Team:William and Mary/Contribution
Contribution
Table of Contents
- Parts Page Characterizations
- Guidance to iGEM Teams Interested in Recombination Detection Program, version 4 (RDPv4)
- Framework for Fatty Acid Export
- Framework for Smart Probiotic Design and Modeling
- Guidance to iGEM Teams on Troubleshooting Challenges
- Troubleshooting Circuit Design
- Troubleshooting Mathematical Modeling
- Troubleshooting Proposed Implementation
.
1. Parts Page Characterization
Using information from the literature, we thoroughly characterized multiple parts in the iGEM Registry of Standardized Parts. On the following parts pages, we made sure to characterize function, protein structure, kinetic parameters, substrate preferences/specificities, location within the cell, genetic regulation, homologous proteins, and synthetic biology applications.
FadL
|
|
Endogenous to Escherichia coli, the FadL gene encodes a 33 kDalton outer membrane protein associated with the import of fatty acids. FadL works as one of multiple proteins necessary for the import and export of fatty acids. For example, FadD activates fatty acids by attaching Coenzyme A shortly after import by FadL. The fatty acid imported by FadL can be metabolized by the cell for carbon.
The FadL protein is inserted into the cell membrane by a signal peptide 27 amino acids in length( Black, P. N. 1991) Once embedded in the membrane, FadL binds to extracellular long-chain fatty acids with high affinity.(Liu, H. 2012). FadL transports these fatty acids across the outer membrane into the periplasm (Liu, H. 2012). FadL is necessary for transmembrane movement of fatty acids, which cannot freely move through the E. coli cell membrane; FadL acts in the facilitated transport of free fatty acids into the cell. The spread of this process has been defined in terms of Vmax= 800 pmol/min/mg enzyme and Km= 87.3 µM for 18:1 fatty acids (Black, P. N., 1988, Maloy et al., 1981).
FadL deletion may be combined with the manipulation of other genes for applications related to fatty acid production and export. If FadL is deleted, fatty acid accumulates extracellularly, which may be desirable when engineering bacteria for fatty acid production. For example, in the paper (Liu, H. 2012), the authors found the amplification of TesA and the deletion of fadL in E. coli BL21 (DE3) improved extracellular fatty acid production. They found that, after promoting TesA and deleting FadL, E. coli produced 4.8 g L−1extracellular fatty acid with a production rate of 0.004 gh−1 g−1 dry cell. In another paper, (Shin, K. S et. al 2017), researchers found that the deletion of FadL increased fatty acid production in most cases. They noted that deletion of ompF or FadL individually, without any additional genetic manipulation, resulted in marginal improvements in FFA production (Shin, K. S et. al 2017). The authors did note that with the marginal increase in total free fatty acid increase was accompanied by a substantial increase in the percentage of extracellular free fatty acid. FadL deletion increased the percentage of extracellular fatty acid to 34% of total fatty acid (Shin, K. S et. al 2017). Interestingly, the researchers found that deletion of FadL did not always increase the production of FFA when combined with other gene deletion. For example, when envR, gusC, and mdlA were deleted in addition to FadL total FFA production was reduced by 10% (Shin, K. S et. al 2017). The graph below shows genes knocked out and how they contributed to FFA production.
Additionally, FadL has been shown to have a significant effect on the integrity of the bacterial outer membrane (Tan Z. et al. 2017). In cases where FadL expression was decreased, researchers attributed a subsequent decrease in membrane integrity to a disruption of the lipid biosynthesis pathway, as fatty acid import is the main carbon source for E. coli. Since lipids are the main component of the cell membrane, even slight changes in lipid concentration can greatly decrease membrane integrity. (Tan Z. et al. 2017). In one study, researchers observed that a decrease in FadL expression actually led to a decrease in fatty acid production after a prolonged period of time, due to the disruption of the lipid biosynthesis. The researchers tested using 6 different promoters that expressed varying levels of FadL mRNA to see if there was a correlation between FadL expression and fatty acid titer. After 72 hours, the researchers found a positive relationship between the abundance of FadL mRNA and the overall fatty acid titer(Tan Z. et al. 2017). Although FadL has been shown to increase extracellular fatty acid, it may decrease the overall production of fatty acid after a long period of time. This concern should be kept in mind when using FadL to increase extracellular fatty acid.
References
Black P. N. (1988). The fadL gene product of Escherichia coli is an outer membrane protein required for uptake of long-chain fatty acids and involved in sensitivity to bacteriophage T2. Journal of bacteriology, 170(6), 2850–2854. https://doi.org/10.1128/jb.170.6.2850-2854.1988
Higashitani, A., Nishimura, Y., Hara, H., Aiba, H., Mizuno, T., & Horiuchi, K. (1993). Osmoregulation of the fatty acid receptor gene fadL in Escherichia coli. Molecular and General Genetics MGG,240(3), 339-347. doi:10.1007/bf00280384
Liu, H., Yu, C., Feng, D., Cheng, T., Meng, X., Liu, W., . . . Xian, M. (2012). Production of extracellular fatty acid using engineered Escherichia coli. Microbial Cell Factories,11(1), 41. doi:10.1186/1475-2859-11-41
Maloy, S. R., Ginsburgh, C. L., Simons, R. W., & Nunn, W. D. (1981). Transport of long and medium chain fatty acids by Escherichia coli K12. The Journal of biological chemistry, 256(8), 3735–3742.
Shin, K. S., & Lee, S. K. (2017). Increasing Extracellular Free Fatty Acid Production in Escherichia coli by Disrupting Membrane Transport Systems. Journal of Agricultural and Food Chemistry,65(51), 11243-11250. doi:10.1021/acs.jafc.7b04521.s001
Tan, Z., Black, W., Yoon, J.M. et al. Improving Escherichia coli membrane integrity and fatty acid production by expression tuning of FadL and OmpF. Microb Cell Fact 16, 38 (2017). https://doi.org/10.1186/s12934-017-0650-8
FadD
| Sequence and brief description of function. |
|
Protein Function & Overview
FadD is a soluble fatty acyl CoA synthetase endogenous to Escherichia coli. It is classified as an AMP-forming fatty acid CoA ligase, meaning that it combines fatty acids with Coenzyme A molecules in a reaction that is powered by converting ATP to AMP. FadD activates both medium and long chain fatty acids into fatty acyl CoA thioesters, which are substrates for beta oxidation, phospholipid biosynthesis, and cellular signalling. Beta oxidation is the pathway that degrades fatty acids, which can be regulated by fatty acyl CoA thioesters (Yoo, 2001).
FadD can be found within the cell by the plasma membrane, where it is non-integrally associated. Its activity is enhanced by the membrane lipids or detergents nearby. The protein has a molecular weight of 62,028 Daltons, though it forms a dimer with a molecular weight of about 120,000. It is contained within a 2.2 kilobase fragment of the E. coli genome, as part of a fatty acid degradative regulon along with fadBA, fadE, and fadL, all under control of repressor FadR. Transcription starts 60 base pairs upstream of the translation start. After the translational stop is a GC-rich inverted repeat and a 8T transcriptional terminator. Two FadR operator sites are found at -13 to -29 and -99 to -115. A rare UUG codon at translation initiation may downregulate expression (Black, 1992).
After long chain fatty acyl-CoA are sequestered inside the cell by vectorial thioesterification with FadD, they can bind repressor protein FadR. This binding to FadR causes it to dissociate from operator sites on the fatty acid degradative (Fad) regulon, relieving the repression on the regulon transcription such that Fad genes can be expressed (Yoo, 2001).
Chemical Reactions, Substrate Specificity, and Kinetics
FA + ATP = FA-AMP (needs Mg2+), FA-AMP + CoASH = FA-SCoA. FadD activates fatty acids by converting their carboxyl group into an acyl-CoA thioester, which is a stronger electrophile. Fatty acids enter the FadD active site from the membrane through a narrow channel that faces the inner membrane while ATP enters through a distinct large channel. Binding these molecules causes FadD to undergo ligand-induced conformational changes. The molecules then form an AMP-FA intermediate, which the flexible C-terminal clamps in order to position the intermediate and prevent its escape. CoA, the final substrate, binds to FadD after the fatty acid and ATP. CoA enters the FadD active site via a third channel and attacks the new bond, generating FA-CoA and AMP. Supposedly, when CoA bonds to a long chain fatty acid, AMP is pushed from the active site by the LCFA-CoA product, but this push is less pronounced with MCFA (Ford, 2015).
FadD belongs to a class of adenylate forming enzymes, whose fatty acid tunnel length determines substrate specificity by accommodating a long hydrophobic tail. FadD has broad chain length specificity with a Vmax ranging from 2632 nmol/min/mg protein for C12 to 135 for C6. However, its maximal activity is reserved for fatty acids with carbon numbers ranging between 12 and 18, activating both mono- and poly-unsaturated fatty acids. The thioesters synthesized are destined for degradation or phospholipid incorporation. FadD has lower activity on medium chain fatty acids with 6 to 12 carbons. Downstream beta oxidation enzymes also have poor activity on MCFA (Black, 1992).
Interactions with other Proteins: Repression by FadR and Cleavage by OmpT
The FadD gene contains two FadR binding sites. The first operator (located from -13 to -29) has 12/17 consensus and the second operator (-115 to -98) has 9/17 consensus. FadD’s operator sites have Keq equal to 10-9 M and 10-8 M, compared to Keq equal to 3 x 10-10 M for FadB’s operator site, the strongest binding site for FadR in E. coli. Long chain fatty acyl CoA prevents FadR from having affinity for operator 1, which would otherwise turn off fadD transcription (Black, 1992).
FadD is a substrate for OmpT in vitro, which cleaves the 62 kDa protein into a 42.97 kDa C-terminal fragment and a 19.39 kDa N-terminal fragment. The cleavage site between residues lysine 172 and arginine 173 is a linker domain that connects the ATP and LCFA binding C-terminal to the N-terminal. The outer membrane serine protease OmpT has dibasic residue specificity with its active site on the cell surface (Yoo, 2001).
Cleavage did not affect binding ability and the fragments remained associated afterwards. OmpT cleaving FadD results in FadD’s new Km and Vmax values twice as high, but catalytic efficiency remains similar. Enzymatic activity is retained because the hydrocarbon chain of fatty acids interacts with the 43 kDa fragment and the AMP binding signature motif binds nucleotides in the same 43 kDa fragment. Adjacent to where the hydrocarbon binds, consensus 25 amino acid fatty acyl CoA signature motif is involved in substrate specificity and binding. While all the binding domains are within the 43 kDa C-terminal fragment, it is unstable alone. The N-terminal remains associated because it is required for structural stability (Yoo, 2001).
When the substrate oleate was present, it inhibited the protease OmpT by changing FadD conformation to protect the cleavage site. Adding ATP allowed FadD to reset its conformation, so the site was exposed again. Sensitivity to proteolysis is correlated with increased mobility and flexibility, so changing the conformation of the flexible hinge by binding LCFA or ATP or interacting with detergent can alter the accessibility of the OmpT cleavage site (Yoo, 2001).
Mutations
Mutations clustered on the face of FadD where AMP exits enhance activity by aiding product exit. Growth rate increased with a wider AMP exit channel from the active site, although some of the mutants had decreased activity on long chain fatty acids. FadD mutations that increase the rate of reaction do not enhance affinity for MCFA. ATP binding precedes and enhances fatty acid binding, so mutants that increase activity facilitate AMP exit and ATP entry from the active site. Removing amino acid side chains surrounding the ATP/AMP channel destabilizes the closed conformation. Mutation eases transition to open state by enhancing AMP exit. These mutations affect the structure of the AMP exit channel or interact with the C-terminal domain (Ford, 2015).
Bioengineering
Synthetic biologists have engineered E. coli to be a biocatalyst for the production of a wide variety of potential biofuels from several biomass constituents. When paired with fadE (acyl-CoA dehydrogenase) disruption and tesA (acetyl-CoA thioesterase) overexpression, overexpression of fadD has enhanced the synthesis of fatty alcohols, olefins, and free fatty acids. For all these biofuels, other steps were involved to maximize production: fatty alcohols required accABCD (acetyl-CoA carboxylase) and acrI (acyl-CoA reductase from A. baylyi) co-overexpression, olefins required oleABCD (beta-ketoacyl-ACP synthase from S. maltophilia) overexpression, and FFA required fatty acid synthase (fabH, fabD, fabG, fabF) and accABCD overexpression (Clomburg, 2010) (Colin, 2011).
Related proteins
Examining fadD’s amino acid sequence alongside rat and yeast acyl CoA synthetases reveals a high degree of similarity in the carboxyl end (353-455), which might be an ATP binding domain because it is also shared with firefly luciferase. High similarity to the other acyl CoA synthetases at a central location (200-273) might indicate a fatty acid or Coenzyme A binding domain (Black, 1992).
There is a second acyl-CoA synthetase in E. coli: FadK. FadK has a higher relative rate on medium chain fatty acids than long chain, but lower absolute activity in comparison to FadD. FadK is only expressed under anaerobic conditions. An unidentified H+/LCFA cotransporter might be present in the inner membrane to interact directly with acyl CoA synthetase. (Ford, 2015)
References
Yoo, J. H., Cheng, O. H., & Gerber, G. E. (2001). Determination of the native form of FadD, the Escherichia coli fatty acyl-CoA synthetase, and characterization of limited proteolysis by outer membrane protease OmpT. The Biochemical journal, 360(Pt 3), 699–706. https://doi.org/10.1042/0264-6021:3600699
Ford, T. J., & Way, J. C. (2015). Enhancement of E. coli acyl-CoA synthetase FadD activity on medium chain fatty acids. PeerJ, 3, e1040. https://doi.org/10.7717/peerj.1040
Clomburg, J. M., & Gonzalez, R. (2010). Biofuel production in Escherichia coli: the role of metabolic engineering and synthetic biology. Applied microbiology and biotechnology, 86(2), 419–434. https://doi.org/10.1007/s00253-010-2446-1
Colin, V. L., Rodríguez, A., & Cristóbal, H. A. (2011). The role of synthetic biology in the design of microbial cell factories for biofuel production. Journal of biomedicine & biotechnology, 2011, 601834. https://doi.org/10.1155/2011/601834
FabH
| The FabH gene endogenous to E. coli has not yet been characterized on the registry, but the FabH gene for B. subtilis has. |
|
Introduction
The FabH gene initiates the Fatty acid synthase II (FAS II) cycle. FAS enzymes can produce short chain fatty acids, long chain fatty acids, and complex fatty acids depending on the substrate availability and enzymes involved. FabH is also referred to as β-ketoacyl-ACP synthase III (KAS III). In general terms, the first step of Fatty acid synthesis involves KAS III acting as a condensation enzyme by “forming a carbon-carbon bond between acetyl-CoA and malonyl-ACP”. Subsequent reactions are carried out by enzymes called KAS I and KAS II to form fatty acids (Nofiani-Mahmoud, 2019).
Usage and Biology
In E. coli, the eFabH enzyme active site has a “catalytic triad” which is an arrangement of 3 amino acids-Cys112, His244 and Asn274. Its function first involves the transfer of the acetyl-CoA to cys112 on the KAS III protein to form an acyl-KAS III intermediate via a process called trans-thioesterification. Then, decarboxylation of Malonyl-CoA forms an enolate anion that performs a nucleophilic attack on the acyl-KAS III carbonyl carbon. This forms the acetoacetyl-ACP product. WhileE. coli has a very high specificity for acetyl-CoA and is unable to use branched chain substrates, many bacteria are able to use different types of starting substrates to form fatty acids (Nofiani-Mahmoud, 2019).
Most bacteria, including E. coli, have only one FabH gene. However some, like B. subtilis, have two. B. subtilis has 2 orfs that correspond with the FabH gene in E. coli, bFabH1 and bFabH2 . The purpose of the presence of two FabH genes has not yet been elucidated. While some bacteria can initiate the FAS cycle using straight and branched substrates and are not restricted to just straight chain unsaturated fatty acids, eFabH like eFabH is. eFabH is incapable of using branched chain substrates as an initial substrate and is highly specific towards the simple 2-carbon Acetyl-CoA.
Therefore, it is thought that the composition of fatty acids produced by broadly specific enzymes, like bFabH2, is determined largely by the initial substrate, and not by the specificity of the enzyme. For example, if the initial substrate is branched, then a high amount of branched fatty acids will be produced (Choi, 2000). Acetyl-CoA, propionyl-CoA, and Butyryl-CoA are substrates that generally lead to the production of short chain fatty acids. Isopropyl-CoA, isobutyryl-CoA and methylbutyryl-CoA are substrates that lead to the production of branched chain fatty acids (Nofiani-Mahmoud, 2019).
Synthetic Biology applications
Research showed that deletion of the FabH gene in E. coli heterologously expressing pfa genes from M. marina had three times the docosahexaenoic acid (a PUFA) production than wild type E. coli. Compared to the control strain which produced 2.8 +/- 0.4 mg DHA per liter of culture, the FabH mutant produced 11.2 +/- 1.9 mg DHA per liter of culture. This happened because the foreign PUFA synthase did not have to compete with the natural fatty acid synthesis pathway because FabH was disabled. Additionally, the FabH mutant also has a decreased C16:0 production, 16.7% compared to 36.5% in the control strain. It also had an increased long chain fatty acid production like C18:1n7 production which went from 23.5% in the control strain to 45.2% in the mutant strain. The Folch method was used to isolate the fatty acids from the sample and gas chromatography was used to quantify each fatty acid. The mutant grew slower than the control and had a higher doubling time, 7.2 hours compared to 5.2 hours of the control. All cultures were grown at 15°C (Robles, 2018).
(^Robles, 2018) The graph highlights the three fold increase in DHA (C22:6n3) production in the FabH mutant compared to the wild type strain. The graph is showing %DHA produced relative to total fatty acids.
Overexpressing FabH reduced fatty acid synthesis and increased Malonyl-CoA levels in the cell. The cellular Malonyl-CoA concentration increased by 49%and Went from 2.45 ± 0.02 nmol/mgDCW in the control to 3.78 ± 0.03 nmol/mgDCW in the FabH overexpressing strain. The total fatty acid production decreased in the FabH overexpressing strain. It went from 27.25 ± 0.08mg/L/OD in the control to 22.53 ± 0.02mg/L/OD in the FabH overexpressing strain (Cao, 2016).
Additionally, because the FabH gene of E. coli is unable to use and produce branched chain fatty acids, heterologous expression of FabH genes from organisms that can use branched chain fatty acid substrates in E. coli led to the production of branched chain 17-carbon fatty acids. This emphasizes the importance of the substrate specificity of FabH in determining the characteristics of the end product (Choi, 2000).
References
Choi KH, Heath RJ, Rock CO. beta-ketoacyl-acyl carrier protein synthase III (FabH) is a determining factor in branched-chain fatty acid biosynthesis. J Bacteriol. 2000;182(2):365-370. doi:10.1128/jb.182.2.365-370.2000
Giner-Robles, Laura et al. “fabH deletion increases DHA production in Escherichia coli expressing Pfa genes.” Microbial cell factories vol. 17,1 88. 8 Jun. 2018, doi:10.1186/s12934-018-0928-5
Nofiani R, Philmus B, Nindita Y, Mahmud T. 3-Ketoacyl-ACP synthase (KAS) III homologues and their roles in natural product biosynthesis. Medchemcomm. 2019;10(9):1517-1530. Published 2019 Apr 29. doi:10.1039/c9md00162j
Weijia Cao, Weichao Ma et al. “Improved pinocembrin production in Escherichia coli by engineering fatty acid synthesis”. Journal of Industrial Microbiology & Biotechnology. Volume 43, Pages 557–566(2016).
OprF
|
|
Usage and Biology
The opportunistic pathogen Pseudomonas aeruginosa detects the inflammatory cytokine IFN-γ through the binding of IFN-γ and OprF, thus inducing the expression of rhlI gene and as a result of that, produces a quorum-sensing dependent virulence determinant PA-I lectin (lecA) (Ding et al., 2010; Wu et al., 2005). Researchers Wu et al. tested whether the expression of rhlI is due to specific cytokines. Strain PLL-EGFP/27853, a PA-I-GFP reporter strain, was individually exposed to IL-2, IL-4, IL-6, IL-8,IL-10, IL-12, IFN-γ and TNF-α, and PA-I promoter activity was assessed. Only IFN-γ was found to induce PA-I promoter activity. Then, the completely sequenced Pseudomonas aeruginosa strain PAO-1 was incubated with 200 ng/ml of IL-2, IL-4, IL-8, IL-10, IFN-γ and TNF-α in cell culture media for 4 hours, and PA-I mRNA was measured using a Northern blot. Relative PA-I mRNA levels were found to be triple in the bacteria group exposed to IFN-γ compared to bacteria groups exposed to other cytokines (Figure 1, Wu et al., 2005). Only IFN-γ and C4-HSL (quorum sensing molecule induced by IFN-γ in P. aeruginosa) resulted in a significant increase in PA-I mRNA. After 6 hours of exposure to different concentrations of IFN-γ, P. aeruginosa was found to have around a 30% increase in relative PA-I lectin expression in 0.1 ng/ml concentration, 50% increase for 1 ng/ml, 210% increase for 10 ng/ml, 310% increase for 100 ng/ml, and 320% increase for 1000 ng/ml (Wu et al., 2005). The rhlI gene is induced by IFN-γ, which leads to the production of C4-HSL, a molecule critical to quorum sensing systems and PA-I production. Activation of the quorum sensing system also leads to the production of pyocyanin (PCN) in addition to PA-I. Then, the researchers tested whether IFN-γ binds to a specific bacterial protein (Wu et al., 2005). By using an enzyme-linked immunosorbent assay (ELISA), epifluorescence photomicrographs showed that IFN-γ preferentially binds to membrane fractions of P. aeruginosa. The binding was found to be diminished upon proteinase K treatment, which indicated that IFN-γ binds to bacterial cell membrane proteins. Membrane proteins were then separated and transferred to polyvinylidene difluoride membranes through non-denaturing gel electrophoresis (Wu et al., 2005). The membrane proteins were then hybridized with IFN-γ followed by biotin-labeled antibody to IFN-γ. Results suggested that it is a 35-kD protein. Using the ESI-TRAP LC-MS-MS ion trap, the researchers identified the 35-kD protein to be the P. aeruginosa outer membrane porin OprF (Wu et al., 2005). A mutant strain with a mutated version of OprF failed to cause an increase in PA-I production as a response to IFN-γ, and reconstructing the mutant strain with a plasmid containing the OprF gene reestablished the responsiveness to IFN-γ.
A possible explanation for why opportunistic pathogen P. aeruginosa developed this function is to allow for an increase in virulence at the time of inflammation in the host, which leads to a weaker host immune system. Also, by binding IFN-γ, P. aeruginosa is able to retain cellular function and replicate in the presence of IFN-γ, which is indicated by experiments (Wu et al., 2005). As a result, with a countermeasure against host immune activation, P. aeruginosa can become more effective at infecting its host.
OprF is an outer membrane porin found in Pseudomonas aeruginosa. It is the major outer membrane protein in Pseudomonas aeruginosa and serves as a transbilayer pore (Brinkman et al., 2000) and a structural protein that maintains cell shape under low osmolarity environment (Freulet-Marrière et al., 2000). OprF consists of two domains: a C-terminal periplasmic domain and a N-terminal domain that forms an eight-stranded antiparallel transmembrane beta-barrel with four extracellular loops. However, OprF exists in two conformations: in two-domain conformation and in a single domain conformation with a 16-stranded beta-barrel with eight extracellular loops (Sugawara et al., 2006; Hughes et al., 1992; Gilleland et al., 1995). Loop 5 (AA 198-237 (Rawling model) (Rawling et al., 1995)) of OprF has the capability of binding IFN-γ (Aurand et al., 2016), and loop 6 (AA 257-275 (Rawling model) (Rawling et al., 1995)) reduces this capability (Aurand et al., 2016).
OprF is also the binding site for complement component C3b. Being structurally similar to OprF, some OmpA proteins are also the binding site for complement component C3b (Mishra et al., 2015). The complement system is a conserved immune system prevalent in vertebrates and some invertebrates that once activated, starts a complement cascade which ultimately forms a cylindrical membrane attack complex that kills invading cells. Complement component C3 acts an important role in the complement cascade. C3b is the product of C3 activation and binds to bacterial surfaces. By binding to OprF, C3b triggers downstream effects which enhance complement-mediated killing.
Part Uses
Researchers have used OprF as an anchor for lipase in E. coli (Lee et al., 2005). Pseudomonas fluorescens SIK W1 lipase gene is fused to the truncated oprF gene by using a C-terminal deletion fusion strategy. Lys 164 is a functional fusion site in the oprF gene, and the researchers were able to successfully display functional lipase on the outer membrane of E. coli. Lipase activity is maximized at 37°C, pH 8.0, and remains 80% active at a temperature range of 20°C and 45°C (Lee et al., 2005). At 37°C, whole cell lipase activity remains over 80% in cells incubated in Tris-HCl, isopropyl alcohol, and hexane for over a week, indicating that E. coli cells displaying lipase can act as an effective biocatalyst for industrial production (Lee et al., 2005).
OprF and Synthetic Biology
OprF is an ortholog of E. coli OmpA, sharing 39% sequence identity and 56% similarity in their C-terminal (periplasmic) domains, but only 15% identity in their N-terminal (TM) domains (Khalid et al., 2006). In the study by Aurand and March, when loop 1 (AA 38-54) of OmpA is replaced with loop 5 (AA 198-237(Rawling model)(Rawling et al., 1995)) of OprF, the β-galactosidase gene lacZ (pPALacZ1) controlled by the system has significantly higher expression when IFN-γ reaches 300 pM, indicating that this version of the OmpA-OprF chimeric protein can detect IFN-γ at 300 pM and in turn control gene expression through the phage shock protein (psp) system. The psp system of E. coli is a conserved system that detects extracellular stress that transduces outer membrane stress to receptors on the inner membrane, PspB and PspC, causing PspA to stop inhibition of PspF, which activates the pspA promoter (Joly et al., 2010). Under non-induced situations, PspA negatively regulates the psp operon by binding PspF (Weiner et al., 1991). β-galactosidase activity was measured in the study through the use of a kinetic assay (Ramsay et al., 2011). When induced, the pspA promoter activates β-galactosidase gene lacZ (pPALacZ1) as designed. Even more intriguing is that when loop 1 (AA 38-54) of OmpA is replaced with loop 5 (AA 198-237 Rawling model) of OprF and loop 2 (AA 79-95) of OmpA is replaced with loop 6 (AA 257-275 Rawling model), the protein can also detect TNF-α at around 150 pM and with increased sensitivity for IFN-γ (200pM) (Aurand et al ., 2016). β-galactosidase activity (measured in RFU/min) was found to be increased at those cytokine concentrations. Other constructs like pTCA9 (loop 2 replaced by loop 6 (AA 241-261, Gilleland model) (Gilleland et al., 1995)) and pTCA10 (loop 1 replaced by loop 5 (AA 200-217 Gilleland model) and loop 2 replaced by loop 6 (AA 241-261 Gilleland model)) require higher concentration of IFN-γ or TNF-α. pTCA9 produces a response to 900 pM of IFN-γ, and pTCA10 requires 800 pM of TNF-α and 900 pM of IFN-γ. As a result, pTCA6 (OmpA loop 1 replaced by OprF loop 5 and OmpA loop 2 replaced with OprF loop 6) is the most effective and sensitive chimeric protein for sensing both cytokines. Thus, the OmpA-OprF chimeric protein can act as a cytokine sensor for IFN-γ and TNF-α, which is useful for identifying regions of inflammation in vivo, and inducing certain genes to synthesize certain compounds to reduce inflammation.
Our project for 2020 is to design a “smart” nasal probiotic that constitutively produces arachidonic acid (AA), which has both pro-inflammatory and anti-inflammatory metabolites, and switches to produce docosahexaenoic acid (DHA) a PUFA with anti-inflammatory metabolites, when excessive inflammation occurs in the nasal cavity, as this can be caused by infection by respiratory viruses (such as SARS-CoV-2). In response to the COVID-19 pandemic, developing an antiviral nasopharyngeal probiotic can be a promising approach to reducing the severity of viral infection. One critical part of our probiotic is its ability to sense a high concentration of cytokines in the extracellular environment, and in turn, induce the expression of genes that regulate antiviral polyunsaturated fatty acid production and secretion. For the sensors in our circuits, we have selected one of the OmpA-OprF chimeric proteins designed by Aurand and March.
References
Brinkman, F., Bains, M., & Hancock, R. (2000). The amino terminus of Pseudomonas aeruginosa outer membrane protein OprF forms channels in lipid bilayer membranes: Correlation with a three-dimensional model (vol 182, pg 5251, 2000). Journal Of Bacteriology, 182(23), 6863.
Ding, B., Von Specht, B., & Li, Y. (2010). OprF/I-vaccinated sera inhibit binding of human interferon-gamma to Pseudomonas aeruginosa. Vaccine, 28(25), 4119-4122.
Freulet-Marrière, M., El Hamel, C., Chevalier, S., Dé, E., Molle, G., & Orange, N. (2000). Evidence for association of lipopolysaccharide with Pseudomonas fluorescens strain MF0 porin OprF. Research in Microbiology, 151(10), 873-876.
Gilleland, Harry, Hughes, Eileen, Gilleland, Linda, Matthews-Greer, Janice, & Staczek, John. (1995). Use of synthetic peptides to identify surface-exposed, linear B-cell epitopes within outer membrane protein F of Pseudomonas aeruginosa. Current Microbiology, 31(5), 279-286.
Hughes, E E, Gilleland, L B, & Gilleland, H E, Jr. (1992). Synthetic peptides representing epitopes of outer membrane protein F of Pseudomonas aeruginosa that elicit antibodies reactive with whole cells of heterologous immunotype strains of P. aeruginosa. Infection and Immunity, 60(9), 3497-3497503.
Joly, Nicolas, Engl, Christoph, Jovanovic, Goran, Huvet, Maxime, Toni, Tina, Sheng, Xia, . . . Buck, Martin. (2010). Managing membrane stress: The phage shock protein (Psp) response, from molecular mechanisms to physiology. FEMS Microbiology Reviews, 34(5), 797-827.
Khalid, S., Bond, P., Deol, S., & Sansom, M. (2006). Modeling and simulations of a bacterial outer membrane protein: OprF from Pseudomonas aeruginosa. Proteins: Structure, Function, and Bioinformatics, 63(1), 6-15.
Lee, Seung Hwan, Choi, Jong‐il, Han, Mee‐Jung, Choi, Jong Hyun, & Lee, Sang Yup. (2005). Display of lipase on the cell surface of Escherichia coli using OprF as an anchor and its application to enantioselective resolution in organic solvent. Biotechnology and Bioengineering, 90(2), 223-230.
Mishra, Meenu, Ressler, Adam, Schlesinger, Larry S, & Wozniak, Daniel J. (2015). Identification of OprF as a complement component C3 binding acceptor molecule on the surface of Pseudomonas aeruginosa. Infection and Immunity, 83(8), 3006-3014.
Ramsay JP, Williamson NR, Spring DR, Salmond GPC. 2011. A quorum-sensing molecule acts as a morphogen controlling gas vesicle organelle biogenesis and adaptive flotation in an enterobacterium. Proc Natl Acad Sci 108:14932–14937.
Rawling EG, Martin NL, Hancock RE. 1995. Epitope mapping of the pseudomonas aeruginosa major outer membrane porin protein OprF. Infect Immun 63:38–42.
Sugawara, E., Nestorovich, E., Bezrukov, S., & Nikaido, H. (2006). Pseudomonas aeruginosa porin OprF exists in two different conformations. Journal Of Biological Chemistry, 281(24), 16220-16229.
Weiner, L, Brissette, J L, & Model, P. (1991). Stress-induced expression of the Escherichia coli phage shock protein operon is dependent on sigma 54 and modulated by positive and negative feedback mechanisms. Genes & Development, 5(10), 1912-1923.
Wu, L., Estrada, O., Zaborina, O., Bains, M., Shen, L., Kohler, J., . . . Alverdy, J. (2005). Recognition of host immune activation by Pseudomonas aeruginosa. Science (New York, N.Y.), 309(5735), 774-777.
OmpA
|
|
Usage And Biology
OmpA, an outer membrane protein, consists of 325 residues (1). This protein consists of a N-terminal region with an eight stranded beta-barrel that anchors four loops which extend into the extracellular area (2). A fifteen-nucleotide long region of this protein connects the N-terminal region to the C-terminal region, which is located in the periplasm (3). OmpA has been shown to form dimers and it is hypothesized to exist in a hybrid state between its monomer and dimer forms (3). While there is conclusive evidence of the existence of the OmpA dimer, the function of this dimer is unclear (3, 4). OmpA has been shown to be heat-modifiable and its molecular mass ranges from 28 kDa to 36 kDa based on the temperature of the protein before the weight has been calculated. There are about 100,000 copies of OmpA per cell (2,5). OmpA is most commonly located in gram-negative bacteria and is best characterized in E. coli (5). However, OmpA has homologs in many different species such as P. aeruginosa and C. trachomatis (5).
OmpA transcription has been shown to be dependent on environmental conditions (2). Factors such as nitrogen availability, adhesion to surfaces, and growth rate all impact the transcription of OmpA (6,7,8). For example, when the cell is growing at a higher rate, it maintains a mRNA half-life of about 15 minutes; however, with a slower growth rate, the mRNA half-life decreases to about 4 minutes (9).
OmpA has many functions within E. coli. OmpA assists the cell in holding together the outer membrane and the peptidoglycan layer (1). While there has been documented nucleotide variability within the loop portion of the protein, the beta-barrel region remains consistent across many different E. coli species, hinting at this region's role in structural support (1). Further, these beta-barrels also play an important role in regulating environmental stress. When mutant bacterial strains without OmpA were formed, they showed greater responses to environmental stress and a greater death rate when exposed to acidic environments or osmotic shock compared to bacterial strains containing OmpA (1). In addition to providing E. coli with structural support, OmpA acts as a non-specific porin with a pore size of roughly 1 nm (10).
While OmpA has many functions in E. coli, it plays a large role in the pathogenesis of E. coli. OmpA is simultaneously a way by which pathogenic bacteria can avoid host immune system defenses and become a target of the immune system (2,5). OmpA provides bacterial cells with methods to avoid the immune system. By providing a binding site for protein C4, OmpA is thought to avoid antibody detection (5). Further, OmpA enables E. coli cells to enter macrophages, replicate, and eventually lyse the macrophages (11). While OmpA helps the cell to avoid immune system responses, it is also an identifier of infection and triggers an immune response. For example, the immune system uses an enzyme called neutrophil elastase to destroy pathogenic bacteria (2,5). Normally, neutrophil elastase is able to effectively kill E. coli. However, in mutant bacterial strains without OmpA, neutrophil elastase is no longer effective, suggesting that OmpA acts as a target of the immune system (12). Furthermore, OmpA is thought to bind to a receptor on macrophages, alerting the infected cell to a potential bacterial invasion (13).
Part Uses
Because of its role in pathogenesis, many scientists have considered the use of OmpA in creating a protein-based vaccine (14,15,16). Gu et al. demonstrated that each loop of OmpA produces an immune response and that each loop has specific antibodies that target it (15). Due to the strong immune response against OmpA, vaccines have been developed using this protein (14,15,16). These vaccines have been shown to decrease bacterial load in the lungs and provide higher levels of antigen-specific antibodies, supporting further research into these vaccines (14,15). The OmpA protein is highly conserved across many different bacterial strains (2). However, there are 22 different known polymorphisms in OmpA sequences (17). Despite the existence of these polymorphisms, OmpA vaccines are able to provide protection against many different strains of bacteria (15).
OmpA and Synthetic Biology Applications
While the N-terminal region of OmpA is highly conserved, the four loops exhibit high variability (1). This variability allows all four of the loops of OmpA to be modified in order to express a different protein (18, 19, 20, 21). By replacing the loop regions of OmpA with other DNA sequences, scientists are able to display heterologous proteins (19). Scientists have used the Lpp-OmpA system to help display these proteins (23, 24). This system consists of a fusion between the Lpp protein, the OmpA protein, and an additional protein, which varies based on the end-goal of the system (24). For example, researchers Scott et. al fused the human O6-alkylguanine DNA alkyltransferase protein, also known as SNAP, into this system (25). The SNAP protein is able to respond to the presence of benzylguanine, allowing it to act as a sensor (25). When the Lpp-OmpA-SNAP sensor is exposed to benzylguanine, it causes the surface of the E. coli to glow (25). This chimeric protein has wide applications in synthetic biology as it can act as a reporter gene (25). Another version of the Lpp-OmpA system was developed by Fasehee et al. (23). This version fused LPQTG (a sortase cleaving sequence), metallothionein, and the chitin binding domain (ChBD) to the Lpp-OmpA system. This fusion allowed metallothionein and ChBD to be expressed on the cell surface of E. coli. When srtA was activated, it cleaved the fused protein at the LPQTG site, releasing metallothionein and ChBD. Fasehee et al. were able to find evidence for the cleaving of this enzyme through the use of gel electrophoresis by comparing the results before and after the addition of srtA. Using this method, scientists can easily produce and purify recombinant proteins (23).
While the Lpp-OmpA system has a myriad of applications, OmpA can be used in conjunction with other proteins as well. Aurand and March created chimeric OmpA proteins that are able to sense the presence of inflammatory mediators interferon-gamma (IFN-gamma) and tumor necrosis factor-alpha (TNF-alpha) (26). This protein consists of OmpA, which is naturally found in E. coli, with some of its extracellular loops replaced by OprF, which is endogenous to P. aeruginosa. When creating their chimeric sensors, Aurand and March tested various combinations of OmpA and OprF loops. They found that the chimeric protein with loop 1 of OmpA replaced by loop 5 of OprF was best able to detect the presence of IFN-gamma. When loops 1 and 2 were replaced with loops 5 and 6, respectively, the system was able to detect levels of TNF-a at 150 pM (26).
References
Smith SG, Mahon V, Lambert MA, Fagan RP. A molecular Swiss army knife: OmpA structure, function and expression. FEMS Microbiol Lett. 2007;273(1):1–11.
Ortiz-Suarez, M., Samsudin, F., Piggot, T., Bond, P., and Khalid, S. Full-Length OmpA: Structure, Function, and Membrane Interactions Predicted by MolecularDynamics Simulations. Biophysical Journal. 2016; 111: 1692-1702.
Kaspersen, J., Jessen, C., Vad, B., Sorensen, E., Andersen, K., Glasius, M., Oliveira, C., Otzen, D., Pedersen, J. Low‐Resolution Structures of OmpA⋅DDM Protein–Detergent Complexes. ChemBioChem. 2014; 15(14).
Confer, A., Ayalew, S. The OmpA family of proteins: roles in bacterial pathogenesis and immunity. Vet Microbiol. 2013;163(3-4):207-222.
Baev MV, Baev D, Radek AJ & Campbell JW. Growth of Escherichia coli MG1655 on LB medium: monitoring utilization of amino acids, peptides, and nucleotides with transcriptional microarrays. Appl Microbiol Biotechnol. 2006; 71: 317–322.
Otto K, Norbeck J, Larsson T, Karlsson KA & Hermansson M. Adhesion of type 1-fimbriated Escherichia coli to abiotic surfaces leads to altered composition of outer membrane proteins. J Bacteriol. 2001; 183: 2445–2453.
Lugtenberg B, Peters R, Bernheimer H & Berendsen W. Influence of cultural conditions and mutations on the composition of the outer membrane proteins of Escherichia coli. Mol Gen Genet. 1976;147: 251–262.
Nilsson G, Belasco JG, Cohen SN & von Gabain A. Growth-rate dependent regulation of mRNA stability in Escherichia coli. Nature. 1984;312: 75–77.
Sugawara E & Nikaido H. Pore-forming activity of OmpA protein of Escherichia coli. J Biol Chem. 1992; 267: 2507–2511.
Sukumaran SK, Shimada H & Prasadarao NV. Entry and intracellular replication of Escherichia coli K1 in macrophages require expression of outer membrane protein A. Infect Immun. 2003; 71: 5951–5961.
Belaaouaj A, Kim KS & Shapiro SD. Degradation of outer membrane protein A in Escherichia coli killing by neutrophil elastase. Science 2000; 289: 1185–1188.
Soulas et. al. Cutting Edge: Outer Membrane Protein A (OmpA) Binds to and Activates Human Macrophages. The Journal of Immunology 2000; 165: 2335-2340.
Lei et. al. DNA vaccine encoding OmpA and Pal from Acinetobacter baumannii efficiently protects mice against pulmonary infection, Molecular Biology Reports. 2019; 46: 5397-5408.
Gu et. al. Rational Design and Evaluation of an Artificial Escherichia coli K1 Protein Vaccine Candidate Based on the Structure of OmpA. Frontiers in Cellular and Infection Microbiology. 2018; 8:172.
Lin L, Tan B, Pantapalangkoor P, Ho T, Hujer AM et al. Acinetobacter baumannii rOmpA vaccine dose alters immune polarization and immunodominant epitopes. Vaccine. 2019;31:313–318.
Nielson et. al. Outer membrane protein A (OmpA) of extraintestinal pathogenic Escherichia coli. BMC Research Notes. 2020; 13:51.
Freudl et. al. Cell surface exposure of the outer membrane protein OmpA of Escherichia coli K-12. Journal of Molecular Biology. 1986; 188(3):491-494.
Freudl, R., Insertion of peptides into cell-surface-exposed areas of the Escherichia coli OmpA protein does not interfere with export and membrane assembly. Gene. 1989; 82(2): 229-236.
Pistor, S., Hobom, G. Expression of viral hemagglutinin on the surface of E. coli. Klin Wochenschr 1998; 66: 110–116.
Schor et. al. Surface expression of malarial antigens in Salmonella typhimurium: induction of serum antibody response upon oral vaccination of mice. Vaccine. 1991; 9(9):675-681
Samuelson et. al. Display of proteins on bacteria. Journal of Biotechnology. 2002; 96:129-154.
Fasehee et. al. Engineering E. coli cell surface in order to develop a one-step purification method for recombinant proteins. AMB Express. 2018; 8:107
Georgiou et. al. Display of pMactamase on the Escherichia coli surface: outer membrane phenotypes conferred by Lpp'-OmpA'-pMactamase fusions. Protein Engineering. 1996; 9(2):239-247.
Scott, Felicia Y. (2015). Surface Displayed SNAP as a New Reporter in Synthetic Biology [Master of Science In Biological Systems Engineering, Virginia Polytechnic Institute].
Aurand and March. Development of a Synthetic Receptor Protein for Sensing Inflammatory Mediators Interferon-g and Tumor Necrosis Factor-a. Biotechnology and Bioengineering. 2015; 113(3): 492-500.
.
2. Guidance to iGEM Teams Interested in Recombination Detection Program, version 4 (RDPv4)
Recombination Detection Program (RDP) is a great tool for identifying possible recombination breakpoints, but the word possible should not be forgotten. Here are a few tactics our team recommends that can be employed to ensure that RDP is used without overestimating the number of recombination breakpoints.
In regards to alignments we recommend that researchers try multiple alignment programs before choosing which one to do the final analysis; the one which gives the least number of recombination breakpoints should be chosen. The RDP handbook notes that errors in alignments can lead to the false identification of recombination breakpoints (Martin 2015) thus whichever alignment gives the fewest recombination breakpoints is likely to have the smallest number of false positives due to alignment errors. Additionally, if there are areas of poor alignment, in that there is relatively little nucleotide overlap between strands, this should be kept in mind when viewing the results of RDP. Each recombinant region should be cross-referenced against the percent nucleotide similarity, if there is low similarity this should call into question the validity of the identified recombination event.
To maximize the performance of RDP simulations should be used. A simulated data set should be made that is as similar as possible to the genome which is to be analyzed. This dataset should evolve from a sample viral strand chosen from the desired genome and should be evolved using parameters that are sourced from the literature for mutation rate and recombination rate. Additionally, the simulated genome can be compared to the original in terms of nucleotide similarity among the strands of each respective dataset. The simulated dataset should then be run through RDP where the false positive and false negative rates must be noted. If RDP does not perform well on the simulated dataset it is a good sign that high precaution should be taken when considering the results of RDP on the real dataset. Additionally, if the simulated dataset performs well on RDP then parameters inside of RDP can be adjusted using the simulated dataset to maximize the performance.
If the exact location of breakpoints is not of dire concern other methods can be used in attempting to find general regions along the genome which are common for recombination. In this case, the outputs of multiple alignment programs can be compared and only regions in which there is a common identification of recombination breakpoints should be considered. In addition to using multiple alignment programs, multiple recombination identification programs can be used. This would look something similar to what the researchers did in the paper Molecular evolution of Zika virus during its emergence in the 20(th) century (Faye et al. 2014) where they compared the results of both RDP and Rec-HMM.
In the case where the absolute breakpoint location is of high importance, future researchers may combine the methods from the papers Characterization of New Recombinant Forms of HIV-1 From the Comunitat Valenciana (Spain) by Phylogenetic Incongruence (Bemund et al. 2019) and Evidence of ancient papillomavirus recombination (Varsni et. al 2006). That is researchers can isolate the parental strands and the daughter strands align them independently then rerun RDP to see if the same recombinant region is identified (Varsni et. al 2006). Then phylogenetic testing can be performed on the three strands in isolation to see if there is movement across the phylogenetic tree (Bemund et al. 2019). This method is extremely useful because both of these ensure that errors contributed by alignments are vastly decreased since they are aligned in isolation. It should be noted that both of the methods independently vastly decrease the number of recombination breakpoints and likely underestimate the total number of recombination events.
References
Faye O, Freire CC, Iamarino A, et al. Molecular evolution of Zika virus during its emergence in the 20(th) century. PLoS Negl Trop Dis. 2014;8(1):e2636. Published 2014 Jan 9. doi:10.1371/journal.pntd.0002636
Martin DP, Murrell B, Golden M, Khoosal A, & Muhire B (2015) RDP4: Detection and analysis of recombination patterns in virus genomes. Virus Evolution 1: vev003 doi: 10.1093/ve/vev003
Varsani A, van der Walt E, Heath L, Rybicki EP, Williamson AL, Martin DP. Evidence of ancient papillomavirus recombination. J Gen Virol. 2006 Sep;87(Pt 9):2527-2531. doi: 10.1099/vir.0.81917-0. PMID: 16894190.
.
3. Framework for Fatty Acid Export
While PUFA synthases have been well-characterized, the export of PUFAs following their synthesis within a bacterial cell has yet to be described. We have designed and modeled modular systems for the export of various PUFAs. Simply swapping PUFA synthases in our phospholipase-based systems could allow for the export of different PUFAs, including AA, DHA, eicosapentaenoic acid (EPA), and linoleic acid (LA). Our FarE efflux pump-based system would allow for the export of AA or LA, depending on the chosen PUFA synthase. We hope that these systems will be helpful to researchers, including future iGEM teams, who wish to produce and export PUFAs. Both systems have applications far beyond the antiviral therapy proposed here. PUFA-secreting probiotics could be utilized to modulate inflammation related to depression, arthritis, and heart conditions. Furthermore, export systems streamline the process of isolating and purifying synthesized PUFAs. These systems could be utilized for the creation of PUFA-based supplements, reducing the need to harvest PUFAs from organisms such as salmon and krill.
For more information on the designs of our PUFA synthesis and export systems, please visit this page.
For more information on the modeling of our systems, please visit this page.
.
4. Framework for Smart Probiotic Design & Modeling
As a novel smart drug, TheraPUFA can respond to varying degrees of inflammation and stages of infection, unlike standard pharmaceuticals such as Dexamethasone, Remdesivir, and Hydroxychloroquine. Our design provides a framework for smart, inflammation probiotics, and could guide future iGEM teams pursuing therapeutic projects. Please visit our Design page for more information.
In addition to design, TheraPUFA is unique in its modeling. Our model extends pre-existing probiotic models by incorporating transcriptional stochasticity and by utilizing a grid to account for spatial heterogeneity within the nasal cavity. Future iGEM teams interested in modeling the effect of a probiotic in a spatially-complex manner could benefit from our equations and code, all of which have been uploaded along with our wiki. Please visit our Model page for more information.
.
5. Guidance to iGEM Teams on Troubleshooting Challenges
Troubleshooting a project begins long before the construction of mathematical models or the execution of wetlab experiments. Simply choosing a research topic can be difficult--the creative process is constrained by resource limitations, time constraints, and the amount of pre-existing literature available. In early spring, the COVID-19 crisis eliminated any opportunity for substantial wetlab experiments by our team. The pandemic also inspired us to pursue an antiviral-focused project.
The timeline below provides a chronological account of obstacles we faced and how we addressed them, divided by categories of circuit design, circuit modeling, and proposed implementation & safety. This troubleshooting process required repeated cycles of broadening and refocusing our research topic. At each step, we assessed the design, safety, and ethical implications of our proposed technology. We reviewed literature and consulted experts to evaluate the feasibility of our design and the validity of our models.
A. Troubleshooting Circuit Design
-
Obstacle: Wetlab opportunities are no longer feasible.
Solution: Prior to the COVID-19 pandemic, the team considered a diverse array of project ideas, including non-hormonal male contraceptives, artificial milk and meat, and improved diagnostic tools for Lyme disease. The COVID-19 pandemic then eliminated the possibility of a wetlab-based project, and the team pivoted to a modeling-based project. Despite the lack of wetlab opportunities, we realized we could still conduct significant computational work with global implications. Additionally, we could pursue intensive literature review and circuit design, rigorous human practices, and engaging education projects.
-
Obstacle: Though previous iGEM teams at our university had focused on the expression of proteins, protein/enzyme-related antiviral therapies may be immunogenic or inefficient to deliver.
Solution: We adopted a lipid approach, as lipids play a fundamental yet relatively understudied role in viral infection.
-
Obstacle: Initial therapeutics ideas were deemed unfeasible. The team investigated 25 hydroxycholesterol (25HC) as a possible broad-spectrum antiviral. However, 25HC can already be produced rather efficiently through standard chemical methods, without the need for synthetic biology. The biosynthetic pathway in mycobacteria is somewhat complicated, and 25HC may have adverse effects in some cases (Gold et al., 2014).
Solution: Continued research into the role of lipids in viral infection and the immune response.
-
Obstacle: The team encountered conflicting research articles. While we researched lipids related to viral infection, we identified docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA) as potential inhibitors of the human phospholipase cPLA2 (Vincenti et al, 2015, Clinical Nutrition; Tajuddin et al., 2014, PlosOne; Shikano, Masuzawa, & Yazawa, 1993, J. Immunol.; ). cPLA2 can enable membrane rearrangement, a common strategy of positive-strand RNA viruses. Studies that utilized chemical inhibitors of cPLA2 in cell lines demonstrated inhibition of coronavirus (HCoV-229, MERS) replication and inhibition of viral membrane rearrangement (Muller et al., 2018, J. Virol.; Kishida et al., 1998, BBA; Su, 2009, Neuro-Signals). However, the relationship between DHA and cPLA2 was unclear from our literature review. Some studies show no effect on cPLA2, or even increased activation of cPLA2 by DHA (Su et al., 2018, Prog Neuropsychopharmacol Biol, Liu et al., 2014, Immunology).
Solution: The team consulted experts and widened our literature search. After consulting Dr. James Shelhamer on the relationship between DHA and cPLA2, we broadened our research topic to investigate the general antiviral and antimicrobial properties of various polyunsaturated fatty acids (PUFAs), including arachidonic acid (AA) and linoleic acid (LA) in addition to DHA and EPA. PUFAs have multiple antiviral effects, potentially capable of leaking viral envelopes and suppressing membrane rearrangement by positive-strand RNA viruses (Kohn, Gitelman, & Inbar, 1980, Arch Virol; Das, Arch. Med. Res., 2020; Yan et al., Viruses, 2019; Yan et al., Int. J. Mol. Sci., 2019). The ability of PUFAs to target multiple aspects of viral infection make them an attractive option for antiviral drugs. When we would later interview ConvaLife drug developer Dr. Xiaokun, who is in the process of developing antivirals effective against SARS-CoV-2, he stated that “most ideal antiviral-drugs, such as an AIDS cocktail, should work on multiple aspects [of viral infection].”
-
Obstacle: Desirable genetic parts were not feasible for therapeutic applications. The team conducted a literature review and identified various PUFA synthases from marine bacteria capable of producing DHA and EPA, even when expressed heterologously in species such as E. coli. However, these marine bacteria inhabit cold environments, and their PUFA synthases produce little DHA or EPA at temperatures above 20C (Yoshida et al., Marine Drugs, 2016; Amiri-Jami et al., FEMS Microbiology Letters, 2015). This temperature sensitivity is unsuitable for a probiotic that must inhabit the human body.
Solution: We researched similar genetic parts, such as PUFA synthases from the protist algae Schizochytrium, as well as PUFA synthases from bacteria that inhabit warmer environments (Yoshida et al., Marine Drugs, 2016). For example, AA-producing bacterium Aurespira marina was isolated from the coastline of Thailand (Hosoya, Int. J. Syst. Evol. Microbiol., 2006).
-
Obstacle: Though we identified a mechanism to synthesize our compound of interest, we struggled to find export mechanisms in the literature. Many papers demonstrate synthesis of PUFAs, but it is unclear whether these PUFAs are exported, incorporated into the bacterial cell membrane, or metabolized. We could not find available literature documenting the export of PUFAs after synthesis within bacterial cells.
Solution: Developing our export mechanisms required us to investigate the export of similar molecules, as well as adopt systems from seemingly unrelated “applications” or purposes. For example, we adopted strategies for similar molecules such as long-chain saturated and monounsaturated fatty acids (LCFA) (Tong et al., BMC, 2019). We also investigated efflux pumps, which bacteria utilize to protect themselves from antimicrobial substances such as PUFAs. We contacted Dr. Tohru Dairi regarding the offloading of PUFAs from PUFA synthases, and Dr. Jeong Lee regarding the export of LCFA from Rhodobacter sphaeroides (Hayashi et al., ACS Chemical Biology, 2020; Tong et al., BMC, 2019).
-
Obstacle: Recently published literature challenged an assumption in our project. Though one paper we had read documented the use of LA and AA to inhibit HCoV-229 and MERS, a BioRxiv preprint brought to our attention by Dr. Shelhamer described the existence of LA-binding pockets within SARS-CoV-2 (Toelzer et al., 2020). It is unclear whether exogenous supplement of LA would help alleviate SARS-CoV-2 infection (by replenishing scavenged LA) or would exacerbate it (by facilitating the binding of the virus to cells).
Solution: As we awaited updates from the research community, we decided to exclude LA from our project.
-
Obstacle: The team considered adverse effects of different PUFAs by conducting a literature review on their metabolites. AA has predominantly pro-inflammatory metabolites at the start of infection, while DHA and EPA contribute to anti-inflammatory metabolites. During an interview, Dr. Shelhamer described possible immunosuppressive effects of anti-inflammatory compounds, causing us to research adverse effects of DHA specifically.
Solution: After consulting an expert in pulmonology and on inflammation, the team realized the need to allow for necessary inflammation but curb excess inflammation. To strike this balance, the team decided that the probiotic must secrete AA to allow for some inflammation, then switch to DHA production to resolve excess inflammation. Such a “switch” requires the probiotic to be smart and responsive to its environment.
-
Obstacle: The team researchers sensors that would allow the probiotic to sense its environment to initiate AA production, then later switch to DHA. Though we initially desired an IL-6 and an IL-10 biosensor protein, we could not find such proteins in the literature.
Solution: In the absence of a suitable IL-6 sensor to differentiate basal inflammation from excess inflammation, we repurposed a sensor we had identified in the literature, an IFN-gamma and TNF-alpha sensor (Aurand & March, Biotechnol. & Bioeng., 2016). By utilizing this sensor as a “high pass filter”, we created a mechanism to differentiate different degrees of inflammation.
-
Obstacle: In Tong et al.’s system for fatty acid export, a phospholipase cleaves fatty acids from the inner membrane of a gram negative bacterium, and these free fatty acids exit into the extracellular environment. Tong et al. concluded that exit from the periplasm through the outer membrane is Tol-C independent, but otherwise could not determine how this transmembrane movement occurred (BMC, 2019). The outer membrane of gram negative bacteria could act as a barrier for PUFA export once PUFAs are released from the inner membrane by a phospholipase.
Solution: We decided to adapt Tong et al.’s system for gram positive chasses, such that there is no outer membrane to prevent free fatty acid export after release from the inner membrane. Additionally, utilizing a gram positive chasses would allow us to use Lactobacillus species in our probiotic. Lactobacilli have been implemented safely in probiotics before (De Boeck et al., 2020), unlike gram negative species native to the nasal microbiome. Dr. Turner from UVA’s Medical Center warned us against utilizing even BSL1 gram negative Neisseria and Corynebacteria species due to their relationship to disease. Consistent with the literature, Dr. Turner advocated for Lactobacillus species.
-
Obstacle: Adapting Tong et al.'s system for fatty acid export to a gram positive chassis proves difficult. Without an outer membrane, it is unclear how to stably localize a phospholipase to the gram positive “periplasm.”
Solution: To prevent loss to the extracellular supernatant, we research the porosity and charge of the gram positive cell wall (Forster & Marquis, Molecular Microbiology, 2012; van Wely et al., FEMS Microbiology Reviews, 2001; Stephenson, Biochem. J, 2000). We identify phospholipases that are too bulky or charged to pass through the cell wall.
-
Obstacle: Again, adapting Tong et al.’s system for fatty acid export to a gram positive chassis becomes difficult. The use of a phospholipase to release fatty acid into the extracellular environment requires that these fatty acids first be incorporated into membrane phospholipids. Gram positive cells utilize a different system for phospholipid synthesis than E. coli and other gram negative cells.
Solution: To target fatty acids for phospholipid synthesis in gram positive cells, we research fatty acid kinase FakB3 from S. pneumoniae (Gullet et al., J. Biol. Chem., 2019).
B. Troubleshooting Modeling
- Obstacle: The team began the search for parameters necessary to model the synthesis and export of PUFAs. While the papers Hauvermale et al. (2006) and Metz et al. (2009) quantified DHA production in E. coli heterologously expressing Schizochytrium genes, they provided relative rather than absolute measurements (such as percentage of PUFA within fatty acid production, ratio of DHA to DPA) (Hauvermale et al., Lipids, 2006; Metz et al., Plant Physiol. Biochem., 2009).
Solution: To estimate molecules of DHA produced, we applied these ratios and percentages to the typical fatty acid production of E. coli. This required us to identify additional parameters regarding the fatty acid composition of E. coli.
- Obstacle: Modeling the system of fatty acid export described in Tong et al. becomes complicated, due to the lack of parameters characterizing the exit of fatty acids from the periplasm. Besides independence from TolC, it is unclear how fatty acids exit into the periplasm (Tong et al., BMC, 2019).
Solution: We assume diffusion (facilitated or otherwise) to continue modeling, but begin research into adopting Tong et al.’s system for a gram positive bacteria, as described in the previous troubleshooting section. Gram positive bacteria lack an outer membrane, meaning that fatty acids would not encounter a second membrane after their release into the “periplasm” by a phospholipase. Throughout the course of the project, we would continue to encounter issues with lack of parameters for ODEs according to Michaelis Menten kinetics. To overcome this obstacle, we utilized parameters from similar species, or collapsed steps in order to consider fewer parameters.
- Obstacle: Start the intrahost model to predict the antiviral effect of probiotic. While there is existing literature about modeling viral dynamics and antiviral drug effects, none of them have tackled things like lipids or probiotics.
Solution: In the literature, we found that if a specific drug had prohibitory effects on both viral replication and viral infection, both aspects were considered separately (Besti 2018, Royal, 2020). For each aspect, the effect was expressed through the equation concentration of drug/(concentration of drug + EC50). We adopted that and let the pufa be the drug here. Biologically pufa is able to prohibit both viral infection and replication. However, after a thorough literature review, we were only able to find the prohibitory effect (EC50) of AA on viral replication (Yan 2019), so we utilized that parameter to represent its overall antiviral effect. For DHA we were not able to find a paper that directly quantifies its activity, but we found its antiviral effect to be very similar to AA but slightly lower (Leu 2004). We adopted the EC50 for AA and increased it by 25% as our best estimate.
- Obstacle: Most literature for modeling viral dynamics used ODE, which assumes homogeneous distribution of all variables. In our design the distribution of viral particles and probiotics matter, so we need to upgrade the model to take spatial aspects into account. In the literature this had been tackled in two ways: PDE and agent-based model.
C. Troubleshooting Proposed Implementation
Our interviews with medical professionals and probiotic and drug experts allowed us to address many of the questions that arose regarding the implementation of our proposed probiotic. Please see our human practice page and our proposed implementation and safety pages .
-
Obstacle: The team identified multiple BSL1 species native to the nasal microbiome for use in the probiotic, but desired additional guidance on choosing the most effective and safe strains.
Solution: Probiotic expert Lydia Mapstone encouraged us to consider the usage of multiple species rather than just one, stating that these could have a greater impact. To determine the safety of these species, Dr. Anders Cervin suggested whole genome searches for concerning sequences. While some medical doctors we interviewed stated that sufficient data could demonstrate the safety data of BSL1 gram negative strains such as Neisseria cinerea, Dr. Turner expressed concerns regarding these strains, due to their relationship with disease. Following their guidance, we considered gram negative strains cautiously, investigating ways to adapt our genetic circuits for use in gram positive strains.
-
Obstacle: The team could not decide between a spray and swab application of the nasal probiotic. We also desired additional guidance on the frequency and timing of probiotic application.
Solution: Multiple medical doctors advocated for a spray, though Dr. Shelhamer expressed concern that a spray could deposit the probiotic too far down within the nasopharyngeal microbiome. After reviewing all the interviews, we decided upon a nasal spray. During interviews, Dr. Cervin and Dr. Mapstone promoted daily application of our probiotic. After consulting results of our mathematical model, we decided upon a nasal spray. Additionally, literature such as Ferguson et al. led us to consider prophylactic application of our probiotic for pandemic containment (Nature, 2005).
-
Obstacle: The team desired additional guidance on regulations that would govern the drug development process, if researchers were able to eventually implement our designs. Additionally, we aimed to assess whether current regulations were sufficient.
Solution: From our interviews with Ms. Lydia Mapstone, Dr. Turner, and an additional M.D., we determined that our probiotic would be regulated as a drug. Though this is the most expensive route for regulation, and more intensive than the route for oral probiotics, our nasal probiotic could not pass as a food supplement. Through this process, we would need to acquire “water-proof” data to affirm the safety of our product. When discussing the regulations he was subject to, drug developer Dr. Xiaokun mentioned strict, standardized procedures he followed prior to clinical trials to demonstrate drug safety and effectiveness. He has experienced incredible success with clinical trials, and has completed phases I and II.
In a conversation with Ms. Gale, regulatory affairs expert at plasma and vaccine manufacturer Emergent Biosolutions, the team asked whether the current regulations are sufficient, especially with regard to Operation Warpspeed. Ms. Gale affirmed her confidence in the current regulations, explaining that Operation Warpspeed expedites vaccine availability due to the concerted efforts of private and public researchers, rather than any “shortcuts” in manufacturing practices. -
Obstacle: The team wondered how we as scientists could contribute to the accessibility of our proposed drug, should it ever be implemented, since pricing is often not determined by the researchers themselves.
Solution: Ms. Mapstone suggested that researchers develop contracts with the regulatory bodies of their respective countries to ensure affordable prices. In countries without such a contract between producers and the government, such as in the United States, her proposed breast milk probiotic could sell for 10x as much. Ms. Mapstone, like other interviewees, also suggested reducing production costs in order to reduce final price. For example, she has considered strains that would improve the efficiency of her probiotic. “One strain could do the work of three,” she explained. Similarly, Dr. Xiaokun stated that the simple production procedures for his designed drug, along with its small molecular weight, would reduce its manufacturing costs. Despite best attempts to reduce manufacturing costs, accessibility issues may still pervade during the pandemic, when there is an urgent, overwhelming need for therapeutics. Ms. Gale suggested that distributors first ensure that priority is placed upon front-line medical workers like ICU nurse K.M, who are at particular risk of exposure. Additionally, though it is difficult to determine which patients will become the most sick, Ms. Gale suggested that priority be placed upon patients at particularly high-risk. Ms. Gale suggests that manufacturers engage in cross-functional discussions with key opinion leaders and government agencies to determine the best distribution strategy.
Finally, despite our efforts to ensure equal accessibility, some patients would be excluded from our probiotic. From our interviews with medical experts, we determined that elderly patients and immunocompromised patients should be excluded from our probiotic due to safety concerns. -
Obstacle: The team sought to determine how to demonstrate the safety of our probiotic, and how to make the probiotic appealing to patients.
Solution: Our interviewees advised us regarding surveys and assays to assess the safety of the probiotic, including SNOT surveys and cytotoxicity assays. Strong safety data, including whole genome analysis of potential chasses, could allow for the implementation of species previously unused in probiotics, such as Neisseria cinerea.
All of our interviewed medical experts stated that they believe their patients would be open to a nasal probiotic, even a genetically-engineered one. "They are always looking for therapeutic options," Dr. Mikita said. Generally speaking, Ms. Mapstone believes that the general public is becoming more open to probiotics, perhaps due to the popularity of foods such as kombucha and yogurts. -
Obstacle: Simulations and literature review demonstrated that mucociliary clearance would pose a major obstacle for the retention of our nasal probiotic in the nasopharynx.
Solution: Dr. Alan Shikani suggested the use of a mucoadhesive poloxamer gel that is liquid when stored in a refrigerator, but that could solidify once heated to body temperature following administration.
References
Aurand, T. C., & March, J. C. (2016). Development of a synthetic receptor protein for sensing inflammatory mediators interferon-γ and tumor necrosis factor-α. Biotechnology and bioengineering, 113(3), 492–500. https://doi.org/10.1002/bit.25832
Das U. N. (2020). Can Bioactive Lipids Inactivate Coronavirus (COVID-19)?. Archives of medical research, 51(3), 282–286. https://doi.org/10.1016/j.arcmed.2020.03.004
Das U. N. (2020). Response to: Bioactive Lipids and Coronavirus (COVID-19)-further Discussion. Archives of medical research, 51(5), 445–449. https://doi.org/10.1016/j.arcmed.2020.04.004
Ferguson, N. M., Cummings, D. A., Cauchemez, S., Fraser, C., Riley, S., Meeyai, A., Iamsirithaworn, S., & Burke, D. S. (2005). Strategies for containing an emerging influenza pandemic in Southeast Asia. Nature, 437(7056), 209–214. https://doi.org/10.1038/nature04017
Forster, B. M., & Marquis, H. (2012). Protein transport across the cell wall of monoderm Gram-positive bacteria. Molecular microbiology, 84(3), 405–413. https://doi.org/10.1111/j.1365-2958.2012.08040.x
Gullett, J. M., Cuypers, M. G., Frank, M. W., White, S. W., & Rock, C. O. (2019). A fatty acid-binding protein of Streptococcus pneumoniae facilitates the acquisition of host polyunsaturated fatty acids. The Journal of biological chemistry, 294(44), 16416–16428. https://doi.org/10.1074/jbc.RA119.010659
Hauvermale, A., Kuner, J., Rosenzweig, B., Guerra, D., Diltz, S., & Metz, J. G. (2006). Fatty acid production in Schizochytrium sp.: Involvement of a polyunsaturated fatty acid synthase and a type I fatty acid synthase. Lipids, 41(8), 739–747. https://doi.org/10.1007/s11745-006-5025-6
Hayashi, S., Ogasawara, Y., Satoh, Y., Maruyama, C., Hamano, Y., & Dairi, T. (2020). Off-Loading Mechanism of Products in Polyunsaturated Fatty Acid Synthases. ACS Chemical Biology, 15(3), 651-656. https://doi.org/10.1021/acschembio.0c00075
Hosoya, S., Arunpairojana, V., Suwannachart, C., Kanjana-Opas, A., & Yokota, A. (2006). Aureispira marina gen. nov., sp. nov., a gliding, arachidonic acid-containing bacterium isolated from the southern coastline of Thailand. International journal of systematic and evolutionary microbiology, 56(Pt 12), 2931–2935. https://doi.org/10.1099/ijs.0.64504-0
Kishida, E., Yano, M., Kasahara, M., & Masuzawa, Y. (1998). Distinctive inhibitory activity of docosahexaenoic acid against sphingosine-induced apoptosis. Biochimica et biophysica acta, 1391(3), 401–408. https://doi.org/10.1016/s0005-2760(98)00015-0 Show less
Kohn A, Gitelman J, Inbar M. (1980) Unsaturated free fatty acids inactivate animal enveloped viruses. Arch Virol. 66(4):301-307. doi:10.1007/BF01320626
Liu, Y., Chen, L. Y., Sokolowska, M., Eberlein, M., Alsaaty, S., Martinez-Anton, A., Logun, C., Qi, H. Y., & Shelhamer, J. H. (2014). The fish oil ingredient, docosahexaenoic acid, activates cytosolic phospholipase A₂ via GPR120 receptor to produce prostaglandin E₂ and plays an anti-inflammatory role in macrophages. Immunology, 143(1), 81–95. https://doi.org/10.1111/imm.12296
Metz, J. G., Kuner, J., Rosenzweig, B., Lippmeier, J. C., Roessler, P., & Zirkle, R. (2009). Biochemical characterization of polyunsaturated fatty acid synthesis in Schizochytrium: release of the products as free fatty acids. Plant physiology and biochemistry : PPB, 47(6), 472–478. https://doi.org/10.1016/j.plaphy.2009.02.002
Müller, C., Hardt, M., Schwudke, D., Neuman, B. W., Pleschka, S., & Ziebuhr, J. (2018). Inhibition of Cytosolic Phospholipase A2α Impairs an Early Step of Coronavirus Replication in Cell Culture. Journal of virology, 92(4), e01463-17. https://doi.org/10.1128/JVI.01463-17
Shikano, M., Masuzawa, Y., & Yazawa, K. (1993). Effect of docosahexaenoic acid on the generation of platelet-activating factor by eosinophilic leukemia cells, Eol-1. Journal of immunology (Baltimore, Md. : 1950), 150(8 Pt 1), 3525–3533.
Stephenson, K., Jensen, C. L., Jørgensen, S. T., Lakey, J. H., & Harwood, C. R. (2000). The influence of secretory-protein charge on late stages of secretion from the Gram-positive bacterium Bacillus subtilis. The Biochemical journal, 350 Pt 1(Pt 1), 31–39.
Su K. P. (2009). Biological mechanism of antidepressant effect of omega-3 fatty acids: how does fish oil act as a 'mind-body interface'?. Neuro-Signals, 17(2), 144–152. https://doi.org/10.1159/000198167
Su, K. P., Yang, H. T., Chang, J. P., Shih, Y. H., Guu, T. W., Kumaran, S. S., Gałecki, P., Walczewska, A., & Pariante, C. M. (2018). Eicosapentaenoic and docosahexaenoic acids have different effects on peripheral phospholipase A2 gene expressions in acute depressed patients. Progress in neuro-psychopharmacology & biological psychiatry, 80(Pt C), 227–233. https://doi.org/10.1016/j.pnpbp.2017.06.020
Tajuddin, N., Moon, K. H., Marshall, S. A., Nixon, K., Neafsey, E. J., Kim, H. Y., & Collins, M. A. (2014). Neuroinflammation and neurodegeneration in adult rat brain from binge ethanol exposure: abrogation by docosahexaenoic acid. PloS one, 9(7), e101223. https://doi.org/10.1371/journal.pone.0101223
Toelzer, C., Gupta, K., Yadav, S. K. N., Borucu, U., Garzoni, F., Staufer, O., Capin, J., Spatz, J., Fitzgerald, D., Berger, I., & Schaffitzel, C. (2020). Unexpected free fatty acid binding pocket in the cryo-EM structure of SARS-CoV-2 spike protein. bioRxiv, 2020.2006.2018.158584. https://doi.org/10.1101/2020.06.18.158584
Tong, X., Oh, E. K., Lee, B.-h., & Lee, J. K. (2019). Production of long-chain free fatty acids from metabolically engineered Rhodobacter sphaeroides heterologously producing periplasmic phospholipase A2 in dodecane-overlaid two-phase culture. Microbial Cell Factories, 18(1), 20. https://doi.org/10.1186/s12934-019-1070-8
van Wely, K. H., Swaving, J., Freudl, R., & Driessen, A. J. (2001). Translocation of proteins across the cell envelope of Gram-positive bacteria. FEMS microbiology reviews, 25(4), 437–454. https://doi.org/10.1111/j.1574-6976.2001.tb00586.x
Vincentini, O., Quaranta, M. G., Viora, M., Agostoni, C., & Silano, M. (2011). Docosahexaenoic acid modulates in vitro the inflammation of celiac disease in intestinal epithelial cells via the inhibition of cPLA2. Clinical nutrition (Edinburgh, Scotland), 30(4), 541–546. https://doi.org/10.1016/j.clnu.2011.02.007
Yan, B., Chu, H., Yang, D., Sze, K. H., Lai, P. M., Yuan, S., Shuai, H., Wang, Y., Kao, R. Y., Chan, J. F., & Yuen, K. Y. (2019). Characterization of the Lipidomic Profile of Human Coronavirus-Infected Cells: Implications for Lipid Metabolism Remodeling upon Coronavirus Replication. Viruses, 11(1), 73. https://doi.org/10.3390/v11010073
Yan, B., Zou, Z., Chu, H., Chan, G., Tsang, J. O., Lai, P. M., Yuan, S., Yip, C. C., Yin, F., Kao, R. Y., Sze, K. H., Lau, S. K., Chan, J. F., & Yuen, K. Y. (2019). Lipidomic Profiling Reveals Significant Perturbations of Intracellular Lipid Homeostasis in Enterovirus-Infected Cells. International journal of molecular sciences, 20(23), 5952. https://doi.org/10.3390/ijms20235952
Yoshida, K., Hashimoto, M., Hori, R., Adachi, T., Okuyama, H., Orikasa, Y., Nagamine, T., Shimizu, S., Ueno, A., & Morita, N. (2016). Bacterial Long-Chain Polyunsaturated Fatty Acids: Their Biosynthetic Genes, Functions, and Practical Use. Marine drugs, 14(5), 94. https://doi.org/10.3390/md14050094