Marine derived bacteria from restinga ecosystems as a platform for antibiofilm agents and next-generation postbiotics and probiotics (NGPPs)

Marine derived bacteria are a rich source of bioactive compounds that have the potential to be used in the control of infectious diseases and as Next-Generation Postbiotics and Probiotics (NGPPs). However, their antibiofilm activity against major human pathogens remains understudied. This study isolated 211 bacterial strains from Restinga ecosystems along the Southern Coast of Rio Grande do Sul, Brazil and screened them for postbiotics and probiotic traits, such as safety, stress tolerance, and antibiotic susceptibility. A subset of the 65 isolates was used for cultivation of 65 supernatants in mono- and 65 supernatants in co-culture conditions in different media to evaluate antibacterial and antibiofilm activities. About 5%–7% of the supernatants inhibited biofilm formation by at least 50% against Pseudomonas aeruginosa and Staphylococcus aureus. Co-culture enhanced the antibacterial activity against P. aeruginosa. Extraction methods strongly influenced bioactivity outcomes. Conventional extraction yielded antibiofilm activity (40%–80%) against S. aureus, though this was often accompanied by increased planktonic growth. Low or no activity was observed against P. aeruginosa. Notably, solid-phase extraction (SPE) enabled the recovery of highly active fractions. Among the isolates, strain 6, identified as Pseudomonas sputi MaRe06, exhibited the most promising profile by combining strong antibiofilm activity with the most apolar fraction (F100) against S. aureus (80% inhibition at 300 μg/mL) and robust tolerance to in vitro gastrointestinal conditions. These findings highlight the Restinga Ecosystems as a valuable source of bioactive metabolites and support the development of novel antibiofilm agents and NGPPs from marine bacteria.

Infectious diseases have always been a major public health problem, in which approximately 20% of deaths worldwide are related (Martens and Demain, 2017). Bacterial resistance has limited our present-day drug library of clinically suitable antibiotics (Igbinosa et al., 2020). Several mechanisms promote the inefficacy of existing antibiotics on bacteria (Wigati et al., 2024), including intrinsic resistance and their ability to acquire antibiotic resistance genes (Uddin et al., 2021; Ciofu and Tolker-Nielsen, 2019). Bacteria also develop adaptive resistance to protect cell viability (Fernández et al., 2011; Miller et al., 2022). Adaptive structures are used by microorganisms as a defensive shield to generate a favorable environment and to show resilience towards hostile conditions (Sharma et al., 2023), being the main defense-type applied by numerous microbial species to communicate with others and interact with their environment (Sauer et al., 2022; Rather et al., 2021a). These protective covers are generally known as biofilms. They are complex communities of microorganisms that are attached to a surface and aggregated together by a self-produced polymer matrix, which is primarily composed of polysaccharides, secreted proteins, and extracellular DNA (Muhammad et al., 2020). Consequently, due to their existence almost everywhere is considered a safe home to microorganisms against available drugs (Rather et al., 2021b). In chronic infections, the National Institutes of Health (NIH) reports about 80% of diseases are related with microbial biofilms (Rather et al., 2021a; Wang et al., 2021). The biofilm lifestyle was reported for the first time in the second half of the 20th century by J. W. Costerton (Costerton et al., 1978), and since then, biofilms have been involved in an increasing number of diseases (Hall-Stoodley and Stoodley, 2009; Costerton et al., 1999) to contribute to the host pathogenesis (Vestby et al., 2020). Such as, Cystic Fibrosis (CF) patients (Høiby, 2017), Chronic Wound (CW) infections (Haesler et al., 2019) and Chronic Rhinosinusitis (CRS) patients (Dlugaszewska et al., 2016). It is now understood that about 40%–80% of bacterial cells on earth can form biofilms (Flemming and Wuertz, 2019), and both Gram-positive and Gram-negative bacteria develop biofilms. The most frequently found biofilm forming bacteria are Staphylococcus aureus, Staphylococcus epidermidis, Escherichia coli and Pseudomonas aeruginosa (Muhammad et al., 2020). It is estimated that about two-thirds of indwelling devices related infections are caused by the St aphylococcal species, being S. aureus biofilm more pathogenic than other bacterial species (Maina et al., 2018). Given the significant impact of biofilm-associated infections on human health, considerable research has focused on the development of novel antimicrobial agents capable of preventing, disrupting, or eradicating bacterial biofilms (Theuretzbacher et al., 2019).

The International Scientific Association of Probiotics and Prebiotics (ISAPP) established the concept for postbiotics as a “preparation of inanimate microorganisms and/or their components that confers a health benefit on the host” (Salminen et al., 2021). Building on this concept, advances in microbiome research have broadened the development of novel microbial-based interventions, leading to the emergence of Next-Generation Postbiotics and Probiotics (NGPPs). NGPPs refer to an emerging class of microbiome-derived therapeutics that includes both novel microbial strains and their functional metabolites, with the potential to provide more targeted health benefits. At present, members of the genera Lactobacillus, Bifidobacterium, Enterococcus, Lactococcus, and Streptococcus are the most commonly used probiotics (Le Morvan De Sequeira et al., 2022), all of which are Gram-positive bacteria. These microorganisms have been historically preferred as probiotics because of their predominantly non-pathogenic nature and their long-established beneficial interactions with the host, in contrast to many Gram-negative bacteria (Sarita et al., 2025). However, the growing interest in NGPPs has expanded the search beyond conventional probiotic taxa, encouraging the exploration of non-traditional microbial candidates and their bioactive products.

NGPPs have gained considerable attention due to their potential advantages in infectious diseases (Newberry et al., 1959; Malagón-Rojas et al., 2020). They have emerged as promising tools for controlling clinically relevant pathogens such as P. aeruginosa and S. aureus, beyond their traditional role in gut health. It has been shown that several Lactobacillus strains interfere with biofilm formation in P. aeruginosa (Alexandre et al., 2014; Drumond et al., 2023) and produce bacteriocin-like compounds that inhibit the growth of S. aureus (Bedir et al., 2025). Highlighting the role of secondary metabolites production in these strains in probiotic mediated antimicrobial effects (Al-Qudah, 2023). Furthermore, non-traditional probiotic candidates, including Bacillus megaterium, have been reported to produce bioactive cyclic peptides with inhibitory activity against both P. aeruginosa and S. aureus (Al-Thubiani et al., 2018). Although the direct role of postbiotics and probiotics in preventing drug-resistant infections remains unclear, these approaches are being investigated as adjunctive therapies or partial alternatives to antibiotics for managing bacterial infections. For the future, more studies are needed to investigate the influence of NGPPs on antibiotic use (Jubeh et al., 2020).

Natural products have played a central role in drug discovery and continue to shape modern medicine (Xie et al., 2019). It is estimated that there are approximately two million species, including plants, animals, fungi, and microorganisms. However, despite significant advances, more than 95% of the world’s biodiversity remains unexplored (Mishra and Tiwari, 2011). Oceans cover around 70% of the earth’s surface and are a significant source of biodiversity (Barzkar et al., 2019; Lahiri et al., 2023). Marine ecosystems have fostered organisms with remarkable genetic and metabolic diversity through long-term evolution, leading to specific adaptive mechanisms to protect themselves (Hanif et al., 2019; Mostafa et al., 2022), and acquiring the ability to produce diversified compounds with unique chemical composition and biological properties (Barzkar et al., 2019; Bertrand and Munoz-Garay, 2019). Microscopic organisms collected from marine-derived resources have been found to produce biologically active chemicals, including antibiofilm properties (Lahiri et al., 2023).

Marine-derived bacteria (Wibowo et al., 2023) are a prolific source for novel bioactive compounds to treat infectious diseases (Carroll et al., 2020) and marine probiotics (Vázquez et al., 2020; Gutiérrez Falcón et al., 2021). Despite significant advances in the field and the development of Next-Generation Postbiotics and Probiotics with activity against pathogens affecting clinically relevant bacteria, there are still few, but emerging, investigations addressing the antibiofilm activity of marine-derived probiotic bacteria or their bioactive products against major human pathogens such as P. aeruginosa and S. aureus (Negi et al., 2025). These findings highlight marine probiotics as a promising and sustainable reservoir of novel antimicrobial and antibiofilm agents, underscoring the need for further investigation targeting clinically relevant bacteria.

In this same context, to further enhance the diversity of bioactive compounds derived from marine bacterial sources, co-cultivation strategies were employed. Several studies have demonstrated the effectiveness of co-culture as an approach to increase secondary metabolite production and chemical diversity (Aulia et al., 2023). Competitive interactions occurring in mixed fermentation systems can induce the activation of previously silent biosynthetic pathways. This strategy has been explored to maximize bacterial chemical diversity from known metabolites, promote the production of previously undetected compounds, and/or enhance antimicrobial activity. Accordingly, co-culturing with heat-killed cells from human pathogens has been proposed as an effective strategy to induce the production of antimicrobial secondary metabolites (Sung et al., 2017), highlighting the importance of co-culture approaches in the activation of silent genes involved in natural product biosynthesis (Shakeri Moghaddam et al., 2021).

The aim of this study was to search for and identify promising marine bacterial candidates from the Restinga ecosystem along the Southern Coast of Rio Grande do Sul, Brazil, with potential application as Next-Generation Postbiotics and Probiotics (NGPPs). Candidate evaluation was based on an integrated screening strategy that included probiotic-related properties, virulence-associated phenotypic traits, tolerance to gastrointestinal-like stress conditions, antibiotic response, and antimicrobial and antibiofilm activity against clinically relevant pathogens. For this purpose, marine bacterial isolates were subjected to preliminary assays to assess safety-related traits, tolerance to low pH and inhibitory gastrointestinal compounds, and survival under simulated gastrointestinal transit conditions. In parallel, selected isolates were evaluated for their capacity to produce bioactive metabolites in mono- and co-culture systems using human pathogens as biological targets. By combining probiotic-related screening with bioactivity assays, this study provides a preliminary framework for identifying marine bacteria from Restinga ecosystems with potential relevance as non-conventional NGPP candidates and as sources of antimicrobial or antibiofilm metabolites.

Coastal marine sand material: Source and bacteria isolation

Coastal marine sand samples were collected and processed as previously described in (Back et al., 2026). Briefly, three subsamples were collected at each Restinga site at 10-m intervals using a sterile shovel and falcon tubes from an area measuring approximately 10 × 25 cm at a depth of 10–15 cm below the surface. To better represent local variability, the three subsamples were pooled to create a composite sample, which was then transported to the laboratory under refrigeration until bacterial isolation. The interval between collection and processing was kept to a maximum of 72 h, with the samples maintained under refrigeration throughout. Sampling was conducted in March 2024 in four Restinga ecosystems located along the south coast of Rio Grande do Sul, Brazil, representing areas with different levels of anthropogenic impact. Sites with lower human influence (-I) included Parque Nacional da Lagoa do Peixe (LP-I) and Cidreira (C-I), while sites with higher human impact (+I) included Balneário Mostardense (BM + I) and Imbé (I + I). The exact dates were as follows: 21 March 2024, for BM + I and LP-I, and 22 March 2024, for C-I and I + I.

Bacterial growth from environmental samples was promoted using multiple culture media under laboratory conditions. Six media were employed: Vegitone Actinomyces Broth (MilliporeSigma, Germany); Difco™ Marine Broth 2,216 (BD/Interlab, United States of America); R2A broth (Biolog, United States of America); Artificial Seawater medium (ASW) (Patil, 2013) supplemented with 2% malt extract (Kasvi, Brazil); YEPD (Combes et al., 2022) medium composed of yeast extract broth (1%; BD/Interlab, United States of America), peptone broth (2%; Himedia, India), and potato dextrose broth (2%; Kasvi, Brazil); and potato dextrose broth (Kasvi, Brazil). Media lacking agar or prepared in diluted form were supplemented with agar to a final concentration of 20 g/L.

For bacterial isolation, the sand sample was first homogenized by vortexing for 10 s. Subsequently, 1 g of the homogenized sand was transferred to 10 mL of sterile distilled water and vortexed again for 10 s. After allowing the sand particles to settle by precipitation, 200 μL of the supernatant was collected and spread onto solid media using a sterile T-shaped cell spreader (Runlab, China). Plates were incubated under laboratory conditions at 28 °C and under cold-room (5 °C) conditions for periods ranging from 1 to 30 days. Colonies displaying distinct phenotypic characteristics were repeatedly subcultured until pure isolates were obtained.

Bacterial strains

A total of 211 isolates were selected based on distinct phenotypic characteristics, including colony morphology, pigmentation, size, and Gram staining, in order to maximize the diversity of the culture collection and avoid redundant isolates with similar profiles. Isolates were preserved in glycerol stocks (25%, v/v), stored at −20 °C for short-term and at −80 °C for long-term storage. Pathogenic bacterial strains used as negative controls (C-) in biological assays were selected to represent clinically relevant, well-characterized biofilm-forming pathogens. These included P. aeruginosa ATCC 27853 (non-multidrug-resistant) and S. aureus ATCC 6538 (non-methicillin-resistant). Strains were stored in skim milk broth (Neogen, United States of America) supplemented with 25% glycerol (v/v) (GE Healthcare, United States of America) at −80 °C. Heat-killed P. aeruginosa PAO1 was used for secondary metabolite production during co-culture experiments. PAO1 was heat-killed by incubating the overnight culture in a boiling water bath (100 °C) for 30 min. For probiotic assays, Lactobacillus rhamnosus ATCC 7469 was employed as a positive control (C+), serving as a well-characterized reference probiotic strain commonly used for benchmarking probiotic potential. This strain was stored under the same conditions as pathogenic controls.

Production of secondary metabolites and evaluation of antibacterial and antibiofilm activities against Pseudomonas aeruginosa ATCC 27853 and Staphylococcus aureus ATCC 6538

A subset of 65 environmental bacterial isolates obtained at 28 °C was selected for secondary metabolite production assays based on their ability to produce visible and abundant colony growth within 24 h under the established incubation conditions. This criterion was used to ensure sufficient biomass and reproducible growth for downstream fermentation, metabolite production, and functional antibacterial and antibiofilm assays. Although this approach may favor fast-growing and easily culturable bacteria and potentially exclude slow-growing isolates with biological relevance, it was considered necessary for the feasibility and reproducibility of the experimental workflow.

Secondary metabolites production: Mono- and co-culture

For inoculum preparation, colonies obtained from fresh cultures agar broth were suspended in sterile 0.9% (w/v) sodium chloride (NaCl) solution, and adjusted according to McFarland’s 0.5 standard, corresponding to approximately 1.5 × 10 8 colony forming units per mL (CFU/mL) and an optical density of OD600 nm = 0.150 (Clinical Laboratory Standards Institute CLSI, 2012; Rosa et al., 2023). Liquid culture media were selected according to the isolation medium originally used for each environmental bacterium. For monoculture fermentation, 1 mL of the standardized bacterial suspension was inoculated into 50 mL of the corresponding liquid medium in a 250 mL baffled Erlenmeyer flask and incubated at 25 °C with agitation at 150 rpm for 5 days (Veiga et al., 2025). Co-culture experiments were performed using a heat-killed bacterial pathogen, following the methodology described by (Shakeri Moghaddam et al., 2021), with minor modification. Heat-killed PAO1 was intended as an exploratory strategy to evaluate whether exposure to pathogen-derived cellular material could modulate secondary metabolite production by environmental isolates. This approach was not designed to reproduce a live microbial interaction. Instead, it was used as a standardized stimulus that could mimic the presence of pathogen-associated cellular components while avoiding overgrowth, direct competition, or interference caused by viable PAO1 cells during fermentation. Briefly, a loopful of pathogenic bacterial biomass from an overnight culture on Mueller Hinton (MH) agar was inoculated into 20 mL of MH broth in a 100 mL Erlenmeyer flask and incubated at 37 °C for 24 h. The culture was subsequently heat-inactivated by immersion in boiling water for 30 min. Inactivation was confirmed by a plate viability assay, in which 10 μL of the heat-killed suspension was spread onto MH agar and incubated at 37 °C for 24 h. Absence of colony growth confirmed loss of viability. Minor modification: The resulting heat-killed P. aeruginosa PAO1 suspension was adjusted to an OD600 nm of 0.580. The adjusted heat-killed bacteria was added to the fermentation medium at 5% (v/v) simultaneously with the inoculation of the environmental bacterial strain. After the cultivation period, the culture supernatant was chosen as the study matrix rather than the bacterial pellet. The bacterial culture was centrifuged at 9.000 rpm for 30 min at 4 °C to remove bacterial cells and residual culture medium (Veiga et al., 2025). Then, supernatants were filter-sterilized using 0.22 µm pore size membrane (Minisart Sartorius Co., Göttingen, Germany) and directly used for preliminary evaluation of antibacterial and antibiofilm activities. Environmental isolates were coded according to their original position in the bacterial collection and were grouped based on the isolation medium as follows: Actinomyces medium (strain codes 1–11; Figures 1A, 2A), YEPD (12–18; Figures 1A, 2A), Marine agar (19–56; Figures 1A–D, 2A–D), R2A (57–59; Figures 1D, 2D), and ASW (80–96; Figures 1E, 2E). The strain codes correspond to identifiers from the original culture collection and are not consecutive for all isolates. In total, 65 environmental isolates were selected and evaluated under mono- and co-culture conditions, resulting in 130 supernatants.

Preliminary evaluation of antibacterial and antibiofilm activities of cell-free supernatants

Pseudomonas aeruginosa ATCC 27853 (Pa)/ S. aureus ATCC 6538 (Sa) was grown overnight on MH at 37 °C. Assays were performed in sterile 96-well plates. Core assay: 40 μL of culture medium was added to each well (tryptic soy broth for Pa and brain heart infusion broth for Sa), followed by the addition of 80 μL of supernatants to be tested and 80 μL of bacterial suspension (standardized for absorbance of 0.150 at OD600 nm, 108 CFU/mL) in sterile NaCl 0.9% to complete 200 μL. Controls: negative control (C-, bacterial growth), 80 μL bacterial suspension, 80 μL ultra pure water and 40 μL culture medium were used; for sterilized control, 40 μL culture medium, 80 μL ultra pure water, and 80 μL NaCl 0.9% were used; and for positive control (C+, antibiotic control), the supernatants were replaced by 80 μL of 20 μg/mL antibiotic; for sterilized control, 40 μL culture medium, 80 μL ultra pure water, and 80 μL NaCl 0.9% were also used to evaluate the sterility of the experiment. Finally, a blank was used to validate the activity of the secondary metabolites without the influence of the growth medium; using 80 μL NaCl 0.9%, 80 μL culture medium used with the environmental bacteria and 40 μL ultra pure water. Then, the 96-well plate was then incubated statically at 37 °C for 24 h. Inhibition of bacterial growth was verified by reading the optical density at 600 nm at 0 h and 24 h. To verify the antibiofilm activity, the contents of the wells were removed and washed three times with 200 μL NaCl 0.9%; the adhered biofilm was fixed for 1 h at 60 °C, and stained with 200 μL of crystal violet (CV, 0.4%) for 15 min at room temperature. After washing, the remaining CV was solubilized with 99% ethanol for 30 min, and the absorbance was measured at 570 nm. Crystal violet assay quantifies total attached biofilm biomass and does not distinguish between viable bacterial cells, dead cells, and extracellular matrix components. C- was considered to represent 100% of planktonic growth and biofilm formation; therefore, values higher than 100% could suggest a stimulation of bacterial growth or biofilm formation in comparison to the control. All experiments in the preliminary screening were performed once with two independent biological replicates (N = 1; n = 2). Supernatants selected for chemical extraction were subsequently validated in an additional independent experiment (N = 1; n = 2) to confirm the reproducibility of the preliminary results. Results are presented as mean ± standard deviation. In total, 130 supernatants of secondary metabolites (65 supernatants in mono- and 65 supernatants in co-culture) were obtained and screened against both opportunistic pathogens. This preliminary assay was intended solely for the initial screening of active supernatants. Therefore, no statistical analysis was performed at this stage. Supernatants were selected based on their ability to inhibit either bacterial growth or biofilm formation by more than 50%. The 50% threshold was established as a conservative criterion to prioritize samples with substantial inhibitory activity for further analyses. However, supernatants that exhibited activity in one assay while resulting in values higher than the negative control in the other were excluded. Thus, only supernatants displaying consistent and promising antibacterial and/or antibiofilm activities were selected for subsequent analyses.

Metabolites extraction

Based on the outcomes of the preliminary assays, nine supernatants were selected from a total of 130 supernatants for further experiments due to their biological activity against P. aeruginosa and S. aureus.

Liquid-liquid extraction (LLE)

Initially, liquid-liquid extraction (LLE) was performed, taking into account key solvent characteristics such as toxicity, volatility, and polarity (Grosso et al., 2015). Therefore, a solvent with intermediate polarity was selected to facilitate the recovery of a broad spectrum of secondary metabolites (Cardoso et al., 2025). Ethyl acetate is widely recognized as a standard solvent for the extraction of natural products, largely due to its relatively low toxicity and suitable physicochemical properties (Schwarz et al., 2021). Then, ethyl acetate (99,5%, NEON, Brazil) was used as the extraction solvent. Supernatants (500 mL) were extracted three times with ethyl acetate (1,500 mL), corresponding to an organic-to-aqueous phase ratio of 3:1 (v/v). The organic phases were pooled and concentrated under reduced pressure at 45 °C using a rotary evaporator. After solvent removal, 1 mL of water was added to the bottom of the flask to resuspend the resulting extract. Then, the samples were evaporated to dryness at 32 °C to obtain crude dry extracts. On average, 3.23% of the extract was recovered from each supernatant.

Solid-phase extraction (SPE)

SPE is an attractive alternative to conventional extraction methods because not only reduces extraction time but also lowers solvent consumption and operational costs (Cardoso et al., 2025). It has been implemented by several studies for the recovery of secondary metabolites from marine-derived bacteria (Magot et al., 2023). Therefore, this technique was selected to process the bioactive supernatants directly, without additional pre-treatment steps. 1 mL (an average of 22.03 mg) of supernatants were loaded onto a Sep-Pak C18 Plus Short Cartridge (360 mg sorbent per cartridge, 55–105 µm particle size). To obtain fractions covering a broad polarity range from highly polar to non-polar compounds, a stepwise methanol/water elution gradient was applied: F05 (5% MeOH), F20 (20% MeOH), F40 (40% MeOH), F60 (60% MeOH), F80 (80% MeOH), and F100 (100% MeOH). Since water is the most polar component of the mixture, the expected elution order was: highly polar compounds > relatively polar compounds > intermediate polarity metabolites > semipolar metabolites > low-polarity compounds > non-polar metabolites (those with the lowest polarity present in the supernatants). Six fractions of 1 mL each were collected and evaporated to dryness at 32 °C, yielding dried extracts for subsequent analyses. On average, these dried fractions represented 18% of the total fraction mass recovered from the supernatant. The extracts and fractions were stored at 8 °C for 24 h before the assay. The extraction and elution procedures were performed in independent experiments under identical conditions, yielding consistent results across three independent experimental replicates (N = 3), with each experiment analyzed in technical triplicate (n = 3). The injected solvent volume resulted in approximately 10 mL of solvent per supernatant (N = 3, n = 3).

Biological assays for extracts: Inhibition of bacterial growth and biofilm formation

Pseudomonas aeruginosa (Pa) ATCC 27853 and S. aureus (Sa) were grown overnight on Mueller Hinton agar (MH) at 37 °C. Selected dried extracts from metabolite extraction obtained from mono- and co-culture experiments were evaluated for their antibacterial and antibiofilm activities using sterile 96-well microplates (Costar 3,599, Corning Inc., United States of America). Core assay: 40 μL culture medium (Tryptic Soy Broth for Pa/Brain Heart Infusion for Sa) were added to each well, followed by the addition of 80 μL of fraction to be tested in a concentration range (20–300 μg/mL) and 80 μL of bacterial suspension (standardized for absorbance of 0.150 at OD600 nm, 10 8 CFU/mL) in sterile NaCl 0.9% to complete 200 μL. Controls: to complete 200 μL - for negative control (C-, bacterial growth), 80 μL bacterial suspension, 80 μL ultra pure water and 40 μL culture medium were used; for positive control (C+, antibiotic control), the chemical samples were replaced by 80 μL of 20 μg/mL antibiotic; for sterilized control, 40 μL culture medium, 80 μL ultra pure water, and 80 μL NaCl 0.9% were also used to evaluate the sterility of the experiment; for vehicle control, 80 μL bacterial suspension, 2% dimethyl sulfoxide (76 μL ultra pure water +4 μL DMSO 100%) and 40 μL culture medium were used. Then, the 96-well plate was then incubated statically at 37 °C for 24 h. Inhibition of bacterial growth was verified by reading the optical density at 600 nm at 0 h and 24 h. To verify the antibiofilm activity after 24 h, the contents of the wells were removed and washed three times with 200 μL NaCl 0.9%; the adhered biofilm was fixed for 1 h at 60 °C, and stained with 200 μL of crystal violet (CV, 0.4%) for 15 min at room temperature. After washing, the remaining CV was solubilized with 99% ethanol for 30 min, and the absorbance was measured at 570 nm. Crystal violet assay quantifies total attached biofilm biomass and does not distinguish between viable bacterial cells, dead cells, and extracellular matrix components. All experiments were performed in triplicate, with three independent biological replicates (N = 3; n = 3). Gentamicin sulfate (Sigma-Aldrich, Italy) and Vancomycin (Sigma-Aldrich, United States of America) were used as positive control (C+) against P. aeruginosa ATCC 27853 and S. aureus ATCC 6538, respectively. Concentrations described in this section refer to the concentration within the well upon completion of 200 μL, stock solutions were prepared in 50% DMSO (dimethyl sulfoxide, Sigma-Aldrich, United States of America), the chemical samples within each well contain a concentration less than 2% DMSO.

Statistical analysis

This section presents the biological data obtained from the assays performed with the extracts and fractions generated in this study. Data are reported as the mean ± standard deviation of three independent biological replicates (N = 3), each analyzed in triplicate as technical replicates (n = 3). Statistical significance was evaluated using one-way analysis of variance (ANOVA), followed by Dunnett’s multiple comparisons test with the negative control (C-) as the reference group. Differences were considered statistically significant when the probability of the null hypothesis was below 0.1% (p < 0.001). All statistical analyses and graphical representations were performed in Python using the SciPy and Statsmodels libraries.

Next-generation postbiotics and probiotics (NGPPs) screening

All environmental bacterial isolates were evaluated as potential NGPPs. As a preliminary screening, bacterial growth at 37 °C for 48 h was assessed by streaking each isolate onto its corresponding specific agar medium. Colony phenotypic characteristics, including color and mucous (exopolysaccharide) production, were visually evaluated. Only isolates capable of growing under these conditions were selected for further analyses (Won et al., 2020).

Growth under anaerobic and microaerophilic conditions

Difco™ sodium thioglycollate medium (BD/Interlab, United States of America) was used to evaluate obligate aerobe (A), microaerophilic (M), and facultative anaerobe (AF) growth (Zimbro and Sparks, 2009). Five milliliters of liquid medium were dispensed into sterile 10 mL glass tubes with caps. Bacterial isolates were first grown overnight at 37 °C on their respective agar media and subsequently inoculated into the thioglycollate medium using a sterile inoculating needle, stabbing the medium to near the bottom of the tube. Cultures were incubated at 37 °C for up to 7 days, and bacterial growth was monitored daily. Only strains capable of growing under microaerophilic and facultative anaerobic conditions were selected for subsequent virulence factor assays. The experiment was performed using three independent biological replicates (N = 3), each analyzed in triplicate as technical replicates (n = 3).

Absence of virulence factors

Gelatinase production

Gelatinase is a type of proteolytic enzyme that when it is expressed in strains intended for human application, gelatinase activity may become harmful. This enzyme can participate in the degradation of host extracellular matrix components and mucosal barriers, facilitating tissue invasion and bacterial dissemination (Salamone et al., 2019). Gelatinase activity was evaluated using a gelatin agar medium composed of peptone (4 g/L, Kasvi, Brazil), yeast extract (1 g/L, BD/Interlab, United States of America), Difco™ gelatin (12 g/L; BD/Interlab, United States of America), and agar (15 g/L). Bacterial isolates were spot-inoculated onto the medium and incubated at 35 °C for 24 h. Gelatinase production was considered positive (A+) when a clear halo was observed around the colony, indicating gelatin hydrolysis. On the contrary, the absence of a transparent halo surrounding the colony was interpreted as a negative result (B-) (Domingos-Lopes et al., 2017). Duplicates were prepared for each isolate under study.

Hemolytic activity

Hemolysis is linked to the production of enzymes capable of degrading phospholipids and disrupting red blood cell (RBC) membranes (Tran et al., 2024). It was assessed using blood agar plates (Yadav et al., 2016) prepared with blood agar base (Kasvi, Brazil) supplemented with 5% (v/v) defibrinated sheep blood (NEWPROV, Brazil; Reg. MS no. 10287910056). Bacterial isolates were spot-inoculated onto the plates and incubated at 35 °C. Hemolysis was evaluated after 24 and 48 h of incubation. Blood agar plates were examined for signs of β-hemolysis (clear zones around colonies), α-hemolysis (green-hued zones around colonies) or γ-hemolysis (no zones around colonies). Duplicates were prepared for each isolate under study. α-Hemolysis results from the oxidation of hemoglobin into green-colored methemoglobin due to bacterial hydrogen peroxide production, whereas β-hemolysis corresponds to complete RBC lysis. In contrast, γ-hemolysis indicates the absence of hemolytic activity and no visible change in the surrounding medium (Pradhan et al., 2020).

Antimicrobial activity testing: Minimum inhibitory concentration (MIC) assay

Antimicrobial activity was determined using the broth microdilution method in 96-well microplates, following the guidelines of the Clinical and Laboratory Standards Institute (“Methods for DIlution Antimicrobial Susceptibility). Briefly, individual colonies grown on agar plates were suspended in sterile saline solution and adjusted to an optical density of 0.150 at 600 nm, corresponding to approximately 1.5 × 10 8 CFU/mL. The bacterial suspension was subsequently diluted 1:1,000 in the appropriate culture medium for each strain. Aliquots of 100 μL of the diluted bacterial suspension were dispensed into each well of a sterile 96-well microplate. Two-fold serial dilutions of the tested antibiotics: gentamicin sulfate (Sigma-Aldrich, Italy), vancomycin (Sigma-Aldrich, United States of America), tetracycline (Sigma-Aldrich, United States of America), chloramphenicol (Sigma-Aldrich, United States of America), and ampicillin (Sigma-Aldrich, United States of America), were prepared in the same medium, covering a concentration range of 0.5–256 μg/mL. Plates were incubated aerobically at 37 °C for 24 h. The minimum inhibitory concentration (MIC) was defined as the lowest antibiotic concentration that completely inhibited visible bacterial growth, as assessed by visual inspection and confirmed by spectrophotometric measurement at 600 nm. Sterility controls (medium only), growth controls (bacteria without antibiotics), and antibiotic controls were included in all assays. The experiment was performed using three independent biological replicates (N = 3), each analyzed in triplicate as technical replicates (n = 3). The selected antibiotic panel included compounds with activity against Gram-positive, Gram-negative, and broad-spectrum bacterial groups. Since the isolates were not taxonomically identified at the species level, the MIC values were not interpreted using clinical breakpoints. Thus, the results are presented as MIC values and growth inhibition patterns under the tested conditions, rather than as formal susceptibility or resistance categories.

Preliminary screening of bacterial tolerance to inhibitory conditions

Microbial tolerance to acidic and alkaline conditions, bile components, and phenol was first evaluated using sterile 96-well microplates. The tested conditions included different pH values (2, 3, 4, 5, 6, 7, and 8), bile tolerance was evaluated using porcine bile extract at 0.3%, which contains a mixture of bile salts and other bile components (B8631-100G, Sigma-Aldrich, United States of America), and phenol (Yadav et al., 2016) at 0.4% (Merck, United States of America). The pH of the culture media was adjusted using 1 M HCl or 3 M NaOH, as required. Methodology proposed by De Caravlho et al. (De Carvalho Maquiné et al., 2024) was used with adjustments. In the present study, a broader range of pH conditions was evaluated, a lower bile salt concentration was used, and the assay was performed directly in 96-well microplates. In addition, the culture medium was adjusted according to the specific growth requirements of each bacterial strain. Bile salts are particularly notable for their role as biological detergents (Begley et al., 2006). Phenolic compounds, on the other hand, are generated in the gut through microbial metabolism of aromatic amino acids and are known to be toxic to potential probiotics (Panda et al., 2017). In both scenarios, the loss of membrane integrity and metabolic function represents a common outcome when bacterial strains lack the ability to withstand these stress conditions. Bacterial suspensions were prepared in sterile NaCl 0.9% and adjusted to an optical density between 0.09 and 0.11 at OD600 nm. For each condition, 100 µL of the bacterial suspension was mixed with 100 µL of the corresponding culture medium in each well, resulting in a final volume of 200 µL. Plates were incubated at 37 °C for 24 h. Bacterial growth was monitored by measuring the optical density at 600 nm at 0 h and 24 h, and tolerance was expressed as the difference in absorbance (ΔOD600 nm) between these time points. Appropriate sterility and growth controls were included. The experiment was performed using three independent biological replicates (N = 3), each analyzed in triplicate as technical replicates (n = 3).

Gastrointestinal transit tolerance assay

Only environmental bacterial isolates that successfully passed the preliminary screening assays, including growth under anaerobic conditions, absence of virulence-related factors, antibiotic susceptibility and tolerance to inhibitory substances, were selected for this experiment. Selection was further supported by the combined results of the antibacterial and antibiofilm assays together with the evaluation of probiotic-related traits. Subsequently, the viability of strains 6, 12, 15, 48, and 90 was assessed following exposure to a dynamic in vitro model designed to simulate gastric and gastrointestinal stress conditions (Fernández De Palencia et al., 2008; Peres et al., 2014). Briefly, bacterial cultures were grown in MH broth at 37 °C with agitation (180 rpm) for 48 h. Cells were harvested by centrifugation (10,000 rpm, 10 min, 25 °C), washed twice with sterile 0.9% (w/v) NaCl, and resuspended in an electrolyte solution (6.2 g/L NaCl, 2.2 g/L KCl, 0.22 g/L CaCl 2, 1.2 g/L NaHCO 3; pH adjusted to 6.2 with 1 M NaHCO 3). The final cell suspension volume was adjusted to 35 mL. An aliquot was withdrawn (A1). To simulate in vivo saliva conditions, 5 mL of a sterile electrolyte solution containing lysozyme (final concentration of 0.01% (w/v), L6876, Sigma-Aldrich) was added to 35 mL of cell suspension, and an aliquot was taken without further incubation (A2). To simulate the gastric environment, 3 mL of electrolyte solution added with 0.3% (w/v) pepsin (P7000-25G, Sigma-Aldrich) was incorporated into the cell suspension. To mimic normal gastric emptying, aliquots of the suspension were collected after sequential incubations of 20 min at 37 °C and 50 rpm at each pH step; pH was gradually adjusted to 4.0 (A3), 3.0 (A4), and 2.0 (A5) using 1 M HCl. To simulate the intestinal environment, samples A3-A5 were adjusted to pH 6.5 with 1 M NaHCO 3, mixed with 4 mL of a sterile electrolyte solution, added with 0.45% (w/v) porcine bile extract (B8631-100G, Sigma-Aldrich, United States of America) and 0.1% (w/v) pancreatin (P3292-25G, Sigma-Aldrich). Then, samples were incubated at 37 °C and 50 rpm for 180 min and fractions of suspensions were collected (Ai3, Ai4, Ai5). Following incubation, serial dilutions for each aliquots were prepared and plated on MH agar for colony forming unit (CFU) enumeration. Bacterial survival throughout the simulated TGI conditions was assessed by viable counts (CFU/mL) to provide a more robust assessment. The experiment was performed using two independent biological replicates (N = 2), each analyzed in duplicate as technical replicates (n = 2).

Bacterial identification

The most promising isolate, strain 6, selected based on the combined results of both biological assays (Sections 2.3 and 2.4), was subjected to whole-genome sequencing. Genomic DNA was extracted using the Qiagen DNeasy Plant Kit®, following the manufacturer’s instructions. This kit was selected because according to plant DNA extraction protocols are designed to efficiently disrupt resistant biological matrices and remove complex chemical inhibitors that may interfere with DNA recovery and downstream analyses. Therefore, its extraction chemistry was considered suitable for obtaining high-quality microbial genomic DNA from the evaluated bacterial isolates. The concentration and purity of the extracted DNA were subsequently assessed using a NanoDrop spectrophotometer prior to downstream genomic analyses, confirming that the DNA obtained was suitable for subsequent procedures. Whole-genome shotgun sequencing was performed on the Illumina MiSeq platform. Raw FASTQ reads were filtered using Scythe v0.994 (Buffalo, 2014) and compared against the NCBI RefSeq genome database using RefSeq Masher v0.1.2 (Ondov et al., 2016). De novo assembly was performed using the Unicycler pipeline (Wick et al., 2017), and assembly quality was assessed using CheckM2 v1.0.2 (Chklovski et al., 2023) and QUAST v5.0.3 (Gurevich et al., 2013). Strain identification was performed using AutoMLST2 (Pourmohsenin et al., 2025), which automatically selected conserved single-copy marker genes for multilocus sequence analysis (MLSA). The concatenated alignment generated by the pipeline was used for phylogenetic inference. The maximum-likelihood phylogenetic tree was constructed using IQ-TREE v2.4.0 (Minh et al., 2020), employing ModelFinder to determine the best-fit substitution model for the alignment (Kalyaanamoorthy et al., 2017). Branch support was assessed using 1,000 ultrafast bootstrap replicates (Hoang et al., 2018). Taxonomic assignment was inferred based on phylogenetic clustering with closely related reference and type strains. The Whole Genome Shotgun project has been deposited in the DDBJ/ENA/GenBank database under the accession number GCF_056897695.1.

In addition, isolates that subsequently showed activity in both biological assays (Sections 2.3 and 2.4), strain 12 and 15 were identified by sequencing the 16 S rRNA gene. Genomic DNA was extracted using a bacterial genomic DNA extraction kit (code 100196; QuatroG Biotecnologia, Brazil). In addition, the concentration and purity of the extracted DNA were subsequently verified using a NanoDrop spectrophotometer before analyses. This quality-control step confirmed that the DNA obtained was suitable for subsequent procedures. The 16 S rRNA gene was amplified by polymerase chain reaction (PCR) using the universal primers 27F and 1492 R. The resulting amplicon sequences were edited and quality-checked using Sequencer software. The curated sequences were then compared with sequences available in the Basic Local Alignment Search Tool (BLAST) database (Macedo et al., 2005). This result is presented in the supplementary data.

Statistical analysis

Statistical significance was assessed using one-way analysis of variance (ANOVA), followed by Tukey’s post hoc multiple comparisons test. This statistical approach was applied to the preliminary inhibitory-substance assays evaluating tolerance to different pH conditions and to the results obtained from the in vitro gastrointestinal transit assay. Due to the amount of data generated in these assays, comparisons were performed among the evaluated experimental conditions. Differences were considered statistically significant when the probability of the null hypothesis was below 5% (p < 0.05). All statistical analyses and graphical representations were performed using Python.

This analysis was not applied to the remaining probiotic-related assays because some of them did not include reference controls for statistical comparison, such as the bile salt and phenol tolerance assays. In addition, other initial assays were qualitative in nature, while assays such as MIC determination did not require this type of statistical analysis.

Secondary metabolites from restinga ecosystem and their evaluation as antibacterial and antibiofilm agents against Pseudomonas aeruginosa ATCC 27853 and Staphylococcus aureus ATCC 6538

In total, 130 supernatants were obtained and tested against opportunistic pathogens (Figures 1, 2). Figure 1 summarizes the antibacterial and antibiofilm activities of all evaluated supernatants against P. aeruginosa. Overall, supernatants 5M, 15C, 31M, 39M, and 42 M exhibited the strongest antibiofilm effects, with inhibition values exceeding 50%. Notably, supernatant 12C, cultivated in YEPD broth, demonstrated antibacterial growth inhibition above 50% regardless of the cultivation strategy employed, whereas the remaining samples showed moderate or no significant activity. These results are particularly relevant considering that the identification of supernatants that inhibit biofilm formation without affecting planktonic viability could contribute as anti-virulence strategies which are promising alternatives to conventional antibiotics (Fleitas Martínez et al., 2019). On the other hand, 90.76% stimulated biofilm formation and 55.08% promoted both bacterial growth and biofilm formation (>100%). The widespread stimulation of biofilm formation and bacterial growth observed for most supernatants suggests that secondary metabolites produced may modulate phenotypic responses in P. aeruginosa. Such interactions are consistent with ecological scenarios in which microbial metabolites act as signaling molecules (Townsley and Shank, 2017), enhancing biofilm development rather than employing inhibitory effects.

Figure 2 also summarizes the antibacterial and antibiofilm activities of the evaluated supernatants against S. aureus. Overall, supernatants 4M, 5M, 6M, 40M, 42C, 50M, 50C, 91M, and 91C exhibited strong antibiofilm activity, inhibiting biofilm formation by approximately 75%. Among these, 91 M was the only supernatant that also inhibited bacterial growth by approximately 50%. Additionally, 20.77% of the evaluated supernatants displayed antibiofilm activity despite simultaneously stimulating bacterial growth (>100%). Furthermore, 27.69% of the samples strongly promoted biofilm formation (>100%), while approximately 80% of all evaluated supernatants stimulated bacterial proliferation.

Similar to the observations for P. aeruginosa, this stimulatory effect was primarily associated with fermentations conducted in ASW culture medium. In contrast, a greater proportion of supernatants exhibited antibiofilm activity against S. aureus. This observation may be associated with structural differences between Gram-positive and Gram-negative bacteria that can influence susceptibility to extracellular bioactive compounds (Rohde, 2023). Nevertheless, the frequent stimulation of planktonic growth observed for S. aureus indicates that biofilm inhibition does not necessarily correlate with reduced bacterial viability. In general, as observed in Figures 1, 2, some supernatants obtained under co-culture conditions showed responses similar to those obtained under mono-culture conditions. This suggests that, for those isolates, the cellular components potentially released by heat-killed PAO1 did not substantially influence metabolite production or biological activity. However, since heat-killed cells may release intracellular components capable of influencing bacterial metabolism, the biological activity observed specifically or more strongly under co-culture conditions in other cases suggests that exposure to pathogen-derived material may have influenced the production or release of bioactive metabolites.

Nine supernatants were selected for further experiments: six active against P. aeruginosa (Figure 1) and three active against S. aureus (Figure 2). Selection was based on their biological activity profiles and their ability to produce visible, abundant colony growth within 24 h under the established incubation conditions. Given the urgent need to address pathogenic Gram-negative bacteria, a greater number of supernatants were selected against P. aeruginosa than against the Gram-positive pathogen S. aureus.

Evaluation of the biological activity of extracts obtained by LLE

To obtain secondary metabolite extracts, the selected supernatants were subjected to liquid-liquid extraction (LLE) using ethyl acetate. The resulting organic extracts were evaluated against the corresponding target strains: 5M, 15C, 31M, 39M, 42M, and 12C against P. aeruginosa (Figure 3), and 4M, 5M, and 6 M against S. aureus (Figure 4), according to their previously observed activity.

The extracts evaluated against P. aeruginosa did not fully reproduce the biological effects previously observed with the crude supernatants. However, some extracts showed mild antibiofilm activity, mainly at the highest tested concentrations. Extracts 5 M (300–150 μg/mL; Figure 3A) and 31 M (300–200 μg/mL; Figure 3D) inhibited biofilm formation by approximately 20%–30%. Similar moderate effects were observed for 15C at 150 μg/mL (Figure 3C) and 42 M at 200–150 μg/mL (Figure 3F). Interestingly, extract 39 M showed antibiofilm activity only at the lowest tested concentration, 20 μg/mL, reaching 22.26% inhibition (Figure 3E).

In contrast, extract 12C, derived from the only supernatant that previously showed antibacterial activity under the co-culture strategy, did not retain this effect after ethyl acetate extraction (Figure 3B). This loss of activity may be explained by several factors. First, the active metabolites responsible for the antibacterial effect may not have been efficiently extracted with ethyl acetate, suggesting that they could be less polar. Second, the compounds may have been unstable during the extraction or solvent evaporation process. Another possibility is that the antibacterial activity observed in the supernatant depended on the combined action of multiple metabolites, and this interaction was disrupted during extraction. Therefore, the absence of activity in the ethyl acetate extract does not necessarily rule out the presence of antibacterial compounds in supernatant 12C, but indicates that the extraction conditions used were not suitable to recover or preserve the active components.

Regarding the extracts evaluated against S. aureus (Figure 4). Extracts 4M and 6 M exhibited consistent antibiofilm activity across all tested concentrations. Biofilm inhibition ranged from 40% to 60% for 4 M (Figure 4B) and 30%–60% for 6 M (Figure 4C). Overall, the performance of these extracts was superior compared to the results obtained against P. aeruginosa. At the highest tested concentration (300 μg/mL), biofilm inhibition reached 60% (4 M), 80% (5 M), and 68.5% (6 M). However, it is important to note that, to extract 5 M (Figure 4A), the highest concentrations tested (300–150 μg/mL), which showed strong antibiofilm activity (70%–80%), also appeared to stimulate planktonic cell growth. This observation raises concerns regarding the potential therapeutic application of these secondary metabolites. Although they inhibit biofilm formation, their simultaneous stimulation of planktonic bacterial growth indicates that their effect may not be exclusively inhibitory. This response could have important implications for their use as antibiofilm agents. Further studies are therefore required to clarify the biological significance of this effect.

Although ethyl acetate has been reported as an effective solvent for the recovery of secondary metabolites from marine-derived bacteria, with minimal degradation during extraction (Dhayalan et al., 2025), and has been associated with strong antibiofilm and antibacterial activities, our results indicate that this approach may not fully account for the complexity of the bioactive compounds involved. Alternative processing techniques are required to improve secondary metabolites recovery.

Evaluation of the biological activity of fractions obtained by SPE

The selected supernatants were fractionated by solid-phase extraction (SPE) using water/MeOH mixtures of different polarities. The resulting fractions were then evaluated against their corresponding target strains based on their previously observed activity: 5M, 15C, 31M, 39M, 42M, and 12C against P. aeruginosa (Figure 5), and 4M, 5M, and 6 M against S. aureus (Figure 6).

In contrast to the results obtained using liquid-liquid extraction (LLE), solid-phase extraction methodology does not show antibiofilm activity against P. aeruginosa (Figure 5) when evaluating the different fractions. These findings underscore the complexity involved in isolating specific metabolites capable of inhibiting virulence-associated traits in Gram-negative pathogenic bacteria (Fleitas Martínez et al., 2019). In particular, the selective recovery of active compounds appeared to be challenging under the extraction conditions used in this study. Considering the relative polarity of the solvents employed for LLE and SPE (water > methanol > ethyl acetate), moderate antibiofilm activity was observed only in the ethyl acetate extracts. In contrast, as solvent polarity increased, the biological activity progressively decreased and became comparable to that of the negative control (C-). These results may be related to differences in compound solubility, stability, or recovery efficiency during the extraction and fractionation processes. However, the chemical nature of the active metabolites cannot be established from these results alone, and further studies are required to validate any interpretation regarding their polarity or molecular characteristics.

In contrast, when evaluating supernatants with antibiofilm activity against S. aureus, distinct active fractions were identified using SPE. For supernatant 5 M (Figure 6B), planktonic bacterial growth was no longer stimulated, unlike what was observed with LLE. Notably, fractions F40 (intermediate polarity metabolites) and F100 (low-polarity metabolites) inhibited biofilm formation at 300 μg/mL by 71.40% and 70.40%, respectively. These values are comparable to those obtained using LLE methodology (80% inhibition at 300 μg/mL). However, fractions obtained from 4 M (Figure 6A) did not exhibit selective activity across the tested polarity gradient. In contrast, the extract obtained by LLE displayed biological activity at all tested concentrations. This result suggests that the absence of detectable activity in the individual fractions, may indicate that the bioactivity depends on the combined presence of multiple metabolites.

On the other hand, extract 6 M (Figure 6C) exhibited significant antibiofilm activity in fractions F80 (33%) and F100 (76%). Importantly, fraction F100 (300 μg/mL) demonstrated biological activity comparable to the positive control (vancomycin 20 μg/mL, 82% biofilm inhibition). This represents the most significant biological outcome of the study, indicating that strain six produces secondary metabolites with selective antibiofilm activity against S. aureus when an optimized SPE-based extraction methodology is employed.

It is important to point out that the antibiofilm effects reported throughout these sections may be associated with different biological processes, including inhibition of microbial attachment, interference with biofilm formation, or disruption of preformed biofilms. However, the crystal violet assay quantifies total attached biomass and does not distinguish between viable bacterial cells, dead cells, and extracellular matrix components. Therefore, this methodology does not allow us to determine the specific mechanism responsible for the observed reduction in biofilm biomass. In this study, crystal violet staining was used as a preliminary screening approach to identify promising candidates with potential antibiofilm activity. Consequently, the results should be interpreted as general changes in biofilm biomass rather than as direct evidence of a specific mechanism of action. Further complementary assays, including viability-based methods, microscopy, matrix-specific analyses, and biofilm dispersal assays, will be necessary to clarify the mechanisms involved. Additionally, this study has one limitation: biofilm formation values were not formally normalized to planktonic growth. Therefore, reductions in biofilm biomass cannot be fully interpreted as effects specific to biofilms, as they may be partially associated with changes in overall bacterial growth. This is particularly relevant in cases where variations in biofilm biomass were accompanied by an increase in planktonic growth.

Screening of next-generation postbiotics and probiotics (NGPPs)

From the 211 marine bacterial isolates obtained from the Restinga ecosystem in RS, Brazil, 16 strains fulfilled the preliminary selection criteria and were considered promising NGPP candidates (Supplementary Figure S1). These isolates represented a reduced subset of strains with favorable characteristics for further safety and functional evaluation. Importantly, none of the sixteen selected isolates showed detectable hemolytic activity on blood agar plates or gelatinase production. However, the absence of these virulence-related phenotypes does not confirm the safety of the strains. Rather, these findings should be interpreted as preliminary phenotypic traits related to safety that require further confirmation through more comprehensive genomic and risk assessments. Therefore, these non-hemolytic isolates were selected for subsequent antimicrobial activity assay.

The response of the sixteen selected strains to five antibiotics was evaluated using the broth microdilution method (Table 1; Figure 7). Therefore, Table 1 summarizes the MIC endpoints, whereas Figure 7 shows the OD600-based heatmap used as a complementary visualization of the growth inhibition patterns across the tested antibiotic concentration ranges. Overall, the isolates showed a heterogeneous response to the antibiotics tested (Table 1). This variability was also evident in the OD600 heatmap (Figure 7), and not only among antibiotics with different mechanisms of action, but also among isolates with the same Gram-staining profile. Among the antibiotics tested, vancomycin showed the clearest Gram-associated pattern. High MIC values were observed mainly among Gram-negative isolates, with strains 3, 12, 13, 15, 24, 30, 51 and 90 frequently showing MIC values >256 μg/mL. In contrast, several Gram-positive isolates, including C+, 48, 188, 200, and 211, showed markedly lower MIC values, frequently ≤1.0 μg/mL; unexpectedly, this response was also observed with the Gram-negative strain 6. This distribution is consistent with the known intrinsic reduced susceptibility of Gram-negative bacteria to vancomycin, which is largely explained by the presence of an outer membrane that limits the access of this large glycopeptide molecule to its peptidoglycan target (Huddleston, 2014; Luqman et al., 2024). Thus, in this dataset, the elevated vancomycin MICs were not randomly distributed across the strain panel, but were mainly concentrated in isolates with a Gram-negative profile.

However, the patterns observed for the remaining antibiotics were less clearly associated with Gram staining and appeared to be more strain-specific. Ampicillin showed a wide range of MIC values, with low MICs in strains 184 and 188 (<0.5 μg/mL), as well as strains 13 and 200 (4.0 μg/mL), while high MICs ≥256 μg/mL were detected in C+, 51, 90, 186, and 211. This heterogeneous response indicates that beta-lactam inhibition was not explained solely by Gram type in this isolated collection. Similarly, tetracycline displayed high MIC values in strains 30, 51, 90, 186, and 188, whereas C+ and strain six showed low MIC values ≤1.0 μg/mL. Gentamicin also showed a broad response range, from <0.5 μg/mL in strains 182 and 186 to ≥256 μg/mL in strains 90, 184, and 188. Chloramphenicol followed the same trend, with low MIC values in C+, 6, 13, 48, and 51, but high MIC values ≥256 μg/mL in strains 30, 90, 184, 186, 200, and 211.

Together, these results suggest that, except for the more evident Gram-associated response to vancomycin, the antibiotic response profiles were mainly isolate-specific. This is particularly relevant for environmental isolates intended as potential Next-Generation Probiotics, because elevated MIC values may reflect either intrinsic traits or acquired resistance mechanisms. Without species-level identification and genomic screening, it is not possible to determine whether the observed high MICs are associated with non-transferable intrinsic resistance (Huddleston, 2014). From a safety perspective, the phenotypic data obtained here provide an initial screening of antibiotic response, but they are not sufficient on their own to confirm the absence of antimicrobial resistance risk.

Interestingly, Gram-negative strains 6, 12, and 15 had previously shown, in biological assays, strong potential for the production of secondary metabolites with antibiofilm and antibacterial activity against clinically relevant pathogens such as P. aeruginosa and S. aureus. These findings suggest that, despite the safety concerns commonly associated with Gram-negative bacteria, these strains may represent relevant candidates for further characterization along with the Gram-positive candidates, particularly to determine whether their bioactive potential can be safely exploited within the framework of Next-Generation Postbiotics and Probiotics (NGPPs).

Preliminary screening of the selected strains for tolerance to inhibitory conditions, particularly in response to different pH levels (2–8), revealed clear strain-specific variability (Figure 8). Overall, strong inhibition of bacterial growth was observed under highly acidic conditions (pH 2–3), with the exception of strains 6, 90, 186, and 188, which maintained measurable growth under these conditions. Notably, strain 48 displayed a survival percentage at pH three comparable to that observed at pH 4–6 (p < 0.05). From pH 5 onwards, a marked recovery in growth was evident across most strains. Optimal growth was generally observed between pH six and 8, particularly for strains 15, 48, and 200, which exhibited high growth percentages (>100%), indicating strong adaptability to neutral and slightly alkaline conditions. In contrast, strains 6, 90, and 186 maintained relatively stable growth across a broad pH range, suggesting greater physiological robustness.

The consistent recovery and enhanced growth observed at pH ≥ 5 indicate that the evaluated strains are well adapted to conditions resembling the intestinal environment (Mazanko et al., 2025). Their robust growth at near-neutral pH further supports active metabolism and suggests potential functionality within the host. In general, most non-extremophilic bacteria are capable of growing across a relatively broad pH range (5.5–9.0), while maintaining a tightly regulated cytoplasmic pH between 7.4–7.8 (Booth, 1985), which is essential for cellular homeostasis.

In contrast, the overall sensitivity observed at pH two to three is expected, as these conditions mimic the highly acidic environment of the human stomach. Exposure to such low pH can lead to intracellular acidification, promoting molecular damage, including increased loss of purines and pyrimidines from DNA (Cotter and Hill, 2003). Despite this, some strains (6, 90, 186, 188, and 200) exhibited partial tolerance, suggesting a potential capacity to withstand gastric transit. This tolerance may be attributed to adaptive mechanisms that allow bacteria to regulate their internal pH relative to the external environment, either through cytoplasmic acidification or alkalinization (Lund et al., 2020). More broadly, bacterial survival under pH stress involves coordinated changes in cell structure, metabolism, and transport systems, enabling maintenance of cytoplasmic pH homeostasis under adverse conditions (Krulwich et al., 2011).

Importantly, the pH tolerance results contribute to the selection of NGPP candidates by incorporating physiological response as an additional criterion alongside the previously reported probiotic-related properties, as well as antimicrobial and antibiofilm activity. Among the strains previously identified as promising, strain six appears particularly relevant, as it combines antibiofilm and antibacterial potential with measurable growth under highly acidic conditions and relatively stable growth across the evaluated pH range. This suggests that strain 6 may have a greater capacity to tolerate gastrointestinal-like stress conditions compared with other bioactive candidates.

Although strains 12 and 15 were previously associated with bioactive metabolite production, their limited tolerance to highly acidic conditions suggests that their potential as NGPP candidates should be interpreted with caution. Further evaluation is required to determine their ability to survive gastric transit, their antimicrobial safety profile, and whether their application would be more suitable as postbiotic preparations rather than as probiotic candidates. Additionally, strains 48 and 90 showed relevant tolerance patterns under specific pH conditions, indicating that they may also represent candidates for further characterization. Overall, the pH tolerance assay helps refine the candidate selection process by identifying strains with greater physiological adaptability under gastrointestinal-like stress conditions.

In the same context, the survival of the strains under inhibitory compounds such as bile salts and phenol was evaluated (Figure 9). Strains 6, 12, 15, and 48 were able to grow under both conditions, demonstrating tolerance to these stressors. In contrast, the control (C+) and strains 186 and 211 showed no detectable growth when exposed to 0.4% phenol. Although isolates 188 and 200 exhibited comparable viability under acidic conditions, they displayed differential responses to bile salts, which may be attributed to the expression of resistance-associated proteins in strains and species (Zhang et al., 2022). Strain 90, despite exhibiting a low growth percentage, maintained detectable viability under both inhibitory conditions, indicating minimal but persistent tolerance. It is important to note that the absence of growth at these concentrations does not necessarily imply a complete lack of resistance, as lower concentrations may still support bacterial survival. In this study, relatively high concentrations were deliberately selected to evaluate the tolerance limits of each strain.

The resistance exhibited by strains 6, 12, 15, and 48 to both bile salts and phenolic compounds suggests that they are capable of surviving passage through the upper gastrointestinal tract, particularly the duodenum, and potentially reaching the colon. Therefore, these strains were selected for further in vitro gastrointestinal transit assay. To simulate physiological conditions of the gastrointestinal tract, an in vitro model was employed consisting of three sequential phases. First, an oral phase was mimicked using a saliva model supplemented with lysozyme (0.01% w/v). This was followed by a gastric phase containing pepsin (0.3% w/v), and finally an intestinal phase consisting of porcine bile extract (0.45% w/v) and pancreatin (0.1% w/v) (Peres et al., 2014).

Among the evaluated strains, distinct survival patterns were observed. The control strain, Lactobacillus rhamnosus (C+, Figure 10A), exhibited an early reduction in viable counts during the saliva phase (A2) compared to the initial inoculum (A1). Upon exposure to gastric conditions (pH 4), a significant decrease in growth was observed; however, the strain maintained relatively stable viability throughout the gastric (A3-A5) and intestinal phases (Ai3-Ai5). A similar trend was observed for strain 48 (Figure 10E), although this strain displayed higher overall growth compared to the control. Despite this, it followed a comparable pattern of decline and stabilization across both gastric and intestinal conditions. In contrast, strains 12 (Figure 10C) and 15 (Figure 10D) showed highly similar responses, characterized by pronounced sensitivity to acidic conditions. As previously observed (Figures 8C,D), these strains were unable to grow at low pH (2–3), which explains their small growth under gastric conditions. However, upon transition to the intestinal phase (pH 6.5), growth was partially restored. Notably, no significant differences were detected between late gastric (A4-A5) and intestinal phases (Ai3-Ai5), despite this recovery. Strain 6 (Figure 10B) exhibited relatively stable growth throughout the saliva and early gastric phases (A1-A3). A marked decline was observed once gastric conditions reached pH 2, after which viability remained consistently through the remainder of the gastric and intestinal phases (A5-Ai5). Finally, strain 90 (Figure 10F) maintained consistent growth across saliva, gastric, and intestinal conditions, with the exception of the final intestinal stage (Ai5), where a significant reduction was observed compared to the initial inoculum (A1). As previously noted, this strain demonstrated notable resistance under stress conditions, including low pH and high antibiotic concentrations (≥256 μg/mL), when compared to the control strain.

Intrinsic resistance to gastric acidity is a relatively uncommon probiotic trait typically associated with well-established probiotic species, such as Lactobacillus acidophilus and Bifidobacterium spp. This characteristic is particularly relevant because it reduces dependence on protective food matrices during gastric transit, favoring the delivery of viable cells to the small intestine (Cotter and Hill, 2003). It is important to note that the gastrointestinal transit assay was performed as an initial in vitro functional evaluation. While viable cell counts offer a more direct evaluation of bacterial survival than optical density measurements, the limited replication of this assay limits the statistical robustness of the results. Therefore, the observed tolerance patterns should be interpreted cautiously and should not be considered definitive evidence of gastrointestinal survival or probiotic efficacy. Additional assays with increased biological replication, together with genomic characterization, safety assessment, and eventually, in vivo models are required before proposing any strain as a safe NGPP candidate.

Overall, strain six emerged as the most robust candidate when the antibacterial and antibiofilm results were considered together with the evaluation of probiotic-related traits. Its ability to tolerate a wide range of pH conditions and inhibitory compounds suggests a physiologically relevant stress-response profile. These properties may increase the likelihood that this strain can withstand gastrointestinal-like conditions and reach target sites where beneficial effects could occur (Lipilkina et al., 2024). However, these findings should be interpreted as preliminary evidence supporting its prioritization for further characterization rather than as confirmation of probiotic functionality.

Bacterial identification

Strain 6 (from I-I), identified as the most robust candidate in all the assays, was isolated from coastal marine sand and classified as Pseudomonas sputi strain (designated MaRe06) based on whole-genome sequence analysis. This species has been recently described following a taxonomic reclassification within the Pseudomonas putida and Pseudomonas fluorescens groups (Tohya et al., 2023).

Evaluation of MaRe06 as promising NGPP candidate

The results showed that MaRe06 exhibited relevant functional traits, supporting its preliminary consideration as a non-conventional microbial candidate. Pseudomonas sputi demonstrated tolerance to gastrointestinal stress conditions, such as acidic pH, bile salts, and phenolic compounds. This response indicates physiological robustness under conditions that partially mimic gastrointestinal transit, an important criterion for selecting probiotic candidates. MaRe06’s ability to tolerate these stressors suggests that it may retain viability and functional potential under host-associated conditions.

Additionally, MaRe06 exhibited bioactivity associated with antibiofilm activity against S. aureus. For this study, it is relevant for NGPP development because candidate strains are evaluated not only for their ability to survive gastrointestinal conditions and their safety profile, but also for their ability to produce functional metabolites with possible antimicrobial applications. Previous studies have reported that marine Pseudomonas strains can produce secondary metabolites with antibacterial or antibiofilm activity (Bhatnagar and Kim, 2010; Vijay et al., 2021), supporting the exploration of MaRe06 as a possible source of bioactive compounds.

However, identifying MaRe06 as P. sputi requires careful consideration of safety concerns. While some Pseudomonas strains have been studied for beneficial applications, the genus includes opportunistic pathogens, and some strains carry clinically relevant traits (De Lorenzo et al., 2024; Yoshino et al., 2011). To further evaluate the suitability of MaRe06 as an NGPP candidate, we assessed its antibiotic response using the MIC values reported in Table 1. These values were interpreted according to the CLSI MIC breakpoints established for non- Enterobacterales, including Pseudomonas spp., as well as other non-fastidious, glucose-nonfermenting, Gram-negative bacilli (Clinical and Laboratory Standards Institute CLSI, 2020). Based on these criteria, MaRe06 was classified as susceptible to tetracycline (0.5 μg/mL) and chloramphenicol (4.0 μg/mL) and, resistant to gentamicin (64 μg/mL). Despite its ability to tolerate gastrointestinal-related stress conditions and the absence of hemolytic activity and gelatinase production, the resistant phenotype represents a critical safety concern, as it could limit its suitability as candidate probiotic due to the potential risk of horizontal gene transfer (Peirano et al., 2025), then it demands a robust genomic safety assessments.

Interestingly, MaRe06 exhibited a low vancomycin MIC value (1.0 μg/mL), indicating growth inhibition by this glycopeptide antibiotic despite its classification as a Gram-negative bacterium. This is noteworthy because vancomycin typically shows limited activity against Gram-negative bacteria. However, a recent study reported that vancomycin can inhibit the growth of P. aeruginosa at low concentrations under low-nutrient conditions (Chan et al., 2023). Therefore, the response observed in MaRe06 may reflect a strain-specific physiological interesting feature that requires additional studies to elucidate the mechanisms involved in this unusual response.

When the MIC values of MaRe06 were compared with the MIC interpretation used for L. rhamnosus C+, based on epidemiological cutoff (ECOFF) values reported in previous studies (Danielsen and Wind, 2003; Ammor et al., 2007), the positive control showed resistance to ampicillin, with an MIC value >256 μg/mL, while it was considered susceptible to the remaining antibiotics evaluated. In contrast, MaRe06 showed a markedly lower MIC value for ampicillin. Although this comparison cannot be interpreted as formal susceptibility due to the absence of a specific breakpoint for ampicillin in P. spp., it represents an interesting antibiotic response pattern worth further investigation, particularly considering the intrinsic reduced susceptibility of Gram-negative bacteria to β-lactams.

The unexpected vancomycin response observed in MaRe06 highlights the importance of exploring the genetic and physiological mechanisms involved in stress response and antimicrobial susceptibility in non-conventional bacterial probiotic candidates. Although Pseudomonas spp. Are not widely accepted as probiotics for human use, mainly because the genus includes both beneficial strains and opportunistic pathogens, some species have been increasingly explored as platforms for metabolic engineering due to their metabolic versatility, genetic accessibility, and stress tolerance (De Lorenzo et al., 2024; Kampers et al., 2019). In addition, emerging evidence suggests potential probiotic-related applications of some Pseudomonas strains, particularly in aquaculture, where they have been associated with improved growth performance, immune response, and overall health in species such as Litopenaeus vannamei (Sutanti et al., 2025).

At this stage, MaRe06 should be regarded as a promising non-conventional probiotic candidate requiring further validation rather than as a confirmed probiotic strain. Its bioactive potential and tolerance to gastrointestinal-like stress conditions support its relevance for further characterization; however, comprehensive genomic and safety assessments, including the evaluation of antimicrobial resistance genes and virulence factors are required before considering its possible application as a NGPP candidate.

Methodological limitations and scope of interpretation

The present study should be interpreted within the scope of its design as a preliminary screening platform to prioritize marine-derived bacterial isolates with potential antibiofilm activity and NGPP-related traits. While the experimental workflow permitted evaluation of a large collection of environmental isolates, some methodological limitations must be considered when interpreting the results.

First, an initial screening of supernatants was performed with a limited number of biological replicates. This step was intended only as a preliminary selection. This approach prioritized bioactive candidates from a large number of samples under mono- and co-culture conditions. However, it does not provide the same robustness as fully replicated confirmatory assays. To address this limitation, we subsequently validated selected supernatants in additional experiments. However, the limited replication of the preliminary screening may have influenced the selection of candidates, so the screening results should be regarded as exploratory rather than definitive.

Second, the values for biofilm formation were not formally normalized with respect to planktonic growth. Therefore, reductions in crystal violet staining cannot be interpreted as exclusively biofilm-specific inhibition, as changes in attached biomass may also result from changes in overall bacterial growth. This limitation is particularly relevant in cases where antibiofilm activity was accompanied by increased planktonic growth. Therefore, the observed effects should be interpreted as changes in total attached biomass under the tested conditions, rather than as conclusive evidence of selective antibiofilm activity.

Third, the crystal violet assay has inherent limitations, as all in vitro methodology. This method quantifies total attached biomass, but it can not distinguish between viable cells, dead cells, and extracellular matrix components. Therefore, it cannot identify the specific biological mechanism responsible for reducing biofilm biomass, such as inhibiting initial attachment, interfering with biofilm development, reducing matrix production, or disrupting preformed biofilms. To define the mechanism of action of the active extracts and fractions, complementary approaches are required, including viability-based assays, microscopy, matrix-specific analyses, and biofilm dispersal assays.

Overall, these inherent limitations do not invalidate the preliminary relevance of the identified candidates, especially MaRe06 and its active SPE fractions, against S. aureus. However, they define the scope of the conclusions that can be drawn. While the results support prioritizing selected marine-derived isolates and their metabolites for further characterization, additional replicated assays, chemical identification, mechanistic studies, and comprehensive safety evaluations are required before proposing these strains or metabolites for applied NGPP or antibiofilm development.

Although the biological outcomes observed in this study were not consistently aligned with the initial expectations, the diversity of responses obtained constitutes a valuable platform for innovation and further investigation. In natural product research, the recovery of biologically active compounds and the attainment of highly potent extracts remain major challenges (Kowalski and Baj, 2025). The selection of an appropriate extraction strategy, including solvent system and methodological approach plays a decisive role in determining chemical yield and bioactivity, in addition to the intrinsic physicochemical characteristics of the biological matrix. Generally, obtaining substantial quantities of bioactive metabolites from marine organisms is more challenging when compared to terrestrial sources (Grosso et al., 2015). Marine extreme environments, such as Restinga ecosystems, may function as natural incubators of structurally unique chemical scaffolds, providing molecular architectures rarely encountered in other habitats (Ng et al., 2015). Continued exploration of these ecosystems increases the likelihood of identifying novel bioactive compounds with therapeutic potential. In an era marked by the rise of chronic diseases and the global escalation of bacterial infections, natural products remain a promising reservoir of new molecular targets and pharmacophores (Kowalski and Baj, 2025).

The combined screening strategy employed in this study identified promising bacterial candidates and emphasized the dual ecological and biotechnological significance of marine bacteria. Of the evaluated isolates, strain 6, identified as P. sputi MaRe06, was the most robust candidate. This strain exhibited strong antibiofilm activity against S. aureus, which is likely associated with the production of secondary metabolites obtained through the extraction approach used in this study. Additionally, MaRe06 demonstrated tolerance to gastrointestinal stress conditions, such as acidic pH, bile salts, phenolic compounds, and digestive enzymes. This supports its physiological robustness under gastrointestinal transit conditions. These traits suggest that MaRe06 is a promising non-conventional NGPP candidate.

Nevertheless, its antibiotic response profile, particularly its resistance to gentamicin and intrinsic response to ampicillin, underscores the importance of cautious interpretation. MaRe06 showed promising functional and stress-tolerance traits, but its potential application as a probiotic cannot be confirmed without further safety validation. Therefore, a comprehensive genomic and risk assessment is required before considering its safe application as a next-generation postbiotic or probiotic candidate, particularly since some strains belong to genera that may include opportunistic species. This assessment should include, at a minimum, screening for antimicrobial resistance genes, virulence factors, and other safety markers. Currently, MaRe06 should be considered a promising non-conventional candidate for further characterization rather than a confirmed probiotic strain.