Choosing our Compound

An important principle in securing the environmental viability of SchistoGONE was selecting the anti-schistosomal compound with the best balance in environmental safety and anti-schistosomal efficiency. To do this, we repeated a process of research and redesign of the project which can be highlighted by our journey through candidate compounds plumbagin, ent-pimaradienoic acid, curcumin, arachidonic acid, and ent-kaurenoic acid.

Our first choice was plumbagin, a plant-based organic compound with proven antischistosomal effects in vitro against the cercarial phase (Zhang et al., 12). Following further research, however, we consulted numerous professors who—using their experience in parasitic diseases and schistosomes—helped us determine the extreme environmental hazards involved in using plumbagin. Plumbagin has detrimental effects on other wildlife including fish, crustaceans, and snails with concentrations of just 0.9mg/L leading to fatality in fish (Tu et al., 20).

Our second choice of ent-pimaradienoic acid had the opposite problem in the balance between anti-schistosomal efficiency and environmental viability. Ent-pimaradienoic acid was tested in-vivo in mice with proof of antischistosomal properties and no signs of toxicity (Mao et al., 20). However, the isomer of pimaradienoic acid that we had the capacity to produce via synthetic biology was syn-pimaradienoic acid, a distinct constitutional isomer that would predictably have different chemical properties and yield different results. Syn-pimaradienoic acid is produced through a different branch of pimarane diterpenoid biosynthetic pathways, and the biosynthetic pathway of ent-pimaradienoic acid is not yet elucidated, meaning we are unable to reliably produce the anti-schistosomal isomer. Moreover, the pimaradienoic acid stereoisomer available to industry for purchase is the enantiomer of ent-pimaradienoic acid, called continentalic acid, which presumably has yet another set of chemical properties. Indeed, in vitro tests of continentalic acid on schistosoma sporocysts through our partnership with UCSD's Center for Discovery and Innovation in Parasitic Disease (CDIPD) confirmed this as zero effects were recorded at all concentrations. The discrepancies between the isomers we could purchase, produce, and the one used in the original paper made it impossible to rely on pimaradienoic acid.

Curcurin was evaluated and found to have antischistosomal properties as it caused the death of all adult schistosoma mansoni adult worms in doses ranging from 5-100 mM (Magalhães et al., 09). However, the issue present was that the biosynthesis pathway to produce curcumin was unclear and had many critical parts missing, meaning that it would be extremely difficult to produce the compound itself. It had little to no precedent, and the gene sequences of the enzymes responsible for catalyzing its production are not yet elucidated.

Arachidonic acid had a similar problem in the replication of the biosynthesis pathway. Though it had been found to be safe and efficacious in killing schistosoma mansoni worms in mice and hamsters, arachidonic acid's pathway was only half known (Tallima et al 20). Through extensive research it was determined that only the second half of the biosynthesis pathway for arachidonic acid was present in literature while the first half was unknown, making this compound unreliable for testing in the SchistoGONE project.

Ent-kaurenoic acid was a candidate compound due to its initial viability as a schistosomide and the clarity of its biosynthesis pathways. After further research it was found that in actuality, ent-kaurenoic acid was not effective against all stages of the worm's life cycle and only derivatives of the compound are considered promising (Oliveira et al 20). This meant that it would be inefficacious for SchistoGONE's goal, which involves killing the worm at its sporocyst stage.

Sanguinarine was selected as the natural product to use in this project. It was found that sanguinarine resulted in 100% schistosome mortality in concentrations of 10 µM (equivalent to 3.68 µg/mL), meeting the World Health Organization's criteria of "hit" compounds for controlling schistosomiasis (Zhang et al 12). The results of the compound testing on specifically schistosoma sporocysts conducted under the UCSD CDIPD revealed that sanguinarine had the highest effectiveness against in-vitro sporocysts compared to plumbagin and the enantiomer of pimaradienoic acid, continentalic acid. After just 0.05 µM concentration of sanguinarine, sporocysts began to die. And by 1 µM concentration of sanguinarine, 100% sporocyst mortality was recorded.

In terms of environmental safety, sanguinarine has been tested for toxicity via in vitro assays and in-vivo lab rat trials revealing insignificant effects on organisms (Mackraj et al., 2008). Furthermore, the unrestrained environmental release of the compound is improbable. The biosynthesis pathway involves multiple modified yeast and E. coli organisms that produce benign substrates of sanguinarine (Fossati et al, 2014). Should the modified organisms be swept into the river or dispersed in the environment, the construction of sanguinarine by a single yeast cell would not be possible. Sanguinarine's efficacy as a schistosomide and its environmental safety led this compound to be chosen for the SchistoGONE project.

Verifying the Compounds

A primary goal of ours at the beginning of the compound selection process was to verify their anti-schistosomal properties, specifically against Schistosoma sporocysts. Many of the compounds had never been tested specifically on sporocysts before; and some, such as the differing isomers of pimaradienoic acid, had little basis in the literature as an anti-schistosomal in general. Testing sporocysts is technically challenging for our high school team as we lack the training and funds and the laboratory safety certifications that are a prerequisite for working with a pathogen of humans.

Fortunately, our contact, Dr. Conor R. Caffrey, at the CDIPD, UC San Diego, specializes in tropical parasitic diseases and already maintains live Schistosoma parasites. Dr. Caffrey generously agreed to conduct the compound testing experiments in our place.

The process that Dr. Caffrey employs is as follows: Schistosoma mansoni eggs, each of which contains one miracidium, are suspended in water. The eggs are tricked to hatch by placing them under a bright light. Once hatched, the miracidia are isolated and then tricked into transforming into the sporocysts by adding a salty medium (to mimic the composition of salts inside a snail). These sporocysts can be maintained for 5-7 days. During that time, an assay using a range of concentrations of continentalic acid, sanguinarine, and plumbagin is conducted on the sporocysts. Data is collected after 24 hours and the antischistosomal effect is given as the percentage of sporocysts surviving.

Optimizing the Biosynthesis Pathway

Sanguinarine is a benzylisoquinoline alkaloid (BIA) whose complex biosynthetic pathway is closely related to those of well-documented morphinans. The production of BIAs using yeast expression systems has precedent in the literature, with the sanguinarine pathway having been successfully reconstituted in Saccharomyces cerevisiae using (R,S)-THP as a fed substrate (Trenchard and Smolke, 2015). However, we endeavored to connect the THP production pathway directly to the sanguinarine production pathway. This would require combining a Escherichia coli expression system with a S. cerevisiae expression system in a concerted multi-strain expression mechanism.

In fundamental terms, most BIAs are derived from L-tyrosine, an amino acid precursor of L-DOPA and dopamine that is naturally available in E. coli. A tyrosinase—in this case RsTYR, isolated from the wilt bacterium Ralstonia solanacearum—catalyzes the formation of L-DOPA from L-tyrosine, a precursor of dopamine that is converted to such by DOPA decarboxylase (DDC, or DODC). A monoamine oxidase (MAO) then converts dopamine to 3,4-dihydroxyphenylacetaldehyde (3,4-DHPAA), a dopaldehyde that spontaneously condenses to form an enantiomeric mixture of (R)- and (S)-tetrahydropapaveroline (-THP) via a Pictet-Spengler reaction. We roughly followed a documented reconstitution of this portion of the biosynthesis in a two-strain E. coli expression system (Nakagawa et al., 2014) while optimizing enzymes chosen.

This first biosynthetic phase is difficult to reconstitute in heterologous expression systems due to the fact that TYR demonstrated o-diphenolase activity toward both L-tyrosine, its intended substrate, and THP, the intended product and the major BIA precursor (Nakagawa et al., 2014)—tyrosinases are known naturally to have both monophenolase and o-diphenolase activity under aerobic conditions. A portion of produced THP would thus be oxidized, which would produce lower yields than desired. We then optimized our choice of TYR by choosing one that possessed a comparatively lower o-diphenolase activity: a RsTYR (Romero et al., 2006). The two-step conversion from L-DOPA to 3,4-DHPAA, catalyzed by DDC and MAO, was bypassed by replacing those enzymes with a DHPAA synthase (DHPAAS) from the mosquito Aedes aegypti (Vavricka et al., 2011). DHPAAS has been identified in multiple species of insect, including A. aegypti, Drosophila melanogaster, and Anopheles gambiae, to directly produce 3,4-DHPAA from L-DOPA in a one-step conversion (Vavricka et al., 2011). The replacement of DDC and MAO with DHPAAS also reduced the number of genes to be introduced into E. coli.

The next portion of sanguinarine's biosynthetic pathway involves the 11-step conversion of (R,S)-THP to sanguinarine through a series of alkaloid intermediates. The enzymes responsible for catalyzing each of these conversions are, respectively, norcoclaurine 6-O-methyltransferase (6OMT); coclaurine N-methyltransferase (CNMT); 4'-O-methyltransferase (4'OMT); berberine bridge enzyme (BBE); cheilanthifoline synthase (CFS), a cytochrome P450 (CYP); a cytochrome P450 reductase (CPR); stylopine synthase (STS); tetrahydroprotoberberine N-methyltransferase (TNMT); cis-N-methylstylopine 14-hydroxylase (MSH); protopine 6-hydroxylase (P6H); and dihydrobenzophenanthridine oxidase (DBOX), which was experimentally observed to be unnecessary in heterologous sanguinarine production (Trenchard and Smolke, 2015). For our design, gene variants (i.e., originating from the opium poppy Papaver somniferum, enzyme abbreviation prefix Ps, or California poppy Eschscholzia californica, prefix Ec) were selected based on optimization data from Trenchard and Smolke (2015). ATR1 is a CYP from Arabidopsis thaliana.

This pathway has previously been reconstituted using a two-strain Saccharomyces cerevisiae expression system, with first producing the intermediate (S)-reticuline using 6OMT, CNMT, and 4'OMT and the second continuing the pathway to produce sanguinarine using the remaining enzymes (Hawkins and Smolke, 2008; Trenchard and Smolke, 2015). A 6OMT-CNMT expression plasmid and 4'OMT expression plasmid were used in the first strain (Hawkins and Smolke, 2008), and Gateway-compatible single expression plasmids containing individual recombinant genes were used in the second strain (Trenchard and Smolke, 2015), with expression cassettes being homologously integrated into the S. cerevisiae genome in each case. Stylopine has also been produced in Pichia pastoris from the fed substrate reticuline using recombinant BBE, CFS, and STS genes (Hori et al., 2016).

The P. pastoris-based stylopine production system cloned all three genes involved into a single vector, but for greater simplicity and due to the entire THP-to-sanguinarine pathway involving ten enzymes, we opted to express pairs of the genes in the THP-to-sanguinarine portion of the pathway (with five yeast plasmids expressing 6OMT-CNMT, 4'OMT-BBE, ATR1-CFS, STS-TNMT, and MSH-P6H).

The THP production pathway plasmids were designed for transformation into standard E. coli strain K-12 due to precedent existing for bacterial expression (Nakagawa et al., 2014). The incompatibility of cytochromes P450 and their reductases with prokaryotic expression systems due to their being membrane-associated proteins led us to select S. cerevisiae strain W303α (genotype MATα: leu2-3,112 trp1-1 can1-100 ura3-1 ade2-1 his3-11,15), which has exhibited good yield of recombinant BIAs in the literature (Hawkins and Smolke, 2010; Trenchard and Smolke, 2015).

Designing the Constructs and Putting It All Together

RsTYR (enzyme reference number EC 1.14.18.1; GenBank accession number NP_518458) and DHPAAS (EC 4.1.1.107) were codon-optimized for expression in E. coli. The coding sequence of each enzyme was cloned into separate pET-28a(+) bacterial vectors with a T7-lac promoter expression system, creating Plasmids A (TYR) and B (DHPAAS). pET-28a(+) is well-documented as a leading choice for recombinant protein expression in E. coli and contains an N-terminal His-tag (Shilling et al., 2020). Each gene fragment was designed to be cloned into the multiple cloning site of the PCR-linearized vector via Gibson assembly.

While the method of Trenchard and Smolke (2015) relies on S. cerevisiae-based plasmids expressing individual genes and uses Gateway cloning procedures, the method of Hori et. al (2014) expresses three genes in one plasmid and still produced desirable titers of stylopine, albeit in P. pastoris. The latter reports P. pastoris as being more effective than S. cerevisiae for multi-gene expression, but this was a comparison based only on the BBE-CFS-STS steps of the biosynthetic pathway with no data on the other steps (Hori et al., 2016). Only Saccharomyces was available to us for experimentation, and following a series of professor consultations we decided to continue with S. cerevisiae while expressing two genes per plasmid, a novel modified combination of the aforementioned methods.

In theory, the integration of each plasmid into the yeast genome via homologous recombination would produce the most optimal and stable yields as compared to the presence of multiple episomal plasmids. In this circumstance, each gene is to be placed into a cassette cloned into a yeast expression vector (YEp), allowing for the measurement of enzyme production and, in the presence of their respective substrates, compound production. After ascertaining the success of this initial cloning process, gene cassettes would be linearized and subcloned into yeast integrative vectors (YIp), which lack origins of replication and will become integrated into the chromosome at specific loci with appropriately-designed homology regions. Nevertheless, we first introduced the YEps episomally into yeast cells to test their fundamental functionality.

We considered multiple methods for the simultaneous expression of two genes in one plasmid for the reconstitution of the THP-to-sanguinarine pathway. Polycistronic (IRES) vectors, in which one promoter controls the expression of multiple genes with internal ribosome entry sites, are not particularly well-characterized in yeast, especially in the context of phytochemical expression. Most studies relating to phytochemistry and biosynthesis use plasmids containing either divergent dual promoters or sequential gene cassettes controlled by different promoters. The use of the same promoter multiple times in one plasmid, however, may be met with complications relating to interference and unwanted spontaneous homologous recombination. A two-cassette expression system was therefore designed, whereby each plasmid consisted of a PCR-linearized pXP418 vector backbone and two gene cassettes: the first gene would be controlled by the TEF1 promoter already present in the vector, followed by an ADH1 terminator fragment; the second would be controlled by a variable promoter fragment optimized for the second gene and the CYC1 terminator in the vector. pXP418 is a high-copy plasmid vector chosen for the presence of a 2-micron origin of replication (2µ ori); URA3 auxotrophic marker; and TEF1 promoter, a strong constitutive promoter commonly used for controlling the expression of plant-derived enzymes previously used in this context (Hawkins and Smolke, 2010). The gene fragments, ADH1 terminator sequence, and variable promoter sequence were then to be assembled via Gibson assembly into the linearized vector.

The yeast plasmids were organized and numbered as follows: Plasmid 1, expressing 6OMT and CNMT; Plasmid 2, 4'OMT and BBE; Plasmid 3, ATR1 and CFS; Plasmid 4, STS and TNMT; and Plasmid 5, MSH and P6H.

Plasmid 1 contains a P_TEF1-6OMT-T_ADH1 cassette followed by a P_PGK1-CNMT-T_CYC1 cassette cloned into a PCR-linearized pXP418 expression vector. The strong PGK1 promoter, another constitutive promoter commonly used in phytochemical applications (Lee et al., 2007), was selected to control expression of CNMT. The same promoter-terminator combination and cassette scheme was used in Plasmid 2 (P_TEF1-4'OMT-T_ADH1 and P_PGK1-BBE-T_CYC1), Plasmid 4 (P_TEF1-STS-T_ADH1 and P_PGK1-TNMT-T_CYC1), and Plasmid 5 (P_TEF1-MSH-T_ADH1 and P_PGK1-P6H-T_CYC1).



Plasmid 3 presented different circumstances due to the explicit identification of an optimal promoter for controlling CFS expression. CFS was found to be optimally expressed under the control of P_TDH3 (synonymous with P_GDP and P_GAP) in a low-copy plasmid or after chromosomal integration (Trenchard and Smolke, 2015). pXP418, being a high-copy YEp, would thus theoretically produce lower initial yields but would, in fact, also potentially improve post-integration expression and yield due to the higher number of copies. ATR1, a CPR, was found to be unnecessary for the production of stylopine in P. pastoris (Hori et al., 2014), but it was included in Plasmid 3 in case of discrepancies between that data and production in S. cerevisiae. The Plasmid 3 cassettes are therefore P_TEF1-ATR1-T_ADH1 and P_TDH3-CFS-T_CYC1.



It would be practical to separate the expression of these genes among multiple strains of S. cerevisiae due to increased ease of yield quantification at each step along the pathway. Collectively, this sequence of plasmids (A-B and 1-5) would reconstitute the full sanguinarine biosynthetic pathway from L-tyrosine to sanguinarine in a concerted E. coli and S. cerevisiae expression mechanism.

Choosing Our Implementation

When considering how to introduce our design into the environment in a safe and effective way, there were a myriad of concerns to address: could we be sure that schistosomes will be reached by our product in the environment? Will the ecosystem be threatened by an environmental release of our constructs?

To answer these questions, we consulted numerous professors, who—using their experience with schistosoma and environmental application—helped us determine what would be feasible given our resources and materials.

Taking in the information from these meetings, our project went through many different stages of development before we decided upon the final iteration. The first iteration involved engineering yeast to secrete plumbagin directly into the water in order to kill the cercariae stage of schistosoma. We soon found that this plan had many large environmental concerns as plumbagin is toxic not only to schistosomes but also to many other organisms. The potential ecological ramifications caused by releasing chemical substances directly into the environment was a shared concern for all our candidate compounds, as even nontoxic natural products could have unforeseen negative impacts on certain organisms. The same went for releasing modified microorganisms into the environment given the unpredictability of how modified DNA would behave and spread. These glaring issues revealed by the input of several experts led us to regroup and rethink the direction of the project.

The implementation method we decided on considers all of these concerns mentioned previously. It starts with the modified yeast and E. coli being situated in specially designed plates along river banks, which then attract Biomphalaria Glabrata snails that consume the yeast and sanguinarine. This results in the sporocysts inside the snails being eradicated without effect on the snail. The ecosystem of these water bodies will not be drastically altered because the snails themselves will survive, sanguinarine will not be released into the water, and the modified microorganisms will not be spread either.

Verifying the Implementation

In order to verify that the implementation method will be effective in reaching schistosomes, we conducted tests on whether snails are actually attracted to our strain of yeast over other food sources. Results from our experiments indicated a significant preference for yeast coated food.

In order to verify the stability of the implementation as well as the environmental effect of the modified constructs, we conducted tests on yeast surviving in simulated environment conditions. Our observations concluded that the W303α strain of yeast we were using would die after 48 hours of exposure, which ensures the stability of the constructs through events such as rising tides and environmental containment should the microorganisms be released.

Fossati E, Ekins A, Narcross L, et al. (2014). "Reconstitution of a 10-gene pathway for synthesis of the plant alkaloid dihydrosanguinarine in Saccharomyces cerevisiae." Nat Commun. 5(1):1—11.

Mao L, Kawaide H, Higuchi T, et al. (2020). "Genomic evidence for convergent evolution of gene clusters for momilactone biosynthesis in land plants." Proc. Natl. Acad. Sci. U.S.A. 117(22):12472—12480.

Magalhães LG, Machado CB, Morais ER, et al. (2009). "In vitro schistosomicidal activity of curcumin against Schistosoma mansoni adult worms." Parasitol Res. 104(5):1197—1201.

Shilling PJ, Mirzadeh K, Cumming AJ, et al. (2020). "Improved designs for pET expression plasmids increase protein production yield in Escherichia coli." Commun Biol. 3(1):214.

Tallima H, Hanna VS, and El Ridi R. (2020). "Arachidonic acid is a safe and efficacious schistosomicide, and an endoschistosomicide in natural and experimental infections, and cysteine peptidase vaccinated hosts." Front Immunol. 2020; 11.

Trenchard IJ and Smolke CD. (2015). "Engineering strategies for the fermentative production of plant alkaloids in yeast." Metabolic Eng.

Tu, Xiao, et al. "Identification of Plumbagin as an Effective Chemotherapeutic Agent for Treatment of Gyrodactylus Infections." Aquaculture, Elsevier, 11 Jan. 2021,
www.sciencedirect.com/science/article/abs/pii/S004484862100034X.

Zhang, Si-Ming, and Kristen A Coultas. "Identification of Plumbagin and Sanguinarine as Effective Chemotherapeutic Agents for Treatment of Schistosomiasis." International Journal for Parasitology. Drugs and Drug Resistance, Elsevier, Dec. 2013,
www.ncbi.nlm.nih.gov/pmc/articles/PMC3638872/.