Value-Added Compounds with Antimicrobial, Antioxidant, and Enzyme-Inhibitory Effects from Post-Distillation and Post-Supercritical CO2 Extraction By-Products of Rosemary

Hydrodistillation is the main technique to obtain essential oils from rosemary for the aroma industry. However, this technique is wasteful, producing numerous by-products (residual water, spent materials) that are usually discarded in the environment. Supercritical CO2 (SC-CO2) extraction is considered an alternative greener technology for producing aroma compounds. However, there have been no discussions about the spent plant material leftover. Therefore, this work investigated the chemical profile (GC-MS, LC-HRMS/MS) and multi-biological activity (antimicrobial, antioxidant, enzyme inhibitory) of several raw rosemary materials (essential oil, SC-CO2 extracts, solvent extracts) and by-products/waste materials (post-distillation residual water, spent plant material extracts, and post-supercritical CO2 spent plant material extracts). More than 55 volatile organic compounds (e.g., pinene, eucalyptol, borneol, camphor, caryophyllene, etc.) were identified in the rosemary essential oil and SC-CO2 extracts. The LC-HRMS/MS profiling of the solvent extracts revealed around 25 specialized metabolites (e.g., caffeic acid, rosmarinic acid, salvianolic acids, luteolin derivatives, rosmanol derivatives, carnosol derivatives, etc.). Minimum inhibitory concentrations of 15.6–62.5 mg/L were obtained for some rosemary extracts against Micrococcus luteus, Bacilus cereus, or Staphylococcus aureus MRSA. Evaluated in six different in vitro tests, the antioxidant potential revealed strong activity for the polyphenol-containing extracts. In contrast, the terpene-rich extracts were more potent in inhibiting various key enzymes (e.g., acetylcholinesterase, butyrylcholinesterase, tyrosinase, amylase, and glucosidase). The current work brings new insightful contributions to the continuously developing body of knowledge about the valorization of rosemary by-products as a low-cost source of high-added-value constituents in the food, pharmaceutical, and cosmeceutical industries.


Plant Material
Dried leaves of rosemary were bought from a local market in Germany; the plant material was authenticated by one of the authors (A.T.). A voucher specimen (RO/220714) was deposited in Biothermodynamics, TUM School of Life Sciences, Technical University of Munich, Freising, Germany.

Extraction 2.3.1. Preparation of Essential Oil
The powdered dried rosemary leaves (50 g) were placed in a Clevenger-type apparatus with 500 mL of deionized water and distilled for 4 h. The obtained rosemary essential oil (REO) fraction was collected and dried over anhydrous sodium sulfate. The hydrodistillation procedure was performed in duplicate.

Preparation of SC-CO 2 Extracts
The SC-CO 2 extractions were performed on a Spe-ed SFE Zoran Extractor (Applied Separations, Allentown, PA, USA) which could operate at a maximum temperature and pressure of 180 • C and 690 bar, respectively. The extraction vessel was loaded with 40 g of powdered dried rosemary leaves, which were compressed to a bed length of 12.0 cm and a diameter of 3.0 cm. The vessel was sealed, placed in a thermostatic mantle, and connected to the extractor. The extraction was started after an equilibration static time of 10 min and performed for 30 min at a constant CO 2 flow (7 standard liter min −1 ). All SC-CO 2 experiments were conducted at a pressure of 100 bar, whereas the temperature conditions were 40 • C, 50 • C, and 60 • C, yielding RC1, RC2, and RC3, respectively. Each experimental setup was performed in duplicate.

Preparation of Total, Spent, and Residual Water Extracts
At the end of the hydrodistillation process, the residual water was filtered and freezedried, affording the rosemary water extract (RWE). Totals of 10 g of the powdered dried rosemary leaves, spent plant material after hydrodistillation, and spent plant materials after the three SC-CO 2 extractions were separately extracted at room temperature in a Bandelin Sonorex Digitec ultrasound bath from BANDELIN Electronic GmbH & Co. KG (Berlin, Germany) with 3 × 100 mL methanol/water 75/25 (v/v) in three repeated ultrasound cycles (35 Hz), each 30 min long. All extractions were performed in duplicate. The following extracts were thus obtained: total extract (RTE), spent plant material extract (RSE), SC-CO 2 (100 bar, 40 • C) spent extract (RSC1), SC-CO 2 (100 bar, 50 • C) spent extract (RSC2), SC-CO 2 (100 bar, 60 • C) spent extract (RSC3). The extract yields are provided in Table 1.

Phytochemical Characterization 2.4.1. GC-MS Analysis
A TRACE gas chromatograph [18] with an ISQ™ mass spectrometer (MS) from Thermo Fisher (Waltham, MA, USA) was used. The chromatographic separations were conducted on a Zebron™ ZB-5MS (30 m × 0.25 mm i.d., 0.25 µm film thickness) from Phenomenex (Torrance, CA, USA). Helium at a flow rate of 1.43 mL/min was the carrier gas. The inlet temperature was 250 • C, the split ratio was 50:1, and the injection volume was 1 µL. The oven temperature was held for 4 min at 60 • C; then it was increased to 280 • C at a rate of 10 • C/min and held for 5 min; finally, it was ramped to 300 • C at a rate of 10 • C/min and held for 10 min. The following MS settings were used: m/z: 50 to 350 amu; ionization energy: 70 eV; transfer line temperature: 320 • C; and source temperature: 230 • C. The linear retention indices were determined for each peak using a C8-C20 standard mixture of n-alkanes and compared with those of the literature. Furthermore, the recorded mass spectral information was compared with that from the NIST11 database. All measurements were performed in triplicate.

LC-HRMS/MS Analysis
An Agilent 1200 HPLC (Agilent Technologies, Palo Alto, CA, USA) with an accuratemass quadrupole time-of-flight MS detector (G6530B) was used. The chromatographic separations were conducted on a Gemini C18 column (100 mm × 2 mm i.d., 3 µm) from Phenomenex (Torrance, CA, USA). The mobile phases comprised water (A) and acetonitrile (B), both acidified with 0.1% formic acid. The run started with 10% B and linearly increased to 60% B in 45 min at a flow rate of 0.2 mL/min; the injection volume was 10 µL. The following MS settings were used: m/z: 100-1700 amu; negative ionization mode; carrier gas (nitrogen) flow rate and temperature of 10 L/min and 275 • C, respectively; sheath gas (nitrogen) flow rate and temperature of 12 L/min and 325 • C, respectively; nebulizer pressure: 35 psi; capillary voltage: 4000 V; nozzle voltage: 1000 V; skimmer: 65 V; fragmentor: 140 V; and collision-induced dissociation: 30 V. The recorded mass spectral information was compared with that from databases and the literature.

Total Phenolic and Flavonoid Content
The total phenolic content (TPC) and total flavonoid content (TFC) were determined spectrophotometrically as described in [19]. Data were expressed as mg gallic acid equivalents (GAE)/g extract in TPC and mg rutin equivalents (RE)/g extract in TFC. All measurements were performed in triplicate.

Enzyme-Inhibitory Assays
AChE, BChE, tyrosinase, amylase, and glucosidase inhibition were determined as presented in [19,21]. The activity data were expressed as mg galanthamine equivalents (GALAE)/g extract in the AChE and BChE assays, mg kojic acid equivalents (KAE)/g extract in tyrosinase assay, and mmol acarbose equivalents (ACAE)/g extract in amylase and glucosidase assays.

Statistical and Data Processing
Data are presented as mean ± standard deviation of the respective number of replicates. One-way analysis of variance with Tukey's post-hoc test was conducted; p < 0.05 was considered statistically significant. The relationship between rosemary compounds vs. antimicrobial, antioxidant, and enzyme-inhibitory activities was assessed by calculating the Pearson correlation coefficient. Principal component analysis (PCA) and clustered image maps (CIM) were also performed, with the phytochemical data log transformed. The statistical analysis was done using R software v. 4.1.2 (R Foundation for Statistical Computing, Vienna, Austria).

GC-MS Characterization of Essential Oil and SC-CO 2 Rosemary Extracts
In this section, rosemary extracts rich in volatile compounds were obtained by hydrodistillation and SC-CO 2 extraction and characterized by GC-MS. The operating SC-CO 2 pressure (100 bar) and temperature range (40-60 • C) were selected based on previous systematic studies that presented a high recovery of rosemary volatiles under these conditions [11,15]. The EO yield was significantly higher than the SC-CO 2 yields (Table 1). This can be related to the different extraction mechanisms between the hydrodistillation and SC-CO 2 extraction. Hydrodistillation allows the recovery of only high-vapor-pressure (volatile, 'distillable') compounds, whereas SC-CO 2 extraction allows the recovery of compounds mostly based on their polarity and to a lower extent based on their vapor pressure. The high lipophilicity of the solvent (SC-CO 2 ) would allow high extraction rates of lipophilic compounds, including the low-polarity terpenes.
Within the three SC-CO 2 extracts, the yield decreased with the increase in temperature. This is in connection with the fact that temperature increments are known to reduce solvent density and negatively impact the solubility and extractability of compounds [22]. According to the GC-MS profiling (Table 2), the rosemary EO (REO) was characterized by 46 compounds, primarily monoterpenes (~97.2% of the total GC-MS peak area). The putative identity of the volatile compounds was established by comparing the linear retention indices with those of the literature data and the recorded mass spectra with those of NIST11 database. However, due to the lack of authentic standards, only a partial structural identification is possible with these resources. Table 2. GC-MS profile (tentative annotation) of the essential oils obtained from thyme, oregano, and basil.
Overall, it can be noticed that hydrodistillation was clearly more selective in recovering rosemary terpenes. However, the selectivity of the SC-CO 2 extraction toward volatiles increased considerably by increasing the temperature from 40 to 60 • C. Most likely, the SC-CO 2 process, especially at 100 bar and 40 • C, allowed the simultaneous extraction of nonvolatile lipophilic compounds (e.g., waxes, fatty acids, lipophilic pigments, etc.). It is also worth emphasizing that the ratio of monoterpenes/sesquiterpenes significantly decreased from 36/1 in the REO to 4/1, 4.4/1, and 1.7/1 in RC1, RC2, and RC3, respectively. Even though various SC-CO 2 extraction conditions allow the efficient recovery of terpenes, the extraction yields, qualitative profile, and quantitative data of terpenes are significantly altered compared to those with hydrodistillation. Similar conclusions were also reported in previous works [12].
The by-product extracts that resulted after the hydrodistillation and SC-CO 2 extraction of rosemary can be regarded as rich sources of phytochemicals, especially phenolic compounds, such as phenolic acids, flavonoids, and diterpenes. Compared to the total (unspent material), no substantial qualitative differences were spotted in the spent material extracts. RTE and the three post-SC-CO 2 extracts showed a very similar metabolite profile. In the RWE, several non-polar diterpenes (e. g., 17, 18, 19, 21, 22, and 25) were not present, which could be linked to the high polarity of the solvent (water). However, RSE showed the highest abundance of compounds. Several hypotheses can be formulated. For instance, constituents found in small amounts in the original (unspent) plant materials could become more accessible to the solvent extraction that follows hydrodistillation. On the other hand, due to the long exposure of the plant material to boiling water, a cell permeation effect can be assumed, favoring the subsequent extraction of the metabolites. In addition, the harsh hydrodistillation conditions (high temperatures and long exposure times) can also lead to the formation of phenolic artifacts in the spent extracts.
To find more significant differences in the six extracts, a CIM analysis was next performed with the logarithmically transformed and scaled semi-quantitative data (peak area extracted from the base peak chromatograms of the LC-HRMS/MS analyses). As shown in Figure 2, the samples were distinguishable from each other, even if they seemed to form four clusters. Moreover, to describe the compounds characterizing each cluster, three blocks (I-III) were defined. In brief, RWE (cluster A) and RSC3 (cluster B) contained low concentrations of compounds grouped in blocks I and III (Figure 2). Cirsimaritin, ladanein, and hydroxybenzoic acid were abundant in cluster C comprising RSE. In contrast, the samples of cluster D (RSC1, RSC2, and RTE) had low levels of the compounds mentioned above. In this cluster, RTE contained the highest concentration of caffeic acid and luteolin-O-acetylglucuronides. spent material extracts. RTE and the three post-SC-CO2 extracts showed a very similar metabolite profile. In the RWE, several non-polar diterpenes (e. g., 17, 18, 19, 21, 22, and  25) were not present, which could be linked to the high polarity of the solvent (water). However, RSE showed the highest abundance of compounds. Several hypotheses can be formulated. For instance, constituents found in small amounts in the original (unspent) plant materials could become more accessible to the solvent extraction that follows hydrodistillation. On the other hand, due to the long exposure of the plant material to boiling water, a cell permeation effect can be assumed, favoring the subsequent extraction of the metabolites. In addition, the harsh hydrodistillation conditions (high temperatures and long exposure times) can also lead to the formation of phenolic artifacts in the spent extracts.
To find more significant differences in the six extracts, a CIM analysis was next performed with the logarithmically transformed and scaled semi-quantitative data (peak area extracted from the base peak chromatograms of the LC-HRMS/MS analyses). As shown in Figure 2, the samples were distinguishable from each other, even if they seemed to form four clusters. Moreover, to describe the compounds characterizing each cluster, three blocks (I-III) were defined. In brief, RWE (cluster A) and RSC3 (cluster B) contained low concentrations of compounds grouped in blocks I and III (Figure 2). Cirsimaritin, ladanein, and hydroxybenzoic acid were abundant in cluster C comprising RSE. In contrast, the samples of cluster D (RSC1, RSC2, and RTE) had low levels of the compounds mentioned above. In this cluster, RTE contained the highest concentration of caffeic acid and luteolin-O-acetylglucuronides.

Total Phenolic and Flavonoid Content of the Total, Post-Distillation, and Post-SC-CO2 Rosemary Extracts
In this section, the TPC and TFC of the total (RTE), post-distillation (RSE, RWE), and post-SC-CO2 (RSC1-3) rosemary extracts were determined. As can be seen from Table 4, the highest TPC was detected in RWE (108.10 mg GAE/g), followed by RTE (99.36 mg

Total Phenolic and Flavonoid Content of the Total, Post-Distillation, and Post-SC-CO 2 Rosemary Extracts
In this section, the TPC and TFC of the total (RTE), post-distillation (RSE, RWE), and post-SC-CO 2 (RSC1-3) rosemary extracts were determined. As can be seen from Table 4, the highest TPC was detected in RWE (108.10 mg GAE/g), followed by RTE (99.36 mg GAE/g), RSC2 (98.02 mg GAE/g), and RSC1 (97.68 mg GAE/g). RSE contained the lowest TPC. Regarding TFC, the highest content was recorded in RSC2 (32.58 mg RE/g) and the lowest in RSE (19.86 mg RE/g). Altogether, RWE can be regarded as an extract with a very high amount of both phenolic and flavonoid compounds. Different results regarding the total bioactive content of rosemary extracts have been reported in the literature [29,30]. In a recent paper by Zeroual et al. [31], the TPC and TFC in rosemary extracts were dependent on extraction methods (Soxhlet and maceration) and solvents (hexane, ethyl acetate, methanol, and ethanol). In their study, the highest TPC (34.98 mg GAE/g) was lower than that of the current findings. In another study [32], sixty Jordanian plants were investigated, with the highest TPC recorded in the rosemary extract (101.339 mg GAE/g).

Post-Distillation and Post-SC-CO 2 Rosemary Extracts as Antimicrobials
The extensive use of antibiotics and the rapid emergence of multi-drug-resistant microbial strains represent severe issues for modern medicine. Numerous approaches are currently under evaluation, such as using novel plant-based antimicrobials with superior efficiency and safety profiles [33]. Various studies have repeatedly brought to attention the antimicrobial activity of rosemary EO and solvent extracts [34][35][36][37][38][39]. Thus, in this section, the activity of the ten rosemary raw and by-product extracts was evaluated by the micro-dilution method in a panel of 14 pathogenic strains. The criteria proposed by Kuete and Efferth [40] were used to categorize the observed activity into significant (MIC < 100 mg/L) and moderate-to-weak (MIC > 100 mg/L) activity. According to the results presented in Table 5, it was observed that REO, RC2, RC3, and RWE showed practically no relevant antimicrobial activity (MIC > 250 mg/L). In connection with the phytochemical composition (Tables 2 and 3), it can be assumed that the rosemary extracts rich in lipophilic compounds (the case of REO, RC2, and RC3) or hydrophilic compounds (the case of RWE) were inactive. Generally, the Gram-negative bacteria and yeasts were not inhibited by any extract. The most sensitive strains (MIC = 15.6 mg/L) were S. aureus after the treatment with RSE and RSC1 and M. luteus after the treatment with RC1. With MIC values of 31.3 mg/L, RC1, RTE, RSC2, and RSC3 also potently inhibited S. aureus. A similar effect was exhibited by RC1 against S. epidermidis, RSE against M. luteus and E. faecalis, and RSC1 against M. luteus. S. aureus MRSA was sensitive (MIC = 62.5 mg/L) to RC1 and RSE, whereas B. cereus was inhibited to the same extent by RC1, RTE, and RSE. The high MIC values for the rosemary EO agree with those of the literature [36,37]. For example, Hussain et al. [38] reported MIC values ranging from 300 mg/L to 1720 mg/L for rosemary EO against various Gram-positive and Gram-negative bacteria, whereas Ojeda-Sana et al. [39] documented values of 1000-2500 mg/L against S. aureus, E. faecalis, E. coli, and K. pneumonia. In contrast, various solvent extracts were more potent as antimicrobial agents. Amaral et al. [34] reported MIC values ranging from 16 to 256 mg/L for rosemary extracts obtained with ethyl acetate, dichloromethane, and ethanol against S. aureus, S. epidermidis, and B. cereus. Karadag et al. [35] showed MIC values between 78 and 156 mg/L against S. aureus, E. faecalis, and H. pylori for a hexane rosemary extract. In summary, it can be stated that the post-distillation and post-SC-CO 2 extracts are more efficient antimicrobial agents than the EO and SC-CO 2 extracts. In addition, some polyphenolic compounds' (e.g., epiirosmanol with S. aureus and C. albicans) volatile metabolites (e.g., camphor with C. parapsilosis) seemed to have been correlated to some extent with the antimicrobial activity (Figures 3 and 4). The high MIC values for the rosemary EO agree with those of the literature [36,37]. For example, Hussain et al. [38] reported MIC values ranging from 300 mg/L to 1720 mg/L for rosemary EO against various Gram-positive and Gram-negative bacteria, whereas Ojeda-Sana et al. [39] documented values of 1000-2500 mg/L against S. aureus, E. faecalis, E. coli, and K. pneumonia. In contrast, various solvent extracts were more potent as antimicrobial agents. Amaral et al. [34] reported MIC values ranging from 16 to 256 mg/L for rosemary extracts obtained with ethyl acetate, dichloromethane, and ethanol against S. aureus, S. epidermidis, and B. cereus. Karadag et al. [35] showed MIC values between 78 and 156 mg/L against S. aureus, E. faecalis, and H. pylori for a hexane rosemary extract. In summary, it can be stated that the post-distillation and post-SC-CO2 extracts are more efficient antimicrobial agents than the EO and SC-CO2 extracts. In addition, some polyphenolic compounds' (e.g., epiirosmanol with S. aureus and C. albicans) volatile metabolites (e.g., camphor with C. parapsilosis) seemed to have been correlated to some extent with the antimicrobial activity (Figures 3 and 4).

Post-Distillation and Post-SC-CO2 Rosemary Extracts as Antioxidants
Over the past decade, antioxidants have become increasingly popular in the treatment of oxidative-stress-related diseases, such as cardiovascular disease, diabetes, and cancer [41]. Thus, intensive efforts are carried out to identify new and safer sources of antioxidants. In this section, the antioxidant properties of rosemary extracts obtained from raw and by-product materials were investigated in six complementary assays, including radical quenching (ABTS and DPPH), reducing power (CUPRAC and FRAP), phosphomolybdenum, and metal chelating. The results are presented in Table 6. In the radical scavenging assays, the best ability was noted in RSC2 (DPPH: 173.49 mg TE/g; ABTS: 202.19 mg TE/g), followed by RWE (DPPH: 164.08 mg TE/g; ABTS: 179.76 mg TE/g),

Post-Distillation and Post-SC-CO 2 Rosemary Extracts as Antioxidants
Over the past decade, antioxidants have become increasingly popular in the treatment of oxidative-stress-related diseases, such as cardiovascular disease, diabetes, and cancer [41]. Thus, intensive efforts are carried out to identify new and safer sources of antioxidants. In this section, the antioxidant properties of rosemary extracts obtained from raw and by-product materials were investigated in six complementary assays, including radical quenching (ABTS and DPPH), reducing power (CUPRAC and FRAP), phosphomolybdenum, and metal chelating. The results are presented in Table 6. In the radical scavenging assays, the best ability was noted in RSC2 (DPPH: 173.49 mg TE/g; ABTS: 202.19 mg TE/g), followed by RWE (DPPH: 164.08 mg TE/g; ABTS: 179.76 mg TE/g), and RTE (DPPH: 144.17 mg TE/g; ABTS: 155.03 mg TE/g). The weakest abilities for both radical quenching abilities were found in the samples obtained from the supercritical CO 2 extraction, and they can be ranked as RC1 > RC2 > RC3. The observed radical scavenging abilities were adversely affected by the increase in temperature during the supercritical CO 2 extraction procedure. In the spent extracts from the SC-CO 2 extractions, the radical scavenging ability decreased in the order RSC2 > RSC1 > RSC3, which corresponds to the level of total bioactive compounds. Additionally, when REO was compared to the SC-CO 2 extracts, REO demonstrated a higher radical scavenging ability than RC3. With values of 396.28 mg TE/g in CUPRAC and 205.38 mg TE/g in FRAP, RWE can be an excellent reducing agent compared to other samples. The reduction power of SC-CO 2 and post-SC-CO 2 extracts followed the same pattern as the radical scavenging activity. From these findings, it could be concluded that similar compounds could play a key role in the assays. As can be seen in Figure 3, some compounds (e.g., rosmarinic acid, luteolin, and caffeic acid) correlated strongly with radical scavenging and reducing abilities. Consistent with our approach, several researchers have already described these compounds as powerful antioxidants [18,42,43]. In addition, some volatile metabolites, such as thymol and carvacrol, could contribute significantly to the observed radical scavenging and reducing activities of REO and SC-CO 2 extracts. active than the polar samples. This fact was also observed in the correlation analysis. As shown in Figure 6, numerous volatile compounds were strongly associated with this propensity. These results are consistent with those of the literature that reported potent phosphomolybdenum properties for EOs [44,45]. Additionally, significant antioxidant properties of rosemary extracts, post-distillation, or essential oils have been reported in several studies [11,46,47].

Post-Distillation and Post-SC-CO2 Rosemary Extracts as Enzyme Inhibitors
In this section, the inhibitory effects of rosemary raw and by-product extracts against AChE, BChE, tyrosinase, amylase, and glucosidase were investigated (Table 7). In the AChE inhibition, the best result was achieved by RSC3 with 3.80 mg GALAE/g. However, its ability was similar to that of RSC1, RC3, and RC2. Interestingly, none of the polar extracts were active on BChE except RSC3. In contrast to the AChE inhibition, RC3 exhibited the most potent BChE inhibitory effect (3.01 mg GALAE/g). As shown in Figure  7, specific terpenoids, including linalool, terpinene-4-ol, and camphor, may be responsible for the anti-cholinesterase properties observed in the EO and SC-CO2 extracts. In this sense, a good agreement with previous studies was found [48][49][50]. In addition, the cholinesterase-inhibiting effects of rosemary have been reported in several studies. Table 7. Enzyme-inhibitory properties of rosemary extracts obtained from raw, post-distillation, or post-SC-CO2 materials.

Post-Distillation and Post-SC-CO 2 Rosemary Extracts as Enzyme Inhibitors
In this section, the inhibitory effects of rosemary raw and by-product extracts against AChE, BChE, tyrosinase, amylase, and glucosidase were investigated (Table 7). In the AChE inhibition, the best result was achieved by RSC3 with 3.80 mg GALAE/g. However, its ability was similar to that of RSC1, RC3, and RC2. Interestingly, none of the polar extracts were active on BChE except RSC3. In contrast to the AChE inhibition, RC3 exhibited the most potent BChE inhibitory effect (3.01 mg GALAE/g). As shown in Figure 7, specific terpenoids, including linalool, terpinene-4-ol, and camphor, may be responsible for the anti-cholinesterase properties observed in the EO and SC-CO 2 extracts. In this sense, a good agreement with previous studies was found [48][49][50]. In addition, the cholinesteraseinhibiting effects of rosemary have been reported in several studies.
The highest tyrosinase inhibition was provided by REO with 59.23 mg KAE/g, followed by RC1, RC2, and RSC1. The anti-tyrosinase activity of post-SC-CO 2 extracts was less potent than their supercritical counterparts. This fact could be explained by some volatile compounds (α-pinene, β-pinene, p-cymene, etc.) and was confirmed as shown in Figure 7. The residual water, spent, and total extracts showed similar anti-tyrosinase abilities.  The highest tyrosinase inhibition was provided by REO with 59.23 mg KAE/g, followed by RC1, RC2, and RSC1. The anti-tyrosinase activity of post-SC-CO2 extracts was less potent than their supercritical counterparts. This fact could be explained by some volatile compounds (α-pinene, β-pinene, p-cymene, etc.) and was confirmed as shown in Figure 7. The residual water, spent, and total extracts showed similar anti-tyrosinase abilities.
REO achieved the most substantial inhibition value (0.39 mmol ACAE/g) in the case of amylase. Surprisingly, the same sample was not active on glucosidase; the highest glucosidase inhibitory activity was found in RSE (1.24 mmol ACAE/g), but the value was similar to RC1 (1.20 mmol ACAE/g) and RC2 (1.17 mmol ACAE/g). These results suggest terpenoids, such as α-pinene, β-pinene, or α-phellandrene, might be attributed to amylase inhibition. At the same time, some phenolics (including luteolin and ladanein) might also be the main players in the glucosidase-inhibitory capacity ( Figure 8). Moreover, the mentioned compounds have been reported to have an inhibitory effect on the enzymes [51][52][53]  REO achieved the most substantial inhibition value (0.39 mmol ACAE/g) in the case of amylase. Surprisingly, the same sample was not active on glucosidase; the highest glucosidase inhibitory activity was found in RSE (1.24 mmol ACAE/g), but the value was similar to RC1 (1.20 mmol ACAE/g) and RC2 (1.17 mmol ACAE/g). These results suggest terpenoids, such as α-pinene, β-pinene, or α-phellandrene, might be attributed to amylase inhibition. At the same time, some phenolics (including luteolin and ladanein) might also be the main players in the glucosidase-inhibitory capacity ( Figure 8). Moreover, the mentioned compounds have been reported to have an inhibitory effect on the enzymes [51][52][53]. The highest tyrosinase inhibition was provided by REO with 59.23 mg KAE/g, followed by RC1, RC2, and RSC1. The anti-tyrosinase activity of post-SC-CO2 extracts was less potent than their supercritical counterparts. This fact could be explained by some volatile compounds (α-pinene, β-pinene, p-cymene, etc.) and was confirmed as shown in Figure 7. The residual water, spent, and total extracts showed similar anti-tyrosinase abilities.
REO achieved the most substantial inhibition value (0.39 mmol ACAE/g) in the case of amylase. Surprisingly, the same sample was not active on glucosidase; the highest glucosidase inhibitory activity was found in RSE (1.24 mmol ACAE/g), but the value was similar to RC1 (1.20 mmol ACAE/g) and RC2 (1.17 mmol ACAE/g). These results suggest terpenoids, such as α-pinene, β-pinene, or α-phellandrene, might be attributed to amylase inhibition. At the same time, some phenolics (including luteolin and ladanein) might also be the main players in the glucosidase-inhibitory capacity ( Figure 8). Moreover, the mentioned compounds have been reported to have an inhibitory effect on the enzymes [51][52][53]

Multivariate Analysis
The application of multivariate tools in biochemical sciences has been proven to be extremely convenient since it enables the cluster of different biological activities of

Multivariate Analysis
The application of multivariate tools in biochemical sciences has been proven to be extremely convenient since it enables the cluster of different biological activities of different samples [54]. To establish the global overview of the similarities and differences between all the rosemary samples in terms of their bioactivities, two multivariate methods (PCA, CIM) were applied. Before the PCA analysis, the data were scaled to ensure the equal influence of all the bioactivities analyzed. In the CIM analysis, clusters were formed by the Ward method, and Euclidean distance was applied as a measure of diversity in the cluster analysis.
In Figure 9A, the first three dimensions represented practically 90% of the variance. The relationship of the three dimensions with the bioactivities is presented in Figure 9B. Due to its high percentage of explained variance (58.9%), the first dimension was linked to several bioactivities compared to the other two dimensions. Indeed, dimension 1 was positively correlated with DPPH, ABTS, CUPRAC, and FRAP and negatively correlated with BChE. The second dimension, which accounted for 21.4% of the variance, was positively bound to phosphomolybdenum, amylase, and tyrosinase and negatively bound to glucosidase. In the third dimension, only AChE showed a significant correlation. by the Ward method, and Euclidean distance was applied as a measure of diversity in the cluster analysis.
In Figure 9A, the first three dimensions represented practically 90% of the variance. The relationship of the three dimensions with the bioactivities is presented in Figure 9B. Due to its high percentage of explained variance (58.9%), the first dimension was linked to several bioactivities compared to the other two dimensions. Indeed, dimension 1 was positively correlated with DPPH, ABTS, CUPRAC, and FRAP and negatively correlated with BChE. The second dimension, which accounted for 21.4% of the variance, was positively bound to phosphomolybdenum, amylase, and tyrosinase and negatively bound to glucosidase. In the third dimension, only AChE showed a significant correlation. Figure 9C shows the division of the tested samples based on the three dimensions. In each scatter plot, the samples were divided into three clusters. In each of these plots, the REO sample formed a standalone cluster. In addition, the samples forming clusters A and B in the first scatter plot differed from those constituting the same clusters in the remaining scatter plot. Thereby to evaluate the accuracy of the PCA classification, the clustering was adequately identified by the CIM analysis. Five clusters grouped into two large clusters were obtained ( Figure 10). Cluster A comprised REO, which had remarkable phosphomolybdenum, anti-tyrosinase, and anti-amylase activities. Cluster B included RC3 and RC2, which demonstrated a relatively high anti-BChE activity. Cluster C contained RC1 and RSE. Clusters D (RSC1, RSC2, and RSC3) and E (RTE and RWE) were distinguished from the other clusters by their antioxidant activity. In short, the terpenecontaining extracts (REO, RC1-RC3) showed better anti-enzymatic activity, while the remaining extracts showed good antioxidant activity.   Figure 9C shows the division of the tested samples based on the three dimensions. In each scatter plot, the samples were divided into three clusters. In each of these plots, the REO sample formed a standalone cluster. In addition, the samples forming clusters A and B in the first scatter plot differed from those constituting the same clusters in the remaining scatter plot. Thereby to evaluate the accuracy of the PCA classification, the clustering was adequately identified by the CIM analysis. Five clusters grouped into two large clusters were obtained ( Figure 10). Cluster A comprised REO, which had remarkable phosphomolybdenum, anti-tyrosinase, and anti-amylase activities. Cluster B included RC3 and RC2, which demonstrated a relatively high anti-BChE activity. Cluster C contained RC1 and RSE. Clusters D (RSC1, RSC2, and RSC3) and E (RTE and RWE) were distinguished from the other clusters by their antioxidant activity. In short, the terpene-containing extracts (REO, RC1-RC3) showed better anti-enzymatic activity, while the remaining extracts showed good antioxidant activity.

Conclusions
In this work, rosemary extracts obtained from raw materials (essential oil, SC-CO2, and total extracts) and post-distillation and post-SC-CO2 materials were comparatively assessed for the first time from a phytochemical (GC-MS, LC-HRMS/MS) and multibiological (antimicrobial, antioxidant, enzyme-inhibitory) approach. Overall, it can be concluded that the by-products can find uses beyond those of the terpene-rich extracts (EO, SC-CO2) that are conventionally used as food preservatives (antioxidants) or aromaactive ingredients. The antimicrobial, antioxidant, and enzyme-inhibitory results could provide initial evidence for the health-promoting effects of the post-distillation and post-SC-CO2 samples, which can constitute novel materials for the pharmaceutical, cosmetic, and nutraceutical industries. Furthermore, this can lead to finding new ways of exploiting the large amounts of waste produced worldwide by the rosemary essential oil industry, with beneficial environmental, technological, and economic advantages.

Conclusions
In this work, rosemary extracts obtained from raw materials (essential oil, SC-CO 2 , and total extracts) and post-distillation and post-SC-CO 2 materials were comparatively assessed for the first time from a phytochemical (GC-MS, LC-HRMS/MS) and multibiological (antimicrobial, antioxidant, enzyme-inhibitory) approach. Overall, it can be concluded that the by-products can find uses beyond those of the terpene-rich extracts (EO, SC-CO 2 ) that are conventionally used as food preservatives (antioxidants) or aromaactive ingredients. The antimicrobial, antioxidant, and enzyme-inhibitory results could provide initial evidence for the health-promoting effects of the post-distillation and post-SC-CO 2 samples, which can constitute novel materials for the pharmaceutical, cosmetic, and nutraceutical industries. Furthermore, this can lead to finding new ways of exploiting the large amounts of waste produced worldwide by the rosemary essential oil industry, with beneficial environmental, technological, and economic advantages.