Modification of Biphenolic Anti-Bacterial to Achieve Broad-Spectrum Activity

Abstract

The gram-positive bacteria, methicillin-resistant Staphylococcus aureus (MRSA) and Gram-negative bacteria, Acinetobacter baumannii, are pathogens responsible for millions of nosocomial infections worldwide. Due to the threat of bacteria evolving resistance to antibiotics, scientists are constantly looking for new classes of compounds to treat infectious diseases. The biphenolic analogs of honokiol that were most potent against oral bacteria had similar bioactivity against MRSA. However, all the compounds proved ineffective against A. baumannii. The inability to inhibit A. baumannii is due to the difficult-to-penetrate lipopolysaccharide-coated outer membrane that makes it challenging for antibiotics to enter gram-negative bacteria. The C2 scaffold was optimized from the inhibition of gram-positive bacteria to broad-spectrum antibacterial compounds that inhibit the dangerous gram-negative pathogen A. baumannii.

Graphical Abstract

A potent biphenolic inhibitor of gram-positive bacteria including MRSA has been permuted to a broad spectrum analog. Incorporation of amine containing groups by O-substitution or by replacement of tert-butyl groups led to a compound both more potent and with a higher therapeutic index than mupirocin, one of the most effective antimicrobial agents against gram negative A. baumannii.Institute and/or researcher

Introduction

Methicillin-resistant Staphylococcus aureus (MRSA) and Acinetobacter baumannii are two of the most potent pathogens responsible for millions of nosocomial infections worldwide.^[1] These multi-drug resistant (MDR) bacteria are commonly found in intensive care units and are able to survive for extended periods of time in hospital environments, facilitating transmission between patients.^[2,3] These two microorganisms are most deadly in late-onset hospital-acquired pneumonia or soft-tissue infections through the contamination of surgical wounds.^[4] Due to the emerging threat of bacteria evolving resistance to more antibiotics, scientists are constantly looking for new classes of compounds that can treat infectious diseases.

Vancomycin, the first glycopeptide antibiotic, was isolated in 1954 from Amycolatopsis orientalis and was originally used to treat postoperative micrococcal colitis.^[5] This drug became widely available after S. aureus evolved resistance to methicillin and is the reference standard for the treatment of many infections caused by drug resistant bacteria.^[6] Still, there are several drawbacks to the usage of vancomycin. This antibiotic is an extremely large, complex molecule, so the generation of analogs is arduous, expensive and time-consuming (Figure 1A).^[7] Further, vancomycin inhibits the cell wall synthesis of bacteria; however, it is only effective against gram-positive species, as it cannot penetrate the outer membrane of gram-negative pathogens.^[8] Finally, this antibiotic has several side effects, such as flushing of the upper body and low blood pressure.^[9]

Figure 1.

Lead gram-positive and broad-spectrum antibiotics.

Advances in antibiotics have given rise to broad-spectrum drugs that are able to be applied topically. One such example is the commercially available mupirocin, a topical aerosol and ointment that has been shown to have strong bioactivity against pathogens that affect the skin flora, such as MRSA and A. baumannii.^[10] This drug is still commonly used to treat Methicillin-resistant infected burn wounds.^[11] Unfortunately, there are several similar disadvantages to mupirocin. The major portion, pseudomonic acid A (PA-A), is only isolatable from submerged cultures of Pseudomonas fluorescens through a procedure with numerous involved steps.^[12] Furthermore, mupirocin contains several stereocenters that makes the synthesis of derivatives difficult (Figure 1B).^[13] The widespread use of mupirocin has also led to many bacteria, including MRSA, garnering resistance and limiting its usage.^[14]

Honokiol, a biphenolic natural product isolated from the bark of magnolia trees, has been used for centuries in traditional eastern medicine as an anti-inflammatory, antidepressant, antiviral, and antibacterial agent. Honokiol and its derivatives have, in the past year alone, been shown to have a wide variety of bioactivities including impeding of breast carcinogenesis, inhibition of oncogenic transcription factor FOXM1, and dose-dependent suppression of the growth of biofilms of A. baumannii.^[15–17] Honokiol is an interesting target due to its structure (Figure 1C) and previously reported strong activity against the etiologic agent of dental caries, Streptococcus mutans. However, under physiologically relevant conditions we discovered that honokiol exhibits poor potency with a minimum inhibitory concentration (MIC) of only 67 μg/mL.^[18] With a library of easily synthesized biphenolic analogs of honokiol that have shown strong inhibitory ability against oral bacteria, their effects against a range of bacteria were of interest.

Results and Discussion

From a library of honokiol analogs, the bactericidal compound 5,5’-(ethane-1,2-diyl)bis(2-(tert-butyl)phenol) (C2) was found to be significantly more potent with an MIC of 0.67–1.3 μg/mL against the oral streptococci bacteria S. mutans as well as Streptococcus gordonii, Streptococcus sanguinis, , and Streptococcus sobrinus. Subsequent structure-activity relationship (SAR) studies on the C2 scaffold determined that in order to retain potency there must be two aromatic rings, at least one free phenol, two tert-butyl groups, and a one or two carbon linker connecting the rings. The next goal was to determine if C2, or any of its most bioactive analogs, are broad-spectrum compounds that can act on the topical bacteria gram-positive MRSA HPV107 and gram-negative Acinetobacter baumannii 2208. If none of the derivatives are able to inhibit gram-negative bacteria, permutation of C2 was planned with the goal of developing a broad-spectrum topical antibacterial agent that is cheaper and easier to synthesize compared to the commercial alternative, mupirocin.

Initially, MIC assays against MRSA and A. baumannii were performed using the most potent oral-targeting biphenolic analogs of honokiol (Figure 2). The bioactivity trends of these compounds against MRSA were similar to those for S. mutans. Compound C2 is the most potent, with an MIC of 1.3 μg/mL. Substitution pattern analog 4H, which had previously matched the activity of C2 against S. mutans, was two dilutions less effective. Compounds 3AE and 4K revealed that activity is retained when one phenol is alkylated, but lost when both phenols are protected, showing the importance of at least one free phenol. Altering the linker size to zero, one, or three carbons (4P, 3N, 4Q) all decreased potency, indicated that two carbons are still the ideal length. However, no analogs were active against A. baumannii at or below 40 μg/mL (Table 1).

Figure 2.

Structures of oral bacterial inhibitors tested against MRSA and A. Baumannii.

Table 1.

Activities of oral bacteria inhibitors against MRSA and A. baumannii.^[a]

Compound	MRSA	A. baumannii
C2	1.3 μg/mL	>82 μg/mL
3N	20 μg/mL	>78 μg/mL
3AE	2.7 μg/mL	>85 μg/mL
3AF	8.7 μg/mL	>68 μg/mL
4F	8.7 μg/mL	>68 μg/mL
4G	1.3 μg/mL	>82 μg/mL
4H	5.2 μg/mL	>82 μg/mL
4K	>89 μg/mL	>89 μg/mL
4M	>41 μg/mL	>41 μg/mL
4P	2.4 μg/mL	>75 μg/mL
4Q	2.7 μg/mL	>85 μg/mL
4R	82 μg/mL	>82 μg/mL
Vancomycin	2.9 μg/mL	>360 μg/mL

^[a]MICs for MRSA and A. Baumannii are reported in μg/mL.

Vancomycin has an MIC of 2.9 μg/mL against MRSA and, similar to the biphenolic honokiol analogs, demonstrated no bioactivity against A. baumannii at or below 360 μg/mL. When accounting for the high molecular weight of vancomycin, analogs C2, 3AE, 4G, 4P, and 4Q (1.4–2.5 μg/mL) all display increased mass potency compared to vancomycin. Furthermore, all of these analogs against MRSA are synthesized through facile 2–4 step syntheses, allowing for easy derivatization compared to the lengthy syntheses of vancomycin derivatives.^[19]

The inability to inhibit A. baumannii is due to the lipopolysaccharide-coated outer membrane that makes it challenging for antibiotics to enter gram-negative bacteria.^[20] The compounds must pass through porins lined with negatively charged amino acids whereupon they are susceptible to efflux pumps, making accumulation inside of the bacterial cell difficult.^[21] Recent comprehensive studies by the Hergenrother group found that polar, rigid, small molecules with ionizable amine groups are the most likely to penetrate the charged outer membrane and accumulate in the cell.^[22,23]

In order to install ionizable primary amines, precursors used in the syntheses of the analogs in previous work were examined for modifiable positions.^[24] The previously synthesized bio-inactive compound 4N (Scheme 1) was chosen due to the electrophilic aldehyde groups that allowed for facile reductive amination reactions using NaBH₄ with N-Boc-1,2-diaminoethane and N-Boc-1,6-diaminohexane to form analogs 5C and 5G. These Boc-protected diamine derivatives were then deprotected using trifluoroacetic acid to afford compounds 5A and 5B.

Scheme 1.

Synthesis of First Broad-Spectrum Phenols

Upon testing the MIC of these four analogs against MRSA and A. baumannii, compounds 5A and 5B, with primary ionizable amines, did not cause inhibition of either bacteria at or below 90 μM (Scheme 1). However, the Boc-protected intermediate 5F inhibited MRSA at 41 μg/mL and A. baumannii at 80 μg/mL, and 5G was even more potent with an MIC of 22 μg/mL against both bacteria. With this first broad-spectrum lead, further analogs of 5G were designed.

In order to determine how the diamine linker length affects bioactivity, Boc-protected derivatives with three, four, and five carbon diamine linkers (5D, 5E, and 5F) were synthesized and subjected to MIC assays (Scheme 1). From this series, compound 5G, with a six-carbon diamine linker, remained the most active. Prior studies with oral bacteria indicate that a larger alkyl group ortho to the phenol increases bioactivity. Thus, analogs were designed with a larger group on the benzylic position ortho to the phenol (Scheme 2). Analog 3D simultaneously underwent demethylation and Friedel-Crafts ortho-acylation by means of TiCl₄ and acetyl chloride. Methyl ketone intermediate 3AF was then subjected to reductive amination with N-Boc-1,6-diaminohexane to afford 5H. The protected acylated intermediate 5J was also subjected to the same sequence to provide 5I. Upon testing the bioactivity, 5H was found to be one dilution (44 μg/mL) less active than 5G against both bacteria, indicating that greater hindrance at the benzylic position ortho to the phenol does not provide additional efficacy. Analog 5I exhibited no potency at or below 100 μg/mL, reinforcing that, in this scaffold, the unprotected phenols are essential to bioactivity (Table 2).

Scheme 2.

Synthesis of Acetyl Acyl Derivatives

Table 2.

Activities of broad-spectrum bacteria inhibitors.^[a]

Compound	MRSA	A. baumannii	HD20	TI
5A	-	>90 μg/mL	360 μg/mL	-
5B	-	>100 μg/mL	-	-
5C	8.9 μg/mL	>100 μg/mL	-	-
5D	73 μg/mL	>100 μg/mL	-	-
5E	77 μg/mL	>100 μg/mL	-	-
5F	41 μg/mL	80 μg/mL	-	-
5G	21 μg/mL	21 μg/mL	340 μg/mL	16
5H	11 μg/mL	44 μg/mL	1400 μg/mL	32
5I	-	>100 μg/mL	-	-
5K	>100 μg/mL	>100 μg/mL	-	-
5L	>100 μg/mL	>100 μg/mL	-	-
5M	>100 μg/mL	>100 μg/mL	-	-
5N	3.8 μg/mL	31 μg/mL	960 μg/mL	32
5Q	14 μg/mL	14 μg/mL	440 μg/mL	32
Mupirocin	≤1.0 μg/mL	250 μg/mL	500 μg/mL	2

^[a]MICs for MRSA and A. Baumannii are reported in μg/mL. Hemolysis values were determined with defibrinated sheep blood and reported as HD₂₀ values which refer to 20% cell lysis compared to the Triton X control. Therapeutic indices (TIs) were calculated as HD₂₀ / MIC A. baumannii.

Potency against MRSA was completely lost when both phenols of the C2 scaffold were protected but was maintained if only one phenol was alkylated.^[18] Thus, phenol alkylation was used to incorporate Boc-protected amines. Compound C2 was treated with 2-(Boc-amino)ethyl bromide and 6-(Boc-amino)hexyl bromide to produce analogs 5K and 5L, respectively (Scheme 3). Unfortunately, neither derivative was effective against MRSA or A. baumannii at or under 100 μg/mL; it appears that both phenols are necessary for inhibition in this system.

Scheme 3.

Synthesis of Boc-Protected Mono-Alkylated Phenols.

Hypothesizing that the steric bulk of the Boc-protected amine provided activity, further analogs with a larger Boc-protected amine were synthesized (Scheme 4). tert-Butyloxycarbonyl-methyl-isothiourea 1 was treated with 1,2-ethanediamine and 1,6-hexanediamine in a nucleophilic displacement of sulfide in order to afford di-Boc guanidine intermediates 5O and 5P, which were then subjected to reductive amination with 4N to yield the bulky analogs 5M and 5N, respectively. Subsequent testing showed that due to the six-carbon diamine linker, 5N had a lower MIC of 31 μg/mL against A. baumannii compared to the shorter two-carbon diamine linker of 5M, which did not exhibit bioactivity. The addition of bulky groups does not appear to increase potency.

Scheme 4.

Synthesis of Guanidine Derivatives

To determine whether the terminal tert-butyl carbamate group was responsible for bioactivity or if inhibition was solely dependent on the steric bulk of the tert-butyl group, compound 5Q was synthesized (Scheme 5). This compound is analogous to 5C, sans the carbamate group, and was found to be the most potent derivative against A. baumannii, with an MIC of 14 μg/mL. Thus, the activity in this system can be retained or enhanced with a benzylic secondary amine and a tert-butyl group.

Scheme 5.

Synthesis of Lead Broad-Spectrum Compound without Boc-Protected Amine.

The commercially available topical antibiotic mupirocin is currently used to treat S. aureus and MRSA in burn victims and is currently being investigated to treat A. baumannii wound infections.^[11,25] Mupirocin has an MIC of ≤1.0 μg/mL against MRSA but an MIC of 250 μg/mL against A. baumannii. These results show that mupirocin is much more potent against the gram-positive bacterium MRSA than any of the broad-spectrum antibacterial compounds described herein. However, lead compounds 5G, 5N, and 5Q all exhibit at least 10-fold greater potency relative to mupirocin for gram-negative bacterium A. baumannii (Table 2). :ead analog 5Q is 18-fold more efficacious than mupirocin and easily synthesizable from the readily available starting materials.

Because these compounds were effective against A. baumannii, their relative toxicities were explored. Hemolysis data, reported as HD₂₀ values, revealed that analogs 5A, 5G, and 5Q had similar hemolytic activity as mupirocin (340–500 μg/mL), while analogs 5H and 5N exhibited less hemolytic activity at 1400 μg/mL and 960 μg/mL, respectively (Table 2). Based on these results, a therapeutic index (TI), expressed as a ratio of the HD₂₀ value to the MIC for A. baumannii, could be determined. Compounds 5N and 5Q have TI values of 32, while 5G has a TI value of 16. All three analogs had TI values much greater than that of mupirocin (TI: 2), suggesting that they have potential for use as topical antibiotics against wounds infected with A. baumannii.

Previous work has shown that the inhibitory action of compound C2 and its analogs against S. mutans involves the permeabilization of the bacterial cell membrane.^[26] Using that information as a starting point, the cell permeability of the most potent broad-spectrum analogs, 5G, 5N, and 5Q, were assessed against A. baumannii over 2 h using the fluorescent dye SYTOX Green Nucleic Acid Stain. All three analogs caused permeabilization of the bacterial membrane at or below their MIC values (Figure 3), with 5G having the most striking effect. The mechanism of action of mupirocin differs from the biphenolic honokiol derivatives, as mupirocin reversibly binds to the bacterial enzyme isoleucyl-tRNA synthetase, inhibiting bacterial RNA and protein synthesis while the biphenolic honokiol derivatives appear to be affecting membrane stability and fluidity.^[27]

Figure 3.

Cellular Membrane Permeability of A. baumannii measured with SYTOX Green over 2 hours.

Conclusion

The most potent biphenolic honokiol analogs against oral bacterial also exhibit strong bioactivity against the gram-positive bacterium MRSA (MICs of 1.3–2.7 μg/mL), comparable in strength to the reference standard antibiotic, vancomycin (MIC of 2.9 μg/mL). However, these compounds are unable to inhibit the gram-negative bacterium A. baumannii. The C2 scaffold then permuted to also inhibit the dangerous gram-negative pathogen A. baumannii. The incorporation of benzylic amines, tert-butyl groups, and free phenols were key moieties that led to bacterial inhibition. Analogs 5G, 5N, and 5Q inhibit A. baumannii at 21 μg/mL, 31 μg/mL, and 14 μg/mL, respectively, a significantly lower concentration than the commercially available topical antibiotic mupirocin (MIC of 250 μg/mL). Mechanism of action experiments revealed that all three derivatives cause bacterial membrane permeabilization within one dilution of the MIC. Notably, 5Q has the lowest MIC (14 μg/mL). Future work includes further study of the discrete mechanism of action. These biphenolic honokiol derivatives merit further study as a new class of broad-spectrum antibacterial compounds that can be used to combat the growing threat of resistant A. baumannii.