Exploring the Co-Crystallization of Kojic Acid with Silver(I), Copper(II), Zinc(II), and Gallium(III) for Potential Antibacterial Applications

Co-crystallization of kojic acid (HKA) with silver(I), copper(II), zinc(II), or gallium(III) salts yielded three 1D coordination polymers and one 0D complex in which kojic acid was present as a neutral or anionic terminal or bridging ligand. All reactions were conducted mechanochemically via ball milling and manual grinding, or via slurry. All solids were fully characterized via single-crystal and/or powder X-ray diffraction. As kojic acid is a mild antimicrobial compound that is widely used in cosmetics, and the metal cations possess antibacterial properties, their combinations were tested for potential antibacterial applications. The minimal inhibition concentrations (MICs) and minimal biocidal concentrations (MBCs) for all compounds were measured against standard strains of the bacteria P. aeruginosa, S. aureus, and E. coli. All compounds exerted appreciable antimicrobial activity in the order of silver, zinc, copper, and gallium complexes.


Introduction
The search for new active antibacterial compounds is attracting increasing interest as the current clinical pipeline remains inadequate to meet the challenges of the growing emergence and spread of antimicrobial resistance [1]. According to the World Health Organization (WHO), the clinical benefit of newly approved products is limited compared to existing treatments, and most of the antimicrobials under clinical development are merely derivatives of molecules that are already known and used on the market, and bacteria have multiple resistance mechanisms that are well-established in the community [1]. Therefore, there is an urgent need to develop new antibacterial drugs. Historically, an increasing number of metal complexes have shown interesting biological and biomedical properties, such as the arsenic-containing organometallic complex for the treatment of syphilis (Salvarsan) [2], the mercuric-based antiseptic agent (Mercurochrome), and the gold complex agent for the treatment of rheumatoid arthritis (Auranofin) [3]. The metal complexes have a wide diversity of metals, ligand types, and geometries, which is an inbuilt advantage for accessing a highly undeveloped chemical space for drug development, especially for designing new antibacterial agents.
Kojic acid (HKA, see Scheme 1) is a common fungal metabolite that was first isolated from fungus. It can be used as a preservative for cosmetics, without one-time or cumulative irritation to the skin, and in a variety of foods, where it is added to prevent enzymatic browning [4]. It can strongly absorb ultraviolet rays and can be used alone or in combination with various sunscreen products and soaps [5]. It also can treat and prevent the formation of skin pigmentation, such as liver spots [6]. HKA also shows antibacterial properties, Silver and its compounds have been widely used as antimicrobial agents [8], for example silver nitrate for the treatment of burn wounds [9] and silver sulfadiazine (Silvadene) as a broad-spectrum antibiotic for burn wounds [10]. The antimicrobial activities of zinc work via direct interactions with microbial membranes, leading to membrane instability and enhanced permeability [11] or via interactions with nucleic acids and the inactivation of respiratory enzymes [12]. In 1931, Levaditi et al. discovered the antibacterial properties of gallium by eradicating syphilis in rabbits and Trypanosoma in mice through the use of gallium tartrate [13]. Since then, various gallium compounds have advanced in preclinical and clinical investigations, such as those studying the potent suppressive activity of GaCl3 and Ga(III)-citrate on P. aeruginosa biofilm formation in vitro and in murine lung infection models [14]. Based on the toxic effects of free copper ions on both bacteria and fungi [15], the complexes conjoining organic molecules and copper were further studied in order to improve the antimicrobial activity; for example, copper complexes with sulfonamide ligands were synthesized for their antimicrobial properties [16], and copper complexes of 5-aminobenzofuran-2-carboxylate Schiff-base ligands have shown excellent antibacterial activity against Mycobacteria Tuberculosis [17].
Our approach is based on the application of crystal engineering strategies to combine, hopefully in a synergistic way, the antimicrobial properties of organic molecules in the solid state with metal complexes to construct co-crystals and complexes. The preparation of co-crystals of known drugs, as well as the quest for new combinations of natural antibacterials, have been shown to be successful in a number of cases [18]. We have also contributed a series of papers in which we have described how these methods have been applied to inhibit enzymatic activity and to enhance the properties of known antibiotics [19,20].
In the present work, we explored the potential antibacterial properties of kojic acid in combination with metal cations by co-crystallizing kojic acid with silver(I), copper(II), zinc(II) and gallium(III) salts. Kojic acid, in combination with metal cations, has been used as a tyrosinase inhibitor [21,22] and as an anticancer agent [23]. The choice of copper and zinc was dictated by the fact that these are essential metal ions; therefore, they have a minor impact on the environment. Silver nitrate is a common antibacterial agent, and it is usually employed in healthcare to treat surfaces [9], but several bacteria have already developed resistance to silver ions [24][25][26]. Gallium salts are toxic, but, as mentioned above, they are already employed in specific treatments, and we are interested in exploring their possible antibacterial effects when used in stable combinations with organic ligands, as we have explored in our recent work [27,28].
Silver and its compounds have been widely used as antimicrobial agents [8], for example silver nitrate for the treatment of burn wounds [9] and silver sulfadiazine (Silvadene) as a broad-spectrum antibiotic for burn wounds [10]. The antimicrobial activities of zinc work via direct interactions with microbial membranes, leading to membrane instability and enhanced permeability [11] or via interactions with nucleic acids and the inactivation of respiratory enzymes [12]. In 1931, Levaditi et al. discovered the antibacterial properties of gallium by eradicating syphilis in rabbits and Trypanosoma in mice through the use of gallium tartrate [13]. Since then, various gallium compounds have advanced in preclinical and clinical investigations, such as those studying the potent suppressive activity of GaCl 3 and Ga(III)-citrate on P. aeruginosa biofilm formation in vitro and in murine lung infection models [14]. Based on the toxic effects of free copper ions on both bacteria and fungi [15], the complexes conjoining organic molecules and copper were further studied in order to improve the antimicrobial activity; for example, copper complexes with sulfonamide ligands were synthesized for their antimicrobial properties [16], and copper complexes of 5-aminobenzofuran-2-carboxylate Schiff-base ligands have shown excellent antibacterial activity against Mycobacteria tuberculosis [17].
Our approach is based on the application of crystal engineering strategies to combine, hopefully in a synergistic way, the antimicrobial properties of organic molecules in the solid state with metal complexes to construct co-crystals and complexes. The preparation of cocrystals of known drugs, as well as the quest for new combinations of natural antibacterials, have been shown to be successful in a number of cases [18]. We have also contributed a series of papers in which we have described how these methods have been applied to inhibit enzymatic activity and to enhance the properties of known antibiotics [19,20].
In the present work, we explored the potential antibacterial properties of kojic acid in combination with metal cations by co-crystallizing kojic acid with silver(I), copper(II), zinc(II) and gallium(III) salts. Kojic acid, in combination with metal cations, has been used as a tyrosinase inhibitor [21,22] and as an anticancer agent [23]. The choice of copper and zinc was dictated by the fact that these are essential metal ions; therefore, they have a minor impact on the environment. Silver nitrate is a common antibacterial agent, and it is usually employed in healthcare to treat surfaces [9], but several bacteria have already developed resistance to silver ions [24][25][26]. Gallium salts are toxic, but, as mentioned above, they are already employed in specific treatments, and we are interested in exploring their possible antibacterial effects when used in stable combinations with organic ligands, as we have explored in our recent work [27,28].
In the following sections, we discuss the reaction of kojic acid with silver(I), copper(II), zinc(II), and gallium(III) salts that afforded the 1D polymers Ag(HKA)(NO 3 )·H 2 O, Cu(KA) 2 , and Zn(KA) 2, and the 0D complex [Ga(KA) 2 (OH 2 ) 2 ][NO 3 ]·H 2 O, respectively, as shown in Scheme 2a-d. All products were characterized via X-ray diffraction, and their antimicrobial efficacy was tested against two Gram-negative organisms, E. coli and P. aeruginosa, and towards the Gram-positive strain S. aureus.

[Ag(HKA)(NO3)]•H2O
The results of ball milling, manual grinding, and crystallization from a solution of HKA and AgNO3 are represented in Scheme 3. While the slurry method resulted in a physical mixture of unreacted starting materials, ball milling and the subsequent crystallization in water/acetone resulted in the formation of the [Ag(HKA)(NO3)]•H2O complex (see Figure 1).

[Ag(HKA)(NO 3 )]·H 2 O
The results of ball milling, manual grinding, and crystallization from a solution of HKA and AgNO 3 are represented in Scheme 3. While the slurry method resulted in a physical mixture of unreacted starting materials, ball milling and the subsequent crystallization in water/acetone resulted in the formation of the [Ag(HKA)(NO 3 )]·H 2 O complex (see Figure 1).

[Ag(HKA)(NO3)]•H2O
The results of ball milling, manual grinding, and crystallization from a solution of HKA and AgNO3 are represented in Scheme 3. While the slurry method resulted in a physical mixture of unreacted starting materials, ball milling and the subsequent crystallization in water/acetone resulted in the formation of the [Ag(HKA)(NO3)]•H2O complex (see Figure 1).  The powder diffraction pattern for [Ag(HKA)(NO3)]•H2O (see Figure S1 in the Supplementary Materials), calculated on the basis of the single-crystal structure, matches the experimental ones, confirming that the single-crystal structure is representative of both the crystallization and ball milling bulk products. Of all the complexes obtained in this work, this was the only one in which the kojic acid was still protonated, i.e., the complex can be described as a hydrated co-crystal of AgNO3 and kojic acid. The coordination around the silver cations in crystalline [Ag(HKA)(NO3)]•H2O is shown in Figure 1a: each silver(I) is chelated by a kojic acid molecule, while a second and third HKA molecule act as terminal ligands via the carbonyl group (with Ag-O bond distances in the range 2.403-2.648(5) Å ); the nitrate ion also coordinates the cation via two oxygen atoms (Ag-ONO3 2.408(4) and 2.604(5) Å ). The HKA molecules bound to silver(I) via the carbonyl group are also chelating the neighbouring silver cations, as shown in Figure 1b, resulting in a 1D ribbon extending along the crystallographic a-axis (see the projections in Figure 1c,d), with the "Ag-O" backbone resembling an n-ladderane; the water molecules fill in the channels in between the 1D ribbons. The powder diffraction pattern for [Ag(HKA)(NO 3 )]·H 2 O (see Figure S1 in the Supplementary Materials), calculated on the basis of the single-crystal structure, matches the experimental ones, confirming that the single-crystal structure is representative of both the crystallization and ball milling bulk products. Of all the complexes obtained in this work, this was the only one in which the kojic acid was still protonated, i.e., the complex can be described as a hydrated co-crystal of AgNO 3 and kojic acid. The coordination around the silver cations in crystalline [Ag(HKA)(NO 3 Figure 1a: each silver(I) is chelated by a kojic acid molecule, while a second and third HKA molecule act as terminal ligands via the carbonyl group (with Ag-O bond distances in the range 2.403-2.648(5) Å); the nitrate ion also coordinates the cation via two oxygen atoms (Ag-O NO 3 2.408(4) and 2.604(5) Å). The HKA molecules bound to silver(I) via the carbonyl group are also chelating the neighbouring silver cations, as shown in Figure 1b, resulting in a 1D ribbon extending along the crystallographic a-axis (see the projections in Figure 1c,d), with the "Ag-O" backbone resembling an n-ladderane; the water molecules fill in the channels in between the 1D ribbons.

Cu(KA) 2
Ball milling and slurrying of HKA and Cu(NO 3 ) 2 in a 2:1 stoichiometric ratio afforded the complex Cu(KA) 2 (see Scheme 4). The powder diffraction patterns of the ball milling and slurrying products are shown in Figure S2 in the Supplementary Materials. Despite numerous attempts, no single crystal could be obtained due to poor solubility of the Cu(KA) 2 crystalline powder in all of the solvents used (water, methanol, ethanol, acetone, chloroform, DMF, and DMSO). The same product was also obtained by reacting HKA with CuCl 2 (see Scheme 4).

Cu(KA)2
Ball milling and slurrying of HKA and Cu(NO3)2 in a 2:1 stoichiometric ratio afforded the complex Cu(KA)2 (see Scheme 4). The powder diffraction patterns of the ball milling and slurrying products are shown in Figure S2 in the Supplementary Materials. Despite numerous attempts, no single crystal could be obtained due to poor solubility of the Cu(KA)2 crystalline powder in all of the solvents used (water, methanol, ethanol, acetone, chloroform, DMF, and DMSO). The same product was also obtained by reacting HKA with CuCl2 (see Scheme 4).
The powder diffraction patterns for the bulk products and the one calculated on the basis of single-crystal data are compared in Figure S3 in the Supplementary Materials. While a period of 60 min of ball milling (see also Section 3) was not sufficient for the reaction to occur quantitatively, peaks of unreacted material were almost negligible in the product of the slurry process. The crystal structure of Zn(KA) 2 is shown in Figure 3. Two kojiate anions bridge two Zn(II) as bidentate (Zn-O COO − distances: 2.05(2) and 2.06(2) Å) and monodentate (Zn-O OH distance: 2.21(2) Å) ligands, thus forming 1D polymers extending in the [101] crystallographic direction and parallel to each other in the crystal (Figure 3a,b); the zinc cations are located on centres of symmetry within octahedra of oxygen atoms (see Figure 3a). The six-membered rings are π-stacked within the 1D polymer, with a distance between the two mean planes of ca. 3.2 Å.

Zn(KA)2
Analogously to Cu(KA)2, Zn(KA)2 was obtained via slurry, crystallization from solution, and ball milling HKA and Zn(NO3)2 in a 1:1 stoichiometric ratio (Scheme 5). The powder diffraction patterns for the bulk products and the one calculated on the basis of single-crystal data are compared in Figure S3 in the Supplementary Materials. While a period of 60 minutes of ball milling (see also the Materials and Methods section) was not sufficient for the reaction to occur quantitatively, peaks of unreacted material

Zn(KA)2
Analogously to Cu(KA)2, Zn(KA)2 was obtained via slurry, crystallization from solution, and ball milling HKA and Zn(NO3)2 in a 1:1 stoichiometric ratio (Scheme 5). The powder diffraction patterns for the bulk products and the one calculated on the basis of single-crystal data are compared in Figure S3

[Ga(KA)2(OH2)2][NO3]•H2O
Ball milling, slurrying, and crystallization from the solution of HKA with Ga(NO3)3 in a 2:1 stoichiometric ratio resulted in the metal complex [Ga(KA)2(OH2)2][NO3]•H2O (see Scheme 6). The powder diffraction pattern for the complex, calculated on the basis of single crystal data, matches the one measured on the bulk products (see Figure S4). Differently from what observed for the silver nitrate co-crystal, when gallium nitrate is reacted with HKA the kojic acid is deprotonated, and acts as a bidentate ligand toward the gallium cation. Two KAanions coordinated the gallium cation (see Figure 4a), and two water molecules completed the octahedral coordination sphere while the nitrate anion was at hydrogen bonding distance from the water molecule that did not belong to the coordination sphere and from the -OH group on the KAanion. Wave layers of cations can be observed extending along the b-axis direction, with nitrate anions and the third water molecule occupying the space among the layers (Figure 4b).

[Ga(KA)2(OH2)2][NO3]•H2O
Ball milling, slurrying, and crystallization from the solution of HKA with Ga(NO3)3 in a 2:1 stoichiometric ratio resulted in the metal complex [Ga(KA)2(OH2)2][NO3]•H2O (see Scheme 6). The powder diffraction pattern for the complex, calculated on the basis of single crystal data, matches the one measured on the bulk products (see Figure S4). Differently from what observed for the silver nitrate co-crystal, when gallium nitrate is reacted with HKA the kojic acid is deprotonated, and acts as a bidentate ligand toward the gallium cation. Two KAanions coordinated the gallium cation (see Figure 4a), and two water molecules completed the octahedral coordination sphere while the nitrate anion was at hydrogen bonding distance from the water molecule that did not belong to the coordination sphere and from the -OH group on the KAanion. Wave layers of cations can be observed extending along the b-axis direction, with nitrate anions and the third water molecule occupying the space among the layers (Figure 4b). The powder diffraction pattern for the complex, calculated on the basis of single crystal data, matches the one measured on the bulk products (see Figure S4). Differently from what observed for the silver nitrate co-crystal, when gallium nitrate is reacted with HKA the kojic acid is deprotonated, and acts as a bidentate ligand toward the gallium cation. Two KA − anions coordinated the gallium cation (see Figure 4a), and two water molecules completed the octahedral coordination sphere while the nitrate anion was at hydrogen bonding distance from the water molecule that did not belong to the coordination sphere and from the -OH group on the KA − anion. Wave layers of cations can be observed extending along the b-axis direction, with nitrate anions and the third water molecule occupying the space among the layers (Figure 4b).
Two KAanions coordinated the gallium cation (see Figure 4a), and two water molecul completed the octahedral coordination sphere while the nitrate anion was at hydrog bonding distance from the water molecule that did not belong to the coordination sphe and from the -OH group on the KAanion. Wave layers of cations can be observed e tending along the b-axis direction, with nitrate anions and the third water molecule occ pying the space among the layers (Figure 4b).

Antimicrobial Assays
Three pathogen-indicator strains were used to assess antimicrobial efficacy. A clas cal disk-diffusion assay was performed to assess whether any antimicrobial activity e isted. This type of assay also evaluates some physical characteristics of the materials. F this type of assay, the compound/crystal is bound/trapped/adsorbed to a filter paper an its polymer strands. This filter disk is then placed on agar bacterial growth media (effe tively, a type of hydrogel). The compound must then be released from the filter paper an diffuse through the agar, leading to a concentration gradient away from the disk. T relative size of the zone of no growth around the disk is then evaluated, and the resulti data are shown in Figure 5. One can observe that, as expected, AgNO3 is an effective an microbial against the two Gram-negative organisms (E. coli and P. aeruginosa), while t best activity towards the Gram-positive strain S. aureus comes from Zn(NO3)2. The copp and gallium salts also showed good antimicrobial activity, as the metal ions diffused rea ily through the agar and are known antimicrobial metals [29]. The kojic acid on its ow showed about one-third the antimicrobial efficacy of AgNO3, validating previous repor of antimicrobial activity. As a comparator, the beta-lactam antibiotic ampicillin show good antimicrobial activity relative to AgNO3 with a normalized zone of inhibition (N of 1.7 for E. coli. Ampicillin is one-half as effective (NZ = 0.5) against P. aeruginosa, b twice as effective against S. aureus (NZ = 2.2).
As for the metal complexes, we observed diverse outcomes. Mechanistically, we d not know if these complexes released from the filter paper equivalently, nor can we certain about their diffusion properties through the hydrogel agar media. Regardle with this assay, we were able to see that both [Ag(HKA)(NO3)]•H2O and Zn(KA)2 d played antimicrobial activity towards all three tested bacterial strains.

Antimicrobial Assays
Three pathogen-indicator strains were used to assess antimicrobial efficacy. A classical disk-diffusion assay was performed to assess whether any antimicrobial activity existed. This type of assay also evaluates some physical characteristics of the materials. For this type of assay, the compound/crystal is bound/trapped/adsorbed to a filter paper and its polymer strands. This filter disk is then placed on agar bacterial growth media (effectively, a type of hydrogel). The compound must then be released from the filter paper and diffuse through the agar, leading to a concentration gradient away from the disk. The relative size of the zone of no growth around the disk is then evaluated, and the resulting data are shown in Figure 5. One can observe that, as expected, AgNO 3 is an effective antimicrobial against the two Gram-negative organisms (E. coli and P. aeruginosa), while the best activity towards the Gram-positive strain S. aureus comes from Zn(NO 3 ) 2 . The copper and gallium salts also showed good antimicrobial activity, as the metal ions diffused readily through the agar and are known antimicrobial metals [29]. The kojic acid on its own showed about one-third the antimicrobial efficacy of AgNO 3 , validating previous reports of antimicrobial activity. As a comparator, the beta-lactam antibiotic ampicillin shows good antimicrobial activity relative to AgNO 3 with a normalized zone of inhibition (NZ) of 1.7 for E. coli. Ampicillin is one-half as effective (NZ = 0.5) against P. aeruginosa, but twice as effective against S. aureus (NZ = 2.2).
As for the metal complexes, we observed diverse outcomes. Mechanistically, we do not know if these complexes released from the filter paper equivalently, nor can we be certain about their diffusion properties through the hydrogel agar media. Regardless, with this assay, we were able to see that both [Ag(HKA)(NO 3 )]·H 2 O and Zn(KA) 2 displayed antimicrobial activity towards all three tested bacterial strains.
Given the range of results of the metal-kojic acid complexes in the disk-diffusion assay, we also performed serial dilution experiments in broth culture to obtain the minimal inhibitory concentrations (MICs) and minimal biocidal concentrations (MBCs) in order to evaluate the bacteriostatic and bactericidal properties, respectively (see Tables 1 and 2). In this case, the metal complexes do not have to dissociate from a filter paper matrix nor diffuse through a hydrogel. The bacteria and the complexes can fully intermingle in the broth medium. Similar to the Kirby-Bauer disk-diffusion assay, for both the MIC and the MBC, the smaller the value, the better the antimicrobial activity, and as typically observed, the MBC is usually 1-4 dilution steps away from the MIC as higher concentrations are required to kill rather than to just inhibit.
These broth experiments reflect the strong antimicrobial efficacy of each of the metal salts and the different efficacies of the different strains [30,31]. We also saw that kojic acid Given the range of results of the metal-kojic acid complexes in the disk-diffusion assay, we also performed serial dilution experiments in broth culture to obtain the minimal inhibitory concentrations (MICs) and minimal biocidal concentrations (MBCs) in order to evaluate the bacteriostatic and bactericidal properties, respectively (see Tables 1  and 2). In this case, the metal complexes do not have to dissociate from a filter paper matrix nor diffuse through a hydrogel. The bacteria and the complexes can fully intermingle in the broth medium. Similar to the Kirby-Bauer disk-diffusion assay, for both the MIC and the MBC, the smaller the value, the better the antimicrobial activity, and as typically observed, the MBC is usually 1-4 dilution steps away from the MIC as higher concentrations are required to kill rather than to just inhibit.
These broth experiments reflect the strong antimicrobial efficacy of each of the metal salts and the different efficacies of the different strains [30,31]. We also saw that kojic acid had intermediate MICs towards all three organisms with equivalent MBCs, except there was no biocidal activity against S. aureus, which may reflect a difference in the penetration of the cell wall of the Gram-positive bacterium. We saw strong MIC and MBC values for the silver-kojic acid complex, while the other metal complexes, although showing intermediate bacteriostatic MIC values, were unable to kill any of the strains.   Overall, our data suggests that the silver-and zinc-kojic acid complexes were able to release their antimicrobial components in both agar (hydrogel) and broth (solution) conditions. The copper complex showed curious results as its antimicrobial activity via disk-diffusion was poor, but it was able to work against the Gram-positive bacteria in the broth culture. Perhaps in an aqueous medium, the crystals can interact more easily with the complex cell wall comprised of peptidoglycans and teichoic acid polymers, which leads to an arrested cellular division. We also saw that, although the [Ga(KA) 2 (OH 2 ) 2 ][NO 3 ]·H 2 O co-crystal did not display antimicrobial activity via disk diffusion, it demonstrated quite good bacteriostatic properties via MIC (see Table 1).

Thermogravimetric Analysis
In this work, a TGA instrument (PerkinElmer TGA7) was employed to determine the behaviour of Cu(KA) 2 , which could be instrumental to the structure solution from powder data. Cu(KA) 2 (1.936 mg) was placed in an aluminium pan and measured at the 305-725 K temperature range at a heating rate of 5 • C·min −1 under an N 2 atmosphere.

Single-Crystal X-ray Diffraction
Single-crystal X-ray diffraction data were collected with an Oxford Diffraction Xcalibur equipped with a graphite monochromator and a CCD detector. Mo-Kα radiation (λ = 0.71073 Å) was used. Single-crystal X-ray diffraction data for Zn(KA) 2 Table S1 in the Supplementary Materials reports the crystallographic details for all of the complexes. The structuresresolution and refinement by the least-squares method against F 2 were carried out using SHELXT [32] and SHELXL [33], respectively, with the Olex2 [34] interface. Nonhydrogen atoms were refined anisotropically. Hydrogen atoms were added in calculated positions. The software Mercury [35] was used to calculate the powder patterns basing on single-crystal data and for molecular graphics. The CCDC numbers 2226706-2226709 contain the supplementary crystallographic data for this paper. These data can be obtained free of charge via http://www.ccdc.cam.ac.uk/ conts/retrieving.html (accessed on 15 December 2022) (or from the CCDC, Cambridge CB2 1EZ, England; Fax: +44 1223 336033; E-mail: deposit@ccdc.cam.ac.uk).

Structural Characterization from Powder Data
The powder diffraction data for Cu(KA) 2 was analysed with PANalytical X'Pert High-Score Plus software. Structure solution was performed with EXPO2014 [36] by simulated annealing in the space group P21 using two KA molecules and one Cu atom. The simulated annealing trials were run 10 times. The best solution was chosen for a subsequent Rietveld refinement using TOPAS4.1 software [37].

Antimicrobial Analysis
The antimicrobial susceptibility of the metal salts and the kojic acid co-crystals were performed in two different ways. As in our previous works [20,27,38], we performed the initial antimicrobial assessment with the Kirby-Bauer disk-diffusion assaywith minor modifications from the standard testing protocol [39]. The following standard strains were used in this study: Pseudomonas aeruginosa ATCC27853, Staphylococcus aureus ATCC25923, and Escherichia coli ATCC25922. All culturing was performed in lysogeny broth (LB) prepared in distilled water with 10 g/L NaCl (VWR International Co., Mississauga, ON, Canada), 5 g/L yeast extract (EMD Chemicals Inc., Darmstadt, Germany), and 10 g/L tryptone (VWR Chemicals LLC, Solon, OH, USA). Agar medium for the disk-diffusion experiments was made by adding 15 g/L bacteriological agar (VWR International LLC, Solon, OH, USA) to the LB. In an aseptic environment, 200 µL of overnight culture of S. aureus, P. aeruginosa, or E. coli was spread on the LB agar plates (two plates per organism for two technical replicates) and left to dry at room temperature for 1 h.
Stocks of 24 mg/mL stock solutions of single components of the metal nitrate salts and suspensions/slurries of the target co-crystal compounds were prepared from powder using the LB media. Four blank antimicrobial disks (Oxoid Ltd., Basingstoke, UK) were placed into a vial with 300 µL of compound stocks and left to soak for 30 min with mixing every 10 min by inversion of the vial. Then, the disks were transferred to the plates, and the plates incubated at 37 • C for 24 h. Every plate had an AgNO 3 disk as a control. This allowed for plate-to-plate and trial variations to be removed through normalization to the zone of inhibition for AgNO 3 . The zone of growth inhibition was measured by a ruler and/or photographed and enlarged in order to quantify smaller zones more easily. Experiments were repeated with 3-5 different biological trials. We note that there may have been differing amounts of co-crystals adsorbed onto each disk within a vial; however, little variation was seen between trials. An additional unknown is whether all of the different co-crystals adsorbed to the filter disks equivalently, or the release of the co-crystals from the disks was equivalent. Yet, the assay is still valid for the interpretation efficacy of antimicrobial applications, such as in a wound dressing.
The second approach was to use the microtiter plate serial two-fold dilutions in LB broth media to obtain the minimum inhibitory concentration (MIC) [39]. Using a standard microtiter plate, each experiment had 8 sterility control wells (150 µL of medium), 8 growth control wells (75 µL of medium), and 2 replicates of dilution series for the different compounds (for a total of 80 wells). A quantity of 75 µL of medium, of which was added 75 µL of the mixture in first well, was then mixed by pipetting up and down a few times. Then, 75 µL was transferred to the next well for a one-half dilution. A 1% inoculant in the LB was prepared from an overnight saturated culture of each of the 3 strains. Inoculation of the microtiter plate was obtained by transferring 75 µL of the 1% culture to each well and mixing for a final volume of 150 µL. Thus, the first-row concentration was diluted from 25 mg/mL to 12.5 mg/mL, and all wells were diluted again by 50%, giving a dilution series ranging from 6.25 to 0.049 mg/mL. The plates were incubated for 24 h at 37 • C and shaken at 150 rpm in a humidity chamber to protect against evaporation. The cultured plates were analysed visually and by optical density at 600 nm, and the MICs were identified as the concentration of the compound in the first well that had no increase in turbidity, adjacent to the well that had such changes present. From this microtiter plate, the minimal biocidal concentration (MBC) was assessed by using a pined replica plater to transfer a 3-5 µL spot to an LB plate and to a fresh microtiter plate with 150 µL per well to assess growth recovery after a further 24 h of incubation at 37 • C. Growth was scored as +/− for each dilution, and the last dilution with no growth recovery was defined as the MBC. Experiments were performed with 3 biological replicates. In Tables 1 and 2, a value of >6.25 indicates the upper limit of the assay, and thus no MBC end point was reached.

Conclusions
In this paper, we have reported on the synthesis, characterization, and evaluation of the antimicrobial activity of a series of novel coordination compounds of the mild antibiotic compounds of kojic acid with metal salts of silver, zinc, copper, and gallium. The new compounds, namely [Ag(HKA)(NO 3 )]·H 2 O, Zn(KA) 2 , Cu(KA) 2 , and Ga(KA) 2 (OH 2 ) 2 ][NO 3 ]·H 2 O, were prepared by mechanochemical, slurry, and solution methods and structurally investigated via single-crystal and powder diffraction. The structures of the silver, zinc, and copper products are 1D coordination polymers, whereby the kojic acid or its deprotonated kojiate anion act as a bridge between the metal centres. The gallium(III) complex formed, instead, a 0D compound bearing two kojiate anions and two water molecules in the coordination sphere.
As the individual components are antimicrobial, we evaluated whether their activity was effected in the co-crystal form, and thus all compounds were tested against the standard pathogen strains of the bacteria P. aeruginosa, S. aureus, and E. coli via disk-diffusion assay measurement. [Ag(HKA)(NO 3 )]·H 2 O was the most effective compound, while [Ga(KA) 2 (OH 2 ) 2 ][NO 3 ]·H 2 O was the least effective. Indeed, the silver and zinc kojic acid 1D coordination polymers were able to release their antimicrobial components in both agar and broth conditions, while the copper complex appeared to work against the Grampositive bacteria in the broth culture. As far as the gallium 0D compound is concerned, while there was no antimicrobial activity via disk-diffusion, a good level of bacteriostatic activity was shown by the MIC.
Overall, our findings have met the objective of this work, mentioned in the introduction, namely the quest for potentially useful, new antimicrobials to meet the growing challenge posed by the increase in antimicrobial resistance by pathogens due to the overuse of antibiotics. We believe that the combination of the antimicrobial properties of organic molecules with those of the metal atoms in the co-crystals and coordination compounds could represent steps forward in this direction. In this same line of work, we have recently shown that the antimicrobial compound proflavine can be co-crystallized with metal salts of silver, zinc, and gallium, yielding an increase in the minimal inhibition concentration of the drug activity towards bacteria [27,38]. However, these are only initial steps as considerable thought is required to understand the mechanisms of action of co-crystals and coordination polymers against bacteria. These studies, which require further, more specialized, and collaborative investigations, are under way.