Neuraminidase Inhibitors from marine-derived actinomycete Streptomyces seoulensis



Ren Xiang Tan, Institute of Functional Biomolecules, State Key Laboratory of Pharmaceutical Biotechnology, Nanjing University, Nanjing 210093, China. E-mail:



This work was performed to characterize new secondary metabolites with neuraminidase (NA) inhibitory activity from marine actinomycete strains.

Methods and Results

An actinomycete strain IFB-A01, capable of producing new NA inhibitors, was isolated from the gut of shrimp Penasus orientalis and identified as Streptomyces seoulensis according to its 16S rRNA sequence (over 99% homology with that of the standard strain). Repeated chromatography of the methanol extract of the solid-substrate culture of S. seoulensis IFB-A01 led to the isolation of streptoseolactone (1), and limazepines G (2) and H (3). The structures of 13 were determined by a combination of IR, ESI-MS, 1D (1H and 13C NMR, and DEPT) and 2D NMR experiments (HMQC, HMBC, 1H-1H COSY and NOESY). Compounds 13 showed significant inhibition on NA in a dose-dependent manner with IC50 values of 3·92, 7·50 and 7·37 μmol l−1, respectively.


This is the first report of two new (1 and 2) and known (3, recovered as a natural product for the first time in the work) NA inhibitors from the marine-derived actinomycete S. seoulensis IFB-A01.

Significance and Impact of the Study

The three natural NA inhibitors maybe of value for the development of drug(s) necessitated for the treatment of viral infections.


Neuraminidase (NA) catalyses the release of progeny influenza virus from infected host cells (Moscona 2005) by cleaving the α-ketosidic bond that links a terminal neuraminic acid scaffold to the adjacent oligosaccharide residue (von Itzstein 2007). Targeting it was therefore recommended as, and has been witnessed to be, a workable strategy to eradicate or control such infections (Hernandez et al. 2011) as exemplified by laninamivir launched in Japan in 2010 as an anti-influenza agent via the NA inhibition (Kashiwagi et al. 2012). However, the limited chemical diversity of the existing NA inhibitors are holding a bottleneck for the discovery of more efficient drugs required for a consolidated controllability of the viral infections in view of the inevitable drug resistance of the currently prescribed drugs. The observations pose a powerful impetus for characterizing more diversified compounds with the desired NA inhibitory activities. Marine microorganisms are important and productive source of new bioactive natural products. During the last few decades, more attention has been paid to antibiotic natural products of marine origin (Maskey et al. 2002). Streptomyces sp. is a potent resource for new lead compounds valuable for the drug development (Kinashi 2011). However, the secondary metabolites of Streptomyces seoulensis have never been investigated since recovered from nature in 1997 (Chun et al. 1997). In our continuing efforts to discover novel natural bioactive products from microbes associated with plants (Zhang et al. 2010), insects (Zhang et al. 2011) and marine creatures (Liu et al. 2006), we explored the metabolite of S. seoulensis IFB-A01 dwelling in the gut of shrimp Penasus orientalis, with an intention to identify new NA inhibitors. We hereby wish to report the isolation, structural elucidation and enzyme inhibition of two new metabolites streptoseolactone (1) and limazepine G (2), along with limazepine H (3), which was obtained in this work as a natural product for the first time (Fig. 1).

Figure 1.

Structures of compounds 1–3.

Materials and methods

General experimental procedures

The optical rotation was measured on a WXG-4 disc polarimeter. The UV and IR spectra were recorded on a Hitachi U-3000 spectrophotometer and a Nexus 870 FTIR spectrometer, respectively. The HR-ESI-MS spectra were recorded on an Agilent 6210 LC-TOFMS instrument (Agilent Technologies, Santa Clara, CA, USA). The 1D (1H and 13C NMR, and DEPT) and 2D NMR experiments (HMQC, HMBC and 1H-1H COSY) were conducted on a Bruker DRX-500 and/or an AV-300 NMR spectrometers. Silica gel (200–300 mesh) for column chromatography and GF254 for TLC were produced by the Qingdao Marine Chemical Factory (Qingdao, China), and Sephadex LH-20 by GE (Fairfield, CT, USA). All chemicals used in the study were of analytical grade.

Strain material

The title strain was isolated from the gut of Penasus orientalis collected in 2009 from the sea nearby Qingdao Port, China, and identified as Streptomyces seoulensis by comparing its morphological characters and 16S rDNA sequence with those of standard records. The strain has been deposited under the number CGMCC3689 in China General Microbiological Culture Collection Center.

Fermentation, extraction and isolation

After growing on malt dextrose agar (MDA) medium at 28°C for 5 days, S. seoulensis IFB-A01 was inoculated into Erlenmeyer flasks (1000 ml) preloaded with 300 ml of PD liquid medium. After incubation for 4 days at 28 ± 1°C on a rotary shaker at 150 rev min−1, the culture liquid (20 ml, each) was transferred as the seed into 250-ml flasks containing an evenly mingled medium composed of grain (7·5 g), bran (7·5 g), yeast extract (0·5 g), sodium tartrate (0·1 g), FeSO4·7H2O (0·01 g), sodium glutamate (0·1 g), pure corn oil (1 ml) and H2O (30 ml). The subsequent cultivation was accomplished by standing for 40 days at 28 ± 1°C with the relative humidity ranging between 60 and 70%.

Extraction and fractionation

The dehydrated solid-substrate culture was exhaustively extracted with MeOH (3 × 10 l) for 12 h at room temperature. Evaporation of the solvent in vacuo gave a residue (15·2 g) that was soaked in MeOH (500 ml) around 60°C. The afforded solution was cooled down gradually to −20°C followed by standing for another 24 h to remove lipids and salts that precipitated. The removal of methanol under reduced pressure yielded a residue (8·0 g) that was further chromatographed over a silica-gel column eluting successively with CHCl3/MeOH (100 : 0, 100 : 1, 100 : 2, 100 : 4, 100 : 8, 100 : 16, and 0 : 100, v/v) to give seven fractions (Frs. 1–7). Fr. 3 (2·4 g) was subjected to gel filtration over Sephadex LH-20 with CHCl3/MeOH (1 : 1) to yield three subfractions (Frs. 3·1–3·3). Fr. 3·1 (0·4 g) was purified further by HPLC [Hypersil BDS C18 5 μm; MeOH-H2O (59 : 41); 2·0 ml min−1] to afford 1 (8·2 mg, tR: 53·9 min). Fr. 4 (1·7 g) and Fr. 5 (1·8 g) were subjected to gel filtration over Sephadex LH-20 with CHCl3/MeOH (1 : 1) to yield four subfractions (Fr.4·1–4·4) and two parts (Fr. 5·1–5·2), respectively. Purification of Fr. 4·1 (0·3 g) by HPLC using MeOH-H2O (51 : 49) afforded 2 (6·8 mg, tR: 24·9 min), and the work-up of Fr. 5·1 (0·5 g) in the same manner yielded 3 (15·2 mg, tR: 16·3 min). The purity of compounds 1–3 was proven to be over 98% ascertained by HPLC analyses.

Neuraminidase inhibition assay

Neuraminidase used for the assay was the enzyme from the bacterium Clostridium perfringens (Sigma N-2876). The NA activity was measured using the substrate 4-methylumbelliferyl-R-D-N-acetylneuraminic acid sodium salt hydrate (4-MU-NANA) (Sigma, M8639) in the enzyme buffer [MES buffer: 32·5 mmol l−1 2-(Nmorpholino) ethanesulfonic acid, 4 mmol l−1 CaCl2, pH 6·5]. All compounds were dissolved in DMSO and diluted to the corresponding concentrations in MES buffer. An enzyme inhibitory assay was carried out in 96-well plates containing 10 μl of NA (0·035 U ml−1) and 10 μl of test compound incorporated in the enzyme buffer at different concentrations to assess their inhibitory activity. The mixture was incubated for 30 min at 37°C, and 30 μl of 4-MU-NANA substrate per well in enzyme buffer was added. The enzymatic reactions were carried out for 2 h at 37°C and quenched by adding 150 μl of the stop solution (25% EtOH, 0·1 mol l−1 glycine, pH 10·7). The fluorescence intensity of the product (4-MU) was acquired with excitation and emission wavelengths of 360 and 440 nm, respectively. The IC50 for reducing the NA activity was then determined. To determine the mode of inhibition, the enzyme assay was performed as previously reported at different substrate concentrations (200, 100, 50, 25 and 12·5 mmol l−1) with compounds 1–3 assayed at 5, 10 and 20 mmol l−1.

Identifications of metabolites 1−3

Streptoseolactone (1)

A white powder (MeOH); math formula −69·8 (= 0·076, MeOH); UV (MeOH) λmax nm (log ε) = 241 (2·78); IR (KBr) νmax 3397, 2930, 2870, 1738, 1702, 1449, 1379, 1255, 1217, 1035, 754 cm−1; 1H and 13C NMR data given in Table 1; HR-ESI-MS m/z: 493·2937 [M+Na]+ (calcd for C29H42O5Na: 493·2930).

Table 1. 1H (500 MHz) and 13C (125 MHz) NMR data of 1 (in CDCl3)
No. δ H δ C No. δ H δ C
  1. a

    Overlapped signals without designating multiplicity.

11·48a33·9164·90 (dd, 11·0, 4·0)81·8
21·72, 1·94a31·417 168·0
33·14 (td, 10·5, 5·5)77·0181·25 (s)24·2
42·01 (m)35·319 124·6
51·44a46·120 176·1
64·03 (brs)72·8212·24,2·34a24·1
7 214·7222·21a27·3
8 52·6235·10 (brs)123·2
92·40a42·524 132·8
10 29·5251·68 (s)25·7
111·50, 1·83a22·8261·60 (s)17·8
121·75, 2·09a22·1271·07 (d, 6·5)15·1
132·87(dd, 12·5, 3·0)43·1281·35 (s)17·2
14 52·5290·87 (s)20·7
151·76, 2·46a36·6   

Limazepine G (2)

Yellowish needles (MeOH); math formula +493·2 (= 0·022, MeOH); UV (MeOH) λmax nm (log ε) = 316 (4·16), 231 (4·30); IR (KBr) νmax 3332, 1688, 1662, 1573, 1408, 1387, 1337, 1253, 1200, 958, 763 cm−1; 1H and 13C NMR data listed in Table 2; HR-ESI-MS m/z: 307·1059 [M+Na]+ (calcd for C16H16N2O3Na: 307·1051).

Table 2. 1H (500 MHz) and 13C (125 MHz) NMR data of 2 and 3 (in DMSO-d6)
δ H δ C δ H δ C
13·40 (dd, 15·5, 3·0)29·33·42 (dd, 15·5, 3·5)28·9
2·80 (dd, 15·5, 11·0) 2·85 (dd, 15·5, 11·5)
2 124·3 123·3
36·90 (s)123·27·40 (s)130·6
5 161·2 161·9
5a 125·5 125·2
67·21 (d, 8·5)120·67·23 (d, 8·0)120·9
77·04 (d, 8·5)126·67·07 (d, 8·0)126·9
8 129·1 129·7
9 145·9 146·2
9-OH9·18 (brs) 9·26 (brs) 
9a 125·1 124·9
109·56 (brs) 9·65 (brs) 
11 168·3 168·1
11a4·61 (dd, 11·0, 3·0)56·14·73 (dd, 11·5, 3·5)56·8
126·32 (br d, 16·0)124·87·32 (d, 15·5)131·1
135·65 (dq, 16·0, 6·0)126·45·94 (d, 15·5)120·0
141·80 (brd, 6·0)18·2 166·8
152·24 (s)16·72·26 (s)16·9
14-NH2  7·0 (s),7·43 (s) 

Limazepine H (3)

Yellowish amorphous powder (MeOH); math formula +797·6 (= 0·076, MeOH); UV (MeOH) λmax nm (log ε) = 336 (4·92), 226 (4·80); IR (KBr) νmax 3365, 3204, 1697, 1682, 1638, 1565, 1424, 1262, 1249, 1232, 1152, 960 cm−1; 1H and 13C NMR data tabulated in Table 2; HR-ESI-MS m/z: 336·1074 [M+Na]+ (calcd for C16H15N3O4Na: 336·1073).


Through a combination of column chromatography, gel filtration and semi-preparative HPLC, the extract of the bacterial culture gave two new metabolites (1 and 2), and compound 3 described previously as a synthetic intermediate (Pena and Stille 1989).

Compound 1 was obtained as white powder. Its molecular formula was evidenced to be C29H42O5 from the quasimolecular ion at m/z 493·2937 [M+Na]+ (calcd for C29H42O5Na: 493·2930) in its high-resolution electrospray ionization mass spectrum (HR-ESI-MS). The 1H NMR spectrum of 1 (Table 1) displayed the signals indicating the presence of an olefinic proton (δ 5·10), two oxygenated methines (δ 3·14 and 4·03) and six methyl (δ 0·87, 1·07, 1·25, 1·35, 1·60 and 1·68). The 13C NMR spectrum of 1 exhibited a total of 29 resonance lines arising respectively from two carbonyl (δC 176·2 and 214·7), four olefinic, three quaternary, seven methine (3 oxygenated ones at δC 72·8, 77·0, and 81·8), seven methylene and six methyl carbons. These characteristic NMR data suggested that compound 1 had a steroidal skeleton as possessed 16-deacetylfusidic acid γ-lactone (Rastrup-Andersen and Duvold 2002). This assumption was substantiated by a set of 2D NMR experiments (1H-1H COSY, NOESY, HMQC and HMBC), which allowed the unambiguous assignment of all NMR signals (Table 1). Thus, the presence of 7-ketone group was indicated by the HMBC correlation of C-7 (δC 214·7) with H-5 (δH 1·44) and H-28 (δH 1·35) (Fig. 2). In compliance with the magnitude of chemical shifts of C-3 (δ 77·0) and C-6 (δ 72·8), the 3,6-dihydroxy functionality was evidenced from the twofold doublet of H-3 at δ 3·14 (J3,4β J3,2β = 10·5 Hz, J3,2α = 5·5 Hz) and the broadened H-6 singlet at δ 4·03. The relative configuration of 1 was disclosed by the NOESY correlations between the proton pairs of H-16/H-13, H-13/CH3-28, H-9/CH3-18, H-9/CH3-29, H-5/H-6, H-6/CH3-27, H-3/H-5 and H-4/CH3-18, demonstrating that 3-OH, H-4, 6-OH, H-9, CH3-18 and CH3-29 were β-oriented whereas H-5, H-13, H-16, CH3-27 and CH3-28 positioned in an α-orientation (Fig. 2). Subsequently, the absolute configuration of 1 was addressed by taking the advantage of the CD bands depending collectively on the chirality of the γ-lactone and cyclohexanone moieties. Compared with 16,21α-epoxy-21-isopropyldigitoxigenin-3-acetate (Fig. 3) (Uchida and Kuriyama 1974), the C-16 configuration of 1 was assigned as S on the basis of the positive Cotton effect at 216 nm (Fig. 4) associated with the π–π* transition in the region 205–235 nm. Relative to the C-16 absolute stereochemistry, the chirality of the other asymmetric carbons of 1 could be deduced from its NOESY spectrum. As anticipated, the 5S-, 6S-, 8S-, 9S- and 10R-configurations mentioned above were identical to those elucidated alternatively according to the octant rule from the negative Cotton effect at 326 nm ascribable to the n–π* transition of cyclohexanone (Fig. 4). Thus, the absolute stereochemistry of 1 was assigned to be 3S, 4S, 5S, 6S, 8S, 9S, 10R, 13R, 14S and 16S.

Figure 2.

Key HMBC, 1H-1H COSY and ROSEY correlations of 1.

Figure 3.

16, 21α-epoxy-21-isopropyldigitoxigenin-3-acetate.

Figure 4.

The CD curve of 1 processed according to the octant rule.

Compound 2, yellowish needles, possessed a molecular formula of C16H16N2O3 as indicated from its quasimolecular ion at m/z 307·1059 [M+Na]+ (calcd for C16H16N2O3Na: 307·1051) in its HR-ESI-MS, 1H and 13C NMR spectra. The 13C NMR spectrum of 2 displayed a total of sixteen carbon resonance lines, which could be grouped by DEPT experiments as two methyl, one methylene, an sp3 methine, five sp2 methine, two carbonyl and five quaternary carbons. In the 1H NMR spectrum of 2, a pair of mutually coupled doublets (= 8·7 Hz) at δ 7·21 and 7·04 required the presence of a 1,2,3,4-tetrasubstituted benzene ring. A 1,1,2-trisubstituted ethane moiety was demonstrated by another proton coupling sequence resonating at δ 4·61 (dd, J = 11·0, 3·0 Hz), 3·40 (dd, J = 16·0, 3·0 Hz) and 2·80 (dd, J = 16·0, 11·0 Hz), and a 1-substituted propene residue by a group of mutually coupled signals at δ 6·32 (br d, J = 16·0 Hz), 5·65 (dq, J = 16·0, 6·0 Hz) and 1·80 (br d, J = 6·0 Hz). These observations, along with the four singlets at δ 2·24 (3H), 6·90, 9·56 and 9·18 (the latter two were D2O exchangeable), suggested that compound 2 was most likely a pyrrolo[1,4]benzodiazepine derivative (Fotso et al. 2009). This proposal was confirmed by its 1H-1H COSY, HMQC and HMBC spectra, which led to the exact assignment of all 1H and 13C NMR signals (Table 2). The substitution pattern of the benzene ring was indicated by the HMBC correlation of H-6 with C-5 (amide carbonyl) C-9a resonating respectively at δ 161·2 and 125·1, and of H-7 with the methyl (C-15 at δ 16·7) and oxygenated (C-9 at δ 145·9) carbons. This observation, along with the architecture of the seven- and five-membered rings, was particularly reinforced by the correlation of the broadened NH (amide proton) singlet at δ 9·56 with C-5a, C-9 and C-11a, and of H-11a with C-2, C-3 and C-5 resonance lines (Fig. 5). The anchorage at C-3 of the 1-propenyl group was evidenced from the HMBC correlation of H-12 with C-1, C-3 and C-14. Finally, compound 2 was found to have a large positive specific rotation (math formula +493·2), which was close to that of other pyrrolo[1,4]benzodiazepine counterparts possessing an S-configuration (Fotso et al. 2009).

Figure 5.

Key HMBC and 1H-1H COSY correlations of 2 and 3.

The HR-ESI-MS of compound 3 exhibited its molecular formula as C16H15N3O4 with eleven degrees of unsaturation. The majority of the 1H and 13C NMR spectra of 3 were quite comparable to those of 2. However, the 12,13-double bond of 3, resonating at δC 131·1 and 120·0, and at δH 7·32 (15·5 Hz) and 5·94 (15·5 Hz), was shown to be more polarized presumably due to its conjugation with an electron-withdrawing group. Furthermore, the three-proton doublet (J = 6·0 Hz) in the 1H NMR spectrum of 2 was replaced by a pair of amide singlets at δH 7·43 and 7·00 in that of 3. Accordingly, an (E)-acrylamide residue had to be proposed for 3. This assumption was confirmed by its 1H-1H COSY, HMQC and HMBC spectra permitting an unambiguous assignment of all 1H and 13C NMR spectral data (Table 2). Again, the large positive specific rotation of 3 (math formula +797·6) suggested its S-configuration at C-11a (Fotso et al. 2009). Compound 3, an intermediate in the synthesis of anthramycin (Pena and Stille 1989), was characterized for the first time as a natural product in the present work.

The three metabolites were assessed for the NA inhibition utilizing oseltamivir as a positive control. Compounds 1–3 were demonstrated to be inhibitory on the enzyme activity in a dose-dependent manner with IC50 values of 3·92 ± 0·20, 7·50 ± 0·26 and 7·37 ± 0·12μmol l−1, respectively, while that of oseltamivir was 0·14 μmol l−1. The curves of compounds 2 and 3 appeared to be superimposable, highlighting that the change of the side chain affected slightly the activity (Fig. 6).

Figure 6.

The neuraminidase (NA) inhibitory effects of 1–3. The sample concentrations are displayed on a logarithmic scale. The IC50 values were identified from the midpoint (NA activity = 50%) of the semilog plot.

To understand the mode of inhibition, the double-reciprocal Lineweaver–Burk and Dixon plots were mapped (Fig. 7). The data indicated that all of the three compounds were noncompetitive inhibitors for the reason that different substrate concentrations led to a series of lines intersected at a nonzero point on the x-axis (−Ki) (Fig. 7a–c). The Ki values for the test compounds complied with the IC50 data (Fig. 7d–f).

Figure 7.

Graphical determination of the type of inhibition for 1–3. (a–c) Lineweaver–Burk plots for the inhibition of the three compounds on the neuraminidase for the hydrolysis of the substrate (image_n/jam12136-gra-0001.png [I] = 0 μmol l−1, image_n/jam12136-gra-0002.png 5 μmol l−1, image_n/jam12136-gra-0003.png 10 μmol l−1, image_n/jam12136-gra-0004.png 20 μmol l−1). (d–f) Dixon plots for the three compounds determining the inhibition constant Ki (image_n/jam12136-gra-0001.png 200 μmol l−1, image_n/jam12136-gra-0002.png 100 μmol l−1, image_n/jam12136-gra-0003.png 50 μmol l−1, image_n/jam12136-gra-0004.png25 μmol l−1, image_n/jam12136-gra-0005.png12·5 μmol l−1).


A growing body of data has clearly underpinned that NA can be adopted as a promising target for the treatment of viral infections (Dobrovolny et al. 2011), with the profitability figured out from some approved anti-influenza agents such as laninamivir (Kashiwagi et al. 2012). However, the structure diversity of the enzyme inhibitors remains very limited, hampering thereby the further development of superior anti-influenza drugs that may be in an urgent need upon the drug resistance of the existing influenza-treating drugs. Although short of the laboratory infrastructure allowing the experimentation with the virulent virus, the present work has characterized three NA inhibitors with structural scaffolds quite different from the existing ones, by using alternatively as target the commercially available bacterial NA. However, the viral and bacterial NA enzymes, all belonging to clan GH-E, contain a sixfold β-propeller as the prominent structural motif, and the release of progeny virus particles from infected host cells requires the presence of the exo-sialidase composed commonly of three acidic amino acids, a tyrosine, and three arginines (Chan et al. 2012). In view of the highly conserved active site of the target enzymes of the viral and bacterial origin (Chan et al. 2012), the three characterized NA inhibitors are of value in terms of drug discovery for the treatment of viral infections such as influenza.

Nature continues to be a big reservoir of bioactive compounds with unforeseeable structures and maybe unique modes of action. It remains an important task to identify new bioactive molecules for the discovery of new drugs, superior pesticides and other beneficial chemicals required for the human survival. As most well-accessed organisms have been investigated chemically and biologically, it has been advised to search for the new functional molecules from poorly explored taxonomic groups and/or form the creatures residing in previously inaccessible regions such as oceans (Leal et al. 2012). The workability of the suggestion has been confirmed by the present investigation reporting three new natural products from the culture of a marine Streptomyces seoulensis, which is terrestrially common as well.


The work was cofinanced by grants from the National Natural Science Foundation of China (21132004, 30901846 and 81121062) and from The Ministry of Science & Technology (Marine 863 Projects: 2011AA09070204 & 2013AA092901).