Conventional antimicrobial agents, such as quaternary ammonium compounds, have been widely applied in industrial disinfection, agriculture, and the food industry (Meireles et al., 2017). Nevertheless, due to the growing safety and environmental concerns associated with these biocides, there is a rising need to explore and develop safer natural and bio-based compounds as alternative antimicrobial agents (Gyawali and Ibrahim, 2014). However, even natural and bio-based antimicrobials can pose risks to ecosystems and human health, considering that some natural compounds are reported to have safety concerns. For example, natural phenolics from plant extracts showed genotoxicity in the Comet test (Demma et al., 2009). These compounds could also exhibit estrogenic activity (EA) by interfering with the estrogen signaling pathway, as most phenolics share a similar structure motif with 17β-Estradiol (E2) (Resende et al., 2013). Resende et al. reported the EA of flavonoids, and our group also discovered that carvacrol exerted weak EA at 10^–12 M (Zhang et al., 2021). Developmental toxicity is another important toxicity endpoint. Flavonoids elicited different levels of adverse effects on the development of zebrafish and chicken embryos (Bugel et al., 2016; Zhang and Wu, 2022). Similar to the phenolic compounds, previous reports have raised concerns about toxicity of lipids, including pro-oxidant activity, mitochondrial toxicity, and aquatic toxicity (Galati and O’brien, 2004; Singh and Chandra, 2019).

Despite all these safety concerns, natural plant monoterpene phenols have been proven to possess antibacterial, antifungal, antioxidant, and anti-inflammatory properties (Khan et al., 2015). For example, the monoterpene phenols carvacrol and thymol are major components of many essential oils and have excellent antimicrobial activity and medicinal properties (Memar et al., 2017). Beechwood creosote consists mainly of natural phenols such as guaiacol and creosol and has been used as preservatives or antiseptics in the pharmaceutical industry (Agency for Toxic Substances and Disease Registry ATSDR, 2024). Natural lipids have also demonstrated antibacterial, antifungal, and anti-inflammatory activities that depend on the chain length and degree of saturation of the fatty acids (FA) (Pohl et al., 2011). For instance, the short-chain lauric acid lipid had excellent antimicrobial activity against viruses and bacteria, while long-chain oleic acid lipid was reported to prevent inflammatory-related and cardiovascular diseases (Matsue et al., 2019; Singh et al., 2020). Some natural antimicrobials are created with balanced benefits and drawbacks, such as ethyl lauroyl arginate (LAE), which is derived from these natural components (lauric acid, L-arginine, and ethanol) and was approved as Generally Recognized as Safe (GRAS) for use as a food preservative (Nübling et al., 2017; Becerril et al., 2013).

To develop efficient and sustainable antimicrobial agents, our group has recently synthesized a family of bio-based phenolic-branched chain fatty acids (phenolic-BCFAs) products derived from two renewable resources (natural phenolics from plants and fatty acids from vegetable oils). Unlike natural phenolic compounds (e.g., carvacrol, thymol), phenolic-BCFAs do not have strong odors because they are covalently ‘carbon-carbon’ bonded between carbon on the phenolic ring and the carbon double bond of the FA (Ngo et al., 2017; Ngo et al., 2014). Thus, they are highly stable and non-volatile, which is in stark contrast to their parent counterparts. They are also a mixture of different isomers where phenolic groups are attached to various positions of the FA chain (represented as dashed lines in Figure 1). Because of this isomeric characteristic, these products are liquid at room temperature, which is ideal for applications requiring this property (Ngo et al., 2017; Ngo et al., 2014). Additionally, the final phenolic BCFA is a saturated FA product because they were prepared through the arylation process. Although some double bonds (unsaturation) can still be present on the FA chain if the process uses soy FA, the amount of unsaturation is still far less than their parent soy FA (Figure 1). In our early study, these phenolic-BCFAs had excellent properties for inactivating Gram-positive bacteria (e.g., Listeria innocua), exhibiting much stronger potency than their individual feedstocks (Ngo et al., 2019). In our more recent study, phenol- and creosote- BCFAs showed great promise in synthesizing antibacterial bio-epoxy polymers. The synthesized bio-epoxy polymers, phenol-soy BCFA/ethylene diamine and creosote-soy BCFA/ethylene diamine, showed good antibacterial activity against both Gram-positive and Gram-negative bacteria (Huang et al., 2021).

For these phenolic-soy BCFA products to be widely adapted, there is a critical need to examine their toxicologic profiles. As our earlier research showed that these phenolic-BCFAs were much more potent antibacterial agents than their parent phenolics (Ngo et al., 2019), it is thus important to determine how the toxicological properties of phenolic-BCFA may differ from the properties of the compounds from which they were derived. In this study, several in silico simulation tools were conducted as the first step for toxicity prediction because of their cost-effectiveness and screening efficiency compared to experimental studies. Although our phenolic-soy BCFAs are mixtures of different isomers (saturated and unsaturated FA chains), they primarily consist of single-phenolic-soy BCFA isomeric products in the saturated form (Figure 1; Table 1). These single-phenolic-branched FA structures were subjected to the in silico simulations. A chicken embryonic assay was used for evaluation of developmental toxicity, as this assay is a promising alternative to traditional in vivo models with increasing applications in pharmacological and toxicological studies (Ghimire et al., 2022; Wu et al., 2018). In contrast to rodent animals (e.g., mice and rats), chicken embryos are isolated biological systems of the early life stage of animals. Unlike rodents, chicken embryos are abundant, cost-effective, and easily manipulated (Ghimire et al., 2022). We investigated the developmental toxicity during chicken embryogenesis for phenol, creosote, iso-oleic acid, and four phenolic-soy BCFAs. Iso-oleic acid is a mixture of methyl-branched-chain C₁₈-fatty acid isomers and is a well-established lubricant (Ngo et al., 2013) (Figure 1). Furthermore, the potential EA of creosote-soy BCFA was assessed by an MCF-7 cell proliferation assay. To our knowledge, this is the first time that the acute toxicity and developmental toxicity of phenolic-BCFAs are investigated by in silico, in vitro, and in vivo methods.

Recently, in silico tools based on various QSAR models have been widely used in toxicity prediction as alternatives to traditional in vivo methods (RK et al., 2019; Zhou et al., 2020). In this study, we applied three in silico tools - T.E.S.T, ProTox-II, and KATE - to predict potential toxicity of three phenolic-soy BCFAs (saturated and unsaturated). Creosote-soy BCFA was not evaluated because the starting material, creosote, is a complex mixture from beechwood tar with at least 36 components, primarily guaiacol, cresols, methoxymethylphenol, and phenol (Figure 1) (Lee et al., 2005). The three tools can only estimate the toxicities of pure compounds; therefore, only the phenol-, thymol-, and carvacrol-soy BCFA were selected. For each of them, three types of pure structures were evaluated: single-phenolic-branched saturated FA (stearic acid), single-phenolic-branched unsaturated FA, and single-phenolic branches attached at different positions of the FA chain.

Our in silico findings suggest a discrepancy in toxicity between three aquatic organism models and rats. T.E.S.T. data show that phenolic-BCFAs exhibited much higher aquatic acute toxicity compared to rats, suggesting different toxicity levels and mechanisms between aquatic organisms and rats (mammals) to warrant further study. The extremely lower T. pyriformis 50% Inhibition Growth Concentration (IGC₅₀) values of three phenolic-BCFAs were related to the inhibitory effect of FAs. FAs have been regarded as a group of pollutants in the aquatic ecosystem and unsaturated fatty acids had more aquatic toxicity concerns (Singh and Chandra, 2019; Wu et al., 2006). Long chain FAs also showed toxicity on microorganisms due to the disruption of the normal function of the cell membrane (Salam et al., 2012). Compared to phenol-BCFA, the thymol- and carvacrol-BCFAs showed higher toxicity on Fathead minnow. Of note, all test compounds were predicted as developmental toxicants from T.E.S.T. which could interfere with normal development of humans or animals. Considering the complex nature of developmental toxicity, further experimental test using a chicken embryonic model was performed in this study.

Chicken embryo is widely used as an alternative animal model within the 3Rs principles for developmental toxicity studies (Ghimire et al., 2022). Since DMSO was used as the solvent for phenolic soy-BCFAs, a limited amount of DMSO (0.1%) without embryotoxicity was administered via egg injection (Luptakova et al., 2021; Hoyberghs et al., 2021). Our previous study showed that carvacrol exhibited higher developmental toxicity in chicken embryos than that of thymol (Zhang et al., 2021). A 45.5% mortality rate was detected for carvacrol at 50 μg/kg dosage, while no death was found at the same treatment concentration for the thymol group. Interesting, carvacrol-soy BCFA in our current study resulted in a lower mortality rate than that of thymol-soy BCFA at the lower dosage (18 μg/kg, Table 4), while at the higher dosage treatment (180 μg/kg) carvacrol-soy BCFA and thymol-soy BCFA showed the same mortality rate. Additionally, our results show that creosote-soy BCFA had the lowest mortality rate among four synthesized phenolic-soy BCFAs, only with a 25% at the lower exposure dosage, although creosote had the most severe deleterious effects at two test dosages (4.1 and 41 μg/kg) on chicken embryo growth. The findings demonstrated that covalently bonding creosote to FA drastically reduced its developmental toxicity. On the other hand, phenol-soy BCFA showed higher adverse effects on chicken embryo growth compared with phenol. Clearly, the choice of phenolic compounds could have significant impacts on the developmental toxicities of the final synthesized phenolic-soy BCFAs and no clear correlation is observed between the toxicities of phenolics and phenolic-soy BCFAs. In other words, developmental toxic phenolics could be used as feedstock to produce phenolic-soy BCFAs with lower or minimal toxicities.

TBARS reflected the lipid peroxidation level in the chicken embryo model and was used as a marker for oxidative damage. The TBARS level in the liver sample varied among different treatments. Carvacrol-soy BCFA at two test dosages significantly increased TBARS values compared with the VC group (p < 0.05). Our previous findings also revealed that carvacrol treatment at 50 μg/kg significantly increased TBARS levels in fetal chicken liver, but thymol did not (Zhang et al., 2021). In another study, carvacrol induced reactive oxygen species generation in human fibroblast (WS-1) and gastric adenocarcinoma cells in a dose-dependent manner at 0–100 µM (Günes-Bayir et al., 2018). These findings suggest that the carvacrol branch may play a significant role in the carvacrol-soy BCFA-induced oxidative stress in the liver. Still, there was no clear relationship between the TBARS results and mortality or malformation rates in our study for other three phenolic-soy BCFAs. More insights concerning actual mechanisms remain to be investigated using other methods which might evaluate different pathways other than lipid peroxidation.

Our in silico findings were obtained on the purified compounds, while the measured responses from the chicken embryonic assay were recorded from a mixture of two compounds with the composition and purity shown in Table 1, which contributed to some discrepancies between the two methods. Furthermore, when compared with experimental approaches using an animal model, computational tests lacked dose response, generally used simplified models and overlooked the complexity of the life development process. The chicken embryo model can thus shed insights on real toxicity impacts in the early life stage, a vulnerable developmental stage with more toxicity concerns. The purity of three phenolic-soy BCFAs (thymol-soy BCFA, phenol-soy BCFA, creosote-soy BCFA) is more than 90% while that of carvacrol-soy BCFA is over 80% (Table 1). No further purification was performed due to the increased cost. In addition, high antimicrobial activities have already been achieved with these bio-based antimicrobials. Considering their cost-effectiveness and future commercial production, we tested them at current compositions to explore their toxicity and application feasibility.

One of the promising applications of phenolic-soy BCFAs is their use as bio-based antimicrobials in the food industry to control foodborne pathogen contamination with minimal environmental impacts (Ngo et al., 2017; Kazem-Rostami et al., 2023). Although the synthesized phenolic-soy BCFAs demonstrated strong antimicrobial properties against Gram-positive bacteria (e.g., L. innocua, B. subtilis, and E. faecium) (Fan et al., 2017), their application in the food industry and other fields will be dependent on their potential toxicity. Our results show that the phenolic-soy BCFAs had varying degrees of adverse effects during chicken embryogenesis with different mortality rates. We selected the injection concentrations of phenolic-soy BCFAs at 4.35–54 μg/mL for the chicken embryonic assays. These concentrations were converted to final dosages of 14.5–180 μg/kg (as indicated in Table 4, considering a 0.2 mL injection volume and an average egg weight of 60 g). It is noteworthy that these concentrations were somewhat higher than their minimum inhibitory concentrations (MIC) against L. innocua (3.6–7.3 μg/mL) (Ngo et al., 2019), since the embryonic environment might reduce the effectiveness of the phenolic-BCFAs. Using a concentration near the MIC is consistent with the use of Thymus extracts or essential oils as a natural antimicrobial preservative in food products at their lowest MIC (Harpaz et al., 2003; Nieto, 2020). Currently, other preservatives such as sodium benzoate may be applied in foods at levels up to 0.1%, or 1,000 μg/mL. Because phenolic-BCFA compounds are effective against Gram-positive bacteria at levels less than 10 μg/mL and have limited aqueous solubility, we anticipate that they would not need to be added at a concentration as high as 1,000 μg/mL.

Furthermore, the potential EA of creosote, creosote-soy BCFA, and iso-oleic acid was investigated via an MCF-7 cell proliferation assay. We chose creosote-soy BCFA as the presentative compound for EA assessment because it demonstrated lower toxicity concerns in the chicken embryonic assay compared with other phenolic BCFAs. The data show that only E2, as the positive control, demonstrated EA at 10 nM and 100 nM. This trend aligns with our previous finding that E2 had the highest EA at 10 nM across the test range from 10^–15 to 10^–4 M (Peng et al., 2018). Creosote-soy BCFA, creosote, and iso-oleic acid did not show EA at the three test concentrations. These findings indicate that creosote-soy BCFA (with a good performance in antimicrobial activity) had no EA concern, in addition to its lowest developmental toxicity among all tested phenolic BCFAs.

This study evaluated the potential developmental toxicities and EA of phenolic soy-BCFAs. Firstly, we predicted the aquatic toxicity, acute toxicity (oral rat LD₅₀), developmental toxicity, and mutagenicity of phenolic soy-BCFAs and phenolic compounds using several in silico tools (T.E.S.T., ProTox-II, and KATE). The phenolic soy-BCFAs (both saturated and unsaturated) had much higher toxicity in aquatic organisms than free, unreacted phenolics. In rats, the phenolics exhibited higher predicted toxicity than the phenolic soy-BCFAs. All the tested compounds were classified as non-mutagenic developmental toxicants. Acute toxicity and developmental toxicity increased slightly when fatty acids contained a double bond. However, the location of the phenolic branch (structural isomers) had no effect on the in silico simulation results. The developmental toxicity of phenolic soy-BCFAs was further investigated by the chicken embryonic assay, which indicated varying levels of developmental toxicity based on the type of phenolic branches. Creosote-soy BCFA had the lowest mortality rate, whereas phenol-soy BCFA showed the highest death rate, malformation number, and altered developmental index (REEW). Thymol and carvacrol soy-BCFAs had comparable mortality rates; however, only carvacrol soy-BCFAs significantly increased the levels of oxidative stress based on TBARS values in the chicken embryonic liver samples. Furthermore, creosote-soy BCFA showed no EA concerns from MCF-7 cell assay. Overall, the phenolic-soy BCFAs, particularly creosote-soy BCFA, presented in this study show promise as potentially safer bio-based antimicrobial products.