Preliminary in vitro antibacterial activity of Globularia trichosantha root extract and associated changes in tet(A) and floR transcript abundance in a single Vibrio anguillarum isolate

Antibiotic resistance among aquaculture pathogens has stimulated interest in plant-derived antibacterial agents. This preliminary in vitro study examined the antibacterial activity of a single-batch methanolic root extract of Globularia trichosantha against a single Vibrio anguillarum isolate from diseased rainbow trout and described the relative transcript abundance of tet(A) and floR after extract exposure. The derivatized extract was profiled by gas chromatography–mass spectrometry (GC-MS); peak assignments are tentative because authentic reference standards were not used. Disk diffusion (three replicate plates per treatment) used extract-loaded blank disks (2 mg extract per disk), florfenicol (30 µg per disk) and enrofloxacin (5 µg per disk); in combined treatments the extract disk and the antibiotic disk were separate disks on the same plate. Mean zone diameters (including the 6 mm disk) were 17.0 mm for the extract and 12.5 mm for enrofloxacin, whereas florfenicol alone and both combined treatments showed no measurable inhibition beyond the 6 mm antibiotic disk. These results are presented descriptively. Minimum inhibitory and bactericidal concentrations (MIC/MBC), checkerboard/fractional inhibitory concentration index (FICI) and time–kill assays were not performed, so the combined-disk observations are not interpreted as evidence of an interaction between the extract and the antibiotics. Reverse-transcription quantitative PCR (RT-qPCR; 2 −ΔΔCt method (Ct, threshold cycle), 16S rRNA as the only reference gene, primer efficiencies not determined) was performed on three parallel replicate cultures per group derived from the single isolate. Relative to the vehicle control, tet(A) and floR showed markedly lower relative transcript abundance at ≈1,000 µg mL - ¹ extract (0.24 and 0.19), intermediate values at ≈500 µg mL - ¹ (0.58 and 0.71) and values close to the control at ≈250 µg mL - ¹ (1.11 and 1.06). No inferential test was applied, and because viability was not measured concurrently, lower transcript abundance cannot be distinguished from reduced viability, growth or RNA yield. The findings are limited to one isolate, one extract batch and the conditions tested; MIC/MBC, interaction testing, viability-controlled expression analysis and multi-isolate validation are required before any application in aquaculture can be considered.

Medicinal plants and their extracts have attracted growing interest as alternatives to synthetic drugs in aquaculture, owing to their broad range of biological activities, including antimicrobial (Erguig et al., 2015; Ispir et al., 2024a; Reverter et al., 2014), antioxidant (Gabor et al., 2012; Syahidah et al., 2015), immunostimulant (Alishahi et al., 2010; Awad et al., 2013; Ispir et al., 2024b; Vaseeharan and Thaya, 2014), appetite-enhancing (Reverter et al., 2014), digestion-promoting (Agus Putra et al., 2013; Lee and Gao, 2012), growth-promoting (Asadi et al., 2012; Ispir et al., 2024a), and survival-enhancing (Hwang et al., 2013; Ispir et al., 2024a) properties. To reduce feed costs, numerous plant species have also been evaluated as partial or complete substitutes for fishmeal protein; such applications have been reported to enhance growth performance and overall biomass production (Bhavan et al., 2012). More recently, an increasing number of countries have begun replacing chemical therapeutants with herbal remedies in aquaculture practice (Iqbal et al., 2016; Kirici et al., 2024), a shift that may also help alleviate the economic and ecological burden associated with chemotherapeutic and antibiotic use (Kirici et al., 2024; Syahidah et al., 2015).

The incorporation of plant-derived extracts as dietary or topical supplements in aquaculture is attractive because of their accessibility, comparatively mild side effects, cost efficiency, biodegradability, biocompatibility, and environmental sustainability. Plant organs—leaves, flowers, fruits, seeds, and roots—are rich sources of bioactive compounds, including flavonoids, phenolic constituents, essential oils, saponins, terpenoids, tannins, alkaloids, polypeptides, polysaccharides, fatty acids, vitamins, carbohydrates, and proteins. Some of these metabolites have been reported to act as antioxidants and immunostimulants in fish (Bulfon et al., 2015); such effects, however, need to be established separately for each plant material and host species.

Globularia trichosantha (Figure 1) is a perennial herbaceous plant of the family Plantaginaceae, characterized by a basal rosette of lanceolate-to-oblong leaves, densely hairy inflorescences arranged in globose heads, and bluish-to-violet tubular flowers. The species is distributed mainly across Anatolia and neighboring regions, typically on rocky slopes and in steppe habitats, and is well adapted to arid and semi-arid conditions. Phytochemical studies have shown that G. trichosantha and closely related species contain iridoid glycosides, sugar esters and phenolic compounds (Calis et al., 2001, 2002), and extracts of these species have exhibited antibacterial activity against various Gram-negative and Gram-positive bacteria, with biological effect varying according to extraction method and solvent. However, evidence on the activity of Globularia extracts against aquaculture pathogens, and on any associated changes in resistance-gene transcripts, remains limited.

Vibrio anguillarum is one of the most clinically and economically important bacterial pathogens in marine and brackish-water aquaculture, causing vibriosis, a hemorrhagic septicemia, in a wide range of farmed fish species (Frans et al., 2011). Although plant-derived extracts are widely investigated as antimicrobial agents, their association with resistance-gene transcripts in aquaculture pathogens remains insufficiently characterized. Here, a methanolic root extract of G. trichosantha from a single extraction batch was tested in vitro against a single V. anguillarum isolate. Antibacterial activity was assessed by disk diffusion with florfenicol and enrofloxacin disks, applied alone and together with a separate extract-loaded disk, and the relative transcript abundance of the resistance genes tet(A) and floR was estimated by reverse-transcription quantitative PCR (RT-qPCR) after extract exposure. The study was designed as a preliminary, hypothesis-generating screen: it did not determine minimum inhibitory concentration (MIC) or minimum bactericidal concentration (MBC) values, did not test antibacterial interactions by checkerboard or time–kill methods, and did not examine effects on resistance phenotypes, host responses or in vivo outcomes.

Plant material

Plant material was collected from the natural flora of Van Province, Türkiye (38°29′ N, 43°22′ E) in May 2024, during the flowering period, and was taxonomically identified as G. trichosantha Fisch. & C.A. Mey. by Prof. Dr. Ali Kandemir. A voucher specimen (herbarium no. VAN-2024-15) was deposited at the Herbarium of Erzincan Binali Yıldırım University, Erzincan, Türkiye.

Extract preparation

Solid–liquid extraction was applied to the dried root tissue of the plant (10 g ± 0.1 g) as described by Akman et al. (2024). Maceration was performed using methanol (Merck, Darmstadt, Germany) (3 × 100 mL, 3 × 12 h); methanol was selected for its efficiency in dissolving both polar and non-polar constituents (Akman et al., 2024). The extract was concentrated under vacuum on a rotary evaporator (IKA, Staufen, Germany) at 40 °C and 100 rpm. Maceration of 10 g of dried root yielded 2.0 g of dried extract, an extraction yield of 20% (w/w). All experiments used this single extraction batch, prepared from pooled, dried root material; the present results therefore characterize this batch and should not be generalized across harvests (see Section 4.1). A single stock solution (100 mg mL - ¹ in methanol) was prepared from the dried extract, sterilized by filtration (0.22 µm), stored in amber vials at −20 °C, and used for all disk-diffusion and gene-expression assays. No separate batch code was assigned. Before each assay, the stock and working solutions were vortexed and inspected for precipitation; when precipitation was observed, they were re-homogenized and used within the same day.

GC-MS analysis

For gas chromatography–mass spectrometry (GC-MS) analysis, 0.2 mL of the methanolic extract was transferred to a vial, 1 mL of n -hexane was added as the organic phase, and the mixture was vortex-mixed. Base-catalyzed esterification was initiated by adding 0.1 mL of 2 M KOH, followed by further vortexing. Particulates were removed by paper filtration, and the clear phase was transferred to a GC-MS vial; this approach was selected for its compatibility with base-catalyzed esterification for rapid fatty acid methyl ester (FAME) production (Ichihara and Fukubayashi, 2010). Analyses were performed by GC-MS with helium as the carrier gas, in pressure flow-control mode; the inlet was operated in split mode at 230 °C with a split ratio of 70:1. A TRB-5MS (Teknokroma Analitica, Barcelona, Spain) column (60.0 m × 0.25 mm i.d. × 0.25 µm film) was used at a column flow of 1.56 mL min - ¹, a linear velocity of 31.8 cm s - ¹, and a purge flow of 3.0 mL min - ¹. The oven program was as follows: 40.0 °C held for 5.00 min; ramped to 180.0 °C at 4.00 °C min - ¹ (2.00 min hold); ramped to 280.0 °C at 8.00 °C min - ¹ (5.00 min hold); and finally ramped to 320.0 °C at 15.00 °C min - ¹ (2.00 min hold), for a total run time of approximately 64.17 min. Peak assignment followed standard FAME-separation principles and the reported performance characteristics of low-bleed 5%-diphenyl–95%-dimethylpolysiloxane columns (Eder, 1995). Peak identities were not confirmed with authentic reference standards, and all assignments are therefore tentative. Because the extract was derivatized before analysis, the chromatogram reflects the esterified fraction and not the underivatized extract used in the bioassays. The GC-MS results are presented as supportive compositional information only and are not used to attribute the biological observations to individual compounds (see Section 4.1).

Bacterial isolate and disk-diffusion testing

A single V. anguillarum isolate recovered from diseased rainbow trout (Oncorhynchus mykiss) was used throughout the study. It had been obtained, with permission, from the collection of an earlier diagnostic study, and its identity as V. anguillarum had been confirmed by PCR before its use in the present study. No isolate code, serotype, virulence-marker profile or previous resistance profile is available for this isolate. All microbiological and molecular analyses were performed at Solver Ar-Ge Ltd. Şti. (Van, Türkiye). Cultures were incubated in Mueller–Hinton broth (MHB) (Merck, Darmstadt, Germany) supplemented with 1.5% NaCl for 24 h at 25 °C. A bacterial suspension equivalent to a 0.5 McFarland standard was prepared and spread onto 90 mm Mueller–Hinton agar plates.

Five treatments were tested, each on three replicate plates (N = 3 plates per treatment): extract alone, florfenicol alone, enrofloxacin alone, florfenicol + extract, and enrofloxacin + extract. Commercial florfenicol (30 µg per disk) and enrofloxacin (5 µg per disk) disks (Oxoid, Basingstoke, UK) were used. For the extract treatments, 20 µL of the 100 mg mL - ¹ stock solution was pipetted onto a sterile 6 mm blank disk, giving 2 mg of dried extract per disk. In the combined treatments the extract was not loaded onto the antibiotic disk; the extract-loaded blank disk and the commercial antibiotic disk were placed as two separate disks on the same inoculated plate. Disk contents and placement were kept identical across the three replicate plates. Plates were incubated for 24 h before zone diameters were measured. Zone diameters included the 6 mm disk; when no inhibition was visible beyond the disk, the value was recorded as 6.0 mm (no measurable inhibition beyond the 6 mm disk). In the combined treatments the recorded value refers to the zone around the antibiotic disk. Around extract-loaded disks the zone appeared brown rather than clear because of the colored extract, and the diameter of this brown zone, including the disk, was recorded as the inhibition zone. Because validated Clinical and Laboratory Standards Institute (CLSI) interpretive breakpoints have not been established for V. anguillarum with these agents, zone diameters were not classified into susceptible, intermediate, or resistant categories; testing otherwise followed the CLSI disk-diffusion framework (CLSI, 2020), and the disk-diffusion methodology followed Onalan and Cevik (2020), who evaluated phytochemical effects on Yersinia ruckeri. Mean zone diameters of the three replicate plates are reported descriptively (Section 2.6). MIC/MBC, checkerboard/fractional inhibitory concentration index (FICI) and time–kill assays were not performed; combined-disk findings are therefore reported descriptively and are not interpreted as interactions between the extract and the antibiotics.

Resistance-gene transcript analysis

Experimental design and extract exposure

Bacteria were incubated in MHB for 24 h, and four experimental groups were prepared from the resulting fresh cultures. The filter-sterilized extract stock solution described in Section 2.2 (100 mg mL - ¹ in methanol) was used for this experiment. Working solutions of 100, 50, and 25 mg mL - ¹ were prepared from the stock (undiluted, and diluted 1:2 and 1:4 with sterile MHB). Final exposure levels are expressed as approximate mass concentrations, calculated from 10 µL of working solution added to approximately 1.0 mL of broth: ≈1,000 µg mL - ¹ (100 mg mL - ¹ working solution), ≈500 µg mL - ¹ (50 mg mL - ¹) and ≈250 µg mL - ¹ (25 mg mL - ¹). A matched methanol (solvent) vehicle control was included at the same final solvent percentage as the highest extract dose. Group 1 served as the vehicle control (1 mL MHB + 10 µL bacterial suspension + matched methanol volume); Group 2 received 1 mL MHB + 10 µL bacterial suspension + 10 µL of the 100 mg mL - ¹ working solution; Group 3 received 1 mL MHB + 10 µL bacterial suspension + 10 µL of the 50 mg mL - ¹ working solution; and Group 4 received 1 mL MHB + 10 µL bacterial suspension + 10 µL of the 25 mg mL - ¹ working solution. Potential adsorption-related losses of hydrophobic compounds during filtration are acknowledged as a possible source of exposure variability. Three replicate cultures were run per group (N = 3), all set up in parallel from the same culture derived from a single colony of the one isolate; they are therefore replicate cultures of the same isolate and not independent isolates. RNA was isolated at the end of incubation and analyzed by RT-qPCR, and the 2 −ΔΔCt method with 16S rRNA normalization was applied, as described by Onalan (2019).

RNA isolation

Total RNA was isolated by acid guanidinium thiocyanate–phenol–chloroform extraction. Fresh cultures were pelleted by centrifugation (10,000× g, 5 min), and the supernatant was discarded. Pellets were kept on ice throughout processing, and RNase-free tubes and filter tips were used. Homogenization was achieved by adding 1 mL of acidic guanidinium thiocyanate–phenol–chloroform mixture per 1 mL culture pellet, with phase separation induced by 0.2 mL of chloroform. The upper aqueous phase was precipitated with 0.5 mL of isopropanol and centrifuged (12,000× g, 10 min); the resulting pellet was washed with 75% ethanol, briefly air-dried, dissolved in 30–50 µL of nuclease-free water, and treated with DNase I at room temperature to remove residual genomic DNA, according to the manufacturer’s instructions. RNA concentration was verified fluorometrically using a Qubit RNA HS Assay Kit on a Qubit 4 Fluorometer (Thermo Fisher Scientific, Waltham, MA, USA), and purity was assessed spectrophotometrically at 260/280 nm using a NanoDrop 2000 (Thermo Fisher Scientific, Waltham, MA, USA); an A260/280 ratio of 1.9–2.1 was considered acceptable, and RNA integrity was verified by 1% agarose gel electrophoresis. Representative Qubit yields ranged from approximately 15 to 64 ng µL - ¹ across preparations; samples below reliable quantification thresholds were re-isolated before proceeding.

cDNA synthesis

A total of 100 ng of DNase-treated RNA was reverse-transcribed using random hexamer primers, in the presence of an RNase inhibitor. The RNA–primer mixture was denatured at 80 °C for 5 min and cooled on ice; single-stranded cDNA was synthesized with a thermostable reverse transcriptase at 37 °C for 10 min and 50 °C for 10–15 min, followed by enzyme inactivation at 80 °C for 10 min. Synthesis used the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Waltham, MA, USA). The resulting cDNA was diluted 1:5–1:10 and stored at −20 °C in single-use aliquots.

Real-time quantitative PCR and relative transcript abundance

The tetracycline efflux gene tet(A) and the phenicol-resistance gene floR were selected as markers for two antibiotic classes that are widely used in aquaculture (tetracyclines and phenicols): tet(A) has been detected among tetracycline-resistant Gram-negative bacteria from fish farms (Miranda et al., 2003), and floR is a well-characterized florfenicol-resistance determinant (Li et al., 2015). Primer sequences were as follows: tet(A) forward 5′-GTAATTCTGAGCACTGTCGC-3′ (melting temperature (Tm) 57 °C) and reverse 5′-CTGCCTGGACAACATTGCTT-3′ (Tm 57 °C), a primer pair previously used for tet(A) detection (Sáenz et al., 2004); floR forward 5′-TATCTCCCTGTCGTTCCAG-3′ (Tm 57 °C) and reverse 5′-AGAACTCGCCGATCAATG-3′ (Tm 54 °C). Expected amplicon sizes, determined by in silico alignment of the primers to GenBank reference sequences, were 957 bp for tet(A) (X00006) and 400 bp for floR (AF118107). The 16S rRNA reference gene was amplified with the universal primers 27F (5′-AGAGTTTGATCMTGGCTCAG-3′) and 1492R (5′-TACGGYTACCTTGTTACGACTT-3′) (Lane, 1991), which cover almost the full length of the gene and give a product of approximately 1.5 kb (Frank et al., 2008). Reactions (20 µL) used SYBR Green (Solver, Van, Turkiye) chemistry on a Rotor-Gene Q 6000 real-time PCR cycler (Qiagen, Hilden, Germany); cycling comprised initial denaturation at 95 °C for 10 min, followed by 45 cycles of 95 °C for 60 s (denaturation), 57 °C for 60 s (primer annealing), and 72 °C for 60 s (extension), with a final extension at 72 °C for 4 min. Primer specificity was confirmed by melt-curve analysis (single peak per amplicon), and representative amplicons were verified by agarose gel electrophoresis. No-template (NTC) and no-reverse-transcription (no-RT) controls were included in every run. Each group comprised three replicate cultures (Section 2.5.1), each measured in technical duplicate; the duplicates were averaged within each culture before ΔCt/ΔΔCt calculation. Group 1 (vehicle control) served as the calibrator; ΔCt and ΔΔCt were calculated, and relative transcript abundance was estimated by the 2 - ΔΔCt method (Livak and Schmittgen, 2001). Standard curves were not run and primer efficiencies were not determined; the calculation therefore rests on the assumption, inherent to the 2 - ΔΔCt method, that amplification efficiencies were similar for the target and reference assays. Reporting follows the Minimum Information for Publication of Quantitative Real-Time PCR Experiments (MIQE) guidelines (Bustin et al., 2009) where possible; however, the assay does not meet several MIQE recommendations (use of a single reference gene, no standard curves or efficiency estimates, and long amplicons), and it is therefore not presented as MIQE-compliant (see Section 4.1).

Data analysis

Disk-diffusion results are reported descriptively as the mean zone diameter of the three replicate plates for each treatment (Table 1); no between-treatment significance test is reported. For RT-qPCR, Ct values of the technical duplicates were averaged within each replicate culture, and ΔCt was calculated by subtracting the Ct of the reference gene (16S rRNA) from that of each target gene, following standard qPCR normalization principles (Vandesompele et al., 2002). ΔΔCt was obtained relative to Group 1 (calibrator), and relative transcript abundance was reported as 2 - ΔΔCt. Relative transcript abundance is presented descriptively as group means (Table 2; Figure 2), without error bars and without a formal between-group significance test. Calculations were performed in R using the “pcr” and “qpcr” packages. Figure 2 was drawn in Python (matplotlib) from the group means reported in Table 2.

GC-MS compositional profile

The total ion chromatogram of the derivatized extract (Figure 3) showed peaks tentatively assigned to benzoic acid methyl ester and aromatic acetals in the early elution region, and to cinnamic acid methyl ester together with long-chain FAMEs (e.g., methyl palmitate, methyl stearate, and methyl linoleate) in the middle and late regions. These assignments are tentative, because they were not confirmed with authentic reference standards, and the methyl esters may have been formed during base-catalyzed derivatization. A peak tentatively assigned to bis(2-ethylhexyl) phthalate was interpreted as a probable laboratory plasticizer contaminant rather than a native phytochemical constituent. The GC-MS results are presented as supportive information only and were not used to attribute the antibacterial or transcript-abundance observations to individual compounds.

Disk diffusion

Mean inhibition-zone diameters (including the 6 mm disk; N = 3 replicate plates per treatment) are given in Table 1. The extract-loaded disk (2 mg) gave a mean zone of 17.0 mm and the enrofloxacin disk (5 µg) a mean zone of 12.5 mm. The florfenicol disk (30 µg) showed no measurable inhibition beyond the 6 mm disk, both when applied alone and when a separate extract disk was placed on the same plate. On plates carrying the enrofloxacin disk and a separate extract disk, no measurable inhibition beyond the 6 mm enrofloxacin disk was recorded, compared with 12.5 mm for enrofloxacin alone. On these combined plates the brown zone of the extract extended toward the antibiotic disk; interaction of the two diffusion fields, or difficulty in reading the zone margin within the brown area, may have contributed to this observation. The results are presented descriptively. Because MIC/MBC and checkerboard/FICI assays were not performed, the combined-disk observations are not interpreted as evidence of an interaction between the extract and the antibiotics.

Relative transcript abundance of tet(A) and floR

Relative transcript abundance was estimated by the 2 - ΔΔCt method after normalization to 16S rRNA, with Group 1 (vehicle control) as the calibrator (1.00) for both genes (Table 2; Figure 2). In Group 2 (≈1,000 µg mL - ¹), both genes showed markedly lower relative transcript abundance than the calibrator (tet(A) 0.24; floR 0.19). In Group 3 (≈500 µg mL - ¹), the values were intermediate (tet(A) 0.58; floR 0.71). In Group 4 (≈250 µg mL - ¹), the values were close to the calibrator level (tet(A) 1.11; floR 1.06). The two genes therefore followed a broadly similar concentration-associated pattern. These values are group means of three replicate cultures derived from the single isolate; no inferential test was applied, and differences between groups are described without statistical inference. Because bacterial viability, growth and RNA yield were not measured at the time of RNA isolation, the lower relative transcript abundance at the higher concentrations cannot be distinguished from reduced viability or growth and is not interpreted as transcriptional regulation.

The Ct values of the 16S rRNA reference gene fell within the range 16.35–16.68 across Groups 1–4. This limited variation is consistent with, but does not validate, the use of 16S rRNA as the reference gene under the conditions tested; reference-gene stability was not assessed with alternative reference genes or dedicated stability algorithms. Amplification curves of the tet(A) and floR assays recorded on the Rotor-Gene Q 6000 are shown in Figure 4.

This preliminary study combined disk diffusion with RT-qPCR-based estimates of relative transcript abundance for one V. anguillarum isolate and one batch of G. trichosantha root extract. Under the conditions tested, the extract-loaded disk (2 mg) gave a larger mean zone diameter (17.0 mm) than the enrofloxacin disk (12.5 mm), and the florfenicol disk showed no measurable inhibition beyond the 6 mm disk. These observations indicate antibacterial activity of this extract batch against this isolate in the disk-diffusion format. They do not permit a quantitative comparison of potency between the extract and the antibiotics: disk-diffusion zones depend on the diffusion of constituents through agar as well as on their activity, the extract and antibiotic disks carried substantially different amounts (2 mg versus 5–30 µg), and no MIC or MBC was determined. The absence of a zone around the florfenicol disk is compatible with the detection of floR transcripts in this isolate, but the resistance phenotype was not characterized further and no interpretive breakpoints are available for V. anguillarum with these agents.

Placing a separate extract disk on the same plate did not increase the zone around either antibiotic disk, and around the enrofloxacin disk no measurable inhibition beyond the disk was recorded in the presence of the extract disk (12.5 mm when applied alone). The present design cannot explain this observation. Overlap of the diffusion fields of the two disks, altered diffusion of the antibiotic in the presence of extract constituents, and difficulty in reading the zone margin within the brown, extract-colored area are possible contributors. Because checkerboard/FICI and time–kill assays were not performed, these observations are not interpreted as an interaction between the extract and the antibiotics, and they do not support any conclusion about the use of the extract together with these antibiotics.

The GC-MS profile of the derivatized extract contained peaks tentatively assigned to aromatic methyl esters and long-chain fatty acid methyl esters. Because these assignments were not confirmed with authentic standards and the methyl esters may partly reflect the derivatization step, the profile cannot identify the constituents responsible for the observed activity. Phenolic compounds and free fatty acids have been reported to act on bacterial membranes (Cowan, 1999; Cushnie and Lamb, 2011; Desbois and Smith, 2010), and iridoid glycosides have been described in G. trichosantha (Calis et al., 2001); whether any of these compound classes contributed to the present observations was not tested. Membrane damage, disruption of oxidative balance or any other mechanism of action was not examined in this study.

At the highest concentration tested (≈1,000 µg mL - ¹), relative transcript abundance of tet(A) and floR was markedly lower than in the vehicle control; it was intermediate at ≈500 µg mL - ¹ and close to the control at ≈250 µg mL - ¹. The Group 4 values (1.11 and 1.06) are close to the calibrator and are not interpreted as an increase. These values are descriptive estimates from three replicate cultures of one isolate, without inferential statistics, normalized to a single reference gene, and calculated without experimentally determined efficiencies for long amplicons. Most importantly, bacterial viability, growth and cell number were not measured at the time of RNA isolation. If the highest extract concentration reduced the number of viable or metabolically active cells, the relative abundance of the target transcripts could change without any specific regulatory effect on these genes, and a change in 16S rRNA content relative to mRNA would likewise affect the normalized values. The data therefore describe an exposure-associated pattern and do not demonstrate an effect on the regulation of these genes or on the resistance phenotype of the isolate. The genes were chosen because tet(A) has been detected among tetracycline-resistant bacteria from fish farms (Miranda et al., 2003) and floR is a well-characterized florfenicol-resistance determinant (Li et al., 2015).

Plant-derived products have been studied in salmonid aquaculture, including in rainbow trout (Awad et al., 2013; Bulfon et al., 2015), with administration via feed or bath (Reverter et al., 2014). No host-related, in vivo or application-related endpoint was examined in the present study, and the data do not address the use of G. trichosantha root extract in fish. Such questions would become relevant for future research only if the in vitro activity is confirmed by standardized susceptibility testing across multiple isolates.

Limitations

Several methodological aspects constrain the conclusions of this preliminary, hypothesis-generating study. (i) Scope: one V. anguillarum isolate was tested; it had been identified by PCR, but no isolate code, serotype, virulence-marker profile or previous resistance profile is available. One extraction batch and one stock solution were used, prepared from root material collected at one site and time point; extract composition and yield can vary with harvest season, provenance and batch. (ii) Replication: N = 3 refers to three replicate plates per treatment (disk diffusion) and to three parallel replicate cultures derived from the same single-isolate culture (RT-qPCR); this design captures assay-level rather than strain-level variability. Replicate-level zone diameters and Ct values are not presented, and no inferential statistics are reported. (iii) Antibacterial testing: MIC, MBC, checkerboard/FICI and time–kill assays were not performed, so neither potency nor interactions with florfenicol or enrofloxacin can be assessed; the brown extract-colored zone may make the zone margin difficult to read, and pigment diffusion may not coincide exactly with growth inhibition. (iv) Chemistry: GC-MS peak assignments were not confirmed with authentic reference standards, the analysis was performed on a derivatized fraction, and no constituent was quantified; confirmation with standards or by liquid chromatography-tandem mass spectrometry (LC-MS/MS) would be required for compound identification. (v) RT-qPCR: 16S rRNA was the only reference gene; standard curves and primer efficiencies were not determined; the amplicons (957 bp for tet(A), 400 bp for floR and approximately 1.5 kb for 16S rRNA) are longer than those usually preferred for SYBR Green assays; reactions were run in technical duplicate; and the assay is not fully MIQE-compliant. The relative values should therefore be read as approximate estimates. (vi) Viability: no colony-forming-unit counts, growth curves or live/dead measurements were made alongside RNA isolation, so changes in tet(A) and floR transcript abundance cannot be separated from effects on viability, growth or RNA yield. (vii) No host, in vivo or application-related endpoint was examined.

Under the in vitro conditions tested, a single batch of methanolic root extract of G. trichosantha (2 mg per disk) produced a mean inhibition zone of 17.0 mm (including the 6 mm disk) against one V. anguillarum isolate from rainbow trout, and exposure to the highest extract concentration tested (≈1,000 µg mL - ¹) was associated with markedly lower relative transcript abundance of tet(A) and floR. Placing a separate extract disk alongside florfenicol or enrofloxacin disks did not increase inhibition around the antibiotic disks. These findings are preliminary and descriptive. They do not establish MIC-level potency, an interaction with the antibiotics, a regulatory effect on resistance genes, or any effect beyond the in vitro conditions tested. MIC/MBC and checkerboard/FICI testing, time–kill and viability-controlled expression analyses with validated reference genes and efficiency-corrected assays, testing of multiple well-characterized isolates and independently prepared extract batches, and confirmation of constituents with authentic standards are required before further conclusions can be drawn.