Mechanism of lactose assimilation in microalgae for the bioremediation of dairy processing side-streams and co-production of valuable food products

This study investigated the mechanism of lactose assimilation in Nannochloropsis oceanica for dairy-wastewater bioremediation and co-production of valuable feed/food ingredients in a circular dairy system (β-galactosidase and omega-3 polyunsaturated fatty acids). Mixotrophic cultivation was found to be mandatory for lactose assimilation in N. oceanica, with biomass production in mixotrophic cultures reaching a fourfold increase over that under heterotrophic conditions. Under mixotrophic conditions, the microalgae were able to produce β-galactosidase enzyme to hydrolyse lactose, with maximum extracellular secretion recorded on day 8 of growth cycle at 41.47 ± 0.33 U gbiomass−1. No increase in the concentration of glucose or galactose was observed in the medium, confirming the ability of microalgae to indiscriminately absorb the resultant monosaccharides derived from lactose breakdown. Population analysis revealed that microalgae cells were able to maintain dominance in the mixotrophic culture, with bacteria accounting for < 12% of biomass. On the other hand, under heterotrophic conditions, native bacteria took over the culture (occupying over 95% of total biomass). The bacteria, however, were also unable to effectively assimilate lactose, resulting in limited biomass increase and negligible production of extracellular β-galactosidase. Results from the study indicate that N. oceanica can be effectively applied for onsite dairy wastewater treatment under strict mixotrophic conditions. This is commercially disadvantageous as it rules out the possibility of deploying heterotrophic fermentation with low-cost bioreactors and smaller areal footprint.


Introduction
Dairy processing and cheese production generate significant amount of wastewater and side products which are rich in lactose (up to 80 wt% of dissolved solids), other carbonbased nutrients (e.g., proteins and lipids) and inorganic nutrients (e.g., nitrates, phosphates and micronutrients) (Slavov 2017;Daneshvar et al. 2019). These waste products pose significant environmental problems due to high organic matter contents, BOD level and COD level (Gramegna et al. 2020), and are generally subjected to conventional wastewater treatment requiring large amounts of chemicals and energy prior to disposal. Wastewater treatment represents a significant economic burden to dairy manufacturers and successful valorisation of dairy waste streams into viable and profitable new co-products is of critical importance (Espinosa-Gonzalez et al. 2014). In the present, however, attempts to commercially use dairy side streams (e.g., whey permeate) have not been highly profitable, with the permeate either sold directly as low-value livestock feed without significant profit or converted into products (e.g., lactose powder or alcoholic beverages) with relatively low market values.
Microalgae can serve as an attractive valorisation strategy for dairy by-products. Dairy side streams, such as whey permeate, can be used as a low-cost carbon and nutrient-rich medium in mixotrophic and heterotrophic cultivation of microalgae (Zimermann et al. 2020;Bentahar and Deschênes 2022). The bioremediation of dairy streams by algal cells will neutralise nutrients in a sustainable manner (compared to chemicalintensive wastewater treatment) while simultaneously generate biomass which can be commercially exploited for food and bioactive compounds (Espinosa-Gonzalez et al. 2014;Bentahar et al. 2019a). Species belonging to the Nannochloropsis genus are particularly attractive for dairy-effluent valorisation because of their high content of omega-3 polyunsaturated fatty acids (ω-3 PUFAs) in the form of eicosapentaenoic acid or EPA (up to ca. 10 wt% of biomass) (Hulatt et al. 2017;Halim et al. 2019a). ω-3 PUFAs are ubiquitously used in the food and feed industry (including dairy industry) as nutrient enrichment to promote immunity and brain development (e.g., in infant formulas) (Kolanowski 2021). A sustainable alternative source of ω-3 PUFAs is, however, urgently needed as commercial production currently relies on fish oil derived from wild-caught fish and places enormous strain on the oceanic fish stock (Poliner et al. 2018). While alternative docosahexaenoic acid (DHA) from Schizochytrium is already commercially available (Sahin et al. 2018;Yin et al. 2019;Półbrat et al. 2021), there is still no commercial source for EPA not derived from fishing. Nannochloropsis can therefore exert dual environmental impact for the dairy processing industry, providing a more sustainable means for wastewater treatment while producing alternative EPA that can then be recycled within the industry in a circular manner to reduce demand for conventional fish oil.
The application of Nannochloropsis in treating dairy waste can also lead to the production of β-galactosidase (or lactase), a key enzyme for the synthesis of galactooligosaccharides (GOS) prebiotic compounds in infant formulas (Suwal et al. 2019), thus adding an extra layer of commercial benefits to the dairy industry. Despite these potential benefits, only few studies have explored the use of Nannochloropsis for the bioremediation of dairy waste (Zanette et al. 2019;Kiani et al. 2023).
Fundamental understanding of the mechanism responsible for lactose assimilation in microalgae is critical to the design and optimisation of microalgae systems for dairy-waste valorisation and the transfer of results from bench-scale to industrial scale. Lactose is a disaccharide composed of glucose and galactose sub-units. It is structurally complex and will likely require the microalgae cells to produce β-galactosidase enzyme to hydrolyse the molecule to monosaccharide sub-units prior to metabolization, resulting in a more intricate assimilation process compared to monosaccharides (such as glucose). Yet studies evaluating the use of microalgae for dairy-waste bioremediation have overlooked this aspect, directly assessing the performance of microalgae in removing nutrients from dairy wastewater without having any understanding of how lactose availability and breakdown can affect the bioremediation kinetics of other nutrients (such as phosphates and nitrates). This has resulted in inconsistent performance reported for microalgae treatment of dairy waste, with wide-ranging nutrient removal efficiency from 30 to 100% (Daneshvar et al. 2018;Biswas et al. 2021;Kiani et al. 2023).
Amongst the few studies to-date that have investigated microalgae utilization of lactose (Suwal et al. 2019;Zanette et al. 2019;Bentahar and Deschênes 2022), to the best of our knowledge, only one study has investigated the mechanism of lactose assimilation in Nannochloropsis. The capacity of Nannochloropsis to assimilate lactose under heterotrophic regimes has never been explored. Heterotrophic fermentation has several key advantages compared to an autotrophic system in the context of dairy waste bioremediation; it eliminates dependence on light irradiation which can often be limiting in northern Europe and reduces areal requirements through simplification of bioreactor design. Standard photobioreactors used for autotrophic/mixotrophic cultivation are constructed with large surface area to volume ratio to enable light penetration, naturally requiring more land footprint compared to heterotrophic fermenters (Tian-Yuan et al. 2014;da Silva et al. 2021). Commercial dairy manufacturers generally have limited space to expand into and cannot afford to implement on-site wastewater solutions requiring large areal footprint.
In addition, the interaction between native bacteria (defined as symbiotic/associated bacteria present in the algal culture since inoculation) and Nannochloropsis cells during mixotrophic and heterotrophic cultivation and their effect on carbon assimilation has not been elucidated, bringing uncertainty to the behaviour of culture under non-axenic conditions encountered in large-scale wastewater treatment. Finally, the dependence of β-galactosidase production in Nannochloropsis on growth conditions and stages of cultivation is not well understood.
This study aims to shed light on the mechanism of lactose assimilation in N. oceanica to harness its potential for dairy-effluent bioremediation and co-generation of valuable EPA and β-galactosidase. The study plans to derive three key insights regarding N. oceanica's lactose assimilation: 1) the capacity of the species to assimilate lactose under both mixotrophic and heterotrophic conditions, 2) the interaction of N.oceanica cells with native bacteria and its effect on lactose assimilation under different cultivation regimes and 3) the extent of β-galactosidase production and the dependence of its kinetics on lactose concentration. To achieve these objectives, the study cultivated N. oceanica on lactose-rich standard solution of different initial lactose concentrations under both mixotrophic and heterotrophic conditions and compared their biomass production and lactose assimilation. The levels of extracellular β-galactosidase secretion as well as lactose breakdown products (i.e., glucose and galactose) in the medium of different cultures were monitored to study the kinetics of β-galactosidase evolution and determine the mechanism of lactose assimilation. Flow cytometry was employed throughout the study to investigate the cell population dynamics in the medium and quantitatively measure the level of bacteria in the cultures, allowing for the determination of their roles in β-galactosidase synthesis and lactose assimilation. Finally, carbon mass balance around the different cultures was conducted to understand the effect of cultivation modes on microalgal carbon metabolism.

Microalgae strain and culture conditions
Microalgae strain was obtained from the Culture Collection of Algae and Protozoa (CCAP, Scottish Association for Marine Science, Oban, Scotland, U.K.). Nannochloropsis oceanica (CCAP849/10 isolated from operational hatchery western Norway; non-axenic stock culture) was assessed in this study. Mixotrophic and heterotrophic cultivation of N. oceanica was performed with f/2 medium (Guillard and Ryther 1962) in synthetic seawater (Red Sea Coral Pro Salt, Rea Sea, USA). Lactose solutions (α-Lactose monohydrate, CAS No. 5989-81-1, Merck KGaA, Germany) were separately prepared, autoclaved, and added to the f/2 medium previously prepared under sterile conditions to form culture media with 10 g L −1 (or 1% (w/v)) and 5 g L −1 (or 0.5% (w/v)) initial lactose concentration. Cultures without sugar addition were used as a control group (i.e., autotrophic control and dark control), leading to a total of six independent experiments ( Table 1). The autotrophic and mixotrophic samples were cultivated in 250 mL Erlenmeyer flasks with a 150 mL working volume under constant orbital shaking (130 rpm) on a 12 h:12 h light:dark cycle at ambient temperature (22 °C ± 1 °C) without additional aeration for 30 days. Cotton caps were placed on each flask to reduce contamination and evaporation of the growth medium during the cultivation period. The dark-control and heterotrophic samples were cultivated under the same conditions with no illumination. All cultivations were performed in duplicate.

Biomass production and growth parameters
Optical density (OD) of each culture was measured at 750 nm using UV-Vis spectrophotometer (Spectramax M3, Molecular devices, Berkshire, UK) every two days. The OD 750 values were then converted to biomass concentration to monitor the growth of microalgae using a previously determined OD 750 -biomass concentration calibration curve for the strain. OD measurement for each collected sample was carried out in duplicate.
The specific growth rate (µ, day −1 ) was calculated using an equation between biomass concentration (Y, g L −1 ) and cultivation time (t, day) based on the model of Zwietering et al. (1990).
where Y m is the maximum biomass concentration (g L −1 ), µ m is the maximum specific growth rate (day −1 ) and the λ is lag time (day). All curve-fitting calculations were performed using Algal Data Analyser (ADA) software (Mapstone et al. 2022).

Chemical analysis
Ten mL of supernatant was collected from each culture every two days by centrifugation at 4000 rpm for 10 min, which was used for the following analyses. Colorimetric methods were employed to quantify the nitrate and phosphate concentrations in the supernatant using commercial kits according to suppliers' instructions. The analysis kits for nitrate and phosphate were obtained from API (Mars Fishcare, USA) and Red Sea (Seahorse, Ireland), respectively. Chemical analyses for each collected sample were carried out in duplicate.
where C N and C P are the nitrate and phosphate concentrations (ppm) at time t, and C N0 and C P0 are the initial concentration of the nitrate and phosphate in the culture.

HPLC analysis
The concentrations of lactose, glucose and galactose in the supernatants were determined by high-performance liquid chromatography (HPLC). The chromatographic separation was achieved using an Agilent ZORBAX NH 2 (250 mm × 4.6 mm containing 5 μm silica particles) column in an Agilent Technologies 1200 Series equipped with a refractive index detector (G1362A RID) at 35 °C. A mixture of acetonitrile and 0.04% ammonium hydroxide in water (70:30, v/v) at pH 6.92 ± 0.05 was employed for isocratic elution at a flow rate of 1.5 mL min −1 (Correia et al. 2014). The running time of each sample was 11 min with a 10 µL injection volume and a column temperature at 30 °C. The sugars were identified by retention time and quantified by comparing their peak areas to a linear calibration curve of known standards.

β-Galactosidase activity assay
The β-galactosidase activity was measured using o-nitrophenyl-β-D-galactopyranoside (ONPG) as a substrate based on the method reported by National Research Council (1981), with some modifications as described by Suwal et al. (2019) and Zanette et al. (2019). ONPG solution (4 mg mL −1 ) was prepared in 0.1 M sodium phosphate buffer (same pH as the supernatant). Then 1 mL of supernatant was added into 2 mL of ONPG solution and incubated at 30 °C for 15 min. The reaction was stopped by adding 1 mL of 1 M Na 2 CO 3 . The absorbance against distilled water at 420 nm was recorded. The enzyme activity, expressed as specific activity (U g biomass −1 ), was calculated based on a standard curve. One unit of β-galactosidase activity (U) was defined as the amount of enzyme that liberated 1 μmol of o-nitrophenol per minute. The measurement of β-Galactosidase activity for each collected sample was carried out in duplicate.

Flow cytometry
The culture was sampled every seven days for flow cytometry analysis. The sample (2 mL) was first harvested by centrifugation at 4000 rpm for 10 min and the supernatant carefully removed. Then 4 mL Phosphate Buffered Saline (PBS, pH 7.4) was added to the cells and mixed thoroughly using a vortex mixer. DAPI (4′,6-diamidino-2-phenylindole) was used to stain the nucleus of both microalgae and bacteria cells in the microalgal suspension. The dye was first dissolved in anhydrous dimethyl sulfoxide (DMSO) at 10 μL mL −1 to enhance the staining efficacy for nucleus fluorescence determination (Doan and Obbard 2011). The dye solution (100 μL) was then added to 1 mL algal suspension, vortexed for 10 s, and incubated at ambient temperature for 5 min to make sure the stain went into the cells. After being thoroughly vortexed, the fluorescence intensity of the suspension was analysed using CytoFLEX LX flow cytometer. Fluorescence measurements were carried out after staining by excitation at 450 nm for DAPI. In addition, forward scatter (FSC) and side scatter (SSC) were measured as indicators of the diameter (or size) and the density of the cell, respectively. The data were subsequently analysed with CytoFLEX software to determine the algal cell population dynamics. Algal and bacteria biomass concentrations were calculated as in Eq. 4 and 5 respectively.
where Y microalgae and Y Bacteria are the contribution to biomass concentrations in the culture by algal cell and by bacteria cells, respectively (g L −1 ), N is algal cell count (cells mL −1 ) detected by flow cytometer, FSC and SSC are forward scatter median and side scatter median of algal cells, respectively. The subscript A represents results of autotrophic cultures. This calculation is based on a hypothesis that only algal cells contributed to biomass concentration in autotrophic cultures.

Carbon flow
Carbon mass balance around each cultivation system was determined in this study. The quantity of carbon stored in both microalgae and bacteria biomass (M C-biomass , g C L −1 ) and carbon assimilated from lactose (M C-lactose, g C L −1 ) were calculated as in Eq. 6 and 7.
where η B (η B = 50.06%) and η L (η L = 42.11%) are the percentage of carbon content of algal-bacteria consortium biomass (Lawford and Rousseau 1996;Chen et al. 2015) and lactose respectively, C L is the lactose concentration, C L0 is the initial lactose concentration of the culture.
Then the carbon mass balance was calculated using Eq. 8.
where M C-capture is the mass of carbon captured during cultivation period. M C-capture represents the difference between the mass of carbon integrated into the microalgal biomass by photosynthesis and the mass of carbon released by microalgal and bacterial cells through aerobic respiration. M C-capture > 0 indicates a net carbon capture via photosynthesis, while M C-capture < 0 indicates a net carbon release by aerobic respiration.

Statistical analysis
Data in this study are expressed as mean ± standard deviation (SD) of a total of four measurements (2 biological replicates × 2 analytical replicates). The statistical analysis was carried out using SPSS Statistics (version 27.0.1, IBM). Statistical significance (p < 0.05) between the results of selected cultures was assessed by one-way analysis of variance (ANOVA) using Post Hoc Tukey's test.

Capacity to grow under mixotrophic or heterotrophic conditions
The actual final biomass concentration, maximum biomass concentration, maximum specific growth rate and lag time for each cultivation are shown in Fig. 1 and Table 1. Nannochloropsis oceanica cultivated mixotrophically with 1% (w/v) initial lactose concentration presented the highest final biomass concentration of 0.71 g L −1 at the end of cultivation period, followed by the culture with 0.5% (w/v) initial lactose concentration at 0.56 ± 0.01 g L −1 in mixotrophic group. This biomass concentration was 4 -4.5 higher than those obtained in the heterotrophic group, which experienced only modest increases in biomass concentration (ca. 0.15 g L −1 regardless of the lactose concentration). The dark-control group understandably experienced almost no growth given that it lacked both light for photosynthesis and organic carbon supply. Previous studies (Espinosa-Gonzalez et al. 2014;Hemalatha et al. 2019;Zanette et al. 2019) have reported a higher final biomass concentration for mixotrophic conditions, which could be probably attributed to higher light intensity and more effective agitation and aeration systems. Nevertheless, it was obvious that lactose was able to enhance algal growth relative to autotrophic system, and the mixotrophic cultivation using lactose showed a strong potential to obtain high biomass concentration. Overall, the results suggested a photosynthetic dependence of lactose assimilation in N. oceanica, where the species appeared to only be able to assimilate lactose under mixotrophic rather than heterotrophic conditions.

Algae-bacteria population dynamics
Flow cytometry with DAPI nucleus fluorescence staining was used to identify the population dynamics in algal suspensions. Both microalgae cells and bacterial cells were stained by DAPI, while only the microalgae cells emitted autofluorescence. Microalgal cell concentration was therefore quantified based on autofluorescence cell count. In the mixotrophic cultures, algal cell count displayed the same increasing pattern as biomass concentration. Both FSC median and SSC median also showed a significant increase by the end of cultivation ( Fig. 2a and b). The shift of FSC median in the mixotrophic culture with 1% (w/v) initial lactose concentration can be observed in Fig. 2d. These results indicated that both microalgae cell concentration and cell size increased during the mixotrophic cultivation, and that the presence of lactose resulted in a positive impact on algae  growth, which was in agreement with not only biomass concentration results presented in previous section based on OD measurements but also other studies investigating lactose assimilation in microalgae (Choi et al. 2018;Suwal et al. 2019;Zanette et al. 2019).
On the other hand, a reduction in microalgae cell counts could be observed in the heterotrophic cultures (Fig. 2c). This was in contrast to the results obtained in previous section which showed a modest increase in biomass concentration in the heterotrophic cultures. This discrepancy can likely be attributed to the proliferation of another set of microbial populations (i.e., bacteria) under heterotrophic conditions which contributed to the OD increase in the culture and thus led to the erroneous impression of algae growth in the heterotrophic system. The flow cytometer results were hence further processed to quantify the level of bacteria in the culture. The concentration of bacterial cells could not be directly quantified by flow cytometry under the current configurations due to the fact that bacterial cells did not emit autofluorescence and DAPI staining alone led to a high background noise that prevented quantification by particle (or cell) counts. Instead, we indirectly deduced bacterial mass contribution in the culture by using autofluorescence measurement of algal cells (from flow cytometry) as the basis of microalgae contribution to the overall biomass concentration in the culture and subtracting such contribution from the measurement of total biomass concentration to determine the mass proportion of bacteria in the culture. Equation 4 normalised the algal cell count in the culture (e.g. mixotrophic or heterotrophic systems) to that obtained in the autotrophic system based on autofluorescence flow cytometry and used the resulting ratio to calculate the algae biomass concentration in the culture. The bacterial biomass concentration was then calculated by subtracting the value of algae biomass concentration from the overall biomass concentration in the culture (Eq. 5). As shown in Fig. 3, the increase in the population of associated bacteria was observed in both mixotrophic and heterotrophic cultures. Cultures grown heterotrophically, however, showed significantly higher bacteria concentration after 3 weeks of cultivation, reaching 0.14 -0.15 g L −1 in concentration. Bacteria emerged as the dominant population in the cultures, occupying ca. 96% of total biomass in heterotrophic cultures. The biomass concentration increase observed in the heterotrophic culture (previous section) can therefore almost entirely be attributed to the proliferation of bacteria rather than microalgae growth. The analysis in Fig. 3 also revealed that bacteria made up only < 12% of biomass in mixotrophic cultures and that microalgae cells were able to suppress bacterial growth and maintain dominance in these cultures. The algal cell counts attained by the mixotrophic cultures (1.85 × 10 7 to 2.00 × 10 7 cells mL −1 ) after 30-day cultivation were ca. 35-fold greater than those obtained under heterotrophic conditions (ca. 6.00 × 10 5 cells mL −1 ) ( Table 1). The results are well aligned with biomass concentration measurements in previous section showing significantly higher biomass growth in the mixotrophic cultures than in the heterotrophic cultures. N. oceanica thus appeared to be able to utilize lactose only

Lactose assimilation
The reduction of lactose and other nutrients in culture medium at the end of cultivation cycle is shown in Table 2. Lactose metabolization was observed in both 1% (w/v) and 0.5% (w/v) mixotrophic cultures, which showed 57.2% and 37.4% reduction in lactose concentration by day 30, respectively (Table 2, Fig. 4c). Glucose and galactose concentrations remained below 0.1 g L −1 at all time points across all mixotrophic cultures, which was in agreement with the study of Zanette et al. (2019). The results supported our findings in previous sections and confirmed that the cells in N. oceanica cultures were able to assimilate lactose under mixotrophic conditions. Despite this, the pathways which the cells utilized to assimilate lactose is currently unknown. We speculate that the cells can either absorb lactose directly from the medium and then break the disaccharides down to glucose and galactose using intracellular β-galactosidase or they can release extracellular β-galactosidase to hydrolyse lactose in the medium into glucose and galactose before subsequently metabolising the resulting monosaccharides or undergo a combination of the two pathways. Assaying the medium for extracellular β-galactosidase activities can provide further insights into lactose assimilation. Regardless, the fact that there was no monosaccharide accumulation observed at any time point across both mixotrophic cultures was suggestive of the fact that Nannochloropsis cells had the capacity to indiscriminately assimilate both glucose and galactose molecules derived from lactose hydrolysis (Girard et al. 2014).
On the other hand, the heterotrophic cultures experienced minimum lactose reduction (ca. < 10% reduction). This supported the initial observation in the previous sections that N. oceanica were unable to activate lactose assimilation pathways in darkness. Interestingly, since the heterotrophic cultures have been overtaken by native bacteria (as evidenced

β-Galactosidase production and kinetics
In order to better understand the underlying pathways associated with lactose assimilation in N. oceanica cultures, extracellular β-galactosidase activity in the medium was assayed every four days. The ability of N. oceanica to produce extracellular β-galactosidase to process lactose when cultivated under mixotrophic conditions were demonstrated in Fig. 5, with the highest specific enzymatic activity found in cultures with 0.5% (w/v) initial lactose concentration on day 8 at 41.47 ± 0.33 U g biomass −1 . The change in specific enzymatic activity in mixotrophic cultures over time followed the same pattern in both mixotrophic experiments, showing a rapid rise at the early stage of cultivation before reaching a peak on day 8 and experiencing a sharp decline to day 14. A low level of β-galactosidase activity persisted in the last two weeks of the cultivation (less than 0.5 U g biomass −1 on day 22). Similar findings were reported by Suwal et al. (2019), who observed β-galactosidase activity in Tetradesmus obliquus when treating whey permeate to reach a plateau before decreasing by half at the end of cultivation. The kinetics of β-galactosidase activity observed in present study could be attributed to the different stages in the microalgae growth cycle (Fig. 1). The initial increase in activity (day 4 -8) can be correlated to the transition from lag phase to exponential phase where cells began to undergo rapid cell division and were hungry for 'carbon' to support growth.
The subsequent decrease in lactase activity (day 8 -16) corresponded to the mid-point in the exponential growth phase where the cells began to experience declining growth rate in anticipation of nitrogen depletion (~ day 12) and the cessation of cell division. The results were in agreement with the study conducted by Bentahar et al. (2019b), indicating a higher enzyme productions in their mixotrophic cultures of T. obliquus after 8 cultivation days and that biomass productivity and enzyme productions were related to the trophic stage. Moreover, a similar behaviour could be observed in yeast cells. Studies conducted by Branco et al. (2020) and Chniti et al. (2017) indicated that young and budding yeast cells showed a higher metabolic rate on sugar consumption for producing ethanol, while maturated cells only consumed smaller amounts of sugar for maintenance. Through these findings, this study showed that cultivation time played an important role in regulating β-galactosidase activity in N. oceanica. Future industrial application planning to use the species for the production of β-galactosidase enzyme will thus have to optimise their harvest time (just before the start of stationary phase) to target maximum enzyme activity.
In contrast, no extracellular β-galactosidase production was observed in heterotrophic cultures, thus validating our initial observation that a) N. oceanica were unable to activate lactose assimilation pathways in darkness and b) the native bacteria that have emerged as dominant population in the culture also lacked the ability to produce extracellular β-galactosidase. The absence of β-galactosidase in the bacteria-dominated heterotrophic cultures served to reinforce the endogenous nature of enzyme production by microalgae in the mixotrophic cultures, ruling out the possibility that the β-galactosidase detected in the mixotrophic medium was derived from bacteria cells or microalgae-bacteria interactions.

Removal of other nutrients
Nitrate and phosphate removals were studied to understand the effect of lactose assimilation on bioremediation performance. Apart from the dark-control group, all of the other cultures were able to completely deplete nitrate and remove 94 -97% of phosphate in their medium, regardless of autotrophy, mixotrophy or heterotrophy (Fig. 4b, c, Table 2). The rate of nitrate and phosphate removal appeared identical for both mixotrophic and autotrophic cultures, despite the higher growth rates observed in the mixotrophic cultures, indicating that the cultures were likely nitrate-and phosphate-limited (Halim et al. 2019b). In comparison with the nitrate removal in mixotrophic cultures, nitrate content was reduced more slowly in the heterotrophic culture and dark-control group. This was expected as these cultures exhibited significantly lower rate of biomass concentration increase relative to their light-experiment counterparts. Since bacteria have overtaken microalgae as dominant population in the heterotrophic cultures, the reduction in nitrate and phosphate concentration in these cultures can likely be ascribed to bacterial utilization (rather than microalgal assimilation). Lactose, nitrate and phosphate are common components of dairy side streams; the high removal efficiency for these nutrients as reported in this study alludes to the promising nature of N. oceanica mixotrophic cultivation as a promising green strategy for valorisation of dairy side streams.

Carbon balance
To further probe microalgal carbon metabolism during mixotrophic and heterotrophic cultivation, carbon mass balance was carried out. The total mass of carbon stored in algal biomass, the mass of carbon assimilated from lactose and the net mass of carbon released or consumed as atmospheric carbon dioxide were shown in Table 3. A net carbon sequestration of 0.23 ± 0.00 g C L −1 was observed in autotrophic cultures. This was expected as the accumulation of carbon in algal biomass in autotrophic cultivation was solely dependent on the difference between CO 2 fixation during photosynthesis and respiratory CO 2 release. In contrast to autotrophic cultures, carbon flow in the other regimes (i.e., dark-control, mixotrophic and heterotrophic cultivations) resulted in a net CO 2 release rather than CO 2 fixation. The highest net CO 2 release was found in mixotrophic culture at 1% (w/v) initial lactose concentration (1.75 ± 0.07 g C L −1 ). The apparent net CO 2 release in mixotrophic cultures indicated that, in the presence of lactose, N. oceanica preferred to obtain carbon through lactose assimilation rather than photosynthesis. This reduced the amount of carbon sequestered via photosynthesis and led to the rate of carbon released by aerobic respiration to exceed that sequestered by photosynthesis. The preferential uptake of dissolved organic carbon over inorganic carbon has also previously been reported in other studies investigating mixotrophy in microalgae (Smith et al. 2015;Zhang et al. 2021) and can be attributed to the low concentration of atmospheric CO 2 present in the air (0.04%) and the associated limitation of diffusive mass transfer into aqueous solution.
In the heterotrophic cultures the dominant carbon metabolism was aerobic respiration by both microalgae (due to the lack of algal photosynthesis) and associated bacteria leading to a net CO 2 release. The inability of N. oceanica and associated bacteria to activate extracellular lactose assimilation in the dark, however, meant that there was limited biomass growth and thus minimal net CO 2 evolution in the heterotrophic cultures (between 0.07 -0.40 g C L −1 ).

Discussion
Findings from previous sections enabled us to develop a working model for N. oceanica lactose assimilation and β-galactosidase production under both mixotrophic and heterotrophic conditions. Under mixotrophic conditions, the cells were able to activate lactose assimilation pathways where they secreted extracellular β-galactosidase into the medium to hydrolyse lactose (up to 57% of available lactose) into glucose and galactose. β-Galactosidase production reached a peak at mid-logarithmic phase (day 8 of cultivation at 41.47 ± 0.33 U g biomass −1 ) before experiencing significant decline, indicating that future industrial application planning to use N. oceanica for β-galactosidase synthesis should aim for a short growth cycle. The cells then rapidly absorbed the resulting glucose and galactose sub-units, likely integrating both monosaccharides into standard glycolytic pathways and tricarboxylic acid (TCA) cycle to fuel further energy and biomass generation. No accumulation of either glucose or galactose was observed in the medium, indicating that the cells were able to indiscriminately uptake both monosaccharides. The metabolization of lactose led to rapid microalgal cell division which depleted all available nitrates and > 95% available phosphates in the medium and resulted in a homogenous culture where microalgae cells remained dominant (> 88% of total biomass). The mixotrophic pathways however came at the expense of photosynthetic efficiency, resulting in a net release of CO 2 to the atmosphere.
Lactose assimilation in N. oceanica was found to be mixotrophic in nature, with the species being unable to secrete any β-galactosidase into the medium under heterotrophic conditions. Heterotrophic conditions led to a decline in the number of microalgae cells, in turn allowing native bacterial cells to take over the culture and emerge as the dominant population (over 95% of total biomass). Interestingly, the native bacterial population also lacked the ability to secrete extracellular β-galactosidase and assimilate lactose in Table 3 The mass of carbon involved in biomass production and lactose assimilation. The difference between mass of carbon assimilated into the biomass by photosynthesis and mass of carbon released by the biomass due to respiration is calculated using Eq. 8 * N/A Not available ** Positive and negative value represents net carbon fixation and net CO 2 release, respectively M C-biomass (g C L −1 ) M C-lactose (g C L −1 ) M C-capture darkness, as evidenced by the absence of the enzyme in the medium. Heterotrophic cultures metabolised less than 13% of available lactose and thus experienced minimal overall biomass growth (final biomass concentration = 0.14 -0.16 g biomass L −1 ) that was ca. four-fold lower than in the mixotrophic cultures (0.56 -0.71 g biomass L −1 ). Despite this slow growth, the cultures were still able to remove all available nitrates and ca. 94% of available phosphates in the medium, likely attributed to the denitrifying capacity of the native bacteria typically isolated in Nannochloropsis culture (Poddar et al. 2018). The underlying reason for the light-dependent nature of lactose assimilation in Nannochloropsis is not well understood but can likely be ascribed to the interdependence enzyme synthesis and adenosine triphosphate (ATP) availability. Photosynthesis provides substrates that are metabolized by algal cells to synthesise energy (i.e., ATP) via aerobic respiration, thus driving other processes in living cells such as cellular maintenance and chemical/enzyme synthesis, including those needed for the assimilation of external nutrients (Knowles 1980). While in heterotrophic cultivation, no additional oxygen was generated due to the lack of photosynthesis and only the dissolved oxygen in the culture could be utilized for aerobic respiration. This likely resulted in a limited energy production; cellular maintenance for basal metabolism therefore took priority over the production of non-essential enzymes, such as the production of extracellular β-galactosidase, leading to undetectable level of enzymatic activity in heterotrophic cultures.
The maximum microalgal enzymatic activity (41.47 ± 0.33 U g biomass −1 ) measured for mixotrophic cultures in this study was lower than those previously reported in other studies with bacteria and yeasts (Carević et al. 2015;You et al. 2017). This can likely be ascribed to species variation and specific cultivation conditions (e.g., temperature, pH and medium composition). Previous studies have reported that ion composition in the medium can affect the activity of lactase derived from microorganisms (Jurado et al. 2004;Otieno 2010). In the present study, since artificial sea water was employed as medium, the mineral ion concentrations were higher than those of freshwater medium and can have an inhibitory effect on the β-galactosidase activity.
The identity of native bacteria associated with N. oceanica cultures was not investigated. However, other studies have shown that bacteria belonging to the Proteobacteria phylum to be dominant in Nannochloropsis sp. cultures (Nakase and Eguchi 2007). Handley and Lloyd (2013) and Kelly et al. (2006) have also indicated that Marinobacter sp. and Paracoccus sp. were ubiquitous in marine environment as they possess the ability to resist high salinity and low nutrient conditions. Poddar et al. (2018) have indicated that native Marinobacter alkaliphilus isolated from their Nannochloropsis cultures had denitrifying capabilities, being able to metabolise organic carbon through the utilisation of nitrate and oxygen as terminal electron acceptors (Robertson and Gijs Kuenen 1984;Liu et al. 2016;Zhang et al. 2016).
In terms of industrial implications, the results from this study highlighted the inherent limitations of using N. oceanica to bioremediate lactose-rich dairy side streams while co-producing valuable β-galactosidase and EPA. The species' inability to utilize lactose under heterotrophic conditions limits its application in treating carbon-rich dairy waste to strict mixotrophic strategy. In the context of wastewater treatment in Europe, however, mixotrophic cultivation is less scalable than heterotrophic systems due to the need for a) light irradiation (a limited natural resource in northern Europe) which leads to increased cost in artificial lighting, b) larger land area to house photobioreactors for which many landlocked food operators (e.g., dairy processors) simply lack the space to accommodate. Photobioreactors for autotrophic and mixotrophic cultivation are also generally considered to more expensive to build and operate than fermenters used for heterotrophic growth. Open-raceway ponds for mixotrophic cultivation is not a viable option due to the increased risk of bacterial contamination and culture crash. Future research effort should focus on screening Nannochloropsis species able to grow heterotrophically in lactose-rich dairy side-streams or further optimisation of mixotrophic conditions to reduce light and heating requirements.

Conclusion
This study investigated the mechanism of lactose assimilation in N. oceanica for dairy-wastewater bioremediation and co-production of β-galactosidase. Lactose assimilation in N. oceanica was found to require mixotrophic conditions, with biomass production in mixotrophic cultures reaching > fourfold that in the heterotrophic systems. Under mixotrophic conditions, maximum extracellular β-galactosidase secretion (41.47 ± 0.33 U g biomass −1 ) was attained during mid-logarithmic phase when cells were still rapidly dividing. The algal cells were able to indiscriminately absorb both glucose or galactose sub-units derived from lactose hydrolysis and no monosaccharide accumulation was observed in the medium. Bacterial growth was supressed, and microalgae cells maintained dominance throughout cultivation (bacteria accounting for < 12% of biomass by the end of cultivation). Moreover, the mixotrophic system investigated in the study has the potential to transform the carbon present in lactoserich dairy wastewater into both microalgae biomass and CO 2 via respiration, thereby reducing the net amount of carbon expelled into the environment relative to the initial carbon in the wastewater.
On the other hand, under heterotrophic conditions lactose assimilation pathways in N. oceanica were inactivated resulting in native bacteria taking over the culture and occupying over 95% of total biomass by the end of cultivation. The bacterial population, however, also appeared to lack the metabolic pathways needed to effectively assimilate lactose in darkness, producing negligible amount of extracellular β-galactosidase and consuming < 10% of available lactose. Results from the study indicate that N. oceanica can be effectively applied for onsite dairy wastewater treatment under strict mixotrophic conditions. Its inability to assimilate lactose under heterotrophic conditions is commercially disadvantageous as it increases areal footprint and bioreactor costs.