Updated on 2022-10-28: We won Best Measurement this year.

Overview

Within a single-cell bacterial factory, exogenous genes are expressed to modify metabolites. Co-expression is often achieved by co-transforming two or more plasmids, each carrying the gene of one subunit and a different selection marker, into the heterologous expression host. However, the number of plasmids used is not infinite, as transforming and maintaining more than four plasmids simultaneously is very challenging.

The uneven copy numbers of each plasmid make protein expression and unpredictable. Another approach is to assemble multiple genes into one plasmid where genes are transcribed from multiple promoters (Scheich, C et al. 2007) or a single promoter to produce long lines of cis mRNA (Tan, S et al. 2005). In our project, we initially inserted crtIYEB into one plasmid, constructing a polycistronic plasmid.

Polycistronic expression suffers from a major problem: translations of cistrons are severely unbalanced. The order of cistrons in the polycistronic mRNA has an unpredictable consequence on the levels of protein expression (Renaud-Gabardos, E et al. 2015). Progressively lower expression was reported with more downstream cistron in the mRNA (Kim, K. J et al. 2004). We witnessed this phenomenon in our project as well, which is depicted by failure in obtaining the intermediate product β-carotene.

To make sure that our single-cell bacterial factory functions well, we must confirm every single step works by successfully detecting target products. Therefore, We integrated a whole set of measurement approaches to validate our parts, which can be applied to other occasions as well.

Verification

PCR to confirm the successful transformation of plasmids into E. coli

Before we verify protein expression we need to confirm whether the target gene has been successfully transformed into the plasmids.

Figure 1. Agarose gel electrophoresis of PCR products, amplified from a bacterial colony.

The first lane was loaded with a D2000 DNA ladder whose sizes were marked on the image. We chose Taq DNA polymerase for its low cost and high reliability, and we designed forward and reverse primers for each carotene synthesis enzyme (crt for short). The PCR reaction was composed of 2 μL 10x Taq polymerase buffer, 16 μL H2O, 0.5 μL Taq polymerase, 0.5 μL dNTP (10 mM each), 0.5 μL forward primer (10 mM), 0.5 μL reverse primer (10 mM), and 1 colony picked from the plate. Using the same forward primer, and different reverse primers, we were able to detect the composition of various crt genes. After PCR, the correct bacterial clones were sent for Sanger sequencing, which take 24-48 hours. While waiting, we start preparing the following experiments. Once the sequencing confirmed, we execute the next.

SDS-PAGE to confirm the successful expression of proteins and the effects of RBS intensity on protein expression

Figure 2. SDS-PAGE of whole cell lysate of bacterial cultures.

Protein expression was induced in parallel cultures by IPTG. Bacterial cultures were monitored by OD600, and 5x10^ 7 cells were harvested by centrifugation and lysis in 1x SDS sample buffer. An equal amount (10 μL, 2x10^6 cells) of whole cell lysate was analyzed by SDS-PAGE (4~20% gradient gel, Tanon brand). Red arrows point to crtI protein. Green arrows point to crtY protein. Black arrows point to crtB protein. Yellow arrows point to crtE protein.

Figure 3. We failed to see IPTG induced expression of the membrane protein BCMO from the electrophoresis result, so we turned to microscopy by tagging fluorescent protein downstream the target protein, which is included on this page.

Color changes to verify β-carotene production

As our intermediate product is the yellow β-carotene, we can confirm its presence simply by observing the color change in the bacterial pellet, sometime even the liquid bacterial culture. It served as the visual assay to tell whether the four crt enzymes cooperated effectively under the same or different types of RBS.
No color change in the bacterial pellet along with the unequal protein expression induced by IPTG pushed us to adjust the strength of RBS in front of each enzyme. By enhancing RBS or weakening it, we successfully obtained the yellow products, as shown in BBa_K4162117 and BBa_K4162118.

Figure 4. Bacterial pellets, after acetone extraction (in-complete), still bear the product β-carotene yellow.

Fluorescence microscopy to confirm the successful expression of membrane proteins

Failure in observing the expression of membrane protein BCMO from SDS-PAGE pushed us to reflect and try to figure out which step went wrong. No band showed after the IPTG induction in BL21(DE3) strain and we surely followed the standardized protocol reported (Jang, H. J et al. 2011). This situation set us thinking maybe it should be the assay’s fault. After consulting Prof. Motoyuki Hattori, who has extensive knowledge in expressing, purifying membrane protein, we knew that SDS-PAGE with CBB staining is not suitable for membrane protein expression detection. So， we changed to this fluorescence microscopy assay.

Figure 5. Fluorescent imaging confirm the expression of both BCMO and ybbO. We observed fluorescence images of E. coli (1. Bacteria expressing BCMO-linker-StayGold; 2. Bacteria expressing BCMO and ybbO-tdMCP-GFP) through an Olympus fluorescence microscope with a 1.45 NA 150× oil objective. GFP and StayGold were excited by a 488 nm laser. Images were taken by the same settings for comparison. For details about sample preparation, please check our protocol here.

Figure 6. Fluorescent imaging revealed that IPTG induction increases the expression of both BCMO and ybbO.

We observed fluorescence images of Hi-control Rosetta strain (1. Bacteria expressing BCMO-linker-StayGold; 2. Bacteria expressing BCMO and ybbO-tdMCP-GFP) through an Olympus fluorescence microscope. Samples induced with IPTG have stronger signal than controls.

HPLC for detection of retinoids

After we observed the color change of the system, for quantitative detection and validation with chemical standards, we applied HPLC to detect β-carotene and retinoids.

Figure 7. Acetone extracted samples were subjected to HPLC (C18 column) as described previously (5), except we are using methanol:water = 96:4 as the moving phase. The extract from bacteria DH5α expressing CAG-MS2, BCMO, ybbO, crtYEBI (the panel above, short name 3-7) compared with the extract from bacteria expressing CAG-MS2, BCMO, ybbO-tdMCP, crtYEBI (the panel below, short name 3td7), there are two peaks in the figure between 2-4 minutes (where our standards appear as well others, detected at 325 nm), the absorbance value of the two peaks of 3td7 in this period is significantly higher than that of 3-7, which depicted the efficiency of TEARS in improving the production of retinol.

Optimization

Verify that adjusting RBS intensity promotes metabolism

In our project, we used ribozyme to separate our crtEBIY BioBrick when it produced in vivo. With the help of self-cleaving ribozyme, self-interaction of the polycistron can be avoided and the translation of each cistron should be basically dependent on the strength of its upstream RBS.

Figure 8. Proof the efficiency of ribozyme-assisted polycistronic expression.

When the RBS strength of the four enzymes in our linear reaction was consistent, the results on the protein gel showed that the expression of upstream proteins (crtI and crtY) was higher than that of downstream proteins (crtE and crtB) after IPTG induction, which was not in line with our expectation of the equal consistent level of the four enzymes. Althrough we use the leaky expression of the promoter, IPTG result suggests that not only RBS but also the protein sequence affects the yield. Moreover, the bacteria pellet was not yellowing, indicating that we do not have the ideal product β-carotene.

When we weakened the RBS strength of the one upstream enzyme (crtY), the electrophoresis result (Figure 8) showed the disappear of strong crtY band, and the bacterial culture as well as the pellet turned yellow, indicating that we had obtained β-carotene. Similar result was obtained when we enhanced the expression of one downstream enzyme (crtE).

Figure 9. IPTG(-/+) = without/with 0.2 mM IPTG for 3-6 hours, adding IPTG to a bacteria culture with OD600 0.2-0.3. M: Protein molecular weight marker ladder. Lane 1~2: pET28 plasmids encoding crtEBIY without any tag were transformed into BL21(DE3) Rosetta strain, single clones (6b) were picked for liquid LB culture. Lane 3~12: pET28 plasmids encoding crtYEBI without any tag were transformed into BL21(DE3) Hi-Control strain, single clones (Hi-7, Hi-7a, Hi-7b, Hi-7c, Hi-7d) were picked for liquid LB culture. Lane 14~17: pET28 plasmids encoding crtB, crtE, crtI, crtY without any tag were transformed into BL21(DE3) Hi-Control strain, single clones (1B1, E1, I7, Y1) were picked for liquid LB culture. Protein expression was induced in parallel cultures by IPTG. Bacterial cultures were monitored by OD600, and 5x10^7 cells were harvested by centrifugation and lysis in 1x SDS sample buffer. Equal amount (10 μL, 2x10^6 cells) of whole cell lysate were analyzed by SDS-PAGE (4~20% gradient gel, Tanon brand). Red arrows point to crtI protein. Green arrows point to crtY protein. Black arrows point to crtB protein. Yellow arrows point to crtE protein.

By combining electrophoresis results with the observation of culture/pellet color changes, we could verify that adjusting the RBS strength of individual mono-cistrons in the presence of ribozyme in polycistronic plasmids could direct linear reactions to favorite metabolite production.

Summary

We integrated a series of common but robust measurements to step-by-step verify and optimze our engineered bacteria. We verify the reaction chain step by step, and employ error correction and optimization, so as to drive bacterial metabolism to the direction we designed, to ensure the efficiency of the system and get the final product. On the basis of the detection results, the reaction was optimized by adjusting RBS strength and introducing TEARS, a recent published (Guo, H et al. 2022) bacterial enzyme enrichment tool, to increase the yield for further commercial use.

References

Scheich, C., Kümmel, D., Soumailakakis, D., Heinemann, U., & Büssow, K. (2007). Vectors for co-expression of an unrestricted number of proteins. Nucleic acids research, 35(6), e43.

Tan, S., Kern, R. C., & Selleck, W. (2005). The pST44 polycistronic expression system for producing protein complexes in Escherichia coli. Protein expression and purification, 40(2), 385–395.

Renaud-Gabardos, E., Hantelys, F., Morfoisse, F., Chaufour, X., Garmy-Susini, B., & Prats, A. C. (2015). Internal ribosome entry site-based vectors for combined gene therapy. World journal of experimental medicine, 5(1), 11–20.

Kim, K. J., Kim, H. E., Lee, K. H., Han, W., Yi, M. J., Jeong, J., & Oh, B. H. (2004). Two-promoter vector is highly efficient for overproduction of protein complexes. Protein science : a publication of the Protein Society, 13(6), 1698–1703.

Jang, H. J., Yoon, S. H., Ryu, H. K., Kim, J. H., Wang, C. L., Kim, J. Y., Oh, D. K., & Kim, S. W. (2011). Retinoid production using metabolically engineered Escherichia coli with a two-phase culture system. Microbial cell factories, 10, 59.

Guo, H., Ryan, J.C., Mallet, A., Song, X., Pabst, V., Decrulle, A., Lindner, A.B. (2020). Spatial engineering of E. coli with addressable phase-separated RNAs. bioRxiv 2020.07.02.182527; doi: https://doi.org/10.1101/2020.07.02.182527

Guo, H., Ryan, J. C., Song, X., Mallet, A., Zhang, M., Pabst, V., Decrulle, A. L., Ejsmont, P., Wintermute, E. H., & Lindner, A. B. (2022). Spatial engineering of E. coli with addressable phase-separated RNAs. Cell, 185(20), 3823–3837.e23.