dobrowolski, A.; Rymowicz, W.; Mirończuk, A.M. Recent advances in biological production of erythritol. Crit. Rev. Biotechnol.
**2018**, 38, 620–633. [Google Scholar] [CrossRef] [PubMed] - Zhuang, X.; Wang, Y.; Li, Q.; Yan, S.; He, N. The production of bioflocculants by Bacillus licheniformis using molasses and its application in the sugarcane industry. Biotechnol. Bioprocess Eng.
**2012**, 17, 1041–1047. [Google Scholar] [CrossRef] - Clarke, M.A. Syrups. In Encyclopedia of Food Sciences and Nutrition, 2nd ed.; Academic Press: Cambridge, MA, USA, 2003; pp. 5711–5717. [Google Scholar] [CrossRef]
- Vignesh Kumar, B.; Muthumari, B.; Kavitha, M.; John Praveen Kumar, J.K.; Thavamurugan, S.; Arun, A.; Jothi Basu, M. Studies on optimization of sustainable Lactic Acid production by Bacillus amyloliquefaciens from sugarcane molasses through microbial fermentation. Sustainability
**2022**, 14, 7400. [Google Scholar] [CrossRef] - Carly, F.; Fickers, P. Erythritol production by yeasts: A snapshot of current knowledge. Yeast
**2018**, 35, 455–463. [Google Scholar] [CrossRef] - Savergave, L.S. Microbial Production of Erythritol and Mannitol: Strain Improvement and Process Optimization. Ph.D. Thesis, National Chemical Laboratory, Pune, India, 2011. [Google Scholar]
- Ghezelbash, G.R.; Nahvi, I.; Malekpour, A. Erythritol production with minimum By-product using Candida magnoliae mutant. Appl. Biochem. Microbiol.
**2014**, 50, 292–296. [Google Scholar] [CrossRef] - Gassara, F.; Brar, S.K.; Tyagi, R.D.; John, R.P.; Verma, M.; Valero, J.R. Parameter optimization for production of ligninolytic enzymes using agro-industrial wastes by response surface method. Biotechnol. Bioprocess Eng.
**2011**, 16, 343–351. [Google Scholar] [CrossRef] - Borkar, V.; Chakraborty, S.; Gokhale, J.S. Fermentative production of naringinase from Aspergillus niger van Tieghem MTCC 2425 using citrus wastes: Process optimization, partial purification, and characterization. Appl. Biochem. Biotechnol.
**2021**, 193, 1321–1337. [Google Scholar] [CrossRef] - Chen, L.; Wei, X.; Chaves, B.D.; Jones, D.; Ponder, M.A.; Subbiah, J. Inactivation of Salmonella enterica and Enterococcus faecium NRRL B2354 on cumin seeds using gaseous ethylene oxide. Food Microbiol.
**2021**, 94, 103656. [Google Scholar] [CrossRef] - Yan, Y.; Zhao, X.; Wang, C.; Fang, Q.; Zhong, L.; Wei, Q. Preparation, optimization, and characterization of inclusion complexes of Cinnamomum longepaniculatum essential oil in β-Cyclodextrin. Sustainability
**2022**, 14, 9513. [Google Scholar] [CrossRef] - Ryu, Y.W.; Park, C.Y.; Park, J.B.; Kim, S.Y.; Seo, J.H. Optimization of erythritol production by Candida magnoliae in fed-batch culture. J. Ind. Microbiol. Biotechnol.
**2000**, 25, 100–103. [Google Scholar] [CrossRef] - Koh, E.S.; Lee, T.H.; Lee, D.Y. Scale-up of erythritol production by an osmophilic mutant of Candida magnoliae. Biotechnol. Lett.
**2003**, 25, 2103–2105. [Google Scholar] [CrossRef] [PubMed] - Savergave, L.S.; Gadre, R.V.; Vaidya, B.K.; Narayanan, K. Strain improvement and statistical media optimization for enhanced erythritol production with minimal by-products from Candida magnoliae mutant R23. Biochem. Eng. J.
**2011**, 55, 92–100. [Google Scholar] [CrossRef] - Krohn, R.I. The colorimetric detection and quantitation of total protein. Curr. Protoc. Toxicol.
**2005**, 23, A-3I. [Google Scholar] [CrossRef] - Kim, K.A.; Noh, B.S.; Lee, J.K.; Kim, S.Y.; Park, Y.C.; Oh, D.K. Optimization of culture conditions for erythritol production by Torula sp. J. Microbiol. Biotechnol.
**2000**, 10, 69–74. [Google Scholar] - Chang, C.S.; Yeh, T.S. Detection of 10 sweeteners in various foods by liquid chromatography/tandem mass spectrometry. J. Food Drug Anal.
**2014**, 22, 318–328. [Google Scholar] [CrossRef] - Hijosa-Valsero, M.; Garita-Cambronero, J.; Paniagua-García, A.; Diez Antolinez, R. By-products of sugar factories and wineries as feedstocks for erythritol generation. Food Bioprod. Process.
**2021**, 126, 345–355. [Google Scholar] [CrossRef] - Park, J.; Seo, B.; Kim, J.; Park, Y. Production of erythritol in fed-batch cultures of Trichosporon sp. J. Ferment. Bioeng.
**1998**, 86, 577–580. [Google Scholar] [CrossRef] - Deshpande, M.S.; Kulkarni, P.P.; Kumbhar, P.S.; Ghosalkar, A.R. Erythritol production from sugar based feedstocks by Moniliella pollinis using lysate of recycled cells as nutrients source. Process Biochem.
**2022**, 112, 45–52. [Google Scholar] [CrossRef] - Rakicka, M.; Biegalska, A.; Rymowicz, W.; Dobrowolski, A.; Mirończuk, A.M. Polyol production from waste materials by genetically modified Yarrowia lipolytica. Bioresour. Technol.
**2017**, 243, 393–399. [Google Scholar] [CrossRef] - Zhang, S.; Wang, J.; Jiang, H. Microbial production of value-added bioproducts and enzymes from molasses, a by-product of sugar industry. Food Chem.
**2021**, 346, 128860. [Google Scholar] [CrossRef] [PubMed] - Adeuya, A.; Price, N. Rapid characterization of polyalcohols by silylation and matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Rapid Commun. Mass Spectrom. Int. J. Devoted Rapid Dissem. Up-Minute Res. Mass Spectrom.
**2007**, 21, 3977–3981. [Google Scholar] [CrossRef] [PubMed] - Yang, L.B.; Zhan, X.B.; Zhu, L.; Gao, M.J.; Lin, C.C. Optimization of a low-cost hyperosmotic medium and establishing the fermentation kinetics of erythritol production by Yarrowia lipolytica from crude glycerol. Prep. Biochem. Biotechnol.
**2016**, 46, 376–383. [Google Scholar] [CrossRef] [PubMed] - Yoon, J.W.; Jeon, E.J.; Jung, I.H.; Min, M.J.; Lee, H.Y.; Kim, M.J.; Moon, T.W. Maltosyl-erythritol, a major transglycosylation product of erythritol by Bacillus stearothermophilus maltogenic amylase. Biosci. Biotechnol. Biochem.
**2003**, 67, 525–531. [Google Scholar] [CrossRef] - Rakicka, M.; Rywińska, A.; Lazar, Z.; Rymowicz, W. Two-stage continuous culture–technology boosting erythritol production. J. Clean. Prod.
**2017**, 168, 420–427. [Google Scholar] [CrossRef] - Forbes, T.P.; Sisco, E. Trace detection and competitive ionization of erythritol tetranitrate in mixtures using direct analysis in real time mass spectrometry. Anal. Methods
**2015**, 7, 3632–3636. [Google Scholar] [CrossRef] - AOAC International. Official Methods of Analysis, 15th ed.; AOAC International: Gaithersburg, MD, USA, 1990; Volume 1. [Google Scholar]
- Daza-Serna, L.; Knežević, K.; Kreuzinger, N.; Mach-Aigner, A.R.; Mach, R.L.; Krampe, J.; Friedl, A. Recovery of salts from synthetic erythritol culture broth via electrodialysis: An alternative strategy from the bin to the loop. Sustainability
**2022**, 14, 734. [Google Scholar] [CrossRef]

**Figure 1.**The contour plots of erythritol yield in molasses and yeast extract concentration landscape at fixed phosphate concertation of 3.5 g/L.

**Figure 2.**LC-MS spectra of (

**a**) 1 mg/L purified erythritol and (

**b**) 1 mg/L standard erythritol crystals.

**Figure 3.**LC-MS overlay chromatogram with (

**a**) 1 ppm standard erythritol and (

**b**) crystalline erythritol obtained after fermentation.

**Figure 4.**Mass spectra of (

**a**) 1 ppm standard erythritol and (

**b**) crystalline erythritol obtained after fermentation.

**Table 1.**The experimental run of the Plackett–Burman design to screen out the most influential variables affecting the responses.

Run | Independent Variables (Coded Value) | Response | ||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|

X_{1} | X_{2} | X_{3} | X_{4} | X_{5} | X_{6} | X_{7} | X_{8} | X_{9} | X_{10} | X_{11} | Y_{1} | Y_{2} | Y_{3} | |

g·L^{−1} | g·L^{−1} | g·L^{−1} | g·L^{−1} | - | mL | °C | h | rpm | mL | mg·L^{−1} | g·L^{−1} | mg | - | |

1 | 300 | 12 | 2 | 0.5 | 7 | 3 | 25 | 48 | 180 | 30 | 80 | 37.2 ± 1.8 | 350 ± 3 | 4.6 ± 0.0 |

2 | 200 | 12 | 5 | 0.1 | 7 | 3 | 30 | 48 | 180 | 20 | 120 | 32.6 ± 1.5 | 451 ± 4 | 4.3 ± 0.1 |

3 | 300 | 9 | 5 | 0.5 | 4 | 3 | 30 | 96 | 180 | 20 | 80 | 105.7 ± 4.3 | 482 ± 4 | 3.0 ± 0.1 |

4 | 200 | 12 | 2 | 0.5 | 7 | 1 | 30 | 96 | 240 | 20 | 80 | 22.5 ± 1.5 | 733 ± 5 | 5.2 ± 0.0 |

5 | 200 | 9 | 5 | 0.1 | 7 | 3 | 25 | 96 | 240 | 30 | 80 | 35.1 ± 1.6 | 934 ± 8 | 4.2 ± 0.0 |

6 | 200 | 9 | 2 | 0.5 | 4 | 3 | 30 | 48 | 240 | 30 | 120 | 36.5 ± 1.7 | 592 ± 4 | 3.1 ± 0.0 |

7 | 300 | 9 | 2 | 0.1 | 7 | 1 | 30 | 96 | 180 | 30 | 120 | 96.6 ± 4.9 | 174 ± 2 | 4.6 ± 0.1 |

8 | 300 | 12 | 2 | 0.1 | 4 | 3 | 25 | 96 | 240 | 20 | 120 | 36.1 ± 1.8 | 664 ± 3 | 3.7 ± 0.1 |

9 | 300 | 12 | 5 | 0.1 | 4 | 1 | 30 | 48 | 240 | 30 | 80 | 94.5 ± 4.7 | 613 ± 3 | 3.9 ± 0.1 |

10 | 200 | 12 | 5 | 0.5 | 4 | 1 | 25 | 96 | 180 | 30 | 120 | 23.8 ± 1.6 | 742 ± 4 | 3.4 ± 0.0 |

11 | 300 | 9 | 5 | 0.5 | 7 | 1 | 25 | 48 | 240 | 20 | 120 | 125.9 ± 4.9 | 585 ± 3 | 4.7 ± 0.1 |

12 | 200 | 9 | 2 | 0.1 | 4 | 1 | 25 | 48 | 180 | 20 | 80 | 33.5 ± 1.5 | 464 ± 3 | 3.0 ± 0.0 |

t-Stat (Y1) | 52.02 | −31.13 | 25.88 | 3.89 | 3.31 | −18.94 | 16.08 | −6.75 | 3.5 | −5.45 | 3.83 | |||

%Cont (Y1) | 53.15 | 19.04 | 13.16 | 0.30 | 0.22 | 7.04 | 5.08 | 0.89 | 0.24 | 0.58 | 0.29 | |||

Rank (Y1) | I | II | III | VIII | XI | IV | V | VI | X | VII | IX |

_{1}= molasses concentration, X

_{2}= yeast extract concentration, X

_{3}= KH

_{2}PO

_{4}concentration, X

_{4}= MgSO

_{4}concentration, X

_{5}= pH, X

_{6}= inoculum volume, X

_{7}= temperature, X

_{8}= time, X

_{9}= agitation, X

_{10}= media volume, X

_{11}= CaCO

_{3}, respectively. Y1, Y2, Y3 are erythritol yield, DCW, and ΔpH, respectively. %cont is the percentage contribution for Y1.

**Table 2.**Rotatable central composite design, corresponding erythritol yield, and predicted results for the polynomial model.

Run | Independent Variable (Coded Value) | Erythritol Yield | Error | |||
---|---|---|---|---|---|---|

Molasses | Yeast Extract | KH_{2}PO_{4} | Actual | Predicted | ||

g·L^{−1} | g·L^{−1} | g·L^{−1} | g·L^{−1} | g·L^{−1} | % | |

1 | 200 (−1) | 9 (−1) | 2 (−1) | 35.1 ± 1.8 | 33.5 | +4.5 |

2 | 200 (−1) | 9 (−1) | 5 (+1) | 96.6 ± 3.9 | 88.1 | +8.7 |

3 | 200 (−1) | 12 (+1) | 2 (−1) | 63.1 ± 2.3 | 60.0 | +4.9 |

4 | 200 (−1) | 12 (+1) | 5 (+1) | 72.5 ± 2.7 | 76.1 | −4.9 |

5 | 300 (+1) | 9 (−1) | 2 (−1) | 59.2 ± 2.5 | 51.6 | +12.7 |

6 | 300 (+1) | 9 (−1) | 5 (+1) | 89.4 ± 3.5 | 86.0 | +3.8 |

7 | 300 (+1) | 12 (+1) | 2 (−1) | 62.3 ± 2.4 | 64.3 | −3.2 |

8 | 300 (+1) | 12 (+1) | 5 (+1) | 65.2 ± 2.9 | 60.2 | +7.6 |

9 | 166 (−1.68) | 10.5 (0) | 3.5 (0) | 28.2 ± 1.7 | 30.9 | −9.6 |

10 | 334 (+1.68) | 10.5 (0) | 3.5 (0) | 68.7 ± 3.1 | 73.4 | −6.8 |

11 | 250 (0) | 7.98 (−1.68) | 3.5 (0) | 57.8 ± 3.2 | 63.5 | −9.8 |

12 | 250 (0) | 13.02 (+1.68) | 3.5 (0) | 65.8 ± 3.6 | 64.1 | +2.6 |

13 | 250 (0) | 10.5 (0) | 0.98 (−1.68) | 80.8 ± 3.3 | 83.3 | −3.1 |

14 | 250 (0) | 10.5 (0) | 6.02 (+1.68) | 78.6 ± 3.1 | 85.2 | −8.4 |

15 | 250 (0) | 10.5 (0) | 3.5 (0) | 89.4 ± 3.6 | 96.6 | −8.0 |

16 | 250 (0) | 10.5 (0) | 3.5 (0) | 105.6 ± 4.6 | 96.6 | +8.5 |

17 | 250 (0) | 10.5 (0) | 3.5 (0) | 104.2 ± 4.5 | 96.6 | +7.3 |

18 | 250 (0) | 10.5 (0) | 3.5 (0) | 89.4 ± 4.8 | 96.6 | −8.0 |

19 | 250 (0) | 10.5 (0) | 3.5 (0) | 97.1 ± 5.0 | 96.6 | +0.5 |

20 | 250 (0) | 10.5 (0) | 3.5 (0) | 95.0 ± 4.9 | 96.6 | −1.6 |

Constituents (Unit) | Quantity |
---|---|

Moisture (g/100 g) | 22.5 ± 4.3 |

Total sugar (g/100 g) | 53.3 ± 5.6 |

Reducing sugar (g/100 g) | 18.1 ± 2.5 |

Non-reducing sugar (g/100 g) | 35.2 ± 3.2 |

Crude protein (g/100 g) | 0.35 ± 0.1 |

Crude fat (g/100 g) | 0.42 ± 0.1 |

Ash (g/100 g) | 11.5 ± 4.0 |

Brix value (°Bx) | 78.1 ± 0.2 |

pH (-) | 5.9 ± 0.1 |

**Table 4.**The analysis of variance (ANOVA) data for the response surface model developed for erythritol yield (Y

_{1}g·L

^{−1}).

Parameter | Coefficient ± 95% CI | p-Value |
---|---|---|

Constant | 96.58 ± 8.41 | 0.0005 |

x_{1} | 12.63 ± 5.61 | 0.0005 |

x_{2} | 0.18 ± 2.52 | 0.9433 |

x_{3} | 0.55 ± 2.51 | 0.8301 |

x_{1}x_{2} | −9.61 ± 3.29 | 0.0153 |

x_{1}x_{3} | −5.05 ± 3.66 | 0.1004 |

x_{2}x_{3} | −3.44 ± 3.28 | 0.3200 |

x_{1}x_{1} | −15.68 ± 5.47 | <0.0001 |

x_{2}x_{2} | −11.57 ± 5.46 | 0.0008 |

x_{3}x_{3} | −4.34 ± 2.05 | 0.0093 |

F-value | 10.52 | - |

p-value (model) | - | 0.0002 |

p-value (lack of fit) | - | 0.4539 |

R^{2} | 0.92 | - |

Adj R^{2} | 0.84 | - |

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

© 2022 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 (https://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Seshadrinathan, S.; Chakraborty, S.
Fermentative Production of Erythritol from Cane Molasses Using *Candida magnoliae*: Media Optimization, Purification, and Characterization. *Sustainability* **2022**, *14*, 10342.
https://doi.org/10.3390/su141610342

**AMA Style**

Seshadrinathan S, Chakraborty S.
Fermentative Production of Erythritol from Cane Molasses Using *Candida magnoliae*: Media Optimization, Purification, and Characterization. *Sustainability*. 2022; 14(16):10342.
https://doi.org/10.3390/su141610342

**Chicago/Turabian Style**

Seshadrinathan, Shruthy, and Snehasis Chakraborty.
2022. "Fermentative Production of Erythritol from Cane Molasses Using *Candida magnoliae*: Media Optimization, Purification, and Characterization" *Sustainability* 14, no. 16: 10342.
https://doi.org/10.3390/su141610342