3. Results and discussion
3.1 Short-term BSA adhesion test
The effect of different coatings and materials on short-term BSA adhesion was evaluated as described in Section 2.3. The maximum adhesion (0.03 μg mm-2) was measured in HX (p <0.05, one-way ANOVA). The BSA adhesion values relative to HX are shown in Figure 1. The minimum adhesion was determined for H-X3, being one order of magnitude lower than that in the HX sample (p <0.05, one-way ANOVA) (Fig. 1A). There were no statistically significant differences between PMMA, PETG and GS (p <0.05, one-way ANOVA).
Protein adsorption on surfaces is the primary event in biofouling. Consequently, low protein adsorption should be recognized as the most important prerequisite for a surface’s resistance to biofouling. As PBR-grown microalgae are known to secrete proteins to the supernatant (Xiao and Zheng, 2016), protein adhesion is an expected process. The importance of this phenomenon in microalgae cultures was revealed in a recent study on N. gaditana adhesion dynamics (Zeriouh et al., 2017a). The surface of N. gaditana was clearly hydrophilic throughout the culture. However, the substrate surface (glass), which was initially hydrophilic, turned almost amphiphilic after two days of culture and strongly hydrophobic a week later. This behavior could only be compatible with the additional adhesion of other materials with hydrophobic properties such as proteins, most likely excreted byN. gaditana .
Consequently, PBR surfaces should be designed or selected to resist protein adsorption. However, the control of this multifactorial process is so complex that there is no known surface with complete protein resistance. The hydrophobicity and hydrophilicity of a surface, generally referred to as wettability, are important factors affecting protein adsorption (Firkowska‐Boden et al., 2018). The static water contact angle θw is commonly used as a descriptor of wettability. Figure 1B shows the relative BSA adhesion versusθw for each surface used in this study. A cutoff of θw ≈ 65° was inserted to distinguish between hydrophilic (θw <65°) and hydrophobic (θw >65°) surfaces, as previously reported (Vogler, 2012). The data did not present a clear pattern. The highest BSA adhesions were observed in the most hydrophobic surfaces (HX and NW). This is consistent with the literature since it is expected that proteins undergo partial unfolding and spreading on strongly hydrophobic surfaces (Firkowska‐Boden et al., 2018). In contrast, the glass surface (GS), which is strongly hydrophilic, did not show a marked efficacy in reducing BSA attachment compared to the other surfaces. It is supposed that hydrophilic surfaces promote easier protein desorption because the proteins are shielded from the surface by a high-density water layer (Firkowska‐Boden et al., 2018). Therefore, GS and PMMA failed as hydrophilic surfaces to inhibit or mitigate the protein adhesion. Surprisingly, BSA adhesion on PS, a moderately hydrophobic solid material, was lower than the hydrophilic surfaces (GS and PMMA). This odd result has so far remained a controversial issue in the field of protein adsorption. Recent studies have revealed that the extent of protein adsorption can be governed by the synergistic effect of the surface hydrophobicity and the relative charge state of the protein and surface (Attwood et al., 2019); so, just because a surface is hydrophilic does not mean it is protein resistant. In fact, globular protein adsorption onto hydrophobic solid surfaces occurs regardless of the protein charge state. Conversely, the adsorption charge of hydrophilic surfaces might be very high under charge-favorable conditions, or very low under unfavorable charge conditions (Attwood et al., 2019). Notwithstanding, in our study, the adhesion propensity of BSA molecules to both hydrophobic and hydrophilic surfaces was significant, the interaction being much stronger on the hydrophobic surfaces. Indeed, this property is exploited in immunology where BSA is used as a blocking agent on both hydrophobic and hydrophilic surfaces (Jeyachandran et al., 2009). Additionally, phosphate groups can be bound with the BSA molecules to form BSA-phosphate surface complexes, the preferred conformation of adsorbed BSA on surfaces (Jeyachandran et al., 2009). Phosphate is the P-source used in the N. gaditanaculture medium.
Interestingly, H-X3 (which is hydrophobic) had the lowest BSA adherence propensity. H-X3 is a fouling release coating based on silicone hydrogel. According to recent studies, comparing the protein adsorption results of hydrogels and other non-hydrogel materials, based on a scale of hydrophobicity and hydrophilicity, could be elusive because hydrogels absorb water and create a deformable surface (i.e. water-swollen coatings) (Vogler, 2012).
Another way to quantitatively represent the surface-thermodynamic hydrophobicity/hydrophilicity scale of condensed-phase materials is the interfacial free energy of interaction between two surfaces, i, immersed in water, ΔGiwi (commonly referred to as the free energy of cohesion) (Van Oss, 2008). Figure 1C shows the relative BSA adhesion values as a function of the corresponding ΔGiwi values for each assayed surface. Although the distribution of the experimental points was different to that of θw (Fig. 1B), the conclusions are similar. Only GS presented a clearly positive ΔGiwi value (ΔGiwi > 0), indicating that it is strongly hydrophilic. PMMA presented an ΔGiwi value close to 0, but as its \(\gamma_{s}^{-}\) value (27.7 mJ m-2) was not significantly greater than the 30 mJ m-2 cut-off, it is most likely hydrophobic, as suggested elsewhere (Van Oss, 2008).
From Figures 1B and 1C, one can conclude that criteria based on hydrophobicity and hydrophilicity are not enough to completely explain the protein adhesion on the surfaces. Protein adsorption behavior is a poorly understood and very complex process that can be influenced by diverse factors, acting either synergistically or antagonistically, such as protein molecular properties, surface chemistry and charge, wettability, and topography (Firkowska‐Boden et al., 2018). For example, polymers exhibiting similar surface hydrophobicity and roughness, differing only in the chemical end groups, can have divergent protein adsorption profiles (Vijaya Bhaskar et al., 2015). What is even more disconcerting is observing how the same polymer provided by different commercial suppliers can present very different protein adhesion values (Vijaya Bhaskar et al., 2015), likely attributed to minor differences in one or more of the above-mentioned factors, or to a variety of chemical residues generated during manufacturing, which cause the protein to interact differently with the groups on the surface (Contreras-Naranjo and Aguilar, 2019). Interestingly, it is well-known that protein adsorption increases along with the water’s ionic strength. Studies have reported that BSA adsorption on glass was markedly higher in seawater than in low ionic-strength buffer (Kirchman et al., 1989). The most common explanation is based on the double layer theory: the surface charge of BSA increases in seawater (i.e. a higher ionic strength), causing a reduced repulsive double-layer interaction and a more globular BSA configuration (Kirchman et al., 1989). The results from Figure 1B are consistent with others obtained at ionic strengths as high as seawater (Ruckenstein and Berim, 2019): the greatest BSA adsorption took place on the hydrophobic surfaces, was substantial on glass and significantly lower on the hydrogels. Recent observations support the hypothesis that BSA in seawater forms a multilayer structure involving Mg2+ cations as bridges between the BSA molecules (Pradier et al., 2002). Furthermore, it has been demonstrated that the salts dissolved in seawater produce micrometer-size surface spots containing mainly metal species, these spots being preferential adsorption sites for BSA (Poleunis et al., 2002).
Figure 2 summarizes in an idealized way how the interaction of both the protein and the surface properties may tune the adsorption process, as conceptualized elsewhere (Contreras-Naranjo and Aguilar, 2019). One can observe how the proteins’ versatile nature is related to their primary structure, ultimately represented by the sequence of amino acids and functional groups available for bonding. This hinders protein adsorption prevention and leads to experimental data spreading, generated from different laboratories studying the same protein-surface system but from different manufacturers.
3.2 Long-term biofouling tests
3.2.1 Culture experiments in the vessels
The biofouling tests consisted of exposing the coupons to the N. gaditana cultures for a long period (50 days), as explained in Section 2.4. Characterizing these cultures is essential for analyzing the adhesion results based on the interaction between the coupon surface and the broth that surrounds it. The representative culture kinetics for all the vessels are displayed in Figure 3. The experiments started as batch cultures (Stage S1) to acclimate the cells to the new set conditions in the vessels. After 18 days, the fed-batch mode was established. Two nutrient pulses were performed - on days 18 (Stage S2) and 33 (Stage S3). Given that the differences between the experiments that had an initial N/P from 45 to 90 were not significant compared to those with an N/P of 5 and 15, Figure 3 shows the mean values together of the three N/P 45 to 90. The highest growth rates and biomass concentrations were obtained at an N/P≥15 in all of the S1 to S3 stages (Fig. 3A).
As can be seen in Figure 3B, the dissolved phosphate was rapidly taken up soon after being added to the cultures with N/P>15 compared to N/P=5. At the end of each stage, the phosphate was almost completely consumed at N/P values above 5 (Fig. 3B) whereas the dissolved nitrate continued to be present in excess in all experiments (Fig. 3C). The culture dynamics presented above describe a growth pattern that is clearly controlled by the phosphate availability in the culture medium for the experiments with an N/P>15. Conversely, growth in the culture medium with an N/P of 5 seemed to be inhibited by phosphate. The Fv/Fmdid not change significantly throughout the culture period, the average value of all vessels being 0.549± 0.035, which is indicative of healthy cells; the one exception was the value of 0.4 with N/P 5 during S1 and S2.
The coupons were detached from the vessels at the end of stage S3 and the cell adhesion intensity was analyzed as a proxy for microalgal biofouling, as indicated in Section 2.4.2. The factors that could potentially affect the fluid dynamics of biofouling formation on the coupon surface are: the coupon position in the culture vessel (bottom or wall), the coupon surface properties represented by the coupon type, and the culture properties to which the coupons are exposed, represented by the N/P ratio. Therefore, to analyze the effect of these factors, and their interactions on the variability of the cell adhesion intensity (AI), a multifactor ANOVA was carried out. All factors and interactions had a statistically significant effect on AI at the 95.0% confidence level (p < 0.05). However, the most significant contributions to AI variability corresponded to the N/P (27.31%), the coupon type (23.78%) and the N/P interaction with the coupon position (12.71%). The N/P factor is a vehicle to combine into one variable the different alterations affecting the environment that directly interacts with the coupons (cells and supernatant) due to the effect of N on cell metabolism. Nutrient availability, biomass concentration and EOS concentration in the culture were taken into account as cellular responses. Figure 4 shows the effect of the above-mentioned factors on AI, EOS, and cell biomass concentrations at the end of the culture experiments.
The cell adhesion intensity and the secreted protein concentration varied inversely to N/P (see Fig. 4A), with no significant effect from N/P ratios above 15 (p <0.05), while an opposite effect on the final biomass concentration was observed. This implies that cell adhesion correlated positively with the concentration of proteins, but negatively with the concentration of biomass in the culture. The latter seems to contradict previous observations, where it was suggested that an increased cell flow density reaching a coupon surface, governed by cell concentration gradients, increased the cell adhesion rate (Zeriouh et al., 2019b). These seemingly contradictory results are easily reconciled if the protein adhesion that occurred during the formation of the conditioning film preceded cell adhesion.
Considering the results observed in the short-term BSA adhesion tests, the above assumption is highly plausible. A recent study supports this point, in which proteins secreted by the cyanobacterium Annabaenasp were first adsorbed onto suitable PBR materials during the formation of the conditioning film (Talluri et al., 2020). Figure 4A demonstrates that protein adsorption on the coupon surface increased as protein concentration increased in the supernatant, as reported for BSA on surfaces immersed in seawater (Kirchman et al., 1989).
The stickiness of proteins on the different surfaces tested was previously evidenced in Section 3.1, even on hydrophilic surfaces. Therefore, it is expected that N. gaditana will adhere to the coupons in proportion to the amount of proteins adhered to them, even on hydrophilic surface (Zeriouh et al., 2017a). This means that N. gaditana attachment on the coupon surface is mediated by prior protein adsorption, which occurs at a rate faster than N. gaditanaadhesion; that is to say, cells grown in media with N/P ratios of 5 and 15 interacted firstly with the coupon surfaces coated with a conditioning film composed of sticky proteins. In this scenario, the importance of the flow densities of nutrients reaching the cells that were attached to the coupon surface was greater in cultures with the lowest N/P ratios, where nitrates and phosphates were in excess (see Fig. 2B), while in cultures with high N/P ratios (above 15), phosphate was present in the supernatant at very low concentrations. As a result, in cultures with an N/P above 15, the microalgae adhering to the coupon surfaces (or embedded in the formed biofilm) continuously experienced significantly reduced phosphate flux densities, which limited their growth. The role played by the excess phosphate present in cultures with an N/P below 15 may also have been relevant during protein adsorption if the proteins secreted by N. gaditana had the capacity to bind with the phosphate groups to form protein-phosphate, as mentioned above for BSA. These explanations are consistent with Figure 4B, where the effect of the coupon surface properties (i.e. the type of coupon) on the cell adhesion intensity is observed. The resulting pattern resembled that of BSA adsorption on the same type of coupon represented in Figure 1A.
The effect of the coupon position on cell adhesion intensity is shown in Figure 4C. One can observe that the slight variability in the cell adhesion intensity associated with the coupons’ position (bottom or wall) mainly occurred in the experiments with an NP = 5, where the highest EOS excretion took place.
It has been suggested that wall shear influences cell attachment to the bottom well plates when agitated on a shaking table (Saleck et al., 2011).The strain rate values were always higher on the coupons attached to the flask wall than those on the bottom (Table 2).
All the wall values were almost twice those on the bottom. The differences were higher for the maximum (∼2.7 times) and minimum (∼6.2 times) values. It has been posited that the initial adhesion might be reversible if the kinetic energy of the microalgae cells in the boundary layer of the solid were superior to the total interaction energy between the solid surface and the microalgae cell. The strain rate (γ) 5 µm from the coupons is indicative of the drag force exerted by the liquid flow on the boundary layer of the solid surface (Zeriouh et al., 2017b). However, it has been suggested that the shear stress factor alone is not sufficient to explain what was observed occurring in the different attachment experiments. Nonetheless, when linked to flow field topology, it helps in understanding the underlying mechanisms of the different biological processes (Salek et al., 2011). Hence, it is necessary to consider both the strain rate values and the different flow structures developed on the bottom and walls of the vessel (Chakraborty et al., 2011). Usually, in orbitally agitated devices, the liquid height is low and a thin liquid film periodically sweeps the bottom, provoking the formation of a radial shear gradient (Salek et al., 2011). The trailing wave formed at the liquid surface forms a recirculation cell that increases the interaction of the cells with the vertical wall of the vessel (Salek et al., 2011). The liquid wetting the coupons on the walls and the bottom of the vessel does not have the flow influence associated with the liquid-gas interphase (Fig. 5B); therefore. the liquid velocity at the bottom is practically constant across the entire cross section (Fig. 5C). The coupons on the vertical wall, however, are closer to the liquid free surface and are subjected to a periodic velocity radial gradient, which generates higher strain rates (Fig. 5D and Table 2), very similar to that occurring in shallow agitated wells (Salek et al., 2011, Berson et al., 2008). In addition, the trailing wave of the liquid surface generates a recirculation cell that repeatedly throws the cells against the wall (Salek et al., 2011, Kim and Kizito, 2009), to some degree centrifuging the cells onto their surface (Kim and Kizito, 2009), which would promote cell adhesion. The different liquid flows developed in the different parts of the flasks apparently counteracted the different strain rate values and, as a consequence, cell attachment was similar on all the coupons.
When the protein concentration in the culture is very high (N/P = 5) (Fig. 4A), the fluid flow also seems to carry more proteins towards the wall coupons; therefore, interaction with the cells is higher, as is the cell attachment, compared to the coupons on the bottom (Fig. 4C).
3.2.2 Nitrogen balance
The synthesis of intracellular components and the release of extracellular organic substances is mediated by nitrogen availability to the cells. As the nitrogen added to a culture is not only fixed in proteins, the nitrogen balance needs to be acceptably closed. The nitrogen taken up by the cells can be distributed amongst the main extracellular and intracellular proteinaceous and non-proteinaceous nitrogenous compounds, namely proteins, amino acids, chlorophylls, nitrate, nitrite and ammonium.
The analyses carried out to determine the total nitrogen present in both the biomass and the supernatant have revealed that, in general, this element has been taken up in a similar way for all the N/P ratios: 10-13% has been used to form part of the biomass (both cells in suspension and those attached to the vessel), while 87-90% was used in the extracellular medium. Furthermore, one should point out that only 25-35% of the nitrogen was released into the medium as nitrate (not consumed), nitrite and ammonium; that is to say, the rest of the nitrogen (65-75%) went to forming part of the organic compounds (proteins, ammonia acids, chlorophylls, DNA fragments and lipids, etc.).
As for organic nitrogen, this element has also been distributed similarly for all the N/P ratios. The most appreciable difference is that at the N/P 5 ratio, 50% of the intracellular nitrogen is present in the adhered cells; this is because there are twice as many adhered cells as there are in suspension. This can be explained by the fact that in this culture, there was 10% more extracellular protein than in the other cultures and, as previously mentioned, this substance has a great influence on cell adhesion.
3.2.3 Analysis in the context of the Baier and Vogler biocompatibility theories
A work recently carried out on N. gaditana has demonstrated that the surface biocompatibility theories of Baier and Vogler may offer interesting insights into developing antifouling surfaces for microalgae photobioreactors (Zeriouh et al., 2019a). Figure 6 shows the relative adhesion values as a function of the water adhesion tension (Vogler’s theory) and the critical surface tension (Baier’s theory). The microalgae adhesion was calculated based on the maximum obtained at the minimum N/P ratio of 5 (4.06 ·105 cells cm2). For clarity’s sake, only the results corresponding to the N/P ratios that led to the highest (N/P=5) and the lowest (N/P=90) cell adhesion and secreted protein concentration values were represented. For the BSA protein model, the values represented in Figure 1 were used. One can observe that the Baier and Vogler theories are only reconciled at the Baier curve minimum attained by the H-X3 coating, based on FRCs‐Hydrogel. The exceptional antibiofouling properties of H-X3 have recently been demonstrated in N. gaditanacultures (Zeriouh et al., 2019a). H-X3, although initially having low energy surface properties (γc = 22 mJ m-2 and τ0 = −21.2 mJ m-2), can acquire amphiphilic properties after exposure to seawater; these properties are close to Vogler’s minimum (τ0 of around 35 mJ m-2), as reported by Zeriouh et al. (2019a) (see Figure 6). The BSA and cell adhesion intensity on the rest of the coupons in the vicinity of Vogler’s minimum were clearly several orders of magnitude higher than that of H-X3. As mentioned above, these analyses, which are based exclusively on surface wettability, are not universal. Even though there is a substantial body of literature dealing with the topic of protein adsorption on surfaces, in a recent excellent review, Vogler confirmed the huge lack of consensus and the impossibility of extracting general conclusions beyond a few general rules of thumb based on the biophysical chemistry of protein adsorption (Vogler, 2012).
Likewise, two further observations from Figure 6 are of particular importance: (i) the pattern of relative BSA adsorption intensity clearly resembles those of the different N/P ratios assayed; and (ii) the higher the concentration of proteins in the supernatant (high for 5 and low for 90, see Fig. 4), the higher the cell adhesion.
These observations support the assumption that the initial event that initiates the biological response of an N. gaditana culture to artificial materials and coatings is the adsorption of proteins secreted by the microalgae into the culture medium. This is in accordance with results from studies on photobioreactor materials used in the cultivation of the cyanobacterium Anabaena sp., where proteins were present on the conditioning films formed on glass and PMMA (Talluri et al., 2020).
3.2.4 Short-term BSA adhesion is indicative of the microalgal biofouling propensity
Given that protein-resistant materials and coatings are also likely to be resistant to microalgae adhesion, two experimental approaches were addressed in this study: the short-term batch adhesion of the sticky BSA model protein under dynamic conditions and the long-term microalgal adhesion under flow conditions that are representative of practical operation. Relative BSA adhesion tended to approximate the Baier curve and successfully mimic the long-term biofouling test results (Fig. 6). These results demonstrate that the short-term batch adhesion test of a sticky model protein such as BSA is a proxy for predicting biofouling in microalgae cultures. Therefore, long-term representative studies may not be mandatory as a preliminary step in assessing potential biofouling control in strategies applied to the massive screening of surfaces.
3.3 Future prospects
The work reported here indicates that, to effectively design or select materials and coatings for use in PBR construction, protein adsorption is a more critical event to consider than that of the cell/clean surface interaction. When PBR surfaces are exposed to cultures containing EOS, including proteins, they are likely to be rapidly coated by a thin layer comprised mainly of proteins dissolved in the supernatant which adsorb to the PBR surface as a monolayer or multilayer. The process of cell adhesion to the surface is indirect because the cells do not bind directly to the PBR surface but instead bind to the adsorbed proteins. According to Vogler (Vogler, 2012), adsorbed protein on water-surface interphases displaced an equivalent volume of interphase water (interphase dehydration), causing adsorbed proteins to concentrate into closely-packed arrangements. Microalgae make contact with, and anchor to, the adsorbed proteins in certain patches and domains.
Although BSA represents an interesting laboratory model for globular protein, the composition, structure, chemistry and functions of the proteins contained in EOS is, as yet, little known (Xiao and Zheng, 2016). Furthermore, the type and profile of the proteins secreted by cultured microalgae may be dependent upon the abiotic conditions and species. Consequently, to gain further insight, studies on the adsorption of other model globular proteins may be needed, as is recommended for animal cells (Attwood et al., 2019). In addition, a conditioning film may also comprise other macromolecules, such as humic substances, polysaccharides, or free amino acids present in the water (Flemming, 2002). All of these can be excreted by the microalgae at high concentrations in PBR cultures. Generally, the adsorption process of these macromolecules onto surfaces (within seconds or minutes) is faster than the cell adhesion (Flemming, 2002). Hence, the degree of initial macromolecule attachment might also be a good indicator of biofouling and its study should therefore be addressed.
Microalgae respond to changing environmental conditions by modulating their extracellular and intracellular metabolites. Within the same microalgal species, the optimal culture conditions vary depending on the target metabolite. Therefore, besides the N/P ratio in the culture medium, the possibilities of combining a wide variety of abiotic factors are vast. Each of these combinations most likely modifies the interaction between the PBR surface and the culture due to changes in the culture and cell properties. From a practical point of view, the biofouling propensity of different PBR surfaces should be studied specifically for each cultivation environment and species in order to the design and/or select the most appropriate surfaces or coatings. Likewise, transparent H-X3-like coatings should be developed and tested for PBR use in microalgae cultures where biofouling is likely dominated by EOS.