Introduction

Tuberculosis (TB), caused by Mycobacterium tuberculosis, has long been considered as one of the most serious infectious diseases. Worldwide, TB is one of the top ten causes of death, and millions of people continue to become ill with TB every year. In 2017, TB caused an estimated 1.3 million deaths among HIV-negative people. Globally, the best estimate is that 10 million people developed TB in 2017: 5.8 million men, 3.2 million women, and 1.0 million children [1]. The first-line antituberculosis drugs for adults recommended by the World Health Organization (WHO) are isoniazid, rifampicin, pyrazinamide, ethambutol, and streptomycin. TB resistance to first-line therapy has increased in recent years, due in-part to long treatment times and poor patient compliance increasing the need for new therapeutic options [2]. In 2017, WHO published its first ever list of antibiotic-resistant, “priority pathogens,” a catalog of 12 bacterial families that pose the greatest threat to human health. They include well-known gram-positive organisms such as Enterococcus faecium and methicillin-resistant Staphylococcus aureus. Thus, it is important to identify novel biological targets and develop new compounds for clinical use against TB and/or against multidrug-resistant (MDR) bacteria [3].

Isono and his coworkers (Antibiotic Laboratory, RIKEN), who isolated and developed the nucleoside pesticide polyoxin in 1965 [4], studied and published comprehensive reviews on nucleoside antibiotics [5, 6]. Dr Isono has also explored the isolation of specific nucleoside inhibitors against bacterial peptidoglycan, modeled after polyoxin, using as substrates— Escherichia coli Y-10 particulate enzyme, UDP-MurNAc pentapeptide from Bacillus cereus, and 14C-UDP-GlcNAc. In 1985, he and his collaborators, including myself, reported the first specific nucleoside inhibitor against peptidoglycan biosynthesis that acted through inhibition of MraY [phospho-N-acetylmuramoyl-pentapeptide-transferase (translocase I, EC 2.7.8.13)], and named it liposidomycin based on its structure [7,8,9,10,11].

This review summarizes the history of peptidoglycan inhibitors, with the discovery and study of liposidomycin and related compounds that are liposidomycin-type liponucleosides that target MraY. Dr Isono’s vision and goal may be realized through this work, because the liposidomycin analog caprazamycin [12] has new antibiotic potential against M. tuberculosis, and CPZEN-45 [13], a new derivative that overcomes the disadvantage of caprazamycin, is a clinical candidate for anti-TB use in the near future.

Bacterial peptidoglycan biosynthesis and peptidoglycan inhibitor molecules

Bacterial peptidoglycan biosynthesis has been a promising therapeutic target for antibiotics ever since the discovery of penicillin in 1929 [14], because no counterpart to bacterial peptidoglycan exists in eukaryotic cells [10, 11, 15]. Bacterial peptidoglycan comprises a β-1,4-linked glycan of alternating N-acetyl-glucosamine (GlcNAc) and N-acetyl-muramic acid (MurNAc) sugars. UDP-GlcNAc is used for the assembly of both peptidoglycan in bacteria and chitin in fungi. UDP-MurNAc is biosynthesized from UDP-GlcNAc by the addition of an enolpyruvyl group to the 3’-hydroxyl, via transferase MurA (UDP-GlcNAc enol pyruvate transferase) [16], followed by reduction to a lactyl sidechain by reductase MurB (UDP-N-acetylglucosamine-enolpyruvate reductase) [17]. Fosphomycin (formerly, phosphonomycin) [18] inhibits Mur A (Fig. 1a).

Fig. 1
figure 1

Bacterial peptidoglycan biosynthesis (a) and MraY (tranlocase I) reaction (b)

Full size image

Assembly of the cytoplasmic peptidoglycan precursor UDP-MurNAc-L-Ala-D-Glu-L-Lys is accomplished using a series of ATP-dependent respective amino acid ligases (MurC, MurD, and MurE) that add l-alanine, d-glutamic acid, and l-lys sequentially to the lactyl side chain of UDP-MurNAc. The final two amino acids (D-Ala-D-Ala) are synthesized by D-Ala-D-Ala ligase (Ddl), and this completes the synthesis of UDP-MurNAc-L-Ala-D-Glu-L-Lys-D-Ala-D-Ala (Park’s nucleotide) through the action of the UDP-MurNAc-L-Ala-D-Glu-L-Lys (or meso-DAP):D-Ala-D-Ala (MurF). D-Cycloserine (D-4-amino-3-isoxazolidone) [19] inhibits both Ddl and alanine racemase (Alr), which is an enzyme catalyzing the conversion of l-alanine to d-alanine [20] (Fig. 1a).

The first step in the intramembrane reactions is the transfer of phospho-MurNAc-pentapeptide from UDP-MurNAc-pentapeptide to undecaprenyl phosphate (C55-P), catalyzed by translocase I (MraY) (Fig. 1a, b). The reaction is reversible, requires the presence of Mg2+ ions, and leads to the formation of undecaprenyl-diphospho-MurNAc-pentapeptide (lipid I) [21]. Bacterial genome sequencing revealed that only one copy of the mraY gene exists in E. coli. Uridyl liponucleoside antibiotics, such as tunicamycin [22], liposidomycin [7], and caprazamycin [12], inhibit this enzyme. Similar liponucleoside antibiotics, such as A-84830A [23], A-97065 [24] A-94964 [25], A-90289 [26], and muraminomycin [27], have also been reported. Uridyl peptide antibiotics, such as capuramycin [28], mureidomicin [29], pacidamycin [30], napsamycin [31], muraymycin [32], A-102395 [33], and sansanmycin [34], are also known as translocase I inhibitors. Although amphomycin [35] is not a nucleoside antibiotic, it was reported to inhibit MraY [36].

A residue of GlcNAc from UDP-GlcNAc is then attached to the undecaprenyl-diphospho-MurNAc-pentapeptide (lipid I). This reaction is catalyzed by a glycosyl transferase enzyme known as translocase II (Mur G), yielding lipid II. Two highly modified cyclopeptides composed of 17 amino acids, enduracidin (or enramycin) [37] and ramoplanin [38], bind to lipid II competitively (Fig. 1a). After the assembly of lipid II is flipped from the cytoplasmic face of the membrane to the external face by MurJ (flippase) [39] (Fig. 1a), it is then transformed into peptidoglycan by penicillin-binding proteins that have transglycosylation and transpeptidation functions. Vancomycin [40] acts by binding the terminal amino acid residue L-Lys-D-Ala-D-Ala-COOH, and β-lactams such as penicillin [14] inhibit the transpeptidase. The cross-linking provides the structural rigidity of a mature peptidoglycan, which is required to maintain the cell shape and prevent cell lysis. Bacitracin [41] inhibits the dephosphorylation of lipid pyrophosphate in the late step of the lipid cycle after completing the biosynthesis of peptidoglycan (Fig. 1a).

Gram-positive bacteria are encompassed by a thick peptidoglycan cell wall. However, gram-negative bacteria are encompassed by an additional outer membrane. Mycobacterium spp. such as M. tuberculosis exhibit cell wall structures with unusual complexity. The mycolyl-arabinogalactan-peptidoglycan (mAGP) complex represents the cell wall core structure creating a lipophilic pseudo-outer membrane [2, 10, 42, 43] (Fig. 2). Small molecules can penetrate the cell membrane of gram-positive bacteria. However, few antibiotics are available against gram-negative bacteria and/or Mycobacterium spp. because of the limited permeability of antibiotics through their outer membranes. In the biosynthetic pathway of peptidoglycan and mycolyl-arabinogalactan in M. tuberculosis, the membrane proteins translocase I (MurX: the ortholog of MraY in various bacteria) and WecA (polyprenyl phosphate-N-acetylglucosamine-1-phosphate transferase; the ortholog of TagO in Bacillus subtilis) are involved and are important drug targets [43]. Caprazamycin inhibits the former enzyme and CPZEN-45 inhibits the latter enzyme, as described in Section “Caprazamycin, A-90289A, and muraminomycin as liposidomycin analogs with specific MraY inhibition activity” (Fig. 2).

Fig. 2
figure 2

Cell wall structure of Mycobacterium tuberculosis

Full size image

Although bacterial peptidoglycan biosynthesis is a well-proven target for antibiotic action, it is surprising that no clinical antibiotics are available that target translocase I (Fig. 1a, b). In the case of mureidomycin A, a translocase I inhibitor, the intrinsic resistance of bacteria is due to low permeability [44]. TB drug resistance has emphasized the need to identify new targets for antibiotic action that lack cross resistance to existing therapies, such as inhibitors of translocase I [10, 11, 15, 43].

Structural features of polyoxin as a chitin synthase inhibitor and tunicamycin as an MraY inhibitor

Chitin is a linear homopolymer of β-(1,4)-linked GlcNAc, formed biosynthetically from UDP-GlcNAc by chitin synthases (chs) 1, chs 2, and chs 3, which were discovered in studies of Saccharomyces cerevisiae mutants with disrupted chitin synthase genes [45]. The absence of all three chs is uniformly lethal. These chitin-containing structures affect the viability of fungi and are not present in mammalian cells. Chs is located in the plasma membrane, requiring that inhibitors be transported into the cell. Isono et al. isolated and developed a peptidyl nucleoside antibiotic, named polyoxin [4,5,6], as a pesticide that inhibits chitin to combat rice blast disease caused by Piricularia oryzae. Compared to the structure with uridyl liponucleoside antibiotics, it mimics only UDP-GlcNAc, a substrate for chs, and is a competitive inhibitor of chs [46] (Fig. 3a).

Fig. 3
figure 3

Structures of polyoxin (a), tunicamycin (b), liposidomycin (c), caprazamycin (d), A-90289 (e) and muraminomycin (f)

Full size image

Tunicamycin was isolated from Streptomyces lysosuperificus (later Streptomyces chartreusis) as an antiviral compound by Tamura and Takatsuki [22]. The process of its discovery and various biological activities of tunicamycin have been reported and reviewed in a book entitled Tunicamycin [47]. Briefly, the structure is composed of uracil, fatty acid, and the two sugars of GlcNAc and undecadialdose (Fig. 3b). Gram-positive bacteria such as the Bacillus genus are sensitive to tunicamycin (MIC: 0.1–20 μg/mL) through inhibition of MraY. It has since been found that tunicamycin inhibits the human enzyme UDP-GlcNAc:dolichyl-phosphate GlcNAc phosphotransferase [GlcNAc-1-P-transferase (GPT or DPAGT1) (EC 2.7.8.15)], which catalyzes the first and committed step of N-linked protein glycosylation in the endoplasmic reticulum membrane. Thus, tunicamycin showed severe toxicity in eukaryotic cells, and is used to study the endoplasmic reticulum stress response as a bioprobe [48]. The recent studies have compared of the structures of tunicamycin bound to DPAGT1 and MraY enabling the design of analogues with altered lipid side chains imparting selectivity for MraY over DPAGT1 [49,50,51]. Chemically modifying tunicamycin, hydrogenating the N-acetyl double bond or hydrogenating both the N-acetyl and uridyl double bonds have also been reported to impart less toxic to eukaryotic cells, while retaining their antibacterial activity [52].

Liposidomycin as the first specific MraY inhibitor in bacterial peptidoglycan biosynthesis

Dr Isono has attempted to isolate a specific nucleoside inhibitor like polyoxin against peptidoglycan biosynthesis using enzyme assays. I had the chance to study natural product chemistry in RIKEN for 2 years as a visiting researcher from Snow Brand Milk Products., Co. Ltd (now Meg Milk Snow Brand Co. Ltd) after graduating from a master’s program in the laboratory of biological chemistry (Profs. Shimura and Mizuno) at Tohoku University in 1984. Initially, Dr Isono provided ten kinds of actinomycete strains that had inhibitory activity against peptidoglycan, mannan, and β-1,3-glucan biosynthesis in order to isolate each inhibitor. To isolate new compounds, a well-designed screening system relying upon basic studies and unique microorganisms (natural sources) is vital, as described by Dr Omura, who won the 2015 Nobel Prize for discoveries concerning a novel therapy against infections caused by roundworm parasites [53]. One of the strains, a liposidomycin producer, called the RK-1061 strain, was later identified as Streptomyces griseosporeus [7]. The RK-1061 strain produces more than ten kinds of liposidomycins that show peaks during high-performance liquid chromatography (HPLC); fortunately, three kinds of new antibiotics, liposidomycins A, B, and C, were isolated in 1985; their structures were elucidated by Ubukata et al. [8]. They each possessed 5′-substituted uridine, 5-amino-5-deoxyribose-2-sulfate, and perhydro-1,4-diazepine moieties but differed in the structure of the lipid side chains, representing one of the most bizarre structures for nucleoside antibiotics [54, 55] (Fig. 3c). The Mycobacterium genus is sensitive to liposidomycin [Mycobacterium phlei: minimum inhibitory concentration (MIC) = 1.6 μg/mL] [7]. The discovery of new liposidomycin compounds, supervised by Dr Isono, led me to the isolation of new biologically active compounds (bioprobes) and opened the door to my work in chemical biology.

The molecular target of liposidomycin C was found to be MraY (translocase I), via the traditional method using 14C-UDP-MurNAc-pentapeptide without UDP-GlcNAc [9]. Liposidomycin B weakly inhibited the formation of Dol-p-p-GlcNAc (IC50, 20 μg/mL) but did not inhibit dolichyl phosphoryl glucose or dolichyl phosphoryl mannose, the precursors for mammalian glycoprotein synthesis, even at 400 μg/mL. By contrast, tunicamycin strongly inhibited the formation of Dol-p-p-GlcNAc (IC50, 0.03 μg/mL) but only weakly inhibited the formation of bacterial lipid intermediate I (IC50, 44 μg/mL) [56]. This was the first specific MraY inhibitor that had no cytotoxicity or inhibitory activity against a similar reaction in eukaryotes. Thereafter, several types of MraY inhibitors were isolated from actinomycetes, and the organic synthetic and/or biosynthetic studies have continued progressing [10, 11].

The molecular mechanism of the inhibition of translocase I by liposidomycin B has been investigated. Using a continuous fluorescent assay, Brandish and Bugg determined that liposidomycin B was noncompetitive versus the UDPMurNAc-dansyl-pentapeptide soluble substrate but competitive against the dodecaprenyl phosphate lipid substrate, which differs from the mechanism of tunicamycin [57].

Each of the original three liposidomycins (A, B, and C) has a sulfate moiety at the hydrophilic 2″-aminopentose. Although they showed potent inhibition activity against peptidoglycan biosynthesis, their antimicrobial activities were weak. Therefore, we tried to produce nonsulfated liposidomycins at 2″-aminopentose by changing the fermentation conditions of the RK-1061 strain [58]. Four new liposidomycins were classified as types I–IV based on their structures. The type I compound is the original liposidomycin that has both the sulfate and 3-methylglutaric acid moieties. The type II compound has the sulfate moiety but not the 3-methylglutaric acid moiety. The type III compound has the 3-methylglutaric acid moiety but not the sulfate moiety. The type IV compound has neither moiety [59] (Figs. 3 and 4, Table 1). Although all liposidomycins have the same inhibition activity against peptidoglycan biosynthesis, only types III and IV liposidomycins showed antimicrobial activity against Mycobacterium phlei (Table 1). Because type III and IV compounds lacking the sulfate moiety are more lipophilic than types I and II, they probably penetrate the lipophilic mycobacterial outer membrane better and inhibit MraY, which exists in the inner membrane [60] (Fig. 1). When the effects of liposidomycin A of types I–IV and tunicamycin on the growth of mammalian BALB/3T3 cells were compared, all liposidomycins showed no cytotoxicity even at 25 μg/ml, whereas tunicamycin inhibited 50% of the growth at 0.05 µg/ml. Two types of liposidomycins, A-III and A-IV, without the sulfate moiety at 2″ are potently active against Mycobacterium spp. and show increased antimicrobial activity against Escherichia coli BE1186 and Bacillus subtilis IFO 3513 [60]. HIV-infected and/or cancer patients have been threatened by the opportunistic infections caused by the mycobacterium avium complex (MAC). Liposidomycin C-III, one of the abundant components, has antimicrobial activity against several types of MACs (MIC = 1.2–12.5 μg/mL) [61].

Fig. 4
figure 4

 Comparison of the biosynthetic pathway (a) and genes (b) of liposidomycin and caprazamycin

Full size image
Table 1 Characteristics of each structure and activity in liposidomycin, caprazamycin, A90289A, and muraminomycin
Full size table

To identify novel antimicrobial agents, it is also important to quickly exclude the known analogs, such as liposidomycin and caprazamycin, by tandem mass spectrometry (MS) and the application of in-line reverse-phase HPLC-electrospray MS (LC-ESI-MS) [62, 63]. A similar strategy could be applied to identify other related antibiotics.

The inhibition activity of deacyl liposidomycin isolated from the fermentation broth (the same as caprazene, which is obtained by the acid hydrolysis of caprazamycins) against translocase I decreased about 200-fold compared with that of the original liposidomycin (Fig. 5). At that time, we did not consider the potential to synthesize simple analogs of liposidomycin such as CPZEN-45 using deacyl liposidomycin without a hydroxyl group.

Fig. 5
figure 5

 Structures of each degradation product of liposidomycin and caprazamycin and its analogue CPZEN-45

Full size image

Because the isolation of liposidomycin is laborious and the recovered quantity is low, precursor-directed biosynthesis is a useful method for the production of a single liposidomycin. Exogenously supplied myristic acid or palmitic acid results in the almost exclusive production of liposidomycin C-III and/or M-III [64]. The four isolated types of liposidomycins are consistent with the subsequent analysis of the biosynthetic genes of uridyl liponucleoside antibiotics [65] (Fig. 4).

Caprazamycin, A-90289A, and muraminomycin as liposidomycin analogs with specific MraY inhibition activity

The liposidomycin-type liponucleoside antibiotic caprazamycin was isolated from Streptomyces sp. MK730-62F2 by Igarashi et al. at the Institute of Microbial Chemistry (BIKAKEN) [12] (Fig. 3d). Takahashi et al. then developed CPZEN-45 as a derivative of caprazamycin as an antibiotic against M. tuberculosis from a core structure of caprazamycin, caprazene (deacyl liposidomycin) [13, 66] (Fig. 5b, c). Igarashi, who won the Sumiki–Umezawa Memorial Award in 2018 from the Japan Antibiotics Research Association with me [67], has summarized the study of caprazamycin, and Takahashi has described the first structure–activity relationship (SAR) of CRZEN-45 in this issue [68, 69]. In addition, two reviews involving caprazamycin and the SARs of the derivatives were published recently in quick succession [65, 70]. They revealed the essential pharmacophores present in the natural caprazamycin scaffold. Caprazamycins differ from liposidomycins in the absence of a sulfate group at the 2″-position of the aminoribose and in the presence of a permethylated l-rhamnose α-glycosidically linked to the 3-methylglutaryl moiety (Figs. 3d, 4, and Table 1).

A-90289A consists of a β-hydroxy palmitic acid moiety and has a structure identical to that of caprazamycin A but also contains a sulfate group, which is characteristic of liposidomycins. However, the sulfate group in A-90289A is attached at the 2’-hydroxy group, as opposed to the 2″-hydroxy group where it is attached in liposidomycins [26] (Fig. 3e and Table 1). The differences in the structures among liposidomycin, caprazamycin, and A-90289A were confirmed theoretically after each gene cluster of biosynthesis was clarified (Fig. 4). The sites of sulfation that differ between liposidomycin and A-90289A may depend on the substrate specificity of the unusual arylsulfate sulfotransferase (ASST) [71].

Muraminomycins have slightly different structures and antimicrobial activities compared to other liponucleoside antibiotics. They possess a 2′-dexyuridine moiety, a 5-amino-2,5-dideoxyribose moiety, a diazepanone ring system, a fatty acid, 3-methylglutaric acid, and succinic-acid-attached 6-methyl-α-rhamnose. The most characteristic structural difference between muraminomycins and liposidomycins, caprazamycins, or A90289A was the loss of hydroxyl groups at C-2′ and C-2″ (Fig. 3f, Table 1) [27, 72].

The primary target of CPZEN-45 in Bacillus subtilis is undecaprenyl phosphate-GlcNAc-1-phosphate transferase (TagO: an ortholog of WecA and involved in the biosynthesis of teichoic acid), which is a different target from that of the parent caprazamycin (MraY). CPZEN-45 also inhibited WecA, which is involved in mycolyl-arabinogalactane biosynthesis in M. tuberculosis and is the ortholog of TagO (Fig. 2) [42, 43, 73], and this is an important result in terms of facilitating the development of new antituberculosis antibiotics.

Organic synthesis and biosynthesis of liponucleoside antibiotics

Two reviews of the organic synthesis and biosynthesis involving nucleoside antibiotics and MraY inhibitors were published in 2003 and 2010 [10, 11]. Since then, organic synthesis involving the total synthesis and biosynthesis of uridyl liponucleoside antibiotics has continued progressing. Tunicamycin V is readily accessible via the longest linear sequence of 24 synthetic steps from uridine and is commercially available from simple materials, with an overall chemical yield of 3.9%. Replacing the GlcNAc and lipid moieties could provide a range of new analogs of tunicamycins [74]. Recently, truncated analogs of tunicamycin V in a series were synthesized, and their MraY inhibition activities were investigated by Yamamoto and Ichikawa. Unfortunately, the MraY inhibition activity of GlcNAc-, lipid-, and uridine-truncated analogs all decreased by more than 100-fold [75]. The complex structure with their binding to MraY from Clostridium bolteae and with tunicamycin-GPT (GPAGT1) will provide new information for seeking potent and/or specific inhibitors against MraY and/or GPT [48,49,50, 76]. Their co-crystal structures indicate that the GlcNAc moiety recognized in tunicamycin is quite different, which shows that the GlcNAc-modified analog is crucial for the specificity between MraY and GPT and/or for potency. Chemically modified tunicamycin with less toxic to eukaryotic cells, but which retain their antibacterial activity were synthesized [52].

The total synthesis of caprazamycin A has been reported and was accomplished in 23 steps. It will guide the synthesis of related uridyl liponucleoside antibiotics such as liposidomycin and A-90289 [77, 78]. Although an analog of caprazamycin, CPZEN-45, has been anticipated as a new antituberculosis drug, the co-crystal structures of caprazamycin analogs to MraY and WecA will also provide information about other new potential drugs (Fig. 6).

Fig. 6
figure 6

Perspective on new antibiotics

Full size image

The caprazamycin gene cluster was reported in 2009 as the first cluster of a translocase I inhibitor [79]. Since then, the gene clusters of liposidomycin, A-90289, and muraminomycin were also described (Fig. 4) [27, 80, 81]. Gene disruption and heterologous expression experiments of the gene clusters allow the generation of novel biologically active derivatives via pathway engineering and will also yield new analog compounds. In addition, a new type of ASST that had low homology with previous ASSTs was identified as Cpz4 during a study on the biosynthetic cluster genes of caprazamycin and liposidomycin [71]. Surprisingly, similar genes are located adjacent to the caprazamycin gene cluster (Fig. 4). Each biosynthetic gene for liponucleoside antibiotics showed the production of various liposidomycin-type liponucleoside antibiotics (Fig. 4).

Polymerase chain reaction screening was performed using a library of ~2500 strains and degenerate primers of a pyridoxal-5-phosphate-dependent l-Thr:uridine-5′-aldehyde transaldolase by Funabashi et al. [82] and a new nucleoside MraY inhibitor, named sphaerimicin, was discovered in 2013.

Perspective

Since the discovery of the transpeptidase inhibitor penicillin, many types of antibiotics targeting peptidoglycan biosynthesis have been developed mainly from microorganisms. Synthetic and semisynthetic technologies of the natural products have also produced clinically effective antimicrobial agents. In addition, biosynthetic technology and identified cluster genes are expected to yield new structures and/or targets for clinical agents. The biological targets of antimicrobial agents include the cell wall, DNA, protein biosynthesis, and cell membranes. Of these, cell wall biosynthesis is a promising target for antibiotics because there is no cell wall in human cells, providing an opportunity for selective toxicity of the antibiotic [10, 11, 15]. In fact, almost all the targets in peptidoglycan biosynthesis serve as clinical antibiotics, except translocase I (MraY) (Fig. 1). Although no clinical antibiotic against translocase I has been available up to this point, the crystal structure of MraY in complex with muraymycin D2 [32, 83, 84] and tunicamycin can inform the design of new inhibitors targeting MraY and/or WecA [22, 46, 48,49,50, 76]. In addition, the total synthesis and gene cluster of tunicamycin, liposidomycin, caprazamycin, A-90289, and muraminomycin open the door to optimal MraY inhibitors for clinical use, inhibitors like CPZEN-45 (Fig. 6).

Simple, specific, and improved biochemical assay systems continue to be critical to isolating and evaluating new MraY and/or WecA inhibitors from natural sources and/or chemical libraries [85,86,87]. Novel structures of MraY inhibitors isolated from natural sources will provide new information for drug-like related compounds using organic synthesis and biosynthetic genes. Tunicamycin and caprazamycin (including liposidomycin and A-90289) have already become candidates for total synthesis and gene cluster biosynthesis, allowing easy examination of the SAR [88,89,90,91]. Amplifications of the entire biosynthetic gene cluster may improve the production of antibiotics. Precursor-directed biosynthesis and mutasynthesis have become useful tools to generate new antibiotic derivatives. The semisynthetic drug candidate CPZEN-45, which is a WecA (TagO) and MurX (MraY) inhibitor in M. tuberculosis, and specific MraY inhibitors against MDR bacteria may become clinically approved in the near future. The combination of organic synthesis and biosynthesis, aided by newly enabled structural studies of liposidomycin/caprazamycin–MraY and/or CPZEN-45–WecA membrane proteins-drug complexes, can be expected to yield new types of antibiotics in the future (Fig. 6) [92, 93].

The vision of Dr Isono, to develop new clinical antibiotics through the discovery of specific nucleoside antibiotics, with polyoxin for antifungals and liposidomycin for antibacterials, is now being realized.