Progress 08/01/18 to 03/31/20

Outputs
Target Audience: Nothing Reported Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?Manus hired a new Research Scientist for the project. The scientist had just graduated from their Ph.D. program and this is their first industry position post-education. They have been fully trained as an industry scientist at Manus and are now independently leading a research program at Manus. How have the results been disseminated to communities of interest? Nothing Reported What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? Created a novel library of 6,500 E. coli strains expressing never-before-seen combinations of genes to create new-to-nature natural products.We succeeded at synthesizing a collection of 724 terpene synthase genes and integrating them into expression vectors with appropriate expression initiation and termination sequences. Each of the terpene syntheses was cloned under three different promoter strengths and then assembled with three different prenyl diphosphate precursor pathways (GPP, FPP, and GGPP) to create 3 x 3 x 724 unique expression constructs for a total of >6,500 unique strains. Assessed library diversity, quantified unique products and determined compound redundancy and novel structures.Each of the 6,500 strains was grown and assayed in triplicate. The assay consisted of chromatographic separation of the product blend followed by elucidation of each compound structure through comparison to the NIST library and reconstruction of the mass spectrometry (MS) fragmentation pattern. We then developed an automated analysis pipeline to deconvolute the mass spectra and NIST confirmation data and calculate out diversity and uniqueness of the library: 38% of the strains (2,476) demonstrated no production of any new recombinant compounds (beyond the prenyl diphosphate precursor produced by the host strain), indicating that about 1/3rd of randomly-generated gene combinations were not functional. A total of 637 unique compounds were identified. These compounds were confirmed as terpenoid based on their characteristic MS spectra and retention times and were binned into monoterpenes, oxygenated monoterpenes, sesquiterpenes, oxygenated sesquiterpenes, diterpenes, and oxygenated diterpenes. 45%of the compounds (286) were assigned a probable structure based on NIST library matches exceeding 70% confidence. The remaining 341 compounds did not have any relevant matches in the NIST library but clearly had m/z peaks confirming terpenoid structures, demonstrating that more than half of the library were novel products, not previously characterized, and likely new-to-nature natural products that do not exist in any organism. The strains that did produce compounds each made between 1 and 18 unique products, with the mode being 5 unique compounds. 64% of the compounds (408) occurred in more than one strain. These compounds were found in 2 to 3,472 strains, with the mode being 17 strains containing any one particular compound. The remaining 36% of compounds (229) occurred in only one strain. 228 of the 229 compounds were classified as one of the 341 unknown and novel terpenoids created in this research program, confirming that the unbiased combinatorial library approach can generate completely novel chemical structures that have high potential to exhibit novel chemical activity.

Publications

Progress 08/01/18 to 07/31/19

Outputs
Target Audience: Nothing Reported Changes/Problems:While the approach has not fundamentally changed, we have refined our tools and expect to be able to accelerate progress of the upcoming work. However, the overall timeline of the project was significantly delayed due to difficulties in hiring a suitable team member to carry out the work. We had screened many candidates, interviewed six, and made two offers that we both rejected before we finally succeeded in landing the right team member for the project in early 2019. As such, the project took a while to get off the ground - while the PI was able to complete Tasks 1-3 with the assistance of the computational team at Manus, the new hire was required for Tasks 4-6. Thankfully, we have been awarded a no-cost extension to the award until March 31st, 2020. We expect to have completed all technical and commercial work by that point and will be ready to submit a Phase II application in Spring of 2020. What opportunities for training and professional development has the project provided?Manus Bio hired a new Research Scientist to perform the work in this project. The scientist recently graduated from her PhD program and has now transitioned to industrial R&D. As a USDA NIFA SBIR project, the scientist is also directly involved in the LARTA Commercial Assistance Program that we have been working on, and has developed a new knowledge base in commercial development while building market research and business intelligence skills to complement her technical skills. How have the results been disseminated to communities of interest?As a private corporation, our results are communicated internally buttightly regulated and not made public without proper confidentiality agreements in place. However, the nature of the project and the resulting library being produced has been communicated to a recently-signed commercial partner. We are actively pursuing the research and development of a novel pesticide with this commercial partner, and they are eager to screen our library for fungicide, insecticide, and herbicide activities in their proprietary screens as soon as we can provide them with a copy of the completed library. What do you plan to do during the next reporting period to accomplish the goals?We will continue to assemble novel pathway constructs and screen them in our recently engineered lead background. This background provides a 'clean' background for screening any terpenoid pathway, be it C10, C15, C20, C30, C35, or C40 coding. It has been stabilized using a new strain editing technique developed at Manus and is capable of g/L precursor flux to our assembled pathways. This has allowed us to identify very weak activities from some of our pathways that are producing hitherto unknown terpenoid scaffolds - it is unlikely these activities could have been discovered in a typical laboratory strain since the extreme carbon flux through the pathway in question has only been made possible through our engineering efforts at Manus. We have nearly completed a novel two-plasmid combinatorial pathway screening system that will allow us to significantly accelerate the P450 screening stage of the project by dramatically cutting down on the time required to build the semi-combinatorial constructs to decorate the terpenoid scaffolds with oxygen. This screening system will be completed by the time we finish screening all of the terpenoid enzymes. Re terpenoid enzymes, we are completing the analysis of the monoterpenoid (C10) and sesquiterpenoid (C15) constructs and identifying the diversity of products formed. We are currently creating 16 new screening strains that will allow rapid combinatorial assembly of Type II and Type I diterpened (C20) synthases in an 'all by all' design to create the most diversity possible from these enzymes and uncover as much chemical space as can be accessed with these constructs. We will then move on to combine the two-plasmid screen with the established terpenoid constructs to screen the novel P450 strains and complete the strain library creation exercise.

Impacts
What was accomplished under these goals? Goal #1: Create Strain Library This goal was broken down into 6 tasks: Bioinformatics and literature search to identify functionally-diverse cytochrome P450 genes (complete) Optimize sequences for expression and function inE. coli(complete) Commercial gene synthesis (ongoing, Phase 2 of 3 complete) Assemble semi-combinatorial biosynthetic pathways to create known and unknown oxygenated terpenoids (progressing) Express pathways in universal high-titer chassis to maximize chances of identifying function (progressing) Analyze strain library and demonstrate low redundancy / good diversity between strains (progressing) Task 1: Manus has completed the in silico analysis and search for terpene synthases (TPS) and cytochrome P450 oxygenases (P450) that are expected to maximize both (1) the proportion of functional genes in our library and (2) the diversity of compounds generated when these genes are combinatorially expressed in our engineered E. coli strains capable of high-titer terpenoid production. The gene diversity in the library is shown for TPS (Figure 2) and P450s. Task 2: Manus has completed the analysis work and successfully executed the full recoding and optimization of all gene sequences to be deployed in this project. The P450 genes, which encode membrane-associated enzymes, needed to have all their N-terminal sequences completely reworked according to our proprietary engineering process to ensure functional expression in vivo. Task 3: Manus has completed the synthesis of 328 TPS genes. These genes have been sequence verified and confirmed to contain full-length genes with no coding errors. An additional 89 TPS genes and 313 P450 genes are currently in synthesis at Twist Bioscience. Task 4: Manus has assembled the 328 TPS genes in relevant expression vectors. In addition to the critical polyprenyl synthase (PPS) enzyme that creates the precursorsubstrates for TPS enzymes, these expression vectors also contain a single TPS gene for monoterpene, sesquiterpene, or sesterterpene synthases or two TPS (both Types I and II) diterpene synthases. These constructs have been combinatorially assembled into 487 unique biosynthetic pathways that are designed so that each creates one or more terpene compounds. Combinatorial assembly of the remaining TPS genes will continue upon receipt of the remaining genes from Twist Bioscience. All functionally characterized TPS pathways will then be combinatorially assembled with the 313 P450 genes. Tasks 5 and 6: In vivo expression of 112 TPS pathways have been completed and analyzed via GC-MS. Subsequent analysis of this pilot dataset has revealed the need for improved analytical methods to more effectively tease apart the product profiles and conclusively identify unique terpene products produced by the strains. Fortunately, our dedicated Analytical Scientist on the Manus team has been able to revamp and improve our methods in Q1 2019. With these improved methods, we have demonstrated 164 unique terpene compounds and are currently assigning compound IDs through analysis of the MS spectra. Re Goal #2:This goal is being tackled as part of the work in Task #6. We are developing streamlined methods to quantify compounds produced by our strains. In addition to quantification of every product, we areassessing the uniqueness of these products by studying the incidence of each compound across each of the constructed strains. Moreover, we have recently purchased the latest NIST MS spectra library and are assigning tentative compound identifications to each of these products. It is clear that we are producing some terpenoids that are known since the hit scores and spectra match very well with data stored in the NIST library. However, many of the compounds have bad matches with poor scores to the NIST database, indicating that we are producing many compounds that are either unknown in nature or not well studied. We are planning to develop methods that will allow us to identify association relationships between compounds so that we can (1) better assign compound identifications to strains, (2) identify trends in enzymatic reactions with respect to the entourage of compounds created and thus be able to predict superior routes for improved enzymes via protein engineering, and (3) more quickly identify the most promising pathways (and the enzymes that form them) so that we can exploit those genes in further library development work to generate the greatest variety of compounds. We have also engaged the computational design team at Manus with the project and with their support we have generated a novel, high-throughput data analysis pipeline that has greatly accelerated the analysis of our ongoing construct development efforts.

Publications

Progress 08/01/18 to 03/31/19

Outputs
Target Audience: Nothing Reported Changes/Problems:While the approach has not fundamentally changed, we have refined our tools and expect to be able to accelerate progress of the upcoming work. However, the overall timeline of the project was significantly delayed due to difficulties in hiring a suitable team member to carry out the work. We had screened many candidates, interviewed six, and made two offers that we both rejected before we finally succeeded in landing the right team member for the project in early 2019. As such, the project took a while to get off the ground - while the PI was able to complete Tasks 1-3 with the assistance of the computational team at Manus, the new hire was required for Tasks 4-6. Thankfully, we have been awarded a no-cost extension to the award until March 31st, 2020. We expect to have completed all technical and commercial work by that point and will be ready to submit a Phase II application in Spring of 2020. What opportunities for training and professional development has the project provided?Manus Bio hired a new Research Scientist to perform the work in this project. The scientist recently graduated from her PhD program and has now transitioned to industrial R&D. As a USDA NIFA SBIR project, the scientist is also directly involved in the LARTA Commercial Assistance Program that we have been working on, and has developed a new knowledge base in commercial development while building market research and business intelligence skills to complement her technical skills. How have the results been disseminated to communities of interest?As a private corporation, our results are communicated internally buttightly regulated and not made public without proper confidentiality agreements in place. However, the nature of the project and the resulting library being produced has been communicated to a recently-signed commercial partner. We are actively pursuing the research and development of a novel pesticide with this commercial partner, and they are eager to screen our library for fungicide, insecticide, and herbicide activities in their proprietary screens as soon as we can provide them with a copy of the completed library. What do you plan to do during the next reporting period to accomplish the goals?We will continue to assemble novel pathway constructs and screen them in our recently engineered lead background. This background provides a 'clean' background for screening any terpenoid pathway, be it C10, C15, C20, C30, C35, or C40 coding. It has been stabilized using a new strain editing technique developed at Manus and is capable of g/L precursor flux to our assembled pathways. This has allowed us to identify very weak activities from some of our pathways that are producing hitherto unknown terpenoid scaffolds - it is unlikely these activities could have been discovered in a typical laboratory strain since the extreme carbon flux through the pathway in question has only been made possible through our engineering efforts at Manus. We have nearly completed a novel two-plasmid combinatorial pathway screening system that will allow us to significantly accelerate the P450 screening stage of the project by dramatically cutting down on the time required to build the semi-combinatorial constructs to decorate the terpenoid scaffolds with oxygen. This screening system will be completed by the time we finish screening all of the terpenoid enzymes. Re terpenoid enzymes, we are completing the analysis of the monoterpenoid (C10) and sesquiterpenoid (C15) constructs and identifying the diversity of products formed. We are currently creating 16 new screening strains that will allow rapid combinatorial assembly of Type II and Type I diterpened (C20) synthases in an 'all by all' design to create the most diversity possible from these enzymes and uncover as much chemical space as can be accessed with these constructs. We will then move on to combine the two-plasmid screen with the established terpenoid constructs to screen the novel P450 strains and complete the strain library creation exercise.

Impacts
What was accomplished under these goals? Goal #1: Create Strain Library This goal was broken down into 6 tasks: Bioinformatics and literature search to identify functionally-diverse cytochrome P450 genes (complete) Optimize sequences for expression and function inE. coli(complete) Commercial gene synthesis (ongoing, Phase 2 of 3 complete) Assemble semi-combinatorial biosynthetic pathways to create known and unknown oxygenated terpenoids (progressing) Express pathways in universal high-titer chassis to maximize chances of identifying function (progressing) Analyze strain library and demonstrate low redundancy / good diversity between strains (progressing) Task 1: Manus has completed the in silico analysis and search for terpene synthases (TPS) and cytochrome P450 oxygenases (P450) that are expected to maximize both (1) the proportion of functional genes in our library and (2) the diversity of compounds generated when these genes are combinatorially expressed in our engineered E. coli strains capable of high-titer terpenoid production. The gene diversity in the library is shown for TPS (Figure 2) and P450s. Task 2: Manus has completed the analysis work and successfully executed the full recoding and optimization of all gene sequences to be deployed in this project. The P450 genes, which encode membrane-associated enzymes, needed to have all their N-terminal sequences completely reworked according to our proprietary engineering process to ensure functional expression in vivo. Task 3: Manus has completed the synthesis of 328 TPS genes. These genes have been sequence verified and confirmed to contain full-length genes with no coding errors. An additional 89 TPS genes and 313 P450 genes are currently in synthesis at Twist Bioscience. Task 4: Manus has assembled the 328 TPS genes in relevant expression vectors. In addition to the critical polyprenyl synthase (PPS) enzyme that creates the precursorsubstrates for TPS enzymes, these expression vectors also contain a single TPS gene for monoterpene, sesquiterpene, or sesterterpene synthases or two TPS (both Types I and II) diterpene synthases. These constructs have been combinatorially assembled into 487 unique biosynthetic pathways that are designed so that each creates one or more terpene compounds. Combinatorial assembly of the remaining TPS genes will continue upon receipt of the remaining genes from Twist Bioscience. All functionally characterized TPS pathways will then be combinatorially assembled with the 313 P450 genes. Tasks 5 and 6: In vivo expression of 112 TPS pathways have been completed and analyzed via GC-MS. Subsequent analysis of this pilot dataset has revealed the need for improved analytical methods to more effectively tease apart the product profiles and conclusively identify unique terpene products produced by the strains. Fortunately, our dedicated Analytical Scientist on the Manus team has been able to revamp and improve our methods in Q1 2019. With these improved methods, we have demonstrated 164 unique terpene compounds and are currently assigning compound IDs through analysis of the MS spectra. Re Goal #2:This goal is being tackled as part of the work in Task #6. We are developing streamlined methods to quantify compounds produced by our strains. In addition to quantification of every product, we areassessing the uniqueness of these products by studying the incidence of each compound across each of the constructed strains. Moreover, we have recently purchased the latest NIST MS spectra library and are assigning tentative compound identifications to each of these products. It is clear that we are producing some terpenoids that are known since the hit scores and spectra match very well with data stored in the NIST library. However, many of the compounds have bad matches with poor scores to the NIST database, indicating that we are producing many compounds that are either unknown in nature or not well studied. We are planning to develop methods that will allow us to identify association relationships between compounds so that we can (1) better assign compound identifications to strains, (2) identify trends in enzymatic reactions with respect to the entourage of compounds created and thus be able to predict superior routes for improved enzymes via protein engineering, and (3) more quickly identify the most promising pathways (and the enzymes that form them) so that we can exploit those genes in further library development work to generate the greatest variety of compounds. We have also engaged the computational design team at Manus with the project and with their support we have generated a novel, high-throughput data analysis pipeline that has greatly accelerated the analysis of our ongoing construct development efforts.

Publications