Team:Edinburgh UG/Design
Our Hydrolyte Design
Designing our device using CAD software
Lab gloves seemingly have multiple purposes
It was clear to us that in any biohydrogen production scenario- a system of simultaneously culturing, collecting hydrogen and achieving maximum surface area for sufficient light exposure were key barriers. We have thus created a small system that allows immediate use of the gas produced while also sustaining our bacterial cells. Our 'artificial leaf', inspired by a 2015 paper on biohydrogen [footnote 1], was kindly 3D printed by the College of Engineering here in the University of Edinburgh. We have affectionately called the device a 'hydrolyte'.
An alginate hydrogel, containing Rhodobacter sphaeroides, will be placed on top of the red, grooved panel. The flow of medium within the grooves will nourish the culture. The white panel will be transparent in the final design to allow photosynthesis, and contains 2 small holes for insertion of syringes. Hydrogen gas extraction will be carried out through these syringes.
Our 3D printed prototype as an initial step towards the final, sleeker design
Grooves in the internal compartment for nutrient flow
Cross-sectional diagram of our Hydrolyte design
Team member Anmol getting in touch with his creative side
Measuring the pore size of different types of PET fabric, which will be coated with the hydrogel. Left: poly-silk. Right: poly-cotton
For the hydrolyte, the R. sphaeroides cells will be immobilized in the alginate hydrogel/fabric composite. We measured the pore size of different types of polyester fabrics to determine which would be best for our system. The 2015 paper used a range of fabrics with different pore sizes, the lower end being 250 micrometres. Das et al. also worked with an algal species that is larger in cell size than our bacterium. We chose to work with poly-cotton and poly-linen as the pore sizes were approximately 100 micrometres. As poly-silk is non-porous, it would be unsuitable due to a restriction of nutrient flow to the bacterial cells.
Overview
We want to introduce a hydrogen producing enzyme (hydrogenase) into R. sphaeroides to increase the rate of its natural hydrogen production. We worked with two different hydrogenases, comparing their activities and deciding which is the most efficient in R. sphaeroides. The enzymes are from different species and have slightly different modes of action: Chlamydomonas reinhardtii and Pyrococcus furiosus.
R. sphaeroides is a species that is difficult to transform (Laibe & Hanson, 2014). As such, we will first transform a conjugative E. coli strain SM10 and S17 with two plasmids pOGG011 and pOGG026 respectively (with different origins of replication) for each design. This will be followed by a conjugation of SM10 and S17 with R. sphaeroides 2.4.1 strain, which will lead to transformation of the two plasmids.
Method 1: Chlamydomonas reinhardtii
We aim to introduce a hydrogenase gene cluster comprising of hydrogenase, ferredoxin and NAD+/ferredoxin oxidoreductase. This gene cluster is taken from C. reinhardtii that has previously been used in E. coli by the Macquirie iGEM team in 2017 (https://2017.igem.org/Team:Macquarie_Australia). In addition, we want to add NADP+/ferredoxin oxidoreductase from Trichomonas vaginalis (Do et al., 2009).
We chose FeFe hydrogenase as it has higher hydrogen production rate than other types of hydrogenases. The algal hydrogenase from C.reinhardtii is very interesting as it has a much simpler structure than bacterial FeFe hydrogenases and has been shown to fold correctly in many different bacterial species (Rumpel et al., 2015).
Initially, ferredoxin is reduced by NAD+ and NADP+ ferredoxin oxidoreductases, which can be represented by
Then [FeFe] hydrogenase can use the electrons from reduced ferredoxin for hydrogen gas production
Fig 1.
Fig 2.
NAD+ and NADH+ oxidoreducatases are encoded by FNOR and FNR and ferrodoxin is encoded by the FDX gene. The [FeFe] hydrogenase is encoded by HydA, and it requires 2 maturase enzymes encoded by genes HydEF and HydG. Altogether the R. sphaeroides will be transformed with two plasmids, one carrying the [FeFe]hydrogenase (HydA) and its maturase enzymes (HydEF and HydG) (Fig.1). The other plasmid will carry genes encoding two oxidoreductases (FNOR and FNR) and ferrodoxin (FNR) (Fig.2). Due to the difference in codon usage difference C. reinhardtii and R. sphaeroides (Kazusa DNA Research Institute, 2018), hydA along with its maturase enzymes will be codon optimized. This will ensure the efficiency of gene expression of the hydrogenase cluster from C. reinhardtii.
Method 2: Pyrococcus furiosus
Our second method will involve the use of NADP-dependent [NifFe]-Hydrogenase termed Sh1 encoded by four genes in a single operon SHI1, SHI2, SHI3, and SHI4 coming from Pyrococcus furiosus. This archaeon species is a hyperthermophile, meaning it can withstand extremely high temperatures. Sh1, can function at a wide temperature range and does not require a ferredoxin intermediate, such hydrogenases are very rare in nature. It can directly oxidise NADPH for hydrogen production in the following manner (Sun et al., 2010).
This hydrogenase requires 12 maturation factors. However, we are introducing only 3 as our BLAST search revealed that R. sphaeroides already has the other 9 maturation factors which help to fold its native NiFe hydrogenase.
As such R. sphaeroides will be transformed with two plasmids, one carrying the 4 genes from Sh1 operon (Fig.3) and another plasmid with the required maturation factors (Fig.4)
Fig 2.
Fig 4.
For Methods 1 and 2, we used 3 different strong rRNA promoters from R. sphaeroides (BBa_J95023, BBa_J95024 and BBa_J95025) alongside synthetic lac promoter (BBa_J95022) which we later characterised in E. coli. We had to create polycistronic transcription units as the number of characterised R. sphaeroides promoters and RBS (BBa_J95016, BBa_J95017, BBa_J95018) is limited and we did not want to reuse the same promoter or RBS twice as this would create a perfect homologous recombination site.
Additional method:
During the modelling stage, Flux Balance Analysis showed that Formate Dehydrogenase overexpression in E. coli should increase hydrogen production rates more than any other hydrogenase. Therefore, we were very eager to try this enzyme as well. Formate dehydrogenase (FDH_h) facilitates this reaction: CO2 + NADH <-> HCOO- + NAD+. Formic acid is a chemical commodity and can be easily stored and transported, making it a stable form of hydrogen fuel. It can be converted into hydrogen gas by E.coli's native formate hydrogen lyase (Yoshida et al., 2005). Although E. coli has its own FDH, we decided to introduce FDH_h from Clostridium carboxidivorans as it does not require any rare cofactors and has high affinity for NADH and CO2 making it very efficient in formate production (Alissandratos et al., 2013).
The enzyme was codon optimised for optimal expression in E.coli. 6xHis-tag was added to the N-terminus for easy purification and the 139th amino acid Selenocysteine was changed into Cysteine to allow expression in regular E.coli strains. We used a strong T7 promoter to overexpress this enzyme in BL21DE3 strain. The transcriptional unit (Fig 5.) was assembled using Golden Gate as described below.
CIDAR MoClo Assembly
Based on Golden Gate technology...
As our design requires an assembly multiple parts, we decided to use a one-pot digestion and ligation multipart assembly method named CIDAR MoClo Golden Gate assembly (Iverson et al., 2015). This method involves 2 levels of assembly of parts containing modular overhanging 4pb fusion sites (coloured circles). MoClo also relies on the alternate use of Type IIS restriction enzymes BbsI and BsaI. These enzymes recognize a 6bp non-palindromic sequence and make a cut ahead of that sequence generating single stranded fusion sites and removing the initially recognized restriction site. CIDAR MoClo provides unique fusion sites, which we replaced with iGEM compatible fusion sites.
Fig.5
a) Level 0 modules are digested with BsaI and assembled with T4 ligase into DVK plasmid to form TU, this is referred to as level 1 assembly.
b) Multiple TUs are digested with BbsI and ligated with T4 ligase into level 2 destination vector
c) this leads to formation of the final plasmid.
The level 0 parts are individual genetic components such as the ribosome binding site, promoter, coding sequence, terminator and spacer enclosed in a plasmid (Fig.5a). Each part is flanked by two fusion sites, which are complementary to one fusion site each of other parts or destination vectors. The fusion sites can only pair once they are digested into single stranded overhangs by a restriction enzyme. Level 1 destination vector (with lacZ) also has BbsI restriction sites (directed in reverse to BsaI sites) flanking each fusion site and antibiotic resistance gene for selection. After digestion (with BsaI) single stranded fusion parts and a destination vector assemble together in a predetermined order to form a transcriptional unit (TU) with remaining BbsI recognition sites for level 2 assembly. Destination vector DVK also contains a lacZ screening cassette which excision and replacement with multigene construct allows identification of successful level 1 constructs via the blue-white screening method.
Several previously assembled TUs with fusion sites (level 1 constructs) are subsequently used for level 2 assembly of the final plasmid (Fig.5b). Level 2 destination vector (either pOGG011 or pOGG026) have two BbsI recognition sites that flank fusion sites. One-pot reaction is repeated, yet this time with a different restriction enzyme than the one used in level 1 assembly such as BbsI. This results in the assembly of TUs into destination vector, which altogether form level 2 assembly vector as shown here in fig.5c. Unfortunately, we do not have a screening method that would allow us to distinguish correctly assembled and self-ligated plasmids at this level of assembly. Additionally pOGG026 and DVK plasmids have the same KanR resistance gene. This requires us to perform multiple purification rounds to identify colonies containing successful constructs. The same procedure is repeated for the remaining three plasmids described before (Fig.1, 2, 4)
Vectors
The following figures depict position of BsaI and BbsI recognition and cleaving sites on DVK plasmid. Both enzymes make a staggered cut at the same position. During level 1 assembly, BsaI cuts at 2 positions, which removes the lacZ cassette and BsaI recognition sites and forms overhanging fusion sites. These are recovered upon ligation with one downstream and upstream fusion sites. During level 2 assembly, BbsI digestion generates the same 2 overhanging fusion sites to allow subsequent ligation.
DVK:
- lacZ: blue-white screening
- KanR: selection
- BsaI and BbsI sites for CIDAR MoClo assembly
pOGG026 and pOGG011
- KanR (pOGG026) or SpecR (pOGG011): selection
- BbsI sites for CIDAR MoClo assembly (Prior to the assembly, the BbsI sites are introduced into both vectors via PCR amplification to make them Golden gate compatible)
- Laibe, P. & Hanson, D., 2014. Transformable rhodobacter strains, method for producing transformable rhodobacter strains. United States Patent Application Publication . Available at:https://patentimages.storage.googleapis.com/43/c3/5c/908a269180a7d8/US20140080176A1.pdf.
- Sun, J. et al., 2010. Heterologous Expression and Maturation of an NADP-Dependent [NiFe]-Hydrogenase: A Key Enzyme in Biofuel Production. PLoS ONE, 5(5).
- Iverson, S.V. et al., 2015. CIDAR MoClo: Improved MoClo Assembly Standard and New E. coli Part Library Enable Rapid Combinatorial Design for Synthetic and Traditional Biology. ACS Synthetic Biology, 5(1), pp.99–103.
- Do, P.M et al., 2009. Engineering Escherichia coli for Fermentative Dihydrogen Production: Potential Role of NADH-Ferredoxin Oxidoreductase from the Hydrogenosome of Anaerobic Protozoa. Applied Biochemistry and Biotechnology, 153(1-3), pp.21–33.
- Kazusa DNA Research Institute, 2018. Codon usage table. Kazusa. Available at: https://www.kazusa.or.jp/en/ [Accessed October 13, 2019].
- Rumpel, S. et al., 2015. Structural Insight into the Complex of Ferredoxin and [FeFe] Hydrogenase fromChlamydomonas reinhardtii. ChemBioChem, 16(11), pp.1663–1669.
- Alissandratos, A., et al. (2013). Clostridium carboxidivorans Strain P7T Recombinant Formate Dehydrogenase Catalyzes Reduction of CO2 to Formate. Applied and Environmental Microbiology 79(2) 741-744. Yoshida A., et al. (2005). Enhanced hydrogen production from formic acid by formate hydrogen lyase‐overexpressing Escherichia coli strains. Appl Environ Microbiol. 2005;71:6762–6768.
- Das, A.A., Esfahani, M.M., Velev, O.D., Pamme, N. and Paunov, V.N., 2015. Artificial leaf device for hydrogen generation from immobilised C. reinhardtii microalgae. Journal of Materials Chemistry A, 3(41), pp.20698-20707