Antimicrobial activity of 1 , 4-naphthoquinones by metal complexation

The effect of metal complexation on the antimicrobial activity of 1,4-naphthoquinones was investigated. Nickel-, chromium-, iron-, copper-, and cobalt-containing metal chelates of 5-amino-8hydroxy-1,4-naphtoquinone (2) and its acyl-derivatives (3-8) were synthesized and characterized, and their antimicrobial activity was evaluated. Data from infrared spectroscopy indicate that naphthoquinones coordinate through oxygen and nitrogen atoms for 2, and through oxygen atoms when ligands were acyl derivatives 3-8. Susceptibility tests for antimicrobial activity showed that 2 and its acyl derivatives were effective on inhibiting the growth of pathogenic bacteria such as Staphylococcus aureus, Streptococcus uberis and Bacillus cereus, but not Gram-negative bacteria. The metal complexation often caused decrease of biological activity. Nickel complex of 2 was the most effective against Gram-positive bacteria, showing MIC values ranging from 375 to 1400 mg/ml. Metal chelates may be useful tools for the understanding of the antimicrobial mechanism of 1,4-naphthoquinones on these bacteria. *Correspondence:

The investigation on new antimicrobial agents is important due to the resistance acquired by several pathogenic microorganisms.The prevalence of strains of Staphylococcus aureus resistant to conventional antibiotics has increased to high levels in some hospitals (Emori, Gaynes, 1993).Increased resistance among strains of streptococci and enterococci are also described (Lipsitch, 2001).
The effect of metal complexation on some antimicrobial agents has been mentioned in the literature.Tetracyclines often lose their antimicrobial activity when coordinated with metals (Chopra, Hawkey, Hinton, 1991).Some cations such as magnesium and aluminium are complexed by fluoroquinolones, resulting in decrease in oral bioavailability, and may cause therapeutic failure (Lecomte et al., 1994;Wallis et al., 1996).However, increased activity has been described for some hydroxyquinone metal-chelates (Christianopoulou, Ecateriniadou, Sarris, 1986) and a copper(II) complex of cephalexin (Iqbal et al., 1999).
In this paper we report on the screening of in vitro activity of naphthoquinones derived from 5-amino-8hydroxy-1,4-naphthoquinone and their metal chelates as part of our interest on the synthesis and characterization of the naphtoquinones derivatives and their transition metal complexes.

Microorganisms
The strains tested for antibacterial activity were clinical isolates obtained at the Faculdade de Medicina Veterinária, UFRGS.Screening tests were carried out on the following bacteria: Staphylococcus aureus, Staphylococcus epidermidis, Staphylococcus intermedius, Enterococcus faecalis, Streptococcus uberis, Escherichia coli, Pseudomonas fluorescens, Bacillus cereus, Salmonella Typhimurium.Strains from collection were S. aureus ATCC 25923 and E. coli ATCC 25922.Stock cultures were kept frozen at -21 °C in nutrient broth (Difco) containing 20% (v/v) glycerol.The organisms were propagated and maintained on nutrient agar dishes at 4 °C during the susceptibility experiments.Strains of S. uberis were submitted to susceptibility tests after isolation on 5% blood sheep agar.

Susceptibility testing
The susceptibility tests were performed following the NCCLS recommendations (NCCLS, 2001).Screening tests regarding the inhibition zone were carried out by the filter paper disk method.Bacterial suspension was prepared from colonies grown overnight on an agar plate, and inoculated into Mueller-Hinton broth (Merck) to give a 0.5 McFarland turbidity standard solution.A sterile swab was immersed in FIGURE 1 -Chemical structures of the naphthoquinones tested.The metal ion coordinates through site B for ligand (2) and site A for the acyl-derivatives (3 to 8).
the bacterial suspension and used to inoculate Mueller-Hinton agar plates.The disks were applied to the surface of inoculated plates using a sterile forceps.The compounds were diluted in phosphate buffered saline solution from a 10 mg/ml stock dimethylsulfoxide (DMSO) solution and then applied on disks (50 mg/disk).The inhibition zone was measured around each disk after 24 h at 37 °C.Strains of E. faecalis and S. uberis were tested on Mueller-Hinton agar plates supplemented with 5% sheep blood.Controls with DMSO were adequately done.
To access the Minimum Inhibitory Concentration (MIC) -defined as the drug concentration at which no growth was visible -96-well sterile microplates (Corning) were filled with 0.1 ml of serial twofold dilutions (2000 to 4 µg/ml) of the different naphthoquinones.The compounds were diluted in Mueller-Hinton broth from a 10 mg/ml stock DMSO solution.A standardized number of bacteria (0.1 ml of a 10 6 CFU/ml suspension in Mueller-Hinton broth) were inoculated into each well.A growth well (broth plus inoculum) and a sterility control well (broth only) were included in each panel.Microplates were incubated at 37 °C for 24 h, and then MIC was determined as the last dilution where no increase in visual turbidity was observed.

RESULTS
The naphthoquinone compounds and their chelates were synthesized as described elsewhere (Martinelli et al., 1999(Martinelli et al., , 2000)).All the chelates were made of bivalent metal, except Cr(III).They are neutral complexes and were isolated as fine powders.They are quite insoluble in common organic solvents despite the lipophylic chain on compounds 4 to 8 and are thermally more stable than the free organic ligands, as should be expected.The most important changes in the IR spectrum of ligands 2 and 3 and their nickel chelates, Ni(2) and Ni(3), can be seen in Table I.
Compounds 2 to 8 and their metal chelates were initially tested against Staphylococcus aureus ATCC 25923 and Escherichia coli ATCC 25922.Antimicrobial activity of 2-8 was observed on S. aureus but not on E. coli (not shown).The acyl-derivatives presented lower inhibition zones than the starting compound 2 (Table 2).Metal chelates showed either decreased antibacterial activity or lack of inhibitory effect.The nickel complexes of 2 and 3 and the cobalt complex of 2 were the most active (Table II).
The biological effect of 2, and their metal-chelates on bacterial growth is shown in Table III.The ligand 2 was capable of inhibiting the growth of S. aureus, S. intermedius, S. epidermidis, B. cereus, S. uberis and E. faecalis.Nickel, cobalt and iron complexes of 2 showed a decreased inhibitory effect when compared with the ligand alone.Lack of inhibitory effect was observed for copper and chromium chelates.The activity of these compounds was also tested on Gram-negative microorganisms, such as Escherichia coli, Pseudomonas fluorescens, and TABLE I -Most relevant infrared frequencies of some characteristic bands of 5-amino-8-hydroxy-1,4-naphthoquinone (2),5-acetylamino-8-hydroxy-1,4-naphthoquinone (3) and their Ni-chelates Ni(2) and Ni(3), respectively  Salmonella Thyphimurium.None of the compounds were able to inhibit the growth of those Gram-negative bacteria (Table III).
The effect of nickel and cobalt complexation of 2 on MIC values was determined against Gram-positive bacteria (Table IV).Compound 2 presented MIC values in the 30 to 125 mg/ml range, whereas the Ni(2) complex resulted in MIC values ranging from 375 to 1400 mg/ml.Nickel complexation of 3 or 5 caused increase of MIC to values higher than 2000 mg/ml (not shown).

DISCUSSION
In the previous work we showed that naphthoquinone ligands have two potential sites to coordinate a metal ion (Martinelli et al., 1999(Martinelli et al., , 2000)).Spectroscopic data allowed us to assign that the metal ion coordinates through site B for ligand 2, while site A is preferred by the metal in the acyl-derivatives 3 to 8 (Figure 1).
Metal chelates of compound 2 and its derivatives 3 to 8 have been synthesized and were tested against Grampositive and Gram-negative bacteria.In Tables 2 and 3 it is possible to observe that all the ligands showed antibacterial activity.Compound 2 showed higher activity compared with its derivatives 3 to 8. All compounds possess the free naphthazarin site (site A, Figure 1), which has been associated with antibacterial properties (Papageorgiou, et al.1999).
Studies on antimicrobial activity of 1,4naphtoquinones demonstrated that those with electronreleasing or weak electron-withdrawing groups at position 2 or 3 may enhance their antimicrobial activity (Ambrogi et al., 1970;Gershon, Shanks, 1975).An explanation for this behavior is related to the electronic effect of the groups directly bonded at the naphthoquinone ring.This effect is attributed to the enhancement of the hydrogen bonding, allowing stronger binding at its site of action.Nevertheless, in our case, the acyl compounds 3 to 8 showed a decrease in activity, although they present the site A free.Such behavior could be explained considering that the group R is not directly bound to the naphthoquinone ring and does not cause a positive electronic effect.Furthermore, the decrease in activity observed for the acyl derivatives may be related to the presence of the lipophylic group, since excess of hydrophobicity causes loss of activity (Gershon, Shanks, 1975).Results are the means of five independent determinations for each isolate.(-) no inhibition was observed.
Antimicrobial activity of metal complexes of naphthoquinones has been described.Some metal chelates of juglone often maintain their antibacterial effect (Joshi, 1986).(II) and Ni(II) complexes of 5-hydroxy-1,4naphthoquinone result in higher antibacterial effect against Bacillus spp.and S. aureus (Christianopoulou, Ecateriniadou, Sarris, 1986).In addition, the Ni(II) complex of vitamin K3-thiosemicarbazole has been recently reported to increase the activity against Grampositive bacteria (Li et al., 2000).Therefore, the effect of metal complexation on antimicrobial activity was a matter of our interest since the chelation of 2 occurred in site B, leaving the naphthazarin site A free.
The data obtained for the metal chelates showed lower activity than their respective ligands, and total loss of activity was observed in some cases.Considering that the naphthazarin site A is the active one, its blocking by complexation may be related with the decrease or loss of activity.In fact, site A remains free after coordination in the case of ligand 2, but it is involved in metal complexation in 3 to 8. Assuming that the antibacterial mechanism is interfering in the electron transport chain, this free site should be required to participate in a redox reaction.This idea may be supported by electrochemical studies on 3 and Ni(3), in which the reduction potential becomes more negative with complexation (Piatnicki et al., 1996), i.e., more energy is required for the reduction process.In addition, the presence of an imino group instead of a keto group in position 1 or 4 results in loss of antimicrobial activity (Riffel et al., 2002).This may indicate that both free C=O groups are required for full activity.
The set of data of compound 2 and its metal chelates showed the highest activity compared to the other chelates.The Ni(2) complex was found to be slightly less active than the ligand itself, nevertheless it followed the behavior of 2 being the most active against all Grampositive bacteria tested.This was an unexpected result since the naphthazarin site is kept uncoordinated in the chelates unless the electrochemical behavior of the ligand has been modified by the metal ion.
As the ligands and some complexes were active against S. aureus, which is a very relevant pathogen, metal chelates may be useful tools for the understanding of the antimicrobial mechanism of 1,4-naphthoquinones on this bacterium.

TABLE II -
Effect of 1,4-naphthoquinones and metal chelates on growth of Staphylococcus aureus* * Inhibition zone in mm.Results are the means of five independent determinations for each isolate.(-) no inhibition was observed.(v) inhibition was not observed against all isolates.

TABLE III -
Susceptibility testing of 5-amino-8-hydroxy-1,4-naphthoquinone metal-chelates* * Inhibition zone in mm.Results are the means of five independent determinations for each isolate.(-) no inhibition was observed.