RESUMO
Pore-forming toxins (PFTs) are proteins that form lesions in biological membranes. Better understanding of the structure and function of these proteins will be beneficial in a number of biotechnological applications, including the development of new pest control methods in agriculture. When searching for new pore formers, existing sequence homology-based methods fail to discover truly novel proteins with low sequence identity to known proteins. Search methodologies based on protein structures would help us move beyond this limitation. As the number of known structures for PFTs is very limited, it's quite challenging to identify new proteins having similar structures using computational approaches like deep learning. In this article, we therefore propose a sample-efficient graphical model, where a protein structure graph is first constructed according to consensus secondary structures. A semi-Markov conditional random fields model is then developed to perform protein sequence segmentation. We demonstrate that our method is able to distinguish structurally similar proteins even in the absence of sequence similarity (pairwise sequence identity < 0.4)-a feat not achievable by traditional approaches like HMMs. To extract proteins of interest from a genome-wide protein database for further study, we also develop an efficient framework for UniRef50 with 43 million proteins.
Assuntos
Bases de Dados de Proteínas , Proteínas Citotóxicas Formadoras de Poros , Proteínas Citotóxicas Formadoras de Poros/química , Proteínas Citotóxicas Formadoras de Poros/metabolismo , Biologia Computacional/métodos , Modelos Moleculares , Algoritmos , Cadeias de Markov , Sequência de Aminoácidos , Estrutura Secundária de Proteína , Aprendizado ProfundoRESUMO
Many pore-forming proteins originating from pathogenic bacteria are toxic against agricultural pests. They are the key ingredients in several pesticidal products for agricultural use, including transgenic crops. There is an urgent need to identify novel pore-forming proteins to combat development of resistance in pests to existing products, and to develop products that are effective against a broader range of pests. Existing computational methodologies to search for these proteins rely on sequence homology-based approaches. These approaches are based on similarities between protein sequences, and thus are limited in their usefulness for discovering novel proteins. In this paper, we outline a novel deep learning model trained on pore-forming proteins from the public domain. We compare different ways of encoding protein information during training, and contrast it with traditional approaches. We show that our model is capable of identifying known pore formers with no sequence similarity to the proteins used to train the model, and therefore holds promise for identifying novel pore formers.
Assuntos
Aprendizado Profundo , Praguicidas , Proteínas Citotóxicas Formadoras de Poros/análise , Sequência de Aminoácidos , Bactérias/metabolismo , Bactérias/patogenicidade , Produtos Agrícolas , Proteínas Citotóxicas Formadoras de Poros/químicaRESUMO
Plant-parasitic nematodes (PPNs) are economically important pests of agricultural crops, and soybean cyst nematode (SCN) in particular is responsible for a large amount of damage to soybean. The need for new solutions for controlling SCN is becoming increasingly urgent, due to the slow decline in effectiveness of the widely used native soybean resistance derived from genetic line PI 88788. Thus, developing transgenic traits for controlling SCN is of great interest. Here, we report a Bacillus thuringiensis delta-endotoxin, Cry14Ab, that controls SCN in transgenic soybean. Experiments in C. elegans suggest the mechanism by which the protein controls nematodes involves damaging the intestine, similar to the mechanism of Cry proteins used to control insects. Plants expressing Cry14Ab show a significant reduction in cyst numbers compared to control plants 30 days after infestation. Field trials also show a reduction in SCN egg counts compared with control plants, demonstrating that this protein has excellent potential to control PPNs in soybean.
Assuntos
Toxinas de Bacillus thuringiensis/genética , Produtos Agrícolas/parasitologia , Resistência à Doença/genética , Endotoxinas/genética , Glycine max/parasitologia , Proteínas Hemolisinas/genética , Tylenchoidea/patogenicidade , Animais , Bacillus thuringiensis/genética , Toxinas de Bacillus thuringiensis/metabolismo , Bioensaio , Caenorhabditis elegans , Produtos Agrícolas/genética , Produtos Agrícolas/metabolismo , Endotoxinas/metabolismo , Feminino , Engenharia Genética , Proteínas Hemolisinas/metabolismo , Melhoramento Vegetal/métodos , Doenças das Plantas/genética , Doenças das Plantas/parasitologia , Plantas Geneticamente Modificadas/genética , Plantas Geneticamente Modificadas/metabolismo , Plantas Geneticamente Modificadas/parasitologia , Glycine max/genética , Glycine max/metabolismo , Tylenchoidea/isolamento & purificaçãoRESUMO
A modified Vip3C protein has been developed that has a spectrum of activity that has the potential to be commercially useful for pest control, and shows good efficacy against Spodoptera frugiperda in insect bioassays and field trials. For the first time Vip3A and Vip3C proteins have been compared to Cry1 and Cry2 proteins in a complete set of experiments from insect bioassays to competition binding assays to field trials, and the results of these complementary experiments are in agreement with each other. Binding assays with radiolabelled toxins and brush border membrane vesicles from S. frugiperda and Helicoverpa armigera show that the modified Vip3C protein shares binding sites with Vip3A, and does not share sites with Cry1F or Cry2A. In agreement with the resulting binding site model, Vip3A-resistant insects were also cross-resistant to the modified Vip3C protein. Furthermore, maize plants expressing the modified Vip3C protein, but not those expressing Cry1F protein, were protected against Cry1F-resistant S. frugiperda in field trials.
