Selective biosynthesis of a rhamnosyl nosiheptide by a novel bacterial rhamnosyltransferase

Abstract

Nosiheptide (NOS) is a thiopeptide antibiotic produced by the bacterium Streptomyces actuosus. The hydroxyl group of 3‐hydroxypyridine in NOS has been identified as a promising site for modification, which we therefore aimed to rhamnosylate. After screening, Streptomyces sp. 147326 was found to regioselectively attach a rhamnosyl unit to the 3‐hydroxypyridine site in NOS, resulting in the formation of a derivative named NOS‐R at a productivity of 24.6%. In comparison with NOS, NOS‐R exhibited a 17.6‐fold increase in aqueous solubility and a new protective effect against MRSA infection in mice, while maintaining a similar in vitro activity. Subsequently, SrGT822 was identified as the rhamnosyltransferase in Streptomyces sp. 147326 responsible for the biosynthesis of NOS‐R using dTDP‐L‐rhamnose. SrGT822 demonstrated an optimal reaction pH of 10.0 and temperature of 55°C, which resulted in a NOS‐R yield of 74.9%. Based on the catalytic properties and evolutionary analysis, SrGT822 is anticipated to be a potential rhamnosyltransferase for use in the modification of various complex scaffolds.

A novel rhamnosyltransferase, derived from Streptomyces sp. 147326, was identified for the glycosylation of 3‐hydroxypyridine in nosiheptide using dTDP‐L‐rhamnose as the donor sugar.

INTRODUCTION

Nosiheptide (NOS), derived from Streptomyces actuosus ATCC25421 (Benazet et al., 1980), is an exceptionally potent thiopeptide antibiotic classified as a ribosomally synthesized, posttranslationally modified peptide (Bagley et al., 2015). With a novel mode of action inhibiting protein biosynthesis (Harms et al., 2008), NOS demonstrates remarkable in vitro antibacterial activity against a range of drug‐resistant Gram‐positive pathogens including methicillin‐resistant Staphylococcus aureus (MRSA), methicillin‐resistant Staphylococcus epidermidis (MRSE) and vancomycin‐resistant enterococci (VRE) (Bagley et al., 2015; Haste et al., 2012), even at concentrations in the ng/mL range. Accordingly, NOS was anticipated to serve as the latest last‐resort antibiotic for the clinical treatment of Gram‐positive bacterial infections; however, due to a lack of in vivo effects, NOS has typically been added to animal feed to prevent or alleviate intestinal bowel disease in livestock (Song et al., 2019; Xie et al., 2019).

To broaden the scope of derivatives for the development of NOS‐based drugs, we previously introduced a glucosyl unit to the hydroxyl group of the hydroxylpyridine core in NOS using a chimeric UDP‐glycosyltransferase originating from OleD and OleI from Streptomyces antibioticus (Zhao et al., 2023). The site and pattern of modification that we used are distinct from the reported structural transformations, which primarily focused on the dehydroalanine side‐chain and side‐ring of NOS (Fan et al., 2021; Liu et al., 2013; Tan et al., 2022; Vinogradov & Suga, 2020; Zhang et al., 2020). Importantly, the created monoglucosyl analogue of NOS not only exhibited improved solubility and stability but also demonstrated a promising in vivo protective effect against MRSA infection in mice. These findings highlight the potential of 3‐hydroxylpyridine in NOS as a viable target for modification with a view to increasing efficacy.

In addition to glycosylation with a glucosyl moiety, rhamnosylation has also been shown to be a valuable approach to improving the physicochemical properties of natural or artificial compounds. This modification would likely enhance or change the biological functions of these aglycones (Li et al., 2022). An illustrative example is the monorhamnosylation of taxifolin, which is a commonly used health supplement (Thuan et al., 2017). The resulting product, astilbin, not only exhibits increased solubility and retains the same anti‐oxidative and anti‐inflammatory activities as taxifolin but also demonstrates new anti‐tumour and anti‐obesity effects (Li et al., 2020; Wang et al., 2022). Therefore, we predicted that the monorhamnosylation of hydroxylpyridine in NOS may create new physicochemical and biological characteristics distinct from those observed for monoglucosyl NOS (NOS‐G) (Zhao et al., 2023; Figure S1). This would be beneficial for the further engineering of NOS by modifying its 3‐hydroxylpyridine core to create a suitable candidate for human/veterinary drug development.

Given the immense challenge in screening or designing an enzyme to glycosylate NOS with its large rigid scaffold, as previously demonstrated by the biosynthesis of NOS‐G (Zhao et al., 2023), we aimed to prepare rhamnosyl NOS (named NOS‐R) using a microbial transformation strategy and subsequently identify the specific bacterial rhamnosyltransferase that transferred the rhamnosyl unit to the 3‐hydroxylpyridine of NOS. To this end, we initially screened 50 strains of Streptomyces and identified Streptomyces sp. 147326 as possessing the ability to biosynthesize NOS‐R with a productivity of 24.6%. The rhamnosyltransferase responsible for the synthesis of NOS‐R was then found and enzymatically characterized through a combination of bioinformatics and activity comparison with glycosyltransferases in Actinobacteria fastidiosa JCM3276 (Figure 1). Therefore, the present study not only expands the range of glycosyl NOSs by attaching a rhamnosyl moiety to 3‐hydroxylpyridine but also presents a newly discovered microbial rhamnosyltransferase capable of selectively glycosylating bulky molecules.

FIGURE 1.

Monorhamnosylation of nosiheptide (NOS) catalysed by SrGT822.

EXPERIMENTAL PROCEDURES

Chemicals and reagents

Escherichia coli strains, plasmids and restriction endonucleases were obtained from Vazyme Biotech Co., Ltd. (Nanjing, China). Actinokineospora fastidiosa JCM3276 (A. fastidiosa CGMCC 4.1172) and Actinoalloteichus sp. AHMU CJ021 (Actinoalloteichus sp. CCTCC AA 2017040) were purchased from the China General Microbiological Culture Collection Center (Beijing, China) and the China Center for Type Culture Collection (Wuhan, China), respectively. All primers (Tables S1–S3) were synthesized by Suzhou Genewiz Biotechnology Co., Ltd. (Suzhou, China). The genomic DNA of Streptomyces sp. 147326 was also sequenced by Suzhou Genewiz Biotechnology Co., Ltd. All other chemicals and (bio)reagents are commercially available. NOS, nocathiacin I and dTDP‐L‐rhamnose were respectively prepared according to the literature (Jin et al., 2017; Yang et al., 2017, 2021).

HPLC and LC–MS analysis

HPLC was conducted on a YMC‐Pack ODS‐A column (5 μm, 150 × 4.6 mm; Kyoto, Japan) with mobile phase A (0.1% formic acid in water) and B (0.1% formic acid in acetonitrile) at 30°C and 330 nm with a flow rate of 1 mL/min. The gradient elution program was as follows: 10–30% mobile phase B for 6 min; 30–50% mobile phase B for 5 min; 50–70% mobile phase B for 5 min and 70–90% mobile phase B for 3 min. Mass spectrometry (MS) parameters were set as described in our previous work (Zhao et al., 2023). A semi‐preparative HPLC with a YMC‐Pack ODS‐A column (5 μm, 250 × 10 mm) was conducted using an isocratic elution of 55% acetonitrile in water (0.1% formic acid) for 36 min at a flow rate of 2 mL/min.

Biotransformation of NOS by Streptomyces and Actinomycetes

A total of 50 Streptomyces strains were isolated from Zijin Mountain in Nanjing, China. Each strain was cultured on solid ISP2 or MS medium at 28°C to promote sporulation. A 50‐mL aliquot of liquid ISP2 medium was then inoculated with 50 μL spore suspension for seed cultivation, which was carried out at 28°C for 2–3 days. Subsequently, 2 mL seed culture was transferred to 50 mL SM25 medium (10 g/L peptone, 21 g/L malt extract, 40 g/L glycerol, pH 6.5) for fermentation at 28°C. After 3–4 days, NOS dissolved in DMSO (20 mg/mL) was added to the culture broth at a final concentration of 0.4 mg/mL. To assess the production of rhamnosyl NOS, a 1‐mL aliquot of the fermentation broth was mixed with two volumes of methanol for 5 min. Following centrifugation at 12,000 rpm for 10 min and filtration, the resulting supernatant was subjected to LC–MS analysis. The same procedures were carried out for the biotransformation of NOS by Actinobacteria fastidiosa JCM3276 (Henssen et al., 1987), Actinoalloteichus sp. AHMU CJ021 (Xie et al., 2020) and Nocardia sp. ATCC 202099 (Ding et al., 2010; Table S4 and Figure S2).

Preparation and structural analysis of rhamnosyl NOS

The scalable biopreparation of glycosyl NOS by Streptomyces sp. 147326 was performed in a 15‐L fermenter in four batches. After the biotransformation of NOS, the reaction mixture was extracted twice with an equal volume of ethyl acetate. The organic phase from the four reaction batches was combined after HPLC and concentrated under reduced pressure. The resulting extract was subjected to semi‐preparative HPLC to purify the monoglycosyl derivative, a yellow powder with a purity greater than 95%, which was dissolved in DMSO‐d6 and then identified using a Bruker Advance AMX600 NMR spectrometer (Rheinstetten, Germany).

Water solubility and activity assessment of rhamnosyl NOS

A standard curve was generated by HPLC using NOS‐R concentrations from 31.25 to 1000 ng/mL. The peak area (Y) and NOS‐R concentration (X) were used to obtain a regression equation, denoted as Y = 2,825,687X − 50,053 (R ² = 0.9999). Subsequently, the aqueous solubility of NOS‐R was determined and compared with that of NOS in solutions with pH values varying from 5.0 to 7.0 using previously described methods (Zhao et al., 2023). The in vitro antibacterial activity of NOS‐R against MRSA, MRSE, Enterococcus faecium and Enterococcus faecalis was determined according to CLSI guidelines.

Gene mining using NosM as a probe

Using the leader peptide NosM (GenBank accession no. ACR48342.1) from the nosiheptide biosynthetic gene cluster (BGC) as a probe (Figure S3), an online BLAST search was conducted under a non‐redundant protein sequence (nr) model in the NCBI database. Genomic data for the excavated strains containing a homologous counterpart of NosM were submitted to the antiSMASH database to display their secondary metabolite BGCs using the default parameters. Then, specific BGCs in the selected strains were selected according to their possession of both a glycosyltransferase and an NosM‐like protein.

Expression and characterization of Af‐GTs from A. fastidiosa JCM3276

The genes encoding Af‐GT1‐GT4 were amplified using the genomic DNA of A. fastidiosa JCM3276 as a template (GenBank accession no. NZ_BMRB01000004.1). Each gene was His‐tagged at the C‐terminus and inserted into pET22b (+) using Nde I and Xho I for soluble expression in E. coli. Subsequently, separate expression of the four enzymes in E. coli BL21(DE3) was induced with 0.5 mM IPTG at 20°C for 12 h. The expression levels of all enzymes were analysed by SDS‐PAGE. Following the same procedure, SrGT822 from Streptomyces sp. 147326 was also expressed and analysed. Enzymes were individually purified from the cell lysate using a His‐TALON metal nickel affinity resin (Takara Bio, Shiga, Japan) in a gradient with Tris–HCl buffer (50 mM Tris–HCl, 200 mM NaCl, pH 9.0) containing 10–300 mM imidazole. The molecular weight of SrGT822 was determined using a gel filtration calibration kit according to the manufacturer's protocol.

The activity of the purified enzymes was assessed using dTDP‐L‐rhamnose as the sugar donor and NOS as the acceptor. The reaction volume was 2 mL, which contained 1 mg/mL NOS dissolved in 0.4 mL DMSO, 2 mM dTDP‐L‐rhamnose, 700 μg purified enzyme and 50 mM Tris–HCl (pH 8.0). The reaction was initiated by the addition of enzyme and allowed to incubate at 30°C for 12 h. After incubation, the GT‐catalysed reaction was terminated by the addition of an equal amount of methanol and then analysed by HPLC after centrifugation at 12,000 rpm for 10 min and filtration using membrane filter (0.22 μm).

Search for the rhamnosyltransferase in Streptomyces sp. 147326

The whole genome of Streptomyces sp. 147326 was sequenced by Suzhou Genewiz Biotechnology Co., Ltd., which was followed by a combined analysis of the antiSMASH and CAZy database. Subsequently, all rhamnosyltransferases from Streptomyces sp. 147326 were aligned with the characterized Af‐GT‐3 from A. fastidiosa JCM3276 based on their amino acid sequences. The gene encoding the rhamnosyltransferase SrGT822 in Streptomyces sp. 147326 was knocked out according to the reported method (Liu et al., 2013).

Investigation of the catalytic properties of SrGT822

Site‐directed mutagenesis of the putative catalytic dyad of SrGT822 was performed using pET22b‐SrGT822 as the template according to the protocol provided with the Axygen Mutation Kit. To determine the optimal pH, assays were conducted at 30°C for 12 h. The pH values ranged from 6.0 to 11.0, using different buffers for each pH scope: 50 mM KH₂PO₄–NaOH buffer for pH 6.0–8.0, 50 mM Tris–HCl buffer for pH 8.0–9.0, 50 mM glycine–NaOH buffer for pH 9.0–10.0 and 50 mM NaHCO₃–NaOH buffer for pH 10.0–11.0. To obtain the optimal temperature, the reactions were performed at pH 10.0 in a final volume of 2.0 mL at temperatures ranging from 20 to 70°C with an interval of 5°C. Using the optimized parameters, the effects of metal ions (Mn²⁺, Zn²⁺, Mg²⁺, Ca²⁺, Fe³⁺, Cu²⁺, Ni²⁺, K⁺ and Na⁺) and EDTA were then investigated at a concentration of 5 mM. Additionally, the thermostability of SrGT822 was examined at 4 and 30°C for 12 h. Finally, a total of 17 compounds belonging to various categories were utilized to assess the substrate promiscuity of SrGT822.

Kinetics analysis of SrGT822 toward NOS

To determine the kinetic parameters of SrGT822 toward NOS, 2.0 mL assay mixture containing 2 mM dTDP‐L‐rhamnose, a varying concentration of NOS (from 0.0156 to 2 mM), 0.7 mg purified SrGT822 and 50 mM glycine–NaOH buffer (pH 10.0) was incubated at 50°C for 45 min. Kinetic analysis was performed using GraphPad Prism 9.0. One unit enzyme activity produced 1 μmol NOS‐R per minute per mg protein.

Sequence alignment and phylogenetic analysis

We selected 72 well‐characterized rhamnosyltransferases deposited in CAZy database, as listed in Table S5, to investigate the evolutionary hierarchy of SrGT822 and Af‐GT3. The amino acid sequences of all enzymes were aligned using MAFFT. The phylogenetic analysis and estimation of phylogeny were conducted in UniProt (https://www.uniprot.org/). Subsequently, a neighbour‐joining tree was constructed and visualized using the iTOL website.

RESULTS AND DISCUSSION

Biosynthesis of rhamnosyl NOS by Streptomyces sp. 147326

As demonstrated in our previous study, O‐glycosylation of the 3‐hydroxylpyridine core of NOS not only enhances its solubility and stability but also confers a new protective effect against pathogen infection in mice (Zhao et al., 2023). Considering the potential of rhamnosylation to change the biological or physicochemical performance of natural and unnatural molecules (Li et al., 2022; Yang et al., 2021), we planned to attach a rhamnosyl unit to the hydroxyl group of 3‐hydroxylpyridine in NOS using a biocatalytic approach, taking advantage of its specific site‐selectivity (He et al., 2022). To address the issue of screening a suitable rhamnosyltransferase, the selective biocatalytic rhamnosylation of NOS was attempted through whole‐cell biotransformation using 50 strains of Streptomyces isolated from Zijin Mountain in Nanjing, China. To evaluate the production of rhamnosyl NOS, each biotransformation mixture was analysed separately by LC–MS. The results revealed the presence of a new compound in the culture broth of Streptomyces sp. 147326, which exhibited a higher hydrophilicity than NOS (Figure 2A). The compound displayed an increased [M + H]⁺ value of 1368.30 accompanied by a characteristic ion of NOS at m/z 1222.20 in positive‐ion mode, indicating the occurrence of O‐monorhamnosylation of NOS (Figure 2B). Further investigation was required to determine the specific site of modification based on the chemical structure of this rhamnosyl derivative.

FIGURE 2.

LC–MS analysis of monorhamnosyl NOS produced by Streptomyces sp. 147326. (A) HPLC analysis of the biotransformation mixture produced by Streptomyces sp. 147326. Reaction 1 refers to the biotransformation mixture produced by Streptomyces sp. 147326 using NOS as a substrate. Reaction 2 represents the biotransformation mixture produced by Streptomyces sp. 147326 without NOS. Reaction 3 is the reaction mixture prepared using purified SrGT22. Monorhamnosyl NOS is indicated with a red asterisk, while NOS is marked with a blue asterisk. (B) Mass spectrum for monorhamnosyl NOS. The ion peak at 1222.20 is produced by the MS fragment of rhamnosyl NOS, which has an [M + H]⁺ value of 1368.30 in positive‐ion mode.

Structural identification of rhamnosyl NOS

To elucidate the structure and anti‐pathogen activity, a biopreparative system for rhamnosyl NOS by Streptomyces sp. 147326 was scaled up in a 15‐L fermenter for four batches, with the substrate NOS at a final concentration of 0.4 mg/mL. After extraction with ethyl acetate, a yellow powder of monorhamnosyl NOS (~150 mg) was obtained using semi‐preparative HPLC, with the purity exceeding 95%. Further NMR analysis confirmed that the monorhamnosyl NOS was the desired derivative, designated as NOS‐R, which possesses an O‐α‐glycosidic bond at the hydroxyl group of 3‐hydroxylpyridine in NOS as compared with NMR data for reported rhamnosyl compounds (Ren et al., 2022; Figure 1 and Table S6). This finding suggests that Streptomyces sp. 147326 has a specific rhamnosyltransferase(s) that can recognize the bulky NOS as substrate and regioselectively rhamnosylate its substructure of 3‐hydroxylpyridine.

Aqueous solubility and biological activity of NOS‐R

In accordance with a previous study on NOS and its monoglucosyl derivative (NOS‐G) (Zhao et al., 2023), we investigated the solubility of NOS‐R in aqueous solution at pH values ranging from 5.0 to 7.0. NOS‐R also displayed maximum solubility at pH 6.0, with a final concentration of 525.8 ng/mL, which was approximately 18.6‐fold higher than NOS (28.3 ng/mL; Figure S4). However, NOS‐R had a slightly lower solubility in water than NOS‐G (596.2 ng/mL; Figure S4). This difference may be attributed to the better hydrophilicity of the glucose unit in NOS‐G, which has four free hydroxyl groups, in comparison with the rhamnose moiety in NOS‐R, which has only three.

Next, the in vitro antibacterial activity of NOS‐R was evaluated against strains of MRSA, MRSE, Enterococcus faecalis and Enterococcus faecium, using NOS, NOS‐G and vancomycin as controls. The in vitro activity of NOS‐R was slightly less than that of NOS against all the tested pathogens, with MIC values ranging from 0.0078 to 0.0625 μg/mL; however, NOS‐R exhibited a significantly higher efficacy than vancomycin (MIC values: 0.125–2 μg/mL) and NOS‐G (MIC values: 0.03125–0.5 μg/mL; Table 1). As anticipated, NOS‐R gave a potential in vivo protection to mice infected with MRSA (data not shown), which was similar to previous observations with NOS‐G (Zhao et al., 2023). The combination of rhamnosylation and glucosylation of NOS further indicates that modifying the hydroxyl group of 3‐hydroxypyridine would be a promising approach to create NOS‐derived analogues with improved efficacy.

TABLE 1.

Antimicrobial activity of the test compound against microorganisms

Strain	MIC (μg/mL)
	Vancomycin	NOS	NOS‐G	NOS‐R
MRSE	0.125–1	0.0078–0.0156	0.0625–0.125	0.03125–0.0625
MRSA	0.25–2	0.0078–0.0156	0.0625–0.25	0.03125–0.0625
Enterococcus faecalis	1–2	0.0078–0.0156	0.0625–0.5	0.0078–0.0156
Enterococcus faecium	1–2	0.0078–0.0156	0.03125–0.25	0.0078–0.0156

Biosynthesis of NOS‐R by A. fastidiosa JCM3276

A comparative analysis of the 16S rDNA sequence revealed that Streptomyces sp. 147326 isolated from Zijin Mountain in Nanjing shares a similarity of 99.93% with Streptomyces avidinii NBRC 13429 (Babalola et al., 2009), suggesting that both strains belong to the same species. After conducting a thorough search in the PubMed database, we were unable to find any research specifically related to rhamnosyltransferases in S. avidinii NBRC 13429. To elucidate the specific rhamnosyltransferase(s), we proceeded to sequence the genomic DNA of Streptomyces sp. 147326 (GenBank accession no. CP134201) and then submitted the results to both the antiSMASH and CAZy databases for comprehensive analysis. However, Streptomyces sp. 147326 does not possess any biosynthetic gene clusters (BGCs) responsible for the production of NOS and its analogues. Additionally, the strain was found to have over 140 putative rhamnosyltransferases, rendering the identification of a specific enzyme through knockout technology unlikely.

Alternatively, bioinformatics analysis was performed to identify the rhamnosyltransferase(s) responsible for the biosynthesis of NOS‐R in Streptomyces sp. 147326. The precursor peptide NosM, which is derived from the BGC responsible for NOS production in S. actuosus ATCC 25421 (Yu et al., 2009), was utilized as a probe to conduct a gene mining search in the GenBank database. A total of 17 strains from various origins were obtained, which are listed in Table S4. To uncover a candidate strain, an intense analysis was conducted on the 17 strains, focusing on the presence of a BGC containing both a homologous NosM protein and a glycosyltransferase. A total of six actinomycete strains were obtained for further testing: A. fastidiosa JCM 3276 (Henssen et al., 1987), Actinoalloteichus sp. AHMU CJ021 (Xie et al., 2020), Actinokineospora sp. UTMC 2448 (Dashti et al., 2022), Nocardia sp. ATCC 202099 (Ding et al., 2010), Nocardia arthritidis (Herisse et al., 2020) and Micromonospora sp. NBC_01655 (GenBank accession no. NZ_JAPEQF010000001; Figures 3 and S4). However, A. fastidiosa JCM3276 and Actinokineospora sp. UTMC 2448 exhibited 100% identity in their 16S rDNA sequences, as well as a similar composition of the BGCs containing a NosM‐like protein, indicating that these may be the same strain with different names and numbers. Furthermore, due to the availability of certain strains, we chose A. fastidiosa JCM3276, Actinoalloteichus sp. AHMU CJ021 and Nocardia sp. ATCC 202099 to evaluate conversion of NOS. Among the tested strains, only A. fastidiosa JCM3276 demonstrated the ability to biosynthesize NOS‐R using NOS as a substrate.

FIGURE 3.

Comparison of the biosynthetic gene clusters containing both NosM‐like peptide and glycosyltransferase. The biosynthetic gene clusters (BGCs) containing both NosM‐like peptide and glycosyltransferase were discovered through gene mining using NosM as a probe. The NosM‐like peptide in each BGC is marked with a red asterisk and a light‐yellow background, while glycosyltransferases are marked with a green asterisk and underlined.

Rhamnosylation of NOS catalysed by SrGT822 in Streptomyces sp. 147326

In addition to possessing the ability to rhamnosylate NOS, A. fastidiosa JCM3276 was also found to produce a NOS congener called NOC‐FG, which consists of four glycosyl units including a modified rhamnosyl moiety linked to the 3‐hydroxylpyridine (Figure S1). This multiglycosylated molecule was first discovered in the fermentation broth of Actinokineospora sp. UTMC 2448, with the published name persiathiacin (Dashti et al., 2022). This further suggested the possible presence of a rhamnosyltransferase(s) responsible for this modification in A. fastidiosa JCM3276; therefore, a more detailed analysis was conducted. Four putative glycosyltransferases were uncovered, namely Af‐GT1–GT4, located within the BGC containing a NosM‐like protein (Figure 3). After expression in E. coli and purification of the four enzymes (Figure 4A), Af‐GT3 was demonstrated to be the α‐rhamnosyltransferase affording the biosynthesis of NOS‐R, but with a negligible conversion of 2.56%. Based on these findings, we conducted an amino acid alignment between Af‐GT3 and all the putative rhamnosyltransferases from Streptomyces sp. 147326. Af‐GT3 exhibited the highest identity of 39.6% to SrGT822 (GenBank accession no. OR542859), suggesting that SrGT822 could potentially be the enzyme responsible for the production of NOS‐R. Additionally, the amino sequence of Af‐GT3 was found to be identical to that of PerS4 (GenBank accession no. WP_189212713.1) in Actinokineospora sp. UTMC 2448, suggesting that PerS4 is the enzyme affording the attachment of rhamnose to produce persiathiacin (Dashti et al., 2022).

FIGURE 4.

SDS‐PAGE analysis of purified Af‐GT1‐4 and SrGT822. (A) SDS‐PAGE analysis of purified Af‐GT1‐4. Lane M, protein marker; Lane 1, His‐tagged Af‐GT1; Lane 2, His‐tagged Af‐GT2; Lane 3, His‐tagged Af‐GT3; Lane 4, His‐tagged Af‐GT4. (B) SDS‐PAGE analysis of purified SrGT822. Lane M, protein marker; Lane 1, His‐tagged SrGT822. SDS‐PAGE was performed in a 10% (w/v) polyacrylamide gel.

For functional analysis in vitro, SrGT822 was expressed in E. coli and purified to achieve an SDS‐PAGE purity (Figure 4B) for constructing the reaction system using NOS as the substrate. The catalytic activity of SrGT822 was examined using dTDP‐L‐rhamnose as the active sugar donor, which is predominantly used by bacterial‐derived rhamnosyltransferases (Yang et al., 2021). The production of NOS‐R confirmed SrGT822 as an enzyme capable of catalysing the rhamnosylation of 3‐hydroxylpyridine in NOS (Figure 2A). The molecular weight of 37.6 kDa determined by gel chromatography further suggested that the active form of SrGT822 is monomeric in nature. After deletion of the gene encoding SrGT822, the mutant strain Streptomyces sp. 147326‐Δsrgt822 completely lost the ability to rhamnosylate NOS, implying that SrGT822 is the sole α‐rhamnosyltransferase responsible for the biosynthesis of NOS‐R in Streptomyces spp. 147326.

Identification of the catalytic dyad of SrGT822

The amino acid sequences of four rhamnosyltransferases with characterized function, SpnG from Saccharopolyspora spinosa (Chen et al., 2009), AraGT from Streptomyces echinatus (Sianidis et al., 2006), ElmGT from Streptomyces olivaceus (Fischer et al., 2002) and Cpz31 from Streptomyces sp. MK730‐62F2 (Kaysser et al., 2010), were aligned with that of SrGT822. According to the high similarity to these four enzymes, SrGT822 is presumed to be a member of the GT‐B topological glycosyltransferase family, which all follow an SN₂‐like catalytic mode (Liang et al., 2015). Additionally, based on the characterized rhamnosyltransferases (Chen et al., 2009; Fischer et al., 2002; Kaysser et al., 2010; Sianidis et al., 2006), two amino acids His13 and Asp308, marked with green asterisks, were identified as the potential catalytic dyad of SrGT822 (Figure 5A), which was later confirmed by the three inactive mutants, H13A, D308A and H13A/D308A (Figure 5B). Accordingly, we propose that His13 acts as the catalytic base to abstract a proton from the hydroxyl group of 3‐hydroxylpyridine in NOS. Subsequently, the C1 of dTDP‐L‐rhamnose is subject to nucleophilic attack by the resulting de‐protonated hydroxyl group, leading to the formation of a new rhamnoside, NOS‐R (Figure 5C).

FIGURE 5.

Identification and verification of the catalytic dyad of SrGT822. (A) Sequence alignment of SrGT822 with other known rhamnosyltransferases. The putative catalytic residues, His13 and Asp308, are indicated by a green asterisk. (B) Relative activity of SrGT822 mutants toward NOS. The relative activity was calculated as the ratio A _V/A ₈₂₂, where A ₈₂₂ and A _V are the enzyme activities of SrGT822 and its mutants, respectively. ND, not detected. (C) Proposed catalytic mode of SrGT822. The catalytic dyad of SrGT822, consisting of His13 and Asp308, is responsible for the rhamnosylation of NOS using dTDP‐L‐rhamnose through an SN₂‐like mechanism.

Enzymatic properties of SrGT822

The basic reaction properties of SrGT822 were evaluated, including the optimal pH, temperature, metal ions, thermostability and substrate specificity. Comparison of the production rate of NOS‐R at different pH values using a series of buffers demonstrated that purified SrGT822 had a relatively broad optimal pH, which was maximum at pH 10.0 using glycine–NaOH buffer (Figure 6A). SrGT822 had a broad optimal temperature centred around 55°C, with a NOS‐R production rate of 74.9% (Figure 6B). The incomplete conversion of NOS to NOS‐R may be attributed to insolubility of NOS in the reaction mixture (Xie et al., 2019). SrGT822 exhibited good thermostability at both tested temperatures over a 12‐h period, without a significant loss of enzyme activity (Figure 6C). However, the relative activity of SrGT822 at 30°C (84.4%) was less than that at 4°C (90.5%), suggesting that a low temperature is preferable for the long‐time storage of this rhamnosyltransferase. Metal ions have been found to significantly impact the catalytic function of rhamnosyltransferases (Feng et al., 2018). Therefore, we conducted a comparative evaluation of the effect of nine different metal ions and EDTA on SrGT822 activity. Among all the ions tested, only Mg²⁺ exhibited a positive effect on SrGT822, resulting in a 23% increase in activity. On the other hand, Cu²⁺ demonstrated the strongest inhibition of this rhamnosyltransferase, which led to a significant loss (~97%) in enzyme activity (Figure 6D). Under optimized parameters, SrGT822 displayed good affinity toward NOS, with an apparent Km value of 107.22 ± 13.38 μM and a V _max of 4.19 ± 0.14 μmol/min/mg.

FIGURE 6.

Enzymatic properties of SrGT822. (A) Effects of reaction pH on the production of NOS‐R. (B) Effect of reaction temperature on the production of NOS‐R. The production rate of NOS‐R was calculated using A _n/A _t × 100%, where A _n is the peak area of produced NOS‐R and A ₀ is the total combined area of NOS‐R and NOS. (C) Effect of temperature on the stability of SrGT822. Relative enzyme activity refers to the percentage ratio of enzyme activity at each specific timepoint to the initial activity of SrGT822. (D) Effect of metal ions on the catalytic activity of SrGT822. The relative enzyme activity was calculated using E _m/E ₈₂₂ × 100%, where E _m represents the activity of SrGT822 when incubated with a metal ion, while E ₈₂₂ is the activity of free SrGT822.

In addition to NOS, 16 other compounds, each of which contains at least one hydroxyl group, were selected according to molecular size for investigation of the substrate spectrum of SrGT822, which may facilitate the development of SrGT822 as a valuable tool. LS–MS analysis showed that, in addition to compound 16 (NOS), only three compounds (5, 7 and 17), each belonging to a different class (coumarin, flavonol and thiopeptide, respectively), can be recognized by SrGT822 as a substrate for monorhamnosylation, with a conversion rate below 60% (Figure S5). The strict substrate specificity of SrGT822 is evident; however, the rhamnosylation of nocathiacin I (compound 17), a thiopeptide antibiotic from Nocardia sp. ATCC 202099 (Constantine et al., 2002), suggests that the active centre of SrGT822 has a large space sufficient to accommodate complex structures with a bulky scaffold. Perhaps SrGT822 can be used as a tool for the modification of natural or artificial drug candidates with a large structural volume in the future.

Phylogenetic analysis of SrGT822 and Af‐GT3

To understand the relationship among SrGT822, Af‐GT3 and the 72 characterized rhamnosyltransferases deposited in the CAZy database, multiple amino acid sequence alignment and evolutionary comparison were conducted and the resulting phylogenetic tree is presented in Figure 7. SrGT822 and Af‐GT3, together with 34 other rhamnosyltransferases, were categorized under the GT‐1 family of glycosyltransferases according to the CAZy system. Further analysis revealed that SrGT822 and Af‐GT3, along with rhamnosyltransferases Cpz31 (Kaysser et al., 2010) and ElmGT (Fischer et al., 2002), were classified in the same clade marked in pink lines under the GT branch of antibiotics and were adjacent to SpnG, a well‐studied bacterial glycosyltransferase (Chen et al., 2009). Cpz31 and ElmGT have the ability to transfer the L‐rhamnose from dTDP‐L‐rhamnose to the bulky aglycone acceptors, resulting in the production of a range of potent antimycobacterials called caprazamycins and elloramycins, respectively. On the other hand, rhamnosyltransferase SpnG from Saccharopolyspora spinosa is responsible for attaching the rhamnosyl moiety to anthracycline antibiotics with a relatively simple structure, such as tetracenomycins. The bioinformatics and verified catalytic functions suggest that enzymes in the clade containing SrGT822, Af‐GT3, Cpz31 and ElmGT could be utilized as potential tools for the rhamnosylation of complex molecules, thereby altering their physicochemical and/or biological properties.

FIGURE 7.

Phylogenetic analysis of SrGT822 and Af‐GT3. SrGT822 and Af‐GT3, in addition to 72 rhamnosyltransferases from plants, bacteria and viruses, were analysed using the Unrooted Tree method for categorization into different families in the phylogenetic tree. SrGT822 and Af‐GT3 were both classified as the GT1 family and marked in red.

CONCLUSIONS

After testing a total of 50 strains, a new glycosyl derivative (NOS‐R) was obtained through biotransformation conducted by Streptomyces sp. 147326, during which a rhamnosyl unit was attached to the 3‐hydroxypyridine core of NOS. This monorhamnosyl derivative exhibited similar improvements in both physicochemical and biological properties as its counterpart NOS‐G, which is a 3‐hydroxypyridine monoglucosylated NOS achieved by a designed chimeric OleD‐10. Subsequently, through gene mining and comparison with the glycosyltransferase Af‐GT3 in A. fastidiosa JCM3276, SrGT822 was identified as the rhamnosyltransferase in Streptomyces sp. 147326 responsible for the glycosylation of 3‐hydroxypyridine in NOS using dTDP‐L‐rhamnose as the donor sugar. Furthermore, the catalytic characteristics of SrGT822 in modifying bulky compounds with complex scaffolds make it a valuable tool for chemical synthesis and human/veterinary drug development in the future.

AUTHOR CONTRIBUTIONS

Yali Du: Data curation (equal); formal analysis (equal); methodology (equal); writing – original draft (supporting). Yuan Xia: Data curation (equal); formal analysis (equal); methodology (equal); writing – original draft (supporting). Lingrui Wu: Methodology (supporting). Lu Chen: Methodology (supporting). Jiale Rong: Methodology (supporting). Junting Fan: Formal analysis (equal). Yijun Chen: Conceptualization (supporting); formal analysis (equal). Xuri Wu: Conceptualization (lead); formal analysis (equal); supervision (equal); writing – original draft (equal); writing – review and editing (lead).

CONFLICT OF INTEREST STATEMENT

The authors declare no competing financial interest.

Supporting information

Data S1.

Du, Y. , Xia, Y. , Wu, L. , Chen, L. , Rong, J. , Fan, J. et al. (2024) Selective biosynthesis of a rhamnosyl nosiheptide by a novel bacterial rhamnosyltransferase. Microbial Biotechnology, 17, e14412. Available from: 10.1111/1751-7915.14412