Bioconversion of a Glycerol- and Methanol-Rich Residue from Biodiesel Industry into 1,3-Propanediol: The Role of Magnesium

Abstract

Biodiesel is one of the most important biofuels worldwide. Besides glycerol, the residual aqueous phase of the transesterification reaction (RAPTR) from the biodiesel industry contains a high concentration of methanol. Here, we propose using RAPTR as substrate for Clostridium beijerinckii Br21 to produce 1,3-propanediol (1,3-PDO). 1,3-PDO is a valuable chemical compound widely used in the production of polymers, cosmetics, and pharmaceuticals. To diminish the methanol content, we pretreated RAPTR by low-pressure evaporation, which minimized water evaporation and prevented other contaminants from being concentrated. We optimized the evaporation conditions by using a 2² central composite rotational design to establish optimal temperature and time of 55 °C and 51.3 min, respectively. Pretreated RAPTR diluted at 20% (v v⁻¹) with a nutrient solution allowed the bacterium to grow, but no glycerol was consumed. Supplementing the nutrient solution with 0.4 g L⁻¹ MgCl₂, defined in another experimental design, led the bacterium to consume glycerol and to produce 1,3-PDO. In the optimized conditions, pretreated RAPTR supplemented with MgCl₂ gave 2.78 ± 0.01 g L⁻¹ 1,3-PDO in higher yield (Y_{1,3-PDO/glycerol}) compared to the theoretical one, 0.61 and 0.50 g g⁻¹, respectively. This result is relevant for biodiesel biorefineries, which could implement the innovative and customized strategy proposed herein to obtain 1,3-PDO, a high-value-added product, from a glycerol- and methanol-rich residue.

1. Introduction

In the biodiesel industry, the production worldwide of 55 billion liters yearly, generates 5.5 billion liters of a glycerol-rich residue, as a by-product from the biodiesel transesterification reaction [1,2,3].

Biodiesel industries primarily use oils (residual or not) and methanol (or other short-chain alcohols) as raw materials in transesterification reactions, giving biodiesel as the main product [4]. To increase the reaction efficiency, catalysts such as sodium or potassium hydroxide are employed [5], along with excess methanol (methanol/oil ratios ranging between 6:1 and 15:1 (v v⁻¹)) [5]. The aqueous phase of this reaction, known as the Residual Aqueous Phase of the Transesterification Reaction (RAPTR), is a glycerol-rich stream [6] where glycerol is produced at a 1:10 glycerol/biodiesel ratio (v v⁻¹). Besides glycerol, RAPTR contains methanol, NaCl or K₂SO₄, free fatty acids, and esters as contaminants even though methanol is usually recovered from the aqueous phase of the transesterification reaction by distillation and returned to the biodiesel production process [7,8]. Depending on the biodiesel industry facilities, the methanol concentration in RAPTR can reach 20% or more [8]. Although RAPTR can be further purified, this can be economically unfeasible because purification is costly, and glycerol has low market value [4,9]. Therefore, valorising this residue is an attractive alternative for the biodiesel industry that could decide how to better manage its waste [10].

Some microorganism species can use residual glycerol from biodiesel production as substrate to obtain value-added products [11,12]. However, these microorganisms may be sensitive to the impurities in residual glycerol [12,13]. Therefore, employing this substrate in microbial processes may require pretreatment or specific culture medium formulations, as well as bioreactor operation modes to mitigate toxicity issues [14,15,16].

Among the limited number of species capable of fermenting glycerol, species belonging to the genus Clostridium are notable. These bacteria consume glycerol and can convert it to 1,3-propanediol (1,3-PDO). 1,3-PDO is a highly demanded commodity used to produce automotive polymers, cosmetics, and refrigerant fluids [14,17]. During glycerol metabolism, Clostridium bacteria produce organic acids and hydrogen as oxidative metabolism by-products [18] and generate reducing equivalents. The latter are re-oxidized through reductive metabolic reactions, including the reaction that yields 1,3-PDO [14].

Nevertheless, some Clostridium species cannot consume glycerol because they lack membrane transporters or the culture medium is deficient in nutrients. In a previous paper [19], our group demonstrated that the culture medium composition determines whether Clostridium species consume glycerol and hence produce 1,3-PDO. By evaluating how different compounds affect glycerol consumption and 1,3-PDO production by C. beijerinckii Br21, we found that magnesium chloride (MgCl₂) exerts one of the most positive effects. Indeed, glycerol transport into C. beijerinckii cells is mediated by the glycerol uptake facilitator protein, whose function and stability depend on Mg²⁺ ions [20,21].

Given the high toxicity of RAPTR to microorganisms and the challenges associated with glycerol being used by Clostridium species, in this study we have conducted two optimization stages to make RAPTR a suitable substrate for C. beijerinckii Br21 to produce 1,3-PDO. In the first experimental design, we pretreated RAPTR via low-pressure evaporation. Such pretreatment aims to reduce the methanol concentration while keeping water removal to a minimum, preventing other toxic compounds from being concentrated in RAPTR. In the second experimental design, we optimized the MgCl₂ concentration in the nutrient solution used to dilute RAPTR to enhance glycerol consumption and 1,3-PDO production. Combining these two approaches, we have achieved an innovative and customized strategy to apply RAPTR within the context of biodiesel biorefineries.

2. Materials and Methods

2.1. Residue and Characterization Methods

RAPTR was provided by a biodiesel production company located in the region of Campinas, state of São Paulo, Brazil. Specifically, this company conducts transesterification of soybean oil and methanol, catalyzed by NaOH (Figure 1). When received by our group, RAPTR was fractionated and stored at 4 °C throughout the experimental period.

2.2. Microorganism and Inoculum

C. beijerinckii Br21, a non-solventogenic strain isolated from sugarcane vinasse treatment sludge [22], was used in the fermentation assays. The genome of this strain had been sequenced and deposited at the DDBJ/ENA/GenBank database under number MWMH00000000 [23]. The strain stock culture was maintained in Reinforced Clostridial Medium (RCM) with 30% cryoprotector at −80 °C. The strain was activated by adding 1 mL of its stock culture to a Hungate tube containing 5 mL of sterile RCM (from which oxygen had been removed by adding nitrogen) and incubated at 37 °C for 18 h. After that, a pre-inoculum was prepared by adding the strain (10% v v⁻¹) to the culture medium at 37 °C and 150 rpm for 24 h, which allowed C. beijerinckii Br21 to adapt to the culture medium. The strain in the pre-inoculum was then employed as the inoculum for the fermentation assays. To this end, the strain was added in a volume that was sufficient to increase the optical density (OD) at 600 nm of the assay medium by 0.1.

2.3. RAPTR Pretreatment Optimization Method

RAPTR was pretreated by low-pressure evaporation at −0.76 bar. For this purpose, 100 mL of RAPTR was added to a 250 mL round-bottom flask attached to the rotary evaporator (Rotavapor RE 111—Büchi). The pretreatment temperature and time were optimized by using a central composite rotational design (CCRD). After low-pressure evaporation, the liquid remaining in the flask was recovered, its volume was measured, and samples were collected to quantify glycerol and methanol. Volume reduction was calculated by applying Equation (1).

$$\mathrm{V}\mathrm{o}\mathrm{l}\mathrm{u}\mathrm{m}\mathrm{e} \mathrm{r}\mathrm{e}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\left(\mathrm{\%}\right)=\left({\displaystyle \frac{\mathrm{I}\mathrm{n}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{a}\mathrm{l} \mathrm{v}\mathrm{o}\mathrm{l}\mathrm{u}\mathrm{m}\mathrm{e}\left(\mathrm{m}\mathrm{L}\right)-\mathrm{R}\mathrm{e}\mathrm{c}\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{e}\mathrm{d} \mathrm{v}\mathrm{o}\mathrm{l}\mathrm{u}\mathrm{m}\mathrm{e}\left(\mathrm{m}\mathrm{L}\right)}{\mathrm{I}\mathrm{n}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{a}\mathrm{l} \mathrm{v}\mathrm{o}\mathrm{l}\mathrm{u}\mathrm{m}\mathrm{e}\left(\mathrm{m}\mathrm{L}\right)}}\right)\times 100$$

(1)

2.4. Optimization of Glycerol and MgCl₂ Concentration in the Culture Medium

To optimize the MgCl₂ concentration, a basal culture medium, named WISMod [19], was used. The medium consisted of 34.14 g of commercial glycerol (Synth), 5 g of potassium hydrogen phosphate (K₂HPO₄), 1 g of yeast extract, 0.055 g of iron (II) sulfate (FeSO₄), and 0.005 g of sodium acetate (CH₃COONa) per liter. The culture media were prepared by following the WISMod medium composition, except that the glycerol concentration in the basal medium was replaced with the concentration specified by the CCRD, and MgCl₂ was added as described by the CCRD. After the medium was prepared, the pH was approximately 8.6 and was adjusted to 7.2 using 1 mol L⁻¹ HCl. Then, 50 mL of the medium was added to a 100 mL penicillin flask. Nitrogen was bubbled, to remove oxygen, and the flask was sealed and autoclaved at 121 °C for 20 min. After inoculation with C. beijerinckii to an OD of 0.1, the medium was incubated at 37 °C and 150 rpm for 48 h. At the start of the fermentation assay and after 48 h, aliquots were collected to measure the OD at 600 nm, pH, and glycerol, butyric acid, and 1,3-PDO concentrations. The fermentation assays for the CCRD were conducted in triplicate.

2.5. Pretreated RAPTR Fermentation

To ferment pretreated RAPTR, the optimum medium containing MgCl₂, as determined in this study, was used. The medium composition per liter included 25 g of pretreated RAPTR, 0.4 g of MgCl₂, 5 g of K₂HPO₄, 1 g of yeast extract, 0.055 g of FeSO₄, and 0.005 g of CH₃COONa.

To prepare the medium, pretreated RAPTR was diluted to 20% (v v⁻¹). Next, the nutrients were added, and the pH was adjusted to 7.2. Afterward, the same procedures described in Section 2.4 were followed.

The substrate-to-product conversion factor (Y_P/S) was calculated as the ratio between the final product (1,3-PDO) concentration and the consumed substrate (pretreated RAPTR) concentration. For Clostridium species, the theoretical maximum Y_P/S is 0.5 g g⁻¹ [24].

2.6. Analytical Methods

Total organic carbon (TOC) was measured on a TOC-VCPH (Shimadzu). The concentrations of total, total fixed, and total volatile solids were measured by following the methodology described by the American Public Health Association [25]. The pH was measured with a pH meter (edge HI 2002, Hanna) directly in the sample. All the analyses were performed in triplicate.

The concentrations of butyric acid, 1,3-PDO, and glycerol in the samples without RAPTR were quantified by gas chromatography (GC) on the GC 2014 instrument (Shimadzu) equipped with a Stabilwax-DA column (30-m length, 0.25 mm diameter, and 0.25 µm bonded film). The method described by Egoburo et al. in 2017 [26] was followed. Nitrogen at 2.5 mL min⁻¹ was used as the carrier gas. A heating ramp (185 °C for 3 min, heating at 40 °C min⁻¹ to 220 °C, and 220 °C for 1 min) was applied. For the analyses, 2 µL of sample was injected into the injector at 300 °C at a 30:1 split ratio and analyzed with a flame ionization detector (FID) at 290 °C.

To quantify methanol, 1,3-PDO, and glycerol in the samples containing RAPTR or pretreated RAPTR, the same equipment and column were used, and adaptations to the method described by Egoburo et al. in 2017 [26] were made. The heating ramp was modified to 40 °C for 2.5 min, heating to 185 °C at 100 °C min⁻¹, 185 °C for 1.5 min, heating to 220 °C at 40 °C min⁻¹, and 220 °C for 2.5 min.

To determine the C. beijerinckii Br21 cell concentration, OD was measured at 600 nm using a UV-VIS M51 spectrophotometer (BEL Engineering). The pH was measured with a pH meter (edge HI 2002, Hanna).

2.7. Statistical Validation

Statistical differences were assessed by using Tukey’s methodology at p < 0.05. The CCRD was analyzed by employing the Statistica 14.0.1 software [27], which provided an ANOVA table (Analysis of Variance table) for pure error and determined the significance of variables through the F-test at p < 0.05 for medium optimization and p < 0.1, for RAPTR pre-treatment. The optimal point was validated by performing the experiment at the optimal point and comparing the result with the predicted result, with a maximum variation of 5%.

3. Results and Discussion

3.1. Residue (RAPTR) Composition

RAPTR is obtained after phase separation at the end of the biodiesel transesterification reaction. In this process, the organic phase is separated and directed for purification, to yield commercial biodiesel. In turn, the aqueous phase (RAPTR) is sent for methanol recovery and residual glycerol disposal (Figure 1) [7,8].

Methanol and glycerol are the main constituents of RAPTR, along with other unidentified organic carbon sources. In addition, on the basis of total fixed solids, RAPTR contains inorganic compounds. The residue contained 294.3 ± 6.6 g L⁻¹ of methanol, an excessively high concentration for microorganism growth. For comparison, crude glycerol, a purified version of RAPTR, contains methanol at concentrations ranging from 0.25 to 17 g L⁻¹ [28,29], whereas other biodiesel production residues, such as wastewater [30], contain methanol at concentrations of up to 40.3 g L⁻¹ [31].

RAPTR (

Table 1. Chemical characterization of RAPTR.

Component	Unit	Value
Glycerol	g L−1	67.9 ± 0.2
	g C L−1	26.5 ± 0.1
	% m m−1	6.6 ± 0.02
Methanol	g L−1	294.3 ± 6.6
	C L−1	110.2 ± 2.5
	% m m−1	28.4 ± 6.4
TOC	g L−1	187.6 ± 1.8
Total solids	g L−1	140.4 ± 2.6
Total fixed solids	g L−1	99.0 ± 0.8
Total volatile solids	g L−1	41.4 ± 2.72
pH	-	1.3

) contained 67.9 ± 0.2 g L⁻¹ glycerol, the main co-product of the transesterification reaction. This concentration is suitable for microorganisms, including C. beijerinckii Br21, to grow [32,33]. The glycerol concentration in RAPTR corresponds to 6.56 ± 0.02% (m m⁻¹), which is low compared to crude glycerol (between 14% and 87% (m m⁻¹)) [34,35]. The glycerol concentration obtained after RAPTR is distilled (Figure 1) can be higher [7,8].

The TOC concentration in RAPTR was 187.6 ± 1.8 g L⁻¹, indicating high organic load. Of this total, 26.5 ± 0.1 and 110.2 ± 2.5 g C L⁻¹ corresponded to glycerol and methanol. The remaining 50.9 g C L⁻¹ likely originated from non-consumed free fatty acids or partially substituted fatty acids formed during transesterification [36].

Finally, RAPTR contained 140.4 ± 2.6 g L⁻¹ of total solids, with total fixed and total volatile solids at 99.0 ± 0.8 and 41.4 ± 2.72 g L⁻¹, respectively. The high concentration of inorganic solids in RAPTR could be due to the raw materials and catalysts used during transesterification. TVS indicate the presence of undissolved organic matter.

3.2. RAPTR Pretreatment Optimization

Aiming at RAPTR fermentation by C. beijerinckii Br21, after adjusting the pH to 7.0, we conducted fermentation assays under the following conditions: (1) direct RAPTR fermentation; (2) direct fermentation of RAPTR diluted 2 and 10 times; (3) fermentation of RAPTR diluted 10 times with 1 g L⁻¹ of yeast extract; or (4) fermentation of RAPTR diluted 10 times with basal culture medium added with nutrients [19]. However, after fermentation at 37 °C for 168 h, no bacterium grew under any of these conditions. Thus, RAPTR was toxic to C. beijerinckii Br21.

The methanol concentration in RAPTR (294.3 ± 6.6 g L⁻¹) could be a source of toxicity. At high concentrations, methanol and other alcohols are toxic to microorganisms [37]. For instance, methanol has an LD₅₀ of 54.5 g L⁻¹ for Acinetobacter calcoaceticus [38], which is five times lower than the methanol concentration in RAPTR (Table 1).

Among the possible strategies to remove methanol, distillation is the most feasible given that this alcohol has a low boiling point (64.5 °C). Nevertheless, removing methanol from RAPTR through evaporation implies increasing the concentration of other potentially toxic compounds or producing new toxic compounds. Therefore, a lower temperature should be used to minimize RAPTR volume reduction. For this reason, we carried out evaporation under reduced pressure, which allows the evaporation temperature to be lowered. This technique also reduces volume loss due to water evaporation.

We established the optimal conditions for RAPTR low-pressure evaporation by applying a CCRD. The variables were evaporation time (from 8.8 to 51.2 min) and temperature (from 40.9 to 69.1 °C) (

Table 2. Time and temperature conditions used during RAPTR evaporation by CCRD and results for volume reduction and methanol and glycerol concentrations in pretreated RAPTR.

Conditions	Pretreatment Condition		Residue Volume Reduction (%)	Methanol (g L−1)	Glycerol (g L−1)
	Time (min)	Temperature (°C)
1	15.0 (−1)	45.0 (−1)	1	291.0	67.5
2	15.0 (−1)	65.0 (+1)	31	128.8	93.3
3	45.0 (+1)	45.0 (−1)	1	295.2	72.5
4	45.0 (+1)	65.0 (+1)	67	93.5	197.9
5	8.8 (−1.41)	55.0 (0)	5	277.8	74.6
6	51.2 (+1.41)	55.0 (0)	27	137.0	93.0
7	30.0 (0)	40.9 (−1.41)	1	295.7	66.4
8	30.0 (0)	69.1 (+1.41)	70	92.7	227.0
9	30.0 (0)	55.0 (0)	20	183.3	88.8
10	30.0 (0)	55.0 (0)	15	205.8	78.8
11	30.0 (0)	55.0 (0)	14	217.8	80.1
Residue	-	-	-	294.3 ± 6.6	67.9 ± 0.2

).

The assays performed under the conditions defined by the CCRD decreased the methanol concentration from 294.3 ± 6.6 g L⁻¹ in RAPTR to 92.7 g L⁻¹ in pretreated RAPTR. Nonetheless, methanol was not completely removed, which was caused by the limitation in volume reduction that prevented other toxic compounds from concentrating in RAPTR. Thus, we analyzed the model by using ANOVA for the responses of volume reduction (

Table 3. ANOVA table for the volume reduction response model.

Factors	Significance	Sum of Squares	Degrees of Freedom	Mean Square	F	p
Linear Time	Significant	563.014	1	563.014	1126.029	0.018966
Linear Temperature	Significant	4684.188	1	4684.188	9368.375	0.006577
Quadratic Temperature	Significant	568.029	1	568.029	1136.057	0.018882
Interaction 1 L and 2 L	Significant	324.000	1	324.000	648.000	0.024996
Lack of Fit	Not significant	5.869	4	1.467	2.935	0.409269
Pure Error		0.500	1	0.500
Total		6145.600	9

) and methanol concentration (

Table 4. ANOVA table for the methanol concentration response model.

Factors	Significance	Sum of Squares	Degrees of Freedom	Mean Square	F	p
Linear Time	Significant	6625.23	1	6625.23	92.0171	0.066127
Linear Temperature	Significant	52,972.74	1	52972.74	735.7325	0.023460
Lack of Fit	Not significant	5005.25	6	834.21	11.5862	0.221180
Pure Error		72.00	1	72.00
Total		64,675.22	9

).

Regarding volume reduction (Table 3), the variables linear time, linear temperature, quadratic temperature, and the interaction between variables (interaction between variables [1 L and 2 L]) were significant. The lack of fit was not significant, so the model is significant and predictive and hence valid for developing the model.

As for methanol concentration (Table 4), only the variables linear time and linear temperature were significant, while the lack of fit was not significant. Therefore, the model is significant and predictive.

To obtain the optimal conditions to pretreat RAPTR, we used the desirability methodology to minimize the responses of volume reduction and methanol concentration (Figure 2 and Figure S1).

On the basis of desirability analysis, the optimal evaporation conditions were 55 °C (0) for 51.2 min (+1.41), which is condition 6 of the CCRD (Table 2). Under these conditions, the RAPTR volume reduced by 27%, with the glycerol and methanol concentrations being 93.0 and 137.0 g L⁻¹, respectively. Thus, compared to untreated RAPTR, the methanol concentration reduced by 53%, and the glycerol concentration increased by 37% in pretreated RAPTR.

After pretreating RAPTR and adjusting the pH to 7.0, we employed it as a substrate in fermentation assays with or without 20% (v v⁻¹) dilution, and with or without culture medium nutrients. Combining diluted pretreated RAPTR with added nutrients caused C. beijerinckii Br21 to grow (final OD at 600 nm = 0.46, Table S1), which had not been observed previously. However, the bacterium did not consume glycerol and hence did not produce 1,3-PDO. Under these conditions, only acetate from the culture medium was consumed for bacterial growth given that it is easily assimilated by C. beijerinckii Br21 [39]. This led us to tune the culture medium composition to facilitate glycerol uptake by the bacterium.

3.3. Optimization of Glycerol and MgCl₂ Concentrations in Culture Medium to Promote Glycerol Consumption

In a previous study, by growing C. beijerinckii Br21 in different culture medium compositions containing glycerol, our group observed that yeast extract, FeSO₄, and MgCl₂ influence 1,3-PDO production the most positively [19]. In this same study, we found that supplementing the culture medium with yeast extract and FeSO₄ increases the 1,3-PDO concentration.

Mg²⁺ ions have been reported to affect microbial metabolism. For example, these ions play a role in ATP synthesis [40], act as enzyme cofactors [41], and underlie cellular membrane stability [42]. In Clostridium, Mg²⁺ ions play a central part in glycerol metabolism: they function as a cofactor and stabilize the glycerol uptake facilitator protein, facilitating glycerol diffusion into the cell [20,21]. Therefore, Mg²⁺ ions are essential for C. beijerinckii Br21 to consume glycerol.

Aiming to increase glycerol consumption by C. beijerinckii Br21, we performed a CCRD to optimize the variables commercial glycerol concentration and MgCl₂ concentration in the culture medium.

During the fermentation assays, the pH decreased from 7 to 6.56 and 5.11 within 48 h. Thus, organic acids, especially butyric acid (

Table 5. CCRD assay conditions with glycerol and MgCl₂ concentrations and the respective OD, pH, consumed glycerol concentration, and concentrations of butyric acid and 1,3-PDO after fermentation for 48 h.

Conditions	Conditions		OD	Final pH	Butyric Acid (g L−1)	1,3-PDO (g L−1)	Consumed Glycerol (g L−1)
	Glycerol (g L−1)	MgCl2 (g L−1)
1	10 (−1)	0.2 (−1)	1.89 ± 0.11	5.93 ± 0.04	0.80 ± 0.12	1.77 ± 0.16	3.73 ± 0.20
2	10 (−1)	0.6 (+1)	1.76 ± 0.11	5.47 ± 0.09	1.07 ± 0.04	3.18 ± 0.13	6.44 ± 0.58
3	40 (+1)	0.2 (−1)	1.75 ± 0.07	5.35 ± 0.09	0.96 ± 0.08	2.67 ± 0.09	5.10 ± 0.24
4	40 (+1)	0.6 (+1)	2.42 ± 0.19	5.11 ± 0.04	1.28 ± 0.05	2.49 ± 0.06	4.30 ± 0.20
5	3.79 (−1.41)	0.4 (0)	0.73 ± 0.20	5.52 ± 0.08	0.47 ± 0.02	0.85 ± 0.12	2.79 ± 0.14
6	46.21 (+1.41)	0.4 (0)	1.61 ± 0.06	5.17 ± 0.09	1.12 ± 0.09	2.13 ± 0.18	3.86 ± 0.76
7	25 (0)	0.12 (−1.41)	1.05 ± 0.09	6.56 ± 0.08	1.02 ± 0.10	2.05 ± 0.02	7.17 ± 0.21
8	25 (0)	0.68 (+1.41)	1.29 ± 0.11	5.64 ± 0.04	1.05 ± 0.03	2.20 ± 0.10	6.03 ± 0.34
9	25 (0)	0.4 (0)	1.52 ± 0.08	5.22 ± 0.07	1.01 ± 0.07	2.18 ± 0.06	7.53 ± 0.37
10	25 (0)	0.4 (0)	1.78 ± 0.06	5.25 ± 0.09	1.10 ± 0.04	2.71 ± 0.18	8.52 ± 0.20

), a product of the glycerol oxidative pathway in C. beijerinckii Br21, were formed.

Glycerol consumption ranged between 2.79 ± 0.14 and 8.52 ± 0.20 g L⁻¹, so varying the concentrations of glycerol and MgCl₂ indeed altered consumption of this substrate (Table 5). ANOVA analysis (

Table 6. ANOVA table for the glycerol consumption response model.

Factors	Significance	Sum of Squares	Degrees of Freedom	Mean Square	F	p
Linear glycerol	Significant	3.1	1	3.1	15.3	0.001
Quadratic glycerol	Significant	57.6	1	57.6	279.7	0.000
Linear MgCl2	Significant	2.1	1	2.1	10.3	0.005
Quadratic MgCl2	Significant	5.5	1	5.5	26.8	0.000
Interaction 1 L by 2 L	Significant	7.4	1	7.4	36.1	0.000
Lack of fit	Not significant	0.3	3	0.1	0.4	0.725
Pure error		3.3	16	0.2
Total		73.0	24

) indicated that all the variables (linear glycerol, quadratic glycerol, linear MgCl₂, quadratic MgCl₂, and the interaction between variables [1 L and 2 L]) were significant, and that changing their concentrations modified glycerol consumption. The lack of fit, on the other hand, was not significant, so the obtained model (Equation (2)) proved predictive and can be used to generate the response surface (Figure 3) and to determine the optimal condition.

$$\mathrm{G}\mathrm{l}\mathrm{y}\mathrm{c}\mathrm{e}\mathrm{r}\mathrm{o}\mathrm{l} \mathrm{c}\mathrm{o}\mathrm{n}\mathrm{s}\mathrm{u}\mathrm{m}\mathrm{p}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\left(\mathrm{g}{\mathrm{L}}^{-1}\right)=7.93+0.42\mathrm{g}-2.32{\mathrm{g}}^2-0.31\mathrm{m}-0.68{\mathrm{m}}^2-0.87\mathrm{g}\mathrm{m}$$

(2)

where g represents the glycerol concentration (g L⁻¹), and m represents the MgCl₂ concentration (g L⁻¹).

The optimal condition indicated by the model (Equation (2)) was 25 g L⁻¹ glycerol and 0.4 g L⁻¹ MgCl₂. This condition resembles the conditions of the central point of the model and matches the conditions used in experiments 9 and 10 of the CCRD. Under this condition, the expected glycerol consumption would be 7.93 g L⁻¹, with an interval ranging from 7.53 to 8.33 g L⁻¹ at 95% confidence.

We conducted a fermentation assay by using the basal medium supplemented with 0.4 g L⁻¹ MgCl₂ and 25 g L⁻¹ commercial glycerol (Figure 4), and obtained glycerol consumption of 8.25 ± 0.43 g L⁻¹. This value fell within the confidence interval of the model, validating the optimal point and optimization results. We termed the optimized medium, supplemented with Mg²⁺ ions to enhance glycerol consumption, as the optimal medium.

Despite MgCl₂ supplementation, C. beijerinckii Br21 only consumed 33% glycerol. Such limited glycerol consumption could be due to accumulated organic acids and consequent decrease in pH [19,32,33]. Such a pH drop should cause ATP to be consumed to maintain cytosolic pH [33,43]. MERMEJO et al., in 2022 [33], demonstrated that controlling the pH increases glycerol consumption and 1,3-PDO production. Thus, controlling the pH by using a bioreactor could further increase the 1,3-PDO concentration.

In the fermentation assays using the basal medium, as reported by ALTAFINI et al. (2022) [19], and the medium with MgCl₂ (

Table 7. Kinetic comparison of the results obtained for the basal medium [19] and the medium with optimal glycerol (commercial) and MgCl₂ concentrations.

	Basal Medium (ALTAFINI et al. (2022) [19])	Optimal Medium (This Work)
Lag phase (h)	4	22
Maximal OD at 600 nm	1.61 ± 0.06	2.40 ± 0.20
Total glycerol consumption (g L−1)	7.2 ± 0.6	8.25 ± 0.43
Maximal glycerol consumption rate (g L−1 h−1)	0.86 ± 0.02	1.80 ± 0.01
1,3-PDO (g L−1)	4.01 ± 0.01	3.54 ± 0.20
Maximal 1,3-PDO production rate (g L−1 h−1)	0.91 ± 0.01	0.96 ± 0.01

), the lag phase increased from 4 to 22 h, respectively. During batch fermentation, extended lag phase lowers productivity. However, the log phase increased from 8 to 11 h in the optimal medium, which may be attributed to increased glycerol consumption and higher cell concentrations, as evidenced by the OD increasing from 1.61 to 2.40.

Upon MgCl₂ addition, substrate consumption increased by 1.05 g L⁻¹, and the maximum substrate consumption rate went from 0.86 ± 0.02 to 1.80 ± 0.01 g L⁻¹ h⁻¹, that is, it more than doubled.

Besides increased glycerol consumption, the maximum production rate of 1,3-PDO rose from 0.91 ± 0.01 to 0.96 ± 0.01 g L⁻¹ h⁻¹. Nevertheless, the 1,3-PDO concentration decreased from 4.01 ± 0.01 g L⁻¹ in the basal medium to 3.54 ± 0.20 g L⁻¹ in the optimal medium. Glycerol uptake probably improved metabolism toward the oxidative pathway, as suggested by increased bacterial growth in the optimal medium.

The preference for the oxidative pathway over the reductive pathway, indicated by the higher substrate consumption rate, increased the cell concentration and reduced the final 1,3-PDO concentration. This could also be associated with the effects of Mg²⁺ ions on proteins involved in oxidative metabolism.

In the genome of C. beijerinckii Br21, the gene encoding glycerol uptake (glpF: locus range 573410 to 574123 on SCAFF01) is followed by the gene encoding glycerol kinase (glpK: locus range 575059 to 576558 on SCAFF01), the enzyme that channels glycerol to the oxidative pathway. Increased glpF expression, promoted by the presence of Mg²⁺ ions, possibly enhances glpK expression.

Increased glycerol consumption and higher maximum consumption rate suggested that adding MgCl₂ positively influenced substrate consumption, which could help uptake of the glycerol-rich RAPTR. Furthermore, effects such as membrane stabilization [42] could make microorganisms more resistant to high concentrations of short-chain alcohols [44].

3.4. Fermentation of Pretreated Glycerol-Rich RAPTR Supplemented with MgCl₂

Finally, we accomplished a fermentation assay by using pretreated RAPTR diluted to 20% v v⁻¹ and supplemented with MgCl₂, aiming to enhance glycerol consumption (Figure 5). This fermentation assay resulted in OD of 0.68 at 600 nm, consumption of 4.57 ± 0.03 g L⁻¹ glycerol, and maximum consumption rate of 0.1 g L⁻¹ h⁻¹. When we conducted the same assay with the residue diluted at 20% v v⁻¹ in the basal medium composition, the final OD was only 0.46. Therefore, combining RAPTR pretreatment with MgCl₂ supplementation favored RAPTR use by C. beijerinckii Br21. However, less glycerol was consumed compared to the medium containing commercial glycerol and the optimal MgCl₂ concentration, so the negative effects of RAPTR still affected the bacterium. This was consistent with the low cell concentration observed in terms of OD, which was only 0.68 ± 0.03, much lower than OD of 2.40 ± 0.20 measured in the medium optimized with commercial glycerol (Figure 4).

At the end of fermentation, the 1,3-PDO concentration was 2.78 ± 0.01 g L⁻¹, which was high considering the low glycerol consumption. This resulted in Y_P/S of 0.61 g g⁻¹, higher than 0.5 g g⁻¹ considered as the theoretical maximum [24]. Thus, C. beijerinckii Br21 might have consumed a second substrate, such as acetate, and directed glycerol to produce 1,3-PDO.

As mentioned earlier, Mg²⁺ ions have various effects on the cell and likely allow pretreated RAPTR to become a fermentable substrate. The first possible effect of these ions is on the glycerol uptake facilitator protein, facilitating glycerol diffusion into the cell [20,21]. A second effect would be promoting ATP synthesis reactions, where Mg²⁺ ions also play a key role in guiding the reaction and allowing inorganic phosphate to bind to the correct position on ADP [40]. Lastly, less reported but equally important, is cell membrane stabilization due to Mg²⁺ ions [42] and increased microorganism resistance to alcohol. In Saccharomyces cerevisiae, addition of Mg²⁺ ions has been reported to increase yeast resistance to ethanol [42,44]. A similar effect may occur for Mg²⁺ ions in C. beijerinckii when the glycerol residue containing methanol is in contact with the bacterium.

WISCHRAL et al., in 2016 [15], reported using residual glycerol to produce 1,3-PDO in the presence of C. beijerinckii. The authors did not purify the residue and only diluted it before accomplishing fermentation to obtain 1,3-PDO. However, these authors used a residue with approximately 400 g L⁻¹ glycerol, so the residue had already undergone methanol recovery processes. The difference in the initial residue explains the significantly differing results: the authors achieved a final 1,3-PDO concentration of 19.7 g L⁻¹, but Yp/s of 0.55 g g⁻¹. Nevertheless, here we have demonstrated that C. beijerinckii strains can produce 1,3-PDO, and that an adequately formulated residue can be used for this purpose.

4. Conclusions

Using a residue obtained from the biodiesel industry directly from the biodiesel transesterification reaction in microbial processes presented some issues: besides glycerol, the residue contains a high methanol concentration and is toxic to microorganisms. Submitting the residue to low-pressure evaporation lowers the methanol concentration and reduces its toxicity, so that microorganisms can grow therein. However, only after MgCl₂ is introduced to the pretreated residue is glycerol consumed and 1,3-PDO produced. In conclusion, this study has demonstrated that Mg²⁺ ions promote glycerol consumption by a C. beijerinckii strain in a complex residue containing short-chain alcohols. Our findings are relevant for biodiesel biorefineries, which could implement our customized strategy combining residue evaporation and MgCl₂ supplementation to obtain a glycerol-rich residue suitable for producing 1,3-PDO.

To make this process more attractive for large-scale integration, further technological enhancements could include optimization of nutrient composition, implementation of continuous or fed-batch fermentation with pH control, and advanced detoxification strategies such as activated carbon treatment or membrane separation to further reduce inhibitory compounds. These approaches could increase productivity, reduce costs, and improve the overall feasibility of integrating 1,3-PDO bioproduction into existing biodiesel biorefineries.