Addressing Antimicrobial Resistance through New Medicinal and Synthetic Chemistry Strategies

Abstract

Over the past century, a multitude of derivatives of structural scaffolds with established antimicrobial potential have been prepared and tested, and a variety of new scaffolds have emerged. The effectiveness of antibiotics, however, is in sharp decline because of the emergence of drug-resistant microorganisms. The prevalence of drug resistance, both in clinical and community settings, is a consequence of bacterial ingenuity in altering pathways and/or cell morphology, making it a persistent threat to human health. The fundamental ability of pathogens to survive in a multitude of habitats can be triggered by recognition of chemical signals that warn organisms of exposure to a potentially harmful environment. Host immune defenses, including reactive oxygen intermediates and antibacterial substances, are among the multitude of chemical signals that can subsequently trigger expression of phenotypes better adapted for survival in that hostile environment. Thus, resistance development appears to be unavoidable, which leads to the conclusion that developing an alternative perspective for treatment options is vital. This review will discuss emerging medicinal chemistry approaches for addressing the global multidrug resistance in the 21st century.

Introduction

In order to pursue chemotherapy successfully we must look for substances which possess a high affinity and high lethal potency in relation to the parasites, but have a low toxicity in relation to the body, so that it becomes possible to kill the parasites without damaging the body to any great extent. We want to hit the parasites as selectively as possible. In other words, we must learn to aim and to aim in a chemical sense. The way to do this is to synthesize by chemical means as many derivatives as possible of relevant substances.

—Paul Ehrlich, Über den jetzigen Stand der Chemotherapie. Berichte der Deutschen Chemischen Gesellschaft, 1909, 42, 17–47.

The effects of Penicillium on bacteria had been known prior to the now famous studies of Alexander Fleming. However, the reports describing the phenomenon received little attention before the clinical importance of penicillin was established. Investigating the possibility that one organism can interfere with another’s growth dates back to the early days of microbiology in the 19th century.^1,2 By the end of the century, this phenomenon was widely accepted and has been given a name: antibiosis. Antibiotics are either bactericidal (kill bacteria) or bacteriostatic (stop bacteria from growing).

Antibiotic mechanisms of action are based on prevention of DNA/RNA, folate, and protein syntheses and destruction of the cell wall/membrane. Bacterial resistance to antibiotics has been known since the 1940s, when a colony of Staphylococci survived lethal action of the compound with a newly determined structure: penicillin. ³ Similarly, resistance to streptomycin was reported shortly after its discovery. ⁴ These findings led to the conclusion that a population of preexisting resistant organisms exists, even in the absence of drug-generated evolutionary pressure. However, it was not until the 1970s when the rate of bacterial resistance accelerated, eventually leading to the current global health problem.^5,6 Looking back from our current understanding of how ancient bacterial resistance toward antibiotics is,^7,8 we realize that resistance is unavoidable and accepting the necessity of developing new strategies for addressing it is pending. Developing resistance is a response of a pathogen to the selective pressure of antibiotics on the existing millennia elements of resistance (e.g., R-plasmids). ⁹ That response is triggered not only by natural antibiotics but also by antibiotics prepared via organic synthesis (e.g., fluoroquinolones). ¹⁰ Resistance mechanisms include expression of hydrolytic enzymes such as the β-lactamases; removal of the antibiotic by efflux pumps; modification of the antibiotic’s molecular target, rendering it unable to bind the drug; and circumvention of antibiotic toxicity.

In addition to bacterial ingenuity in altering pathways and/or cell morphology, human behavior (e.g., overutilization of miracle drugs, modest commercial opportunities of producing future infectious disease medicines) has contributed to reaching the magnitude of antimicrobial resistance (AMR) we face today with clinically relevant antibiotics. It is estimated that by 2050, the lives lost to AMR will exceed the 8.2 million a year currently lost to cancer, ¹¹ and the economic burden caused by AMR worldwide might be as high as $100 trillion. ¹¹ In 2017, the Centers for Disease Control and Prevention (CDC) detected more than 220 cases of a rare type of nightmare bacteria that are virtually untreatable, according to a report released on April 3, 2018. Although the CDC has warned of the danger of antibiotic-resistant bacteria for years, this new report helps illustrate the scope of the problem. The CDC’s principal deputy director, Anne Schuchat, said she was surprised by the extent resistance has spread. “As fast as we have run to slow (antibiotic) resistance, some germs have outpaced us,” Schuchat said. “We need to do more and we need to do it faster and earlier.” ¹²

Challenges in antimicrobial drug discovery are imposed by the vast microbial genetic diversity and the fact that it is technically difficult and takes a long time to make a novel antibiotic: “Converting an early chemical prospect into a medicine that can be used in people is a profound scientific challenge, the difficulties of which are not going to be mitigated by a change in the commercial landscape or public policy.” ¹³

In addition, the design of new antibiotics ought to make them less prone or even impervious to resistance development. This is a challenging goal, since the distribution of AMR is not always predictable. Illustrative to that is the finding that the gene encoding for a hydrolyzing enzyme (New Delhi metallo-β-lactamase, NDM-1) of β-lactam antibiotics is a chimeric gene. The latter resulted from the fusion of a progenitor type metallo-β-lactamase gene with a partial sequence of a gene conferring resistance to aminoglycosides, probably created de novo in Acinetobacter baumannii. ¹⁴

The chemical approaches used to create new lead molecules (aside from efforts in new natural product isolation) are (1) combinatorial chemistry generally resulting in primarily sp²-hybridized molecules, (2) methods that develop diverse libraries of complex scaffolds, and (3) modifications to previously identified antibiotic compounds. This review will focus on recent developments encompassing the two latter approaches attempting to address AMR through synthetic and medicinal chemistry means.

The Antibiotic Tree: Where Are We At?

The majority of today’s most effective antimicrobials are natural products isolated in the 1940s from soil-colonizing bacteria and fungi. Of the principle antibiotic scaffolds in current clinical use, fluoroquinolones, sulfonamides, trimethoprim, and more recently oxazolidinones are products of organic synthesis. Natural product scaffolds of clinical significance include β-lactams, aminoglycosides, tetracyclines, glycopeptides, and macrolides.^15,16 Since their initial discoveries, the antimicrobial industry has focused on fine-tuning existing classes of antibiotics to improve their antimicrobial spectrum, efficacy, and safety as well as to combat clinical resistance in the later years of the 20th century. The semisynthetic approach for achieving the latter has been the most commonly used approach for most of the past century. ⁷ However, multiple variation of natural product scaffolds with a mode of action on conserved targets has contributed to the development of multidrug-resistant (MDR) bacterial strains. High frequencies of MDR bacteria have been grouped under the acronym ESKAPE: Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter spp. ESKAPE pathogens are responsible for the majority of nosocomial infections and are capable of escaping the biocidal action of antimicrobial agents. By the mid-1990s, the enthusiasm for making yet another incremental improvement to a β-lactam, macrolide, or quinolone had diminished (example of diversification of existing scaffold, of vancomycin, is shown in Fig. 1 ). Then, in 1995, determination of the complete DNA sequence of a bacterial genome from Haemophilus influenzae revived interest in the development of antibacterials. “The prospect of hundreds of new genes to explore as possible targets sparked new interest in antibacterial discovery and fired the imagination.” ¹³ From then on, attempts at finding a novel antibacterial agent in the pharmaceutical industry have been generally directed toward three discovery approaches: new molecular targets, found by comparative genomics; new structures for old molecular targets; and cell-based screening for novel structures. There are several excellent reviews on the subject, with lessons learned from the era of the great expectations of comparative genomics, nicely presented in the reflective papers of the GlaxoSmithKline (GSK) and AstraZeneca experiences.^13,17

Figure 1.

Traditional (semisynthesis and combinatorial) and innovative (diversity-oriented synthesis, complexity-to-diversity) synthetic approaches.

Antimicrobials Are in a Drug Class of Their Own: Building the Right Antimicrobial Molecule Is the Hard Part

An extremely important factor when searching for new drugs is the chemical diversity of compounds available to screen. Comparative genomics give new targets; however, finding the suitable lead for a new target has been disappointing. Chemical collections of most large pharmaceutical companies have been exhausted for possible antibacterials mainly by using high-throughput screening (HTS). ¹³ The analysis of intensive HTS campaigns from that time by research groups at GSK ¹³ and AstraZeneca ¹⁷ offer some explanation for this lack of success. The reasons for the latter are complex. However, the limited diversity of chemical space of their libraries as sources of difficulty in identifying new antibacterials has been recognized in both reports.^13,17 The initial excitement for the opportunity to find chemically diverse compound libraries has been spurred by the development of combinatorial chemistry ( Fig. 2 ). The latter has been extensively used in industry to generate lead compounds. From 1999 to 2008, 45 of the 50 Food and Drug Administration (FDA) approvals for first-in-class small-molecule new molecular entities originated from a screen. ¹⁸ However, this approach has not paid off for the class of antimicrobials as expected, since it generally results in the sp²-hybridized molecules.^19,20 Essential gene screens and HTS technologies allow for testing of millions of compounds in a short period of time, but they have not resulted in new antimicrobial drugs. ²¹ The reason is the limited makeup of screening libraries.^22,23 In addition, Lipinski’s rules, which are set for combinatorial chemistry, do not include the drug class of antibiotics. ²⁴ The lack of success to fit antibacterials in the combinatorial approach is because they (marketed drugs and natural products) tend to be larger and more complex than the average screening compound.^13,25 In addition, gram-positive agents have larger molecular weight, larger polar surface area (PSA) of 243 Å, and lower ClogD_7.4 (–0.2) than comprehensive medicinal chemistry or normal drugs, which have a PSA of 70 Å and clogD_7.4 of 1.6.^26–28 The gram-negative agents are generally smaller, are more polar, have a PSA of 165 Å, and have a ClogD_7.4 of (–2.8).^26,28

Figure 2.

Representative classes of clinically relevant and newly discovered antibiotics (teixobactin and cystobactamides). The fragments shown in bold represent the peripheral structural changes of vancomycin that led to improved antimicrobial potency and provided additional synergistic mechanisms of action, including those that are associated with membrane-binding properties.

The physicochemical properties of antimicrobial drugs^26,29 have not been met by the majority of libraries of compounds from most commercial and pharmaceutical sources.^26,27 The aforementioned approaches do not allow for coevolution of a high affinity for a biological target and the ability to reach the cellular site of action. These are two distinct advantages of starting from a natural product scaffold during lead optimization.^27,30 Therefore, it is only conceivable that nature will continue to serve as a prolific source of new antibacterial leads. ³¹

New Structures for Old Molecular Targets: “A Good Target Is Better than a New Target” ¹³

There are several recent reports confirming the revived interest in collection of new^29,32 and utilization of previously rejected (due to either toxicity or poor pharmacokinetics/pharmacodynamics, narrow antimicrobial spectrum) natural products. ²⁹ Kibdelomycin, a novel natural product inhibitor of bacterial type II topoisomerase ³² that is unrelated to fluoroquinolone antibiotics, is such an example. This confirms the notion that the search for naturally produced yet unsuspected scaffolds for established drug targets expands our antimicrobial arsenal. A novel class of natural products from Cystobacter spp. has been recently identified ³³ as part of a search for new compounds with activity against gram-negative bacteria. Cystobactamides ( Fig. 2 ) have structures of arylamides, and in a sense, they are peptides with unusual aromatic moieties, such as p-aminobenzoic acid. Cystobactamides are inhibitors of topoisomerases and thus can be regarded as natural quinolones. ³³ They strongly inhibit the growth of various ESKAPE pathogens, ³⁴ including the gram-negative Acinetobacter baumanii and gram-positive Enterococcus faecalis and Streptococcus pneumoniae at concentrations comparable to and even exceeding the activity of ciprofloxacin, a clinically relevant second-generation fluoroquinolone antibiotic.

Another representative of naturally occurring peptides, the cyclic depsipeptide teixobactin, has been discovered in a screen of uncultured bacteria (from the previously unknown β-proteobacterium Eleftheria terrae). ³⁵ Notably, no teixobactin-resistant bacterial strain has been detected so far, regardless of directed efforts. The molecular target of teixobactin apparently is the same as that of vancomycin, lipid II, the precursor of peptidoglycan, but the binding site differs from the one of vancomycin. ³⁶ Because teixobactin has shown excellent antimicrobial activity against several ESKAPE pathogens, including vancomycin-resistant strains, (0.5 to 0.03 µg/mL), it is considered to be one of the very promising lead compounds for antimicrobial drug development. ³⁶

Synthetic approaches directed toward reengineering of natural products into antibiotics continue to rejuvenate the established scaffolds of the golden age of antibiotics. These approaches include diversity-oriented synthesis (DOS) and complexity-to-diversity (CtD). They allow for preparation of libraries of compounds through relatively few chemical transformations in a highly divergent manner.^22,37–51 The DOS^38–47 and CtD^22,49–51 are directed toward addressing the specific requirements for a successful antibacterial structure. As mentioned earlier, average log D values (log P or log D values are the parameters for measuring hydrophobicity) for drugs targeting these organisms are −0.2 and −2.8, respectively, and 1.6 for other drugs. Additional data from approximately 3200 compounds emerging from AstraZeneca antibacterial discovery campaigns has expanded the aforementioned findings. ²⁸ The hydrophobicity has been found to be highly distinct between active antibacterials and average log D in the screening compound library. The trend to improve in vitro activity of the lead scaffolds frequently led to increased log D, which in turn resulted in lowering in vivo activity. The balance between cell penetration and efflux ²⁸ is especially important for gram-negative organisms such as Pseudomonas, which has numerous efflux pumps. Understanding the role of the cell envelope in compound entry and retention in the cell is needed, ⁵² if efforts in building tailored chemical libraries with improved whole-cell bioactivity are to be supported. ²¹ However, a high number of pharmaceutical companies are still synthesizing compounds with a mean ClogP >4. ⁵³

DOS

The DOS strategy (in which simple starting materials are coupled to make diverse structures; Fig. 1 ), originally introduced by Schreiber, ⁴¹ has led to the discovery of complex bioactive scaffolds. ⁴² Compounds produced by DOS synthesis are comparable to natural products by their higher ratio of sp³-hybridized atoms and stereocenters as compared with compounds found in conventional (produced by combinatorial chemistry) screening collections. The descriptor Fsp³ is the number of sp³-hybridized carbon atoms in a compound divided by the sum of the sp³- and sp²-hybridized carbon atoms. ⁴⁸ The benefits of a higher Fsp³, which include lower melting points and enhanced aqueous solubility, has been demonstrated. ¹⁹ Three antimicrobial compounds (out of 242 screened) with inhibition activity against S. aureus, including two UK epidemic methicillin-resistant strains (EMRSA 15 and EMRSA 16), have been discovered through the DOS approach. ⁴⁴ The most potent compound, (±)-gemmacin (1; Fig. 3 ), is weakly active against resistant strains, including vancomycin-resistant enterococci (VRE). Its low antifungal activity indicated a degree of selectivity. The DOS strategy has been successfully employed toward previously identified antibacterials. ⁴⁵ The enopeptin family (2 and 3; Fig. 3 ) has been shown to have excellent antibacterial activity in vitro but limited activity in vivo because of its poor solubility. Structure-activity relationship (SAR) studies by Bayer Pharmaceutical Research identified key hydrogen bonding interactions that constrained the peptidolactone core and identified a lead compound called acyldepsipeptide 4 (4; Fig. 3 ). Using the DOS strategy, compounds 5a–5c ( Fig. 3 ) were prepared and proven to be the most potent against both MRSA and VRE strains. Recently, lead compounds with representative 6 ( Fig. 3 ) with novel mechanism of action (inhibitors of glutamate racemase) against Clostridium difficile ⁴⁶ and as inhibitors (7a-d; Fig. 3 ) of metallo-β-lactamase ⁴⁷ have been identified through the DOS approach. Compound 6 (BRD0761) has greater potency against several C. difficile isolates (MIC of 0.06–1 µg/mL) than vancomycin. ⁵⁴ The suggested target of BRD0761 is glutamate racemase, an essential protein involved in cell wall biosynthesis. It is interesting to note that while this enzyme is a known target in Helicobacter pylori, Mycobacterium tuberculosis (Mtb), and other gram-positive bacteria, ⁵⁵ it has not yet been validated in C. difficile. ⁵⁶ About 83,000 compounds of the Broad DOS collection for activity against green fluorescent protein–expressing Mtb ⁵⁷ have been screened through a collaborative effort of researchers at the Broad Institute, the University of Chicago, and Argonne National Laboratory. ⁵⁷ This screen identified the chiral azetidine 8 (BRD4592) ( Fig. 3 ) to have bactericidal activity against Mtb (MIC₉₀ of 3 µM). Because none of the other seven stereoisomers of 8 has demonstrated activity, it appears that its activity is due to specific target binding. Both subunits of tryptophan synthase (IC_50α of 71 nM; IC_50β of 23 nM), a previously untargeted essential metabolic enzyme, has been identified as the molecular target of azetidine 8. ⁵⁷ A thorough discussion of the DOS approach for discovering compounds with novel mechanisms of action can be found in the excellent recent review. ²³

Figure 3.

Structures of antibacterials from diversity-oriented synthesis (DOS) and complexity-to-diversity (CtD) approaches. Compounds 1–8 are examples of the DOS approach. Despite their structural similarities, compounds 6 and 7 are inhibitors of different bacterial enzymes, glutamate racemase and metallo-beta-lactamase, respectively. Compound 8 has highly specific engagement with its target, tryptophan synthase of M. tuberculosis. Compounds 9–14 are representatives of the CtD strategy. Compound 9 has no antimicrobial properties; compounds 10 and 11 have modest activity against S. aureus.

CTD Approach

The CtD approach introduced by Hergenrother and coworkers ²² uses chemoselective reactions to systematically alter core structures of natural products via reactions that distort the ring system. In contrast to traditional optimization campaigns, the goals of the CtD method are to enhance the inherent biological activity/improve druglike properties of a natural product.^22,49–51 Several natural products from different structural classes such as gibberellic acid, adrenosterone, and quinine have been used as core structures to demonstrate the potential of the CtD approach. ²² Because there are many readily available natural products, the possibilities to achieve diverse and complex molecules using this approach are plenty. The CtD strategy has been inspired by how nature creates complex molecules using a common intermediate to generate a multitude of compounds that are very different from one another. ²² The resulting compounds from the CtD approach appear to meet the aforementioned goals: they have an average Fsp³ of 0.59, which is higher than that in the commercial collection (0.23), and the average ClogP is 1.1 log units lower than that in the commercial screening set (2.90 versus 3.99), which corresponds to a 12-fold reduction in hydrophobicity. ²²

Using this strategy, small-molecule compounds with antimicrobial properties have been constructed from readily available natural products. ⁵⁸ The indole alkaloid yohimbine is an example of such a starting natural product. Yohimbine, (9; Fig. 3 ) has no antibacterial activity; however, two compounds, 10 and 11 ( Fig. 3 ), with modest inhibitory activity against S. aureus have been identified. The ring distortion method demonstrates the ability to bring about antibacterial activity by introducing substantial structural modifications, a strategy that has been used to prepare quorum-sensing (QS) modulators based on the scaffold of a natural marine product (12; Fig. 3 ). ⁵⁹ The QS activity of compounds 13 and 14 ( Fig. 3 ) has been determined against an opportunistic pathogen bacterium of tropical marine organisms, Vibrio harveyi. ⁵⁹

Fragment-Based Drug Discovery

Fragment-based lead generation (FBLG) is an alternative to traditional HTS. FBLG allows the discovery of novel inhibitors of known drug targets through a unique binding site, chemistry, or mechanism of inhibition, which allows for the evasion of existing resistance mechanisms. Most major pharmaceutical and biotech companies use fragment-based approaches as part of their lead generation strategies. ⁶⁰ The underlying concepts of FBLG were implemented in silico in the 1980s and early 1990s.^61,62 However, lead generation driven by experimental screening of fragments has been receiving widespread interest only since the introduction of SAR by nuclear magnetic resonance in the mid-1990s.^63,64 This approach initially focused on the use of two-dimensional nuclear magnetic resonance techniques to screen for small compounds binding to the target protein, preferably to different adjacent subpockets in an active site. It appears, however, that most proteins have only a single energetic focal point, dominating the binding energy available in the active site. This so-called hot-spot^65–67 is comparatively small and easily exploited with fragment-sized compounds. Pyrrolamides, prototypes of a novel class of DNA gyrase inhibitors, with activity against predominantly gram-positive bacterial pathogens, including resistant strains, have been identified using the FBLG approach by an AstraZeneca research team. ⁶⁸ A representative compound from the pyrrolamide series (compound 15; Fig. 4 ) demonstrated efficacy in a mouse lung infection model. ⁶⁸

Figure 4.

Antibacterials identified by the fragment-based lead generation (FBLG; compounds 15–28) and DNA-encoded chemical library (compound 29) approaches. Compound 16 could be considered nature’s approach to FBLG. Compounds 25 and 26 represent the first known bacterial phosphopantetheine adenylyltransferase inhibitors with cellular activity against wild-type gram-negative bacteria.

Simocyclinone D8 (SD8, 16; Fig. 4 ), an aminocoumarin antibiotic, acts also as an inhibitor of GyrB unit ATPase activity. SD8 itself is not druglike because of its poor solubility and cell permeability. Its novel mode of action, however, is an attractive feature for drug development. The design of new DNA gyrase inhibitors, coumarin-quinolone hybrids (17; Fig. 4 ), is a result of the SD8-inspired FBLG approach. The coumarin moiety proved necessary for the observed inhibitory activity (IC₅₀ ~3 µM) of the hybrid compound, which is in part mediated by stabilization of a cleaved-DNA intermediate. ⁶⁹

The type II fatty acid biosynthetic pathway is an excellent target for antibacterial drug discovery. ⁷⁰ A Pfizer research group described pyridopyrimidine-based potent biotin carboxylase inhibitors using the FBLG approach. ⁷¹ This strategy has led to identification of a structurally diverse collection of weak-binding but ligand-efficient fragments for biotin carboxylase adenosine triphosphate–competitive inhibitors (compound 18; Fig. 4 ) ⁷² and a series of amino-oxazoles with antibacterial activity (compound 19; Fig. 4 ). ⁷² Research in the area of biotin carboxylase inhibitors by Mochalkin et al. ⁷² laid the groundwork for addressing several challenges, including modifications of fragments that are effective against gram-positive bacteria.^73,74

Recently, the FBLG approach has been directed at another enzyme of the fatty acid synthesis pathway: FabI in the gram-positive bacterium, S. aureus. Type II fatty acid biosynthesis, as has already been mentioned, is a drug target that has been validated by several studies^75,76 with clinically relevant molecules such as AFN-1252 (compound 20; Fig. 4 ) originally from Affinium and now under further development by Debiopharm.^76–78 The type II fatty acid biosynthesis cyclic pathway is characterized by the key enzyme enoyl acyl-carrier-protein reductase (FabI). ⁷⁹ FabI does not reveal a broad spectrum of druggability since it is known that not all gram-positive strains are susceptible to FabI inhibition. These strains use fatty acids directly from host serum rather than de novo synthesis. ⁸⁰ However, the same independence from fatty acid biosynthesis is not the case for S. aureus. S. aureus FabI (SaFabI) has been targeted in the past decades by several inhibitors, such as the classical drug, triclosan ( Fig. 4 ),^81,82 and analogs ⁸³ and via new drug classes. ⁸⁴ Recently, aminopyridine derivatives have been proposed as promising competitive inhibitors of SaFabI. ⁸⁵

A molecular drug target associated with coenzyme A is phosphopantetheine adenylyltransferase (PPAT; also known as CoaD). PPAT catalyzes the penultimate step in coenzyme A biosynthesis. PPAT is essential for bacterial growth. It shares relatively little sequence homology with the human ortholog, which makes is a good target for antibacterial drug development.^86–88,89 Two compounds from AstraZeneca (compounds 21a and 21b; Fig. 4 ) have demonstrated in vitro and in vivo inhibition of gram-positive bacterial growth. ⁹⁰ A research group from Novartis has identified inhibitors of gram-negative bacterial PPAT using an FBLG approach. ⁹¹ All fragment hits bind at the 4′-phosphopantetheine site of Escherichia coli PPAT. Lead compounds 23a and 23b ( Fig. 4 ), which have nanomolar IC₅₀ values, displayed modest cellular activity against E. coli ΔtolC. Of particular interest is methoxytryptamine derivative (fragment 22; Fig. 4 ) because of its ability to overlap with compound 23a but to bind differently with the enzyme. ⁹² Incorporation of a weakly basic amine in the lead structures led to enhanced gram-negative permeability (compound 24; Fig. 4 ). Combination of the amine moiety with groups able to engage in π-stacking interactions (attractive, noncovalent interactions between aromatic rings, due to the presence of p-bonds) with the enzyme has yielded compounds with subnanomolar binding affinity and minimal inhibitory concentration (MIC) against wild-type E. coli as low as 32 µg/mL (compound 25; Fig. 4 ). Compounds such as 25 and 26 ( Fig. 4 ) represent the first known bacterial PPAT inhibitors with cellular activity against wilt-type gram-negative bacteria. ⁹² However, these series of compounds has not been developed further, because of the lack of sufficient potency against other gram-negative species.

Identification of compounds using the FBLG approach for the DNA-G–SSB interaction has been reported. ⁹³ SSB protects single-stranded DNA during replication.^94,95 A fragment-based screen has led to the identification of inhibitors (compound 27 and 28; Fig. 4 ) of the primase/SSB-Ct interaction. ⁹³ Indole- and 1H-tetrazole–containing scaffolds have been identified as first-generation hits. The observation by the authors that compounds selected for binding to the C-terminal domain of DnaG primase also bind to other SSB-interacting proteins indicates that, in the future, compounds may be derived that bind to other SBB pockets in multiple protein targets. This feature might prove important for development of antibacterial compounds with a very low propensity for development of resistance. ⁹³

Additional examples of FBLG in the antibiotic area, as well as advances that have supported the evolution of FBLG methods, have been nicely summarized in a recent review. ⁹⁶

DNA-Encoded Chemical Libraries

A recent addition to lead-generation technologies is DNA-encoded chemical libraries or encoded library technology (ELT) to discover small-molecule binders to targets of interest. ⁹⁷ Use of ELT has led to identification of a large number of molecules obtained via chemical synthesis that bind to proteins from diverse chemotypes, as well as to an understanding of SAR directly from the screening output. ⁹⁷ A variety of methods for creating large DNA-encoded combinatorial libraries have been developed in the past two decades. Regardless of the individual synthesis techniques, ELT involves screening very large collections (typically >10⁸) of small-molecule compounds using a technology that involves the conjugation of chemical compounds to DNA fragments. ⁹⁸ This technology, however, has led to fewer clinical candidates as compared with other lead-generation approaches. This is mainly due to the current limited access to ELT by only several large pharmaceutical companies and contract research organizations ⁹⁸ and by limitations posed by the types of chemistry that can be used to construct libraries on DNA. The latter, coupled with the challenges associated with building of large libraries with low molecular weight and low lipophilicity, has been recently highlighted. ⁹⁹ However, successes in the area of infectious diseases^100,101 using this technique might be an indicator of its future impact on drug discovery. A pool of 11 DNA-encoded libraries comprising more than 66 billion on-DNA compounds has been used for selections against various forms of InhA (enoyl-acyl-carrier protein reductase, the primary target of isoniazid). Cofactor-specific inhibitors of InhA that do not require activation by KatG (unlike isoniazid), many of which had bactericidal activity in cell-based assays (e.g., 29; Fig. 4 ), have been identified. ¹⁰¹ There are several excellent recent reports summarizing the advances and outcomes using DNA-encoding library technology.^98,102,103

Old Structures: Delaying the AMR

Modifications to Previously Identified Antibiotic Compounds

The idea of taking a deeper look at some previously undeveloped leads is supported by the emerging view of the mechanism of action of daptomycin, a natural product (lipopeptide antibiotic), which, even though currently a therapeutic, initially was given little interest. ¹⁰⁴ Other examples of a revived interest toward several natural products include hygrobafinomycin ¹⁰⁵ (a member of a novel hygrolide [macrolide]) and tetrodecamycin ¹⁰⁶ (a member of the tetronate family, whose antimicrobial activities and biosyntheses have been recently reported).

The structural diversity and complexity of natural product antibiotics have inspired many successful total synthesis campaigns for their preparation. ¹⁰⁷ These syntheses are challenging, and aside from prompting the development of a synthetic methodology for overcoming a host of common obstacles, are for the most part impractical from the point of view of antibacterial development. However, the impressive achievements in synthetic methodology over the decades allow current efforts in the development of total synthesis of antibacterials to change in synthetic strategy. A shift from the original goal of reaching solely the molecule of interest (the synthetic target) to preparation of target analogs in order to establish SAR has started to occur. The newest developments in total synthesis of the main classes of antibiotics have been thoroughly reviewed, ¹⁰⁷ and a special issue of the Journal of Organic Chemistry has been recently published that focused on both synthetic and biosynthetic chemistry of natural products, with an emphasis on antibiotics. ¹⁰⁸

One recent success in the deliberate design of an antibacterial compound (vancomycin) based on the total synthesis may serve as a confirmation for applicability of the modification of existing natural product scaffolds in making the latter less prone to resistance development. ¹⁰⁹ Certain features of glycopeptides have been identified that enable this class of antibiotics to avoid many mechanisms of resistance.^110,111 Based on this knowledge, incorporation of peripheral structural modification in vancomycin led to enabling multiple synergistic mechanisms of action. Bacteria sense the presence of the antibiotic, ¹¹² which triggers a change in the peptidoglycan termini from D-Ala-D-Ala to D-Ala-D-Lac. ¹¹³ This remodeling of the peptidoglycan allows for a 1000-fold reduction of the binding affinity of vancomycin for the altered target,^114,115 resulting in a 1000-fold loss in antimicrobial activity. The three modifications made in the structure of vancomycin (triply modified vancomycin; Fig. 1 ), two peripheral ones (with synergistic mechanisms of action) and one pocket modification, have led to a vancomycin analog with three independent mechanisms of action, only one of which is dependent on D-Ala-D-Ala/D-Ala-D-Lac binding. This triply modified vancomycin demonstrated increased antimicrobial potency against VanA VRE (>6000-fold) as well as reduced susceptibility to resistance. ¹⁰⁹ This accomplishment is a result of the total synthesis of the starting pocket modified aglycon(s) (26 steps), ¹¹⁶ enzymatic installation of the disaccharide (2 steps), ¹¹⁷ and subsequent addition of the two peripheral modifications (2 steps), demonstrating an impressive synthetic methodology. Peripheral structural changes lead to improved membrane binding. Traditionally, drug development has been focused on optimizing target affinity, selectivity, and protein binding. ¹¹⁸ A more recently explored approach is adding structural motifs, especially to drugs acting on membrane-associated targets (e.g., vancomycin). This could lead to increased drug concentration at the target site, in addition to improved membrane binding.^119,120

New Life for Old Friends: β-Lactam Drug Combinations

Triply modified vancomycin is an example of a single molecule having three independent mechanisms of action, which allow for retention of its antibacterial activity in both susceptible and resistant bacteria. Identifying a single molecule able to block several resistance activities is challenging. That challenge could be attenuated by identifying prominent modes of resistance in order to develop specific inhibitors for them. ¹²¹ Illustrative to that approach is the success of the serine β-lactamase inhibitors. The first compound of this class to find clinical use in combination with amoxicillin (Augmentin) or ticarcillin (Timentin), clavulanic acid, ¹²² is produced by Streptomyces clavuligerus. ¹²³ Clavulanic acid was discovered in 1976 by Beecham. Since then, different types of β-lactamase inhibitors have been explored, which led to the clinical implementation of penicillanic acid sulfones, sulbactam and tazobactam, in combination with various penicillins. Continued interest in β-lactams as antimicrobials is due to their low toxicity in humans and the good pharmacokinetics/pharmacodynamics profiles they enjoy. That is confirmed by the FDA approval of three new combinations of β-lactam (two representatives of the cephalosporins and one meropenem with intrinsic antibacterial activity)/β-lactamase inhibitor for the past 5 y, as opposed to one cephalosporin in the past 8 y. β-lactams with intrinsic activity of the β-lactam/β-lactamase combinations consist of a cephalosposin, Zerbaxa (2014), Avycaz (2015), and one of meropenem (2017). The non–β-lactam diazabicyclooctane (DBO) avibactam (30; Fig. 5 ), is the first in class approved for clinical use. Similar to clavulanic acid and sulbactam/tazobactam, DBO covalently inhibits the active site serine in the serine β-lactamases. A tetrazole-based non–β-lactam inhibitor (31; Fig. 5 ) binds noncovalently to the CTX-M, the most prevalent family of extended spectrum serine β-lactamases (K_i 89 nM). ¹²⁴

Figure 5.

Examples of β-lactamase inhibitors (compounds 31–34) used in combination with β-lactam antibiotics to prolong the clinical relevancy of the latter. Molecular hybrids (35–44) prepared by coupling of the different entities to improve the effectiveness of the individual antibiotics.

Although the attenuation of resistance to serine β-lactamases has continued for more than 30 y of clinical implementation, identification of a clinical candidate to the metallo-β-lactamases (MBLs) has been challenging. MBL inhibitors with good in vitro activity have been reported^125–131 but have not demonstrated efficacy in animal infection models. An exception to this is the natural product aspergillomarasmine A (WAC0138, 32; Fig. 5 ), first isolated from Aspergillus sp. ¹³² and later synthesized^133,134 with in vivo activity. WAC0138 is a strong metal ion chelator, whose further development might be limited by toxicity (the 50% lethal dose in mice is 159.8 mg/kg). ¹³⁵ Another MBL inhibitor 33, ANT431, (Antabio SAS; Fig. 5 ) with in vivo activity against New Delhi MBL 1 (NDM-1), can potentiate the activity of meropenem against a broad range of carbapenem-resistant Enterobacteriaceae and can restore its efficacy against MBL-producing E. coli. ¹³⁶ ANT431 has demonstrated good selectivity for the MBLs over nonbacterial metallo-enzymes, which, in addition to its druglike physicochemical properties, makes it a good candidate for future optimization. The metal-chelating thioether-based MBL inhibitor (34; Fig. 5 ; IC₅₀ value of 0.41 µM against L1 MBL) has been reported ¹³⁷ to restore the antimicrobial properties of cefazolin to that of susceptible, non–MBL-producing E. coli cells. Virtual screening methodologies^138,139 have also been applied for identification of non–β-lactam BML inhibitors, including nonmetal chelating structures, in addition to the use of a recombinant NDM-1 strategy. ¹⁴⁰

Molecular Hybridization

Molecular hybridization is the chemical fusion of multiple pharmacophores via covalent bonds into a single molecule. The main reason for implementing chemical hybridization, as opposed to combining biologically active molecules, is associated with efforts to achieve a benefit in terms of therapeutic potential, potency, mode of action, pharmacokinetics, and so forth.^141,142

An important incentive for applying the molecular hybridization approach is the delay of antibiotic resistance. Undoubtedly, there will always be an element of uncertainty as to whether combinations of different entities may alter the effectiveness of individual antibiotics. However, this strategy holds the promise of achieving structural novelty, which is a great challenge for medicinal chemistry and an even greater challenge for the design of antimicrobial compounds. ¹⁴² Several examples of hybrids^141–148 as antibacterials are discussed briefly below.

The aziridine-based hybrid molecules (35; Fig. 5 ), in which the aziridine moiety has been introduced as a synthetic precursor of biologically relevant amino propane side chains, have demonstrated antiviral (while coupled with β-lactams) and antimalarial activity (when coupled with quinolines). ¹⁴¹

β-lactams/benzothiazole hybrids ¹⁴⁶ such as 36 ( Fig. 5 ) have shown modest antimicrobial activities against gram-positive and gram-negative bacterial strains. The presence of a naphthyl group on C-4 and a phenyl on N-1 of the β-lactam ring (37; Fig. 5 ) lead to antimicrobial activity against both gram-positive S. aureus and gram-negative bacteria E. coli or P. aeruginosa. ¹⁴⁶ The broad activity of hybrids 36 and 37 may be indicative of a lack of specificity and a potential toxicity liability. However, all synthesized compounds show no significant destructive effect on red blood cells and a relatively reduced cytotoxicity effect on the hepatocellular carcinoma cell line (HepG2) even at a final concentration of 250 µM for 72 h. ¹⁴⁶

Earlier examples using phenothiazine and 1,3,4-thiadiazole hybrids have demonstrated anti-Mtb activity with an MIC of 0.8 µg/mL of 38 ( Fig. 5 ; R² = methyl or n-propyl; R¹ = H or Cl). SAR studies showed that alkyl groups (methyl or n-propyl) and substituted phenyl groups (4-methyl, 4-Cl and 4-F) enhance activity. ¹⁴⁷

Triazole hybrids have demonstrated promising antibacterial activity. Sugar-triazole hybrids show a broad-spectrum antibacterial (in addition to antifungal) activity. Compound 39 ( Fig. 5 ) has the best antibacterial activity of this series, with an MIC of 0.78 µg/mL against S. aureus and Klebsiella pneumoniae and 3.12 µg/mL against methicillin and vancomycin-resistant S. aureus. ¹⁴⁸ From the bioinformatics studies performed, it appears that penicillin binding protein-2 (PBP-2) is the most likely target of these compounds. ¹⁴⁸

Dual anti-Mtb activity (growth inhibition and efflux pump inhibition) and a synergistic effect with antitubercular agents of hybrid triazoles has been reported, ¹⁴⁹ with an MIC of 4 µg/mL and 8 µg/ mL, respectively.

The dehydroacetic acid-chalcone-1,2,3-triazole hybrids have demonstrated an enhanced antimicrobial activity as compared with the dehydroacetic acid alone. ¹⁵⁰ Several compounds (40; Fig 5 ) have activity against E. coli with an MIC of 0.003 µM/mL, which is similar/better than the reference drug ciprofloxacin (MIC of 0.0047 µM/mL). Based on molecular modeling studies, these series of compounds appears to be DNA gyrase inhibitors. ¹⁵⁰

As a continuation of earlier studies on quinolone-containing hybrids with activity against both gram-positive and gram-negative organisms, ¹⁵¹ quinolone-thiourea, ¹⁵² quinoline-oxadiazole, ¹⁵³ and quinolone-benzimidazole ¹⁵⁴ hybrids have been prepared. In the former series, MICs of 41 ( Fig. 5 ) for E. coli (MIC₅₀ of 2.63–3.95 µM) and S. aureus (MIC₅₀ of 3.95–5.26 µM; Fig. 5 ) are similar. This compound also had anti-MRSA activity with an MIC₅₀ of 11.44 µM. Hybrid 41 is nontoxic to Galleria mellonella larvae (concentrations of up to 1000 µg/mL).

The quinolone-oxadiazole hybrids (e.g., compound 42; Fig. 5 ) have demonstrated anti-Mtb activity. ¹⁵³ Structures with aliphatic groups or alkyl attached to the aromatic ring groups obtained better results (MIC of 0.4–1.05 µM). This compound shows the highest anti-Mtb activity, with an MIC of 0.4 µM and a selectivity (SI) of >610. ¹⁵³

Benzimidazole-quinoline hybrids (e.g., 43; Fig 5 ; MIC of 1 µg/mL) do not intercalate into DNA isolated from drug-resistant P. aeruginosa but are able to cleave DNA effectively. ¹⁵⁴

Anti-Mtb activity (MIC of 0.12 µg/mL) has been also reported for coumarin-theophylline hybrids (44; Fig. 5 ). ¹⁵⁵ Molecular modeling studies against Mtb 4DQU enzyme show good binding interactions and are in agreement with the in vitro results.

New Targets Inspired by Natural Product Leads

All of the clinically relevant β-lactam antibiotics, including those recently approved by the FDA, contain an ionizable group either in the proximity (carboxylic acid, bicyclic, penicillin-like structures) or on (sulfonic acid, monobactams) the lactam nitrogen of the β-lactam ring ( Fig. 6 ). ¹⁵⁶ Until recently, it was generally accepted that for β-lactams to exert bactericidal activity, they must contain a scaffold, which specifically has an ionizable group at the lactam nitrogen within 3.6 Å of the β-lactam carbonyl carbon. The aforementioned ionizable group allows for β-lactams to be recognized by transpeptidases. However, there now appear to be exceptions to this scaffold requirement, since N-alkylthiolated β-lactams 45 ( Fig. 6 ) possess inhibitory, although not cidal, narrow-spectrum antimicrobial activity. ¹⁵⁷ Subsequent reports have confirmed that once that ionizable group is removed from the lactam nitrogen, a variety of novel molecular targets begin to emerge.^158,159 These N-alkylthiolated lactams have inspired preparation of hydroxamate-containing lactams 46 ( Fig. 6 ) with anti-Mtb activity and as inhibitors of β-lactamases. ¹⁶⁰ β-lactamase–resistant monocyclic β-lactams (47; Fig. 6 ) that carry an arylthio group at C4 have been also reported.^161,162 These N-unsubstituted lactams exhibit inhibitory and cidal activity against serine β-lactamase producing wild-type strain Mtb and multiple β-lactamase producing Moraxella catarrhalis clinical isolates. ¹⁶² These examples of nontraditional β-lactams, none of which possess an ionizable group at the lactam nitrogen, confirm that natural products inspire the design and preparation of new synthetically accessible structures in addition to strategies. ¹⁶³ It has been demonstrated that a given function can be achieved with many different structures, as a “design-for-function strategy becomes a powerful tool for creating totally new targets inspired by natural product leads or by abiological needs.” ¹⁶³ β-lactam antibiotic structures are one illustration of our understanding that while preserving/improving a function/activity, chemists can move from the natural realm to the design of structures, often from complex to very simplified ones. In essence, form (structure) follows function, and this approach has been called “function-through-synthesis informed design.” ¹⁶³ The very attractive feature of this approach is simplification of the structure (as compared with the natural structure that inspires that design), which allows for a practicality and economy in achieving the desired bioactivity. ¹⁶³ Another confirmation of how the function affects the antimicrobial activity/molecular target identification is the use of β-lactones as chemical probes. β-lactams occur relatively rarely in nature. There are, however, numerous bioactive β-lactones found in nature. Similarly to β-lactams, β-lactones are of microbial origin ¹⁶⁴ and are acylating agents of serine enzymes. The latter chemical reactivity allows for β-lactones to inactivate β-lactamases. ¹⁶⁵ Synthetic β-lactones (48; Fig. 6 ) have been shown to inhibit proteins involved in metabolism, antibiotic resistance, the β-lactamase PBP4* in Bacillus subtilis, and a virulence-associated protein, caseinolytic protease P (ClpP).^166–171 The determination of D,D-carboxypeptidase as one of the 13 targets of a β-lactone probe in Mycobacterium smegmatis ¹⁷¹ led to the recent discovery of β-lactones as a privileged scaffold for the generation of PBP-selective probes. These β-lactones (49; Fig. 6 ) used for imaging of the essential proteins, PBP2x and PBP2b, in Streptococcus pneumonia, similarly to the aforementioned β-lactams, also lack an ionizable group, which is considered a prerequisite for binding to trans-peptidases (PBP). ¹⁶⁵

Figure 6.

Simplified structures as compared with their naturally occurring counterparts, β-lactam and b-lactones (compounds 45–49), with altered mode/target of action and antimicrobial peptides (50–54) exemplifying design-to-function strategies.

Another example that could be considered a design-to-function strategy is the synthetic mimic of first-line host defense mechanisms, such as the antimicrobial peptides (AMPs). These peptides are evolutionary conserved and interact with bacterial membranes, causing cell death. Their main mode of action is to impair the function of the membrane. ¹⁷² Most AMPs are amphiphilic, with a positive face interacting with the negatively charged bacterial membrane and a hydrophobic face that allows for insertion into the membrane and interaction with the apolar acyl chains of the bilayer.^173–177 Since the membrane is their the primary target, development of resistance is expected to be limited. ¹⁷⁸ Several scaffolds, such as D-L peptides, β-amino acid helices, and antimicrobial polymers,^179–181 as well as tunable peptide macrocycle assemblies introduced into known AMPs (e.g., buforin II), ¹⁸² have been explored to improve bioavailability and proteolytic degradation of the AMPs. ¹⁸³ Another significant pharmaceutical issue with AMPs is systemic toxicity as well as the difficulty and expense of manufacturing, which have a negative effect on their clinical progress. ¹⁸³ Because of these limitations, therapeutic development of the peptides has been limited largely to topical or local administrations. One of the clinically advanced (completed phase 1 and two phase 2 clinical trials, with a phase 3 trial currently planned) small mimics of host defense proteins is brilacidin 50 ( Fig. 6 ; PMX30063, Cellceutix, Inc, Beverly, MA). It is being developed for acute bacterial skin and skin structure infections as an intravenous agent. Low cytotoxicity of brilacidin (and analogs) has qualified it for entering clinical studies as a systemic intravenous antibiotic. Brilacidin is highly active against staphylococcal and other gram-positive bacteria with MIC₉₀ values ≤1 µg/mL against methicillin-sensitive S. aureus and MRSA and coagulase-negative S. aureus. ¹⁸⁴ To address the aforementioned physicochemical requirements of the AMPs, brilacidin is an arylamide having two guanidino and two pyrrolidine cationic end groups and two trifluoromethyl hydrophobic sidechains. It appears that efforts directed at developing amphiphilic compounds that summarize the structural and biological properties of AMPs on small abiotic scaffolds have paid off. Successful examples of small-molecular mimics for AMPs include Destiny Pharma Ltd.’s exeporfinium chloride (XF-73; phase 1) ¹⁸⁵ and Lytix Biopharma’s ¹⁸⁶ LTX-109, 51 ( Fig. 6 ; phase 2).^186,187

Another alternative for development of AMPs is to incorporate unnatural functionalities into amino acid side chains to disguise the cationic AMPs.^17,25 Amphiphilic AMPs are mimetic with antimicrobial activity against gram-positive bacteria, based on the natural xanthone of α-mangostin and its semisynthesized xanthone-based derivatives, including 52 (AM016) and 53 (AM218), Fig.6, have been reported.^188–191 Compounds of scaffold 54 ( Fig. 6 ) have shown good antimicrobial activity against gram-positive bacteria, including MRSA and VRE (MICs of 0.78–6.25 µg/mL) with a rapid bactericidal effect, low toxicity, and no emergence of drug resistance. The total synthesis of compounds 54 ( Fig. 6 ) is advantageous because it allows for synthetic simplicity, low cost, and versatility. ¹⁹² The strategy to prepare amphiphilic mimic, rather than modification via semisynthesis of amphiphilic xanthones using natural α-mangostin, provided several poly(guaninylureas) 55 ( Fig. 6 ) with antimicrobial activity. ¹⁹³ This new polymer design introduces both positive charge and hydrophobicity into the flexible polymer backbones. The initial SAR showed that poly(guanylurea piperazine)s exhibit excellent antimicrobial activity against different types of bacteria with high selectivity. ¹⁹³

Addressing AMR by Other Combinatorial Approaches

Bacteria can adapt to a single antibiotic during treatment, leading to different outcomes: either bacteria become more resistant or more susceptible to other antibiotics. ¹⁹⁴ Antibiotic combinations have been used for decades to delay the emergence of resistance. It has been demonstrated that resistance to certain antibiotics in vitro, particularly aminoglycosides, enhances the sensitivity to other antimicrobial agents, as indicated by an increase in the MIC.^195,196 However, antibiotic combinations may also promote the evolution of drug resistance by increasing the danger of superinfection due to the coevolution of MDR variants and sensitization to other antibiotics. ¹⁹⁷ Therefore, careful design of effective antibiotic combinations is an important approach to delaying the AMR. In addition to the aforementioned antibiotic combinations, which directly inhibit resistance (e.g., β-lactam/β-lactamase combinations), there are other antibiotic combinations that attempt to delay it. Some are directed toward inhibition of the bacterial resistome ¹⁹⁸ and others to bypass resistance mechanisms. Several examples of the latter strategy are given below. Loperamide (an antidiarrheal drug, nonantibiotic) enhances the activity of tetracycline by increasing its intracellular concentration. ¹⁹⁹ Tetracyclines require a pH gradient to cross the gram-negative outer membrane. Loperamide increases the DpH gradient, which in turn leads to increased uptake of tetracyclines. ¹⁹⁹

The flavonoid gallic acid enhances the activity of halogenated quinolones by up to 11,000-fold against S. aureus through an unknown mechanism. ²⁰⁰ Synergistic with kanamycin are synthetic diamines, which, analogously to the aforementioned AMPs, depolarize the cytoplasmic membrane and increase the permeability of the bacterial outer membrane. These diamines are bactericidal against both gram-positive and gram-negative bacteria. ²⁰¹

Thioridazine in combination with antibiotics to which Mtb isolate was previously resistant for therapy of MDR, extensive drug resistance, and most likely, totally drug resistance, leads to potentiation of the Mtb antibiotic therapy. ²⁰² Thioridazine and methylene blue belong to the phenothiazine class of compounds. Thioridazine is therapeutically safe, and similarly to chlorpromazine, it has been in use for more than six decades to control psychosis. Shortly after the synthesis on methylene blue (1883), Paul Erlich used it for staining live cells and found that it could reduce movement of microorganisms. ²⁰³ Thioridazine has an MIC of 8 to 15 μg/mL against Mtb (H37Rv, susceptible strain). Most importantly, it can totally reverse the resistance of Mtb toward isoniazid, ²⁰⁴ mainly but not limited to interference with the overexpressed efflux pump genes of resistant strains. ²⁰²

A recent profiling of close to 3000 dose-resolved combinations of antibiotics, human-targeted drugs, and food additives in six strains from gram-negative pathogens (E. coli, Salmonella enterica serovar Typhimurium, and P. aeruginosa) has been accomplished. ²⁰⁵ The goal of this systematic study was to identify general principles for antibacterial drug combinations and understand their potential. More than 70% of detected drug-drug interactions are species specific, and 20% display strain specificity, leading to the conclusion that there is a potential for narrow-spectrum therapies. Most importantly, this study shows that synergies are more preserved and are enhanced in drugs that target the same process. ²⁰⁵ This study has demonstrated that nonantibiotic drugs hold promise as adjuvants, even against MDR clinical isolates. Whether such synergies would have clinical relevance remains to be determined.

Conclusion

The discovery of scaffolds best fitted for antibacterial drug development is challenging. Differences in the requirements for drug development for gram-negative versus gram-positive increases the challenges innate to any drug discovery effort, especially for broad-spectrum antibacterials. The AMR complicates the drug discovery process even further. That alone has reduced the enthusiasm for antibiotic drug development in the past decades. However, there appears to be a resurgence of antibacterial efforts. New antibiotic scaffolds, such as the ones described here, are being synthesized, as well as discovered. More than a century has passed since Paul Ehrlich introduced the discovery of a magic bullet, able to destroy tumor cells and microorganisms, with the help of organic chemistry. Making as many as possible derivatives of a given compound might have a somewhat different interpretation today. There are limited time and resources to synthesize every possible compound. Synthetic chemistry is a prerequisite of antimicrobial research. It has and will always play a key role in this area. The development of new synthetic strategies allows for efficacious preparation of new antibiotics. The existence of an expeditious feedback mechanism between the preparation of a given synthetic organic entity and changes in cell morphology/gene expression that this synthetic product could trigger becomes imperative for future success of the synthetic efforts of antimicrobials. This rapid feedback mechanism between chemistry and biology might allow for delaying of AMR for as long as possible.