Combinatorial metabolic engineering of Streptomyces sp. CB03234-S for the enhanced production of anthraquinone-fused enediyne tiancimycins.

BACKGROUND: Anthraquinone-fused enediynes (AFEs) are excellent payloads for antibody-drug conjugates (ADCs). The yields of AFEs in the original bacterial hosts are extremely low. Multiple traditional methods had been adopted to enhance the production of the AFEs. Despite these efforts, the production titers of these compounds are still low, presenting a practical challenge for their development. Tiancimycins (TNMs) are a class of AFEs produced by Streptomyces sp. CB03234. One of their salient features is that they exhibit rapid and complete cell killing ability against various cancer cell lines. RESULTS: In this study, a combinatorial metabolic engineering strategy guided by the CB03234-S genome and transcriptome was employed to improve the titers of TNMs. First, re-sequencing of CB03234-S (Ribosome engineered mutant strains) genome revealed the deletion of a 583-kb DNA fragment, accounting for about 7.5% of its genome. Second, by individual or combined inactivation of seven potential precursor competitive biosynthetic gene clusters (BGCs) in CB03234-S, a double-BGC inactivation mutant, S1009, was identified with an improved TNMs titer of 28.2 ± 0.8 mg/L. Third, overexpression of five essential biosynthetic genes, including two post-modification genes, and three self-resistance auxiliary genes, was also conducted, through which we discovered that mutants carrying the core genes, tnmE or tnmE10, exhibited enhanced TNMs production. The average TNMs yield reached 43.5 ± 2.4 mg/L in a 30-L fermenter, representing an approximately 360% increase over CB03234-S and the highest titer among all AFEs to date. Moreover, the resulting mutant produced TNM-W, a unique TNM derivative with a double bond instead of a common ethylene oxide moiety. Preliminary studies suggested that TNM-W was probably converted from TNM-A by both TnmE and TnmE10. CONCLUSIONS: Based on the genome and transcriptome analyses, we adopted a combined metabolic engineering strategy for precursor enrichment and biosynthetic pathway reorganization to construct a high-yield strain of TNMs based on CB03234-S. Our study establishes a solid basis for the clinical development of AFE-based ADCs.

Chalkophomycin Biosynthesis Revealing Unique Enzyme Architecture for a Hybrid Nonribosomal Peptide Synthetase and Polyketide Synthase.

Chalkophomycin is a novel chalkophore with antibiotic activities isolated from Streptomyces sp. CB00271, while its potential in studying cellular copper homeostasis makes it an important probe and drug lead. The constellation of N-hydroxylpyrrole, 2H-oxazoline, diazeniumdiolate, and methoxypyrrolinone functional groups into one compact molecular architecture capable of coordinating cupric ions draws interest to unprecedented enzymology responsible for chalkophomycin biosynthesis. To elucidate the biosynthetic machinery for chalkophomycin production, the chm biosynthetic gene cluster from S. sp. CB00271 was identified, and its involvement in chalkophomycin biosynthesis was confirmed by gene replacement. The chm cluster was localized to a ~31 kb DNA region, consisting of 19 open reading frames that encode five nonribosomal peptide synthetases (ChmHIJLO), one modular polyketide synthase (ChmP), six tailoring enzymes (ChmFGMNQR), two regulatory proteins (ChmAB), and four resistance proteins (ChmA'CDE). A model for chalkophomycin biosynthesis is proposed based on functional assignments from sequence analysis and structure modelling, and is further supported by analogy to over 100 chm-type gene clusters in public databases. Our studies thus set the stage to fully investigate chalkophomycin biosynthesis and to engineer chalkophomycin analogues through a synthetic biology approach.

Characterization of Chalkophomycin, a Copper(II) Metallophore with an Unprecedented Molecular Architecture.

Metals play essential roles in life by coordination with small molecules, proteins, and nucleic acids. Although the coordination of copper ions in many proteins and methanobactins is known, the coordination chemistry of Cu(II) in natural products and their biological functions remain underexplored. Herein, we report the discovery of a Cu(II)-binding natural product, chalkophomycin (CHM, 1), from Streptomyces sp. CB00271, featuring an asymmetric square-coordination system of a bidentate diazeniumdiolate and a conjugated 1H-pyrrole 1-oxide-oxazoline. The structure of 1 may inspire the synthesis of Cu(II) chelators against neurodegenerative diseases or Cu(II)-based antitumor therapeutics.

177 Saponins, Including 11 New Compounds in Wild Ginseng Tentatively Identified via HPLC-IT-TOF-MSn, and Differences among Wild Ginseng, Ginseng under Forest, and Cultivated Ginseng.

Wild ginseng (W-GS), ginseng under forest (F-GS, planted in mountain forest and growing in natural environment), and cultivated ginseng (C-GS) were compared via HPLC-DAD and HPLC-IT-TOF-MSn. A total of 199 saponins, including 16 potential new compounds, were tentatively identified from 100 mg W-GS (177 saponins in W-GS with 11 new compounds), F-GS (56 saponins with 1 new compound), and C-GS (60 saponins with 6 new compounds). There were 21 saponins detected from all the W-GS, F-GS, and C-GS. Fifty saponins were only detected from W-GS, including 23 saponins found in ginseng for the first time. Contents of ginsenosides Re (12.36-13.91 mg/g), Rh1 (7.46-7.65 mg/g), Rd (12.94-12.98 mg/g), and the total contents (50.52-55.51 mg/g) of Rg1, Re, Rf, Rb1, Rg2, Rh1, and Rd in W-GS were remarkably higher than those in F-GS (Re 1.22-3.50 mg/g, Rh1 0.15-1.49 mg/g, Rd 0.19-1.49 mg/g, total 5.69-18.74 mg/g), and C-GS (Re 0.30-3.45 mg/g, Rh1 0.05-3.42 mg/g, Rd 0.17-1.68 mg/g, total 2.99-19.55 mg/g). Contents of Re and Rf were significantly higher in F-GS than those in C-GS (p < 0.05). Using the contents of Re, Rf, or Rb1, approximately a half number of cultivated ginseng samples could be identified from ginseng under forest. Contents of Rg1, Re, Rg2, Rh1, as well as the total contents of the seven ginsenosides were highest in ginseng older than 15 years, middle-high in ginseng between 10 to 15 years old, and lowest in ginseng younger than 10 years. Contents of Rg1, Re, Rf, Rb1, Rg2, and the total of seven ginsenosides were significantly related to the growing ages of ginseng (p < 0.10). Similarities of chromatographic fingerprints to W-GS were significantly higher (p < 0.05) for F-GS (median: 0.824) than C-GS (median: 0.745). A characteristic peak pattern in fingerprint was also discovered for distinguishing three types of ginseng. Conclusively, wild ginseng was remarkably superior to ginseng under forest and cultivated ginseng, with ginseng under forest slightly closer to wild ginseng than cultivated ginseng. The differences among wild ginseng, ginseng under forest, and cultivated ginseng in saponin compositions and contents of ginsenosides were mainly attributed to their growing ages.

Genome mining of Streptomyces sp. CB00271 as a natural high-producer of ß-rubromycin and the resulting discovery of ß-rubromycin acid.

ß-rubromycin (ß-RUB) (1) is an efficient inhibitor of human telomerase possessing a unique spiroketal moiety as a potential pharmacophore and regarded as a promising anticancer drug lead. But the development of (ß-RUB) (1) has long been hampered by its low titer and very poor water solubility. By adopting a genome mining strategy, an FAD-dependent monooxygenase RubN involving with the formation of the spiro system was applied as the probe and Streptomyces sp. CB00271 was screened out from our strain collection as an alternative natural high producer of ß-RUB (1). After a series of fermentation optimizations, CB00271 could produce 124.8 ± 3.4 mg/L ß-RUB (1), which was the highest titer up to now. Moreover, the enhanced production of ß-RUB (1) in fermentation broth also led to the discovery of a new congener ß-RUB acid (7), which was structurally elucidated as the acid form of ß-RUB (1). Comparing to ß-RUB (1), the substituted carboxyl group endowed ß-RUB acid (7) much better solubility in serum and resulted in its higher activity towards tumor cells. Our work set up a solid base for the pilot-scale production of ß-RUB (1) and its congeners to facilitate their future development as promising anticancer drug leads, and also provide an alternative and practical strategy for the exploitation of other important microbial natural products.

MutagenPred-GCNNs: A Graph Convolutional Neural Network-Based Classification Model for Mutagenicity Prediction with Data-Driven Molecular Fingerprints.

An important task in the early stage of drug discovery is the identification of mutagenic compounds. Mutagenicity prediction models that can interpret relationships between toxicological endpoints and compound structures are especially favorable. In this research, we used an advanced graph convolutional neural network (GCNN) architecture to identify the molecular representation and develop predictive models based on these representations. The predictive model based on features extracted by GCNNs can not only predict the mutagenicity of compounds but also identify the structure alerts in compounds. In fivefold cross-validation and external validation, the highest area under the curve was 0.8782 and 0.8382, respectively; the highest accuracy (Q) was 80.98% and 76.63%, respectively; the highest sensitivity was 83.27% and 78.92%, respectively; and the highest specificity was 78.83% and 76.32%, respectively. Additionally, our model also identified some toxicophores, such as aromatic nitro, three-membered heterocycles, quinones, and nitrogen and sulfur mustard. These results indicate that GCNNs could learn the features of mutagens effectively. In summary, we developed a mutagenicity classification model with high predictive performance and interpretability based on a data-driven molecular representation trained through GCNNs.

Ensemble Learning Models Based on Noninvasive Features for Type 2 Diabetes Screening: Model Development and Validation.

BACKGROUND: Early diabetes screening can effectively reduce the burden of disease. However, natural population-based screening projects require a large number of resources. With the emergence and development of machine learning, researchers have started to pursue more flexible and efficient methods to screen or predict type 2 diabetes. OBJECTIVE: The aim of this study was to build prediction models based on the ensemble learning method for diabetes screening to further improve the health status of the population in a noninvasive and inexpensive manner. METHODS: The dataset for building and evaluating the diabetes prediction model was extracted from the National Health and Nutrition Examination Survey from 2011-2016. After data cleaning and feature selection, the dataset was split into a training set (80%, 2011-2014), test set (20%, 2011-2014) and validation set (2015-2016). Three simple machine learning methods (linear discriminant analysis, support vector machine, and random forest) and easy ensemble methods were used to build diabetes prediction models. The performance of the models was evaluated through 5-fold cross-validation and external validation. The Delong test (2-sided) was used to test the performance differences between the models. RESULTS: We selected 8057 observations and 12 attributes from the database. In the 5-fold cross-validation, the three simple methods yielded highly predictive performance models with areas under the curve (AUCs) over 0.800, wherein the ensemble methods significantly outperformed the simple methods. When we evaluated the models in the test set and validation set, the same trends were observed. The ensemble model of linear discriminant analysis yielded the best performance, with an AUC of 0.849, an accuracy of 0.730, a sensitivity of 0.819, and a specificity of 0.709 in the validation set. CONCLUSIONS: This study indicates that efficient screening using machine learning methods with noninvasive tests can be applied to a large population and achieve the objective of secondary prevention.

Discovery of gas vesicles in Streptomyces sp. CB03234-S and potential effects of gas vesicle gene overexpression on morphological and metabolic changes in streptomycetes.

Gas vesicles are usually found in aquatic microbes to help against adverse growth conditions and also present applied potentials in various biotechnological areas. Although gvp genes encoding gas vesicle proteins are occurred in many bacteria, the formation of gas vesicles in streptomycetes has not been reported yet. During the transcriptome analysis of Streptomyces sp. CB03234-S, a high producer of 10-membered enediynes tiancimycins (TNMs) generated in our previous studies, the notable activation of a gvp gene cluster (gvp3234) was detected, which may be related to the morphological change and yield increase in CB03234-S. Subsequent bioinformatic analysis revealed that gvp3234 possessed some unique features different from other gvp gene clusters, and its relative expression level was much higher than homologues in Streptomyces coelicolor. The overexpression of gvp3234 further improved the total titer of TNM-A and TNM-D from 12.3 ± 0.1 mg/L up to 21.4 ± 0.8 mg/L in CB03234-S, and TEM analysis clearly showed the emergence of gas vesicles, while the heterologous expression of gvp3234 in Streptomyces sp. CB09001 also resulted in the similar formation of fluffy mycelial pellets with long hyphae like in CB03234-S and enhanced the production of xanthone derivative. Our work reported a functionally expressed gvp gene cluster and the corresponding formation of gas vesicles in streptomycetes for the first time. The revealed potential effects of gas vesicles on morphological and metabolic changes shall provide an alternative strategy for strain improvement of industrial streptomycetes to obtain better performances in liquid cultivation and enhance production of secondary metabolites.

Streptomycin-induced ribosome engineering complemented with fermentation optimization for enhanced production of 10-membered enediynes tiancimycin-A and tiancimycin-D.

Tiancimycins (TNMs) are a group of 10-membered anthraquinone-fused enediynes, newly discovered from Streptomyces sp. CB03234. Among them, TNM-A and TNM-D have exhibited excellent antitumor performances and could be exploited as very promising warheads for the development of anticancer antibody-drug conjugates (ADCs). However, their low titers, especially TNM-D, have severely limited following progress. Therefore, the streptomycin-induced ribosome engineering was adopted in this work for strain improvement of CB03234, and a TNMs high producer S. sp. CB03234-S with the K43N mutation at 30S ribosomal protein S12 was successfully screened out. Subsequent media optimization revealed the essential effects of iodide and copper ion on the production of TNMs, while the substitution of nitrogen source could evidently promote the accumulation of TNM-D, and the ratio of produced TNM-A and TNM-D was responsive to the change of carbon and nitrogen ratio in the medium. Further amelioration of the pH control in scaled up 25 L fermentation increased the average titers of TNM-A and TNM-D up to 13.7 ± 0.3 and 19.2 ± 0.4 mg/L, respectively. The achieved over 45-fold titer improvement of TNM-A, and 109-fold total titer improvement of TNM-A and TNM-D enabled the efficient purification of over 200 mg of each target molecule from 25 L fermentation. Our efforts have demonstrated a practical strategy for titer improvement of anthraquinone-fused enediynes and set up a solid base for the pilot scale production and preclinical studies of TNMs to expedite the future development of anticancer ADC drugs.

The discovery and development of microbial bleomycin analogues.

The bleomycins (BLMs) belong to a subfamily of glycopeptide antibiotics and are clinically applied in combination chemotherapy regimens to treat various malignancies. But the therapeutic applications of BLMs are restricted by the accompanied dose-dependent lung toxicity and potential incidence of lung fibrosis. Many efforts have been devoted to develop novel BLM analogues, for seeking of drug leads with improved antitumor activity and/or reduced lung toxicity. The progresses in the biosynthetic studies of BLMs have greatly expedited the process to achieve such goals. This review highlights the discovery and development of microbial BLM analogues in the past two decades, especially those derived from engineered biosynthesis. Moreover, the summarized structure-activity relationship, which is specifically focusing on the sugar moiety, shall shed new insights into the prospective development of BLM analogues.