Thiopeptides are a class of peptide antibiotics produced by some bacteria, especially Streptomyces. They have potent activity against various bacterial pathogens, and exhibit potential as lead compounds for antimicrobial drug against methicillin-resistant Staphylococcus aureus and penicillin-resistant Streptococcus pneumoniae, etc [1, 2]. Natural thiopeptides such as thiostrepton and nosiheptide are veterinary antibiotic and animal feed additive, and semi-synthetic thiopeptide LFF571 is under clinical trial against Clostridium difficile infection in humans [3].

Thiopeptides are structurally comprised of a macrocyclic core moiety and a tail moiety. They are divided into four families according to the size of macrocyclic core: 26-membered, such as thiocillin, thiostrepton, and nosiheptide; 29-membered, such as GE37468, amythiamicin, and GE2270A; 32-membered, such as lactazoles; and 35-membered, such as berninamycin, TP-1161, and geninthiocin [4,5,6]. Most thiopeptides have in their tails modified amino acid residues such as thiazole, oxazole, oxazoline, and dehydroalanine (Dha), and a terminal amine group.

Thiopeptides are a growing class of antibiotics with various biological activities [2, 7,8,9,10,11,12]. Among them, geninthiocin (hereafter designated as geninthiocin A, Fig. 1) is a 35-membered thiopeptide discovered from Streptomyces sp. DD84 with tipA promoter inducing activity and potent antibacterial activity, and val-geninthiocin (hereafter designated as geninthiocin B, Fig. 1) is a close analog of geninthiocin A identified from Streptomyces sp. RSF18 with antibacterial activity comparable to that of geninthiocin A [13].

Fig. 1
figure 1

Chemical structures of geninthiocins A-D. Pyr (pyridine); Thz (thiazole); Thr (threonine); Oxa (oxazole); Dha (dehydroalanine); Hyval (hydroxyvaline); Val (valine); Ala (alanine)

In the course of exploring secondary metabolites from a soil isolate Streptomyces sp. CPCC 200267 with strong anti-G+ bacteria activity, we identified two new congeners of geninthiocin A. Herein, structures of the two congeners and their antibacterial activities were described, which revealed a preliminary structure-activity relationship (SAR) concerning the tail moiety and antibacterial activity of 35-membered thiopeptides.

The agar culture of Streptomyces sp. CPCC 200267 was extracted with ethyl acetate (EtOAc). The EtOAc extract revealed three peaks with similar UV absorption profiles by reversed-phase HPLC (27.4, 27.0, and 25.9 min; Fig. S1), which corresponded to three structure-related compounds, 1 (major component), 2 and 3 (minor components), respectively. They were purified by a procedure of EtOAc extraction, silica gel and ODS column chromatography and semi-preparative HPLC (Supplementary material), and then identified by NMR as geninthiocin A (1) and its two new congeners (2-3).

Compound 1 was obtained as white amorphous powder. Its molecular formula was determined as C50H49N15O15S by HR-ESIMS (Fig. S4). The 1D and 2D NMR data of 1 (Fig. S7-S12) were highly consistent with those of geninthiocin A reported in the literature [6], which indicated that 1 was geninthiocin A. The NMR data of 1 were assigned completely as indicated in Table S1.

Compound 2 was obtained as white amorphous powder. Its molecular formula was determined by HR-ESIMS as C50H51N15O15S (Fig. S15), two hydrogen atoms more than 1. Compound 2 showed very similar NMR data to 1 except the appearance of new signals for a methyl group (δH 1.32 (d, J= 7.2 Hz), δC 18.0) and a sp3-hybridized methine (δH 4.34 (q, J= 7.2 Hz), δC 49.3), and the loss of signals for a terminal methylidene (δH 6.04 (s) and 5.70 (s), δC 106.4; δC 135.4) in 1. The COSY correlations of Ala-NH (δH 8.62)/Ala-Hα (δH 4.34)/Ala-H3β (δH 1.32), and HMBC correlations from the terminal -NH2 (δH 7.39, 7.02) to Ala-C = O (δC 174.2) and Ala-Cα (δC 49.3) further revealed that an Ala residue had replaced the Dha4 residue in 1 (Fig. 2). All the other structural parts of 2 were the same as 1 deduced from their highly similar 1D and 2D NMR data (Fig. S18-S24). Therefore, 2 was determined as an analog of 1 with an Ala residue taking the place of Dha4 residue in 1. The MS2 spectra comparison between 1 and 2 further supported the amino acid replacement (Fig. S3, S14). The stereochemistry of 2 was established by C3 Marfey’s method and ROESY spectrum. The configuration of Ala, Thr, and Hyval in 2 were determined to be l- by C3 Marfey’s method (Fig. S36) [14]. The vinyl methyl group in Oxa1 was determined as Z-form based on ROESY correlation between the methyl proton (δH 1.76) and the amide proton (δH 9.63) in Oxa1 (Fig. 2). Compound 2 was designated by us as geninthiocin C. Its NMR data were assigned completely as indicated in Table 1.

Fig. 2
figure 2

Key COSY, HMBC, and ROESY correlations of 2 and 3

Table 1 NMR data of geninthiocins C (2) and D (3)

Compound 3 was obtained as white amorphous powder. Its molecular formula was determined by HR-ESIMS as C44H43N13O13S (Fig. S26), which was C6H6N2O2 (equivalent to Dha-Dha residues) less than geninthiocin A. Compound 3 also showed very similar NMR data to 1 except the loss of signals for two Dha residues. It had an identical 35-membered macrocylic core moiety to 1 because of its highly similar 1D and 2D NMR data to 1 (Fig. S29-S35). The HMBC correlations from the terminal -NH2 (δH 7.90, 8.41) to the carbonyl carbon (δC 165.3) and C-6 (δC 150.9) of pyridine-2-carboxylate proved that 3 lost Dha-Dha residues in its tail (Fig. 2). Therefore, 3 was determined as another analog of 1 without Dha-Dha residues in its tail. The ROESY correlation between the methyl proton (δH 1.75) and the amide proton (δH 9.59) in Oxa1 of 3 was observed, indicating that the vinyl methyl group in Oxa1 was in Z-form (Fig. 2). Both Thr and Hyval residues in 3 must take the l-configuration (the same as 2) because 3 shared the same prepeptide with 2 in biosynthesis (Fig. 3). Compound 3 was designated by us as geninthiocin D (Fig. 1). Its NMR data were assigned completely as indicated in Table 1.

Fig. 3
figure 3

The proposed biogenesis of geninthiocins A, C, and D (1-3)

We proposed a possible production relationship of geninthiocins A, C, and D (13) based on the biosynthetic mechanism of geninthiocin A and the structural differences of them (Fig. 3). Besides, we eliminated the very small chance that geninthiocin C (2) may come from a specific mutant carrying a base change in the structural gene encoding the core peptide for geninthiocin biosynthesis (leading to the translation of Ala instead of Ser in geninthiocin biosynthesis), by choosing randomly well separated single colonies of Streptomyces sp. CPCC 200267 for culture and then analysis of geninthiocin(s) they produced. All the ten single colonies produced the same profile of geninthiocins as the original Streptomyces sp. CPCC 200267 did (Fig. S37), i.e. geninthiocins A (1) as major component and C-D (2-3) as minor components. Thus, we demonstrated that geninthiocin C (2) was not the metabolite from a specific mutant of Streptomyces sp. CPCC 200267. It was a congener in geninthiocin A (1) production, and should be the product of an unidentified enzyme-catalyzed hydrogenation of geninthiocin A (1) at Dha4 residue, which generated the only l-Ala residue in geninthiocin C (2, Fig. 3). We believe that geninthiocin D (3) is a shunt product in geninthiocin A (1) biosynthesis. It may come from the incorrect removing of two more Ser residues from C-terminal of the core peptide in geninthiocin A (1) biosynthsis (Fig. 3).

Geninthiocins A, C, and D (13) differed only at tail moiety of their structures, which provided us the opportunity to explore tail variations on their antibacterial activity. In vitro MIC assay revealed that geninthiocins C and D (23) lost activities against G+ bacteria (Table S2) [15, 16]. Thus, the -Dha-Dha-NH2 tail of geninthiocin A (1) was essential for its potent anti-G+ bacteria activity.

The structure-activity relationship (SAR) of thiopeptides is very complex. While the macrocyclic core moiety is essential for the anti-G+ bacteria activity, the tail moiety is also indispensible for some thiopeptides (Table S3 for a brief summary concerning the SAR of Dha-containing tail moiety and anti-G+ bacteria activity of some thiopeptides). For example, tail modification of 26-membered thiopeptides such as siomycin, nocathiacin, and thiostrepton resulted in various effects on anti-G+ bacteria activity (nearly no changes, decreases or increases) [17,18,19,20], and tail modification of 29-membered thiopeptides such as thiomuracin A and baringolin exerted little influences on anti-G+ bacteria activity [8, 9]. Our study of geninthiocins provided the first example for 35-membered thiopeptides concerning the SAR of tail moiety and anti-G+ bacteria activity. It may also provide some references for predicting the SAR of berninamycins [4], another 35-membered thiopeptide with very similar structure to geninthiocins.

Geninthiocin A (1): white amorphous powder; UV (MeOH) λmax (log ε) 240 (4.78) and 324 (3.79) nm; IR (KBr) vmax 3344.5, 2916.1, 1763.2, 1665.4, 1503.3, 1201.7, 1106.3, 995.6, and 890.1 cm−1; 1H NMR (DMSO-d6, 600 MHz) and 13C NMR (DMSO-d6, 150 MHz) data, see Table S1. ESIMS m/z 1132.46 [M + H]+, 1154.57 [M + Na]+; ( + ) HRESIMS m/z 1132.3354 [M + H]+ (calcd for C50H50N15O15S, 1132.3326, ppm = 2.5).

Geninthiocin C (2): white amorphous powder; UV (MeOH) λmax (log ε) 238 (4.83) and 324 (3.90) nm; IR (KBr) vmax 3342.8, 2921.0, 1765.9, 1660.8, 1503.6, 1202.4, 1105.7, 1025.6, and 889.1 cm1; 1H NMR (DMSO-d6, 600 MHz) and 13C NMR (DMSO-d6, 150 MHz) data, see Table 1. ESIMS m/z 1134.51 [M + H]+, 1156.58 [M + Na]+; ( + ) HRESIMS m/z 1134.3490 [M + H]+ (calcd for C50H52N15O15S, 1134.3483, ppm = 0.7).

Geninthiocin D (3): white amorphous powder; UV (MeOH) λmax (log ε) 239 (4.89) nm; IR (KBr) vmax 3338.7, 2975.8, 1765.2, 1668.9, 1503.9, 1387.8, 1202.3, 1100.9, and 887.8 cm−1; 1H NMR (DMSO-d6, 600 MHz) and 13C NMR (DMSO-d6, 150 MHz) data, see Table 1. ESIMS m/z 994.31 [M + H]+, 1016.42 [M + Na]+; (+) HRESIMS m/z 994.2906 [M + H]+ (calcd for C44H44N13O13S, 994.2897, ppm = 1.0).