Design

In order to enable our bacteria to help microalgae grow, the primary purpose of our experiment was to test whether our bacteria can produce IAA using the implemented production pathway.

Based on the literature, we found an engineered pathway for IAA production from glucose (Guo et al., 2019)(Figure 1). To boost the biosynthesis of IAA and simplify the process, we used L-tryptophan as the precursor for IAA production, which is catalysed by three enzymes:ARO8, KDC, and AldH (Figure 1). The three enzymes perform different functions: Saccharomyces cerevisiae aminotransferase ARO8 for the conversion of L-tryptophan to indole-3-pyruvic acid, S. cerevisiae decarboxylase KDC for the decarboxylation of indole3-pyruvic acid to indole-3-acetaldehyde, and E. coli AldH for the oxidation of indole-3-acetaldehyde to the corresponding IAA (Guo et al., 2019). The coding sequences were retrieved from GenBank, optimised for E. coli codon usage and synthesised as gBlocks. For the plasmid design, the RBS for the three auxin genes are found in the iGEM page and they are the same for ARO8, KDC and AldH.

Figure 1: Metabolic pathway for the production of IAA from glucose (Figure was taken from Guo et al., 2019)

To enable the expression of the three pathway genes and subsequent phytohormone production, we designed a plasmid that will subsequently be built and transformed into the bacterium. We chose the E. coli DH5a strain and pBb8Ec-RFP (www.addgene.org, n.d.) as the vector backbone as they are well characterized. As microalgae take up auxin in the medium, an E. coli transporter may be required for exporting auxin into the medium. We found a member of the auxin efflux carrier (AEC) family that is expressed in E. coli but uncharacterized (Guo et al., 2019), as well as additional family members that are expressed in either Gram-positive bacteria or yeast with reference to the auxin secretion in the medium (‌Vandeputte et al., 2005). However, since there may already be pumps or channels where the auxin is externalized and there are not any properly characterized auxin efflux carriers expressed in any E.coli strains, we did not put an efflux carrier within our plasmid.

For the design of modeling, our plan was to both quantify auxin and determine the impact on the bacterial growth rate. We aimed to generate data showing the exact rate of auxin produced by our engineered bacterium. If the auxin expression is successfully tested, this could be done under a plate reader with different concentrations of cells and monitored at various timepoints consecutively. Another possible modeling was to check if the concentration of auxin reaches a certain amount, the accumulated auxin might impact bacteria growth rate. To do this, the engineered bacteria could be cultured with normal bacteria in different concentrations of auxin. The growth rate obtained by the experiment will be used to model the possible impact of auxin on bacteria growth. 

Build

In this phase, the main focus was to ligate the g-blocks and the vector to a complete plasmid.

Figure 2: Plasmid map of BBa_K4123001) carrying ARO8, KDC, and aldH under control of the arabinose-inducible promoter.

We used the Golden Gate DNA assembly method to insert ARO8, KDC, and aldH into the vector backbone pBbE8c-RFP. The vector backbone, lacking the rfp reporter gene, was amplified by PCR. We also added a FLAG tag to each insert as we might do western blot to check the expression of certain proteins.

To assemble the plasmid, we designed primers to remove the FLAG-tag by PCR and did the agarose gel extraction. According to the agarose gel (Figure 3), the size of the two fragments (normal Aro8 and vector backbone) were roughly the same as we expected (∼1.6kb and ∼2.6kb), so we performed an agarose gelDNA extraction and measured the DNA concentration using the Nanodrop.

Figure 3: Agarose gel results under a transilluminator. (Lane 1: 1kb plus DNA Ladder. Lane 3: PCR mix of Aro8. Lane 5: PCR mix of vector)

The other two gBlocks, encoding KDC and aldH, were resuspended in nuclease-free water. The Golden Gate assembly reaction was performed by mixing 1:5 molar ratio of g-blocks and vectors and incubating them in the thermocycler for the cycle (37°C, 1 min → 16°C, 1 min) x 30 cycles→ 60°C, 5 min. In total, two Golden Gate reactions were set up. One reaction contains all three inserts required to build the IAA biosynthesis pathway. To the second reaction only two of the inserts, ARO8 and aldH, were added. This reaction should not yield a circular plasmid and was used as negative control. LB-agar plates were supplemented with L-arabinose to induce gene expression.

Among the 8 plates, we expect colony growth on plate 1 as the mixture contains all the three inserts and were plated on the plate with the correct selective marker (chloramphenicol) and were induced by arabinose. Plate 2 was the negative control with only two of the inserts, ARO8 and aldH. Growth was not expected on this plate as it should not yield a circular plasmid. Plate 3 and plate 4 were cell competency tests, so we expected to see a lawn of bacteria on the plate as pBbE8c-RFP was chloramphenicol resistant. Plates 5 to 8 were positive controls and we expected a lawn of bacteria on the plates as they were plated on plates without chloramphenicol. Based on the results, we could conclude that the transformation worked as we expected and our bacteria were likely to have the plasmid we made at this stage.

In the next stage of the experiment, the main goal was to test whether our bacterium produces auxin. To check for correct assembly, we performed a diagnostic digest. The transformed cells were used to measure the level of auxin production using UPLC. Ideally, the UPLC result should show the production of auxin in the medium, indicating that the engineered bacteria was producing auxin as intended.

To do this, 6 transformants were tested for correct assembly. We extract the plasmids inside the bacteria using miniprep and we also use the original vector pBbE8c-RFP as a control. We chose two restriction digestion enzymes to cut the plasmid. One was Ndel which cut a single site on the plasmid (Figure 4). The other one was HindIII which cut twice on the plasmid, giving two DNA fragments (Figure 5).

Figure 4: Unique restriction sites of the final auxin plasmid. Ndel was used to digest the plasmid into linear strand.

Figure 5: Unique and double restriction sites of the final auxin plasmid. HindIII was used to digest the plasmid into 2 fragments.

We ran the restriction digestions in parallel and in the same agarose gel (Figure 6). The results indicate that the plasmids were assembled correctly with 3 fragments as they gave correct sizes. The 6 transformants also showed roughly the same results on the gel.

Figure 6: Agarose gel results. (Lane 1: 1kb plus DNA ladder. Lane 2-7: 6 linearized samples with expected size of 8131bp. Lane 8 & 15: intact vector backbone (undigested). Lane 9-14: 6 samples cutting into 2 fragments with expected sizes of 2737bp and 5394bp.)

Test

After checking the correctness of plasmid assembly, we carried on checking if the bacteria with the plasmid are able to produce auxin (IAA). The 6 transformants with 3 inserts and the original undigested vector backbone were taken from the previous step. Referring to the synthesis pathway of IAA, we added 100μl of 10 g/L L-tryptophan as the limiting factor for our bacteria to produce the final product, IAA. 1% Arabinose stock was also added into the 6 samples to make up 0.1% arabinose for gene expression. Then we put them on the shaker for 24h, and checked if the medium contains IAA under UPLC.

To compare the wavelength of peaks between our samples and the control, we included pure IAA and L-Tryptophan to see if our samples contain these compounds. Based on the literature, IAA has a UV absorption spectrum characteristic of substituted indoles, with a strong maximum at 220 nm (E220 = 33 200) and a characteristic of three overlapping peaks with maxima at 274, 282, and 288 nm (E282 = 6060) (www.sciencedirect.com, n.d.), while tryptophan absorbs more light at 280nm, which is distinct from IAA (Biotek.com, 2020). Since we aim to detect IAA in the samples, all 6 samples (G1-G6, G stands for golden gate product) were focused on 280 nm which corresponds to IAA.

We ran the pure sample of both IAA (Figure 7) and L-Tryptophan (Figure 8) under the same settings, and the intact vector (Figure 9) was included as a control.

Figure 7: UPLC-DAD chromatogram of IAA at 280 nm. The retention time of IAA is roughly between 0.55 to 0.6 min (0.563 in specific) .

Figure 8: UPLC-DAD chromatogram of L-Tryptophan at 280 nm. The retention time of L-Tryptophan is roughly between 0.3 to 0.35 min (0.313 min in specific).

Figure 9: UPLC-DAD chromatogram of the vector (pBbE8c-RFP) at 280 nm. The peaks were roughly between 0.25 to 0.3 min (0.258 in specific) and between 0.8 to 0.85 (0.820 in specific).

Figure 10: UPLC-DAD chromatogram of the 6 samples (G1-G6) at 280 nm.

According to the graphs (Figure 10), G1, 2, 5 of the 6 samples show differences with the vector control, while the other 3 samples showed a similar pattern with the vector control. Moreover, the peaks were not in a perfect overlap with IAA.

Learn

According to the results of UPLC, the results are inconclusive to show that our bacteria successfully produced IAA after the addition of L-tryptophan. In this case, we were considering additional tests to get more information from the experiment.

We could add more tryptophan for auxin synthesis. As we only added 0.1M L-tryptophan to the medium, this might not be sufficient for our bacteria to use as substrate for IAA production pathway. After comparing the excitation wavelength of the intermediates in the data analysis section, it is difficult to tell the exact compounds responsible for other peaks. Therefore, MS may be performed to see the details of the compounds corresponding to other peaks.

Another possible circumstance was that our plasmid did successfully assembled while several key sequences might have changed as the sequence was not checked. Although we checked the size of the plasmid before adding L-Tryptophan for IAA secretion, the results of the gel cannot guarantee the accuracy of sequences. The gel is not able to show us the specific difference if there were mutations in several sequences. To further check the sequences, we could do Sanger sequencing to the three samples (G1, G2, and G5) shown differences with the control in UPLC. If the sequences are what we desired, we could conclude that the plasmid was built successfully. As we still don’t know if the gene is expressing KDC, Aro8, and AldH, it might be necessary to check if the genes are functional. Therefore, presence of ARO8, KDC, and AldH may be validated by western blot analysis of FLAG-tagged proteins.