Eco-Friendly Solution Based on Rosmarinus officinalis Hydro-Alcoholic Extract to Prevent Biodeterioration of Cultural Heritage Objects and Buildings

Biodeterioration of cultural heritage is caused by different organisms capable of inducing complex alteration processes. The present study aimed to evaluate the efficiency of Rosmarinus officinalis hydro-alcoholic extract to inhibit the growth of deteriogenic microbial strains. For this, the physico-chemical characterization of the vegetal extract by UHPLC–MS/MS, its antimicrobial and antibiofilm activity on a representative number of biodeteriogenic microbial strains, as well as the antioxidant activity determined by DPPH, CUPRAC, FRAP, TEAC methods, were performed. The extract had a total phenol content of 15.62 ± 0.97 mg GAE/mL of which approximately 8.53% were flavonoids. The polyphenolic profile included carnosic acid, carnosol, rosmarinic acid and hesperidin as major components. The extract exhibited good and wide spectrum antimicrobial activity, with low MIC (minimal inhibitory concentration) values against fungal strains such as Aspergillus clavatus (MIC = 1.2 mg/mL) and bacterial strains such as Arthrobacter globiformis (MIC = 0.78 mg/mL) or Bacillus cereus (MIC = 1.56 mg/mL). The rosemary extract inhibited the adherence capacity to the inert substrate of Penicillium chrysogenum strains isolated from wooden objects or textiles and B. thuringiensis strains. A potential mechanism of R. officinalis antimicrobial activity could be represented by the release of nitric oxide (NO), a universal signalling molecule for stress management. Moreover, the treatment of microbial cultures with subinhibitory concentrations has modulated the production of microbial enzymes and organic acids involved in biodeterioration, with the effect depending on the studied microbial strain, isolation source and the tested soluble factor. This paper reports for the first time the potential of R. officinalis hydro-alcoholic extract for the development of eco-friendly solutions dedicated to the conservation/safeguarding of tangible cultural heritage.


Introduction
Cultural heritage includes tangible heritage assets (buildings, monuments, books, material and decorative objects, natural heritage), and intangible heritage assets, such as language arts, rituals, folklore, oral traditions, and knowledge [1]. The biodeterioration of cultural heritage objects is determined by different biological systems capable of inducing complex alteration processes caused by their biological and metabolic activity, the type of constituent materials of art works and the environmental conditions. In order to inhibit or eradicate biological colonization, chemical methods are usually used in combination with physical ones. However, chemical compounds are generally toxic and non-degradable, persistent in the environment and could cause uncontrollable contamination [2,3]. Thus, over the last decades, trends in control of the biodeterioration process have indicated an increased necessity to develop non-toxic preservation and conservation solutions without adverse effects on cultural heritage objects, human health, and the natural environment [4].
Since ancient times, natural compounds have been used as alternative solutions in the treatment of infectious diseases, although their antimicrobial properties and mechanisms have been only recently proved scientifically. This has prompted the research of green biocides, such as different vegetal natural compounds and essential oils (EO), towards new applications beyond the clinical field, which, due to the biocides microbicidal, antioxidant and anti-biofilm properties, might include conservation and preservation of cultural heritage [5].
Among the vegetal biocides, EOs are among the most studied. This is because they contain different secondary metabolites able to inhibit the growth of bacterial, yeast or filamentous fungi by inducing the deterioration of the cellular walls and plasma membrane, and by interfering with microbial metabolism by affecting the protein synthesis or the structure and function of macromolecules, such as DNA, RNA, proteins and lipids [6,7]. Many plant extracts and essential oils have been described as being alternative solutions for heritage protection against biodeterioration. Although these solutions have a high antimicrobial efficiency it is important that they do not degrade the heritage object itself. Matusiak et al. [8] have described cinnamon essential oil as being safe for heritage textile disinfection since it had a limited impact on the optical, mechanical and structural properties that were observed. Lavin et al. [9] showed that Origanum vulgare L. and Thymus vulgaris L. EOs exhibited antifungal activity against Scopulariopsis spp. and Fusarium spp. deteriogenic strains, recommending these EOs for the protection of paper heritage objects against aesthetic and structural damage. Rosmarinus officinalis L. is recognized as a medicinal and aromatic plant from the Lamiaceae family native to the Mediterranean region [10][11][12]. EOs extracted from R. officinalis L. showed a higher effectiveness against fungal sporulation and growth than, e.g., the commercial biocide benzalkonium chloride [13]. Besides antimicrobial effects, the preventive conservation agents must ensure the reproducibility of the product's functionality and the optimal potential for inhibiting oxidative degradation. In preventive conservation, the focus is on the environment, mainly temperature, relative humidity and pollutants, which, if not controlled, could accelerate natural decay processes [14]. Thus, antioxidant agents can prevent the oxidation of substrates (formation of side groups, aldehydes and ketones) that may lead to the destabilization of the structure and thus, to easier hydrolysis. Free radicals generated by oxidation or light can contribute to the increase in acid concentration, because by photooxidation, primary alcohol groups in polymers can be oxidized to aldehydes and later to carboxylic acid species [15].
Due to their extremely erosive capacity, fungi are one of the most important biodeteriogenic organisms associated with stone buildings. Black fungi can cause biopitting, which appear as black spots and are resistant to antimicrobial treatments and biocides, owing to their thick walls [16]. Artworks from museums are one of the most affected heritage patrimony objects, with several species of fungi being associated with their biodeterioration. Some genera and species have been reported in association with various types of material and objects from art museum, such as: Penicillium spp. (parchment, keratinous substrates, paintings, papers), Alternaria spp., Aspergillus spp., Chaetomium spp., Mucor spp. (keratinous substrates, paintings, papers), Cladosporium spp. (parchment, keratinous substrates, paintings), Fusarium sp. (paintings, paper), Aureobasidium pullulans (paintings, keratinous substrates), Epicoccum nigrum (parchment, keratinous substrate), Rhizopus spp., Stachybotrys chartarum and Trichoderma spp. (paper, keratinous substrate), while others were isolated from only one type of substrate, such as: Toxicocladosporium irritans (paper), Phlebiopsis gigantea (parchment) and Coniosporium spp. (keratinous substrates) [17]. Aspergillus was reported on cultural heritage objects from churches and temples worldwide, the most frequently encountered being Aspergillus niger and A. versicolor [18]. Among bacterial species involved in biodeterioration B. subtilis, B. megaterium and B. mojavensis are most frequently encountered in Transylvania. Members of these species are highly resistant to nutrient deficiency and other environmental stress conditions exhibiting different characteristics such as the ability to form capsules, spores or biofilm. Most of the biodeteriogenic bacterial strains are able to secrete cellulases, which could be involved in the biodegradation process of wooden objects or esterases, enzymes that affect the murals in stone churches [19].
The present study aimed to evaluate the capacity of the hydro-alcoholic extract of R. officinalis to inhibit the growth and to modulate the production of soluble factors of deteriogenic fungal and bacterial strains previously isolated from the cultural heritage museum collections and churches in Romania. For this, the physico-chemical characterization of the extract, it's antimicrobial, antibiofilm and antioxidant activity, as well as it's modulatory activity on the production of soluble factors involved in biodegradation were demonstrated on a representative number of filamentous fungi and bacterial strains isolated from biodeteriorated wooden and stone heritage churches from Arad, Hunedoara Bucharest and Tulcea counties and from objects included in museum collections.

Chemical Composition and Antioxidant Activity
The obtained extract has a total phenol content of 15.62 ± 0.97 mg GAE/mL which approximately 8.53% were flavonoids (Table 1). The polyphenolic profile of this extract is characterized by the presence of carnosic acid, carnosol, rosmarinic acid and hesperidin, as major components [26]. HPLC-MS identified approximately 2.04% of the phenolic content and 14.19% of the total flavonoids (Table 2). Among the compounds identified by RP-HPLC-MS, the highest concentrations were obtained for hesperidin (10.39 mg/L), vanillic acid (7.51 mg/L) and quercetin (7.10 mg/mL). By MS, the compounds carnosol and rosmarinic acid were identified, the latter being the one of the main compounds. The main compound from R. officinalis extract, previously known for its broad antimicrobial spectrum, could be considered the rosmarinic acid ( Figure S1).
The best antioxidant activity for R. officinalis extract was obtained by the TEAC method followed by CUPRAC (Table 1), probably due to the similarity of redox potential between the redox couple Cu (II, I) -Nc (0.6 V) and ABTS+/ABTS (0.68 V) [27]. Another favourable reason to use TEAC and CUPRAC is that these methods are able to measure both hydrophilic and lipophilic antioxidants as opposed to the FRAP and DPPH methods [28]. In addition, the TEAC method can measure the antioxidant activity over a wide pH range, and the increase in acidic conditions due to the high content of phenolic acids facilitates the electron transfer [29]. The considered positive control was ferulic acid. This compound was found in the extract composition, the proportion of its antioxidant activity can be found in Table 1. The antioxidant activity of the extract is given by the synergisms/antagonisms between the amounts of compounds. The total ion current (TIC) chromatogram of R. officinalis extract by UHPLC-MS/MS detection is presented in Figure S1. Quantitative UHPLC-ESI/MS results indicate that the main phenolic acids identified in the alcoholic extract of R. officinalis were vanillic acid (7512.58 µg/L), caffeic acid (1637.63), chlorogenic acid (999.51µg/L), and syringic acid (836.12), while hesperidin (10,389.09 µg/L) and epi-catechin (664.45 µg/L) were the predominant flavonoids ( Table 2). The identification of the main quantified phenolic acids and flavonoids are presented in Figure S2.
The analytical approach based on a non-target UHPLC-Q-Orbitrap HRMS analysis allows the identification of other bioactive compounds and specialized metabolites which are also responsible for the bioactive potential of R. officinalis extract. Data processing analysis using Compound Discoverer software following a metabolomics working template includes RT alignment, molecular formula prediction, evaluation of adducts, the assignment and comparison of fragmentation pattern, background annotation, and an automated library and database search for identification purposes, including mzCloud (MS2 fragments), Chemspider, MzVault and Mass List Matches [30].
The most represented classes of bioactive compounds that were identified in R. officinalis extract are terpenoids (di-and tri-terpenoids) and flavonoids (flavones, flavanones, and flavonols). The compounds name, molecular formula, retention time, exact mass and accurate mass of m/z adduct ions, and MS/MS fragment ions in negative HESI mode are shown in Table 3. Diterpenoids compounds, such rosmanol, rosmaquinone and sophorol isomers, carnosol and carnosol derivatives, with strong anti-cancer and anti-inflammatory properties [31] were identified as predominant compounds in the rosemary extract ( Figure S3). Oleanolic and ursolic triterpenic acids, typical constituents in medicinal plants, with antibacterial, anti-inflammatory and anti-proliferative [32] activities, were also identified. Among flavonoids, flavones represent the predominant group, being present as glycosylated forms of luteolin, apigenin, isorhamnetin, hispidulin, and also dihydroxy-dimethoxy flavone, compounds with strong antioxidant and anti-proliferative activities [33]. Phenolic acids such as rosmarinic acid and methyl rosmarinate derivate and usnic acid were also identified.  A total number of 41 filamentous fungi strains isolated from different wooden and stone churches and from museum objects (paper and paintings) were identified according to presumptive examination and MALDI-TOF mass spectrometry as belonging to A. niger, A. flavus, A. pseudoglaucus, A. nidulans, A. flavus oryzae, A. versicolor, A. montevidensis, A. clavatus, P. chrysogenum, P. corylophylum, P. expansum, P. digitatum, P. brevicompactum, P. namylowskii, Rhizopus oryzae, A. alternata, Fusarium proliferatum, Mucor circinelloides, Purpureocillium lilacinum species [19].
The antifungal activity of R. officinalis hydro-alcoholic extract against 41 filamentous fungi strains, mainly belonging to Penicillium (49.9%) and Aspergillus genera (31.70%) was performed by an adapted diffusion method which demonstrated a good inhibitory effect of the extract as compared to solvent ( Figure S4). Therefore, the extract inhibited 60.97% of the tested filamentous fungal strains while the solvent only inhibited 21.95% (Figure 1).

Antibacterial Activity
For the antibacterial activity a total number of 15 bacterial strains isolated from wooden and stone churches and from museum collections were selected as representative for the microbial communities developed on cultural heritage objects from Romania [19].
The qualitative screening of the antimicrobial activity of the hydro-alcoholic extract of R. officinalis showed that the bioactive compounds inhibited the growth of all tested bacterial strains, evidenced by the total ( Figure S5) and increased diameter of the induced growth inhibition zone compared to the solvent ( Figure 2).

Antibacterial Activity
For the antibacterial activity a total number of 15 bacterial strains isolated from wooden and stone churches and from museum collections were selected as representative for the microbial communities developed on cultural heritage objects from Romania [19].
The qualitative screening of the antimicrobial activity of the hydro-alcoholic extract of R. officinalis showed that the bioactive compounds inhibited the growth of all tested bacterial strains, evidenced by the total ( Figure S5) and increased diameter of the induced growth inhibition zone compared to the solvent ( Figure 2).

Antifungal Activity
The distribution of the MIC values obtained using the binary serial dilution method by the isolation sources revealed that the most susceptible strains were recovered from the deteriorated textiles (MIC = 25.03 mg/mL) followed by stone and wooden churches, while the most resistant corresponded to other museum objects (MIC = 100 mg/mL) ( Figure 3).

Antifungal Activity
The distribution of the MIC values obtained using the binary serial dilution method by the isolation sources revealed that the most susceptible strains were recovered from the deteriorated textiles (MIC = 25.03 mg/mL) followed by stone and wooden churches, while the most resistant corresponded to other museum objects (MIC = 100 mg/mL) (Figure 3). The MIC distribution by the most frequently deteriogenic genera of the cultural heritage churches and museum collections (Penicillium spp. and Aspergillus spp.) ( Figure  S6a,b) has shown that R. officinalis hydro-alcoholic extract demonstrated the highest efficiency against A. clavatus (MIC = 1.2 mg/mL) and Aspergillus spp. strains (MIC = 6.25 mg/mL), while the P. namyslowskii and P. brevicompactum were the most resistant (MIC = 100 mg/mL).

Antibacterial Activity
The quantitative evaluation of the antimicrobial efficiency of R. officinalis hydro-alcoholic extract against bacterial strains recovered from different sources revealed that the highest efficiency was detected for strains isolated from paintings and textile objects (MIC = 1.56 mg/mL) ( Figure 4). Significantly lower MIC values (p < 0.01), except for strains isolated from textiles and paintings, were obtained for the extract compared to the solvent, demonstrating that the antibacterial activity was mainly due to the bioactive compounds contained by R. officinialis extract.
The MIC value ranges within each isolation source probably due to different physiological characteristics of the bacterial strains isolated from the respective source ( Figure  4).
Concerning the distribution of MIC values by bacterial species, the most susceptible strains belonged to Arthrobacter globiformis (MIC = 0.78 mg/mL), followed by Bacillus cereus, B. thuringiensis and A. aurascens (MIC = 1.56 mg/mL) and the most resistant corresponded to Pseudomonas koreensis recovered from stone churches (MIC = 6.25 mg/mL) (Figure S7). The same significantly lower MIC values were recorded for the R. officinialis extract as compared to the solvent (p < 0.01), except for B. atrophaeus and A. aurascens strains, suggesting the large spectrum antibacterial activity of the bioactive compounds contained by the tested extract. The MIC distribution by the most frequently deteriogenic genera of the cultural heritage churches and museum collections (Penicillium spp. and Aspergillus spp.) ( Figure S6a,b) has shown that R. officinalis hydro-alcoholic extract demonstrated the highest efficiency against A. clavatus (MIC = 1.2 mg/mL) and Aspergillus spp. strains (MIC = 6.25 mg/mL), while the P. namyslowskii and P. brevicompactum were the most resistant (MIC = 100 mg/mL).

Antibacterial Activity
The quantitative evaluation of the antimicrobial efficiency of R. officinalis hydroalcoholic extract against bacterial strains recovered from different sources revealed that the highest efficiency was detected for strains isolated from paintings and textile objects (MIC = 1.56 mg/mL) ( Figure 4). Significantly lower MIC values (p < 0.01), except for strains isolated from textiles and paintings, were obtained for the extract compared to the solvent, demonstrating that the antibacterial activity was mainly due to the bioactive compounds contained by R. officinialis extract.
The MIC value ranges within each isolation source probably due to different physiological characteristics of the bacterial strains isolated from the respective source ( Figure 4).
Concerning the distribution of MIC values by bacterial species, the most susceptible strains belonged to Arthrobacter globiformis (MIC = 0.78 mg/mL), followed by Bacillus cereus, B. thuringiensis and A. aurascens (MIC = 1.56 mg/mL) and the most resistant corresponded to Pseudomonas koreensis recovered from stone churches (MIC = 6.25 mg/mL) ( Figure S7). The same significantly lower MIC values were recorded for the R. officinialis extract as compared to the solvent (p < 0.01), except for B. atrophaeus and A. aurascens strains, suggesting the large spectrum antibacterial activity of the bioactive compounds contained by the tested extract.

The Inhibition of Microbial Adherence to the Inert Substratum
To determine the potential of R. officinalis extract to inhibit the ability of fungal and bacterial strains to adhere to the inert substrate, we selected six fungal and three bacterial strains which, according to previous studies, have shown a high capacity of biofilm development on the surface of heritage objects [19]. For this purpose, the extract and the solvent were diluted to final concentrations corresponding to MIC, MIC/2 and MIC/4. The R. officinalis extract with a concentration equivalent to the MIC value or concentrations lower than the MIC value has also been able to inhibit the adhesion to the inert substrate of the most tested fungal strains ( Figure 5). Apart from the P. chrysogenum S2042 strain, in most of the cases, different concentrations of the extract proved to be significantly more efficient than the solvent (p < 0.05). The best results were obtained against P. chrysogenum MTR3 and S1787strains isolated from wooden (p < 0.001) or textiles objects (p < 0.01) for all tested concentrations ( Figure 5).

The Inhibition of Microbial Adherence to the Inert Substratum
To determine the potential of R. officinalis extract to inhibit the ability of fungal and bacterial strains to adhere to the inert substrate, we selected six fungal and three bacterial strains which, according to previous studies, have shown a high capacity of biofilm development on the surface of heritage objects [19]. For this purpose, the extract and the solvent were diluted to final concentrations corresponding to MIC, MIC/2 and MIC/4.
The R. officinalis extract with a concentration equivalent to the MIC value or concentrations lower than the MIC value has also been able to inhibit the adhesion to the inert substrate of the most tested fungal strains ( Figure 5). Apart from the P. chrysogenum S2042 strain, in most of the cases, different concentrations of the extract proved to be significantly more efficient than the solvent (p < 0.05). The best results were obtained against P. chrysogenum MTR3 and S1787strains isolated from wooden (p < 0.001) or textiles objects (p < 0.01) for all tested concentrations ( Figure 5). According to Figure 6, the chemical compounds from the composition of R. officinalis have also inhibited by different degrees the biofilm developed by the three selected bacterial strains, as revealed by the low MBEC values, the anti-biofilm effect depending on the bacterial species and the tested concentration. This might possibly be explained by the different composition of biofilm matrix formed by different bacterial strains, influencing the rate of anti-biofilm compounds diffusion inside the bacterial biofilm. Thus, for B. pumilus, the best anti-biofilm activity was recorded at MIC/2, followed by MIC/4, for B. thuringiensis at the MIC value and for P. koreensis, all tested concentrations exhibited an antiadherence effect, although not statistically significant as compared to the solvent. According to Figure 6, the chemical compounds from the composition of R. officinalis have also inhibited by different degrees the biofilm developed by the three selected bacterial strains, as revealed by the low MBEC values, the anti-biofilm effect depending on the bacterial species and the tested concentration. This might possibly be explained by the different composition of biofilm matrix formed by different bacterial strains, influencing the rate of anti-biofilm compounds diffusion inside the bacterial biofilm. Thus, for B. pumilus, the best anti-biofilm activity was recorded at MIC/2, followed by MIC/4, for B. thuringiensis at the MIC value and for P. koreensis, all tested concentrations exhibited an anti-adherence effect, although not statistically significant as compared to the solvent.
The anti-biofilm effect seems to correlate, at least in some cases with the inhibition of the NO release (Table 4). The anti-biofilm effect seems to correlate, at least in some cases with the inhibition of the NO release (Table 4).
Starting from these results, strains producing different soluble factors were further selected to assess the influence of the rosemary extract on the production of each of them. Regarding the production of cellulases, no significant inhibition was noticed for any of the tested strains. Regarding caseinases, the enzymatic production was significantly
Starting from these results, strains producing different soluble factors were further selected to assess the influence of the rosemary extract on the production of each of them. Regarding the production of cellulases, no significant inhibition was noticed for any of the tested strains. Regarding caseinases, the enzymatic production was significantly inhibited by the extract in comparison to solvent in the cases of the P. chrysogenum strain isolated from textiles and the A. nidulans strain isolated from stone churches (p < 0.05). The production of amylase was significantly inhibited in the case of A. nidulans strain from stone churches and of A. montevidensis from textiles (Table 5 (a)).     Thus, the impact of the extract on the biodegradative enzymes production differs depending on the studied strain.

R. officinalis
In the case of the five fungal strains producing organic acids, when they are cultivated in the presence of R. officinalis extract, the production and secretion capacity of these types of compounds increases (the obtained values being higher than 100%). However, the stimulating role seems to be due to the solvent used for extraction and not to the extract itself. This conclusion is sustained by the fact that, in the case of the P. chrysogenum strain isolated from wooden churches, the presence of the extract diminishes the stimulating effect of the solvent from 147.62% to 104.76% (p < 0.001) (Table 5 (b); Figure S8).
The impact of R. officinalis extract on the enzymatic activity was tested on three bacterial strains belonging to the genus Bacillus. The presence of R. officinalis extract does not significantly influence the ability of bacterial strains to produce caseinases or amylases. The general effect was the stimulation of the production of these enzymes but the effect is rather due to the solvent, excepting the B. cereus strain isolated from a museum object, for which the presence of chemical components in the structure of the extract decreases the stimulatory effect of the solvent (p < 0.05) (Table 5 (a)).

Discussion
The present study is part of an important current research direction of our group, aiming at the development of effective strategies to combat deteriogenic microorganisms and preserve cultural heritage objects in Romania. According to the latest research on the restoration and conservation of heritage objects, the use of biocides based on chemicals of natural origin, to the detriment of biocides obtained by chemical synthesis, has a number of advantages, e.g., diminished risk of side effects for human and environmental health [34]. Several studies have shown that naturally occurring compounds have antimicrobial and antioxidant activities that depend on several factors: organic/inorganic artifacts; indoor/outdoor environments; bacterial or fungal colonizers; the biological activity of different solutions; and the mechanisms of action [35]. Most of these have focused on the development of biocides based on EOs [36,37] and less on hydro-alcoholic plant extracts. Even so, it has been proven that various extracts derived from plants such as A. sativum [21], Clinopodium nepeta [38] or Robinia pseudoacacia [39] can be efficient for treating different cultural heritage objects.
In this context, our study aimed to demonstrate and characterize the efficiency of R. officinalis hydro-alcoholic extract to inhibit the growth of deteriogenic fungal and bacterial strains previously isolated from cultural heritage objects and buildings from Romania. In order to achieve the proposed objective, the first stage consisted in obtaining the hydro alcoholic extract. The chosen extraction method was based on the study conducted by Švarc-Gajić et al., in 2013 [40], where the polyphenol content differs depending on the solvent used. Higher phenol yields were obtained by using methanol, which is the most widely used solvent in the literature [41,42] but with a strong toxic effect, so it is not recommended to use [43]. The lowest extraction yield was obtained for ethyl acetate, even if the dipole moment is not very different from methanol (1.78 and 1.69, respectively) [40]. This effect can be explained by the much higher dielectric loss coefficients observed in methanol than in ethyl acetate [44]. The next high phenol content was obtained using ethanol; thus, we have selected this solvent and the extraction was finally made in 50% ethanol. The next step was to evaluate the antimicrobial efficacy of the obtained extract against 41 fungal strains and 15 bacterial strains isolated from the surface of cultural heritage objects and buildings in Romania. The antimicrobial activity of R. officinalis extract differs depending on the isolation substrate and the microbial species. The best results were obtained for Gram-positive bacterial strains such as A. globiformis (MIC = 0.78 mg/mL) or B. cereus, B. thuringiensis and A. aurascens (MIC = 1.56 mg/mL), most likely due to the chemical composition of the cell wall and membrane, while in the case of fungi the most sensitive strains belong to A. clavatus species (MIC = 1.2 mg/mL). Further, a study confirmed that R. officinalis extracts are more efficient against Gram-positive compared to Gram-negative bacteria, mostly due to the presence of carnosic acid, carnosol and rosmarinic acid [45].
The antimicrobial activity of R. officinalis hydro-alcoholic extract against both fungal and bacterial strains seem to be correlated with the presence of hesperidin (10.39 mg/L), vanillic acid (7.51 mg/L) and quercetin (7.10 mg/mL). These compounds have a large spectrum of antimicrobial activity; for example, in Citrus spp., hesperidin was considered the most active compound against Staphylococcus aureus, Escherichia coli and B. subtilis [46].
Hesperidin is a strong antifungal and is active against A. parasiticus, A. flavus, F. semitectum and P. expansum [47]. Vanillic acid exhibits antimicrobial activity against Enterobacter cloacae and E. hormaechei strains; further, it induces a delay in the development of Trichoderma spp., Athelia rolfsii, Phaeoacremonium parasiticum, A. niger and F. oxysporum [48][49][50]. The R. officinalis extract also contains other derivatives of caffeic acid, that could act as ligands for metal ions; thus, they react with peroxide radicals and stabilize free radicals. The catechol group is responsible for radical electron capturing formed as a result of oxidation. The skeleton formed by the three rings allows the charge relocation. The presence of the carboxylic group increases this conjugation, especially in aqueous systems. However, in slightly polar environments, such as fats, the structure of lactone seems to provide greater stability [51].
Although a limited number of studies have focused on demonstrating the efficiency of R. officinalis extract against biodeteriogenic microorganisms, the antifungal activity of R. officinalis extracts (water and ethanol) against four fungal strains (A. flavus, A. versicolor, Penicillium spp., P. purpurogenum) isolated from different archaeological artefacts in an Egyptian Museum demonstrated the efficiency of both extracts [52]. Several other authors have demonstrated the antimicrobial activity of R. officinalis EO or of the crude extract against several food-borne pathogens, for i.e., against Macrophphaseolina phaseolina to control charcoal rot in Glycine max [53]. The R. officinalis showed strong antifungal activity by its disruption of the cell wall and the loss of cellular components and by its inhibition of fumonisin B1 and B2 production in a F. verticillioides Sacc. strain previously isolated from corn residue used in animal feed (MIC = 150 µg/mL) [54].
The role of antioxidants in an artifact's preservation depends on their physico-chemical nature and possible oxidation reactions that may occur on the surface depending on environmental conditions. In the case of woody artifacts, it was found that the initial stage of wood degradation by fungi occurs through oxidative reactions. Free radical species are involved in the initial stage of wood degradation by fungi. To increase the pore size that facilitates the penetration of extracellular enzymes, the wood treated with an antioxidant becomes more resistant to damages caused by cellulase-producing fungi [55]. In the case of paintings and textiles, it has been found that with the passage of time the flavonoids from the paint's compositions are transformed into phenolic acids, thus suffering an oxidative degradation. Thus, the use of natural antioxidants could contribute to a better conservation of these types of cultural heritage [56].
The establishment of surface-associated assemblages of microorganisms on different heritage objects is an essential step in the biodeteriogenic process, providing them enhanced resistance to physical and chemical stress [57]. Thus, a better understanding of the processes that allow cells to attach to different surfaces could allow the clarification of the relevance of biofilm formation and subsequently induced biodeterioration. Although it is of great importance to develop biocidal solutions for the decontamination of heritage objects, the prevention of secondary contamination brings a special economic advantage. For this reason, the R. officinalis extract has also been tested as a solution to prevent the adherence of deteriogenic microbial strains to the inert substrate. Our study reports for the first time in the literature the anti-adherent effect of R. officinalis hydroalcoholic extract against biodeteriogenic strains. The results show that the anti-adherence effect of R. officinalis extract occurs at different concentrations depending on the tested fungal or bacterial strain. No specific correlation with the isolation source or species could be noticed. However, the R. officinalis extract has shown a good anti-adherent activity against P. chrysogenum strains isolated from wooden or textile materials or B. thuringiensis strains.
The high efficiency of R. officinalis extract against oral planktonic bacteria recommended it as an adjunct in the treatment and prevention of biofilms in the oral cavity [58]. Further, hesperidin impedes the adherence of S. aureus to inert substratum by a molecular mechanism involving the down regulation of genes such as: polysaccharide intracellular adhesion gene (icaA and icaD), fibronectin-binding protein coding gene (fnbA and fnbB) and staphyloxanthin production (crtM) [59]. Vanillic acid also acts at the molecular level, reducing the expression of genes involved in biofilm development in the case of Vibrio alginolyticus strains [60].
Intracellular reactive species play critical roles in microbial vitality, including stress management and differentiation [61]. Among all reactive species, NO has been considered a universal signaling molecule in most organisms, including microorganisms. However, for fungi, the synthesis and functions of NO have not been fully elucidated. In most fungi, endogenous NO generation has been indirectly revealed by identifying NOS activity or using NO absorbers. Few studies have directly shown the presence of intracellular NO [62][63][64].
It has been found that oxidative and nitrosative stress could occur inside biofilms, affecting their development and maturation under various conditions [65]. The production of ROS and RNI and accumulation of these radical oxidants in the extracellular environment could affect the biofilm matrix. Low, non-toxic endogenous NO concentrations, could also induce the dispersion of the biofilm [66]. In bacteria, endogenous NO can be generated from L-arginine by bacterial NOS. Alternatively, NO can be produced by NOS-independent pathways, such as the reduction of nitrites to NO by nitrite reductases in denitrifying bacteria such as Bacillus spp., Serratia spp. and Pseudomonas spp. [67]. In the case of bacteria, the production of NO is closely associated with the formation and dispersion of bacterial biofilm [66]. Low NO concentrations (25 pM to 2.5 µM NO) inhibited biofilm formation, those above 2.5 µM NO, promoted biofilm and concentrations greater than 1 mM NO had a bactericidal effect [68]. Low NO concentrations regulate a widely conserved genetic pathway that controls the conversion between planktonic and sessile lifestyle through the action of c-di-GMP [69]. A consequence of NO-mediated dispersal is that, due to the reversal of the genetic program, both the biofilm and the dispersed cells lose their antimicrobial resistance [70]. The effect of NO on the inhibition and dispersal of biofilm has been studied in both Gram-negative and Gram-positive bacteria, such as Serratia marcescens, Vibrio cholerae, E. coli, Fusobacterium nucleatum, B. licheniformis and S. epidermidis.
The method proposed in our study allows quantification of the NO content in a liquid medium without distinguishing between specific sites or structures in which the NO is produced or accumulated. Endogenous NO has been detected in fungi during the formation of conidia on solid medium and the branching of hyphae in liquid medium by a fluorescence method [62]. In addition, high fluorescence detected inside vacuole-like structures in fungal hyphae may indicate the subcellular site of NO synthesis or may be the result of fluorogenic compound accumulation in vacuoles. Moreover, although the highest NO level was observed during the formation of microconidia, fluorescence was largely detected in vegetative hyphae, not in phialides and microconidia. Vacuoles, in general, act as a reservoir for the storage of basic amino acids, such as arginine, in filamentous fungi [71]. Thus, NOS-dependent NO synthesis may be responsible for the increased NO concentration.
The successive colonization of heritage objects and the action of secondary metabolites produced by biodeteriogenic microorganisms lead to the loss of the quality of heritage objects, affecting both their integrity and appearance. The main enzymes involved in biodeterioration are esterases, lipases, proteases, cellulases, glucanases or laccase. Depending on the material from which the heritage objects are made, the relevance of the enzymes involved in biodeterioration is different. For paintings or lacquered church objects, esterases or lipases are the most aggressive. Meanwhile, for wooden or paper objects [72], cellulases and proteases have the most harmful effect [72,73]. Similarly, in the case of stone objects or derivatives the organic acids have a significant impact [74].
The microbial strains included in this study presented various characteristics regarding the production of enzymes involved in the biodeterioration. Thus, for each type of enzyme, several representative microbial strains were selected to determine whether R. officinalis extract influences their ability to produce and secrete the enzyme of interest in the culture medium. According to the obtained results, the treatment of microbial cultures with sub inhibitory concentrations of extract or solvent leads either to the inhibition of the enzymatic activity or to its stimulation, the results being different depending on the studied strain, the type of tested enzyme or the isolation source. From a statistical point of view, in most situations it is observed that the solvent is that which determines the modification of the enzymatic activity. There are few exceptions to this, in which the tested extract inhibited the capacity of A. nidulans (stone church) to secrete amylases or caseinases, of A. montevidensis from textiles to secrete amylases, of P. chrysogenum isolated wooden church to produce organic acids or of the B. cereus strain isolated from a museum object to produce esterases.
Although not tested in biodeterioration studies, chlorogenic acid, an important compound present in the R. officinalis extract has been shown to inhibit carbohydrates-hydrolyzing enzymes such as amylases and glucosidases being currently investigated as hypoglycaemic adjuvant [75]. Other phenolic compounds such as p-coumaric acid, vanillic acid, syringing acid, benzoic acid, and ferulic acid are known for their ability to inhibit the production of enzymes such as laccase (that might be involved in textile discoloration), as demonstrated for Botryosphaeria genera [76,77]. Further, it has been shown that caffeic acid, p-coumaric acid, and ferulic inhibit protease, cellulase and esterase production in F. oxysporum, while p-coumaric acid and ferulic acid strongly decrease cellulase activity in the case of T. reesei strains [78,79]. Similarly, rutin, quercetin, apigenin or kaempherol inhibit cellulase production in F. solani [80]. Green alternatives used for heritage protection may refer to the prevention of microbial contamination or the development of control strategies that are based only on inhibiting maturation or dispersal of biodeteriogenic microbial strains [34]. Microbial contamination prevention can be achieved either by treating the affected area with specific biocidal compounds or by exposing the heritage objects to inhibitory vapours existing in the repository environment. For example, essential oils with a high content of volatile antimicrobial compounds can be used to prevent biodeterioration without being directly applied to the surface of the affected cultural heritage objects [81]; however, exposure to the vapours can cause significant damage to the structure of the heritage object. Such a situation has been described for thymol, the main chemical component of thyme essential oil which was used frequently in the past in so-called "thymol cabinets". Although it has great biomedical potential (having antimicrobial, anti-inflammatory or antioxidant properties) [82] it cannot be used for heritage protection since it softens the varnishes and resins, renders parchment brittle and degrades paper objects as well iron gall ink or watercolour binders [34,83]. Therefore, an important step for describing new biocidal solutions to be used for biodeterioration control or prevention is to determine the threshold above which the solution is effective against microbial strains but at the same time does not imply any risk for the object itself. Regarding rosemary extract, there is a limited number of studies concerning its possible use as an ecological alternative for heritage protection. Even so, Di Vito et al., have stipulated that the rosemary extract can be included in gellan hydrogel to obtain a cleaning agent for paper artwork with no negative impact on heritage object structure or integrity [84]. Taking into account that the antimicrobial efficiency of the plant extract proven by this work is sufficiently high and that the minimum concentration of plant extract necessary to inhibit the growth of microbial strains, their adhesion to the substrate and their enzymatic activity being relatively low, it can be recommended to proceed further with future studies for testing the effectiveness of rosemary extract or polymeric matrices containing rosemary extract as antimicrobial agents. Further studies should also include their impact on different substratum models such as exterior deteriorating wood, stone, mural paintings or paper by SEM, ATR-FTIR or gravimetrical methods.

Microwave Assisted Extraction (MAE)
A microwave-assisted extraction (MAE) procedure was carried out for the extraction of R. officinalis bioactive compounds using a microwave extractor (Ethos SEL, Milestone, Brondby, Denmark). For extraction, plant material powder was mixed with selected solvents (ethanol: ultrapure water = 50:50) at a ratio of 1:10 (w/v) and underwent MAE for 1 h at 100 • C. After the MAE procedure, the R. officinalis extract was filtered with Whatman filter paper No.1, collected in a volumetric flask and stored refrigerated at 4 • C until further use.

Total Flavonoids Content (TFC)
TFC was determined by the aluminium chloride method [85]. Specifically, 0.1 mL sample/standard with 0.1 mL of 10% sodium acetate. An amount of 0.12 mL of 2.5% AlCl 3 solution was added to this mixture and made up to a final volume of 1 mL with 70% ethanol. After stirring, this solution was left for 45 min in the dark and the absorbance of the mixture was read at 430 nm. The calibration curve was performed for quercetin concentrations in the range of 10-250 µg/mL (R 2 = 0.9968).

Determination of Reducing Sugar Content (RSC)
RSC was determined using the 3,5-dinitrosalicylic acid (DNSA) method, as previously described with slight modification [86]. For the measurement, 330 µL of DNSA reagent was pipetted into a test tube containing 170 µL of plant extract and kept at 90 • C for 15 min. After cooling, a volume of 150 µL was pipetted from each reaction mixture in 96-well plates and the absorbance of the resulting solution was measured at 540 nm using a UV-VIS spectrophotometer. The reducing sugar content was calculated from the calibration curve of standard D-glucose (0.1-1 mg/mL) (R 2 = 0.9972).

Characterization of the Extract by UHPLC-MS/MS
Polyphenolic compounds quantification was performed using a high-resolution Q Exactive mass spectrometer™ Focus Hybrid Quadrupole-OrbiTrap (Thermo Fisher Scientific, Waltham, MA, USA) equipped with HESI, coupled to a high-performance liquid chromatograph UltiMate 3000 UHPLC (Thermo Fisher Scientific). Chromatographic separation was performed on a Kinetex ® C18 column (100 × 2.1 mm, 1.7 µm particle diameter) at 30 • C, under a gradient elution of two mobile phases: A-water with 0.1% formic acid and B-methanol with 0.1% formic acid, at a flow rate between 0.3 and 0.4 mL/min, as presented by [87]. Full scan data in negative mode covering a scan range of m/z 75-1000 was acquired at a resolving power of 70,000 FWHM at m/z 200, while variable dataindependent analysis MS2 (vDIA) was performed at the resolution of 35,000, isolation windows and scan ranges being set as follows: 75-205 m/z, 195-305 m/z, 295-405 m/z, 395-505 m/z and 495-1000 m/z. Nitrogen was used as collision gas and auxiliary gas at a flow rate of 11 and 48 arbitrary units, respectively. The applied voltage was 2.5 kV, the capillary temperature was 320 • C and the energy of the collision was set as 30 eV. Polyphenolic compound quantification and calibration was performed in the concentration range between 0 and 1000 µg/L for each of the phenolic acids and flavonoids by serial dilution with methanol of the standard mixture of concentration 10 mg/L. Data processing, analysis and interpretation were performed with Xcalibur software package (Version 4.1) and Compound Discoverer v. 2.1 software (Thermo Fisher Scientific) using an untargeted metabolomics working template.

Antioxidant Activity
The measurement of DPPH assay was performed by a method adapted from [88].
To 100 µL of sample/standard was added 100 µL of 0.3 mM DPPH radical solution. The absorbance was read at λ = 517 nm after 30 min of incubation in the dark. The calibration curve was achieved for concentrations between 80-5 µM Trolox/mL (R 2 = 0.9988). The results were expressed in mM Trolox equivalent/mL extract.
The CUPRAC method is based on the reduction of a cupric complex, neocuproin, by antioxidants in copper form. Copper ion reduction was performed as previously described [89]: 60 µL of sample/standard solutions of different concentrations were mixed with 50 µL CuCl 2 (10 mM), 50 µL neocuproin (7.5 mM), and 50 µL ammonium acetate buffer 1 M, pH = 7.00. After 30 min, the absorbance was measured at 450 nm. The stock Trolox solutions required for the calibration curve were 2 mM, and the working concentrations were between 0.24 and 2.0 mM Trolox/mL (R 2 = 0.9991). The results were expressed in mM Trolox equivalent/mL extract.
The determination of the antioxidant capacity of iron reduction was performed by the FRAP assay method [90]. The stock solutions included 300 mM acetate buffer, pH 3.6, 10 mM 2,4,6-tripyridyl-s-triazine (TPTZ) solution in 40 mM HCl, and 20 mM FeCl 3 6H 2 O solution in a volume ratio of 10:1:2, which were warmed to 37 • C before use. After incubation, the absorbance was read at 593 nm. A 1 mM Trolox stock solution was used to plot the calibration curve, the concentration ranging between 30 and 250 µM Trolox/mL (R 2 = 0.9971). The results were expressed in mM Trolox equivalent/mL of extract.
Trolox equivalent antioxidant capacity (TEAC) assay was performed according to Re et al., [91] with few modifications. A stable stock solution of ABTS+ was produced by mixing a solution of 7 mM ABTS in 2.45 mM potassium persulphate. Then, the mixture was left standing in the dark at room temperature for 12-16 h before use. An ABTS+ working solution was obtained by dilution with ethanol to an absorbance of around 0.70. The reaction mixture consisted in 20 µL of sample/standard and 180 µL of ABTS+ working solution and was incubated 30 min in the dark. The standard curve was linear between 20 and 200 µM Trolox (R 2 = 0.9981). The results were expressed in mM Trolox equivalent/mL extract.  [19][20][21]. After 5-7 days of incubation at room temperature/24 h at 37 • C the growth inhibition diameters were measured and converted in arbitrary units as previously described [21] The influence of R. officinalis on the microbial adherence ability to the inert substratum (96-well plate, untreated polystyrene) was measured at 490 nm after running the quantitative analysis of the antimicrobial activity through the microtiter method [93]. The minimum biofilm eradication concentration on inert substrate (MBEC) was calculated as follows:

Antimicrobial Activity
where: As = the absorbance of the microbial adherence treated with R. officinalis extract/solvent and Ac = the absorbance of the microbial adherence untreated.

Extracellular NO Release
Quantification of extracellular nitric oxide (NO) release was undertaken for MIC, MIC/2 and MIC/4. NO was rapidly converted to nitrite in aqueous solutions and therefore total nitrite can be used as an indicator of NO concentration. This was measured using a total nitrite spectrophotometric analysis with Griess reagent as previously described with some modifications [65,94]. To the microbial supernatant obtained after 24 h incubation for bacteria and eight days for fungi (50 µL), 50 µL of 2% sulphanilamide in 5% (v/v) H 3 PO 4 and 50 µL of 0.13% N-(1-naphthyl)-ethylenediamine aqueous solution were added. Azo dye was measured after 30 min at λ = 540 nm. For the quantification of nitric oxide, a calibration curve with different NaNO 2 concentrations in the range of 5-100 µM was performed (R 2 = 0.9998). The results were expressed according to the following equation: where: A sample-the absorbance of MIC/2 extract/solvent A blank-the absorbance of culture media (RPMI or MH broth) A control-the absorbance of culture media inoculated 4.5.5. The Modulation of Enzymatic Activity by R. officinalis Hydro-Alcoholic Extract Microbial strains grown treated with subinhibitory concentration of hydro-alcoholic extract (MIC/4) and solvent control were evaluated for their capacity to secrete organic acids and enzymes involved in the biodegradation process (cellulase, phenoloxidase, esterase, caseinase and amylase) as previously described [21,81,95,96]. Briefly, different culture media Sabouraud (Sab) supplemented with cellulose and Congo red for cellulase production was highlighted as a clear zone around the colonies; Sab malt agar with tannic acid for phenoloxidase activity was revealed as a brown zone on the colony's reverse; Sab peptone agar with Tween 20/Tween 80 for esterase detection was highlighted as a precipitation zone on the colony's reverse; Sab agar supplemented with milk for caseinase detection was revealed by a white-yellow precipitate of calcium para-caseinate around the colonies; Sab agar supplemented with rice starch and hydrolysis was highlighted by flooding the plate with Lugol solution (yellow ring around the culture spot, while the rest of the medium turns blue); and Sab agar supplemented with sodium nitrate, dipotassium phosphate, potassium chloride, glucose, and phenol red for acid production was revealed by a colour change from red to yellow near the inoculum that was spotted with 10 µL suspension adjusted to 1/0.5 McFarland (fungal/bacterial strains) prepared from each microbial strain (untreated culture, serving as growth control and treated with R. officinalis) and then incubated for 5-8 days at 26-28 • C.
The influence of R. officinalis extract and solvent control on cellulase, phenoloxidase, esterase, caseinase, amylase and organic acid production was evaluated by using the following relation: where: C1-colony diameter of strain control, D1-clear/brown/white/yellow precipitate/ring surrounding the culture spot/yellow zone diameter of strain control, C2-colony diameter of sample, D2-clear/brown/precipitation/yellow zone diameter of sample.

Statistical Analysis
Data were expressed as means ± SD determined by triplicate analysis. The statistical analysis was conducted using GraphPad Prism 9. Data were analysed using ordinary twoway ANOVA with two-stage linear step-up procedure of Benjamini, Krieger and Yekutieli, with individual variances computed for comparison between extract and solvent biological activities. The level of significance was set to p < 0.05.

Conclusions
To the best of our knowledge, this is the first study conducted to demonstrate the efficiency of R. officinalis extract against deteriogenic strains isolated from the surface of cultural heritage objects and buildings. The antimicrobial activity of R. officinalis extract differs depending on the species and substrate, the best results being obtained for the Grampositive bacterial strains and for the A. clavatus fungal species. The antimicrobial activity of R. officinalis hydro-alcoholic extract against fungal and bacterial strain was associated with the presence of hesperidin, vanillic acid and quercetin. The R. officinalis extract exhibited anti-biofilm activity, correlating with the release of the NO stress signalling molecule against P. chrysogenum strains isolated from wooden or textile materials or bacteria belonging to B. thuringiensis species. The inhibitory concentrations of extract or solvent modulated the enzymatic and organic acids profile of the tested biodeteriogenic strains. Taken together, these results recommend the R. officinalis hydro-alcoholic extract for further studies (with an emphasis on determining its impact on the heritage object itself) for the development of natural biocides used in cultural heritage conservation.