Bacterial adaptation to essential oils and their main compounds in food systems: mechanisms related to antibacterial resistance

Abstract

Essential oils (EOs) are complex mixtures of volatile metabolites with strong biological properties. Their multi-target mechanisms confer rapid bactericidal activity, lowering the likelihood of resistance development compared to antibiotics. This review examines bacterial responses to EO-induced damage, including adaptation, tolerance, resistance, cross‑adaptation, and collateral sensitivity, which are often overlooked in existing reviews of EOs’ biological efficacy. Additionally, it links EO-adaptation to food, including application methods, biofilms, and interactions with food components.

Introduction

The rapid emergence of antimicrobial-resistant (AMR) bacterial strains is one of the greatest threats to human and animal health^1,2. Clinical and agricultural environments are facing growing challenges, primarily due to the so-called ESKAPE pathogens (Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter spp.), which cause severe infections and economic losses^3,4. These pathogens, among many others, are known for their alarming resistance profiles, as they routinely acquire multidrug resistance (MDR), extensively drug resistance (XDR), and pan-drug resistance (PDR) phenotypes^5–7. Therefore, there has been intense interest in identifying alternatives to conventional antibiotics across medical, agricultural, and food‑processing sectors.

Many investigations have shed light on the antimicrobial effects of numerous natural alternatives, particularly essential oils (EOs)^8–10. EOs are aromatic, lipophilic, and highly volatile secondary metabolites of plants, mainly composed of terpenes and phenylpropanoids^11–13. Unlike single-target antibiotics, many EOs contain complex mixtures of compounds that act on multiple cellular targets simultaneously, including membrane disruption, inhibition of protein synthesis, and interference with DNA replication^14–16.

To avoid confusion among bacterial resistance, tolerance, and adaptation, it is worth noting that resistance is usually a stable, heritable trait that persists even after stress removal. By contrast, tolerance is a physiological survival state that does not necessarily increase MIC and, foremost, may disappear after stress removal. Finally, adaptation refers to a bacterial adjustment induced by external stress, so it is a broad term that may include tolerance and, at times, resistance development^17,18.

Although EOs are generally considered less likely to induce stable resistance, bacterial resistance cannot be entirely ruled out. Recent evidence suggests that, among adaptive responses, genetic resistance may develop, particularly after repeated exposure¹⁹. In fact, studies have shown that continuous or sublethal exposure may select tolerant variants and/or, in rare cases, lead to stable genetic resistance to individual bioactive components such as carvacrol and cinnamaldehyde^20–22. This response involved activating antioxidant systems and fatty acid biosynthesis pathways in the treated bacteria. Consequently, these mutations significantly enhance membrane repair and reduce reactive oxygen species (ROS) accumulation, thereby increasing tolerance^20–22.

Moreover, given their antimicrobial, anti‑inflammatory, and antioxidant properties, EOs have long been exploited by the food, fragrance, and pharmaceutical industries^23–25. Especially due to consumers’ demand for natural, eco‑friendly preservatives, many manufacturers are replacing synthetic antimicrobials with EOs for food safety applications^26,27.

These observations highlight the need to assess whether adaptation to EO stress is transient or stable and to obtain a clear understanding of whether bacterial responses are damage‑specific or stress-specific. However, the biological reason for the rarity and low persistence of resistance to EOs stress, unlike antibiotic resistance, remains poorly understood.

This review critically examines recent research on bacterial adaptation mechanisms towards EOs. First, summarizing the various adaptation types observed in response to EOs and their individual compounds; notably, homologous adaptation, cross-adaptation, and forward and backward collateral sensitivity (CS). Second, compiling methodologies for detecting and monitoring adaptation in repeated‑dose experiments, from phenotypic assays to omics approaches. Third, discussing factors that influence the reversibility and stability of EO‑induced adaptation. Also, exploring how bacteria perceive and respond to EO‑induced damage, and how these stress responses intersect with the bacterium’s virulence and fitness. Finally, examining the main factors influencing the development and dissemination of EO-reduced sensitivity in food industry settings. These include interactions with food components (such as lipids, proteins, and pH), improper use of encapsulation methods, biofilm formation in food and processing environments, and the impact of repeated sublethal exposures. Gaining a clearer understanding of these factors can facilitate the safe and effective implementation of EO-based treatments within the food industry.

By synthesizing evidence from recent investigations, the present review highlights the rarity of stable, heritable resistance to EOs and emphasizes the distinction between stress‑specific and damage‑specific responses to EO exposure and their relationship to the concepts of cross‑adaptation and CS in the context of EOs. This provides a framework for predicting how exposure to EOs may influence susceptibility to antibiotics and other natural antimicrobials^28–30. Furthermore, factors that modulate adaptation to EOs are being analyzed outside the Petri dish, under real‑world conditions, such as food matrices, biofilms, and encapsulation applications, and the implications for safety, efficacy, and regulatory acceptance of EO use^31–34.

Consequently, this review shifts the focus from biological efficacy to adaptation dynamics, providing a better understanding of the risks of EO misuse and the emergence of EO adaptation, thereby enabling more confident application of EOs in medicine, the food industry, and various biotechnological fields.

Literature search strategy

The literature included in this review was collected from Google Scholar, PubMed, and Scopus, with a focus on studies published from 2020 to 2026 to provide a current overview of advances in EO-induced adaptation and antimicrobial resistance, particularly in the food industry. The search used various combinations of keywords including “essential oil”, “bacterial adaptation”, “tolerance”, “homologous adaptation”, “cross-adaptation”, “CS”, “biofilm”, “VBNC”, and “food industry”. While recent articles were emphasized, earlier research was also reviewed during the conceptual phase, even if not directly cited.

Adaptive mechanisms of bacteria toward EOs pressure

When exposed to chemical or physical stressors, such as antimicrobials, oxidative agents, pH, heat, high hydrostatic pressure, pulsed electric fields, and other stressors, microorganisms tend to either enter a transient tolerance state or develop resistance through various intracellular mechanisms³⁵. More specifically, this adaptive response involves both transient non-genetic physiological modifications and heritable genetic changes. Some of these adaptive changes include: heterogeneity in efflux pump expression, variation in cellular metabolic activities, adjustments of nutrient utilization, and, in some cases, the activation of enzymes that normally play a recessive role, depending on each bacterial strain and the specific stress it encountered^6,36. In stressful environments, bacteria can survive due to three key features: an inherent cellular trait, an immediate physiological response, and, lastly, the emergence of de novo genetic mutations¹.

On one hand, the innate cellular characteristic, referred to as the baseline resistance, includes features such as an impermeable cell envelope, natural active efflux pumps, and membrane composition^1,37. Variations in membrane fluidity, phospholipid composition, and surface charge affect how EO components insert into membranes, often reducing their effectiveness or rendering them completely ineffective³⁸. One example is P. aeruginosa, known for its low-permeability outer membrane and consistently active efflux pumps, which make it naturally resistant to many antimicrobials without prior exposure^39,40. However, these pre-existing traits are not considered adaptive responses since they are not triggered by the stressor; instead, they are innate and already present⁴¹. On the other hand, the immediate physiological and genetic adaptive responses are considered the two major adaptive responses in stressful environments^42,43.

Physiological and phenotypic adaptive responses

When exposed to stressful conditions, bacteria often activate rapid, non-genetic physiological adaptations, also known as immediate cellular responses, in order to improve short-term survival⁴⁴. This mainly includes bacterial membrane remodeling, oxidative stress responses, efflux pump overexpression, and metabolic slowdown⁴⁵.

Indeed, changes in membrane fatty acid profiles are among the most potent phenotypic adaptations in pathogenic microorganisms when faced with environmental stress, especially when natural treatments such as Cinnamaldehyde are used at non-lethal concentrations⁴⁶. For instance, in their study, Wijesundara et al.⁴⁷ stated that exposure of Streptococcus pyogenes to sub-lethal concentrations of carvacrol, a well-known strong bioactive EO compound⁴⁸, has caused remarkable changes in cell membrane phospholipids. Hence, S. pyogenes has increased levels of unsaturated long-chain fatty acids that enhance membrane fluidity, unlike cell membranes composed of saturated fatty acids, which are thicker and less flexible⁴⁹.

Generally, membrane fluidity boosts tolerance to bioactive compounds, especially EO and antibiotics, by enhancing efflux and transport activity^50,51, enabling faster membrane repair and reorganization⁵², altering how antimicrobials bind or insert^53,54, also increasing outer membrane vesicle production and HGT^55,56, along with changing membrane-linked signaling that activates protective pathways^57,58. However, in some cases, membrane fluidization often makes bacterial cells more vulnerable and increases their susceptibility, depending on the bioactive agent and the targeted species^59,60.

Furthermore, sublethal EO exposure induces upregulation of efflux pumps⁶¹. One example is P. aeruginosa, which, when following sublethal exposure to EO, showed an upregulation of the MexAB-OprM and AcrAB-TolC efflux pumps^61,62. These changes can decrease the intracellular concentration of EO constituents and reduce susceptibility not only to the EO itself but also to other stressors⁶³. Thus, maintaining efflux pump activity within microorganisms represents a non-specific, intrinsic bacterial stress tolerance and a metabolic state that is not inherited, rather than a form of acquired resistance⁴¹.

Moreover, EO constituents often trigger oxidative stress by generating ROS, which can damage DNA, proteins, and membranes^64,65. To overcome this EO-induced oxidative stress, bacteria activate defense mechanisms by increasing antioxidants like catalase and activating regulators like soxS^66–68. These regulators enhance the expression of additional ROS-detoxifying genes, helping restore redox balance. For example, to surpass oxidative stress caused by Origanum vulgare L. EO, Yersinia enterocolitica has upregulated genes involved in GSH-Px and methionine synthesis⁶⁹.

Indeed, physiological adjustments are translated into phenotypic adaptations such as an increase in MIC, a reduction in growth rate, or transient tolerance. Additionally, researchers have observed no uniform genotypic or phenotypic patterns in S. enterica Typhimurium strains after carvacrol treatment, underscoring the complexity and unpredictability of these non-genetic adaptation responses⁷⁰.

An important point is that bacterial populations in real-world environments are typically heterogeneous, comprising subpopulations that exhibit diverse physiological and metabolic states. These include persister cells, viable but non-culturable (VBNC) cells, and biofilm-associated cells, with some subpopulations surviving stress better than others^71–73.

Previous results⁷⁴ demonstrated that repeated exposure to Mentha piperita L. EO leads to changes in the physiological functions of the S. enterica Typhimurium PT4 strains, creating a bacterial subpopulation that maintains membrane integrity while transitioning into a VBNC state, as viable counts during the experimentation stop increasing. In this cellular state, the bacteria remain viable but fail to form colonies on standard media; instead, they can be revived under appropriate conditions⁷⁵. This is a physiological adaptation state in which bacterial metabolic activities and gene expression continue at low levels, even with altered cell envelopes, making antibiotic targets less active⁷⁶. Also, studies show that VBNC cells, once resuscitated, become sensitive again⁷⁷. Therefore, entry into VBNC is a transient and reversible stress‑induced survival strategy rather than a heritable resistance trait.

Another form of bacterial adaptation mechanism is biofilm-mediated adaptive protection, a unique collective strategy that confers robust protection against EOs and other antimicrobials. Bacterial survival in biofilms depends on structural, physiological, and genetic factors⁷⁸. Biofilms allow microorganisms to create communities by adhering to surfaces and encasing themselves in a self-produced matrix of extracellular polymeric substances (EPS)⁷⁹. Indeed, these EPS have the capacity to entrap and immobilize both single- and multi-species microbial communities, including living and non-living cells, as well as biopolymers, proteins, polysaccharides, lipids, extracellular RNA, and DNA⁸⁰. Consequently, this matrix plays a crucial role in protecting bacteria from external stressors by supplying nutrients and simultaneously limiting access of biocides, oxidants, antimicrobials, antiseptics, and other stressors inside the biofilm^81,82. Recent studies have highlighted the critical role of quorum sensing (QS) in biofilm formation and its contribution to bacterial resistance to external stressors^83,84. QS acts as a communication system that regulates gene expression and the production of infectious agents, notably extracellular enzymes and lysins (cell‑lysing toxins such as hemolysins), in accordance with external conditions^85–87. This communication system plays a vital role in pathogenicity, inflammation, and biofilm development, and also contributes to AMR development^33,88,89.

Therefore, within biofilm communities, numerous characteristics can enhance bacterial adaptation, making bacteria encased in biofilms more tolerant to EOs and other antimicrobials than planktonic bacteria^90,91.

Given the high density and close proximity of bacteria within the biofilm, plasmid transport occurs quickly and easily, resulting in a higher rate of HGT, particularly for antibiotic resistance^92,93. This process accelerates the spread of resistant bacteria. Nevertheless, cases of genetic transfer of EO-resistance, when it occurs, are underreported in the literature.

Furthermore, post-treated biofilm often harbors subpopulations of dormant bacteria, known as persister cells, that survive antimicrobial treatment. These are bacterial subpopulations that remain metabolically inactive in a temporary dormant state^35,94. This dormant state is linked to a prolonged lag phase as they wait for more favorable conditions to recover and repair themselves^95,96. Hence, since EO mainly affects metabolically active cells, this temporary physiological shutdown decreases EO’s efficacy and thereby confers increased tolerance rather than permanent resistance⁷⁴. Indeed, it is very important to recognize this physiological state during lag-phase extension as a mechanism of tolerance rather than genetic resistance.

For instance, after using Coleus amboinicus L. EO at sub-inhibitory concentrations to treat S. enterica Typhimurium biofilms, signs of bacterial re-growth and post-biofilm formation have appeared⁹⁷. This bacterial regrowth and post-treatment biofilm formation might be due to the presence of persisters within the biofilm. These cells are not genetically resistant but can undergo regrowth once the treatment is no longer present⁹⁸. It is worth noting that these cells could serve as a sheltered evolutionary reservoir for the transfer of resistance genes that have been acquired from previous biofilm bacteria to future cells joining the biofilm^3,99.

Nevertheless, according to many studies, EOs and their compounds play a crucial role in reducing biofilm formation and in downregulating virulence genes, even at subinhibitory concentrations. According to a previous study⁹⁷, genetic analyses of biofilm bacteria after exposure to sub-lethal concentrations of Coleus amboinicus L. EO identified four significantly downregulated virulence genes, namely flhD, fljB, fimD, and csgD, implicated in motility control, along with suppression of the hilA invasion gene. These genes are crucial for biofilm formation and survival, as reducing the expression of flhD, fimD, and fljB impairs bacterial motility, which is crucial for initial surface attachment and colonization. Additionally, lowering csgD significantly reduces curli and cellulose production, weakening the matrix and resulting in fragile or nonexistent biofilm with poor maturation and stability. Furthermore, hilA normally suppresses the expression of biofilm genes^100–102.

Subsequently, this investigation not only shows genetic changes in biofilm bacteria before and after EO treatment but also highlights how wisely combining antimicrobials with EOs can help reduce the spread of biofilm formation, particularly in the food industry, and aid in controlling biofilm‑mediated infections¹⁰³.

Furthermore, although the responses to EO treatment mentioned in earlier studies do not appear to involve stable genetic mutations and are reversible once stress is removed, repeated exposure of the same foodborne pathogens to sublethal EO doses can likely result in more stable resistance traits¹⁰⁴.

Genetic adaptive responses

In most cases, EO resistance evolves slowly because EOs contain multiple active compounds acting on diverse targets, meaning that bacteria would need to accumulate multiple genetic mutations to counter all their mechanisms of action^105,106. Nevertheless, evolutionary experiments demonstrated that the emergence of de novo genetic mutations under stressful environments is possible under prolonged selection¹⁰⁴.

Therefore, by analogy to antibiotic-induced resistance, when EO-induced resistance occurs, the induced antimicrobial resistance genes (ARGs) may potentially be transferred to other members via vertical gene transfer (VGT) or horizontal gene transfer (HGT)^107,108.

In the case of antibiotic resistance, once incorporated into a cell’s plasmid or transposons, ARGs can transfer spontaneously through three key mechanisms: conjugation, which occurs via direct physical contact between a resistant bacterium and a wild-type bacterium; transformation, which occurs when bacteria incorporate an external “free/naked DNA” fragment into their genome; and transduction, in which a bacteriophage injects a new ARG into a non-resistant strain^109,110. Moreover, current evidence suggests a new mechanism in the HGT process called “Vesiduction”, enabling ARG transfer using nanoscale spherical units produced during bacterial growth^111,112. These units, known as outer membrane vesicles, carry small plasmids and DNA fragments of ARG from the host bacteria to new microorganisms^113,114.

Previous investigation¹¹⁵ demonstrated that treatment with Citrus medica L. var. sarcodactylis Swingle EO has exhibited potent antibacterial activity, notably causing morphological changes in Listeria monocytogenes ATCC19115 strains. Then, to better understand how L. monocytogenes responds to the EO treatment, a comparative genome-wide transcriptome analysis was performed. Results have revealed the upregulation of some stress-response genes, including those involved in oxidative stress, chaperones, and membrane repair, along with significant and notably enhanced changes in chemotaxis (including cheV, cheA, cheR, motA, motB, fliM, fliN, and fliG), flagellum assembly, motility (including fliD, flgE, flgK, flgG, flgB, flgC, motA, and motB, among others), and the uptake of specific carbohydrates and metallic cations¹¹⁵.

Similarly, in a related investigation by Berdejo et al.²² using whole-genome sequencing (WGS) of S. enterica Typhimurium strains cyclically exposed to sublethal and lethal doses of carvacrol revealed single-nucleotide polymorphisms (SNPs) in transcriptional regulators associated with the oxidative stress response. Specifically, the identified SNPs include yfhP in the S. enterica Typhimurium strain exposed to sublethal doses of carvacrol, and soxR in the strain exposed to lethal doses of the same compound²². These genetic changes are thought to confer increased tolerance to carvacrol by boosting the cellular stress response, which helps improve carvacrol efflux or limits its intracellular accumulation.

These two studies explain how treated bacterial strains exhibited tolerance and increased survival after EO treatment: enhancing mobility by improving flagella rotation, triggering cell tumbles, and reorienting to escape from the EO, along with improving the absorption of carbohydrates to get more energy, and changing the uptake of various metallic cations like cobalt, iron, nickel, and zinc. Nevertheless, short-term exposure alone cannot confirm whether this EO-induced genetic variation is heritable. Instead, a prolonged treatment period should be used to better understand the possibility of EO resistance acquisition.

For instance, after prolonged cyclic treatments at low concentrations of S. aureus with Citrus sinensis EO, results revealed an enhanced survival rate under lethal treatments compared to the wild-type strain¹¹⁶. Additionally, using WGS, specific point mutations in hepT and accA were identified in the evolved strains, highlighting the roles of oxidative stress and fatty acid biosynthesis in the bacterial response to EO¹¹⁶. These metabolic mutations play a major role in improving membrane repair and reducing ROS accumulation, thereby increasing tolerance. Importantly, according to this study, these genetic variants were stable and heritable, demonstrating de novo genetic adaptation in response to EO exposure and, therefore, EO-resistance acquisition¹¹⁶.

Unlike antibiotics, EO does not directly increase mutation rates; instead, it may select for mutants with better growth fitness during extended sublethal exposure¹¹⁶. In other words, EO would select spontaneous, random mutants that emerge during the evolution assay and show enhanced growth fitness under EO stress.

Consequently, although direct bacterial resistance is limited, phenotypic adaptation, such as activation of efflux pumps, could eventually lead to stable resistance if exposure is frequent and prolonged. Therefore, continued monitoring of these adaptations is essential to gain a better understanding of EO-resistance acquisition, especially within foodborne bacteria.

Collectively, these studies make it clear that bacterial survival and adaptation are very likely, especially among foodborne pathogens. This poses a significant risk to food safety, economic losses, and, especially, consumer safety. When sub-lethal doses of EO are used against foodborne bacteria, the emergence of tolerance and resistance, even if rare, remains a major risk. The EO-resistant gene may then be transferred among microorganisms in real food systems.

This can be illustrated by a previous study¹¹⁷ on Vibrio cholerae strains, which, when exposed to a sublethal dose of carvacrol, showed downregulation of flrC expression and most class III flagellar synthesis genes as a way of stress tolerance. Furthermore, given that bacterial flagella, mainly flagellin, trigger the host’s immune response during infection^118–120. Therefore, using carvacrol at sublethal concentrations in food conservation could lead to V. cholerae becoming aflagellate and undetectable by the immune systems of infected individuals, which is totally undesirable. Furthermore, to thoroughly understand these adaptive mechanisms, research should involve repeated exposure to lethal and sublethal doses of EOs, using appropriate screening methods and ongoing monitoring.

Screening approaches to reveal bacterial adaptation to continuous EO exposures

Similar to what has been observed with antibiotics, recent research has highlighted concerns about the potential development of EO-resistant variants (RVs), referring to bacterial subpopulations that show stable and heritable reduced susceptibility to EOs or their components after repeated exposure¹¹⁶. Hence, to study the habituation of microorganisms and their potential to acquire resistance after being treated with different EO doses, multiple techniques are employed, including phenotypic, molecular, and multi-OMIC approaches, which are highly complementary to one another^121–123. As illustrated in Fig. 1, most of these studies begin by (1) repeatedly exposing the studied microorganism to lethal or sublethal concentrations of the EO or the EO-derived bioactive compound, then (2) carefully monitoring the minimal inhibitory concentration (MIC) and minimal bactericidal concentration (MBC) evolution over time^15,74, while also (3) tracking the membrane composition and its biophysical measurements^47,124, and (4) conducting efflux activity assays and phenotypic resistance tests²², along with (5) checking biofilm formation, motility, and virulence phenotyping⁹⁷. In addition, molecular techniques majorly utilized in this field include: (i) gene expression profiling using qRT-PCR, (ii) transcriptomics analysis (RNA-sequencing)⁷⁰, (iii) WGS and mutant selection^123,125, in addition to (iv) the use of multiomics approach^{28,121,126,127}. These approaches would improve the understanding of bacterial adaptation to EOs by analyzing data from genomic, transcriptomic, and metabolomic perspectives¹²⁸. Therefore, combining biological layers may effectively enhance the comprehensiveness and understanding of complex biological systems and their mechanisms, compared with studying and interpreting each layer individually.

Fig. 1. General workflow to analyze and understand microbial adaptation to EOs and their bioactive compounds.

This schematic outlines a general five‑step strategy for studying microorganisms’ habituation to repeated EO exposure and their potential to develop adaptation. The steps involve: (1) repeatedly exposing the microorganism to lethal or sublethal concentrations of the EO or the bioactive compound being studied; (2) carefully monitoring the evolution of MIC and MBC over time; (3) tracking membrane composition and its biophysical measurements; (4) conducting efflux activity assays and phenotypic adaptation tests; and (5) assessing biofilm formation, motility, and virulence phenotyping. The figure also highlights four key molecular techniques widely used in this field: (i) gene expression profiling using qRT-PCR; (ii) transcriptomics analysis (RNA-sequencing); (iii) WGS and mutant selection; and (iv) multi-omics approach.

In another study⁷⁴ using, plate counting and flow cytometry techniques revealed a decrease in cell culturability of epidemic S. enterica Typhimurium PT4 strain cyclically exposed to two different concentrations of Mentha piperita L. EO (Menthol representing 43.75%) for ~252 hours. Specifically, at a constant-dose exposure, culturability declined during the first cycles, indicating that the studied strain had developed tolerance traits. Results reveal that repeated exposure leads to changes in the physiological functions of the S. enterica strains, resulting in VBNC cells, as viable counts during the experiment stop increasing. Furthermore, using flow cytometry techniques, this study revealed that bacterial cells maintain activity of the ArcAB-TolC efflux pump under constant sub-MIC exposure. This efflux pump plays a crucial role in expelling antimicrobial agents to prevent intracellular accumulation and support membrane repair, thereby ensuring stress tolerance and bacterial survival¹²⁹.

This study combined two techniques to investigate EO-induced adaptation monitoring. First, plate counting methods are simple, inexpensive tools used to measure microbial survival after antimicrobial treatments and are considered reference methods^130,131. However, this approach detects only a small fraction of the total microbial population. Notably, studies on drinking-water quality have demonstrated that plate counts detect <1% of microbes, as most environmental and stressed cells cannot form colonies¹³². Consequently, following subinhibitory EO treatment, plate counts may fail to detect stressed bacteria and VBNC cells, thereby underestimating the survival rate.

Second, the flow cytometry method, which enables single-cell analysis of physiological responses and heterogeneity, making it possible to count injured and VBNC cells^133,134. This approach involves monitoring morphological and physiological changes in bacteria after treatment and reveals subpopulations with damaged membranes, altered membrane potential, and reduced esterase activity⁷⁴. Consequently, combining plate counting with flow cytometry provides a more complete picture of EO‑induced adaptation.

Meanwhile, after repeated exposure to lethal and sublethal concentrations of carvacrol, the comparison of single-nucleotide variants and structural variations in L. monocytogenes and S. enterica Typhimurium, each with their wild-type strains, has revealed the emergence of RVs in both strains^22,123. More specifically, using WGS technology, two single-nucleotide variations and three non-silent mutations were identified in L. monocytogenes after sublethal and lethal treatments, respectively. Among them are genes encoding the transcriptional regulators RsbT and ManR, which may have contributed to their enhanced resistance to carvacrol²². Similarly, WGS has enabled the identification of single-nucleotide variations in transcriptional regulators of the oxidative stress response, notably YfhP and SoxR, in S. enterica following treatment with carvacrol¹²³. This WGS approach not only provides information on the mode of action of the EO’s individual compounds against each microorganism, but also enables tracking of how RVs appear under varying conditions. This approach, commonly applied in AMR surveillance, can deliver specific genetic-level analysis, using a standardized method that allows results to be easily compared across different laboratories^135,136.

Furthermore, previous investigation⁷⁰ have examined the potential adaptation in S. enterica Typhimurium LT2 strains following 10 cycles of exposure to lethal doses of carvacrol. Phenotypic characterization showed that MIC values increased in four adapted strains compared to their wild-type strains, indicating a higher survival rate at lethal concentrations. Results reveal that across the four strains, WGS detected 10 mutations in rrsH, lon, sseg, rob, wbaV, nirC, flhA, fliH, upstream argR, and upstream yfhP genes. Notably, the flhA, lon, rob, and wbaV genes are known for their roles in developing adaptive traits not only to carvacrol but also to many other stressors, including antibiotics and heat treatment^137,138. Consequently, when a microorganism develops adaptation to a particular stressor, it can simultaneously become less vulnerable to other stressors that target the microorganism through similar mechanisms; this is known as the “Cross-adaptation” phenomenon^36,138,139. This observation was also demonstrated in a previous investigation³⁶ on the emergence of both cross- and unique antibiotic resistance after exposing bacteria to antiviral drugs. A detailed discussion on the cross-adaptation phenomenon and its impact on public health is provided in the next section.

Although many technologies are used to investigate and monitor microorganisms’ adaptation to EOs, each has its own strengths and limitations. Notably, the serial passage model is useful for revealing MIC shifts, especially under sublethal EO concentrations, which also mimic industrial repeated exposure^27,140,141. However, an increase in MIC alone is insufficient to determine whether the acquired adaptation is a resistance or a tolerance trait. As the MIC shift may reflect only temporary stress responses and may disappear after stress removal^78,142.

Furthermore, given the heterogeneity of bacterial populations, MIC determination has significant limitations, especially when examining EO-induced adaptation, as it captures only the overall population response rather than insights into individual cell behavior¹⁴³. This significantly affects the results obtained and complicates their reproducibility, since some cells may start growing sooner, while others may die, recover after initial inhibition, or tolerate concentrations above the MIC.

Moreover, omics technologies are very useful for identifying stress-response pathways by revealing oxidative stress responses, shifts in efflux pump activation, and membrane remodeling^144–146. Yet these bacterial changes do not necessarily indicate stable adaptation, as they can reflect a general stress response rather than an EO-specific resistance mechanism⁴⁴. Genomic approaches, such as WGS, can detect stable genetic mutations that induce inherited resistance¹²³. Nevertheless, studies often fail to identify stable genetic changes following EO habituation, supporting the hypothesis that EO adaptations are often phenotypic and reversible rather than stable mutations⁷⁶.

Another challenge in tracking, comparing, and interpreting results from EO-adaptation experiments is the significant variability in EO chemical composition. The chemical profiles of EOs can vary even within the same botanical species, affected by factors such as chemotype, geographic origin, environmental conditions, harvesting time, plant developmental stage, and extraction technique^10,147,148. Thus, each EO has its own major and minor compounds, potential synergies among these compounds, and specific physicochemical properties^149–152. Therefore, inducing variability in chemical profiles can lead to dissimilar bioactive properties for the same botanical species across separate studies^153,154. A recent study on Cinnamomum cassia leaves found that spring EO had the highest trans-cinnamaldehyde content, about 92%, compared to EOs from other seasons. The spring EO also showed antibacterial activity twice as strong as the EOs from other seasons against S. aureus and L. monocytogenes¹⁵⁵. These significant variations in antimicrobial activity indicate that results across EO studies, including those examining EO-induced adaptation, may lack reproducibility and that meaningful comparisons may be difficult to make.

To sum up, many recent studies support the idea that bacteria may undergo genetic adaptation as a survival mechanism when exposed to EO over extended periods, and could also play a role in the emergence of AMR^137,138. However, to the best of our knowledge, no clear standardized pattern has been directly linked to bacterial resistance in strains exposed to repeated doses of EOs or their main bioactive compounds.

Adaptive responses of bacteria to EO exposure: Homologous, cross-adaptation, forward and backward CS

Many studies over the last twenty years have explored the bacteria’s potential to adapt to EOs exposure. These studies tested how repeated exposure of susceptible bacterial strains to fix or increase lethal or sublethal doses of EOs or their bioactive compounds may affect their genomic and phenotypic traits^22,140,156. This repeated exposure reveals two types of adaptation: (i) homologous adaptation, where bacteria adapt to the same stressor over time, and (ii) heterologous adaptation, also known as cross-adaptation, where bacteria gain adaptation to stressors they have never encountered before^46,74.

Homologous adaptation was observed in S. enterica Typhimurium LT2 and S. pyogenes strains after exposure to lethal and sublethal doses of carvacrol for 1–4.5 hours^22,47. After treatment, WGS identified six single-nucleotide variants, notably in genes encoding transcriptional regulators of RNA 16S ribosomal components, HTH IscR transcriptional regulator, redox-sensitive regulator, membrane transporter (nitrite transport), membrane receptor, and protease, specifically in the rrsH, yfhP, soxR, nirC, fepA, and lon genes²². Hence, the composition of membrane phospholipids and fatty acids, as well as the oxidative stress response, were modified in treated strains compared to untreated S. enterica Typhimurium, resulting in a restored redox balance that improved bacterial survival under sublethal doses, despite the rise in membrane fluidity^157,158. These results demonstrate that repeated exposure to carvacrol has led bacteria to adapt and modify their typical physiological characteristics, enabling them to tolerate carvacrol-induced damage.

Moreover, these same S. Typhimurium strains that have developed tolerance to carvacrol also showed cross-adaptation to a number of antibiotics, notably tetracycline, chloramphenicol, nalidixic acid, rifampicin, norfloxacin, novobiocin, trimethoprim, and ampicillin²². These antibiotics are known for their mechanisms of action that target cell wall integrity, ribosomes, RNA, and DNA synthesis, as well as the thymidine synthesis pathway, which are the same cellular targets as carvacrol^159,160. This raises the question of whether bacterial adaptive responses are specific to the type of stressor or damage; in other words, do bacteria recognize the stressor’s identity, or simply respond to the damage they detect on themselves?

Another study⁴⁶, has found that 16 hours of cyclic exposure of S. aureus GL 5674 and GL 8702 strains to sublethal concentrations of cinnamaldehyde resulted in a homologous adaptation, allowing the habituated strains to grow at the MBC, unlike the wild-type strains. However, neither strain grew at concentrations above its MBC. Additionally, cross-adaptation has occurred, showing enhanced adaptive plasticity and pH tolerance; the latter dropped from 4.5 and 4.0 to 3.0 and 3.5, respectively, for the S. aureus GL 8702 and S. aureus GL 5674 strains. Correspondingly, continuous exposure of the wild-type strains to the minimum growth pH led both S. aureus strains to develop cross-adaptation to cinnamaldehyde. However, this time the strains were able to grow at concentrations above the cinnamaldehyde MBC successfully.

In this case, cross-adaptation to cinnamaldehyde following continuous exposure to the minimum growth pH has enabled strains acclimated to the acidic environment to tolerate higher concentrations of cinnamaldehyde (above the cinnamaldehyde MBC) than those adapted solely by constant exposure to cinnamaldehyde (not exceeding the cinnamaldehyde MBC). This observation indicates that cross-adaptation may sometimes confer higher stress tolerance to the secondary stressor than homologous adaptation achieved through direct exposure to the same stressor.

Furthermore, adaptation to one specific stressor can sometimes lead to increased sensitivity to another³. This appears as the opposite form of cross-adaptation and is referred to as “collateral sensitivity” (CS)^161–163. One example is Enterococcus faecium strains that developed resistance to vancomycin have simultaneously become more vulnerable to pleuromutilin antibiotics¹⁶¹. This can be explained by van gene expression in vancomycin-resistant strains interfering with msrC, which encodes the ABC-F ribosomal protection protein, thereby making the bacteria more sensitive to pleuromutilin antibiotics, notably lefamulin, which targets bacterial ribosomes^163,164. Although most studies reporting cases of CS emergence mainly concentrate on antibiotics. These findings can be highly valuable for developing smart, sustainable strategies for using EO treatments alone or combined with other bioactive molecules across various domains, particularly to address AMR challenges. The CS phenomenon may be a key factor in designing alternative usage of EOs or an EO-antibiotic combination treatment to address the evolution of AMR^165,166. For instance, a previous study¹⁶⁷ has demonstrated that the combination of carvacrol with antibiotics produced better results than monotherapy against Methicillin-resistant S. aureus (MRSA), particularly when combined with linezolid, minocycline, or sulfamethoxazole, leading to increased bacterial susceptibility to antibiotics. Moreover, using thymol (512 µg/mL), a well-known EO’s bioactive compound, with ciprofloxacin (0.125 µg/mL) showed strong synergistic effects against both planktonic and biofilm cultures of two P. aeruginosa GC14 and ATCC 9027 strains, which could be a promising therapeutic approach against infections caused by P. aeruginosa biofilms¹⁶⁸. Nevertheless, even after a CS acquisition, susceptibility may not necessarily persist in subsequent evolutionary stages, since CS’s potential to prevent MDR development or to predict and promote antimicrobial sensitivity restoration has yet to be experimentally verified¹⁶⁹.

Beyond classical forward CS, where resistance to one antimicrobial agent increases susceptibility to another, the relation between two antimicrobials can also work in the opposite direction: an inverse phenomenon known as “backward collateral sensitivity” (BCS) has been reported in the literature, particularly in sequential treatment regimens³⁰. In this case, resistance to the second antimicrobial administered increases susceptibility to the first, thus offering a potential strategy to help reduce existing resistance levels^30,169,170.

Indeed, it was previously demonstrated that E. coli strains often become more sensitive to β-lactam antibiotics, such as piperacillin, after developing resistance to aminoglycosides, including gentamicin¹⁶⁹. This is due to incompatible modifications, notably increased efflux pumps and weakened proton motive force¹⁷¹. Furthermore, after exposure to tigecycline, a reduction in MIC and a 100% restoration of polymyxin B sensitivity in E. coli were observed. Such a re-sensitization could be highly relevant for chronic diseases where tigecycline and either polymyxin B or the polymyxin B analog colistin are required, such as in cystic fibrosis cases¹⁷². Moreover, it was previously stated that “re-sensitization occurs even when bacteria are adapted through plasmid-bound mutations instead of chromosomal changes”¹⁶⁹, which indicates that restored susceptibility can possibly occur via both VGT and HGT. However, this phenomenon has been reported only in the context of antibiotics, and to the best of our knowledge, no case study on the emergence of BCS to EOs has been reported. Still, these results can be very inspiring for improving the benefits of EO applications, especially for mitigating the emergence of adaptation to EOs and antibiotics across sectors, notably in healthcare and the food industry.

Figure 2. provides a schematic overview of the main adaptive responses of bacteria after exposure to antimicrobial agents.

Fig. 2. Bacterial adaptive responses to antimicrobial exposure.

This figure illustrates the major adaptive responses observed in bacteria after exposure to antimicrobial agents. (1) Homologous adaptation: repeated exposure to a given stressor increases adaptation to the same agent. (2) Cross-adaptation: adaptation to stressor (A) leads to reduced susceptibility to a different agent. (3) Forward collateral sensitivity: acquiring adaptation to an initial antimicrobial (A) enhances bacterial susceptibility to a subsequently applied agent (B). (4) Backward collateral sensitivity: adaptation to the second antimicrobial (B) induces re-sensitization to the first agent (A). Arrows indicate the direction of antimicrobial exposure and the associated changes in susceptibility.

Overall, evidence suggests that EOs and their individual compounds’ antibacterial activities can be positively or negatively affected by adaptation to other bacterial stressors, such as antibiotics, pH, temperature, phytocompounds, and others, leading to cross-adaptation or forward CS^29,173. This happens because adaptation acquisition often induces physiological changes in bacteria, which can increase or decrease their vulnerability to other treatments¹⁶³.

Indeed, bacterial adaptation to EOs tends to occur more slowly and less frequently than antibiotic resistance development¹⁷⁴. For instance, results from E. coli and S. enterica Typhimurium pre-exposed to two commercial EOs at subinhibitory levels have shown a minor increase in MIC by about 14%–28%. In contrast to the pre-exposure to amoxicillin and colistin, which have led to significant MIC increases, reaching up to 100% for amoxicillin-selected E. coli and 200% for colistin-selected S. enterica Typhimurium¹⁷⁴. This low adaptation rate may be due to the fact that EOs are complex mixtures of bioactive compounds that act simultaneously on multiple targets in the microorganism, and to their vast compositional variability, which depends on plant chemotype and other previously discussed factors^10,175,176.

Additionally, EOs’ use is less standardized and more challenging to track and study, unlike standard antibiotics that are broadly used in both humans and animals, which often leads to clinical resistance^177–179. This could explain the lack of clinical studies on EO adaptation and why most case studies focus on laboratory-adaptive evolution or experiments with sublethal exposure^46,126,174. Another point is that standard serial-passage comparisons are uncommon in these laboratory-adaptive evolution experiments, where the same strain, passage number, and sub-MIC concentrations should be used. Instead, studies evaluate EOs and antibiotics using different techniques, leading to variable conditions and results that are hard to compare^{116,180–182}.

Reversibility and stability of EO-induced adaptation

After evolutionary treatment using EOs or their derived compounds on different strains, studies have revealed various forms of adaptation^29,46,162. This observation can be attributed to the nature of microorganisms, which, when exposed to fluctuating environmental conditions, tend to adapt in order to maintain their vital physiological traits and metabolism, mainly through natural selection, genetic recombination, HGT, and DNA damage repair^22,140,156. Notably, two major types of EO-adaptation were reported: primarily physiological adaptation, and less frequently, genetic resistance^22,140,156.

Physiological adaptation, also referred to as phenotypic or structural adaptation, is recognized as a short-term response^183,184, characterized by temporary changes in gene expression^185,186. These genetic variations can impact cell membrane fluidity, thereby affecting permeability and transport proteins, which in turn influence the oxidative stress response and membrane stress regulons^8,62,91,185. For instance, previous investigation¹⁷⁴ has reported that genotypic analysis of E. coli and S. enterica after repeated exposure to subinhibitory doses of two commercial EOs revealed increased tolerance against oxidative stress and a shift in cell energy toward efflux pump activity. This physiological adaptation only permits cells to survive and overcome the stressor. However, it is genetically unstable and cannot be transferred among bacteria, unlike the genetic resistance trait.

Under favorable conditions, free planktonic bacteria tend to form biofilms by attaching and organizing themselves within self-produced EPS, creating a matrix that provides physiological protection from environmental stress^88,93. This process provides an effective survival strategy, which is defined as an innate form of phenotypic tolerance to antimicrobials^87,187,188. Besides, biofilm plays a crucial and highly threatening role in the transfer of resistance genes, especially via HGT, which has been reported to be 700 times more efficient than among free-living, planktonic microbial cells^187,189,190. This organized community represents a form of transient phenotypic tolerance mechanism that decreases bacterial susceptibility to EOs without requiring stable genetic resistance^191,192.

In general, stable EO-induced adaptation is often rare and less likely to persist in the absence of EO-induced pressure¹⁷⁴. One example is Yersinia enterocolitica, which, when facing Origanum vulgare L. EO, activates genes related to glutathione peroxidase expression and amino acid synthesis to prevent cell oxidation damage, while also triggering the phage shock protein response and reducing unnecessary energy use⁶⁹. These phenotypic adaptations are temporary and expected to revert once the EO pressure is removed.

In phenotypic adaptation, gene expression returns to baseline once the stress is eliminated¹⁹³. Consequently, the bacterial membrane composition normalizes, and the microbial population regain their original susceptibility. This explains the variation in phenotypic adaptation among bacterial populations, especially when facing changing environments, such as antibiotic or EOs treatments^194,195.

Alternatively, genetic adaptation involves changes in the microorganisms’ DNA to ensure survival and is viewed as a permanent mutation that remains even after the removal of EOs selective pressure^22,174. This was illustrated by a previous study¹¹⁶, which reported genetically encoded changes in S. aureus strains after prolonged cyclic treatments with Citrus sinensis EO at sublethal doses, notably in hepT and accA genes. This variation was regarded as a stable genetic adaptation. Furthermore, an emerging mutant population of L. monocytogenes was observed after repeated exposure to sublethal concentrations of Thymbra capitata EO, revealing a single nucleotide polymorphism in the plsC gene, which encodes the enzyme of phosphatidic acid production, a precursor for cell membrane synthesis¹²⁵. Similarly, a related study found that the selective pressure of carvacrol on L. monocytogenes strains led to genetic resistance, suggesting a likely stable variant¹²³.

To date, most experimental studies have reported physiological or chromosomal regulatory adaptations in response to EO pressure, rather than plasmid genetic adaptations^22,70,174. However, by analogy with how plasmids can help preserve ARGs after antibiotic pressure¹⁹⁶, an EO-induced resistance mutation can most likely also be carried by microbial plasmid-encoded DNA. Plasmid DNA is known to carry genes that may not be essential for survival but provide beneficial traits, such as AMR¹⁹⁷. Moreover, plasmid mutations are more likely to be maintained in the same cell line through VGT or transferred to other bacterial cells through HGT^198,199. In fact, HGT involves the transfer of genetic material between organisms, which can either replace existing genes or introduce new ones into the genome^108,112. Furthermore, this genetic transfer is a very common process in prokaryotic cells, where it plays an essential role in evolution and adaptation to environmental changes⁴². However, the plasmids can be lost over generational transfer once the selective pressure is removed^200,201. Once these genetic adaptations are incorporated into the cells’ genome, the newly acquired ability may persist, and an emergence of resistant mutants could be documented¹²⁵.

Nevertheless, not all transfer events are biologically significant; instead, the outcome may reflect an ongoing evolutionary process that is only occasionally beneficial²⁰². In some cases, once in a stress-free environment, the acquired resistance gene makes it very difficult for the bacteria to survive; this is known as the fitness cost²⁰³. As a result, the bacteria eliminate their expression through compensatory chromosomal mutations, recognized as second-site resistance mutations, which bypass the need for high-level gene amplification²⁰⁴. Briefly, fitness cost occurs when the expression of the adaptation gene negatively impacts the characteristics of microorganisms, causing them to grow more slowly, become more vulnerable to other bacteria, or survive less effectively under normal growth conditions^205,206.

Previous study¹³⁸, illustrates the fitness costs associated with S. enterica Typhimurium using cyclical exposure to prolonged sublethal concentrations of carvacrol at 0.1 µL/mL. This process generated a resistant variant, noted RV-CAR, characterized by mutations in the sseG, flhA, wbaV, and lon genes, which have also exhibited cross-adaptation to heat and antibiotics compared to their wild-type (WT) counterparts. Under optimal growth conditions (37 °C), RV-CAR demonstrates a slight fitness cost, as evidenced by a reduced growth rate and an extended lag phase, indicating delayed recovery after adaptation. These findings align with observations of fitness cost reported in the study carried out by ^{[NO_PRINTED_FORM] 125} on L. monocytogenes RV to T. capitata EO. However, under cold conditions (8 °C), the growth rate of S. enterica RV-CAR strain has doubled compared to S. enterica wild-type. Notably, improved growth rate at low temperatures indicates better growth fitness at refrigeration temperatures, potentially attributable to a mutation in the lon gene, which encodes the Lon protease responsible for protein quality control during stress responses, thereby preventing cellular damage caused by low temperatures¹³⁸. Consequently, the stability of cross-adaptation traits depends on whether they result from temporary phenotypic changes or fixed genetic modifications.

Many studies have shown that bacterial adaptation to EOs and their bioactive compounds causes genetically lasting resistance less often than with antibiotic use^2,196. This is one of many reasons why researchers consider EOs as a promising option to substitute for or combine with other antimicrobial agents that have long-term antimicrobial potential against the proliferation of AMR^2,10.

Nevertheless, the potential reversibility and stability of microbial adaptation to EOs’ treatment are often overlooked when investigating the antimicrobial potential of EOs, especially in identifying the fundamental forces responsible for this complex process and the metabolic pathways involved. Instead, gaining a better understanding of the mechanisms by which bacteria develop temporary or permanent adaptation to EOs is crucial for a more accurate evaluation of their safety and long-term use.

Bacterial perception and interaction with its environment: is bacterial adaptation stressor-specific?

Once in the environment, bacteria use all their mechanisms and abilities to survive and grow, which primarily depend on perceiving and recognizing chemical, mechanical, and biological signals in their surroundings^207–209.

Among the primary mechanisms bacteria use to achieve this goal is chemotaxis, which permits their movement toward or away from a molecular stimulus, thereby enhancing nutrient and energy uptake^210,211. This is mediated by specialized transmembrane proteins, known as chemoreceptors, that detect and respond to chemical gradients in the environment²¹². Additionally, chemotaxis helps bacteria in sensing the structure and texture of their surroundings²⁰⁹. Moreover, bacteria are known to be equipped with machinery that enables them to sense the environment’s physical forces through mechanotransduction, thereby stimulating multiple behavioral adaptations in the cell^213–215. Besides, to defend themselves effectively, these prokaryotic cells must identify several biological entities, like phages, that influence and shape bacterial communities’ survival strategies^207,216.

For EO specifically, once it enters a bacterial cell, it induces extensive physicochemical damage, including destabilization of the membrane, cell-wall lesions, and the generation of ROS and metabolic imbalance^217,218. Bacteria recognize this induced damage via conserved stress sensors and subsequently trigger adaptive responses to restore cellular homeostasis⁶⁹.

In Staphylococcus aureus, exposure to Litsea cubeba EO causes cell wall and membrane damage and upregulates the VraSR two-component system²¹⁹. The membrane-bound histidine kinase VraS undergoes autophosphorylation when peptidoglycan integrity is compromised and transfers this phosphate to VraR. This process induces the expression of peptidoglycan biosynthesis genes, such as pbp4, and increases the production of histidine and lysine biosynthesis genes to repair the cell wall²¹⁹.

Furthermore, to coordinate and communicate among microorganisms present in the same environment, the intercellular communication system, known as the QS system, takes place^220,221. QS, mediated by the production of self-inducing molecules, also known as bacterial pheromones, is a key communication mechanism used to harmonize their interactions, allowing cell-to-cell communication that permits coordinated bacterial collective activities, mainly: virulence and pathogenicity, biofilm formation, antibiotic resistance, and other cellular functionalities^85,222. However, bacteria continue to produce small amounts of signaling molecules called autoinducers (AIs) to measure the surrounding cellular density, and the gene expression of these autoinducers only occurs if the bacterial population reaches the critical cell density required for quorum^33,87,223. By combining these mechanisms, bacteria typically survey their environment to anticipate changes, respond appropriately, and survive as long as possible.

Most studies report that, in the presence of stress agents, the bacterial response primarily depends on the damage induced¹⁸⁴. Meaning that bacteria do not recognize the exact stressors (e.g., carvacrol, thymol, O. vulgare L. EO, ampicillin, colistin, etc.), instead, they typically sense physicochemical shifts in their environment (e.g., membrane modification, redox shifts, DNA damage, nutrient deprivation, heat, and cold) and thereby activate damage-specific gene expression^184,224,225. As illustrated in Table 1, different types of stressors can trigger shared damage-driven stress responses in bacteria. For example, when confronted with stressors targeting the cell envelope, such as EO, antibiotics (ramoplanin, daptomycin, and teicoplanin), or cold, among others, bacteria activate shared gene regulatory systems in response to cell membrane and envelope perturbation^{219,226–228}. As is the case with the LiaFSR (lipid II interacting antibiotic) gene expression, which contributes to the development of resistance in Gram-positive bacteria, including group A Staphylococcus spp., Lactococcus lactis, B. subtilis, and E. faecalis^229–231. Another example of a damage-specific response is the SOS system, a DNA repair pathway in bacteria that is activated by DNA damage caused by high pressure, acid, oxidants, EOs, and antibiotic exposure^232,233. For instance, exposing E. coli, P. aeruginosa, and Proteus spp. strains to R. officinalis L. EO, camphor, and fluoroquinolones such as ciprofloxacin, enrofloxacin, and moxifloxacin have revealed a significant upregulation of recA expression that coordinates the self-cleavage of the SOS transcriptional repressor lexA^233–235. Both these genes are SOS regulators that activate the expression of genes encoding various DNA repair proteins²²⁵. Bacterial gene expression appears to be more damage-specific than stress-specific.

Table 1.

Distinct stressors triggering shared damage-driven stress responses in bacteria

Damages	Stressing agent	Microorganism	Activated regulators	Gene function	References
Cell envelope and Cytoplasmic membrane perturbation	Laspartomycin C	B. subtilis	LiaFSR	Regulates and coordinates cell membrane remodeling	229
	Vancomycin, polymyxin B, and colistin	Group A Streptococcus	LiaFSR/SpxA2		231,298
	Daptomicin	E. faecalis	LiaFSR/LiaY		227
	Vancomycin, oxacillin, and fosfomycin	S. aureus	VraT/VraS	Phosphorylates its own histidine residues into VraR and then sends signals for peptidoglycan regulation	299
	L. cubeba L. EO	S. aureus	VraS		219
Oxidative stress	Paraquat and H2O2 (oxidizing agents)	Paraburkholderia xenovorans	SoxS, SoxR, and OxyR	Induces expression of sodA, sodB, oxyR, ahpC, and katG, and other genes that protect from oxidative damage	300
	Fluoroquinolones	-	OxyS		301
	Quinolones	Acinetobacter spp.	OxyR/KatG		302
	Cinnamomum burmannii L. EO	S. aureus	AhpC/F, TrxR,		303
	p-cymene	P. xenovorans	AhpCF/AhpCD/Ohr		304
Metabolite and energy disfunction	Gentamicin, kanamycin, and tetracycline	P. protegens	(p)ppGpp regulon	Plays a key role in suppressing primary metabolism and activating secondary metabolism	305
	Norfloxacin	S. sonnei	RelA/SpoT	Encodes the enzyme responsible for ppGpp synthesis	306
	Ampicillin, norfloxacin, ang gentamycin	Acinetobacter spp.			307
Protein misfolding	High-temperature	E. coli	sigma factors: σE/σ32 Chaperones: DnaK/DnaJ and GroEL/GroES	Regulates protein homeostasis via repair or degradation of misfolded proteins	308
		Pseudomonas spp.	Sigma factor: RpoH/(σ32) Chaperones: GroEL and DnaK		309
	Citral	A. baumannii	DnaK and GroEL		310
	O. vulgare L. EO, thymol, and carvacrol	S. enteritidis	ClpB, HtpG, LuxS, IbpB		126
Nucleic acids damage	Enrofloxacin, moxifloxacin, and ciprofloxacin	E. coli	LexA/RecA	SOS regulators that activate the expression of genes encoding various DNA repair proteins	234
	Ciprofloxacin	Proteus spp.	RecA		235
	Camphor	S. mutans, E. faecalis, S. aureus, and E. coli	LexA/RecA/RecN/UmuC		311
	Myristic acid	L. monocytogenes and S. aureus
	R. officinalis L. EO	P. aeruginosa	LexA		233

It is also noted that several EOs’ individual constituents impose oxidative stress on bacteria. For example, S. Typhimurium evolved under carvacrol stress has mutations in the redox‑sensing transcription activator soxR that enhance tolerance to carvacrol‑generated ROS²².

Another point to consider is that this similarity in the expression of resistance genes in response to damage is one of the primary reasons explaining cases of cross-adaptation between different bioactive stressors, even between those of different types^46,159,160. For instance, S. Typhimurium RV after carvacrol treatment exhibits cross‑adaptation to heat and several antibiotics, notably ampicillin, cephalexin, and ciprofloxacin, among others⁷⁰.

Therefore, if the bacterial response depends on the induced damage rather than the stressor itself, what might be the biological reasons making resistance to EOs stress uncommon and less likely to persist, unlike antibacterial resistance¹⁷⁴?

Impacts on food industry

Nowadays, in response to the growing consumer demand for more natural, eco-friendly products, particularly within the food sector, the substitution of artificial food preservatives with natural alternatives, such as EOs, is becoming increasingly popular^10,236–238. One of the main reasons EOs and their bioactive compounds are particularly famous is their wide-spectrum antibacterial and antioxidant properties^{25,186,239,240}. This is why many food processing industries began incorporating them into their processes to ensure additive-free, safe, and pathogen-free food^241–243.

Recent studies have demonstrated that the emergence of EO-adapted strains in the food sector following repeated exposure is highly likely and poses an important and dangerous issue^125,178,244. This adaptation would directly affect food salubrity and thereby endanger consumers’ safety, even though the chance of developing genetically stable EO resistance is considered relatively low^245,246.

The emergence of EO adaptation in the food sector is often linked to the misapplication (improper application methods) of EOs or their derived compounds when added to food in order to control spoilage and prolong shelf life^247,248. An example of this misapplication is the introduction of EOs into the food system at sub-inhibitory doses. This may be intentional to reduce the strong aroma of these natural bioactive molecules in food or may occur unintentionally due to interactions between EOs and food components^249–251. Indeed, the antimicrobial activity of food preservatives strongly depends on the food’s physical traits, such as its state (solid, liquid, or porous), and its chemical composition, including lipids, proteins, carbohydrates, and water activity^252,253. Therefore, an in vitro MIC may be lower or higher than necessary depending on the food matrix. In particular, the antibacterial activity of EOs can be enhanced by the presence of certain food compounds, such as organic acids and alcohols, which promote bacterial death by lysing the cell phospholipid membrane²⁵⁰.

Furthermore, in an acidic food matrix (low pH), the lipophilicity of EOs increases, thereby improving their ability to penetrate the bacterial phospholipid membrane^122,254. Proteins and carbohydrates, on the other hand, can interfere with the antibacterial activity of EOs’ bioactive compounds, given their ability to bind with the EOs’ phenolic compounds, which stops the hydroxyl groups’ antimicrobial activities and reduces their bioavailability^255,256. In a lipidic matrix, EOs’ lipophilic compounds tend to dissolve into the fat phase of the food matrix, reducing their availability in the aqueous phase, where microorganisms are generally dominant²⁵¹. Also, since lipids usually form a protective barrier around the bacterial membrane, blocking interactions between the organisms and EO’s functional groups^257,258, resulting in an insufficient lethal dose of EO.

Consequently, higher concentrations of EO in the food matrix would be required to achieve the same bactericidal effect observed in vitro^10,259. These interactions reduce the concentration of free EO components, thereby exposing bacteria to sublethal doses rather than achieving complete inhibition²⁶⁰. In such cases, sublethal exposure can produce phenotypic adaptations that decrease susceptibility to antimicrobials, heat, and oxidative stress⁶⁷.

EOs’ implications in the food industry do not only include direct food applications but are also widely used in environmental sanitation and surface disinfection^261–263. A recent study indicates that using sub-MIC doses of various bioactive molecules from EOs in food processing sanitation could increase the risk of developing EO-tolerant bacteria²⁸. In their study, the authors showed that S. enterica serovar Enteritidis can develop cross-adaptation to continuous sublethal linalool exposure, resulting in tolerance to previously lethal citral concentrations²⁸.

In the food industry, balancing practical effectiveness with sensory acceptance is crucial^248,249. More specifically, the selection of bioactive compounds, whether used alone or in combination with other molecules in food antimicrobial treatments, should be done carefully to avoid bacterial adaptation and therefore treatment failure^179,259. For instance, although many EOs and their bioactive compounds can have a significant effect on the same bacterium, notably in food conservation, choosing a treatment that maximizes consumer safety and acceptance would be optimal^147,236,264. One example is the application of Thymus vulgaris L. and Origanum vulgare L. EOs to carp fillets (Cyprinus carpio), which have been shown to enhance sensory qualities, induce beneficial effects, and prolong the shelf life of fillet samples by five to six times when spraying 112 ± 13 mg of each EO on both sides of the fillets^248,265.

Nevertheless, some EOs and their individual compounds may favor bacterial survival and, surprisingly, enhance their tolerance to EOs treatment^266,267. Namely, rosmarinic acid, a bioactive molecule in Rosmarinus officinalis EO, when encountering P. aeruginosa PAO1, acts by mimicking AIs, competing with the natural ligand C4-HSL of P. aeruginosa. By doing so, it activates a broad QS response, leading to bacterial tolerance through increased biofilm formation. However, this same compound effectively inhibits biofilm formation of Aeromonas hydrophila³³. This means that EOs’ bioactivity is largely context-dependent; the same concentration may be effective in one condition yet favor tolerance in another^179,266. Consequently, precautions should be taken when targeting both biofilm and planktonic bacteria to ensure effective antimicrobial activity and prevent the emergence of EO adaptation^267,268.

It is widely known that the direct application of EOs in food is often limited by their volatility, intense aromas, poor diffusion, and chemical instability^269–271. Hence, to overcome these issues, the food industry is increasingly adopting encapsulation strategies, namely microencapsulation and nanoencapsulation, to protect EOs and enhance their release, thereby improving their stability and efficacy in food conservation^272–274.

However, in practice, even EOs that are successfully encapsulated and exhibit strong in vitro activity may still fail to achieve the desired antibacterial effect in real food systems^32,275. This can be due to several factors, namely, a slow encapsulation system release or an inability to maintain sufficient EO concentration over an extended period, which would result in sublethal exposure to food-borne bacteria^271,275,276. Furthermore, when encapsulated, the effects of food compounds can limit the amount of EOs delivered and hinder the achievement of adequate antimicrobial effectiveness²⁷⁷. According to numerous studies, the release profiles of EO from capsules are highly dependent on food conditions, such as pH, humidity, and temperature^32,278. Two case studies of nanocapsules containing Carum copticum EO and carvacrol revealed that at acidic pH, both encapsulation systems released significantly more bioactive compounds than at alkaline pH^279,280. Similarly, high humidity boosts the release of encapsulated bioactive molecules by accelerating carrier degradation^281,282. Also, varying environmental temperatures caused chitosan polymer chains to alter their shape, decreasing the pore space between particles and consequently obstructing release from microcapsules²⁸³.

Consequently, it is imperative to choose the right encapsulation method and carrier material, as they may directly affect the release behavior of the encapsulated bioactive agent and, eventually, its antimicrobial performance in the food system³². Any mismatch between the encapsulation design and the food application may undermine the system’s intended benefits and increase the risk of bacterial adaptation. Likewise, treatment with sublethal doses of EOs can induce the appearance of sublethally injured cells^35,284, which can regrow normally under favorable environmental conditions, posing a significant health threat^285,286. Subsequently, once ingested, sublethally injured microorganisms can cause infections, intoxications, and, in some cases, evade detection by the immune system^117,119,120.

Furthermore, bacteria in food systems are often not found as planktonic cells but rather protected within biofilms on food and food-contact surfaces, such as vegetable skins, dried meat surfaces, or polysaccharide-based edible biofilms^287–289. Biofilms are well known for blocking antimicrobials: their EPS creates an impenetrable shield that prevents agents from diffusing to deeper bacterial cells^91,221. This will hinder both free and encapsulated EOs from diffusing into the biofilm^78,267. Consequently, limited diffusion of EOs through these barriers poses a significant challenge, potentially causing sub-lethal exposure and increasing the risk of EO-induced adaptation within biofilms^94,290,291. As a result, the risk of EO-adaptation may significantly increase and spread over time, as is the case now with AMR^140,181.

Taken together, these problems could reduce the effectiveness of EO-based packaging and EOs as natural alternatives to industrial preservatives in the food sector in the future. Subsequently, more robust spoilage organisms and food pathogens will emerge, ultimately posing a threat to public health. Research on this topic, however, remains limited, especially in long-term evolution studies. Moreover, most research on bacterial adaptation to EOs is conducted in culture media rather than in food matrices^116,123,125; therefore, it is difficult to extrapolate the results to real food preservation scenarios.

In summary, it is very important to develop more effective, context-dependent strategies for the application of EOs in the food industry to prevent bacterial adaptation, which can cause conservation treatments to fail. One example is combining EOs with other selected treatments, such as heat, pressure, or even other biological treatments^292–294. This combination of multiple preservation methods is known as hurdle technology^295–297.

Conclusion

The reviewed evidence shows that bacterial adaptation to EOs is indeed real, but it is generally highly context-dependent and often only temporary. Adaptive responses rarely translate into stable genetic resistance; instead, they are physiological adjustments and stress-induced tolerance that can disappear once the stressor is removed. This indicates that the EO application can stay effective when used wisely and in combination with other methods, but its effectiveness depends on understanding how adaptation changes over time and in various conditions.

Importantly, it has been shown that bacteria respond to EO-induced damage rather than to the specific chemical identity of the EO. This means that the type of cell damage determines which repair and survival pathways are mobilized within the microorganism. Hence, recognizing these damage-specific responses helps explain why exposure to different stressors can produce overlapping phenotypes and why cross-adaptation to unrelated antimicrobials sometimes emerges. Conversely, the collateral sensitivity phenomenon highlights the potential to exploit these damage-driven pathways to develop more effective preservation strategies.

In food safety, these adaptive phenomena can create both risks and opportunities. Cross-adaptation to EO could contribute to the persistence of adapted pathogens in processing environments, thereby reducing sanitizer efficacy or increasing biofilm resilience. At the same time, documented cases of collateral sensitivity suggest that carefully chosen combinations or rotations of EOs with conventional antimicrobials may be an opportunity to resensitize bacteria and reduce the likelihood of stable tolerance. Developing food preservation strategies that reduce unwanted cross-adaptation while taking advantage of collateral sensitivity will require a deeper understanding of EO interactions and adaptive trade-offs.

Looking ahead, further studies should focus on three major areas. First, a mechanistic understanding of EO adaptation to explain regulatory networks and stress-signaling pathways associated with different types of damage. Second, adaptive responses must be validated under realistic conditions, in complex food matrices, and with variable parameters, rather than relying solely on in vitro observations. Third, standardizing protocols and risk assessment frameworks is essential to harmonize findings across studies evaluating the longevity and reversibility of adaptation, and to ensure the safe incorporation of EOs into food preservation. Overcoming these limitations will improve conceptual understanding and support the safe, responsible use of EOs in many sectors, notably the food sector.

Author contributions

Conceptualization: Nadia Ziyate and Abdelhakim Bouyahya. Methodology: Abdelhakim Bouyahya, Tarik Aanniz, and Rim Rais. Data curation: Rim Rais, Hiba Mahir, and Zineb Soubai. Software: Rim Rais and Oumayma Aguerd. Writing—original draft: Rim Rais. Investigation: Tarik Aanniz, Waleed Al Abdulmonem, Mohammed Alorini, Learn-Han Lee, and Abdelhakim Bouyahya. Funding acquisition: Waleed Al Abdulmonem and Mohammed Alorini. Writing—review & editing: Waleed Al Abdulmonem, Mohammed Alorini, Learn-Han Lee, and Abdelhakim Bouyahya. Supervision: Nadia Ziyate and Abdelhakim Bouyahya. Format Analysis: Abdelhakim Bouyahya and Rim Rais. Validation: Tarik Aanniz, Waleed Al Abdulmonem, Mohammed Alorini, and Learn-Han Lee. Project Administration: Abdelhakim Bouyahya.

Funding

The Researchers would like to thank the Deanship of Graduate Studies and Scientific Research at Qassim University (www.qu.edu.sa) for financial support (QU-APC-2026).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.