Flavonoid Glycosides with a Triazole Moiety for Marine Antifouling Applications: Synthesis and Biological Activity Evaluation

Over the last decades, antifouling coatings containing biocidal compounds as active ingredients were used to prevent biofouling, and eco-friendly alternatives are needed. Previous research from our group showed that polymethoxylated chalcones and glycosylated flavones obtained by synthesis displayed antifouling activity with low toxicity. In this work, ten new polymethoxylated flavones and chalcones were synthesized for the first time, including eight with a triazole moiety. Eight known flavones and chalcones were also synthesized and tested in order to construct a quantitative structure-activity relationship (QSAR) model for these compounds. Three different antifouling profiles were found: three compounds (1b, 11a and 11b) exhibited anti-settlement activity against a macrofouling species (Mytilus galloprovincialis), two compounds (6a and 6b) exhibited inhibitory activity against the biofilm-forming marine bacteria Roseobacter litoralis and one compound (7b) exhibited activity against both mussel larvae and microalgae Navicula sp. Hydrogen bonding acceptor ability of the molecule was the most significant descriptor contributing positively to the mussel larvae anti-settlement activity and, in fact, the triazolyl glycosylated chalcone 7b was the most potent compound against this species. The most promising compounds were not toxic to Artemia salina, highlighting the importance of pursuing the development of new synthetic antifouling agents as an ecofriendly and sustainable alternative for the marine industry.


Introduction
Marine biofouling, resulting from the accumulation of marine micro and macroorganisms on submerged surfaces, has been a huge problem for maritime industries, causing several technical and economic problems, including corrosion of materials and the increase in fuel consumption. Moreover, marine biofouling is associated with environmental and health problems, due to an increase in gas emissions and the spread of invasive species [1, 2].
Biocidal paints containing organotin compounds, namely tributyltin (TBT), were widely used for decades in the maritime industry to prevent biofouling. However, due to their negative effect on the environment and on live organisms, these substances were completely banned in 2008 by the international maritime organization [3]. Since then, some booster biocides, such as Irgarol 1051 or Sea-nine 211, in combination with copper, have been used; nevertheless, even these compounds have demonstrated toxicity on living organisms. The first step in the synthetic process to obtain glycosylated chalcones (Scheme 2) was the propargylation of 2,4-dihydroxyacetophenone with propargyl bromide. As for the synthesis of flavones 2a-2b, firstly this reaction was accomplished with propargyl bromide, in the presence of anhydrous Cs 2 CO 3 and tetrabutylammonium bromide (TBAB). Nevertheless, in addition to the desired 4-O-monosubstituted acetophenone (5), the 2,4-disubstituted acetophenone was obtained. Therefore, this reaction was performed in the presence of anhydrous K 2 CO 3 , as described by Zhao et al. [23], with slight modifications, and the 4-O-monosubstituted acetophenone (5) was successfully obtained as expected, with a 76% yield. Afterwards, the base-catalysed aldol reaction of this propargylated acetophenone with benzaldehydes afforded chalcones 6a and 6b with moderate yields, which were subsequently submitted to MW assisted CuAAC with azide sugar derivatives, affording triazole linked glycosylated chalcones 7a, 7b, 8a and 8b with 45-65% yield.

Mussel (Mytilus galloprovincialis) Larvae Anti-Settlement Activity
Mussels are one of the main macrofouling organisms present on ships and submerged maritime structures worldwide; thus, they are a target species used in settlement inhibition bioassays [28,29]. Due to the presence of a muscular sensory foot, mussel plantigrade larvae are highly specialized in adhesion to the submerged surfaces and the fixation is made through the production of byssal threads [30], which constitutes the endpoint of this bioassay. Therefore, for the evaluation of the AF activity of the compounds towards macrofouling species, the ability of the synthetized flavonoids to inhibit the settlement of Scheme 3. Synthesis of chalcones 10a-10b and 11a-11b. (i) K 2 CO 3 , acetone, 60 • C, 1 h, 84%; (ii) 40% NaOH, methanol, microwave (MW), 4 h, 33-47%; (iii) p-Toluenesulfonic acid (PTSA), methanol, 50 • C, 5 h, 24-31%.

Mussel (Mytilus galloprovincialis) Larvae Anti-Settlement Activity
Mussels are one of the main macrofouling organisms present on ships and submerged maritime structures worldwide; thus, they are a target species used in settlement inhibition bioassays [28,29]. Due to the presence of a muscular sensory foot, mussel plantigrade larvae are highly specialized in adhesion to the submerged surfaces and the fixation is made through the production of byssal threads [30], which constitutes the endpoint of this bioassay. Therefore, for the evaluation of the AF activity of the compounds towards macrofouling species, the ability of the synthetized flavonoids to inhibit the settlement of Mytilus galloprovincialis larvae at 50 µM was assessed. In this screening bioassay, in addition to glycosylated flavones 3a, 3b, 4a and 4b and chalcones 7a, 7b, 8a and 8b, non-glycosylated flavones 1a-b and 2a-b and chalcones 6a-b, 10a-b and 11a-b were tested in order to perform SAR studies. Results showed that among 18 tested flavonoids (10 chalcones and 8 flavones), seven chalcones (6a, 6b, 7b, 8a, 8b, 11a and 11b) and only three flavones (1b, 4a and 4b) presented a percentage of settlement ≤ 40%, suggesting that chalcone scaffold seems to be more promising for anti-settlement activity. These 10 compounds were further selected for dose-response studies in order to determine LC 50 /EC 50 values.

Quantitative Structure-Activity Relationship
Quantitative structure-activity relationship (QSAR) studies have been used for several years to point out small molecules' properties that are relevant for activity, and to forecast the activity of new compounds [32]. Therefore, a QSAR model was built to highlight the descriptors that are being relevant for anti-settlement activity against M. galloprovincialis plantigrades of the tested flavonoids. In this work, a 2D-QSAR model was elaborated using the Comprehensive Descriptors for Structural and Statistical Analysis (CODESSA 2.7.2) software package, which calculates approximately 500 descriptors. The heuristic method performs a pre-selection of descriptors by eliminating descriptors that are not available for each structure, that have a small variation in magnitude, that are correlated pairwise, and that have no statistical significance. The heuristic method is a very useful method for searching the best set of descriptors, without restrictions on the data set size [33].
The correlation coefficient (R 2 ), squared standard error (S 2 ), and Fisher's value (F) were used to evaluate the validity of regression equation [34]. As the rules of QSAR establish that there must be one descriptor for each five molecules used to build the model [34], three descriptors were used to build the QSAR equation. The multilinear regression analysis using Heuristic method for 15 compounds in the three-descriptor model is shown in Figure 1. The compounds are uniformly distributed around the regression line (Figure 1), which suggests that the obtained model has satisfactory predictive ability. The correlation coefficient (R 2 ), squared standard error (S 2 ), and Fisher's value (F) were used to evaluate the validity of regression equation [34]. As the rules of QSAR establish that there must be one descriptor for each five molecules used to build the model [34], three descriptors were used to build the QSAR equation. The multilinear regression analysis using Heuristic method for 15 compounds in the three-descriptor model is shown in Figure 1. The compounds are uniformly distributed around the regression line ( Figure  1), which suggests that the obtained model has satisfactory predictive ability. The best QSAR equation had a R 2 of 0.7945, Fisher value of 14.18, and S 2 of 0.0243, which reveals that the proposed model has statistical validity [35]. The R 2 is higher than 0.6, which is an indicator of a good fit to the regression line [36], representing close to 80% of the total variance in AF activity shown by the test compounds. The QSAR model is significant at a 95% level, as shown by the Fisher F-test (F = 14.18), which is higher than the tabulated value (3.59), as desired for a statistically significant model [35]. The squared standard deviation S 2 is small and close to zero (s 2 = 0.0243), proving that the model is significant and has low variation about the regression line [37]. The reliability of the resulting QSAR model was explored using two different types of validation criteria: external validation by using a test set and internal validation by leave-one-out (LOO) cross-validation [38]. The model was able to predict the activity of an external test set with an average difference of 0.19 from the experimental value [39]. Moreover, the cross-validated R 2 (Q 2 = 0.5953) from the LOO internal validation process is higher than 0.5 and smaller than the overall R 2 , as expected, and the difference between R 2 and Q 2 is lower than 0.3, which The best QSAR equation had a R 2 of 0.7945, Fisher value of 14.18, and S 2 of 0.0243, which reveals that the proposed model has statistical validity [35]. The R 2 is higher than 0.6, which is an indicator of a good fit to the regression line [36], representing close to 80% of the total variance in AF activity shown by the test compounds. The QSAR model is significant at a 95% level, as shown by the Fisher F-test (F = 14.18), which is higher than the tabulated value (3.59), as desired for a statistically significant model [35]. The squared standard deviation S 2 is small and close to zero (s 2 = 0.0243), proving that the model is significant and has low variation about the regression line [37]. The reliability of the resulting QSAR model was explored using two different types of validation criteria: external validation by using a test set and internal validation by leave-one-out (LOO) cross-validation [38].
The model was able to predict the activity of an external test set with an average difference of 0.19 from the experimental value [39]. Moreover, the cross-validated R 2 (Q 2 = 0.5953) from the LOO internal validation process is higher than 0.5 and smaller than the overall R 2 , as expected, and the difference between R 2 and Q 2 is lower than 0.3, which indicates that the model does not suffer from overfitting [40].
By interpreting the molecular descriptors in the regression model (Figure 1), it is possible to have some insight into structural characteristics that are likely to be responsible for AF activity of the studied compounds. There are three descriptors included in the regression model, which proved to be important features and provide statistically significant contributions to the QSAR equation.
As indicated by the higher t-test value, hydrogen bonding acceptor ability of the molecule (HACA1) is a charged partial surface area (CPSA) descriptor that appeared as the most significant descriptor for the obtained QSAR model, contributing positively to the AF activity [41]. HACA1 is determined by the equation: where S A stands for solvent-accessible surface area of H-bonding acceptor atoms, selected by threshold charge. This descriptor proves the importance of the hydrogen bonding acceptor properties for the activity of the test compounds [42].
The topological descriptor average complementary information content of order 2 (CIC2) descriptor is predicted as being negatively implied in the AF activity of the test compounds [41]. The CIC2 descriptor represents the difference between the maximum possible complexity of a molecule and its real topological information. It belongs to the multi-graph information content indices and it describes neighborhood symmetry of second order [43]. The constitutional descriptor number of triple bonds is also responsible for a decrease in activity.
The molecular descriptors used in the QSAR model demonstrate that the mechanism underlying the AF activity of flavonoids is mainly related to their HACA1, and it may be prejudiced by topological CIC2 and by the presence of triple bonds. Interestingly, the triazolyl glycosylated chalcone 7b, with the most promising anti-settlement activity, is one of the compounds with more hydrogen-bonding acceptors. In contrast, propargylated flavones 2a and 2b had a percentage of settlement higher than 40% at 50 µM, and therefore were not selected for dose-response studies and for the determination of the LC 50 /EC 50 values. Moreover, propargylated chalcones 6a and 6b showed the lowest activity. Overall, the examination of the molecular descriptors reported in this work can lead to a better understanding of the relation between the structure and AF activity of flavonoids.

Biofilm-Forming Marine Bacteria Growth Inhibitory Activity
Although the macrofouling species represent the most problematic component of fouling in terms of biomass and negative repercussions, the first micro-colonizers are also of extreme importance, since they represent the basis of the fouling community, and ultimately, they may modulate the colonization of further species by inducing or inhibiting species adhesion via biochemical cues [44]. Thus, synthesized flavonoids were further evaluated for their ability to inhibit the growth of five marine biofilm-forming bacteria, Vibrio harveyi, Cobetia marina, Halomonas aquamarina, Pseudoalteromonas atlantica and Roseobacter litoralis.
Results showed that only the bacterial growth of Roseobacter litoralis was meaningfully compromised by tested compounds, with significant inhibitory activity for propargylated chalcones 6a and 6b ( Figure 2). These compounds were selected for concentration-response analysis ( Figure 3). Mar. Drugs 2021, 19, x FOR PEER REVIEW 8 of 22    Compounds 6a and 6b presented low anti-bacterial activity towards R. litoralis with EC30 values of 135 and 83.5 μM, respectively. Compounds 6a and 6b presented low anti-bacterial activity towards R. litoralis with EC 30 values of 135 and 83.5 µM, respectively.

Biofilm-Forming Marine Diatoms Growth Inhibitory Activity
The most promising compounds regarding anti-settlement activity (1b, 7b, 11a, 11b) were further evaluated for their ability to inhibit the growth of the biofilm-forming Mar. Drugs 2021, 19, 5 9 of 21 microalgae Navicula sp. This marine diatom is a major biofouling species that very effectively colonizes submerged surfaces by secreting adhesive extracellular polymer substances (EPS), and thus is a good representative of fouling microalgae.
Only triazolyl glycosylated chalcone 7b showed significant inhibitory activity with the concentration-response analyses revealing an EC 50 value of 41.76 µM; 30.94 µg·mL −1 , suggesting the ability of this compound to act also as a promising AF agent against microfouling species.

In Vitro Acetylcholinesterase (AChE) and Tyrosinase (Tyr) Activities
The identification of the mechanism of action associated with AF activity remains a challenge for the scientific community. According to Qian et al. (2013) antifoulants appear to affect settlement through distinct pathways, which can be classified roughly into several categories such as inhibitors of ion channel function, inhibitors of quorum sensing, blockers of neurotransmission or inhibitors of adhesive production or release [45]. Moreover, some specific target molecules in fouling organisms have been determined, such as AChE, which seems to be involved in cholinergic neural signaling during the settlement [46]. It is known that the commercial booster biocide Sea-Nine 211 acts by this mechanism [47,48], as well as two natural compounds isolated from marine organisms, territrem A and pulmonarin [49,50]. For this reason, the ability of the most promising compounds to modulate the activity of AChE was evaluated (1b, 7b, 11a and 11b). AChE activity was significantly induced for chalcones 7b and 11b ( Figure 4). Induced AChE activity has been described as an exposure effect that is in some cases associated with apoptosis [51], and thus the specific target behind these compounds' bioactivity should be further explored in future work.

Biofilm-Forming Marine Diatoms Growth Inhibitory Activity
The most promising compounds regarding anti-settlement activity (1b, 7b, 11a, 11b) were further evaluated for their ability to inhibit the growth of the biofilm-forming microalgae Navicula sp. This marine diatom is a major biofouling species that very effectively colonizes submerged surfaces by secreting adhesive extracellular polymer substances (EPS), and thus is a good representative of fouling microalgae.
Only triazolyl glycosylated chalcone 7b showed significant inhibitory activity with the concentration-response analyses revealing an EC50 value of 41.76 μM; 30.94 μg·mL −1 , suggesting the ability of this compound to act also as a promising AF agent against microfouling species.

In Vitro Acetylcholinesterase (AChE) and Tyrosinase (Tyr) Activities
The identification of the mechanism of action associated with AF activity remains a challenge for the scientific community. According to Qian et al. (2013) antifoulants appear to affect settlement through distinct pathways, which can be classified roughly into several categories such as inhibitors of ion channel function, inhibitors of quorum sensing, blockers of neurotransmission or inhibitors of adhesive production or release [45]. Moreover, some specific target molecules in fouling organisms have been determined, such as AChE, which seems to be involved in cholinergic neural signaling during the settlement [46]. It is known that the commercial booster biocide Sea-Nine 211 acts by this mechanism [47,48], as well as two natural compounds isolated from marine organisms, territrem A and pulmonarin [49,50]. For this reason, the ability of the most promising compounds to modulate the activity of AChE was evaluated (1b, 7b, 11a and 11b). AChE activity was significantly induced for chalcones 7b and 11b ( Figure 4). Induced AChE activity has been described as an exposure effect that is in some cases associated with apoptosis [51], and thus the specific target behind these compounds' bioactivity should be further explored in future work. A well-known pathway in the production of biological adhesives of mussels is the 3,4-dihydroxyphenyl-L-alanine (L-DOPA) metabolism that functions in the production of DOPA-containing mussel byssal plaques by the action of Tyr that catalyses the conversion of DOPA precursor into DOPA residues [46,52]. Considering this, the most promising compounds in the inhibition of mussel adhesion were tested for their ability to inhibit Tyr ( Figure 5). Results show that flavone 1b is able to significantly decrease Tyr activity at all the concentrations tested, reaching 23.5% of inhibition at 100 μM. Therefore, the inhibition A well-known pathway in the production of biological adhesives of mussels is the 3,4-dihydroxyphenyl-L-alanine (L-DOPA) metabolism that functions in the production of DOPA-containing mussel byssal plaques by the action of Tyr that catalyses the conversion of DOPA precursor into DOPA residues [46,52]. Considering this, the most promising compounds in the inhibition of mussel adhesion were tested for their ability to inhibit Tyr ( Figure 5). Results show that flavone 1b is able to significantly decrease Tyr activity at all the concentrations tested, reaching 23.5% of inhibition at 100 µM. Therefore, the inhibition of this enzyme, with a crucial effect in the formation of mussel adhesive, could be one of the mechanisms involved in the inhibition of the mussel settlement. This also highlights a specific AF mode of action related with mussel adhesion and explains the absence of activity against bacteria and diatoms. of this enzyme, with a crucial effect in the formation of mussel adhesive, could be one of the mechanisms involved in the inhibition of the mussel settlement. This also highlights a specific AF mode of action related with mussel adhesion and explains the absence of activity against bacteria and diatoms.

Environmental Fate Parameters: Artemia Salina Ecotoxicity Bioassay
Ecotoxicity assays carried out on non-target organisms aim to understand how tested compounds can affect sensitive non-target organisms and influence the health status of the surrounding ecosystem [53]. Artemia salina is a species of small crustaceans that live in salty marine environments and are used as test organisms because of their easy culture, short generation time, cosmopolitan distribution and commercial availability of their eggs in latent form [54].
Ecotoxicity results showed that the most promising compounds 1b, 7b, 11a and 11b are non-toxic to Artemia salina (less than 10% mortality) at both concentrations tested (25 and 50 μM) (Figure 6), in contrast to the commercial AF agent ECONEA ® which was previously shown by our group to cause 100% lethality at the same concentrations and conditions [9]. These results suggest that any of the tested compounds could be a good alternative, being more environmentally compatible antifoulants.

Environmental Fate Parameters: Artemia Salina Ecotoxicity Bioassay
Ecotoxicity assays carried out on non-target organisms aim to understand how tested compounds can affect sensitive non-target organisms and influence the health status of the surrounding ecosystem [53]. Artemia salina is a species of small crustaceans that live in salty marine environments and are used as test organisms because of their easy culture, short generation time, cosmopolitan distribution and commercial availability of their eggs in latent form [54].
Ecotoxicity results showed that the most promising compounds 1b, 7b, 11a and 11b are non-toxic to Artemia salina (less than 10% mortality) at both concentrations tested (25 and 50 µM) (Figure 6), in contrast to the commercial AF agent ECONEA ® which was previously shown by our group to cause 100% lethality at the same concentrations and conditions [9]. These results suggest that any of the tested compounds could be a good alternative, being more environmentally compatible antifoulants. the mechanisms involved in the inhibition of the mussel settlement. This also highlights a specific AF mode of action related with mussel adhesion and explains the absence of activity against bacteria and diatoms.

Environmental Fate Parameters: Artemia Salina Ecotoxicity Bioassay
Ecotoxicity assays carried out on non-target organisms aim to understand how tested compounds can affect sensitive non-target organisms and influence the health status of the surrounding ecosystem [53]. Artemia salina is a species of small crustaceans that live in salty marine environments and are used as test organisms because of their easy culture, short generation time, cosmopolitan distribution and commercial availability of their eggs in latent form [54].
Ecotoxicity results showed that the most promising compounds 1b, 7b, 11a and 11b are non-toxic to Artemia salina (less than 10% mortality) at both concentrations tested (25 and 50 μM) (Figure 6), in contrast to the commercial AF agent ECONEA ® which was previously shown by our group to cause 100% lethality at the same concentrations and conditions [9]. These results suggest that any of the tested compounds could be a good alternative, being more environmentally compatible antifoulants.

Synthesis of Flavones 1a and 1b
A mixture of phloroglucinol (0.175 g, 1.39 mmol) and ethyl 3,4-dimethoxybenzoylacetate (0.700 g, 2.78 mmol) or ethyl 3,4,5-trimethoxybenzoylacetate (0.739 g, 2.78 mmol) was heated at 240 • C in muffle furnace (Thermo Fisher Scientific, Oeiras, Portugal) for 60-100 min. Afterwards, the crude mixture was dissolved in 10% NaOH (20 mL) and washed with diethyl ether (2 × 20 mL), and the product was precipitated by adding 37% HCl. The solid was filtered and washed with water, and the flavones 1a and 1b were obtained with 74% and 77% yields, respectively. The structure elucidation of compounds 1a and 2b was established by 1 H and 13 C NMR techniques and data were in accordance with previously reported results [19].

Synthesis of Propargyloxychalcones 6a and 6b
To a solution of 5 (0.350 g, 1.84 mmol) in methanol (20 mL) was added a solution of 40% NaOH in methanol, until pH 14, under stirring. Afterwards, 3,4-dimethoxybenzaldehyde (0.612 g, 3.68 mmol) or 3,4,5-trimethoxybenzaldehyde (0.772 g, 3.68 mmol) was slowly added to the reaction mixture. The reaction was submitted to MW irradiation at 180 W at 70 • C for 4 h. After, a solution of 10% HCl was added until pH 5, and the obtained solid was filtered, washed with water, and purified by crystallization with methanol, giving rise to chalcone 6a and 6b with 41% and 43% yield, respectively. The structure elucidation of compounds 6a and 6b was established by 1 H and 13 C NMR techniques and data were in accordance with previously reported results [23].  20.8,20.7,20.6,20.3 (2 ,3 ,4 ,

Synthesis of Chalcones 10a and 10b
To a solution of 9 (0.500 g, 2.55 mmol) in methanol (20 mL) a solution of 40% NaOH in methanol was added until pH 14, under stirring. Then, a solution of 3,4dimethoxybenzaldehyde (0.847 g, 5.10 mmol) or 3,4,5-trimethoxybenzaldehyde (1.00 g, 5.10 mmol) in methanol was slowly added to the reaction mixture. The reaction was submitted to MW irradiation at 180 W at 70 • C for 4 h. Then, a solution of 10% HCl was added until pH 5, and the obtained solid was filtered and washed with water and purified by crystallization with methanol, giving rise to 10a and 10b with 33% and 47% yield, respectively. The structure elucidation of both compounds was established by 1 H and 13 C NMR techniques and data of 10a were in accordance with previously reported results [24]. Although the synthesis of compound 10b has been previously reported [25], the NMR data are described here for the first tim€(E)-1-(2-hydroxy-4-(methoxymethoxy)phenyl)-3- To a solution of 10a (0.200 g, 0.58 mmol) or 10b (0.250 g, 0.67 mmol) in methanol (10 mL), p-toluenesulfonic acid monohydrate (0.110 g, 0.58 mmol or 0.127 g, 0.67 mmol) was added. The reaction was submitted to conventional heating at 50 • C for 5 h. After the addiction of 10 mL of water, methanol was evaporated, and the aqueous solution was extracted with ethyl acetate (2 × 20 mL). The organic phase was washed with water (1 × 20 mL), dried over anhydrous sodium sulphate and concentrated under reduced pressure, giving rise to an orange solid. The crude product was purified by flash column chromatography (n-hexane: ethyl acetate, 8:2) (11a) or crystallization with chloroform (11b), giving rise to chalcone 11a and 11b with 24% and 31% yield, respectively. The structure elucidation of compounds 11a and 11b was established by 1 H and 13 C NMR techniques and data were in accordance with previously reported results [56,57].
The flavonoids were screened at 50 µM in 24-well microplates with 4 well replicates per condition and 5 larvae per well, for 15 h, in the darkness at 18 ± 1 • C, following   [58]. Test solutions were obtained by dilution of the compounds stock solutions (50 mM) in DMSO and prepared with filtered seawater. All bioassays included a negative control with DMSO and a positive control with CuSO 4 , a potent AF agent. After the exposure period, the anti-settlement activity was determined by the presence/absence of attached byssal threads produced by each individual larvae.
All compounds that caused more than 60% of settlement inhibition (≤40% of settlement) in the screening bioassay were considered active and selected for the determination of the semi-maximum response concentration that inhibited 50% of larval settlement (EC 50 ), at compounds concentrations of 3.12, 6.25, 12.5, 25, 50, 100, 200 µM.

Quantitative Structure-Activity Relationship
The eighteen flavonoid derivatives (1a, 1b, 2a, 2b, 3a, 3b, 4a, 4b from Scheme 1; 6a, 6b, 7a, 7b, 8a, 8b from Scheme 2; and 10a, 10b, 11a, 11b from Scheme 3) were used to build a QSAR model using the experimental data obtained from the mussel (Mytilus galloprovincialis) larvae anti-settlement activity in vivo bioassay (AF activity = log(100/%settlement). AF activity was selected as a dependent variable in the QSAR analysis. The 18 molecules were randomly distributed into a training set (15 molecules) and a test set (3 molecules). CODESSA software (version 2.7.10, University of Florida, Gainesville, FL, USA) was used to calculate more than 500 constitutional, topological, geometrical, electrostatic, quantum-chemical and thermodynamical molecular descriptors [59]. The heuristic multilinear regression methodology was chosen to perform a complete search for the best multilinear correlations with a multitude of descriptors of the training set [60]. The 2D-QSAR model with the best square of the correlation coefficient (R 2 ), F-test (F), and squared standard error (S 2 ) was selected. The final model was further validated using the test set and leave-one-out (LOO) internal validation.

Inhibitory Activity against Biofilm-Forming Marine Bacteria Growth
For anti-bacterial screening, five strains of marine biofilm-forming bacteria from the Spanish Type Culture Collection (CECT): Cobetia marina CECT 4278, Vibrio harveyi CECT 525, Halomonas aquamarina CECT 5000, Pseudoalteromonas atlantica CECT 570, and Roseobacter litoralis CECT 5395 were used. Bacteria were inoculated and incubated for 24 h at 26 • C in marine broth (Difco) at an initial density of 0.1 (OD600) in 96 well flat-bottom microtiter plates and exposed to the test compounds at 15 µM. Test solutions were obtained by dilution of the compounds stock solutions (50 mM) in DMSO. Bacterial growth inhibition in the presence of the compounds was determined in quadruplicate at 600 nm using a microplate reader (Biotek Synergy HT, Vermont, USA). Negative and positive controls used were a solution of marine broth with DMSO, and a solution of marine broth with penicillin-streptomycin-neomycin, respectively. Compounds exerting a significant antibacterial activity (Dunnet test, p < 0.05) in the screening bioassays were selected for the determination of the effective inhibitory concentration (EC 50 ).

Inhibitory Activity against Biofilm-Forming Marine Diatom Growth
The anti-microalgal activity of the most promising compounds was also evaluated against a benthic marine diatom, Navicula sp., purchased from the (Telde, Gran Canaria) Spanish Collection of Algae (BEA). Diatom cells were inoculated in f/2 medium (Sigma) at an initial concentration of 2-4 × 106 cells mL −1 and grown in 96-well flat-bottom microtiter plates for 10 days at 20 • C. Navicula growth inhibition in the presence of each compound at 15 µM was determined in quadruplicate and quantified based on the difference in cell densities among the treatments, and cells were counted using a Neubauer counting chamber. A positive control with cycloheximide (3.55 µM) and a negative control with f/2 medium 0.1% DMSO were included. Compounds that showed significant inhibitory activity in the screening assay (Dunnet test, p < 0.05) were selected for further determination of their effective inhibitory concentrations (EC 50 ).

In Vitro Acetylcholinesterase (AChE) and Tyrosinase (Tyr) Activities
The ability of the most promising compounds to inhibit AChE and Tyr was tested to assess their potential mode of action related with neurotransmission disruption or impairment of adhesive metabolism pathways, respectively.
Tyr inhibition assay was performed using Agaricus bisporus Tyr (EC1.14.18.1) according to Adhikari et al. (2008) [63] with some modifications [8]. The enzymatic reaction follows the catalytic conversion of L-Dopa to dopaquinone and the formation of dopachrome by measuring the absorbance at 475 nm. Briefly, Tyr (25 U/mL) was added to 50 mM phosphate buffer pH 6.5 and the tested compounds at 25, 50 and 100 µM (final concentrations, 1% DMSO). The enzymatic activity was triggered by the addition of L-dopa (25 mM). Kojic acid (1.4 mM, water) was included as positive control and 1% DMSO in water as negative control.

Environmental Fate Parameters: Artemia Salina Ecotoxicity Bioassay
The brine shrimp (Artemia salina) nauplii lethality test was used to determine the toxicity of promising AF compounds to non-target organisms [64]. Artemia salina eggs were allowed to hatch in seawater for 48 h at 25 • C. Bioassays were performed in 96-wells microplates with 15-20 nauplii per well and 200 µL of the compounds test solution. Compounds were tested at final concentrations of 25 and 50 µM (filtered seawater with 1% DMSO). All tests included K 2 Cr 2 O 7 (13.6 µM) as positive control and DMSO (1%) as negative control. Bioassays were run in the dark at 25 • C, and the percentage of mortality was determined after 48 h of exposure.

Statistical Analysis
Datasets from anti-settlement, antibacterial and anti-microalgal bioassay, and determination of AChE and Tyr activities, were analysed by one-way analysis of variance (ANOVA) followed by a multi-comparisons Dunnett's test against negative control (p < 0.05). For the AF bioassays, the half maximum response concentration (EC 50 ) values for each compound, when applicable, were calculated using Probit regression analysis. Significance was considered at p < 0.01, and 95% lower and upper confidence limits (95%LCL; UCL). The software IBM SPSS Statistics 26 (Armonk, New York, USA) was used for statistical analysis.

Conclusions
In this study, eight new triazole-flavonoid hybrids were synthesized using the click chemistry approach. From the series of synthesized compounds, flavone 1b and chalcones 7b, 11a and 11b showed significant anti-settlement activity towards the macrofouling species Mytilus galloprovincialis adhesive larvae. Regarding the compounds' structures, HACA1 was the most significant descriptor for the obtained QSAR model, contributing positively to the AF activity. Particularly, triazolyl glycosylated chalcone 7b, with a high number of hydrogen-bonding acceptors, showed the most effective anti-settlement activity (EC 50 = 3.28 µM; 2.43 µg·mL −1 ) with the highest therapeutic ratio (LC 50 /EC 50 > 60.98), exhibiting also a significant inhibitory activity against the marine diatom Navicula sp.
(EC 50 = 41.76 µM; 30.94 µg·mL −1 ), suggesting potential in the suppression of biofouling colonization succession. Flavone 1b, which was effective against the settlement of mussel larvae, also showed capacity to inhibit the activity of Tyr, which might explain the specific AF activity against mussel larvae. Ecotoxicity studies on the non-target species Artemia salina revealed that the flavonoids 1b, 7b, 11a and 11b did not show ecotoxicity to the nauplii of this sensitive crustacean, even at 50 µM, a concentration much higher than their EC 50 . These results disclosed synthetic flavonoids, particularly a new chalcone incorporating a 1,2,3-triazole ring (7b), with potential to be a good environmentally compatible alternative to the majority of the antifoulants in use. Flavonoids are ubiquitous in Nature and, therefore, they come with the advantage that they have been selected during evolution to have high specificity, high efficiency and some might be potential nontoxic inhibitors of fouling. Natural compounds are usually biodegradable, not leaving residue in the environment, and are thus considered one of the most promising alternatives to the biocides in use. However, the yields of natural compounds from marine organisms are generally poor, hindering their development as AF agents. Moreover, optimizing a micro-organism for enhanced production of antifoulant is generally laborious and time consuming. Synthesis of nature-like antifoulants seems to be a more sustainable way to create an opportunity to produce commercial supplies for the antifouling industry.