Cloning & Expression of sMMO

The soluble methane monooxygenase (sMMO) from Methylococcus capsulatus is an enzyme complex capable of oxidizing the greenhouse gas methane to methanol. The complex consists of three main components all of which are required for proper enzyme function. The hydroxylase component itself is made up of the three subunits α, β and γ which are arranged as a heterodimer (α₂β₂γ₂) (Figure 1). The α subunit houses the active site of the enzyme where the hydroxylation of methane takes place. The reductase component MMOR oxidizes NAD(P)H and transfers the electrons to the hydroxylase. The electron transfer is facilitated by a [2Fe-2S] cluster and a FAD cofactor bound to the reductase. It is assumed that the third component, the regulatory protein MMOB, has a function in modulating electron transfer as it binds to the same site of MMOH as MMOR [1]. Another open reading frame within the mmo operon encodes MMOD whose function has not yet been clearly determined [2,3]. Despite this uncertainty, we included all components of the mmo operon in our cloning strategy.

Growth kinetics of the MMO host organism Methylococcus capsulatus

For the purpose of ensuring a significant degradation of methane in the cow’s rumen, our organism has to meet several requirements: It has to be well characterized to simplify the laboratory handling while ensuring low safety risks. Furthermore, the ability to grow fast in the respective environment and high levels of heterologous protein expression have to be given. Another aspect to consider is the ubiquitous occurrence in the cow´s rumen.
To evaluate the benefit of using Escherichia coli as a heterologous expression system for sMMO instead of sticking with the natural methanotroph M. capsulatus we decided to compare their growth rates using a cultivation curve at the respective optimal cultivation conditions.
As shown in figure 2, E. coli does not only reach a much higher cell density in comparison to M. capsulatus (OD₆₀₀ = 7 and OD₆₀₀ = 1, respectively) but furthermore shows a much faster growth. The time span for the organism to reach the exponential stadium is of particular interest for the choice of the organism. Whereas M. capsulatus reaches the exponential stadium after 20 h, E. coli does after only 3 h .

The difference in growth is caused by the differences in carbon metabolism. While E. coli is able to metabolise a wide range of carbon sources, M. capsulatus depends on C1 compounds as the sole carbon source, which results in a slower metabolism linked with decreased protein biosynthesis. To degrade a significant amount of methane in the cow’s rumen, it is necessary to choose an organism capable of producing the sMMO in high amounts. Furthermore, E. coli is a natural resident of the rumen [4], the best-characterised model organism for synthetic biology and the standard chassis for the iGEM competition. These factors facilitated the choice for E. coli as a heterologous expression system. Therefore, we provided E. coli with all necessary mmo genes from M. capsulatus.

Cloning

The first step of our experimental procedure was to isolate the genes for all the six subunits of the enzyme complex from M. capsulatus (Bath) - mmoX, mmoY, mmoZ (coding for subunits α, β and γ, respectively), mmoB (coding for the regulatory protein), mmoC (coding for the reductase) and mmoD (Table 1). We chose the vector pSB1A3 in view of our future plans to coexpress the subunits with chaperone proteins which were contained on vectors with chloramphenicol resistance. Thus the use of the shipping vector pSB1C3 would not have been a practical approach.

We put each of the genes under the control of the inducible lacl promoter (R0011) which has a low leakiness and is easily inducible with IPTG. This strong, frequently used promotor is well-characterised in the iGEM registry and is reported to be well functioning. We also added the extensively documented weak ribosome binding site (B0032) with a strength of 33.96 % compared to B0034 to diminish the formation of inclusion bodies and a double terminator (B0015). Furthermore, a His-tag was added to the C-terminus of each subunit to simplify the identification and purification process.
Additionally we also constructed a vector (final construct) which included all six subunits, each under control of the above-mentioned promoter and RBS. Compared to the natural cluster of M. capsulatus our final construct does not include the naturally occurring transposases and hypothetical proteins. However, the arrangement order of the subunits is the same as in the host organism.
Next, we tested expression of the different subunits in E. coli strain XL1-Blue MRF’ by Western Blot analysis using anti-His antibodies for detection. We separated the soluble and insoluble fractions and analyzed them separately. However, expression could only be shown in two cases, of which MMOC had mostly been produced solubly whereas MMOX had been produced in inclusion bodies (Figure 3).

Under natural conditions, protein folding is facilitated by chaperone proteins. This ability of molecular chaperones has also been taken advantage of for research purposes, for example in the case of the protein Cryj2 which gained stability if coexpressed with certain sets of chaperones [5]. Therefore, we decided to use the TaKaRa Clontech Chaperone Plasmid Set and coexpress the contained chaperones with the MMO subunits. This kit has previously been used in combination with the E. coli strains JM109 and BL21(DE3) for the expression of a different monooxygenase [6].
Expression of the chaperone plasmids (designated C1-C5, Table 2) in these two strains was tested via SDS-PAGE (Figure 4).

Expression of chaperones in E. coli JM109 resulted in more distinct bands corresponding to the expected molecular weights in comparison to strain BL21(DE3) (Figure 4, cf. Table 2). However, in some cases the expression levels differed substantially even if the same chaperone protein was encoded on the plasmid. In the case of plasmid C2 in E. coli JM109 the two expected bands at about 10 and 60 kDa are clearly visible, but in the case of C3 the band at 60 kDa is much less distinct.
We continued working with both strains and cotransformed them with the different chaperone plasmids and the different subunits (except MMOC which had already been expressed without the aid of chaperones, see Figure 3). With E. coli BL21(DE3) no cell growth occurred which might be due to high metabolic stress caused by maintenance of two plasmids. This observation is supported by literature data [6]. In contrast, E. coli JM109 grew normally allowing Western blot analysis for detection of the subunits followed by immunostaining.
The use of chaperone plasmid C1 resulted in partly soluble expression of the subunits MMOX, MMOY and, to a smaller degree, MMOZ. No bands corresponding to the other subunits were visible (Figure 6, upper panel). A slightly different result was obtained for the subunits coexpressed with chaperone plasmid C3. In this case, all subunits except MMOD could be detected. However, bands corresponding to subunits MMOB and MMOZ were still very weak (Figure 6, lower panel)

In contrast, using chaperone plasmid C2 resulted in expression of all MMO subunits. Although a substantial amount was still detected in inclusion bodies these results indicate soluble expression of all His-tagged components of the enzyme complex (Figure 6). Therefore, we used E. coli JM109 cotransformed with chaperone plasmid C2 and our final construct encoding all six MMO subunits for further experiments and activity assays.

The combination of GroEL and GroES led to synthesis, of all subunits in soluble form suggesting a functional similarity to the chaperone genes contained in the M. capsulatus genome. A closer look at the sequence of the mmo operon reveals that the operon is followed by an open reading frame whose sequence shows similarity to those of chaperones [7]. This proximity indicates a role of the respective chaperone in folding of the mmo gene products. Based on this, further research could be aimed at creating a library of presumed chaperone genes from M. capsulatus. These chaperone proteins could then be coexpressed with the sMMO subunits in order to find optimal expression conditions.
An important question for possible applications of E. cowli is whether its growth is inhibited by the two plasmids. Therefore, we tested E. cowli in comparison to the strain only carrying chaperone plasmid C2. The growth curves (Figure 7) show that cotransformation in fact results in a more distinct lag phase, probably due to greater replication stress owing to the additional plasmid. However, as the bacteria reach the stationary phase both systems grow to the same OD. Therefore, the cotransformation does not seem to limit the final OD of the cultures and no toxic substances are involved.

References

1. Wang W, Iacob RE, Luoh RP, Engen JR, Lippard SJ (2014) Electron Transfer Control in Soluble Methane Monooxygenase. J Am Chem Soc 136:9754-62
2. Merkx M, Lippard SJ (2001) Why OrfY? Characterization of MMOD, a long overlooked component of the soluble methane monooxygenase from Methylococcus capsulatus (Bath). J Biol Chem 277:5858-65
3. Semran JD, Jagadevan S, DiSpirito AA, Khalifa A, Scanlan J, Bergman BH, Freemeier BC, Baral BS, Bandow NL, Vorobev A, Haft DH, Vuilleumier S, Murrel JC (2013) Methanobactin and MmoD work in concert to act as the “copper-switch” in methanotrophs. Environ Microbiol 15:3077-86
4. Irle A (2011) Untersuchungen zum Einfluss von Clostridiengaben bei Grassilagen mit auffällig niedrigen Reineiweißanteilen auf die Pansenfermentation (in vitro). Dissertation, Tierärztliche Hochschule Hannover
5. Nishihara K, Kanemori M, Kitagawa M, Yanagi H (1998) Chaperone Coexpression Plasmids: Differential and Synergistic Roles of DnaK-DnaJ-GrpE and GroEL-GroES in Assisting Folding of an Allergen of Japanese Cedar Pollen. Appl Environ Microbiol 64:1694-9
6. Rehdorf J, Kirschner A, Bornscheuer UT (2007) Cloning, expression and characterization of a Baeyer-Villiger monooxygenase from Pseudomonas putida KT2440. Biotechnol Lett 29:1393-8
7. Blazyk JL (2003) Electron Transfer and Protein Engineering Studies of the Soluble Methane Monooxygenase from Methylococcus capsulatus (Bath). Dissertation, Massachusetts Institute of Technology

Delivery

The next step was to evaluate whether E. coli is able to survive in environments such as the rumen. Therefore, we tested growth of E. coli on rumen fluid, which was kindly provided by one of our cooperation partners. That way we were able to conclude whether rumen fluid is a sufficient medium for growth of E. coli or if it would inhibit its growth in any way. The results (Figure 1 A-F) show that the microorganisms of the cow’s natural microbiota were able to grow on every plate and that the degree to which the growth occurred increased with the rumen fluid content in the agar plate. Supplementation with glucose did not seem to have any substantial effect. In contrast, E. coli did not form colonies on rumen fluid. The lack of a carbon source can be excluded as a reason for this finding as the same result was obtained for plates supplemented with glucose. Furthermore, the pH value in each case was about 6.5. It has been shown that at this pH value the growth of E. coli is inhibited by 69 % owing to volatile fatty acids (VFAs) present in rumen fluid [1]. Such VFAs could also be responsible for the growth inhibition observed here. Another factor possibly leading to inhibited growth in future application but not tested in our experiments are ciliates feeding on bacteria like E. coli [2].

Due to this result we had to consider possibilities to introduce E. co(w)li into the rumen without harming the cells. In the past, encapsulation techniques have often been applied in cases when free probiotic cells could not survive to a sufficient degree in the desired environment [3]. One of the most widely used substances for encapsulation is alginate. Alginate beads could represent a low-priced and easy-to-use way to introduce the bacteria into the cow’s rumen without exposing the cow to a health risk while simultaneously protecting E. coli from getting in contact with the harmful surrounding [4]. In the course of our experiments, we produced beads with a diameter of about 6 mm (Figure 2). To provide sufficient mechanical stability, the alginate concentration was adjusted.

However, the encapsulated bacteria still need to be able to take up methane. To demonstrate that methane could permeate the beads’ alginate structure, we dyed one portion of alginate beads with the food colorant Patent Blue V and another with carmine. The two portions were mixed and, after incubation overnight, had changed their colour mostly to purple, indicating that the two colorant molecules had diffused between the beads of different colors (Figure 3). As the alginate beads consist of water to a high percentage and methane is soluble in water it can be assumed that methane can readily enter the alginate beads.

To evaluate the survival rate of E. cowli in rumen fluid, we determined the plating efficiency for the supernatant and dissolved beads. We entrapped equal numbers of bacteria in alginate beads, incubated them for 24 h in 2YT medium or rumen fluid, and separated the beads from the supernatant, i.e. the medium in which the beads had been cultivated. The beads were dissolved in PBS.
In our analysis we normalised the number of cells grown in rumen fluid to the number of cells grown in 2YT medium, representing ideal culture conditions. The ratio therefore, poses an indicator as to how many cells survived in rumen fluid compared to ideal conditions. We set the value obtained for non-encapsulated cells as 1 and treated the results obtained for encapsulated cells and the supernatant in respect to this value. The higher the calculated ratio is, the more cells survived in rumen fluid in comparison to 2YT medium under the given conditions.
Bacteria encapsulated in beads survived to a much higher degree than those cultivated in rumen fluid without beads (Figure 4). Unexpectedly, the number of survived cells in the supernatant outreached the non-encapsulated cells. One possible explanation of this finding is that the beads serve as a reservoir for the cells. As the cells populate the beads they can leave the protecting beads and become non-encapsulated cells, which are fully exposed to the perils of the rumen fluid. These non-encapsulated cells cannot survive as long as those cells still entrapped in the beads. But as more and more bacteria can grow inside the beads and thus may leave them, the number of cells surviving in the supernatant is higher than the number of those which were non-encapsulated from the beginning. Another explanation would be partial disintegration of the beads by phosphate compounds. In a possible application and in view of safety issues this might be prevented by further coating of the beads, for example by adding polyethylene diamine. This was done by the iGEM Team Paris Bettencourt for their project in 2012 [5].

In conclusion, non-encapsulated E. cowli does not seem to be able to grow in rumen fluid (Figure 1). However, bacteria immobilized in alginate beads showed a higher survival rate possibly due to a better nutrition supply [6]. The alginate matrix hypothetically protects the bacteria from ciliates feeding on them [2]. Moreover, the beads are easier to administer to the animal in comparison to a bacterial suspension. One could also consider modifications of the alginate beads, for example by including a buffer substance providing an appropriate pH value (see modeling) which does not inhibit growth of the cells. Another possible supplement is paraffin oil which has been shown to increase methane uptake and could thus also be incorporated into the beads [7]. Furthermore, the beads could be produced using a common growth medium providing the cells with nutrients.

Reference

1. Wolin, MJ (1969). Volatile fatty acids and the inhibition of Escherichia coli growth by rumen fluid. Appl Microbiol, 17, 1:83-7.
2. Irle A (2011) Untersuchungen zum Einfluss von Clostridiengaben bei Grassilagen mit auffällig niedrigen Reineiweißanteilen auf die Pansenfermentation (in vitro). Dissertation, Tierärztliche Hochschule Hannover
3. de Vos P, Faas MM, Spasojevic M, Sikkema J (2010) Encapsulation for preservation of functionality and targeted delivery of bioactive food components. Int Dairy J 20:292-302
4. Chávarri M, Marañón I, Villarán C (2012) Encapsulation Technology to Protect Probiotic Bacteria. In: Rigobelo EC, ed. (2012) Probiotics. ISBN: 978-953-51-0776-7
5. https://2012.igem.org/Team:Paris_Bettencourt/Encapsulation Date:16.10.2014
6. Qi WT et al. (2006) Behavior of microbial growth and metabolism in alginate–chitosan–alginate (ACA) microcapsules. Enzyme Microb Technol 38:697-704.
7. Jiang H et al. (2010) Methanotrophs: Multifunctional bacteria with promising applications in environmental bioengineering. Biochem Eng J 49: 277-88

Activity

The probably most important aspect of our project was to proof the functionality of the sMMO complex in E. cowli. Therefore, activity tests based on the degradation of methane by our modified bacterium were performed. With this measuring setup, we could gather experimental data on the time-dependent methane degradation. Calibration with defined concentrations of methane revealed linear dependency as shown in figure 1.

The measurements were performed with bacteria cultures of an OD_600nm value of 4 in NMS medium [1,2] and a wet biomass of 0.5 g for E. cowli or with cultures of an OD₆₀₀ value of 1 in NMS medium and a wet biomass of 0.2 g for M. capsulatus.
First, we performed an experiment to compare our engineered E. cowli to the strain JM109 and M. capsulatus and to determine the time span necessary for the following experiment (Figure 2). From this measurement, it becomes clear that our negative control (E. coli transformed with chaperone plasmid C2) showed a stable methane concentration in the first 6,000 seconds of the measurement. Both E. cowli and M. capsulatus showed a decrease of methane concentration during the same time span.

These results suggest that E. cowli is indeed able to degrade methane.
The difference in degradation rates of E. cowli and M. capsulatus is not indicative because of dissimilar wet biomass. Furthermore, it is obvious that degradation of methane does not stop after 6,000 sec. For that reason, we decided to extend the time span of the following measurements up to 8,000 sec and to characterize E. cowli in more detail.
E. coli solely transformed with chaperone plasmid C2 served as a negative control. The results obtained here show that E. cowli is capable of degrading methane even for a time span of 8,000 sec, although a slight flattening of the curve becomes visible from about 5,000 to 6,000 sec on.

To further approach the possible application of E. cowli we subsequently tested the methane degradation rates of bacteria encapsulated in alginate beads with an OD₆₀₀ of 4. Cultivation was carried out either in NMS medium or in rumen fluid to determine the influence of the medium on methane degradation. In both cases the methane concentration decreased, although the slope is not as steep as in the aforementioned experiments. There was only a slight difference between the decrease of methane concentrations in NMS medium and the decrease in rumen fluid (Figure 4), indicating that no inhibition is caused by cultivation in rumen fluid compared to cultivation in NMS-medium.

Finally, we standardized our previously obtained data in reference to the bio wet mass of M. capsulatus to compare all the different strains and conditions used. This includes the loss of enzyme activity due to the pH and the survival rate of bacteria inside the alginate beads. As expected, no methane degradation was observed in the negative control. E. cowli degrades a similar amount of methane as M. capsulatus (about 30 µM/s), whereas the degradation rate of encapsulated bacteria is decreased by about one third. The reason for this might be that the cultivation of the beads does not allow for proper mixing and thus leads to less diffusion of methane into the beads. Nevertheless, an activity independent from the surrounding medium could be shown.

M. capsulatus grows optimally in atmospheres containing 45 % methane according to [3] or 20 % according to [4]. The methane concentration in the rumen is about 27 % [5]. However, as the explosivity margin of methane is 5 % we decided to stick to a methane concentration of 2 % due to safety considerations. We took this discrepancy into consideration in our modeling approach.
This problem could be circumvented by including alternative activity assays into further investigations which do not rely on methane but instead make use of the broad substrate range of sMMO [6]. A convenient possibility would be the use of naphthalene [7]. A well-established naphthalene assay is based on a colour reaction which is easy to detect either by the naked eye or photometrically for quantitative analysis. Enzyme activity could also be detected indirectly by measuring a decrease in oxygen concentration. For quantification, of expressed enzymes antibodies directed against the different sMMO subunits could be developed. This would be a valuable contribution to the iGEM Registry as well as facilitate future research on the sMMO. Better quantification of the enzyme would also enable the determination of degradation rates.
To confirm the results obtained with the methane sensor other measurements such as gas chromatography could be performed in the future to examine the activity of the sMMO. Nevertheless, the results suggest functionality of our final construct and the encoded enzyme complex.

References

1. http://www.dsmz.de/microorganisms/medium/pdf/DSMZ_Medium1179.pdf
2. http://www.dsmz.de/microorganisms/medium/pdf/DSMZ_Medium632.pdf
3. http://www.ncimb.com/BioloMICS.aspx?Table=NCIMBstrains&Rec=2758&Fields=All
4. Pilkington SJ, Dalton H (1990) Soluble Methane Monooxygenase from Methylococcus capsulatus Bath. Methods Enzymol 188:181-90
5. Irle A (2011) Untersuchungen zum Einfluss von Clostridiengaben bei Grassilagen mit auffällig niedrigen Reineiweißanteilen auf die Pansenfermentation (in vitro). Dissertation, Tierärztliche Hochschule Hannover
6. Green J, Dalton H (1989) Substrate specificity of soluble methane monooxygenase. Mechanistic implications. J Biol Chem 264:17698-703
7. Canada KA, Iwashita S, Shim H, Wood TK (2002) Directed evolution of toluene ortho-monooxygenase for enhanced 1-naphthol synthesis and chlorinated ethene degradation. J Bacteriol 184:344-9
8. http://www.pololu.com/file/0J311/MQ4.pdf (16.10.2014)

Interlab Study

Measurement of Fluorescence

The samples were measured according to the guidelines of the manufacturer using the TECAN ULTRA microplate reader. The fluorescence measurements were performed in a qualitative way due to lack of a fluorescence standard.
Bacterial cultures were directly used for determination via fluorescence emission measurements. The cultivation media Luria-Broth (LB) and cells without fluorescent protein were used as standard and control, respectively.

Two different set-ups were used for measurements depending on the fluorescent marker. Due to variation in fluorescent protein expression gains had to be adjusted throughout the measurement. Thus, plain LB-medium was used for standardization. Configuration data is shown in table 2.

The fluorescence measurements were obtained as a triplicate. Therefore cultivation of E. coli up to the optical density (OD₆₀₀) needed was performed three times. A scheme and obtained measurements for green fluorescence and red fluorescence are shown in table 2 and figure 1, respectively. Furthermore, a calibration curve for red and green emission values using XL1-Blue-MRF’ cells was determined and is shown in figure 2 and figure 3.

The calibration curves were used to estimate the influence of the optical density on the fluorescence. Thus, the fluorescence caused by the medium was subtracted from the emission.

For determination, of the effects caused by various promoters, the measured emission data (minus the fluorescence caused by the media) was compared. For a comparative analysis between the experiments with different gains plain media was used for standardization. This was done for all three replicates. Data obtained from the triplicate is shown in figure 4 for GFP and figure 5 for mRFP1. Remaining schemes and emission data are shown in the appendix.

The observed results were expected due to variations in promoter and backbone. The pSB1C3 is described as a high-copy plasmid, whereas pSB3K3 is a low-copy plasmid. High-copy plasmids occur with a high number of copies inside the bacterial cell. This can be observed in fluorescence emission data wherein the construct GFP0, consisting of a low-copy plasmid with a stronger constitutive promoter showed substantially lower amounts of GFP expression compared to GFP1, which consisted of a high-copy plasmid with a strong constitutive promoter (figure 4). The same tendency could be observed for RFP (figure 5). RFP0 being GFP0 red fluorescent counterpart also expressed much lower amounts of RFP compared to RFP1 consisting of a high-copy plasmid, and RFP driven by a strong constitutive promoter. Beside the backbone, the promoter as well had a strong effect on protein expression. The fluorescence emission being relative to the amount of expressed protein showed higher values in constructs with GFP or RFP under the control of a strong constitutive promoter rather than a weak constitutive promoter. Therefore, protein expression was dependent on two main factors: (a) strength of the promoter and (b) copy number of the plasmid. Both affected the fluorescent protein yield substantially. The comparative analysis furthermore showed that a weak promoter resulted in fluorescent protein expression than a low-copy plasmid. Thus it can be concluded that the right choice of promoter is inevitable for high protein yields.

Comparison of GFP and RFP – Extra Credit

For a comparative analysis of GFP and RFP expression both fluorescent proteins were tested under the same conditions. Therefore, we first constructed all correspondent analogues with RFP instead of GFP as protein of interest. Furthermore, the fluorescence emission of GFP and RFP were determined after identical cultivation and measurement conditions. As shown in figure 5 observations similar to GFP were made for RFP expression. The fluorescence emission and thus RFP yield is affected by the promoter strength as well as varying backbone. A high-copy backbone as well as a strong promoter (RFP1) resulted in the strongest red fluorescence. The copy number of the plasmid had a similar effect as already observed in GFP expression and led to a lower RFP yield. The strongest effect was caused by the promoter strength. The weak constitutive promoter led to the lowest fluorescence emission in RFP as well as GFP expression.