No abstract

Microbial biosynthesis of plant natural products from simple building blocks is a promising approach toward scalable production and modification of high-value compounds. The pathway for biosynthesis of noscapine, a potential anticancer compound, from canadine was recently elucidated as a 10-gene cluster from opium poppy. Here we demonstrate the de novo production of noscapine in , through the reconstruction of a biosynthetic pathway comprising over 30 enzymes from plants, bacteria, mammals, and yeast itself, including 7 plant endoplasmic reticulum (ER)-localized enzymes. Optimization directed to tuning expression of pathway enzymes, host endogenous metabolic pathways, and fermentation conditions led to an over 18,000-fold improvement from initial noscapine titers to ∼2.2 mg/L. By feeding modified tyrosine derivatives to the optimized noscapine-producing strain we further demonstrated microbial production of halogenated benzylisoquinoline alkaloids. This work highlights the potential for microbial biosynthetic platforms to support the synthesis of valuable and novel alkaloid compounds, which can advance alkaloid-based drug discovery and development.

This paper focuses on structured-output learning using deep neural networks for 3D human pose estimation from monocular images. Our network takes an image and 3D pose as inputs and outputs a score value, which is high when the image-pose pair matches and low otherwise. The network structure consists of a convolutional neural network for image feature extraction, followed by two sub-networks for transforming the image features and pose into a joint embedding. The score function is then the dot-product between the image and pose embeddings. The image-pose embedding and score function are jointly trained using a maximum-margin cost function. Our proposed framework can be interpreted as a special form of structured support vector machines where the joint feature space is discriminatively learned using deep neural networks. We test our framework on the Human3.6m dataset and obtain state-ofthe-art results compared to other recent methods. Finally, we present visualizations of the image-pose embedding space, demonstrating the network has learned a high-level embedding of body-orientation and pose-configuration.

Genetic sensors capable of converting key metabolite levels to fluorescence signals enable the monitoring of intracellular compound concentrations in living cells, and emerge as an efficient tool in high-throughput genetic screening. However, the development of genetic sensors in yeasts lags far behind their development in bacteria. Here we report the design of a malonyl-CoA sensor in Saccharomyces cerevisiae using an adapted bacterial transcription factor FapR and its corresponding operator fapO to gauge intracellular malonyl-CoA levels. By combining this sensor with a genome-wide overexpression library, we identified two novel gene targets that improved intracellular malonyl-CoA concentration. We further utilized the resulting recombinant yeast strain to produce a valuable compound, 3-hydroxypropionic acid, from malonyl-CoA and enhanced its titer by 120%. Such a genetic sensor provides a powerful approach for genome-wide screening and could further improve the synthesis of a large range of chemicals derived from malonyl-CoA in yeast.

Abstract

Saccharomyces cerevisiae is considered one of the most promising organisms for ethanol production from lignocellulosic feedstock. Unfortunately, pentose sugars, which comprise up to 30% of lignocellulose, cannot be utilized by wild type S. cerevisiae. Heterologous pathways were introduced into S. cerevisiae to enable utilization of d-xylose, the most abundant pentose sugar. However, the resulting recombinant S. cerevisiae strains exhibited a slow growth rate and poor sugar utilization efficiency when grown on d-xylose as the sole carbon source. d-xylose uptake is the first step of d-xylose utilization. d-xylose can only enter yeast cells through hexose transporters, which have two orders of magnitude lower affinity towards d-xylose compared to hexoses. It was also shown that inefficient pentose uptake is the limiting step in some d-xylose metabolizing yeast strains. Here we report the cloning and characterization of two novel d-xylose-specific transporters from Neurospora crassa and Pichia stipitis. These two transporters were identified from a total of 18 putative pentose transporters. They were functionally expressed and properly localized in S. cerevisiae as indicated by HPLC analysis and fluorescence confocal microscopy, respectively. Kinetic parameters of the d-xylose-specific transporters were determined using a (14)C-labeled sugar uptake assay. Use of pentose-specific transporters should improve d-xylose consumption and ethanol production in fast d-xylose assimilating strains, thereby lowering the cost of lignocellulosic ethanol production.

Middle East respiratory syndrome coronavirus (MERS-CoV) with pandemic potential is a major worldwide threat to public health. However, vaccine development for this pathogen lags behind as immunity associated with protection is currently largely unknown. In this study, an immunoinformatics-driven genome-wide screening strategy of vaccine targets was performed to thoroughly screen the vital and effective dominant immunogens against MERS-CoV. Conservancy and population coverage analysis of the epitopes were done by the Immune Epitope Database. The results showed that the nucleocapsid (N) protein of MERS-CoV might be a better protective immunogen with high conservancy and potential eliciting both neutralizing antibodies and T-cell responses compared with spike (S) protein. Further, the B-cell, helper T-cell and cytotoxic T lymphocyte (CTL) epitopes were screened and mapped to the N protein. A total of 15 linear and 10 conformal B-cell epitopes that may induce protective neutralizing antibodies were obtained. Additionally, a total of 71 peptides with 9-mer core sequence were identified as helper T-cell epitopes, and 34 peptides were identified as CTL epitopes. Based on the maximum HLA binding alleles, top 10 helper T-cell epitopes and CTL epitopes that may elicit protective cellular immune responses against MERS-CoV were selected as MERS vaccine candidates. Population coverage analysis showed that the putative helper T-cell epitopes and CTL epitopes could cover the vast majority of the population in 15 geographic regions considered where vaccine would be employed. The B- and T-cell stimulation potentials of the screened epitopes is to be further validated for their efficient use as vaccines against MERS-CoV. Collectively, this study provides novel vaccine target candidates and may prompt further development of vaccines against MERS-CoV and other emerging infectious diseases.

The development of high-throughput phenotyping tools is lagging far behind the rapid advances of genotype generation methods. To bridge this gap, we report a new strategy for design, construction, and fine-tuning of intracellular-metabolite-sensing/regulation gene circuits by repurposing bacterial transcription factors and eukaryotic promoters. As proof of concept, we systematically investigated the design and engineering of bacterial repressor-based xylose-sensing/regulation gene circuits in Saccharomyces cerevisiae. We demonstrated that numerous properties, such as induction ratio and dose-response curve, can be fine-tuned at three different nodes, including repressor expression level, operator position, and operator sequence. By applying these gene circuits, we developed a cell sorting based, rapid and robust high-throughput screening method for xylose transporter engineering and obtained a sugar transporter HXT14 mutant with 6.5-fold improvement in xylose transportation capacity. This strategy should be generally applicable and highly useful for evolutionary engineering of proteins, pathways, and genomes in S. cerevisiae.