Motivation of Hydrogenase

Figure 1A The reversible oxidation of molecular hydrogen.

Hydrogenase is an enzyme that catalyses the reversible oxidation of molecular hydrogen (H2). (Figure 1)

Hydrogenase can be sub-classified into three different types based on the active site metal content: iron-iron hydrogenase ([FeFe] hydrogenase), nickel-iron hydrogenase ([NiFe] hydrogenases), and iron hydrogenase. In contrast to [NiFe] hydrogenases, [FeFe] hydrogenases are generally more active in production of molecular hydrogen. Turnover frequency (TOF) in the order of 10,000 s−1 have been reported in literature for [FeFe] hydrogenases from Clostridium pasteurianum.[1] This has led to intense research focusing on the use of [FeFe] hydrogenase for sustainable production of H2.[2]

Normal E. coli bacteria contain [NiFe] hydrogenase, but the activity and expressive rate is non-obvious. For the above reasons, we decided to construct [FeFe] hydrogenases gene cluster for sustainable production of H2.

Figure 1B The inner structure of [FeFe]-hydrogensase.

The main functional catalytic group in [FeFe]-hhydrogenase is considered to be an iron-sulfur cluster domain with a di-iron center covalently linked to a dithiolate group.

Hydrogenases

At molecular level, the gene sequences involved in producing hydrogenase in different species vary wildly. In our study, we focus on hydrogenase gene cluster from Clostridium. acetobutylicum. The important genes include hydA, hydEF, hydG, which are expressed as HydA, HydE and HydF, HydG respectively. We will briefly introduce these enzymes below. （Tip：click enzymes to have fun:）

Figure 2A HydA is the main catalytic unit, whereas the rest of the hyd genes are co-expressed to achieve a stable maturation of the final functional HydA.

Figure 2B HydE, as well as HydG have a radical-SAM motif. In most of the cases, these two enzymes might form a complex to fulfill their functions in helping the HydA mature.

Figure 2C HydF, whose N-termiatal domain is homoligous to the GTPase family and C-terminatal domain putatively contains a iron-sulfur center bingding motif CxHx45HCxxC,is considered to provide energy during the process.

Figure 2D HydG,besides a radical-SAM motif Cx3Cx2C,it also has motif Cx2Cx22C to bind the redox center- [4Fe4S] cluster. The [4Fe4S] cubic cluster might transer to HydA.

Our goal is to transplant the gene clusters of [FeFe]-hydrogenase from Clostridium. acetobutylicum into E. coli, and produce a strain that could effectively produce hydrogen. This seemingly novel idea has been actually fulfilled by Yuki Honda, et al. [4] However, the methods and the result of gene manipulation was not efficient. They used the pETDuet-1+pCDFDuet-1 system to carry the hydEA and hydFG sequence separately. This method in cloning is not only laborious but also inefficient. Firstly, the expression of HydA, HydE, HydF, HydG are not controlled in a synchronized way; secondly, the two-plasmid system runs certain risk in the stability of the strain[4]. Thus to explore the strength of synthetic biology, we made certain improvements on the system from the level of gene manipulation.

Construction of [FeFe]-hydrogenases gene cluster

Principle of Molecular Cloning

To ensure normal enzyme activity, we need to make sure that these four enzymes are simultaneously expressed in E. coli with a moderate amount. The well-established high-efficiency Acembl system [5] came into our sight.

We adopted this Acembl system as a multi-expression system with special DNA replication origin and Cre-loxP site, which utilizes Cre recombinase to integrate four basic plasmid backbones into one. (Figure 3) Descriptions are as follows.

The Acembl system in our project involves four plasmids, pACE, pDC, pDS, and pDk, and each contains one of the four gene sequences we would like to fuse (Figure 3A-D).

Figure 3 Integration of four basic plasmid backbones into one.

Figure 3A 1.Histag-TEV-HydA-Spytag in pACE(pACE-HydA-Tag in abbreviaFon/pladmid 1) Figure 3B 3.HydE in pDC(pDC-HydE in abbreviaFon/plasmid3) Figure 3C 4. HydF in pDK (pDK-HydF in abbreviaFon/plasmid4) Figure 3D 5. HydG in pDS(pDS-HydG in abbreviaFon/plasmid5)

Figure 3. The single plasmids to fuse by Acembl system. We obtained five sequence-confirmed single plasmids including the RBS, promoter region and loxP site. More detailed information about the sequence files could be seen on our wiki_parts.)

In particular, pACE is the “acceptor” plasmid with hydA sequence, while others are the “donor” plasmids with the auxiliary protein sequences. With one-step Cre recombination and subsequent transformation into BL21 or DH5a, we would obtain strictly fused plasmid with either all gene circuits integrated in one big plasmid or non-fused single plasmids. The screening of successful assembly involves different resistance (Ampicillin / Chloramphenicol / spectinomycin) and different kinds of origin. In pACE1, it has a replication origin that can be recognized by common DH5a or BL21. In pDC,pDS,pDk, it has a special origin (R6K gamma ori) can be recognized only by a mutation strain of E. coli. (PirHC or PirLC, which can express pir gene product for its replication.) Only a successful fusion into the acceptor plasmid can it propagate, using the accepters ori. Therefore, we efficiently put all four hyd sequences on one single plasmid, avoiding the potential problems imposed by the two-plasmid system.

The basis of our constructs, the four sequences, are not directly obtained from bacteriaBut they are all codon-optimized to ensure high-level expression. (The original sequences of hydrogenase are found on www.genome.jp.)

Results of cloning

As mentioned before, we basically relied on the Acembl system for hydrogenases gene cluster construction. In using the system, however, we can either fuse 4 single plasmids with one step of Cre recombination or do it step by step, integrating each plasmid one at a time. In order to gain higher success rate, we choose the second way.

First step:Fusion of plasmid 1/2 and plasmid 4

We fused pACE-Histag-TEV-HydA-Spytag/pACE-Histag-TEV-HydA-Spycatcher with pDK-HydF together as the first step. To test if we successfully fused the two, we use single restricted endonuclease digestion of XhoI. The restriction gives two bands on a 1% TAE Gel, in accordance with the band predicted by SnapGene®.

Figure 4A Fusion of plasmid 1 and plasmid 4.

Single restricted-endonuclease digestion of Xhol in pACE-Histag-TEV-HydA-Spytag x pDK-HydF gives two bands. The left pic refers to expected results based on SnapGene® software prediction, with two bands at 5427bp and 2146bp, respectively. The right figure refers to the experimental results, which is in good agreement with the software prediction.

Figure 4B Fusion of plasmid 2 and plasmid 4.

Single restricted-endonuclease digestion of Xhol in pACE-Histag-TEV-HydA-Spycatcher x pDK-HydF gives two bands. The left pic refers to expected results based on SnapGene® software prediction, with two bands at 5427bp and 2455bp, respectively. The 2455bp is larger than 2146bp due to the larger SpyCatcher. The right figure refers to the experimental results, which is in good agreement with the software prediction.

Figure 4A/B shows that plasmid1/2 and 4 are successfully fused.

Second step:Fusion of plasmid in step one and plasmid 3.

We test through the selection of LB solid plate with three resistance, Ampicillin, Chloramphenicol, and kanamycin. Then we use single restricted endonuclease digestion of XhoI. There should be two kinds of ways in fusing. Comparing our electrophoresis band with the prediction by SnapGene®, we confirmed the kind we obtained.

Figure 4C Fusion of the plasmid in step one(4A) and plasmid 3.

After the fusion of the plasmid in step one and plasmid 3, there will be one more enzyme restriction site of XhoI. Single restricted-endonuclease digestion of Xhol in pACE-Histag-TEV-HydA-SpyTag x pDK-HydF x pDC-HydE gives two bands. The left pic refers to expected results based on SnapGene® software prediction, with three bands at 5427bp, 2897bp and 2249bp, respectively. The right figure refers to the experimental results, which is in good agreement with the software prediction.

Figure 4D Fusion of the plasmid in step one(4B) and plasmid 3.

After the fusion of the plasmid in step one and plasmid 3, there will be one more enzyme restriction site of XhoI. Single restricted-endonuclease digestion of Xhol in pACE-Histag-TEV-HydA-SpyCatcher x pDK-HydF x pDC-HydE gives two bands. The left pic refers to expected results based on SnapGene® software prediction, with three bands at 5427bp, 2897bp and 2558bp, respectively. The right figure refers to the experimental results, which is in good agreement with the software prediction.

Figure 1C/D shows that plasmids obtained in step 1 and plasmid 3 are successfully fused.

Final step:Fusion of plasmid in step 2 and 5.

This fusion was conferred many possibilities due to the multiple loxP sites that are potentially recognized by Cre, and the fact that some fused loxP sites are reversely separated. However, since the plasmid in step 2 and plasmid 5 are put into the reaction in equal molar, the fully fused plasmid has a better chance. In parallel, we mixed four (plasmid 1/2, 3, 4, 5) plasmids together. After characterization by endonuclease restriction, we obtained the final plasmid. In addition, we find that the mixing of four in one reaction is not efficient.

Figure 4E Fusion of the plasmid in step (4C) and plasmid 3.

For a whole fused plasmid, It becomes hard to analyze it with just Xho I single enzyme. The bar at 3k actually accounts for two bars, with a separation of 20bp. In the picture, although the four bands predicted by SnapGene® can be found on our real gel, it is less clear.

Given the inconvenience with testing by restriction, we turned to resistance screening. The result is that it is resistant to four antibodies (Ampicillin, Chloramphenicol, kanamycin and Spectinomycin). Figure 4E shows that plasmids obtained in step 2 and plasmid 4 are successfully fused. Thus, we obtained a plasmid with all four subunits, HydA, HydE, HydF, HydG, fused together. The next step is to induce the expression of the hydrogenase.

Hydrogenases Expression and Enzyme Activity Assay

(1) Principles and Methods

In the activity assay of the hydrogenase in producing hydrogen, we repeated three parallel experiments to test the activity and validated the repeatability of our rudimentary system. In each parallel experiment, the system goes through three periods of “light-on and light-off”. The results (see below) shows the stability of the system and the reversible catalytic activity of the hydrogenase of the reaction, 2H+ + 2e- ⇿ H2 .

The three parallel systems consist of E. coli with engineered hydrogenase (wet weight 100ug) resuspended in PBS, 200ul quantum dots/nanorods (7.72*10^-9 M) resuspended in PBS, 150Mm NaCl, 100mM VitaminC, and mediator solution (5mM Paraquat dichloride, for mediating the electrons across the cell membrane). The whole solution including bacteria is adjusted to pH=4 by 100mM Tris-HCl(pH=7.0), given that the pH of 4 was reported to be an optimal environment. Prior to the assay, the E. coli was induced with IPTG overnight at room temperature. The whole system is based on former study.

In addition, we did a fourth assay with resuspended microspheres covered with quantum dots/nanorods bound biofilm in PBS, in place of the resuspended quantum dots/nanorods solution. The fourth set is the actual system we are proposing, since it is as efficient and allows the recycling of quantum dots/nanorods.

In our experiment, we find that despite the reported affected catalytic ability of FeFe hydrogenase due to oxygen, non-strict anaerobic and short-term exposure to oxygen does not cause detrimental effects on the enzyme activity of producing hydrogen. This can be explained by the high catalytic ability and the segregation layer from the atmosphere provided by the hydrogen it produces. Meanwhile, the electron sacrificial agent VitaminC also adds to the “protection layer” of the hydrogenase in our system.

(2) Instrument

Figure 5 Apparatus of the hydrogen production assay.

It contains (1) an anaerobic reaction container which is a transparent circular cuvette that allows light to go through; (2) a light source in our hydrogen production assay acting as a substitute for the real sun. (We chose a high-power white LED light, set 28cm away from the reaction container for a even distribution of photons); (3) a hydrogen electrode linked to its inner sensor inserted into the reaction container to measure the realtime concentration of hydrogen; (4) a date hub; (5) a computer connected to the hub to record the data and generate the curve of concentration variation within a period of time.

Results

a) Contribution of each component of the hydrogen production system

The first hydrogen production data using our system is the pink curve (curve 1) in Figure 5. It shows that lighting can induce hydrogen production in a closed system with nano rods (NR), mediator Methyl Viologen, and IPTG-induced bacteria transformed with fused plasmid. To prove that every element of the system is necessary and that it is our hydrogenase that produced the hydrogen rather than NR, we conducted a series of experiments.

To see whether NR is necessary and whether the hydrogen is produced by the reaction between NR and water under lighting rather than our hydrogenase, we conducted the experiment where the system does not contain nano rods or contain only nano rods. The data is summarized in Figure 5A. The red curve (curve 2) represents the system with the transformed bacterial suspension but without nano rods (NR). The flat curve shows that the system without NR could not produce hydrogen with light; NR is necessary for the system. The black curve (curve 3) represents a system in which only NR and mediators are present, with no bacteria. The flat curve shows that it could not produce hydrogen, which proves that the elements of the bacteria is necessary in the synthesis of hydrogen.

Figure 6

click to enlarge the figure

Hydrogen production evolution curve (Sensor Data/ Hydrogen amount vs Time) with different components. The pink curve (curve 1) in all pictures is the hydrogen production with all the components, nano rods (NR), IPTG induction, and the bacteria transformed with our hydrogenase plasmid. The rest are data with one or two components missing. In particular, data in the integrated picture are categorized into Figure 6A and 6B. Figure 6A shows the system with or without nano rods or with nano rods alone, and Figure 6B represents the system with or without induction. The curve 3 in each of the specific figure is the blank control with not transformed E. coli BL21. This series of experiments show that only when both nano rods (NR) and IPTG-induced transformed bacteria are present can the system produce hydrogen in a stable way.

Another step in proving that it is that the hydrogenase is indeed responsible for hydrogen production is to contrast the production level between the induced and un-induced bacteria suspension. The experiment we conducted are summarized in Figure 6B In this set of experiment, the blue line (curve 4) acts as our blank control. In this group, we use the wild type BL21 cells without plasmid. Although we can see a positive oscillation during a short time in the curve, the production was not at high rate and is likely due to the native hydrogenase in E. coli. The green curve (curve 5) represents the transformed bacterial with no induction of IPTG after 12h cultivation. The flat curve shows that it could not produce hydrogen, which proves that the induction of the hydrogenase expression is necessary. To further confirm, we did another experiment using bacteria that have grown 36 hours with no induction. The purple curve (curve 6) clearly contrasts the induced BL21 and the non-induced one. With curve 4 to 6, we have demonstrated that, with the help of NR, it was our hydrogenase in the system that produced the hydrogen we detected.

b) Bidirectional catalytic property of [FeFe] hydrogenase

As mentioned earlier, hydrogenase catalyzes the reversible oxidation of molecular hydrogen (H2). Thus, when we “turn off” the production mode, we should be able to see the consumption of hydrogen by hydrogenase. In testing this bidirectional catalytic property, conducted an experiment where we turned on and turned off the light alternately. The data is shown below in Figure 7. During lighting period, the hydrogen production increases, until we shut off the light at points that correspond to the tips. The curve then goes downward, showing that the hydrogen concentration is lowered, an evidence of the consumption of hydrogen. It is noteworthy that the hydrogenase shows the greatest production rate at the beginning of lighting: a transient sharp rise can be observed at the valleys. It is also obvious that each period of “light-on light-off” gives similar curves, which implies that our hydrogenase is stable.

Figure 6 Verifying the bidirectional catalytic property of [FeFe] hydrogenase.

During the period under lighting, the hydrogen production increases, until we shut off the light at points that correspond to the tips. The curve then goes downward, showing that the hydrogen concentration is lowered, an evidence of the consumption of hydrogen.

c) Hydrogen production with nano rods suspension replaced by nano rods bound to biofilm beads.

Given the difficulty in recycling the nano rods due to their small size, we utilize biofilm to immobilize nano rods and aggregate them into larger assemblies that allow filtration or other ways of recycling including centrifugation. However, testing whether the NR aggregate work in our system is needed. We conducted experiments with nano rods suspension replaced by nano rods bound to biofilm beads. The biofilm, whose subunit was CsgA engineered with HisTag on N-termial and SpyCachter-HisTag on C-terminal, was grown on microspheres, 25 micrometers in diameter for 48 hours. NR’s were then added and given 30 min to bind to the HisTag on CsgA subunit. (The engineered SpyCatcher was used for future pure hydrogenase binding.) The solution was centrifuged and the sediments contained biofilm beads covered with NR. This sediment was resuspended in PBS and was added to the reaction system. The data is in Figure 7 In this experiment, we did the same “light-on light off” actions to the system and the pattern is similar to the one with NR suspension (Figure 6) During lighting, the rapid production of hydrogen can be clearly observed. Some other characteristics pertain, such as the sharp rise at the beginning of lighting.

Figure 8 Hydrogen production with nano rods suspension replaced by nano rods bound to biofilm beads.

We replaced the nanorods suspension with nano rods bound to biofilm beads. During the period with lighting, the hydrogen production increases, until we shut off the light at points that correspond to the tips. The curve then goes downward, showing that the hydrogen concentration is lowered, an evidence of the consumption of hydrogen, as in Figure 7.

Comparing Figure 6 and Figure 7

In the process of hydrogen generation, a stir bar with a necessary speed of 800 RPM was used to generate the curve in Figure 7. But in Figure 8, a stir bar was not used. It is likely because the aggregates of NR have a bigger chance in colliding with E. coli to transfer electrons. We therefore propose this model as our final model, although further optimization of the system is still under way.

d) Repeatability

Figure8. Repeatability of our hydrogen production system. The three curves conform. This demonstrates our preliminary prototype as repeatable and robust.

To further prove our system as a reliable one, we did three sets of hydrogen production assays in one day in a row (Figure8, curve 10-12). The system was mainly made up of resuspended CdS nonorods, E. Coli BL21 transformed with the plasmid containing all the four [FeFe]hydrogenase subunits from Clostridium. acetobutylicum, and mediator, methyl viologen. (See Figure5 or the section of Principles and Methods for details of the reaction system.) From the data shown, we clearly see the conformation of the three curves. This demonstrates our preliminary prototype in Figure5 as repeatable and robust.

Reference

[1]Madden C, Vaughn MD, Díez-Pérez I, Brown KA, King PW, Gust D, Moore AL, Moore TA (January 2012). "Catalytic turnover of [FeFe]-hydrogenase based on single-molecule imaging". Journal of the American Chemical Society. 134 (3): 1577–82. doi:10.1021/ja207461t. PMID 21916466.

[2] Smith PR, Bingham AS, Swartz JR (2012). "Generation of hydrogen from NADPH using an [FeFe] hydrogenase". Int. J. Hydrogen Energy. 37: 2977–2983.

[3] Madden C, Vaughn MD, Díez-Pérez I, Brown KA, King PW, Gust D, Moore AL, Moore TA (January 2012). "Catalytic turnover of [FeFe]-hydrogenase based on single-molecule imaging". Journal of the American Chemical Society. 134 (3): 1577–82. doi:10.1021/ja207461t. PMID 21916466.]

[4] Honda, Y., Hagiwara, H., Ida, S., & Ishihara, T. (2016). Application to photocatalytic h 2, production of a whole-cell reaction by recombinant escherichia coli, cells expressing [fefe]-hydrogenase and maturases genes. Angewandte Chemie.

[5] Bieniossek, C., Nie, Y., Frey, D., Olieric, N., Schaffitzel, C., & Collinson, I., et al. (2009). Automated unrestricted multigene recombineering for multiprotein complex production. Nature Methods, 6(6), 447-450.

[6] King, P. W., Posewitz, M. C., Ghirardi, M. L., & Seibert, M. (2006). Functional studies of [fefe] hydrogenase maturation in an escherichia coli biosynthetic system. Journal of Bacteriology, 188(6), 2163-72.

[7] Cao Y, Bai X F. Progress in Research of Preparation of Loaded Nano-CdS and H_2 Production by Photocatalytic Decomposition of Water[J]. Imaging Science & Photochemistry, 2009, 27(3):225-232.

[8] Honda Y, Hagiwara H, Ida S, et al. Application to Photocatalytic H2, Production of a Whole-Cell Reaction by Recombinant Escherichia coli, Cells Expressing [FeFe]-Hydrogenase and Maturases Genes[J]. Angewandte Chemie, 2016

Team:ShanghaitechChina/Hydrogen