Correlation between Perturbation of Redox Homeostasis and Antibiofilm Capacity of Phytochemicals at Non-Lethal Concentrations

Biofilms are the multicellular lifestyle of microorganisms and are present on potentially every type of biotic or abiotic surface. Detrimental biofilms are generally targeted with antimicrobial compounds. Phytochemicals at sub-lethal concentrations seem to be an exciting alternative strategy to control biofilms, as they are less likely to impose selective pressure leading to resistance. This overview gathers the literature on individual phytocompounds rather than on extracts of which the use is difficult to reproduce. To the best of our knowledge, this is the first review to target only individual phytochemicals below inhibitory concentrations against biofilm formation. We explored whether there is an overall mechanism that can explain the effects of individual phytochemicals at sub-lethal concentrations. Interestingly, in all experiments reported here in which oxidative stress was investigated, a modest increase in intracellular reactive oxygen species was reported in treated cells compared to untreated specimens. At sub-lethal concentrations, polyphenolic substances likely act as pro-oxidants by disturbing the healthy redox cycle and causing an accumulation of reactive oxygen species.


Sub-Lethal Concentrations of Antibiofilm Phytochemicals as an Innovative Strategy against Microbial Resistance
Biofilms are the multicellular lifestyle of microorganisms and are present on potentially every type of biotic or abiotic surface. Some biofilms are beneficial to humans in a variety of fields, including energy production [1], bioremediation [2] and organism health [3]. In contrast, other biofilms are a threat to humans [4] and other organisms [5], and to human-made systems [6]. Focusing on the underlying mechanisms of biofilm formation in response to environmental cues is pivotal, as a critical characteristic of detrimental biofilms is their extremely enhanced capability to resist antimicrobial agents in comparison to planktonic cells.
Antimicrobial agents are the traditional approach to combat biofilms that cause deterioration or diseases or that compromise functional performance within an industrial setting. However, biofilms can develop resistance to antimicrobials, meaning that a substance is no longer effective in targeting the detrimental biological agents [7,8]. For instance, some Listeria monocytogenes strains are now increasingly resistant due to the excessive use of antimicrobial agents in animal and food production and human medicine [9]. The idea that is emerging is one of simply disarming detrimental microorganisms rather than killing them, to avoid antimicrobial resistance.
Many natural bioactive substances targeting virulence/detrimental factors seem to be an exciting strategy to control biofilms, since they are less likely to impose selective pressure leading to resistance, as they can work as biocide-free agents [10]. This strategy addresses the sustainable management of biofilm challenges. Many organisms produce antibiofilm Different compounds have different MICs against specific target microorganisms. In the work by Rózalska and colleagues [17], ursolic acid always showed a lower MIC than ellagic acid, epicatechin and quercetin when S. aureus ATCC 43300, S. aureus H9, Candida albicans ATCC 10231 and C. albicans A4 were considered. A concentration below the MIC (4 µg/mL) of ellagic acid reduced biofilm development of Streptococcus dysgalactiae NCTC 4671 up to 27% and S. dysgalactiae ATCC 27957 up to 21%, but 4 µg/mL had no effect on S. aureus biofilm growth [18]. In addition, even with the same phytocompound, it is not always possible to gather unique conclusions as different investigations were often carried out with variations in the methodologies, i.e., differences in sub-lethal doses, microbial targets, microbial cell density, and culture media [19]. Time could also be a key factor. Some plant-derived compounds were proven to exhibit a time-dependent antibiofilm effect. For instance, cannabidiol at the sub-MIC dose of 6.25 µg/mL did not decrease the biofilm mass after 1 day, while inhibition of biofilm mass by 28% and 39% was seen after 2 and 3 days, respectively [20]. Antibiotics alone at sub-MIC doses are claimed to promote biofilm formation and the development of antibiotic resistance [21]. The synergic effect of phytochemicals at sub-lethal doses and drugs has been tested and found very promising. Thyme oil and its major component thymol displayed significant synergy with fluconazole against Candida biofilms. The findings were remarkably relevant, as the above drugs, when used alone, may require too high doses, resulting in increased adverse side effects such as in vivo toxicity [22]. Similarly, non-lethal zosteric acid concentrations strongly enhanced sensitivity towards chlorhexidine digluconate, with most of the Candida cells being killed within 2 h [23]. Upon 40 µg/mL quercetin treatment, Klebsiella pneumoniae, P. aeruginosa and Yersinia enterocolitica displayed a >90% increase in sensitivity to tetracycline, kanamycin, and gentamycin [24]. According to Abdelraheem et al. [19], vitamin C (ascorbic acid) had a synergistic effect with antibiotics against P. aeruginosa both in vitro and in vivo. The authors suggested that vitamin C should be consistently prescribed with antibiotics to treat Pseudomonas infections to reduce treatment duration and antibiotic dose.
Some papers report the employment of plant extracts. The violacein production in Chromobacterium violaceum 12472 was reduced by more than 80% by using an extract of Plumbago zeylanica, without inhibiting planktonic cell growth [25]. In the above research, as in others involving extracts [26,27], many plant metabolites were detected [28][29][30] but the isolation and study of the most bioactive compounds were not always pursued. Essential oils are also used at sub-lethal concentrations, but they are not always composed of a single compound [31].
Many studies on the effects on biofilms of phytochemical treatment at sub-MICs were retrieved from the literature, and a few are listed in Table 1. Importantly, the characteristics of extracts are largely influenced by the source and the extraction method used [32] and it is often impossible to reproduce the same results in other laboratories. Consequently, in Table 1 as well as in the list of individual phytochemicals below, only the literature considering the investigation of individual plant compounds is reported. For some individual phytocompounds, specific effects have been reported that did not compromise the microbial growth of the target organism while exhibiting antibiofilm activity. The following alterations were detected after exposure to the plant-based compound below the MIC: (1) QS (cinnamaldehyde in Niu et al. [56]; tannic acid, ellagic acid and (−)-epigallocatechin gallate in Huber et al. [57]; cyanidin in Gopu and Shetty [37]; caffeine in Maisarah Norizan et al. [58]); (2) ability to adhere to a surface (shikonin in Li et al. [15]); (3) mobility (piperine in Das et al. [47]); (4) production of extracellular polymers (limonene and beta-caryophyllene in Tomaś et al. [59]); (5) translocation apparatus (glycone (myricitrin, hesperidin and phloridzin) and aglycone flavonoids (myricetin, hesperetin and phloretin) in Lopes et al. [48]); (6) expression of protective proteins (zosteric acid in Villa et al. [60]) and genes related to pathogenicity (salicylic acid, p-hydroxybenzoic acid, cinnamyl alcohol, p-coumaric acid, and hydrocinnamic acid in Hu et al. [61]); (7) replicational/transcriptional activity (berberine in Wang et al. [41]). Da et al. [62] studied the effects of salicylic acid on P. aeruginosa PAO1 and its QS mutant. Interestingly, they reported that the mutant also exhibited a salicylic acid-dependent biofilm inhibition. The researchers suggested that the antibiofilm effect of salicylic acid depended on more than QS disruption and a possible explanation was the iron chelating ability of the phytochemical.
In this review, we want to explore whether there is an overall mechanism that can explain the effects of individual phytochemicals at sub-lethal concentrations.

Reactive Oxygen Species (ROS)
The atmospheric oxygenation of Earth started ∼2.43 billion years ago [63]. Oxygenic photosynthesis led to the oxidation of reductants. At that time, soluble and relatively bioavailable ferrous iron (Fe 2+ ) was the prevailing marine inorganic reductant. Therefore, oceans experienced a 4-fold decrease in iron level due to the insolubility of the oxidized ferric form [64]. Since iron in the ferric form is extremely insoluble, it became scarce and a limiting factor within several ecosystems [65]. Despite the reliance on iron-based chemistry being threatened by oxygenic photosynthesis, contemporary enzymes still rely on iron as a cofactor. Iron is involved in the catalysis of all the redox transitions necessary for sulfur and nitrogen assimilation and many metabolic pathways, e.g., all types of respiration [64].
The increase in O 2 not only led to changes in metal iron availability but also forced cells to cope with its toxicity. Oxygen has a tendency to acquire electrons from other molecules, although the potency of oxygen to acquire one electron at a time leads, firstly, to the formation of the superoxide anion O 2 •− [65]. The monovalent reduction of oxygen provides the superoxide anion, which is considered both a radical and an anion with the radical sign (•) and a charge of −1 (Equation (1)) [66].
Biologically, in respiring organisms, O 2 •− can be generated enzymatically in catalyzed reactions such as in the mitochondrial respiratory chain and in phagocytic nicotinamide adenine dinucleotide phosphate oxidase (NADPH oxidase) [67].
In the second single reduction, the superoxide anion dismutase spontaneously or through enzyme-catalyzed reactions-for example, the superoxide dismutase enzymes (SODs) act to produce hydrogen peroxide (H 2 O 2 ) and O 2 (Equation (2)) [68].
Even if the superoxide anion is not considered a strongly oxidizing agent, it is able to univalently oxidize the solvent-exposed [4Fe-4S] cluster of the enzyme dihydroxy acid dehydratase during substrate dehydration. This causes cluster instability, degradation, and loss of the catalytic iron atom, leading to the inactivation of the enzyme [69]. In continuation, the formatted ferrous atom (Equation (3)) is able to reduce the hydrogen peroxide (Equation (2)), based on the Fenton reaction, to form the highly reactive and toxic hydroxyl radical (OH • ) (Equation (4)). Initially, the Haber-Weiss reaction was considered to be a source of hydroxyl radicals but later this hypothesis was disproved [70]. Although the superoxide anion can reduce aqueous Fe 3+ , even if ferric ions are characterized by poor solubility, to provide Fe 2+ (Equation (5)), which can enter the Fenton reaction (Equation (4)), the rate of the reduction of Fe 3+ by the superoxide anion is very slow and for this reason, other cellular reductants are considered for this reaction [71].

Individual Phytocompounds Used at Sub-Lethal Doses
The following is a list of antibiofilm compounds used at sub-lethal doses. The chemical structure of these compounds is reported in Figures 2-5. The antioxidant activity of compounds is generally evaluated in multiple radical scavenging assay systems such as 2,2 -diphenyl-1-picryl-hydrazyl (DPPH•), 2,2 -azinobis(3-ethylbenzothiazoline-6-sulfonic acid (ABTS •+ ), hydroxyl radical, superoxide anion, hydrogen peroxide, the Oxygen Radical Absorbance Capacity (ORAC), lipid peroxidation, galvinoxyl, and linoleic acid peroxidation. The chelating ability of the compounds for Fe 2+ is measured by the ferrozine assay.     Considering that microbial response to environmental stimuli can be dose dependent is extremely important. While repressing biofilm formation, sub-inhibitory concentrations of carvacrol and thymol (in the essential oil of oregano) led to treated E. coli O157:H7 cells upregulating the genes related to membrane, heat, and oxidative stress responses and iron uptake, thus increasing direct and cross-resistance [72]. At 0.33 mM sub-lethal concentration, both carvacrol and thymol presented antioxidant activity in the linoleic acid emulsion assay, with values of 95.3% and 96.8%, respectively, but almost no inhibition in the DPPH• assay at the same concentration [73].

Cathecol
Cathecol is strictly linked to salicylic acid (SA), as the enzyme salicylate monooxygenase converts salicylate into catechol. The PGPR Bacillus subtilis protects plant roots from pathogenic bacteria, also forming biofilms. The non-colonization and suppression of B. subtilis biofilm formation on the roots of A. thaliana line NahG, a transgenic line-containing gene for salicylate hydroxylase, which hydrolyzes SA and results in the overproduction of catechol, were observed [74]. The authors speculated that B. subtilis colonization and biofilm inhibition was due to the high ROS concentrations generated by catechol augmentation on NahG roots. The antioxidant activity of catechol is associated with the consecutive two-electron oxidation of the compound to form the corresponding quinone [75]. Moreover, catechol is capable of forming complexes with Fe(III), and its stoichiometry depends on the pH at which the reaction is performed. However, if the complexes undergo an internal redox reaction giving rise to iron (II) and quinone, iron (II) will produce hydroxyl radicals through the Fenton reaction [76].

Tannic Acid and Ellagic Acid
A concentration of 20 µg/mL (11.8 µM) of tannic acid was found to strongly inhibit the biofilm formation of S. aureus [77]. Auto-oxidation of catechins leads to their polymerization and forms tannins [51]. Tannic acid presented good radical scavenging potential, very near to sub-lethal concentrations, with 50% of radical scavenging to be achieved on 20.42 and 20.19 µM in the DPPH• and superoxide assay, respectively. Moreover, tannic acid showed 50% ferrous chelation at 11.63 µM, with the highest dose of tannic acid inhibiting complex formation up to 77% [78]. The plant phenolics ellagitannins encompass more than 1000 identified natural bioactive compounds [79]. In the study by Dalvi and colleagues [80], spectral analyses of iron-ellagic acid complexes showed that ellagic acid seizes ferric ions from EDTA within hours, and from citrate within 1 min. In fact, ellagic acid showed a moderate radical scavenging potential but was found to induce a dose-dependent inhibition of the ferrozine-Fe 2+ complex formation up to 70% at the highest dose of 100 µM [81].

Cannabidiol
At sub-lethal doses, cannabidiol showed a prevention and control activity against C. albicans SC5313 biofilm [20]. Cannabidiol modified mitochondrial activity and induced intracellular ROS production [20]. Cannabidiol was able to reverse iron-induced expression of the mitochondrial fission protein DNM1L in rats [82].

Resveratrol and ε-Viniferin
Resveratrol occurs in grapes and grape-derived beverages such as red wine. A statistically significant and concentration-dependent effect was proved for resveratrol with >80% reduction for all Propionibacterium acnes strains studied without affecting planktonic cell growth [53]. The mechanism by which resveratrol exerts its activity was not investigated. Sub-MICs of trans-resveratrol and ε-viniferin, a resveratrol dimer, inhibited the biofilm formation of P. aeruginosa PAO1 and PA14 [83]. In particular, trans-resveratrol at 50 µg/mL (219 µM) decreased P. aeruginosa PAO1 biofilm formation by 92%, and ε-viniferin at 50 µg/mL (110 µM) decreased P. aeruginosa PA14 biofilm formation by 82%. The ability of trans-resveratrol and ε-viniferin to scavenge the hydroxyl radical formation was determined in the 2-deoxyribose degradation and in the rat liver microsomal lipid peroxidation assays. At sub-lethal concentrations, trans-resveratrol, trans-ε-viniferin and cis-ε-viniferin were able to inhibit 50% of the produced radicals at 7.35, 0.17 and 0.43 µM, respectively, in the 2-deoxyribose degradation assay and 2.58, 0.41 and 1.08 µM, respectively, in the lipid peroxidation assay [84]. Additionally, 50% of scavenging activity on the superoxide radicals of trans-ε-viniferin was obtained at 140 µM [85]. Trans-resveratrol, which is by far the most potent chelator of copper, does not chelate iron [86].

Eugenol
In addition to inhibiting biofilm formation, eugenol at sub-MIC concentrations inhibited the production of virulence factors, including pyocyanin and pyoverdine, in P. aeruginosa PAO1 [55,87]. Eugenol inhibited the generation of the superoxide anion by 50% at a sub-lethal concentration of 250 µM and the generation of hydroxyl radicals to an extent of 70% at 200 µM [88]. In the DPPH• assay, 50% of the radical scavenging activity of eugenol was obtained at 510 µM [89].

Vanillin
Reduced production of ergosterol in presence of sub-MIC vanillin indicated that inhibition of ergosterol biosynthesis may be a probable target in C. albicans [34] and modulation of ergosterol content had been already proven critical in adaptation to oxidative stress [90].
Vanillin showed no activity in the DPPH• radical, ORAC and galvinoxyl radical scavenging assays but showed stronger activity than ascorbic acid and Trolox in the ABTS •+ scavenging assay with a sub-lethal concentration at 19.4 µM for the effective scavenging of 50% of the formatting radicals [91]. On the other hand, vanillin did not display any Fe 2+ chelating ability [92].

Quercetin and Fisetin
In the study carried out by Lee et al. [77], only 1 µg/mL of quercetin inhibited biofilm formation by methicillin-resistant S. aureus by >80% and methicillin-sensitive S. aureus strains ATCC 6538 and ATCC 25923 by >50%. In a later investigation, upon treatment with sub-MIC quercetin, overexpression was proved for the proteins DnaK, EF0080 and OsmC, showing that E. faecalis cells were under oxidative and general stress [39]. In addition, the NADH peroxidase (npr), an oxidoreductase enzyme playing a critical role in maintaining cellular redox homeostasis, was overexpressed as well as chaperone protein DnaK and chaperonin GroS, both proteins involved in stress management [39]. Quercetin can completely suppress Fenton chemistry both at sub-lethal micromolar levels (10 µM concentration) and in the presence of major cellular iron chelators such as ATP or citrate in hydroxyl radical formation based on the 2-deoxyribose degradation assay. However, the radical scavenging activity of quercetin provides only partial protection against Fenton chemistry-mediated damage while Fe chelation by quercetin can completely inhibit Fenton chemistry, indicating that the chelation may be key to its antioxidant activity [95]. Fisetin (5-deoxyquercetin) is a flavonoid commonly found in several fruits and legumes. At a very low concentration (16 µg/mL, 55.9 µM), fisetin inhibited the biofilm development of S. aureus 8325 by 90% [18]. Fisetin increased its radical scavenging percentages in a dosedependent manner in various antioxidant assays at sub-lethal concentrations. The values for the 50% radical inhibition in hydroxyl, superoxide, DPPH• and ABTS •+ scavenging assays were 47.41 ± 4.50, 34.05 ± 0.87, 9.69 ± 0.53 and 2.43 ± 0.14 µM, respectively [96].

(−)-Epigallocatechin-3-gallate
(−)-Epigallocatechin-3-gallate (EGCG), which is the major polyphenolic component of tea, decreased Vibrio mimicus autoaggregation and swimming motility, favored membrane permeability and ROS production led to cell membrane damage and caused potassium leakage [49]. Approximately at tea cup amount (200 µg/mL, 436 µM), EGCG eliminates Escherichia coli K12 biofilm matrix by interfering with the assembly of curli subunits into amyloid fibers and by provoking the σ E cell envelope stress response [50]. Interestingly, this phytocompound can sometimes promote biofilm formation and tolerance to specific antibiotics [97]. The concentration of EGCG for 50% radical scavenging was decreased with decreasing pH value in the DPPH• assay. At pH 10 the scavenging activity was 0.7 µM, at pH 7 of 0.6 µM, and at pH 4 of 1.4 µM, albeit in sub-lethal doses [98]. Additionally, EGCG presented a 50% radical scavenging at concentrations of 1.8 and 1.0 µM against the superoxide anion and hydroxyl radical, respectively [99]. Moreover, EGCG might have iron-chelating activity since it was able to inhibit paraquat Pq-induced MDA production, at 40 µM concentration, an inhibition that disappeared when excess amounts of FeSO 4 were added to the reaction mixture [100].

Cyanidin
The anthocyanin cyanidin was proved to affect K. pneumoniae strain PUFST23 biofilm development at 50-150 µg/mL (0.17-0.52 mM) [37]. Quercetin and other polyphenols form co-pigments via intermolecular interaction with anthocyanins [101]. Cyanidin and its 3-glucoside are able to chelate iron and reduce both iron and copper [102]. Qian and colleagues [101] evaluated the antioxidant effects of anthocyanin complexes of cyanidin-3diglucoside-5-glucoside (CY3D5G), rutin and Mg(II)/Fe(III). The antioxidant properties were affected by the molecular combination with CY3D5G-rutin-Fe(III) showing much lower activities than CY3D5G-rutin-Mg(II). Cyanidin and its 3-glucoside showed a dosedependent antioxidant activity in the DPPH• assay with the 3-glucoside being more potent at sub-lethal concentrations. At 30 µM, the cyanidin-3-glucoside demonstrated the same activity as 30 µM of the positive control Trolox. Additionally, the two compounds exhibited the same dose-dependent activity in the inhibition of the superoxide anion formation with, again, the 3-glucoside being more potent. The cyanidin-3-glucoside at 1 µM concentration displayed a similar inhibition as the superoxide dismutase at 80 mU/mL [103]. Cyanidin presented 50% chelating activity on Fe 2+ at 60 µM concentration and 60% hydroxyl radical scavenging at 25 µM concentration [104].

Hesperetin, Hesperidin, Myricetin and Myricitrin
Hesperetin and hesperidin showed a significant superoxide radical scavenging effect in a dose-dependent manner at sub-lethal concentrations (40-100 µM), although, at the same range of concentrations, in the DPPH• assay, hesperidin had a slight observed free radical scavenging effect and hesperetin a moderate one. Hesperetin exhibited a weak Fe 2+ chelating activity at the same concentrations [105]. Myricetin showed a slight radical scavenging effect of 6.2% in the ABTS •+ assay at 98 µM [106] but was able to strongly chelate Fe 2+ at 25 µM [107]. On the other hand, myricitrin exhibited a strong scavenging activity in the DPPH• and hydrogen peroxide assays with 50% of radical inhibition at 3.0 and 65.2 µM, respectively [108]. (Figure 4)

Salicylic Acid
Mishra and Baek [109] wrote a comprehensive review on salicylic acid (SA). SA is produced by bacteria, fungi and plants. In plants, it is a hormone-regulating plant growth, environmental stress and defense responses against pathogens. At low iron availability in SA-producing bacteria, SA is critical for the biosynthesis of salicyl-derived siderophores or catecholates, e.g., anachelin, pyochelin, bacillibactin, petrobactin, enterobactin, photobactin, amychelin, salmochelin, vibriobactin, vulnibactin and mycobactin. The antioxidant activity of SA is based on the redox deactivation of iron through chelation and not by hydroxyl radical scavenging. The voltametric results indicated that the iron-salicylate complex does not have the thermodynamic driving force to act as an effective Fenton reagent necessary for the production of damaging oxygen-containing radicals [110]. In the DPPH• assay, salicylic acid did not exhibit any antiradical activity even at concentrations of 800 µmol/assay [111].

Zosteric Acid
After proving its ability to contrast biofilm formation at sub-lethal doses [23], zosteric acid or p-(sulfoxy)cinnamic acid, made by the seagrass Zostera marina, was shown to interact with the E. coli protein NADH:quinone reductase, WrbA [112]. Investigation of the antibiofilm activity revealed that the para-sulfoxy ester group is not responsible, whereas the cinnamic acid scaffold carries the antibiofilm performance. A study by Kurth et al. [113] demonstrated that the biofilm-inhibiting effects of zosteric acid on Vibrio natriegens can be entirely attributed to coumaric acid, which is released from zosteric acid by sulfatase activity. The scavenging ability of coumaric acid in the DPPH•, hydrogen peroxide, superoxide and ABTS •+ radical assay, as well as the ferrous chelating ability [114] at concentrations similar to the antibiofilm sub-lethal concentrations of zosteric acid, suggest that zosteric acid could also have an efficient antioxidant activity at sub-MIC.

Ascorbic Acid
The antibiofilm effect of vitamin C on carbapenem-resistant hypervirulent K. pneumoniae at sub-MIC was associated with the induction of ROS production, by Xu et al. [38]. Ascorbic acid functions primarily as a donor of single hydrogen atoms, and the radical anion monodehydroascorbate react with radicals. Its crucial role as a reducing agent involves enzymatic reactions, such as for dopamine β-monooxygenase and peptidyl-glycine α-amidating monooxygenase. Additionally, it reduces Fe 3+ to Fe 2+ in enzymes that contain Fe 2+ as a cofactor such as procollagen-proline dioxygenase and procollagen-lysine dioxygenase. Moreover, ascorbate oxidase and ascorbate peroxidase form H 2 O by reducing O 2 and H 2 O 2 , respectively, by using ascorbate as a single-equivalent donor. The antioxidant and free-radical scavenging activity of ascorbate are associated with its non-enzymatic reduction of superoxide, hydroxyl, alkoxyl, peroxyl, and other radicals [115]. Ascorbic acid is a known potent antioxidant and is used as a positive control in the DPPH• radical scavenging assay and in the ABTS •+ radical scavenging assay since the percentage of the antioxidant activity of ascorbic acid at the sub-lethal concentration of 100 µM in both assays is near 100% [116]. In the lipid peroxidation assay, the value of the potency of ascorbic acid to inhibit 50% of the produced radicals was 14 µM [117]. In contrast to its antioxidant use, ascorbic acid under certain conditions can also act as a pro-oxidant and a source of free radicals. The alpha oxo-hydroxy and di-hydroxy ligands in the chemical structure of ascorbic acid create complexes with ferric ions but generally, ascorbic acid is considered a weak chelating agent and cannot form strong iron complexes. Thus, the reduction of ferric ions from ascorbic acid to ferrous irons and ascorbic radical will turn on the cyclic oxidation process [118].

Caffeic Acid
Cattò and colleagues [112] showed that 18.3 µM of caffeic acid reduced E. coli adhesion to a hydrophobic surface without killing the cells. The antibiofilm performance of caffeic acid was expressed at a concentration 10-fold lower than that of zosteric acid. The free radical scavenging activity of caffeic acid was near 60% at the sub-lethal concentration in the DPPH• assay [119].

Piperine
Piperine, a bioactive component of pepper, functions as a potent antibiofilm agent to inhibit S. aureus biofilm formation by accumulating ROS, but showed no toxicity against WI 38 line at the tested concentrations [47]. Piperine derivatives were investigated as green corrosion inhibitors on iron surfaces, leading to Fe-inhibitor complex in the hetero-atom centers [120].

Berberine and Roemerine
The maximum scavenging concentration of berberine was found at a sub-lethal concentration of 0.76 mM whilst the minimum scavenging activity was found at 5.9 µM in the DPPH• assay, with 50% of radical scavenging at 0.12 mM. In the ABTS •+ assay, the free radicals were scavenged in a concentration-dependent manner with the maximum scavenging activity at 1.5 mM and the minimum at 5.9 µM with 50% of radical inhibition at 0.11 mM [116]. The superoxide assay showed 76.3% radical inhibition at 1.5 mM. Roemerine was able to inhibit only 1.8% of the radical generation at 100 µM concentration in the ABTS •+ assay, while was not active in the DPPH• assay at the same concentration [116]. (Figure 5)

Coumarins
More than 1300 coumarins have been identified in plants [36]. Non-grass plants and grass plants possess different iron responses, named Strategy I and Strategy II, respectively, activated in roots under Fe deficiency [121]. The chelation-based mechanism for Strategy II involves the exudation of phytosiderophores that bind ferric Fe in the soil [122]. Then, the complex is taken up by transporters in the root epidermis [121]. The strategy I is a reduction-based mechanism and involves acidification of the rhizosphere via proton secretion, and enzymatic reduction of iron chelates at the root surface [122]. In the reduction strategy, large amounts of coumarins and protons are released in the rhizosphere to reduce Fe 3+ to Fe 2+ before transport across the plasmalemma of root epidermal cells [123]. Thus, coumarins (scopoletin, fraxetin and sideretin), improve plant health by engaging microorganisms involved in iron nutrition [124]. Conversely, these specialized metabolites contrast pathogens. 6-methylcoumarin showed significant antibiofilm activity against P. aeruginosa PAO1 biofilm at 125 µg/mL [46]. Several virulence factors were inhibited including pyocyanin (which promotes glutathione oxidation that results in ROS formation) and pyoveridine and pyochelin siderophores. Finally, oxidized coumarins are often more reactive than their counterparts, as oxidative hydroxylation and dimerization increase phenolic groups and bring new chelating features [125]. At the sub-lethal concentration of 100 µM, scopoletin and fraxetin were able to inhibit the superoxide anion generation at 100% and 67%, respectively. Moreover, at the same concentration, scopoletin showed a small but significant reduction in the scavenging of hydroxyl radicals [126].

Shikonin
Sub-lethal concentrations of shikonin, a naphthoquinone retrieved from the roots of Lithospermum erythrorhizon, effectively decreased L. monocytogenes biofilm biomass on polystyrene and adherence to glass slides. Further, the transcription of biofilm-associated genes and virulence-associated genes was downregulated/repressed [15]. At sub-lethal concentrations, shikonin was able to inhibit 50% of the radical formation in the superoxide anion, hydroxyl, and tert-butylperoxyl radical assays, with values of 7.2, 40 and 27 µM, respectively [128].

trans-Cinnamaldehyde
Despite the ability of trans-cinnamaldehyde (a major component of bark extract of cinnamon) to suppress biofilm formation, sub-inhibitory concentrations led to E. coli O157:H7 treated cells to upregulate the genes related to membrane, heat, and oxidative stress responses and iron uptake hence increasing direct and cross-resistance [72].

Curcumin
The QS-regulated siderophore production in Aeromonas sobria was inhibited by 420 µg/mL (1.14 mM) curcumin liposomes (formulated to increase bioavailability of the poor water solubility of curcumin) but not by 280 µg/mL (0.76 mM) free curcumin [54]. Curcumin exposure affected iron homeostasis and oxidative stress response of P. aeruginosa PAO1 [129]. In the presence of curcumin, L-ornithine N (5)-monooxygenase, pyoverdine biosynthesis protein and Fe 3+ pyochelin receptor protein were upregulated, whereas ferroxidase, bacterioferritin, and isochorismate pyruvate lyase were downregulated. These results suggested the insufficient iron acquisition by P. aeruginosa PAO1 upon treatment with curcumin. Moreover, curcumin downregulated the expression of the antioxidant enzymes catalase, peroxidase, superoxide dismutase (SOD) and alkyl hydroperoxide reductase, increasing the ROS level. The 50% of the scavenging effect of curcumin on the DPPH• radical assay was achieved at the sub-lethal concentration of 94.6 µM, and curcumin displayed 50% effective radical cation scavenging activity in the ABTS •+ assay at 49 µM. The inhibition of superoxide radical generation was 42.7% at 40.7 µM and the scavenging of hydrogen peroxide was 28.4% at the same concentration. Moreover, curcumin exhibited 56.7% ferrous ion chelation at 40.7 µM concentration [130].

Conclusions
Antimicrobial resistance impacts the achievements of several of the 17 United Nations Sustainable Development Goals (SDGs), in particular SDG 3 'Good health and well-being' (https://sdgs.un.org/goals (accessed on 9 December 2022)). To address the antimicrobial resistance problem, the effects of natural compounds on biofilm formation at sub-lethal concentrations have been extensively investigated. This overview is significant as it gathers together the literature on individual compounds, rather than on extracts, of which the use is difficult to reproduce. Thus, to the best of our knowledge, this is the first review that targets only individual phytochemicals below inhibition concentrations against biofilm formation.
In vitro experiments cited here reported that the phytochemicals are efficient ROS scavengers at sub-MIC antibiofilm concentrations. Due to their features and quantity (production and disposal) inside cells, ROS are recognized as universal signaling chemical species [131]. ROS, either associated with abiotic stresses or produced by host plants, determine the outcome of the plant-bacteria interaction [132], e.g., the optimal formation of legume-rhizobial interaction, especially in the development of symbiosis [131]. It is speculated that a high level of polyphenol oxidase activity in N 2 -fixing nodules relates to the similarity of the response of plants to pathogens since the protective responses of legume plants to the invasion of rhizobia are analogous to the pathogenic process [133]. Ong et al. [134] stated that compounds that could target oxidative stress regulators, such as antioxidants, could potentially be exploited as a novel strategy for biofilm control. General mechanisms of how ROS inhibition influences biofilm growth are still unknown, but mechanisms of how sessile cells cope with the increase in oxidative stress have been reported. For instance, Chua et al. [135] demonstrated that sub-lethal hydrogen peroxide doses are essential for the selection of pro-biofilm-forming pathogenic variants by modulating cyclic-di-GMP levels in P. aeruginosa.
Interestingly, in all experiments reported here in which oxidative stress was investigated, a modest increase in intracellular reactive oxygen species was reported in treated cells in comparison with untreated specimens [20,38,47,49]. The above-mentioned increment of ROS is in line with other outcomes obtained with polyphenols [136,137] and could be related to bacterial wall damage. At lethal concentrations, polyphenols cause bacterial wall damage and a subsequent increase in free radicals; this pro-oxidant effect is accompanied by lipid peroxidation and DNA lesions [136]. In order to respond to this damage, the cell starts a self-protective mechanism that could be further enhanced by the antioxidant and radical scavenger activities of the polyphenols themselves [136]. In this respect, one of the major misinterpretations in the field of oxidative stress concerns the scavenging of superoxide (O 2 •− ) or H 2 O 2 by small molecules. In fact, the antioxidant enzymes react thousands to millions of times faster with those oxidants than the small molecules, and for this reason, they are the predominant cellular antioxidant mechanism [138]. Polyphenolic antioxidants might act as pro-oxidants by disturbing the healthy redox cycle and causing an accumulation of reactive oxygen species (ROS) (i.e., hydrogen peroxide, superoxide and hydroxyl radicals) [134], whereas a healthy redox cycle promotes microbial attachment, thus favoring biofilm formation [139]. High levels of ROS derived from the presence of polyphenols acting as pro-oxidant compounds were seen to cause the following effects: ascorbic acid was able to suppress the biofilm exopolysaccharide and to inhibit the efflux pump in the KP1088 and HvKP3 K. pneumoniae strain by, at the same time, increasing the accumulation of ROS [38]; piperine raised the level of ROS and decreased the bacterial motility of the S. aureus (MTCC 96) [47]; an inhibition in the bacterial motility was observed by epigallocatechin gallate in Vibrio mimicus accompanied by increased levels of ROS [49]; an accumulation of ROS was detected in the C. albicans SC5313 strain when cannabidiol was used, which was followed by downregulation of genes involved in the biofilm maintenance and development, and in the maturation of factors associated with the EPS synthesis [20].

Future Perspective
As a future perspective, phytochemicals can not only be used to control biofilm formation, but also to study how cells react when a modest and controllable rise in oxidative stress is experienced. Furthermore, while there are many recent manuscripts on biofilm prevention and removal, we foresee that this review will give also clues on promoting biofilm growth for biotechnological applications, a research topic of growing interest [140]. Nevertheless, a considerable number of issues remain to be clarified or investigated further [141]. In this respect, the incomparability of studies due to different model microorganisms, experimental conditions, and laboratory methods still limit the understanding of the overall mechanism of action. The pro-or antioxidant effects of some molecules in living systems are difficult to determine, as well as the cascade of events triggered by oxidative stress. For instance, the scientific literature reported different sub-MIC concentrations of phytochemicals for each target microorganism. This finding makes challenging to determine the concentration that causes the perturbation of redox homeostasis and, consequently, the antibiofilm activity of the natural compound. Furthermore, the perturbation of redox homeostasis is not exclusive to a specific cellular compartment and it may occur at the same time in different and multiple locations. Finally, the lack of in vivo assays with individual phytochemicals and the phytochemicals in situ persistence must be thoroughly addressed before the more extensive use of plant-based compounds at sub-MIC doses to contrast biofilm.
Many questions remain to be answered: What is the role of oxidative stress in the antibiofilm capacity of phytochemicals? What are the mechanisms through which the perturbation of redox homeostasis can contribute to the antibiofilm performance of phytochemicals? What are the biochemical/molecular targets of ROS? Which biological compartments are affected by oxidative stress? What are the external factors (e.g., UV radiation, heavy metal, growth conditions) that influence the ROS mode of action?
Given the rapid rate at which our knowledge in this area has increased in recent years, it is likely that answers to many of these questions will be forthcoming in the decades. These answers will undoubtedly help to determine whether ROS are merely consequences of cellular response to treatment or, instead, are central regulators of the phytochemicals' antibiofilm capacity.  Data Availability Statement: United Nations Sustainable Development Goals (SDGs), in particular SDG 3 'Good health and well-being' (https://sdgs.un.org/goals (accessed on 9 December 2022)).

Conflicts of Interest:
The authors declare no conflict of interest.