Results
In our project, we aimed to create a library of genes encoding for cytochrome P450 enzymes (CYPs), for production in Chlamydomonas reinhardtii. We successfully expressed five P450 genes of different origin and tried to exemplarily detect the activity of one of them. Successfully expressed genes include those encoding human NADPH cytochrome P450 oxidoreductase (POR), two human P450 enzymes (3A4 and 2D6), CYP enzyme 101 (CYPCamC) from Pseudomonas putida, and 9Q3 from the oriental honeybee Apis cerana cerana. A gene encoding the CYP enzyme CYP81A10V7 from the sweet grass Lolium rigidum was also designed and tested.
We investigated different approaches to produce these enzymes in Chlamydomonas reinhardtii and to measure their activity. Use the navigation bar on the left side to navigate through our results to get to whatever part you are interested in.
Our MoClo Collection
During the experimental iGEM phase we generated six level 0 parts (being basic parts), nine level 1 parts (intermediate parts/transcriptional units) and 25 level 2 parts (functional multigene constructs) all of which are standardized and compatible with the Modular Cloning System (MoClo) so that they can be integrated seamlessly. Further, these parts are codon optimized and contain RBCS2-introns in their coding sequence to ensure optimal expression in Chlamydomonas. Overall, we contribute 39 distinct parts to the iGEM community. All our level 2 constructs shown in Figure 1 were transformed into Chlamydomonas reinhardtii and analyzed for expression.
(a) A visual explanation of the MoClo system with Level 0, 1, and 2 parts (Weber et al., 2011). (b) These level 2 constructs were designed according to the standards of the Modular Cloning System (Phytobrick) and are codon optimized for Chlamydomonas reinhardtii. We employed resistance cassettes (Hygromycin (HR), Paromomycin (HR) and Spectinomycin (HR)) for selection of positive transformants. The numbers 1-11 represent specific MoClo fusion sites. To create multigene constructs, we used Dummies (D61 and D63) and end linkers (El37 and El39) to connect our parts with the destination vector. For a more detailed part description check out our parts registry, especially our best basic part.
Cytosolic Expression
P450 enzymes are produced in the cytoplasm and integrate into the ER membrane. Therefore, our first approach was to produce the different P450 enzymes in the cytoplasm, hoping that they would integrate into the ER membrane of our microalgal host just as they do in their source organism. To accomplish this, we introduced constructs A-I (Fig. 1) into the Chlamydomonas strain UVM4, which is a strain suitable for high transgene expression and is easily transformable. These constructs were designed for producing cytosolic proteins harboring C-terminal 3xHA or 3xFlag tags for detection via immunoblotting.
(1a-7a) Level 2 MoClo constructs for the production of enzymes CYP3A4, 9Q3, 2D6, the POR and CYPCamC containing either the FLAG or HA-tag were designed. (1b-7b) The UVM4 strain was transformed with the construct in (a). Up to 30 antibiotic-resistant transformants (depending on the construct) were cultivated in TAP-medium and samples taken after 3 days. Whole-cell proteins were extracted and analyzed by SDS-PAGE and immunoblotting using an anti-HA antibody. In the resultant blot, the black arrow marks our enzymes and the white arrow marks a cross reaction of antibodies. Signals were obtained for CYP3A4 (~57 kDa), CYP 9Q3 (~59 kDa), CYP2D6 (~ 56 kDa), the POR (~ 77 kDa) and CYPCamC (~ 47 kDa). The UVM4 recipient strain was used as negative control and a strain expressing the HA-tagged ribosomal chloroplastic 50S protein L5 (RPL5) or FLAG-tagged VIPP1 as positive control.
Notably, we managed to produce five out of six enzymes (Fig.2), with highest expression levels obtained for CYP3A4 and the reductase (POR).
Chloroplast-targeted enzymes
Another idea was to target the proteins encoded on constructs T-W (Fig.1) to the chloroplast to use electrons from the light reactions to photosynthesis to fuel our enzymes. Constructs for CYP3A4, CYP2D6, CYPCamC and the POR employed an N-terminal transit peptide to target the proteins to the chloroplast and a C-terminal 3xHA tag for detection via immunoblotting.
(1a-4a) Level 2 MoClo constructs for the production of CYP3A4, 2D6, POR, and CYPCa,C with the CTPPSAD-transit peptide. (1b-4b) The UVM4 strain was transformed with the construct shown in (a). Up to 30 antibiotic-resistant transformants (depending on the construct) were cultivated in TAP-medium and samples taken after 3 days. Whole-cell proteins were extracted and analyzed by SDS-PAGE and immunoblotting using an anti-HA antibody. The white arrow marks cross-reactions of the antibody. No specific signals were detected for CYP2D6 (~ 56 kDa), CYP3A4 (~ 57 kDa), POR (~ 77 kDa) and CYPCamC (~ 47 kDa). The UVM4 recipient strain was used as negative control and a strain expressing HA-tagged ribosomal chloroplastic 50S protein L5 (RPL5) as positive control.
Unfortunately, we were not able to detect signals for any of our enzymes targeted to the chloroplast. This could be attributed to the unnatural environment for the enzymes leading to their degradation or to inefficient chloroplast import. Alternatively, the 3xHA-tag might be cleaved off in the chloroplast, which our lab has frequently observed for other chloroplast-targeted proteins. Since cytosolic expression was successful, we focused on the cytosolic enzymes in further experiments.
The reductase
To optimize the supply of reducing equivalents to our CYPs, we sought to co-express them with a reductase. Our initial approach involved designing a multi-gene construct (X in Fig. 1) for producing CYP3A4 together with a POR.
(a) Level 2 MoClo multigene construct for the co-expression of CYP3A4 and POR, both containing 3xHA tags. (b) The UVM4 strain was transformed with the construct shown in (a). 30 hygromycin-resistant transformants were cultivated in TAP-medium and samples were taken after 3 days. Whole-cell proteins were extracted and analyzed by SDS-PAGE and immunoblotting using an anti-HA antibody. The white arrow marks a cross-reaction of the antibody. CYP3A4 (~ 57 kDa) and POR (~ 77 kDa) could both be detected. The UVM4 recipient strain was used as negative control and a strain producing HA-tagged ribosomal chloroplastic 50S protein L5 (RPL5) as positive control.
To address this problem, we pursued an alternative approach known as a “supertransformation”. Here, previously transformed positive CYP3A4 strains were again transformed with the reductase gene (POR). Advantages of this approach were the ability to pre-select strains with known CYP3A4 expression levels and a higher probability to find a strain which shows high expression for both genes, since genes on tandem constructs (multi-gene constructs) are expressed independently of each other. The enzymes were once again designed with a C-terminal 3xHA-tag for detection.
(a) (a) Level 2 MoClo constructs for the production of CYP3A4 and POR, both with a C-terminal 3xHA tag. (b) A CYP3A4-producing strain was transformed with the POR construct in (a). Up to 30 hygromycin-resistant transformants were cultivated in TAP-medium and samples were taken after 3 days. Whole-cell proteins were extracted and analyzed by SDS-PAGE and immunoblotting using an anti-HA antibody. The white arrow marks a cross-reaction of the antibody. Production of CYP3A4 (~ 57 kDa) was observed, but POR (~ 77 kDa) could not be detected. The UVM4 recipient strain was used as negative control and a strain producing HA-tagged ribosomal chloroplastic 50S protein L5 (RPL5) as positive control.
As shown in Fig.5, this approach proved to be unsuccessful as our transformants showed lower levels of CYP3A4 than the original CP3A4-producing recipient strain. Possibly, the construct for CYP3A4-expression was silenced. In an interview with Dr. Hugues Renault from the University of Strasbourg, an expert on P450 in plants, we learned that overexpression of the reductase can lead to a disruption of the ER. To investigate if the production of POR is harmful for the ER in Chlamydomonas, we changed our focus on measuring activity of our CYPs without the support by a human P450 reductase.
Freeze Thaw
To be functional, CYPs must be correctly embedded within the membrane (Šrejber et al., 2018). To verify if our CYPs produced in Chlamydomonas are membrane incorporated, a freeze-thaw assay was performed, where rapid freezing and thawing leads to fragmentation of the cells. During this process, large membranes vesicles are formed which can be separated from soluble fractions by centrifugation. Soluble and membrane fractions were analyzed by immunoblotting for our strains producing 3xHA-tagged CYP3A4.
(a) Level 2 MoClo construct for the production of 3xHA-tagged CYP3A4. (b) The UVM4 strain was transformed with the construct shown in (a). An anti-HA antibody was used to detect CYP3A4 by immunoblotting (A). For controls, antibodies against cytochrome b6f (Cyt f) and CGE1 were used, which serve as markers for membrane and soluble proteins, respectively (B).
30 spectinomycin-resistant transformants were cultivated in TAP-medium and samples were taken after 3 days. For reference, whole-cell proteins were extracted and analyzed by SDS-PAGE and immunoblotting. Cells were separated into soluble (S) and membrane (P) fractions and analyzed by SDS-PAGE and immunoblotting. CYP3A4 (~ 57 kDa) was detected in the membrane pellet and not in the supernatant (S), confirming its membrane association. CYP3A4 positive strains E2, E13 and E26 were analyzed. The UVM4 strain was used as a negative control.
We were able to detect CYP3A4 in the membrane fraction (Fig.6), indicating correct membrane targeting that is essential for its activity.
Activity Assays
Since CYP3A4 was produced at highest levels in Chlamydomonas, we used CYP3A4-producing transformants to conduct first activity tests. CYP3A4 has many known substrates, one of which is erythromycin. If CYP3A4 was functional in Chlamydomonas, we expected our transgenic lines to exhibit better growth in the presence of erythromycin compared to the wild type, as Chlamydomonas is sensitive to erythromycin. We incubated our transgenic algae in TAP-medium containing 10 mg/L of erythromycin.
(a) Level 2 MoClo construct for the production of 3xHA-tagged CYP3A4. (b) CYP3A4-producing transformants E2 and E26 were incubated for up to 192 h in TAP medium containing the indicated concentrations of erythromycin (Ery). Cells were counted once a day. The UVM4 strain was used as negative control.
The growth curves shown in Figure 7 suggest that one CYP3A4-producing strain (E2) is more sensitive to erythromycin and one is less sensitive (E26). However, this does not necessarily indicate enzyme activity, as the E26 line also exhibits faster growth in the absence of erythromycin.
In parallel, we investigated a potential action of CYP3A4 on erythromycin via a spot assay. Here, Chlamydomonas cells were spotted onto plates with and without erythromycin. The erythromycin concentration was chosen to be high enough that the UVM4 recipient strains cannot grow. Therefore, only lines with active CYP3A4 should grow on plates containing erythromycin.
(a) Level 2 MoClo construct for the production of 3xHA-tagged CYP3A4.
Four different concentrations of CYP3A4-producing transformants were spotted onto TAP plates lacking (b) or containing 8 mg/L erythromycin (c). The UVM4 strain was used as negative control.
Figure 8 shows that several transformants showed high resistance against erythromycin. Since this did not correlate with CYP3A4 expression, we concluded that spontaneous mutations in ribosomal proteins were responsible. Hence, erythromycin is no suitable substrate to test for P450 activity.
A visually less intuitive but probably more reliable activity assay involves the detection of erythromycin degradation by HPLC. For this, we first needed to detect erythromycin via our HPLC protocol.
Attempt at measuring Erythromycin with our HPLC protocol and column. 10 mg Erythromycin were dissolved in Methanol and the absorption was measured at 210 nm.
Unfortunately, we were not able to detect a significant peak for erythromycin with our method and column. For this reason, we had to choose a new substrate. The synthetic estrogen estradiol, among other uses, is found in birth control pillls and makes its way into the environment, where it continues to have hormonal effects, which is a strong incentive to degrade it.
Cultures of the recipient strain UVM4, as well as the two CYP3A4 expressing transformants E13 and E26, were grown to a defined cell number of 6*10^6 cells/ml. Then, TAP culture medium was exchanged via centrifugation and subsequent resuspension in TAP medium containing 5 µM estradiol. Cultures were then incubated for 24 h at 25°C. Shown is the absorption of the culture supernatant of the three stains as well as a TAP medium control containing estradiol in dependency on their retention time within the HPLC. Estradiol shows a peak at a retention time of about 7.9 minutes. Another prominent peak was observed at about 8.2 minutes. This peak negatively correlates with the amount of estradiol and is therefore suspected to be a degradation product of estradiol .
(a)Measurements of different estradiol concentrations for finding the right amount of substrate. (b) E26 with and without estradiol to identify estradiol related peaks. (c) HPLC-signals of TAP medium, and medium from UVM4, E13, and E26 cultures containing 5 µM estradiol. (d) Enlarged section of (c) without TAP medium signal.
(e-g) Evaluated data from (c). (e) Absolute estradiol content (area of the estradiol peak at 7.9 min). (f) Total estradiol content (sum of both areas of estradiol (7.9 min) and potential product (8.2 min)). (g) Relative estradiol content (amount of estradiol in relation to the sum of both areas (7.9 min + 8.2 min)).
Before starting activity tests using estradiol as a substrate to investigate the activity of CYP3A4, pre-tests were performed to elaborate an optimal protocol for measuring estradiol containing samples via HPLC. As a first step we monitored whether estradiol could be measured by HPLC and at which concentrations. Indeed, a peak could be detected at 210 nm after 7.9 minutes (retention time) (Fig. 10(a)). Based on this experiment, we chose an estradiol concentration of 50 µM for the measurement, which relates to 5 µM estradiol concentration during cultivation. Next, we identified peaks which specifically occur due to estradiol degradation, that can therefore not be seen in the control without estradiol (Fig. 10 (b)). Thereby two peaks occurred at about 7.7 min and 8.2 min additionally to the estradiol peak (7.9 min). When incubating the culture supernatant of our CYP3A4 producing transformants with 5 µM estradiol for 24 hours at 25°C, we discovered that our recipient strain UVM4 already shows high activity against our test substrate estradiol. This shows that Chlamydomonas itself is already an excellent platform for bioremediation. But this also makes it harder to distinguish between activity that is due to Chlamydomonas and CYP3A4 related activity. Nevertheless, both transformants, E13 and E26, exhibit even higher degradation of estradiol, which was shown to be significant compared to UVM4 (Fig. 10 (g)). This proves our concept of the functional production of CYP enzymes in Chlamydomonas using CYP3A4 as an example.
These results provide the entry point for further optimization regarding culture and assay conditions, other types of P450 enzymes, P450 enzymes lacking tags, or co-expression of reductases.
Producing CYPs without Tag
We reasoned that the C-terminal tags we used might interfere with P450 activity. Therefore, we designed constructs N-R (Fig. 1). However, due to fixed sequences in our constructs producing fusion sites for the MoClo system, a sequence encoding for a Gly/Ser linker was left between the CYP coding sequence and the stop codon. To remove that sequence, the coding sequences for CYP3A4, 9Q3, and 81A10V7 were amplified by PCR.
Agarose gel to verify PCR products for CYP3A4, 9Q3, and 81A10V7 for the assembly of level 2 constructs with stop codons (mStop) directly downstream of the coding sequence. The primers were specifically designed to remove the GS-linker.
Figure 11 shows successful amplification of the coding sequences for CYP3A4, 9Q3, and 81A10V7 lacking the linker-encoding sequence. Removal of the linker sequence was verified by Sanger sequencing. Unfortunately, these constructs could not be further investigated due to time reasons.
Chlamydomonas is sensitive to the drugs chlorsulfuron and atrazine. Both are known substrates of CYP81A10V7. We therefore transformed Chlamydomonas with a construct for CYP81A10V7 production and looked for resistant colonies.
(a) Level 2 MoClo construct for the production of CYP81A10V7 lacking any tag. (b) UVM4 cells were plated on HMP plates with and without atrazine. Moreover, UVM4 cells were transformed with the construct shown in (a) and plated on an HMP plate containing atrazine. (c) UVM4 and UVM4 transformed with the construct shown in (a) were plated on TAP plates containing chlorsulfuron.
No atrazine or chlorsulfuron resistant transformants were obtained. Sequencing revealed that, following the removal of the GS linker, mutations had occurred in the middle of the DNA sequence, which would explain the enzyme’s lack of functionality. Also, we did not succeed to find positive transformants by immunoblotting for the HA-tagged version of CYP81A10V7.
The screening of transformants generated with constructs producing P450 without tags required CYP-specific antibodies. For this, we used a commercial antibody against CYP3A4 and tested it on our positive transformants producing HA-tagged CYP3A4.
(a) Level 2 MoClo constructs for the production 3xHA-tagged CYP3A4.
Whole-cell proteins of transformants E2, E13, and E26 producing 3xHA-tagged CYP3A4 were extracted and analyzed by SDS-PAGE and immunoblotting using a commercial CYP3A4 antibody at different dilutions: (b) 1:1000, (c) 1:5000, (d) 1:10000, (e) 1:20000. The black arrow marks CYP3A4 and the white arrow marks a cross-reaction of the antibody. UVM4 served as negative control. (f) To confirm the expression of HA-tagged CYP3A4, the blot was detected with an anti-HA antibody.
A After testing several antibody concentrations, we can conclude that the commercial CYP3A4 antibody is not capable of detecting CYP3A4 produced in Chlamydomonas.
Finally, we attempted to screen transformants generated with mStop constructs spectroscopically. CYPs show a peak in absorption at 420 nm due to their heme domain. If transgenic CYPs would be produced in our transformants, this absorption should increase and changes should therefore be detectable. Initially, our transformants producing HA-tagged CYP3A4 were analyzed to see if absorption differences were detectable.
Native whole-cell protein samples of transgenic lines E7 and E26 producing 3xHA-tagged CYP3A4. Absorption was measured using a Nano Drop 2000. The ratio of absorption at 418 nm (Heme) to the total protein content at 280 nm is depicted. UVM4 served as negative control. n = 10.
By normalizing the absorption at 418 nm by the protein content (absorption at 280 nm), an elevated ratio could be observed for CYP3A4-producing lines compared to the UVM4 control (Fig. 14). Even though there is still room for improvement of this method, it indicates that this method might be employed for the screening of CYP450-producing transformants. According to Dr. Hugues Renault, tags could severely interfere with the activity of CYP enzymes. Hence, this method would allow for identifying transformants that express untagged versions of CYPs. Nevertheless, we need to cross-validate these results with another method, as we tested it only for one specific CYP enzyme (3A4). Due to time constraints, we were not able to screen transformants producing untagged CYP variants with this method.
Successes and failures
- Successes: We were able to produce five out of six of our P450s and reductases in the cytosol of Chlamydomonas reinhardtii. Among them were two CYP enzymes of human origin, which is a significant achievement in Chlamydomonas research, independent of our project goal.
Using the high flexibility of the MoClo system, we generated several multi-gene constructs and successfully achieved co-expression of the genes encoding for CYP3A4 and a reductase (POR) in Chlamydomonas.
We confirmed the targeting of CYP3A4 to membranes, which is essential for its activity.
We tested several visual methods to detect enzyme activity, such as growth and survival assay. These should be repeated with other antibiotics and herbicides. .
Furthermore, we have learned much about HPLC and have further developed it to analyze substrates such as estradiol. We have successfully implemented a protocol for measuring estradiol via HPLC, a known substrate for CYP3A4. We are confident that we can apply our knowledge to other potential CYP substrates, as long as they can be detected by UV-vis.
- Failures: We were unable to co-express the reductase (POR) in combination with our CYPs. Unfortunately we were not able to proof activity of our P450 enzymes against the chosen substrates. These were chosen because they represent threats to our environment.
Future plans
We intend to continue working on this project and apply other substrates for activity testing, first focusing on known substrates of our CYPs, which would allow for an easier detection of products via HPLC. We will continue measuring enzyme activities that can be monitored in survival assays. Furthermore, we would like to expand the library of CYPs with new, highly specific enzymes in contrast to the enzymes of broad substrate specificity like CYP3A4. We will strive to refine our algal chassis and explore additional methods for its improvement, for example by optimizing the supply of reducing equivalents to facilitate enzyme activity. There are many possibilities for enhancing our system and we will continue our research on this topic as we see it as our duty to combat man-made destruction of our planet.
Considerations for replicating the experiments
We uploaded all our parts on our registry with the specific DNA sequence. Further you can finde all the protocols we used on our experiments page.
