Design

Background

We wanted to engineer a probiotic bacterium able to reduce cholesterol levels in the body, as a preventative for cardiovascular disease.

We started off by choosing the gut as our target, as it represents a “bottleneck” in cholesterol metabolism. The majority of cholesterol is synthesized in the liver, after which it is transported along with bile to the small intestine for absorption¹. Simultaneously, dietary cholesterol is transported through the gastrointestinal tract to the small intestine, where it is also absorbed². Absorbed cholesterol is then packaged in lipoproteins and sent for use around the body¹. No matter what, the cholesterol that ends up in our cells will always at some point have been in the gut first².

As such, if we can prevent cholesterol absorption in the gut, we can target excess amounts of both dietary and endogenous cholesterol².

More and more, scientists are starting to uncover the role that our gut microbiome plays in our body, including in cholesterol metabolism. In the ‘90s, scientists discovered that certain species of bacteria in the gut convert cholesterol to coprostanol, through a three-step metabolic pathway³.

While cholesterol is absorbed by the gut, coprostanol is not due to its chemical structure³. As such, cholesterol that once may have been absorbed by the body, would be excreted harmlessly as waste if converted into coprostanol in the gut.

Humans with a higher prevalence of bacteria that are able to perform this conversion show significantly lower serum cholesterol levels, and are at lower risk of cardiovascular disease.

Our objective

We ask ourselves: what if we could engineer a safe, scalable and effective probiotic bacterium that could convert cholesterol to coprostanol in the gut as a preventative for cardiovascular disease?

Our probiotic would have to:

1. Import cholesterol
2. Encode three enzymes that together convert cholesterol to coprostanol
3. Be safe, tolerable and survive the journey through the body

Enzymatic conversion of cholesterol to coprostanol

We were tasked with identifying three enzymes to sequentially catalyze each of the conversion steps from cholesterol to coprostanol

For the first step of converting cholesterol to cholestenone, we opted to use ismA, a hydroxysteroid dehydrogenase enzyme from the coprostanol-forming bacterium E. coprostanoligenes. ismA was previously validated by Kenny et al. to perform the cholesterol to cholestenone conversion³.

Unfortunately, the enzymes that do steps two and three of the enzymatic pathway are unknown in any coprostanol-converting bacteria.

To find enzymes for steps 2 and 3, we looked to natural sterol biosynthetic processes in the hopes that we could repurpose existing enzymes that do the same enzymatic reactions.

We dove deep within ourselves and studied the bile acid synthesis pathway in humans. Bile acids are synthesized from cholesterol through a multi-step metabolic pathway⁴. Interestingly, the synthesis of bile acids from cholesterol involves the same removal of cholesterol’s double bond (Δ4,5) as in the synthesis of coprostanol from cholesterol⁴.

Given a seemingly similar synthesis pathway, we reasoned that bile acid synthesis would be a good place to look for enzymes that could participate in the conversion of cholesterol to coprostanol.

Within the human bile acid synthesis pathway, we found the enzyme AKR1D1, a 5β-reductase that does the exact same enzymatic reaction as the putative second enzyme of the cholesterol to coprostanol conversion pathway4,5. Specifically, it reduces the double bond on the third carbon of the sterol backbone4.

The next step in the pathway is catalyzed by a 3α-hydroxysteroid dehydrogenase enzyme called AKR1C4⁶. AKR1C4 converts the ketone group on the 3rd carbon of the sterol backbone into a hydroxyl group; the exact same reaction as the third enzyme of the coprostanol synthesis pathway⁶,⁷.

The only difference between the substrates of AKR1D1 and AKR1C4 and our intended substrates is the presence of an extra hydroxyl group on the 7th carbon of the sterol backbones in the case of bile acid synthesis. However, given the striking similarity between our intended substrates - cholestenone for step 2 and coprostanone for step 3, we decided to repurpose AKR1D1 and AKR1C4 for the second and third steps of cholesterol to coprostanol conversion.

An importer for cholesterol

Since our proteins were expressed intracellularly in B. subtilis, we needed a way to import cholesterol intracellularly into our probiotic bacterium so it could be converted into coprostanol. There were two options: either cholesterol is transiently transported across the cell membrane through non specific importers, or certain bacteria that metabolize cholesterol have specialized machineries that import cholesterol for use.

The first option was unlikely, because it is known that not all bacteria are able to import cholesterol. So, we assumed bacteria such as E. coprostanoligenes have evolved specific importers that purposefully import cholesterol, and we went searching for such transporters.

Bacteria tend to co-regulate genes that are part of the same metabolic pathways by housing them in operons. A well known example of this is the lac operon, which encodes three genes responsible for processing exogenous lactose, all under one promoter. We asked ourselves: what if ismA, which plays a role in the cholesterol metabolism pathway in E. copro, was part of an operon containing other genes responsible for cholesterol metabolism? Could one of these genes in this hypothetical operon be our desired cholesterol importer?

We located ismA in the genome of E. coprostanoligenes and used the NCBI ORFfinder to identify all open reading frames greater than 75bp within 10 kilobase pairs in either direction of the ismA gene. Shockingly, we observed another gene going in the same direction as ismA, that starts just 91bp downstream of ismA. Since 91bp is significantly shorter than any possible promoter, it was clear to us that this gene was in the same operon as ismA.

When we blasted this adjacent gene against the UniProtKB database of proteins, the top hit was an uncharacterized protein part of the MFS transporter protein family. The function of MFS transporters is to import and export small molecules such as sugars in and out of the bacterial cell. Since this MFS transporter was in the same operon as the ismA gene, we reasoned that it could possibly be a cholesterol importer.

We decided to test this hypothesis by expressing the MFS transporter in B. subtilis and measure cholesterol uptake through multiple experimental approaches. We chose B. subtilis as our testing organism not only because it was our probiotic chassis of choice, but also because both B. subtilis and E. coprostanoligenes are gram positive bacteria. Since the cell walls of gram positive and negative bacteria vary widely, we wanted to be sure that the expressed MFS transporter would have a high likelihood of being properly inserted inside our bacterium’s cell wall. Choosing a gram positive bacterium such as B. subtilis would ensure this.

A chassis for cholesterol reduction

To be able to make this pathway work in the gut, we needed a mini “bioreactor” in which to put these enzymes, so that they could be in the right conditions to function. We opted to go with Bacillus subtilis, a human-friendly, widely abundant bacterium that is known to survive the harsh environment of the gastrointestinal tract⁷.

Unlike in the petri dish, it is not possible to provide selective pressure for our probiotic bacterium to continuously express our genes of interest. If we encoded our three genes on a plasmid, many of these bacteria would lose the plasmid by the time they reach the gut, rendering our probiotic bacterium effectively useless. Therefore, we chose to stably integrate ismA, AKR1D1 and AKR1C4 into the genome of B. subtilis, where they would be continuously expressed without the need of selective antibiotic pressure.

Luckily, the 2012 Munich iGEM team designed a repertoire of standardized BioBrick components for B. subtilis, that we used⁸. We used their integrative vectors to integrate ismA, AKR1D1 and AKR1C4 into targeted regions of the genome, and expressed all three genes in B. Subtilis from a constitutive promoter.

Since B. subtilis, along with other gram positive bacteria, is a naturally difficult species to transform, we used a strain of B. Subtilis that is super competent. This super competent Bacillus subtilis expresses comK, a master regulator of competence and exogenous DNA uptake, under a xylose-inducible promoter. Upon addition of xylose to the growth media of Bacillus, the bacterium begins to express high levels of the comK transcriptional regulator and quickly starts taking up large amounts of DNA from its surroundings - primarily the DNA we provide for transformation.

References

  1. Rajaratnam, R.A., Gylling, H., Miettinen, T.A. Cholesterol Absorption, Synthesis, and Fecal Output in Postmenopausal Women With and Without Coronary Artery Disease (2001) Arteriosclerosis, Thrombosis, and Vascular Biology. 21:1650–1655. https://doi.org/10.1161/hq1001.097019
  2. Kruit, J. K., Groen, A. K., van Berkel, T. J., & Kuipers, F. (2006). Emerging roles of the intestine in control of cholesterol metabolism. World journal of gastroenterology, 12(40), 6429–6439. https://doi.org/10.3748/wjg.v12.i40.6429
  3. Kenny, D. J., Plichta, D. R., Shungin, D., et al. (2020). Cholesterol Metabolism by Uncultured Human Gut Bacteria Influences Host Cholesterol Level. Cell Host & Microbe, 28(2), 245–257.e6. https://doi.org/10.1016/j.chom.2020.05.013
  4. Reshetnyak V. I. (2013). Physiological and molecular biochemical mechanisms of bile formation. World journal of gastroenterology, 19(42), 7341–7360. https://doi.org/10.3748/wjg.v19.i42.7341
  5. Charbonneau, A., Van-Luu The, V. Genomic organization of a human 5β-reductase and its pseudogene and substrate selectivity of the expressed enzyme. (2001) Biochimica et Biophysica Acta (BBA) - Gene Structure and Expression, 1517, Issue 2,Pages 228-235, ISSN 0167-4781, https://doi.org/10.1016/S0167-4781(00)00278-5.
  6. Uniprot. P51857 · AK1D1_HUMAN. https://www.uniprot.org/uniprotkb/P51857/entry
  7. Khanna, M., Qin, K. N., Wang, R. W., Cheng, K. C. (1995). Substrate specificity, gene structure, and tissue-specific distribution of multiple human 3 alpha-hydroxysteroid dehydrogenases. The Journal of biological chemistry, 270(34), 20162–20168. https://doi.org/10.1074/jbc.270.34.20162
  8. Uniprot. P17516 · AK1C4_HUMAN . https://www.uniprot.org/uniprotkb/P17516/entry
  9. Cutting, S. M., Hong, H. A., Baccigalupi, L.. et al. (2009). Oral Vaccine Delivery by Recombinant Spore Probiotics. International Reviews of Immunology, 28(6), 487–505. https://doi.org/10.3109/08830180903215605
  10. Mckenney, P. T., Driks, A., Eichenberger, P. (2013). The Bacillus subtilis endospore: assembly and functions of the multilayered coat. Nature Reviews Microbiology, 11(1), 33–44. https://doi.org/10.1038/nrmicro2921
  11. LMU-Munich. (2012) Bacillus Promoters. Bacillus Bio Brick Box, https://2012.igem.org/Team:LMU-Munich/Bacillus_BioBricks/Promoters

Explore