Introduction

Macrolide antibiotics are active against Gram-positive bacteria, especially Streptococcus pneumoniae, and their safety as an oral agent has already been proved. Therefore, macrolide antibiotics are regarded as very important chemotherapeutic agents against bacterial respiratory infections as in the case of β-lactam antibiotics or new quinolones. Although clarithromycin and azithromycin, which are representatives of widely used macrolides, exhibit enhanced antibacterial activities and characteristic pharmacokinetics compared with those of erythromycin, they are not active enough against resistant bacteria of S. pneumoniae with erm gene. Recently these resistant bacteria have widely spread, especially in European and Asian countries, and caused severe social problems. Chemical modifications of erythromycin provided the clinical site with two novel ketolides, telithromycin and cethromycin, which are effective against resistant bacteria of S. pneumoniae with erm gene. No oral antibiotic, however, has been launched so far, which is effective against the resistant bacteria of S. pneumoniae and does not have any problems in safety or taste.

Lincomycin (LCM) isolated as a secondary metabolite from fermentation broth of Streptomyces lincolnensis was chemically transformed to a useful oral antibiotic, clindamycin (CLDM) (Figure 1), which inhibits bacterial protein synthesis in a similar manner to macrolides. Although telithromycin is still influenced by efflux pumps of resistant S. pneumoniae with mef gene, CLDM is not influenced by efflux pumps. We focused not only on its safety and effectiveness against efflux pumps but its possibility of switch therapy by CLDM analogs, and planned to generate a novel oral antibiotic, which is effective against resistant bacteria of S. pneumoniae with erm gene or mef gene by chemical modifications of LCM.

Figure 1
figure 1

Chemical structures of LCM, CLDM, their reported analogs and novel LCM derivatives.

Full size image

Background and Molecular Design

Regarding chemical modification of LCM, Argoudelis et al.1 investigated fundamental SAR analysis focused on the C-6 (N-1′ and C-4′) and C-7 positions in 1960s and Magerlein et al.2, 3 generated U-24,279 A, (Figure 1) which exhibited much stronger in vitro and in vivo activities compared with LCM. On the other hand, Lewis et al.4, 5 disclosed VIC-105555 (Figure 1) and proved that its in vitro activity against Enterococcus faecalis was much stronger than that of CLDM and its half-life in plasma was greatly improved compared with that of CLDM in a variety of animal species. As a pioneer work, Sztaricskai and Ōmura et al.6 reported introduction of a heterocycle via sulfur atom onto the LCM carbon framework (1 and 2 shown in Figure 1) with (7R)-configuration, by means of SN2 reaction analogous to the synthesis of CLDM. All these derivatives introduced as shown in the above exhibited substantially same antibacterial spectra as those of LCM or CLDM, and there was no description about their antibacterial activities against resistant bacteria of S. pneumoniae with erm gene. Because 3D structural interaction of CLDM-rRNA complex by X-ray single crystallographic analysis7 indicated that there were several interactions, mainly a hydrogen bonding, between the peptidyl transferase cavity (A2058Ec, G2520Ec and A2059Ec) and hydroxyl groups at the sugar moiety of CLDM, chemical modification at the sugar part was supposed to be rather difficult for improving activity. As a matter of fact, 2-deoxy-LCM8 shows only 1% activity of LCM, and this result can be explained by lack of the hydrogen bonding between A2058Ec and 2-OH of LCM. We consequently pursued modification at C-7 and then C-6 (N-1′ and C-4′), focusing on hydrophilic interaction in extra 3D spaces in the above analysis.

Results and Discussion

We paid attention to the fact that stereochemistry at the C-7 position affected antibacterial activity, and synthesized a variety of (7S)-7-sulfur-aromatic derivatives of LCM firstly, in application of 2,3,4-tris-(trimethylsilyl)-LCM9 as a key intermediate prepared via 2,3,4,7-tetrakis-(trimethylsilyl)-LCM. Precise molecular design, practical synthetic procedures and full package of physicochemical properties of each derivative will be reported elsewhere. Although every derivative exhibited almost the same in vitro activities, introducing a substituent to an aromatic ring (1,3,4-thiadiazole or a benzene ring) enhanced antibacterial activity (3 to 4 and 7 to 8) (Table 1, all antibacterial evaluations were performed as hydrochloride). Representative 1H NMR data and MASS data of each analog are shown in Table 2. Then, we selected compounds 3 and 7 as lead compounds and started systematic synthesis of derivatives of 3 and 7 through introducing a substituent onto 1,3,4-thiadiazole or a benzene ring.

Table 1 Antibacterial activities of novel LCM derivatives against S. pneumoniae, S. pyogenes and H. influenzae. Minimum inhibitory concentration (μgml−1)a
Full size table
Table 2 Representative 1H NMR data and MASS data of each analog.a
Full size table

In the thiadiazole analogs synthesis, 5-phenyl-1,3,4-thiadiazol-2-yl analog (5) exhibited moderate activities against resistant bacteria of S. pneumoniae and Streptococcus pyogenes with erm gene. Then, optimization of 7-sulfur-substituent by the introduction of an additional substituent onto the phenyl group of 5 gave key compound 6. In the benzene analogs synthesis, locally optimized compound 9 showed stronger activity against the target compared with 8, but there seemed to be limitation for further enhancement of in vitro activity. Then, basic functionality or one more aromatic ring (by SAR of compound 5) was introduced to the para-position of the benzene ring of compound 7, and partially optimized compounds 10 and 11 were prepared. Because compounds 6, 10 and 11 exhibited strong antibacterial activities against clinically important pathogens for respiratory infections (Table 1), we performed final optimization of the whole molecule, especially by handling the proline part.

Ring expansion strategies (from a proline framework to a pipecolic acid framework, see VIC-105,555) have already been reported by Birkenmeyer et al.,10, 11 and we expanded a five-membered ring to six- or seven-membered ring and simultaneously optimized structure of a side chain at the C-4′ position. As a conclusion,12, 13, 14 compounds 12 and 13 to 15 were generated as candidates for further biological evaluations. Introduction of a cyclopropylmethyl group to the pipecolic acid moiety in the LCM framework has already been reported by Lewis et al.15, 16 As shown in Table 1, these derivatives possessed very strong antibacterial activities against not only resistant bacteria of S. pneumoniae with erm or mef gene, but other clinically important pathogens for respiratory infections such as S. pyogenes. Among them, compounds 12, 14 and 15 unexpectedly exhibited extremely potent antibacterial activities against resistant S. pneumoniae with erm gene even when compared with telithromycin and these derivatives were recognized as the first examples which were effective enough in vitro against the target resistant S. pneumoniae apart from ketolides.

Experimental Procedure

In vitro antibacterial activity

Minimum inhibitory concentration was determined by the agar dilution method. Test strains were subjected to seed culture using sensitivity test broth (STB, Nissui Pharmaceutical, Tokyo, Japan) cultured on blood agar plate for S. pneumoniae, S. pyogenes and H. influenzae. A 5 μl portion of cell suspension of the test strains having about 106 CFU per ml was inoculated into sensitivity disk agar (SDA, Nissui Pharmaceutical) supplemented with 5% horse blood and incubated at 37 °C for 20 h. Then, minimum inhibitory concentration was measured.