Potential for Biomass Production and Remediation by Cultivation of the Marine Model Diatom Phaeodactylum tricornutum in Oil Field Produced Wastewater Media

: While oilﬁeld produced water (PW) is one of the largest, unclaimed wastewater streams of the oil industry, it could potentially be used as a cultivation medium for microalgae. Microalgae could help with the remediation of this water while also delivering biomass that can be transformed into valuable byproducts such as biofuels. The coupling of these two purposes is expected to cut production costs of biofuels while aiding environmental protection. In this study, we compared the cultivation capacity of the marine model diatom Phaeodactylum tricornutum in media at varying salinities and in media composed of PW from two oilﬁelds in the Central Valley of California that differed drastically in the concentration of inorganic and organic constituents. Speciﬁcally, we measured the carrying capacity of these media, the maximum growth rates of P. tricornutum , its cellular lipid accumulation capacity, and its capacity to remediate the most polluted PW source. Our study shows that P. tricornutum can successfully adjust to the tested cultivation media through processes of short-term acclimation and long-term adaptation. Furthermore, the cultivation of P. tricornutum in the most heavily polluted PW source led to signiﬁcant increases in cell yield and improved photosynthetic capacity during the stationary phase, which could be attributed chieﬂy to the higher levels of nitrate present in this PW source. Chemical water analyses also demonstrated the capability of P. tricornutum to remediate major nutrient content and potentially harmful elements like ﬂuorine and copper. Because P. tricornutum is amenable to advanced genetic engineering, which could be taken advantage of to improve its cultivation resilience and productivity in an economic setting, we propose this study as a step towards essential follow-up studies that will identify the genetic regulation behind its growth in oilﬁeld PW media and its remediation of the PW constituents.


Introduction
During oil extraction, water trapped with the oil in underground reservoirs is brought to the surface. This water is referred to as produced water (PW) and it is the largest waste stream of the oil industry. Disposal of PW poses environmental challenges due to the presence of organic and inorganic constituents that can contaminate the environment. To accommodate the risks, oil operators commonly dispose of the PW in evaporation ponds, re-inject it into subsurface reservoirs, or transport it to offsite storage facilities [1]. However, these practices do not achieve repurposing of the water for beneficial uses such as land irrigation or groundwater augmentation. Remediation of PW for such beneficial uses relies on costly physicochemical procedures like filtration, reverse osmosis, flotation, adsorption, oxidation, or electrodialysis [2,3]. An alternative to these procedures is bioremediation, the use of biological systems to achieve water purification. For example, natural or constructed wetlands have shown to remove nutrients and petroleum-derived toxic organic the variable salinities of PW media. The strain that performed best at a wider range of salinities was then used for the assessment of growth and cellular lipid accumulation in media prepared with PW that was collected from two oilfields in the Southern San Joaquin Valley of central California. Finally, chemical analysis before and after cultivation in the most polluted PW source allowed evaluating the possible causes for P. tricornutum's growth cessation in the stationary phase, as well as assessment of the PW remediation achieved by P. tricornutum cultivation.

Strain Maintenance and Cultivation
The P. tricornutum strains CCAP1055/1 (Pt1) and CCAP1052/6 (Pt4) were obtained from the Culture Collection of Algae and Protozoa (Argyll, Scotland, UK), and the Culture Collection of Algae at the University of Texas in Austin (USA), respectively. Stock cultures of both strains were maintained by weekly reinoculations of cultures in synthetic ESAW medium at normal salinity (31 psu) [47]. All procedures were conducted in an AirClean ® 600 PCR Workstation equipped with HEPA filters (AirClean, Creedmoor, NC, USA). For cultivation, strains were incubated in a Percival E-41L2 incubator (Percival Scientific, Perry, IA, USA) at 18 • C with a 14 h:10 h light:dark cycle provided by top-mounted cool white fluorescent lamps (Philips, TL841) at a photon flux of 120-140 µmol m −2 s −1 . Routine microscopic observations were made by investigation of cell suspensions in 12-multiwell plates at 400× magnification on an Leica DM-IL inverted microscope (Leica, Wetzlar, Germany) and at 1000× magnification using an Olympus CH30standard light microscope (Olympus, Shinjuku City, Japan).

Media Preparation
"East-side" PW (EaS-PW) was collected from an oilfield PW collection and treatment facility near Edison, California. PW from several small, independent oil leases in the Edison Oil Field is transported to the facility by pipeline and flows into gunite-lined ponds where crude oil is skimmed from the water. PW was collected at the final stage before it is pumped to another facility into unlined sumps and onto irrigation areas. "McFarland" PW (McF-PW) was collected from leases in the Dyer Creek oil field, located approximately 15 miles northeast of Bakersfield, California. The treatment of the water includes the separation of oil by heat treatment after which the remaining PW flows through a series of gravity separators. The water for this project was taken from the last separator before flowing into an unlined pond.
The PW from both oil field sources were stored in tempered, autoclaved polypropylene carboys after filtration with Whatman type 1 filters (11 µm pore size) to remove large insoluble fractions. This filtrate was then used without further modifications like autoclavation or ultrafiltration to allow that the outcomes of subsequent cultivation and chemical analyses (which could be altered by further processing of the PW) represent authentic field settings. Nonetheless, the PW sources were enriched with typical phytoplankton cultivation media enrichments to allow meaningful P. tricornutum growth, needed for PW remediation (see Section 3.4) and for allowing investigation of the impacts of PW constituents on P. tricornutum growth (see Section 3.3). Specifically, the cultivation media were enriched with "F/2" formulations for iron and the major nutrients (without silicate; [48]) and with the enriched artificial seawater (ESAW) enrichment for the trace metals [47]. The experimental condition called "McF/EaS-PW" was prepared by diluting the EaS-PW (at 12 PSU) with McF-PW until reaching a salinity of 5 psu, as measured with a salinity refractometer (RF20, Extech ® , Nashua, NH. USA). Synthetic medium for preparation of the control media was made following the standard procedure by Berges et al. (2001) and the reported salinity was achieved by dilution of this synthetic base medium with HPLC grade purified water.

Salinity and PW Cultivation Experiments
Cultivation was done on an orbital shaker (ca. 200 rpm) in 125 mL Erlenmeyer flasks (Pyrex, No. 4980) with 30 mL cultivation medium. For the salinity experiment, cell inocula originated from stock cultures that had acclimated to the respective salinity medium during three consecutive batch cultivation cycles (trial 1). Cell inocula in the salinity experiment 'trial 2' had received three additional batch cultivation cycles during which cells continued to acclimate. The growth rate analysis experiment in response to salinity levels was initiated by a 40-fold dilution of the starter cultures. Afterwards, both strains were maintained in 0, 5 and 12 psu control media by weekly inoculations at a 40-fold dilution of the cell suspension. The PW experiments were initiated three months after the salinity experiments and the cell inocula that were used (for both the controls and PW treatments) originated from the acclimated strains that had been maintained in the ESAW media at 0, 5, or 12 psu. Experimental cultures for testing growth in the PW media (both 'trial 1' and 'trial 2') were initiated by inoculation with ca. 3 × 10 4 cells mL −1 as estimated from microscopic cell counts.

Cellular Growth and Photophysiology Measurements
Cell density estimations were made by performing cell counts of culture aliquots using microscope counting slides (Quick-Read TM , Globe Scientific Inc., Mahwah, NJ, USA) at 100× magnification using a standard light microscope (CH30, Olympus). Following the manufacturer's recommendation, cell density estimates for each sample were based on two technical replicate counts that were performed in separate counting wells. During cell counting, qualitative observations of cell morphology and bacterial presence were made. Cellular fluorescence parameters were monitored using Pulsed-Amplitude-Modulation based fluorometry (AquaPen-C 100; Photon Systems Instruments, Drasov, Czech Republic) after 15 min dark incubation of 2 mL cell suspensions. This method yields the cells' photosynthetic capacity (the quantum yield of Photosystem II) as F v /F m and the darkadapted minimal fluorescence parameter (F 0 ) as a reliable proxy for biomass during the exponential growth phase [49]. Maximum growth rates during batch cultivation were calculated by multiplication of the doubling time k with the natural logarithm of 2, where k was determined by regression of either F 0 values or cell density values during the exponential growth phase according to the equation k = log 2 (N t /N 0 )/∆t [50]. For each condition, the phase of exponential cell proliferation was determined from the Log 2transformed growth curves.

Neutral Lipid Content Measurements
Intracellular neutral lipid content was estimated by the staining of cells with Nile Red (NR, Sigma-Aldrich). The NR stain forms a fluorescent complex with cytoplasmic oil bodies and quantification of this fluorescence is a proxy for neutral lipids in various unicellular organisms, including P. tricornutum [51]. Our method was developed by modification of the procedure from Sitepu et al. [52]. Briefly, 1.8 mL cell suspensions were sampled from each culture, pelleted by centrifugation to remove the supernatant cultivation medium (5000× g, 6 min), and resuspended in corresponding fresh medium that contained 10% (v/v) DMSO and 4 µg mL −1 Nile Red from a 0.5 mg mL −1 working stock (parent stock was made at 1 mg NR mL −1 acetone).
Because fluorescence-based cell assays are sensitive to cell density effects (i.e., a limitation of sensitivity at the low densities and a saturation and/or shading at high densities), the volume added for resuspension of the cell pellets was varied based on the cell counts in each culture at the time of sampling so that consistent cell densities were achieved that ranged between 5 × 10 6 cell mL −1 and 30 × 10 6 cells mL −1 . Immediately afterwards, the centrifuge tubes were mixed by inversion and three 200 µL aliquots were transferred to black, clear-bottom 96-well plates to serve as technical replicate reads for the calculation of an average value per sample. The plates were read in a fluorescence plate reader (Flx-800, BioTek Instruments, Winooski, UT, USA) using the bottom probes with appropriate filters for NR fluorescence detection (Ex: 525/30 nm; Em: 671/20 nm). Plates were scanned every minute for 30 min to identify the maximum fluorescence in each well. The background fluorescence, determined from a mock NR-stained medium sample (no cells), was subtracted from this value. Because samples for chemical characterization were taken from the PW source before using them for media preparation, and thus did not include the reported medium enrichments, we calculated the RC based on the estimated analyte concentrations after enrichment. Supplementary Table S1 shows the calculations of these estimated concentrations based on the used enrichment formulations. Characterization of the spent EaS-PW cultivation medium was done after removal of the P. tricornutum biomass by vacuum filtration on Grade A Borosilicate Glass Microfiber filters (Sterlitech, Auburn, AL, USA). These filters have a with a nominal pore size of 1.6 µm, thereby allowing retention of P. tricornutum cells while allowing passage of possible bacterial cells into the filtrate.

Statistical Analyses
Statistical analysis was performed using Prism 9.0.0 (GraphPad Software). Normality of data was tested with the D'Agostino-Pearson "omnibus K2" normality test. Based on this, two-way ANOVA tests were performed on raw values (cell yields) or Log 2 -transformed values (NR fluorescence, and growth rates) to satisfy the assumptions of the statistical model. Model factors in the salinity experiments were 'Strain' (Pt1, and Pt4) and 'Salinity' (5, 10, 15, 25, 31 psu), and in the PW experiments were 'Treatment' (PW, and control) and 'Salinity' (0, 5, 12 psu). Post hoc multiple comparisons of the dependent variable were made between groups as mentioned, and with Bonferroni's correction for calculation of adjusted p-values.

Growth Comparison of Phaeodactylum tricornutum Strains Pt1 and Pt4 at Various Salinities
A growth comparison of the P. tricornutum strains CCAP1055/1 (Pt1) and CCAP1052/6 (Pt4) was conducted to assess variation in their capacity to grow and acclimate to media with varying salinities (Figure 1). In the first experimental run (trial 1), initiated after only a three batch cultivation cycles at the respective salinities (see Section 2), Pt4 grew significantly faster in 10 and 15 psu media than in 5, 25 and 31 psu media, but differences in growth rate between the strains were not significant ( Figure 1a). In trial 2, initiated after three additional acclimation cycles at the same conditions, the growth rate differences at 25 and 31 psu disappeared ( Figure 1b).
(Pt4) was conducted to assess variation in their capacity to grow and acclimate to media with varying salinities (Figure 1). In the first experimental run (trial 1), initiated after only a three batch cultivation cycles at the respective salinities (see Section 2), Pt4 grew significantly faster in 10 and 15 psu media than in 5, 25 and 31 psu media, but differences in growth rate between the strains were not significant ( Figure 1a). In trial 2, initiated after three additional acclimation cycles at the same conditions, the growth rate differences at 25 and 31 psu disappeared ( Figure 1b). For the freshwater condition (0 psu), both strains exhibited an extended lag phase during days 1-5 in both trial 1 and trial 2 ( Figure 1b). This indicated that cultures were likely not fully acclimated to this freshwater condition during the provided six growth cycles (see Section 2) and made it impossible to calculate growth rates for this condition. Given the absence of significantly higher growth rates in Pt4, and an apparent broader optimal salinity range for Pt1, we selected Pt1 for subsequent experiments.

Chemical Characterization of Two Oil Field Produced Water Sources
We physicochemically characterized oilfield PW from two locations in California's San Joaquin Valley. One PW source originated from oil fields near Edison, California, which we called "East-side" PW (EaS-PW), and the other PW source originated from oil fields near For the freshwater condition (0 psu), both strains exhibited an extended lag phase during days 1-5 in both trial 1 and trial 2 ( Figure 1b). This indicated that cultures were likely not fully acclimated to this freshwater condition during the provided six growth cycles (see Section 2) and made it impossible to calculate growth rates for this condition. Given the absence of significantly higher growth rates in Pt4, and an apparent broader optimal salinity range for Pt1, we selected Pt1 for subsequent experiments.

Chemical Characterization of Two Oil Field Produced Water Sources
We physicochemically characterized oilfield PW from two locations in California's San Joaquin Valley. One PW source originated from oil fields near Edison, California, which we called "East-side" PW (EaS-PW), and the other PW source originated from oil fields near McFarland, California, which we designated "McF-PW". Quantities of analytes for the PW sources, as they were collected, are shown in Table 1.
With 12,600 mg/L total dissolved solids (TDS), the EaS-PW had a much lower quality than the McF-PW (1160 mg/L TDS). The higher TDS value of the EaS-PW was reflected by the higher salinity (12 psu vs. 0 psu) and electrical conductivity (20 mmhos/cm vs. − ), a much less studied nutrient for diatoms [53], was present at higher levels in EaS-PW, but was also present in McF-PW (390 mg/L vs. 180 mg/L). The alkalinity, i.e., the sum of bicarbonate (HCO 3 − ), carbonate (CO 3 2− ) and hydroxide (OH − ) concentrations, was ca. 6-fold higher in EaS-PW (630 mg/L vs. 100 mg/L) and this seemed to result mainly from higher HCO 3 − (490 mg/L vs. 100 mg/L) and CO 3 2− (140 mg/L vs. 0 mg/L) concentrations. This was also in line with the higher pH value of EaS-PW (8.56 pH units vs. 7.9 pH units). Finally, the Water 2021, 13, 2700 7 of 18 chemical analysis showed higher levels of total hydrocarbon contamination in EaS-PW (1.13 mg/L vs. 6.69 mg/L). The main fraction existed as C 28-35 hydrocarbons (4.1 mg/L vs. 1.13 mg/L), followed by C 10-28 hydrocarbons (2.59 mg/L vs. <0.05 mg/L), while shorter C 6-10 hydrocarbons were not detected in either PW source. Table 1. Analyte quantities in the produced water (PW) source and the relative change in EaS-PW analytes after Phaeodactylum tricornutum cultivation. Analyte quantities below the practical detection limit (PDL) are indicated with a smaller-than sign followed by the detection limit. Negative relative change (RC) values indicate a reduction and positive RC values indicate an increase in the quantity of the analyte after cultivation in EaS-PW medium. In cases where the PDL was reached in one of the EaS-PW samples, the RC was calculated using the PDL as a conservative estimate of the analyte's concentration. Bolded numbers are those with a reduction exceeding 10%, while italic numbers are those with an increase exceeding 10%. a Values between parenthesis correspond to the quantity of the analyte that was supplemented after chemical analysis as part of the medium enrichment (see Supplemental Table S1 for calculations). For these analytes, the relative change (RC) is calculated using the total estimated concentration in the medium (i.e., sum of the measured concentration and supplemented concentration).

Growth Characterization and Lipid Accumulation of Pt1 in Three Distinct PW Media Formulations
To assess the impact of PW on P. tricornutum growth and physiology, PW media were prepared from the characterized PW sources by enriching them with nutrients typically used for creation of synthetic seawater media, and the same additions were used for the preparation of the synthetic control media at the corresponding salinities. Three media types were compared: McF-PW (0 psu), EaS-PW (12 psu) and McF/EaS-PW, a mixture of both PW-media at intermediate salinity (5 psu). Four replicate sets of cultures were monitored, with cultivation carried out in two runs (trial 1, trial 2), each lasting 9 days. These cultures were initiated with inocula from salinity-acclimated stock cultures that had acclimated for an additional 3 months after the earlier salinity experiments shown in Figure  1 (see Section 2). Cell densities, photosynthetic capacities and cellular lipid accumulation were determined each day. In addition, qualitative microscopic observations were made while performing cell counts to assess bacterial growth and cell morphology.
Growth curves showed that cells in most conditions immediately resumed division upon re-inoculation in their respective medium (Figure 2a,b). Only the McF-PW cultures and their 0 psu controls showed a prolonged lag-phase compared to cultures at higher salinities. However, the lag-phase was shorter than in the earlier salinity experiments (2 days vs. 5 days; compare Figures 1c and 2b) indicating that the cell population was able to adjust to the 0 psu medium during the 3 months between both trials. Comparison of growth rates (Figure 2c) showed no significant difference between the treatment and control at each salinity level, thereby indicating that cells were successfully coping with the oilfield PW constituents. Furthermore, comparison of cell yields (Figure 2d) showed that both EaS-PW and McF/EaS-PW cultures had significantly higher carrying capacities compared to their controls, with a mean cell density increase of 159% (i.e., 1.67 × 10 6 additional cells ± 643,447 SE) and 149% (i.e., 1.41 × 10 6 additional cells ± 3,033,021 SE), respectively. Interestingly, these cell yield increases coincided with higher photosynthetic capacities during the late-exponential and stationary growth phases (Figure 2e Figure 3b). However, this difference disappeared when comparing the lipid content per culture volume (Figure 3c).

Characterization of the EaS-PW Medium after P. tricornutum Cultivation
To gain further insight into the limiting growth factors of the EaS-PW media, as well as to assess the remediation capacity of P. tricornutum cultures, we chemically analyzed the spent cultivation media of stationary-phase EaS-PW cultures with the same analytical methods as before. These results were used to calculate the relative change (RC) in the concentration of each analyte (Table 1; "EaS-PW medium after" and "Relative Change" column).
Foremost, while EaS-PW contained higher levels of HCO 3 − (which is essential for photosynthetic carbon fixation) its presence was not detected in the spent medium. This indicated that the uptake of HCO 3 − outpaced the dissolution of atmospheric CO 2 , an observation that was also reflected by the pH increase and the associated increase in biounavailable CO 3 2− . Cultivation also led to a significant reduction, but not exhaustion, of the major nutrients NO 3 − and PO 4 3− (92% and 76% RC, respectively). According to other studies, the remaining levels of NO 3 − and PO 4 3− (90 µM, and 20 µM, respectively) are well above the level at which cells can continue importing these compounds [54,55]. Interestingly, silicon (Si), the main component of the diatom cell wall in most species, but only facultatively required by P. tricornutum [56,57], decreased with 60%. Other inorganic EaS-PW constituents that were significantly reduced by P. tricornutum cultivation were Fe (85% RC), F (72% RC), Mg (56% RC), SO 4 2− (21% RC), and Cu (18% RC). Inorganic analytes that remained largely unchanged were B, Ba, Ca, Cl, K, and Na. Finally, surprisingly, we measured an increase in C 10-28 and C 28-35 hydrocarbon fractions after P. tricornutum cultivation.   Individual data points are shown as dots, the median is represented by a dashed line, and quartiles are represented by dotted lines. The starred comparison marks indicate significant differences in the dependent variable (*: adj. p < 0.05; **: adj. p < 0.005) based on two-way ANOVA post hoc multiple comparisons between the PW treatments and their controls (n = 4). One outlying growth rate value was removed from the 0 psu control condition (see Supplemental File S1). (e) Photosynthetic capacities (F v /F m ) of cultures in (a,b). For graphing clarity, panels (a,b,e) show data from the second trial only (n = 2); Error bars represent standard deviations around the mean. The first trial showed comparable growth dynamics and its results are available in Supplementary File S1.
Water 2021, 13, x FOR PEER REVIEW 10 o Figure 3b). However, this difference disappeared when comparing the lipid content p ture volume (Figure 3c).

Acclimation and Adaptation of Phaeodactylum tricornutum to Variation in Salinity
From results of trial 1 of the salinity experiment, it appears that the optimal salinity of Pt4 ranged around 10-15 psu, while Pt1 showed a salinity tolerance across the 10-31 psu range. This higher growth rate of Pt1 at 25 and 31 psu is in accordance with the higher salinity level at its origin of isolation (Irish Sea, 30 psu) compared to that of Pt4 (Baltic Sea; ca. 6 psu). However, we had hypothesized that the habitat of Pt4, i.e., a supralittoral rock pool, would have acted as selective pressure on this strain to equip it with a higher tolerance to variation in salinity common to these complex habitats [58,59]. While the improved growth of Pt4 at the 25 and 31 psu condition after additional cultivation cycles at these salinities does indicate that this strain has a high capacity to acclimate to variation in salinity, its comparable growth rates after acclimation did not support superiority of Pt4 for growth in PW sources of variable salinity. Of relevance here is a recent study that showed a distinct pattern of nucleotide polymorphisms in Pt1, setting it apart from the other P. tricornutum strains [46]. While this indicated that Pt4 did indeed experience a stronger past selective pressure compared to other accessions, the functional analysis of the affected gene groups pointed at main effects from variation in light intensity and nitrate availability, instead of variations in salinity. Given these considerations, and our results, we can thus not support the hypothesis that Pt4 is better adapted for growth at the variable salinity levels that characterize PW sources.
While acclimation is the physiological adjustment of an organism to its environmental conditions, adaptation is a longer-term evolutionary process that involves changes to the species' genome sequence followed by selection. While both strains did not successfully acclimate to the freshwater condition (0 psu) within the six provided batch cultivation cycles, the Pt1 strain adjusted successfully to conditions after three additional months of cultivation in the 0 psu control medium (Figure 2). The longer time scales needed for improvement of growth in the freshwater condition indicates the Pt1 strain relied on genetic adaptation instead of physiological acclimation [60]. At first sight, given that P. tricornutum is a non-sexual species, it is surprising that genetic adaptation would have occurred so rapidly. Recently, however, it was shown that under stressful conditions, P. tricornutum executes mitotic homologous recombination events between homologous chromosomes at a rate more than 10 times higher than that of the yeast S. cerevisiae [61]. The authors of that study suggest that this process underlies the phenotypic plasticity that led to its adaptation to stressful conditions, even in monoclonal strains after just one month of exposure to the stressor. Based on this information, we consider it very likely that homologous recombination events were responsible for the observed adaptation of the strains to the 0 psu and the McF-PW conditions. Together, these results show that prolonged cultivation of P. tricornutum might be a strategic approach to optimize its performance in the specific conditions pertaining to economic mass cultivation in wastewater sources, including oilfield PW.

Cultivation of P. tricornutum in Oilfield PW
While the carrying capacity of McF-PW medium trended lower compared to that of the media at higher salinity, statistical comparisons showed that this medium performed equally well compared to its control. Furthermore, P. tricornutum growth rates were also not significantly different in this condition, meaning that cells were successfully coping with the constituents of the McF-PW source. In contrast, the carrying capacities of the two media types prepared with the EaS-PW source were significantly elevated compared to their control medium. Because nutrient depletion typically constrains cell yield during diatom batch cultivation (as opposed to light or space limitation), these results suggest that P. tricornutum acquired nutrients from the EaS-PW source. Chemical analyses showed that there was a substantial amount of nitrate (NO 3 − ) present in the EaS-PW source. NO 3 − is the main source for N assimilation in diatoms and its depletion is well-known to lead to decreases in photosynthetic capacity, similar to those observed in the control media [62].
We, therefore, suggest that the surplus NO 3 − in EaS-PW, which still was not fully depleted in the stationary phase, allowed cellular maintenance of the photosynthetic machinery for a longer time, thereby stimulating additional cell divisions, and yielding higher cell densities the EaS-PW cultures compared to the control cultures. Furthermore, given the absence of bioavailable bicarbonate (HCO 3 − ) from the spent EaS-PW medium, the higher cell densities, in turn, appear to have shifted the culture to a C-limited state as opposed to a N-limited state. While HCO 3 is needed for photosynthetic carbon fixation, it is known not to lead to severe impact P. tricornutum's photosynthetic capacity [63,64]. As such the higher observed F v /F m values are in line with our interpretation for HCO 3 − limitation in EaS-PW cultures. It should be noted that EaS-PW cultures were also possibly experiencing a slight co-limitation by iron (Fe) given its low concentration in the spent medium (i.e., below the practical detection limit of 0.1 mg/L) but severe Fe limitation can again be excluded (similar to N limitation) based on the observed high photosynthetic capacities in the stationary phase of these cultures [65].
NO 3 − is one of the main nutrients that regulates lipid metabolism in diatoms; when the medium is depleted of these nitrogen compounds, the cells initiate proteins degradation (mainly photosynthetic light harvesting complex proteins) while storing excess carbon in the form of lipids. Then, when nitrogen becomes available again, these lipids can generate energy needed in these photosynthetic impoverished cells to restore growth [66][67][68][69][70]. Likely because of this phenomenon, together with the fact that NO 3 − is progressively depleted during batch cultivation, cellular lipid content measurements showed a progressive increase in the cellular lipid pool (Figure 3a).
On the final day of cultivation (day 9), cells in McF/EaS-PW and EaS-PW cultures appeared to have accumulated significantly less lipid per cell. These lower levels of cellular lipid accumulation are not surprising since these cultures contained additional NO 3 − , which counteracts lipid synthesis as explained above. In addition, the absence of CO 2 in these cultures prohibits the synthesis of the fatty acid precursor Acetyl-coA through carbon fixation [33]. This suggests that higher cellular lipid yields in the EaS-PW medium could likely be achieved by bubbling cultures with CO 2 or atmospheric air, thereby shifting cells to a N-limited state instead of a C-limited state.
Interestingly, however, compared to the lipid content per unit of cell, the lipid yield per unit of culture volume was similar between the EaS-PW-containing cultures and their controls (Figure 3c). This means that the higher cell yields in these cultures offset the cell's limited capacity for lipid accumulation. Thus, while the oilfield PW did not improve lipid yield per cell, it did allow for comparable amounts of lipids to be produced in both cultures.. Because the NR cell assay is an approximate method for lipid quantification, future experiments using more precise analytical methods would be helpful to validate these insights.
Together, our analysis shows that diatom cultivation in PW media can achieve substantial biomass that can be used for downstream processing. Nevertheless, its downstream applicability may be limited to certain purposes. For example, while it might be suitable for biofuel production, the specific contaminants of the oilfield PW could pose health or environmental risks that make it unsuitable as a nutritional feedstock in, e.g., the aquaculture industry. Thus, it should be noted that from this perspective, the downstream purpose of the biomass for certain oil field PW sources could be inversely related to the bioremediation efficiency discussed in Section 4.3.

Remediation of Oilfield PW by P. tricornutum Cultivation
There was a partial remediation of the EaS-PW by the cultivation of P. tricornutum. Foremost, the major nutrients NO 3 − and PO 4 3− were drastically reduced, as is expected from cellular assimilation of these nutrients for the synthesis of protein and nucleic acid macromolecules. To drive these anabolic pathways, the cell needs an adequate supply of HCO 3 − for photosynthetic carbon fixation. While this HCO 3 − originates from the dissolution of CO 2 into the water, removal of the surplus of HCO 3 − that was already present in the EaS-PW shows that this was also utilized for growth by P. tricornutum, and it demonstrates the capacity of microalgal cultivation to aid in atmospheric carbon sequestration. To accommodate this photosynthetic process, microalgae rely heavily on Fe, a main component of the photosynthetic apparatus. The significant removal of Fe (85% RC) from the EaS-PW medium is thus in line with the exceptional growth of P. tricornutum in this medium. Furthermore, partial removal of SO 4 2− from the medium was also achieved. Diatoms assimilate SO 4 2− by reducing it to the amino acid cysteine [53]. Sulfate is the oxidation product of sulfur, and while it is usually harmless to the environment, it can be converted to more environmentally harmful forms [71]. The 21% RC of SO 4 2− thus shows that diatoms can also contribute to the remediation of this nutrient.
We found a significant reduction in the concentration of silicate (SiO 2 ; 60% RC) after cultivation. SiO 2 sequestration by diatoms commonly happens as a means to synthesize their Si-based cell walls (frustules). However, in the case of P. tricornutum, only one of the morphotypes, i.e., the oval morphotype, uses SiO 2 when it is available in the medium; The other morphotypes consist of purely organic frustules, even in the presence of SiO 2 [57]. The significant reduction in SiO 2 thus indicates that the EaS-PW might have stimulated transformation of P. tricornutum into the oval morphotype type. Indeed, this has already been shown to happen in stressful conditions such as during changes in salinity levels [56]. In addition, as an alternative explanation for the reduction of SiO 2 , it is also possible that the organic fractions of the P. tricornutum frustule in the non-silicified cells adsorbed some of the SiO 2 [10]. Future quantification of cell morphotypes during cultivation in PW will help to shed a more complete picture on the mechanism of SiO 2 removal and the value of morphotype transformations in the bioremediation by P. tricornutum [72,73].
P. tricornutum cultivation also led to reduced amounts of some elements that occurred at trace levels in the EaS-PW medium. Fluorine content was significantly reduced (72% RC: from 4.6 mg/L to 1.3 mg/L). A common form of fluorine in water is the fluoride ion F − . F − has several molecular targets that can impact cellular respiration and photosynthesis [74,75], as well as be detrimental to human health [76]. The WHO limit on concentration of F − in drinking water is 1.5 mg/L [77] and, depending on the ordinance levels, 1-2 mg/L are recommended for long-term soil irrigation [78]. At the same time, several phytoplankton species exhibit dose-dependent growth stimulation by F − (by as yet unknown mechanisms), even at concentrations well above those in the PW media (e.g., up to 200 mg L −1 in the diatom Chaetoceros gracilis) [79]. The successful remediation of F − by P. tricornutum cultivation, to levels surpassing, or close to, the legal thresholds, indicates a similar resistance of P. tricornutum to F − and points at a capacity for cellular F − uptake, or an adsorption to its frustules.
For copper (Cu), toxicity has been reported in microalgae and P. tricornutum at free Cu 2+ ion concentrations 10-times below the measured elemental concentration present in the EaS-PW medium (i.e., LC 50~0 .006 mg L −1 ) [80][81][82]. However, P. tricornutum cultures that were grown in the presence of high levels SiO 2 , such as those present in the EaS-PW medium, have been shown to exhibit greater Cu adsorption and associated resistance [73]. It is thus possible that this, together with chelation of Cu 2+ ions by other PW constituents, has likely contributed to the absence of any observable PW toxicity and the concomitant reduction (18% RC) in Cu concentration after P. tricornutum cultivation.
Additionally, magnesium (Mg) concentrations were significantly reduced (56% RC), but this is less essential from an environmental protection standpoint since there are no toxicity concerns for this element. Mg is the central atom that chelates chlorophyll molecules composing the light-harvesting complexes of the chloroplast membranes. In addition, it is also a common cofactor of ATP-driven enzymatic reactions. Therefore, reduction in Mg concentration could be explained by biomass-driven sequestration.
With respect to the metalloid boron (B), while some diatoms have been reported to require B for growth [83], the finding that P. tricornutum only requires B in the absence of other trace metals [84] (which was not the case) could explain why the B concentrations remained largely unchanged.
The concentrations of various remaining analytes did not change substantially (i.e., <10% RC for Ba, Ca, Cl, K). This could mean that, while some of these elements are common solutes of the diatom cytoplasm, their removal from the PW closely matched the removal by volume that was taken up by the wet diatom biomass, as this would not result in major concentration differences in the spent medium. In contrast, the substantial reduction in the PW constituents described above results from cellular sequestration in excess of the volume that was removed upon filtration. This explanation is in line with the observation of phytoremediation in plants where some species are geared towards preferential sequestration of specific components [85,86].
Finally, the higher concentration of hydrocarbons after cultivation can be attributed to the escape of lipids synthesized by P. tricornutum. This could happen naturally during the stationary growth phase when cell membranes lose integrity in advance of cell death [87]. Alternatively, while frequent microscopic observations during cultivation did not show any obvious bacterial growth in any of the media, including the EaS-PW spent medium, we judge it possible that the higher hydrocarbon values originated from a marginal growth of bacteria that were originally present in the EaS-PW source, fed on P. tricornutum photosynthate, before being passed into the collected spent medium during the separation of P. tricornutum biomass on the used glass microfiber filters, which had a nominal pore size of 1.6 µm. The fact that only the longer hydrocarbon species (C 10 -C 35 ) were elevated after cultivation is in line with such biogenic origin since these hydrocarbons typically represent a larger fraction of diatom and bacterial lipid content [88,89]. Importantly, however, we would like to note that this filtration strategy (desirably) precluded attribution of the remediation of inorganic constituents to bacterial populations, thus pointing at a chief role of P. tricornutum in this process.

Conclusions
This study highlights the potential use oil field produced water (PW) for the photosynthetic generation of oleaginous biomass through cultivation of the model diatom species Phaeodactylum tricornutum. Our results indicate that the nutrients present in this wastewater source can increase the cell yields of cultivation, pushing the lipid production capacity of these cultures to similar levels compared to those grown in pure-water media. At the same time, chemical analyses show that P. tricornutum growth can make improvements in the water quality by removal of major nutrients and potentially harmful elements. Cultivation experiments further show that P. tricornutum strains successfully adapt to these low-salinity media over the course of several months. This result is in alignment with a recent study by Bulankova et al. (2021) showing the exceptional ability of P. tricornutum to adapt to stressful environments through genomic evolution. Together this study raises questions about the gene repertoires underlying P. tricornutum's capacity to grow on oil field PW and its bioremediation of elements like fluorine and copper. Such follow-up studies could provide the essential know-how for the generation of genetically modified strains that can be used for future economic production of biomass using oil field PW sources.