Valuation of agro-industrial wastes as substrates for heterologous production of α-galactosidase

Background The recycling of agro-industrial wastes is at present limited by the availability of efficient and low-cost enzyme cocktails. The use of these materials as culture media to produce the enzymes can contribute to the profitability of the recycling process and to the circular economy. The aim of this work is the construction of a recombinant yeast strain efficient to grow in mixed whey (residue of cheese making) and beet molasses (residue of sugar manufacture) as culture medium, and to produce heterologous α-galactosidase, an enzyme with varied industrial applications and wide market. Results The gene MEL1, encoding the α-galactosidase of Saccharomyces cerevisiae, was integrated (four copies) in the LAC4 locus of the Kluyveromyces lactis industrial strain GG799. The constructed recombinant strain produces high levels of extracellular α-galactosidase under the control of the LAC4 promoter, inducible by lactose and galactose, and the native MEL1 secretion signal peptide. K. lactis produces natively beta-galactosidase and invertase thus metabolizing the sugars of whey and molasses. A culture medium based on whey and molasses was statistically optimized, and then the cultures scaled-up at laboratory level, thus obtaining 19 U/mL of heterologous α-galactosidase with a productivity of 0.158 U/L h, which is the highest value reported hitherto from a cheap waste-based medium. Conclusions A K. lactis recombinant strain was constructed and a sustainable culture medium, based on a mixture of cheese whey and beet molasses, was optimized for high productivity of S. cerevisiae α-galactosidase, thus contributing to the circular economy by producing a heterologous enzyme from two agro-industrial wastes. Electronic supplementary material The online version of this article (10.1186/s12934-018-0988-6) contains supplementary material, which is available to authorized users.


Background
Agro-industrial activities generate huge amounts of wastes susceptible to be transformed in culture media for bio-productions with high added value, thus contributing to the circular economy. One outstanding bioproduction is the microbial fermentation to ethanol [1,2] or other biofuels [3] as an alternative to current threat of fossil fuels depletion and growing environmental concerns associated to their massive consumption. The production of heterologous proteins, including enzymes, using microbial cell factories is also remarkable [4].
The fermentation of agro-industrial waste materials to biofuels is hitherto still a challenge that requires the development of efficient and low-cost enzyme cocktails to be profitable [11]. The use of residues as culture media to produce the required enzymes can help with the economy of the process. Whey is a sub-product of cheese making, rich in the milk sugar lactose, that is produced in increasingly high amounts world-wide [15], while molasses are very abundant sub-products of sugar manufacture, rich in sucrose and raffinose [16,17]. Large research has been performed aimed to the biotechnological use of these waste materials, thus avoiding their polluting impact, being a main destine their fermentation to bioethanol [1]. An increase in yield has been reported by mixing both sub-products [15,[18][19][20].
Kluyveromyces lactis and S. cerevisiae are two foodgrade yeasts widely used in biotechnology [21,22]. Apart from other advantages for heterologous protein production, K. lactis is able to use a greater variety of carbon sources than S. cerevisiae, being well documented the case of lactose, i.e., the ability to metabolize lactose of the former and the inability of the second species, while both can use sucrose [23,24]. Due to its special features, K. lactis has been used successfully as a cell factory for recombinant protein secretion in large-scale cultures; however, the use of agro-industrial wastes such as whey-molasses mixed media as substrates has not been reported [21]. Mixtures of molasses or bagasse and milk or cheese whey have been used for ethanol production by Kluyveromyces sp. [18][19][20], but not for recombinant protein production.
With the aim to develop a sustainable and cheap system for production of ScAGal, in this paper we describe the work performed to construct recombinant K. lactis strains that secrete high levels of ScAGal and to optimize a culture medium based on two agro-industrial waste materials, cheese whey and beet molasses.
Phusion High-Fidelity DNA Polymerase (for gene amplification), DreamTaq polymerase (for colony PCR analysis), and the DNA purification kits GeneJEt Gel Extraction Kit and GeneJET Plasmid Miniprep Kit, were used following the specifications of the supplier (Fisher Scientific).
Low salt LB (1% bactotryptone, 0.5% yeast extract, 0.5% NaCl) and YPD (1% yeast extract, 2% peptone, 2% glucose), were used as culture media for growth and maintenance of bacteria and yeasts, respectively. Low salt LB was supplemented with 100 mg/L ampicillin (Sigma Aldrich) for the propagation of plasmids in bacteria. YCB medium (30 mM Phosphate buffer, pH 7, 1.17% yeast carbon base) supplemented with 5 mM acetamide (Sigma Aldrich) was used for the selection of transformant yeasts. To solid media, 2% agar was added. All the media components were sterilized by autoclave at 121 °C for 20 min, except ampicillin and acetamide that were filtered through 0.22 μm microfiltration membrane (Sartorius AG).

Construction of recombinant K. lactis strains
Two different constructions with the gene MEL1His, i.e. MEL1 that encodes ScAGal (Uniprot P04824) fused to a Poli-His sequence by the C-terminal, were cloned into the integrative K. lactis expression vector pKLAC-1 (New England Biolabs). The complete MEL1His ORF with and without the N-terminal K. lactis α-MF domain sequence (encoding 18 amino acids) were amplified with the primer pairs P1/P3 and P2/P3, respectively (Table 1). P1 was designed to encode the excision site of the endoprotease Kex, and thus assure the correct processing of the fusion protein with the signal peptide of the K. lactis α-MF. P2 contains the HindIII site directly in frame with the ORF MEL1His, excluding the α-MF domain but with the native signal sequence of MEL1. The PCR products were purified and cloned between the XhoI-BglII and HindIII-BglII sites of pKLAC-1, giving rise to pKLACMEL1His-1 and pKLACMEL1His-2, respectively (Fig. 1). The recombinant plasmids were verified by sequencing (Servizos de Apoio á Investigación, Universidade da Coruña) with primers P4 and P5 (Table 1). Both plasmids were linearized with SacII, to obtain the expression cassettes that, after being purified, were integrated by homologous recombination in competent cells of K. lactis GG799 and Y1140 that were transformed by the lithium acetate procedure [28]. The transformants were selected in YCB-acetamide plates after incubation at 30 °C for 4 days. Recombinant colonies were checked by PCR [29] with the primers P5, P6 and P7 ( Table 1) that anneal inside and outside the expression cassette, to confirm that the integration in single and multiple copy was correct (Fig. 2a).

Determination of the number of copies of the integrated cassette
qPCR and Southern blot analysis were carried out to determine the copy number of the integrated exogenous DNA. Genomic DNA isolation from yeasts was performed by the method described in [30]. Amount and quality of DNA were estimated by 0.7% agarose gel electrophoresis and spectrophotometry (Biospectrometer Kinetic Eppendorf ). Samples with a ratio 260 nm/280 nm equal or higher than 1.8 were selected. qPCR was performed using an Eco Real-Time PCR System with software version 4.0.7.0 (Illumina, Inc., San Diego, California, USA). The threshold cycle (Ct) was determined by the 'Fit Points Method' in the software. The primers were designed with Tm 60.59-60.49 °C to synthesize amplicons of 203 bp. Reaction mixture was prepared with KAPA SYBR FAST Kit (Kapa Biosystems) that uses SYBR Green as fluorophore. Serial  [32]. The kit DIG DNA Labelling and detection Kit (Roche) was used for Southern blot assays. The probe was prepared using the primers P12/13 (Table 1) to obtain an amplicon of 423 bp that was purified and labelled with digoxigenin. Genomic DNA (3 ng/μL) was incubated with SpeI and ApaI at 37 °C overnight. Digested samples and GeneRuler High Range DNA LAdder (Thermo Scientific) were loaded, with buffer containing GelGreen Nucleic Acid Gel Stain (Biotium), in a 0.4% agarose gel and electrophoresed. Separated fragments were transferred to a positively charged nylon membrane Hybond-N+ (GE Healthcare) using the protocol as per [26], and hybridization was performed at 52 °C for 12 h with the labelled probe denatured at 80 °C for 3 min. Then, the membrane was washed 10 min at 52 with 2X SSC (0.6 M NaCl, 60 mM sodium citrate, pH 7) containing 0.1% SDS, and a second wash of 10 min at room temperature. After immunological detection and colorimetric reaction, performed as described in the kit protocol, the membrane is guided by the 3´and 5´ends of P LAC4-PBI (black boxes) to the LAC4 promoter region that is reconstructed after the integration. P6 primer anneals upstream P LAC4 of the original strain and P7 primer anneals upstream P LAC4-PBI inside the expression cassette. Primer pairs P6-P5 and P7-P5 were used to analyze if the integrations were single or multiple copy, respectively. b During Southern blot analyses, the genomic DNA was digested with SpeI and ApaI, and bands separated by electrophoresis before Southern transference. The probe (white box) when hybridizes with single copy bands reveals a fragment of 12.4 Kb that increases in 7.4 Kb per additional cassette integrated in multicopy strains. c Immunological detection and colour reaction allowed to determine the number of copies of the integrated cassette in KGM strains. MW, molecular weight marker (Kb) was photographed, and the visualized fragments analyzed with the program Molecular Imager Gel Doc XR+ (BioRad).

Culture media and conditions for ScAGal expression in recombinant K. lactis strains
Culture media used to induce the expression of the heterologous protein ScAGal in the K. lactis recombinant strains constructed in this work were: YPG (1% yeast extract, 2% peptone, 2% galactose), YPL (1% yeast extract, 2% peptone, 2% lactose) and YPW (1% yeast extract, 2% peptone, 2% cheese whey). Also, a culture medium based in the mixture of the food industry subproducts cheese whey and beet molasses was used to optimize the production of the heterologous protein ScAGal, as described below. Cheese whey (concentrated ultrafiltration permeate) and beet molasses were provided by QUEIZUAR, SL (Bama, A Coruña, Spain) and AB Azucarera Iberia, formerly Azucarera Ebro (Spain), respectively. Molasses were diluted in distilled water ratio 1:1 (v/v) to facilitate handling and, before autoclave, both molasses and cheese whey were centrifuged at 10,000 rpm for 15 min to remove solid impurities. Sugars present in molasses and cheese whey were identified and quantified by HPLC, using as standards mixtures of raffinose, sucrose, glucose, galactose and fructose in the case of molasses; and mixtures of lactose, glucose and galactose in the case of cheese whey. Both standards were prepared with five points from 4 to 0.06 mg/mL of each analyte.
To prepare the inoculum of the cultures, one isolated K. lactis colony was grown in YPD at 30 °C and 250 rpm until reaching an OD at 600 nm of 10. This pre-culture was used to inoculate 100 mL Erlenmeyer flasks containing 20 mL of YPG, YPL or YPW up to an initial optical density at 600 nm (OD 600 ) of 0.5 (30 °C, 250 rpm). For the cultures with the optimized medium, an initial inoculum of 10 OD 600 was used, keeping the rest of the here described conditions.

Heterologous expression of ScAGal in recombinant K. lactis strains
Cultures of selected transformants isolated in YCB-acetamide were initially performed in 96-well microtiter plates (TC Plate 96 Well, Sarstedt) with 200 μL YPD per well and incubated at 30 °C and 300 rpm. After 24 h of incubation, 20 μL of each well were used to inoculate a second plate with 200 μL YPG per well, to induce the expression under the same incubation conditions. After 72 h of incubation, the plates were kept at 4 °C to favor cell decanting and then analyze extracellular α-galactosidase activity with the cell-free supernatant. In each case, the plates contained three cultures of a recombinant strain with single copy integration. Then, the candidate with single copy integration and those with multiple copy integration that showed the highest activity were chosen for the cultures in shaken flasks with YPG as described above. Samples were taken at predetermined time intervals to measure growth (biomass) and enzymatic activity.
Biomass was measured as OD 600 with a UV-Visible spectrophotometer (Biospectrometer Kinetic Eppendorf ). Culture samples were centrifuged (5000 rpm, 5 min) and supernatants were used to measure extracellular activities of heterologous α-galactosidase and native invertase, whereas the cell pellet was permeabilized with 15% chloroform and 0.01% SDS to measure native β-galactosidase.

Optimization of ScAGal production by recombinant K. lactis strains
The chosen K. lactis recombinant strain was first cultured in YPL and YPW. Then, the surface response methodology (SRM) was applied to maximize ScAGal production in a medium composed by cheese whey (concentrated ultrafiltration permeate) and beet molasses. A factorial experimental design 3 2 allowed to study the influence, individual and combined, of the factors (independent variables) cheese whey and beet molasses concentration, on the response (dependent variable) that is quantified as the extracellular α-galactosidase activity of the performed cultures under the assayed conditions. Each factor was added at the concentration determined by the experimental design as percentage of total sugars, and all the cultures were supplemented with 1% yeast extract to provide nitrogen source. Real and coded values for the factors at three different levels, − 1, 0, + 1, being 0 the central point of the experimental domain, are shown in Table 2. A second order quadratic model was used to correlate the response or dependent variable with the independent variables or factors, and the optimum point of the experimental domain was obtained solving the regression equation and analyzing the contour response surface graphic. An ANOVA was applied, with confidence intervals of 95%, to determine the statistical significance of the effects of the factors. Beet molasses (%), X 1 6 7.5 9 Cheese whey (%), X 2 2 3 4

Laboratory scale bioreactor
The selected recombinant K. lactis strain was cultured in a Biostat-MD 2 L (Braun-Biotech) bioreactor with a 1 L working volume of the optimized production medium. The bioreactor was sterilized by autoclave at 121 °C for 30 min, inoculated at an initial OD 600 of 10, and aseptically supplemented with 300 mg/L of ampicillin. Culture temperature was 30 °C, air flow rate 2 L/min and stirring 300 rpm, conditions maintained during 120 h. Cell biomass was measured as dry weight from 5 mL culture samples, centrifuged at 5000 rpm for 5 min, washed with distilled water, centrifuged again and dried in oven at 105 °C until constant weight. The supernatant was used to measure extracellular α-galactosidase activity and the conversion of substrates into secondary metabolites. Sugar consumption and metabolites production was analyzed by HPLC using as external standard raffinose, sucrose, glucose, galactose, fructose, glycerol and ethanol at five concentrations ranging from 4 to 0.06 mg/mL.

Enzymatic activity assays
α-galactosidase and β-galactosidase activities were measured at 40 °C using as substrates 10 mM PNPG in McIlvaine buffer pH 4 [33] and 13 mM ONPG in Z buffer [34], respectively. Both substrates were supplied by Sigma Aldrich. At two consecutive time intervals, the reaction was stopped with 0.5 M Na 2 CO 3 and released p-nitrophenol and o-nitrophenol were quantified at 400 nm (ε = 18.2 L/mmol/cm) or 420 nm (ε = 4.5 L/mmol/cm), respectively. Invertase activity was measured with 100 mM sucrose as substrate in 50 mM acetate buffer pH 5 for 10 min at 40 °C, the reaction was stopped by boiling (95 °C, 5 min) and the increment in reducing sugars was quantified by the DNS method [35]. Assays were performed in triplicate and spectrophotometric readings were performed in flat bottom microtiter plates with Synergy H1 Hybrid Multi-Mode Reader (BioTEk). One enzymatic unit (U) is defined as the amount of enzyme that releases one μmol of product per min under the assay conditions.

HPLC analysis
A Sugar Pack Waters (6.5 mm × 300 mm) column with Refractive Index Detector (Cienytech) was used. The samples were clarified with the cartridges HyperSep Silica SPE Column (Fisher Scientific). Running conditions were, column temperature 80 °C, detector temperature 37 °C, sensitivity 32 and using 100 μM EDTA-Calcium (Sigma Aldrich) at a flow rate of 0.5 mL/min as mobile phase. Eluted compounds were identified and quantified using the corresponding external standards. To each sample and external standard 1 mg/mL sorbitol was added as internal standard [36].

Statistical analysis
Statistical data treatment and graphics drawing was performed with the program StatGraphics Plus version 5.1.
p-values were calculated with Student's t-tests and results were considered significant for p-values less than or equal to 0.05.

Construction of recombinant K. lactis strains
The expression cassettes, linearized by SacII digestion from the plasmids pKLACMEL1His-1 and pKLAC-MEL1His-2 ( Fig. 1), were successfully integrated in the genome of the K. lactis strains GG799 and Y1140. The integration occurred, by homologous recombination, in the promoter region of the LAC4 locus [37]. The amdS gene was used as selection marker, this gene encodes the enzyme acetamidase that allows the growth of the transformants in a minimal medium with acetamide as single nitrogen source. The transformants contained either one or several tandem copies of the MEL1 gene that were expressed under the LAC4 promoter, while the native LAC4 locus, that drives the expression of the enzyme β-galactosidase, was reconstructed (Fig. 2a). Single copy and multiple copy integrations were identified by PCR with the primers pairs P6/P5 and P7/P5 (Table 1), respectively. The frequency of multiple copy integration (50-60% of total transformed clones) was similar for the K. lactis strains GG799 and Y1140 (Table 3).  Table 3 shows the values of extracellular ScAGal activity measured from 96-well microtiter cultures in YPG medium, which allowed to classify the multiple copy transformants in two groups, probably with two or with more than two integrated copies of the MEL1 gene. The extracellular ScAGal activity of the strain GG799MEL1His-2 was fourfold (4 ± 0.465) higher than that of GG799MEL1His-1, whereas the extracellular ScAGal activity of the strain Y1140MEL1His-2 was 20-fold (20 ± 2.178) higher than that of Y1140MEL1His-1 ( Table 3). These results show that the extracellular ScA-Gal activity was notably higher when the endogenous signal sequence of S. cerevisae MEL1 was used (MEL1His-2 cassette) in comparison with the K. lactis α-MF domain sequence (MEL1His-1 cassette).

Heterologous expression of ScAGal
However, in all cases, GG799 was the strain that secreted higher levels of extracellular ScAGal activity regardless of the number of copies of MEL1 integrated into its genome. Although these results confirm the advantage of the pKLAC-1/Klactis GG799 expression system recommended by the manufacturer (K. lactis protein expression kit, New England Biolabs) vs another wild type strain, such as Y1140, we also demonstrated the interest of carrying out a study of the native signal peptide of the heterologous protein before being replaced by an exogenous signal peptide (K. lactis α-MF) to drive its secretion to the culture medium.
The multiple copy transformants of both strains with MEL1His-2 showing the highest extracellular activity were selected to analyze the expression of ScAGal at a higher scale in shake flask cultures (Fig. 3a). Higher values of extracellular ScAGal activity were obtained with GG799MEL1His-2 transformants (abbreviated KGM from now on) than with Y1140MEL1His-2 transformants (abbreviated KYM from now on), which is corroborated by the preliminary results shown above in the same medium. On the other hand, the activity of the native enzymes β-galactosidase and invertase was also measured from single copy transformants of KGM and KYM in comparison with the corresponding wild type strains (Fig. 3b). KGM expressed an average of β-galactosidase activity of 23 U/mL that was maintained until 144 h, while KYM obtained 15 U/mL decreasing abruptly in the same time range. Further, KGM reached the stationary phase at 48 h with a higher cellular biomass and both strains secreted around 30-35 U/mL of invertase. This result also shows the highest production of β-galactosidase by both wild strains, and suggests that integration of the heterologous gene in the LAC4 promoter negatively affects the expression of the LAC4 gene, as was confirmed in a chymosin producing strain vs wild strain using proteomic analysis [24]. For all these reasons, KGM was selected as the best K. lactis expression system of ScAGal for further research.

Copy number distribution of MEL1 in KGM multiple-copy strains
A total of 10 KGM strains were selected for qPCR analysis, eight bearing multi-copy integration and 2 bearing single-copy integration. The single copy strains KGM28 and KGM2, were used as calibrator to generate the standard curves and control of the calibrator, respectively. The vast majority of qPCR data in the literature is derived from the use of relative quantification by employing reference genes, which must be at a similar concentration to the target In order to get valid results [32,38]. The house-keeping gene TAF10, TATA Binding Protein (TBP) Associated Factor 10, was shown to be a reliable endogenous control. TAF10 is a single copy gene similar to S. cerevisiae YDR167W TAF10 subunit (145 kDa) of TFIID and SAGA complexes involved in RNA polymerase II transcription initiation and in chromatin modification [39]. Both for the gene MEL1 and the reference gene TAF10 they were not curve but straight lines (R 2 > 0.994) and showed a high amplification efficiency (Additional file 1: Figure S1A). The lineal regression analysis of the ΔCt variation between both genes versus the template DNA concentration gave a slope close to zero (0.0755), which confirms that amplification efficiencies are similar enough to validate the use of the 2 −ΔΔCt method [31] (Additional file 1: Figure  S1B). Since KGM28 contains one copy of the specific sequence of each gene, the copy number ratio MEL1/ TAF10 for the calibrator is 1 and, therefore, the ΔCt of the calibrator (1.98 ± 0.38) was employed to normalize the copy-number ratio MEL1/TAF10 for the analyzed strains. Table 4 shows the results of the relative quantification and the copy number estimation by the 2 −ΔΔCt method.
The number of copies of the MEL1His-2 gene integrated in the KGM strains was also analyzed by Southern blot using a specific MEL1 probe labelled with digoxigenin (Fig. 2b). In this analysis, the number of copies was calculated from the length of the hybridized DNA fragments (Fig. 2c). In addition, the southern blot analysis allowed to exclude the integration of further expression cassettes in other orientations or in other loci. Up to 4 copies of the MEL1His-2 gene were integrated in the KGM strains, as determined by qPCR (Table 4) and corroborated by Southern blot (Fig. 2c). KGM21 was the strain selected for further research, due to its highest number of integrated MEL1His-2 copies ( Table 4) and levels of ScAGal extracellular activity (Fig. 3a).

Optimization of a medium based on cheese whey and beet molasses for extracellular ScAGal production by the strain KGM21
With the aim to develop a sustainable and cheap culture medium for heterologous ScAGal extracellular production, while valuating both wastes of the food industry, mixtures of cheese whey and beet molasses were assayed. As shown above, the strain KGM21 expresses four copies of the gene encoding ScAGal under the LAC4 promoter that is inducible by lactose. Recombinant ScAGal allows the hydrolysis of raffinose and its use as carbon source for growth. This strain also produces native β-galactosidase and thus is able to use lactose as carbon source for growth. Moreover, KGM21 produces the native enzyme invertase, allowing the use of sucrose as carbon source for growth. HPLC analyses confirmed that lactose (99% w/v) is the major component of cheese whey, while sucrose (60% w/v) followed by raffinose (3% w/v) are the main components of beet molasses, in the lots used in this work.
The extracellular production of ScAGal by KGM21 in media with pure lactose and whey, YPL and YPW respectively, both with a 2% final lactose concentration, was compared. The levels of ScAGal were more than double in the whey medium (3.7 ± 0.15 U/mL) than in the lactose medium (1.4 ± 0.25 U/mL), during 144 h of culture (Additional file 2: Figure S2). In addition, YPW with a 4% final lactose concentration allowed Fig. 3 Time course of cultures from the strains KGM and KYM in shaken flasks using YPG medium. a Extracellular α-galactosidase activity of selected multicopy transformants. b Typical profile of growth (circles), native β-galactosidase (squares) and invertase (triangles) of wild type (empty) versus single copy strains (full). KGM28 and KYM7 are single copy transformants with MEL1His-2. Data shown are average ± SD, N = 3 a considerable increase of ScAGal in the same time assayed (7.5 U/mL). Therefore, cheese whey proved to be a good substrate for ScAGal production by KGM21.
Then, beet molasses were assayed as an additional carbon source for ScAGal production by KGM21, mixed with cheese whey, by means of an experimental factorial design 3 2 that was composed of 12 experiments with three replicates of the central point (Table 5). We selected, as values of the two variables at the central point of the design, a 3% of whey based on the data observed above and a 7.5% of molasses because Kluyveromyces sp. at higher concentrations may exhibit catabolite repression in these mixtures and not be able to consume lactose [20]. Since nitrogen fraction of whey and molasses is low [15], the culture media were supplemented with yeast extract. Commercial yeast extract can be replaced by an autolysate of the yeast biomass grown in the cultures [40], and thus this supplement does not represent an economic load. Cultures were inoculated at a DO 600 = 10 (1 mg dry matter/mL) to favor the consumption of all available sugars.
The results of extracellular ScAGal activity obtained of each experiment and predicted from RSM are shown in Table 5. Multiple regression analysis was performed to fit the response function to the experimental data, which resulted in the following quadratic polynomial equation: Y = − 57.09 + 15.37x 1 + 5.92x 2 − 1.019x 1 x 1 − 0.357x 1 x 2 − 0 .549x 2 x 2 , where Y is the extracellular ScAGal activity (U/  mL) and x 1 and x 2 are the coded values for the concentrations of beet molasses and cheese whey, respectively. The R 2 statistic indicated that the model explains the 95.51% of the variability of the response (Y) and the rest of the total variance is attributed to deviations different from the experimental factors. The "lack of fit" was not statistically significant (p value = 0.338), which confirms that the model is suitable to predict the extracellular production of ScAGal into the experimental domain used (Additional file 3: Table S1). The Pareto chart of effects of each independent variable on the response-extracellular ScAGal activity shows that the concentration of beet molasses was the highest negative effect on the response (Fig. 4a). A good correlation between the experimental results and the values predicted from the model was confirmed (Fig. 4b), so that the statistically not significant effects were not excluded. Also, it is observed an inflection point of extracellular ScAGal activity, from which and to both sides, it decreases when molasses concentration varies up or down, with independence of whey concentration (Fig. 4c). Finally, the response surface graphical representation and its corresponding contour plot allowed to determine the optimal point, the conditions that yield the highest value of the response (Fig. 4d). In this case, the theoretical optimal conditions were found close to the central point of the experimental domain: 7% beet molasses and 3% cheese whey, yielding a value predicted from the model of 5.95 U/mL of extracellular ScAGal activity.

Scale-up of extracellular ScAGal production by the strain KGM21
Using the culture medium optimized as described above, the production of extracellular ScAGal by KGM21 was scale-up to shake flask and bioreactor levels. The substrates present in the optimized medium were 74 g/L sucrose, 13 g/L fructose, 12 g/L glucose, 10 g/L lactose, 6 g/L raffinose and 1 g/L galactose, which were identified and quantified by HPLC. Figure 5a shows that in shake flask cultures, extracellular ScAGal production increased in parallel to cellular growth up to reaching the stationary phase at 72 h, the average enzyme activity value of 5.83 U/mL was maintained during the stationary phase and is concordant with the model prediction. Figure 5b shows that in bioreactor cultures KGM21 arrived at the stationary phase in 48 h showing an extracellular ScA-Gal activity value of 10 U/mL that continued increasing until 19 U/mL during the stationary phase, which is more than threefold higher than the optimal value predicted by the model, and 4.6-fold higher than the activity obtained using a whey medium without molasses, at the same culture time in shake flasks. From the beginning of the cultures, the expression is induced of recombinant ScAGal and native β-galactosidase and invertase, which allows the hydrolysis of raffinose, lactose and sucrose, respectively. The conversion of the sugar substrates during the bioreactor cultures was analyzed by HPLC and confirmed that the strain KGM21 consumes the 79.5% (w/v) raffinose and 98% (w/v) of the sum of lactose and sucrose. At the same time, 22.4 g/L ethanol was produced by fermentation and was then almost all consumed as carbon source, which is a behavior characteristic of the respirefermentative metabolism of K. lactis [41]. A maximum peak of ethanol production appears at 24 h of culture and is quickly consumed without inhibiting the growth of the strain or the ScAGal production (Fig. 5b). Although it is described that, in industrial yeasts, glucose and fructose repress the metabolism of other less preferred sugars [29], we observed that KGM21 efficiently used the sugars in the mixture of whey and molasses and produced recombinant ScAGal. It has been reported that a K. lactis mutant in the gene GAL1, shows a decreased galactose consumption and, in consequence, an increased availability of this monosaccharide to induce the LAC4 promoter [24]. A possibility for further research may be to test the hypothesis that this mutation increases ScAGal expression in KGM21.
Since the recombinant ScAGal is secreted to the extracellular medium, it has the advantage that it can be separated from the impurities of the production medium through tangential filtration membranes. In previous studies, we have successfully confirmed, using SDS-PAGE analysis, the concentration and purification of the recombinant protein with the tangential flow filtration technique (TFF, Millipore), suggesting that it was an efficient and effective process to avoid the use of purification systems that make the industrial production process more expensive [42]. In addition, we modified ScAGal to create a fusion protein with a poly-His tag to obtain a higher purity final product by affinity chromatography, if necessary.
On the other hand, the parameters of productivity (Qp) and yield from substrate of ScAGal (Y p/s ) calculated from the cultures in YPL, YPW and the optimized wheymolasses medium are shown in Table 6. For the bioreactor culture, Q p was 0.158 U/L/h, notably higher than the other shake flask cultures and Y p/s was similar to the culture in shake flasks with YPW and also higher than the other cultures. The improved productivity in bioreactors is a common trait, generally attributed [4] to the fact that bioreactors are closed systems with a stricter control of aeration, agitation and temperature than other systems. Moreover, the mechanical agitation of bioreactors, compared to the orbital agitation of shake flasks, favors the continuous supply of oxygen to the cells and in consequence growth and bioproductions.
The proposed medium, based on whey and beet molasses, allowed the environmental innovation of the production process of ScAGal by K. lactis strains and, in turn, this could possibly be used in the heterologous production of other proteins with similar features to the object of study.

Conclusions
We have engineered the K. lactis KGM21 strain harboring up to four copies of the gene coding for ScAGal (with its endogenous signal peptide) integrated at the LAC4 chromosomal locus. An optimized mixture of milk whey and beet molasses turned out to be a cheap, sustainable and highly recommendable culture medium for high yield heterologous ScAGal production, directed to the extracellular medium by KGM21, thus allowing the valuation and ensuing reduction of the polluting impact of these agro-industrial wastes.