Plants of Rosmarinus officinalis L. were obtained at a local market in Masafi, Ras Al Khaimah, UAE and then grown in Al Atheab, Umm Al Quwain, under greenhouse conditions using UAE sand as a substrate. The harvest was at a uniform vegetative pre-flowering stage of the plants. Fresh healthy leaves were picked in the evening, rinsed with distilled water, dried at 45˚C to constant weight and then ground to fine powder using the short-pulse electric grinding. Sampled material was placed in amber glass jars that were stored under refrigeration. Morphological leaf shape, aroma and growth habit were used to confirm species identity as R. officinalis.

The most common extraction solvent was analytical-grade absolute ethanol (99.9%). Calibration was carried out using a RA reference standard (96% purity). All microbiological work was performed using Mueller-Hinton agar, nutrient broth and nutrient agar. Antibiotic discs (MASTDISCS^TM, Mast Group Ltd., UK) included Penicillin G (10 units), Ampicillin (AMP, 10 µg), Ciprofloxacin (CIP, 5 µg), Ceftriaxone (CRO, 30 µg), Imipenem (IMI, 10 µg), and Erythromycin (E, 15 µg).

Fresh rosemary leaves (~5 g per replicate, n = 4) were oven-dried at 100˚C ± 5˚C for 24 hours, cooled in a desiccator, and reweighed. Moisture content (percent) was calculated as

The primary stock solution of RA was prepared in 70 (v/v) ethanol at 1000 µg/mL, and the purity of 96% was corrected. Ten-fold dilution was used to prepare a 100 µg/mL working solution. UV-Vis scanning (200 - 450 nm) in quadruplicate (BioTek Epoch 2 microplate reader) was used to determine the maximum analytical wavelength (λmax). Calibration standards were prepared between 5 and 30 µg/mL (n = 4 per level) in 70% ethanol and absorbance was recorded at 330 nm with 200 µL per well. Linear regression was used to produce the calibration equation that was used to quantify RA.

Three parameters were optimised sequentially using ultrasound-assisted extraction (UAE) in an ultrasonic water bath at 45˚C with 70% ethanol: 1) plant-to-solvent ratio (1:40, 1:30, 1:20, 1:10 w/v, fixed at 10 mL; n = 4 per ratio); 2) extraction time (20, 30, 45, 60 min at 1:20 w/v; n = 4); and 3) number of successive extraction cycles (up to 3 cycles; 0.25 g dry material, 5 mL per cycle; n = 4). All extracts were filtered through 0.2 µm syringe filters, diluted 30-fold with 70% ethanol, and absorbance was measured at 330 nm. RA-equivalent concentrations were calculated from the calibration equation corrected for the 30-fold dilution factor. Precision was assessed by

Optimum extraction conditions were selected descriptively based on three criteria: higher RA-equivalent recovery, higher absorbance suitable for RA-equivalent calculation, and lower coefficient of variation among four replicates.

Ten grams of dried rosemary powder was extracted under optimised UAE conditions. The pooled extract was concentrated by rotary evaporation at 35˚C under reduced pressure, and the flask was rinsed with 90% ethanol to recover residual extract. All fractions were transferred to pre-weighed Falcon tubes and freeze-dried (lyophilised) over approximately four days. Crude extract mass was determined gravimetrically using the formula:

RA-equivalent content in the crude extract was estimated spectrophotometrically and expressed as mg RA equivalents per gram of crude extract. Because rosemary contains other phenolic compounds that may absorb at 330 nm, the values are reported as RA equivalents rather than specific RA-equivalent concentrations.

Local isolates of K. pneumoniae, P. aeruginosa and E. coliwere used alongside ATCC reference strains (LYFO DISK^®): E. coliATCC 25922, K. pneumoniae ATCC 13883, and P. aeruginosa ATCC 27853. All the 3 tested strains were isolated from urine samples and stored in the microbiology laboratory for routine work with the alphabetical serial number D2, E1, G1. Stocks were maintained at −80˚C in 50% (v/v) glycerol. Local isolates were identified by VITEK^® 2 (bioMérieux) using GN identification cards following Gram staining.

Susceptibility was assessed by Kirby-Bauer disk diffusion on Mueller-Hinton agar per CLSI M100, 36th edition [12]. Bacterial suspensions were adjusted to 0.5 McFarland (~1.5 × 10⁸ CFU/mL) by DensiCheck. Antibiotic discs were placed on inoculated plates and incubated at 37˚C for 24 h. Zone diameters were measured and interpreted (S/I/R) per CLSI organism-specific breakpoints. Four independent replicates (n = 4) were performed. ATCC reference strains served as quality controls.

The freeze-dried rosemary crude extract was dissolved in 70% ethanol to make four concentrations of 100, 200, 300, and 400 mg/mL. The concentrations of these crude extracts were made depending on the weight of the crude extracts. For 2.50 mL final volume, 0.2500, 0.5000, 0.7500, and 1.0000 g of crude extract were dissolved in 2.50 mL of 70% ethanol to prepare the 100, 200, 300, and 400 mg/mL solutions, respectively.

The RA-equivalent content of the crude extract was determined to be 55.74 mg RA equivalents/g crude extract. Thus, the corresponding pure RA-equivalent concentrations were determined by the following equation:

Based on this, the corresponding equivalent pure RA-equivalent concentrations of 100, 200, 300, and 400 mg/mL crude extract were 5.574, 11.148, 16.722 and 22.296 mg/mL, respectively. As an illustration, the RA-equivalent concentration of 400 mg/mL crude extract was computed as follows:

Sterile filter paper discs were loaded with 7.5 µL of each solution, air-dried, and placed on Mueller-Hinton agar plates inoculated with standardised suspensions of six strains: E. coli ATCC 25922, E. coli local (E1), P. aeruginosa ATCC 27853, P. aeruginosa local (G1), K. pneumoniae ATCC 13883, and K. pneumoniae local (D2). The negative control was a 70% ethanol disc and the positive control was a standard antibiotic disc. Incubation of plates was done at 37˚C over 24 h and inhibition zones were measured.

The highest crude extract concentration (400 mg/mL) and its equivalent pure RA were combined with antibiotics selected by strain susceptibility: CIP 5 µg for E.coliand Pseudomonas, CRO 30 µg for Klebsiella. Three-disc conditions were tested per plate: 1) antibiotic alone; 2) antibiotic + crude extract (7.5 µL of 400 mg/mL applied onto the antibiotic disc); 3) antibiotic + equivalent RA. Assays were conducted on three ATCC strains and three local isolates in triplicate.

Findings are presented in the form of mean ± standard deviation. Coefficient of variation (CV percent) was calculated as:

Linear regression was used to construct a calibration curve. The extraction optimisation and antibacterial assays were evaluated descriptively based on mean values, CV%, and inhibition-zone comparisons. No inferential statistical test was applied. All calculations were performed using Microsoft Excel.

The fresh rosemary leaves were analysed using the gravimetric analysis to identify the moisture content of the leaves in four replicates. Table 1 summarises the results. The values of individual moisture content were between 85.03 and 85.46 percent among replicates. The mean moisture level was 85.27 + 0.18%, which implied that the moisture level was high and reproducible.

Table 1. Moisture content of fresh rosemary leaves (Rosmarinus officinalis) (n = 4).

The UV-Vis absorption spectrums of R measured in the wavelength range of 200-450 nm exhibited similar spectral patterns in four replicates with a maximum absorbance (lmax) of 330 nm (Figure 1). At this wavelength, the measured absorbance values for calibration standards increased with concentration from 5 to 30 µg/mL, yielding mean absorbance values of 0.50 ± 0.01, 0.77 ± 0.01, 1.02 ± 0.01, 1.24 ± 0.01, 1.66 ± 0.02, and 1.82 ± 0.01, respectively (Table 2). The analysis of the calibration data using linear regression yielded the equation y = 0.0542x + 0.2193 with a coefficient of determination (R²) of 0.9905 (Figure 2).

Figure 1. UV-Vis absorption spectrum of rosmarinic acid in 70% (v/v) ethanol and ethanol as control over the wavelength range of 250 - 450 nm, expressed as mean absorbance ± SD from four independent replicates (n = 4).

Table 2. Rosmarinic acid calibration standards and corresponding mean absorbance values at 330 nm measured using a BioTek Epoch 2 microplate reader (n = 4).

Figure 2. Standard curve of rosmarinic acid at 330 nm.

When extraction was performed on equivalent dry matter (0.50 g dry matter per 10 mL 70% ethanol), dried rosemary produced a marginally higher mean absorbance (0.912 ± 0.058; CV = 6.41%) than fresh rosemary (0.854 ± 0.092; CV = 10.79%) (Table 3). Converting these absorbance values using the calibration equation, the RA-equivalent yield from dried material was 7.67 mg/g dry weight compared to 7.03 mg/g dry weight from fresh material.

Table 3. Comparison of absorbance values (330 nm) and calculated RA-equivalent yields from fresh and dried rosemary extracts on a dry-weight basis (n = 4).

Four plant-to-solvent ratios were evaluated to determine the optimal balance between total RA-equivalent recovery and extraction efficiency (Table 4). Mean absorbance at 330 nm increased progressively from 0.521 ± 0.082 (1:40 ratio, 0.25 g dry mass) to 0.901 ± 0.038 (1:10 ratio, 1.00 g dry mass). However, when efficiency was expressed as RA-equivalent yield per gram of dry plant material, the inverse relationship was observed: yield decreased from 6.68 mg/g at the 1:40 ratio to 3.77 mg/g at the 1:10 ratio, consistent with solvent saturation effects at higher solid loadings. The 1:20 ratio (0.50 g dry mass, 10 mL 70% ethanol) achieved an acceptable balance between total RA-equivalent recovery (2.53 mg per extraction) and extraction efficiency (5.07 mg/g).

Table 4. Effect of solid-to-solvent ratio on absorbance at 330 nm, total RA-equivalent recovery, and extraction yield from dry rosemary powder (n = 4).

Mean absorbance values obtained at 20 min (0.712 ± 0.060, CV = 8.40%), 30 min (0.718 ± 0.049, CV = 6.80%), and 45 min (0.719 ± 0.036, CV = 5.00%) were closely comparable, suggesting that the majority of extractable RA-equivalent content is released within the first 20 minutes of sonication (Table 5). The longest extraction time (60 minutes) yielded the largest mean absorbance (0.804 ± 0.095) but with a significantly higher CV (11.80%).

Table 5. Absorbance values at 330 nm obtained from the first extraction of dry rosemary using ultrasound-assisted extraction at different extraction times (n = 4).

Three successive extraction cycles were performed on the same 0.25 g dry rosemary material using 5 mL 70% ethanol per cycle (Table 6). The highest percentage of RA-equivalent content was obtained in the first extraction (mean absorbance 0.8945 ± 0.0438, CV = 4.90%; yield 7.47 mg/g), which recovered 61.04% of the total RA-equivalent content recovered in three cycles. The second extraction yielded an additional 25.38% (mean absorbance 0.500 ± 0.030, CV = 6.0%), and the third cycle contributed only 13.58% of the total (mean absorbance 0.3695 ± 0.0343, CV = 9.3%). The cumulative RA-equivalent recovery after two cycles was 86.42% of the total.

Table 6. Mean absorbance values and calculated RA-equivalent yields from successive ultrasound-assisted extraction cycles of dry rosemary (n = 4, per cycle).

Large-scale extraction of 10.0296 g dried rosemary yielded 1.69025 g freeze-dried crude extract (Table 7). The crude extract (diluted 1:30) was spectrophotometrically quantified at 330 nm with an absorbance of 0.320. The use of the calibration equation gave an RA-equivalent concentration of 1.858 µg/mL in the diluted sample, which is equivalent to 55.74 mg RA equivalents/g crude extract (0.55738 mg/0.01 g). The detailed calculation is presented in Table 8.

Table 7. Weights of freeze-dried crude rosemary extract collected in individual Falcon tubes and calculated total yield.

* Tube 4 represents the additional fraction recovered from rinsing the rotary evaporator flask with 90% ethanol.

Table 8. Calculation of crude extract yield and RA-equivalent content of freeze-dried rosemary extract.

Tables 9-11 provide data on antibiotic susceptibility between local isolates and ATCC reference strains. The three local Gram-negative isolates were all ampicillin-resistant (zone diameter < 13 mm). The local E. coli and K. pneumoniae isolates were susceptible to imipenem, ceftriaxone, and ciprofloxacin. Intermediate susceptibility of local P. aeruginosa to imipenem and strong susceptibility to ciprofloxacin were observed. However, ceftriaxone was also tested against P. aeruginosa; however, no S/I/R interpretation was assigned because no CLSI breakpoint was applied for this organism-antibiotic combination. All ATCC QC strains performed within acceptable CLSI ranges (CV% < 7%).

Table 9. Antibiotic susceptibility of local Klebsiella pneumoniae versus ATCC 13883.

S = susceptible; I = intermediate; R = resistant.

Table 10. Antibiotic susceptibility of local Escherichia coli versus ATCC 25922.

S = susceptible; R = resistant.

Table 11. Antibiotic susceptibility of local Pseudomonas aeruginosa versus ATCC 27853.

I = intermediate; R = resistant.

The antibiotic control yielded clear inhibition zones in all the bacterial strains tested whereas the 70% ethanol control did not yield any inhibition. No inhibition zones were observed for either rosemary crude extract or equivalent RA-equivalent concentrations (100 - 400 mg/mL) against E. coli ATCC, E. colilocal isolate,Pseudomonas local isolate, and Klebsiella pneumoniae local isolate, with all measurements recorded as 0.00 mm. However, weak inhibition was observed against Pseudomonas aeruginosa ATCC treated with crude extract, with mean inhibition zones of 8.25 ± 0.12 mm, 8.52 ± 0.11 mm, 8.92 ± 0.12 mm, and 9.24 ± 0.13 mm at concentrations of 100, 200, 300, and 400 mg/mL, respectively. Detectable inhibition was also observed only in Klebsiella pneumoniae ATCC treated with crude extract, with mean inhibition zones of 11.19 ± 0.31 mm, 11.11 ± 0.60 mm, 11.59 ± 0.57 mm, and 12.30 ± 0.29 mm at concentrations of 100, 200, 300, and 400 mg/mL, respectively. No inhibition was recorded for equivalent RA at any concentration against any of the tested strains (Table 12).

Table 12. Inhibition zone diameters (mm) of crude rosemary extract and equivalent pure rosmarinic acid against selected bacterial strains. Data are expressed as mean ± standard deviation (SD), with coefficient of variation (CV%).

The antibiotic control resulted in quantifiable inhibition areas on all strains tested with an average diameter of 24.60 + 0.12 mm to 31.54 + 0.26 mm. When used in combination, both antibiotic + crude extract and antibiotic + RA showed a decrease in the inhibition zones in all strains. For Klebsiella ATCC, inhibition zones decreased from 27.98 ± 0.54 mm (antibiotic) to 27.95 ± 0.23 mm (antibiotic + crude) and 25.52 ± 0.12 mm (antibiotic + RA). Similar reductions were observed in PseudomonasATCC (24.60 ± 0.12 mm to 22.13 ± 0.54 mm and 21.59 ± 0.42 mm), E. coli ATCC (31.22 ± 0.66 mm to 30.10 ± 0.06 mm and 27.80 ± 0.27 mm), Klebsiella (local) (30.13 ± 0.75 mm to 29.34 ± 0.70 mm and 27.73 ± 0.64 mm), Pseudomonas (local) (31.54 ± 0.26 mm to 27.84 ± 0.53 mm and 24.88 ± 0.11 mm), and E. coli (local) (30.09 ± 0.92 mm to 27.59 ± 0.60 mm and 25.56 ± 0.53 mm), respectively (Table 13).

Table 13. Inhibition zone diameters (mm) of antibiotic alone and in combination with rosemary crude extract (400 mg/mL) or equivalent pure RA, tested against ATCC reference strains and local isolates.

The current experiment demonstrated that the Rosmarinus officinalis grown in the UAE is a quantifiable source of RA, with dry leaves containing slightly more RA-equivalent content than fresh leaves (7.67 vs. 7.03 mg/g dry weight) and lower variability. This is in accordance with the ultrasound extraction study by Albu et al. [13], who reported the enhancement of the dried rosemary extraction as compared to the fresh material and with Paniwnyk et al. [14], who demonstrated that sonication enhances the extraction efficiency of rosemary antioxidants, as well as reducing the time of extraction. Likewise, Zu et al. [15] also found that RA is recoverable under the ultrasound-assisted conditions in rosemary and could be observed at 330 nm, which is the same wavelength in the current study.

The optimisation findings were also in agreement with earlier extraction studies. The findings of the current investigation indicated that as the amount of plant loading was increased, the total material recovered rose; however, the yield per gram decreased. 20 - 45 min of extraction times yielded comparable absorbance readings, and two cycles yielded 86.42% of the total extractable RA-equivalent content. Such results are consistent with Paniwnyk et al. [14], who determined that ultrasound primarily increases the initial extraction phase, with lesser improvements in extraction, and with Bellumori et al. [16], who observed that the recovery of rosemary phenolic under ultrasound is highly condition-dependent and selective instead of merely being dependent on extension of extraction. Overall, these findings justify the chosen protocol as a promising approach to extracting RA-equivalent-rich extracts of UAE-grown rosemary.

The crude extract yield of 16.85 percent and an estimated 55.74 mg /g RA equivalents crude extract also suggest that rosemary grown under the UAE conditions has retained its useful phenolic content. Previously reported rosemary studies have also indicated that ethanol-based ultrasound extraction could yield phenolic-rich fractions expressed as RA equivalents, but with varying absolute values depending on cultivar, pretreatment, solvent composition, and analytical platform due to variation in studies [13]. In this respect, the current information positions rosemary cultivated in the UAE as a viable local source of RA-containing extract in the expanded literature on extraction.

In the current experiment, there was mild direct antibacterial activity of the rosemary preparations against Gram-negative strains tested. Crude extract demonstrated weak inhibition of Klebsiella pneumoniae ATCC with zones of approximately 11 - 12 mm, and weak inhibition of Pseudomonas aeruginosa ATCC with zones of approximately 8 - 9 mm, whereas pure RA had no inhibition of that strain, and neither preparation inhibited the tested Escherichia coli strains or the local Pseudomonas aeruginosa isolate in the standalone test. The profile is more similar to the results of Khalil and Al-Lebathi [17], who found only significant direct activity of ethanolic rosemary extract against multi-antibiotic-resistant Gram-negative bacteria and only determined inhibition zones of 7.50 ± 0.50 mm of P. aeruginosa and 6.00 ± 0.20 mm of K. pneumoniae at 100 mg/mL.

At the same time, the present results are weaker than those of Husein et al. [18], who reported measurable antibacterial activity of rosemary extract against urinary isolates, including inhibition zones up to 16 mm in E. coli, MIC values of 4 mg/mL for E. coliand 32 mg/mL for Enterococcus faecalis, and marked biofilm inhibition. This comparison puts the existing results in the literature with great variability in rosemary extract activity depending on the set of strain and assay format. The present study, therefore, aligns more closely with reports of modest Gram-negative activity than with studies showing stronger inhibition in broth-based or biofilm-focused systems.

A further result of the present study is that weak activity was observed with the crude extract against K. pneumoniae ATCC and P. aeruginosa ATCC, whereas equivalent pure RA showed no activity. This difference is in line with the larger RA literature synthesized by Kernou et al. [19], who found that antibacterial effects of RA and its derivatives are strongly reliant on the bacterial model and formulation, and with Ivanov et al. [10], who reported RA activity to be variable among organisms and experimental systems. In that scenario, the current evidence indicates that any low activity detected in this case can rather be ascribed to the crude phytochemical mixture than RA.

The combination disk diffusion assay revealed that the combination of the chosen antibiotics with either rosemary crude extract or the same amount of pure RA did not enhance the antibacterial activity against any of the tested strains. The combined treatments showed no interaction under the current experimental conditions as the inhibition zones of the combined treatments were always equal or smaller than those of the antibiotic alone in both ATCC reference strains and local isolates ofKlebsiella pneumoniae,Pseudomonas aeruginosa, andEscherichia coli. Furthermore, in several strains, the antibiotic-plus-RA condition had smaller inhibition zones than the antibiotic-plus-crude-extract condition, which suggests that pure RA did not show a greater antibiotic adjuvant effect than the crude extract in this assay. These results are in contrast to the earlier reports of combination effects between rosmarinic acid and antibiotics against Staphylococcus aureus and MRSA, suggesting that these effects are highly dependent on bacterial species, antibiotic selection, and assay format [11]. Likewise, differences from more recent rosemary-based combination studies may reflect variation in organism type, extract preparation, and test design [17].

Overall, these results present valuable evidence that the antibiotic-modifying effect of rosemary crude extract and RA is assay- and strain-specific, which provides a clear foundation to continue the research with more sensitive techniques and wider bacterial models.