This page will discuss the detailed trouble-shooting steps during DNA cloning and transformation. For the summative results of DNA cloning, transformation, and protein expression, please visit the Result Page. For the enzymatic assays and ELISA test of our fusion proteins, please visit the Proof of Concept Page.
Introduction
With SepScan, our goal is two-fold. Firstly, we want to make nanobody-reporter protein fusions to detect Cystatin C (Cys C), a predictive biomarker for sepsis and acute kidney injury. Secondly, we want to improve the sensitivity of the fusion protein so that this detection system is applicable for developing detective tools for other septic biomarkers with a lower limit of detection in the future. We set out by exploring three types of nanobody-reporter protein fusions:
- Nanobody fused with a chromoprotein
- Nanobody fused with a single NanoLuiferase
- Nanobody fused with two NanoLuciferases
Highlights From the Cycles
Nanobody (NB)-chromoprotein designs: Four cycles were carried out. NB-gfasPurple, NB-tsPurple, and gfasPurple gBlocks were all successfully cloned into plasmids and transformed into BL21 (DE3). During protein expression, NB-gfasPurple appeared to be trapped in the cytoplasm, while NB-tsPurple showed a promising lilac colour but the size was not proved to be correct from SDS-PAGE. However, tsPurple was successfully expressed.
Single NanoLuciferase (NL) designs: Four cycles were carried out. Both NB-NL and NL were successfully expressed, with their sizes confirmed by SDS-PAGE.
Double NL Designs: Two cycles were carried out. We successfully assembled and cloned NL-NL, while NB-NL-NL was expressed and verified via SDS-PAGE.
Since the design, cloning, transformation, and expression of all our gBlocks followed a similar process, we have summarised the general design and build phases below. For more details, please refer to the specific cycles later on this page.
Design
We obtained the sequence for the anti-Cys C NB from the literature [1]. For chromoproteins, we searched one with a high extinction coefficient, which measures light attenuation through a medium. A high extinction coefficient means that chromoprotein can absorb and give off light efficiently even at low concentration. A study on coral-derived chromoproteins identified gfas-Purple as having a very high molar extinction coefficient of 205,200 [2]. However, during our research, we discovered that expressing gfas-Purple in E. coli halves the strain’s growth rate [3]. Therefore, as a back-up plan, we selected tinsel-Purple (tsPurple), which is well characterised and familiar to UCL’s Department of Biochemical Engineering. Most importantly, when expressed in E. coli, it has relatively minimal impact on the growth rate [2]. We obtained the sequences for gfas-Purple (BBa_K1033918) and tsPurple (BBa_K1033906) from the iGEM registry. We obtained the NL sequence from a database and codon-optimised it for E. coli [4].
The well-characterised, high-copy plasmid pSB1C3-FB, derived from UCL’s 2018 iGEM team (BBa_K2842666), was used as the plasmid backbone. Since the E. coli cytoplasm is a reducing environment unsuitable for NB formation, we targeted the fusion protein to the periplasm by attaching a pelB signal sequence upstream of the NB-reporter protein for periplasmic expression. For the linker, we selected a long flexible (GGGGS)₃ linker based on a study comparing linkers in Luciferase-GFP proteins [5], to minimise interference between the NB and reporter proteins. A His-tag was added to the reporter protein’s C-terminus using a short GGSG linker (BBa_K243004) for purification. As pSB1C3-FB contains BsaI restriction sites, we designed the gBlocks with BsaI sites that create overhangs complementary to the plasmid backbone.
Three chromoprotein gBlocks were assembled as shown in Figure 1.1. The gfas-Purple without NB will act as a control in ELISA affinity tests. Dr Jack Jeffries provided us with the glycerol stocks of tsPurple. Hence, we did not design and clone the gBlock for tsPurple.
Figure 1.1: The gBlock designs of NB-gfasPurple (BBa_k5357019), NB-tsPurple (BBa_k5357021), and gfasPurple (BBa_k5357013).
Four NL gBlocks are assembled as shown in Figure 1.2. NL and double NL (without the NB) served as controls for NB-NL and NB-NL-NL in ELISA affinity tests and enzymatic models. We used the linker-NL gBlock to assemble NL-NL ourselves. A new reverse primer was designed to remove the His-tag, T7Te terminator, and original BsaI reverse site on the NL gBlock. This also introduced a new BsaI reverse site with an overhang complementary exclusively to the BsaI forward site on the linker-NL.
Figure 1.2: the gBlock designs of NB-NL (BBa_k5357016), NL (BBa_k5357015), NB-NL-NL (BBa_k5357022), and linker-NL (for assembling double NL: BBa_k5357018).
Except that linker-NL was ordered from IDT as gBlocks and NB-NL-NL were assembled into the plasmid by Genscript, all of the rest of the designs were ordered as gBlocks from Genscript.
Build
We cloned the gBlocks into pSB1C3-FB using Golden Gate cloning with BsaI restriction enzymes. Specifically for NL-NL, PCR was performed on the NL gBlock using a universal forward primer and our designed reverse primer before the Golden Gate cloning. The plasmids were then transformed into E. coli DH5α for amplification. Transformed cells were plated on IPTG/X-gal agar for blue-white screening. White colonies indicated successful gBlock insertion, as the LacZ sequence was disrupted, while blue colonies indicated no insertion, as LacZ remained intact, hydrolysing X-gal into a blue product. Chloramphenicol was added to LB-agar for contamination control and selection. We performed PCR cleanup to remove sample impurities.
Next, the plasmids were extracted and transformed into the protein expression strain BL21 (DE3). For protein expression, we included a negative control using the pSB1C3-FB plasmid without any inserts. Periplasmic extraction was carried out with TES buffer, and immobilised metal ion affinity chromatography was used for protein purification via the His-tag. Specifically for tsPurple, ion exchange chromatography was used for purification.
Nanobody-Chromoprotein
Cycle 1
Test:
The results of DH5α plates are shown in Figure 2.1. Colonies were seen on the plates with NB-tsPurple and gfasPurple but not on the NB-gfasPurple plate.
Figure 2.1: the plates of NB-tsPurple with white colonies, gfasPurple with white colonies, and NB-gfasPurple without colonies.
Learn:
It was unusual to see no blue colonies on the NB-tsPurple and gfasPurple plates, as some blue colonies are typically expected. On a positive note, this could indicate a highly efficient Golden Gate cloning and transformation process. No colonies were observed for NB-gfasPurple, which may be due to the larger gBlock insert causing stress during cell growth.
Cycle 2
Build:
We replated NB-gfasPurple under the same conditions to test if its non-growing is due to chance. Meanwhile, we added another condition by doubling the amount of Golden Gate product during transformation, to increase efficiency. This means instead of a 1: 10 Golden Gate plasmid product to competent cell ratio, a 1: 5 ratio was used. Furthermore, a control was also added to determine if it was a human mistake during the plating process. (Figure 2.2)
Test:
From plates, it has been observed that more colonies were formed when the original plasmid to cell ratios is used (1: 10). Doubling the golden gate plasmids to competent cell ratio (1: 5) has further increased cell stress resulting in fewer colonies.
Figure 2.2: the plate of control containing pSB1C3-FB plasmid without any insert; the plate of NB-gfasPurple with all white colonies; the plate of NB-gfasPurple with double plasmid to cell ratio during transformation with all white colonies.
Now that colonies were obtained for all strains, we extracted the plasmids and carried out a diagnostic digest to determine if the correct sequences were inserted into the plasmid (Figure 2.3). One cutting site was selected in the plasmid backbone while the other cutting site was selected in the insert. The length of all fragments matched with the expected length of virtual digest generated by Benchling (Figure 2.4).
Figure 2.3: Diagnostic Digest Results of NB-tsPurple, gfasPurple, NB-gfasPurple, and doubled concentration NB-gfasPurple.
Figure 2.4 Virtual digest result.
Learn:
Therefore, we concluded that 1: 10 would be the more preferred ratio, and the absence of colonies in Figure 2.1.1 might have resulted from uncertainties.
Cycle 3
Test:
As shown in Figure 2.5, the transformed BL21(DE3) plates of NB-tsPurple and gfasPurple did not form colonies, while NB-gfasPurple did. Consequently, the successfully grown NB-gfasPurple colonies were selected to made into starter cultures for protein expression. We expected to observe a purple-coloured culture after protein expression, but this was not the case. To test for the presence of protein, we conducted a quick qualitative test using Bradford reagent, which resulted in a blue colour for both the negative control and the purified NB-gfasPurple sample due to imidazole reacting with the reagent. Therefore, we performed SDS-PAGE and western blotting to confirm the presence of proteins. Samples included the purified sample (E), the cell pellet after periplasmic extraction (P), and the raw protein supernatant before purification (S). This approach aimed to determine the presence of proteins: if proteins were found in S but not in E, the elution process might have been unsuccessful; if proteins were present in P but not in E, it would suggest that periplasmic extraction was ineffective.
Figure 2.5: transformed BL21 (DE3) plate of NB-tsPurple, gfasPurple without colonies, and NB-gfasPurple with colonies.
Comparing the SDS-PAGE results of the negative control (pSB1C3-FB), no significant difference was observed in the NB-gfasPurple sample (Figure 2.6), suggesting that the protein may not have been expressed. However, the His-tag western blot (Figure 2.7) revealed a clear band in lane P at approximately 50 kDa, similar in size to the NB-gfasPurple protein, including the pelB sequence (42.8 kDa). Since all our fusion protein designs contain a His-tag, this result may indicate the presence of the NB-gfasPurple protein trapped inside the cytoplasm. Additionally, the western blot of NL expressed in the same batch showed a band at about 19 kDa, aligning with the expected size of NL (19.1 kDa). The successful expression of NL suggests that the failure to express NB-gfasPurple was likely not due to human error in the procedure but rather a random occurrence.
Figure 2.6: SDS-PAGE result of NB-gfasPurple protein expression sample. S is the supernatant after lysing the cells with TES buffer; P is the cell pellet; E is the purified protein sample.
Figure 2.7: His-tag western blot result of NB-gfasPurple protein expression sample. S is the raw protein sample before purification; P is the cell pellet; E is the purified protein sample.
Learn:
As the plates containing the NB-tsPurple and gfasPurple plasmids did not yield any colonies, we sent the miniprep products of NB-tsPurple, gfasPurple, and NB-gfasPurple for sequencing. The results indicated that the NB-tsPurple sample was impure, containing a high percentage of concatemers. We expected the plasmids to measure 3461 bp; however, as shown in Figure 2.8, a significant proportion of the sample consisted of plasmids around 6900 bp, suggesting multiple fragments had been inserted. Similarly, the gfasPurple miniprep product was impure (Figure 2.8). We anticipated the plasmid length to be 3014 bp, but the presence of concatemers at approximately 6000 bp, along with some improperly assembled plasmids at 3000 bp, was noted. In contrast, the sequencing length of NB-gfasPurple matched our expected length of 3440 bp and aligned perfectly with our assembled plasmid in Benchling (Figure 2.9). Therefore, we concluded that the presence of concatemers in NB-tsPurple and gfasPurple inhibited the growth of BL21 (DE3).
Figure 2.8: sequencing result of NB-tsPurple and gfasPurple.
Figure 2.9: alignment of sequencing result with NB-gfasPurple design.
Cycle 4
Build:
We sent more miniprep samples of NB-tsPurple and gfasPurple for sequencing. After receiving the confirm result (Figure 2.10 and Figure 2.11), we transformed the correctly inserted plasmids of gfasPurple and NB-tsPurple into BL21 (DE3) separately. Buffer exchange was carried out to eliminate the effect of imidazole from protein purification buffers on concentration measurement. Meanwhile tsPurple was expressed using the same method. It was then harvest via sonification and purified via ion exchange chromatography.Figure 2.10 Alignment of the NB-tsPurple miniprep sample with our gBlock design on Benchling.
Figure 2.11 Alignment of the gfasPurple miniprep sample with our gBlock design on Benchling.
Test:
As shown in Figure 2.12, both two repeats of gfasPurple and the repeat of NB-tsPurple had colonies. NB-gfasPurple and NB-tsPurple are similar constructs designed for our goal of detecting Cys C using NB-chromoprotein. Due to the previous unsatisfactory expression results with NB-gfasPurple, we proceeded with protein expression using only NB-tsPurple and tsPurple, and no further transformation or protein expression steps were conducted for the NB-gfasPurple and gfasPurple designs. During purification, the chromatography column of NB-tsPurple turned purple and samples presented a lilac colour while tsPurple presented a deep purple colour (Figure 2.13). In SDS-PAGE (Figure 2.14), no band was observed around the size of NB-tsPurple (41.2 kDa). Therefore, it is likely that tsPurple was expressed without NB fusing to it. From the Bradford assay (Figure 2.15), the concentration of NB-tsPurple sample was 35.92 µg/mL. and the concentration of tsPurple purified sample ranged from 15353.33 µg/mL to 522.67 µg/mL depending on the order in elution flowthroughs.
Figure 2.12: plates of gfasPurple and NB-tsPurple all with colonies.
Figure 2.13: purification picture of NB-tsPurple and elution samples of tsPurple.
Figure 2.14, SDS-PAGE result of NB-tsPurple.
Figure 2.15: Concentration of NB-tsPurple determined from Bradford assay
Learn:
The observation of purple colour during purification confirms the presence of tsPurple in the sample. Interestingly, no protein of matching size was detected in the SDS-PAGE results. This may be due to the breakage of the (GGGGS)³ linker during protein expression and purification. The faint lilac colour can be explained by the presence of the retained tsPurple insert in the plasmid. Unfortunately, due to the project's timeframe, we were unable to verify this hypothesis. Consequently, no further analyses, such as enzymatic assays or ELISA, were conducted, as there is existing enzymatic assay for tsPurple demonstrating its activity, and without the NB, there should be no affinity to Cys C.Single NanoLuciferase
Cycle 1
Test:
Similar to chromoproteins, colonies were not found in one of the plates, NL, when we transformed DH5α and plated for the first time. (Figure 3.1)
Figure 3.1: Plate of NL with no colonies present, and plate of NB-NL with all white colonies.
Learn:
This result is unexpected because NL is a smaller insert compared to NB-NL. Given that white colonies were observed for NB-NL, the cells should not be stressed to express NL. Hence, we suspected that it was because an old LB agar plate was used, leading to uncertainties in cell growth.
Cycle 2
Build:
We made fresh LB agar plates. Similarly to the chromoprotein troubleshooting process, we repeated transformation using a 1: 10 plasmids to cell ratio and a 1: 5 plasmids to cell ratio.
Test:
The results turned out to be similar to repeated NB-chromoprotein plasmid transformation discussed in NB-chromoprotein section. The original plasmid to cell ratio (1: 10) during transformation has more colonies. The two blue colonies in Figure 3.2 has also suggested a reasonable uncertainty in plasmid transformation.
Figure 3.2: Left, repeated transformation and plating of NL with two blue colonies. Right, plate of NL, using doubled the plasmid to cell ratio during transformation, with all white colonies.
Learn:
Diagnostic digest was carried out to determine if the correct sequence was inserted into the plasmid. From the diagnostic result of NB-NL and NL, the length of the fragments matches with the expected length shown in virtual digest. (Figure 3.3 & 3.4)
Figure 3.3: Diagnostic Digest results of NB-NL and NL.
Figure 3.4 Virtual digest result of NB-NL and NL generated by Benchling.
Cycle 3
Build:
We carried the NL and NB-NL sample confirmed from diagnostic digest forward to transformation into protein expression strain BL21 (DE3), protein expression, and purification.
Test:
As shown in Figure 3.5, both NB-NL and NL had colonies. The successfully grown colonies got picked and made into starter culture and proceeded to protein expression.
Figure 3.5: plate of NL and NB-NL with colonies.
In SDS-PAGE, a band at about 20 kDa was observed in E lane of NL which matches with its expecting size of 19.1 kDa (Figure 3.6). This suggests the successful expression of NL. The result is further confirmed by His-tag western blot. For NB-NL, no significant difference were observed in sample NB-gfasPurple (Figure 3.6). Similarly, no band was observed in the western blot result of NB-NL. This suggests that NB-NL has not been expressed.
Figure 3.6 SDS-PAGE and His-tag western blot results of NL and NB-NL protein expression sample.
Learn:
Although the NB-NL plasmids were confirmed correct through previous diagnostic digests, protein expression did not succeed. Therefore, we sent the NB-NL miniprep product for sequencing. The results revealed that the transformed and expressed NB-NL plasmid was a large concatemer (Figure 3.7), explaining the unsuccessful outcome. Reflecting on the diagnostic digest, we had selected one restriction enzyme that cuts within the gBlock and another that cuts the plasmid backbone. Therefore, even if the plasmid is a concatemer, this would not be picked up by the diagnostic digest, as all multiple inserts would have been cut to the same size. To improve this process, we suggest that future iGEM teams select two sets of restriction enzymes: one set that only cuts the backbone and another set that cuts both the backbone and the insert. This will allow for thorough verification that the gBlocks have been correctly inserted.
Figure 3.7, plasmid map and bar chart of NB-NL concatemer. 3 copies of the NB-NL gBlocks were inserted into the plasmid.
Interestingly, the expressed NL strain was sequenced and found to be concatemer while western blots confirmed that NL has been expressed (Figure 3.8). The expected length of NL plasmid is 2867 bp but the sequencing result reflected that plasmids had a length about 5600 bp. Diagnostic digest did not detected this concatemer for the same reason as NB-NL mentioned above. Hence, We suggested that concatemer may not completely hinder protein expression of the strain depending on the number of inserts that was repeated.
Figure 3.8: The sequencing result of NL.
Cycle 4
Build:
We selected new DH5α colonies of NL and NB-NL plasmid for sequencing. The sequences of NB-NL and NL aligned completely with our gBlock designs (Figure 3.9). Hence, we proceeded to transformation of plasmids into BL21 (DE3) and protein expression.Figure 3.9: Alignment file of NL and NB-NL
Test:
As shown in Figure 3.10, both plates of transformed NB-NL and NL in BL21 (DE3) had colonies.
Figure 3.10 plates of NB-NL and NL with colonies
In SDS-PAGE (Figure 3.11), a band was observed in lane 1 at about 20 kDa, matching the size of NL (19.1 kDa). For NB-NL, lane 2, a band was observed at about 35 kDa, consistent with the expected size of 34.8 kDa. From Bradford assay, the concentration of NB-NL and NL was 5.05 μg/mL and 236.75 μg/mL respectively.
Figure 3.11 SDS-PAGE result of NL (lane 1) and NB-NL (lane 2).
Learn:
Because no concatemer was present for the newly pickedDH5α colonies of NB-NL and NL, the first-time concatemer results were likely happened due to chance.As the bands on SDS-PAGE match with our designs, NB-NL and NL have been expressed successfully. Reflecting back to the first-time failure, it stressed the the importance of transforming the correctly inserted plasmid into protein expression strain. To perform the lab more efficiently, it is worth confirming the plasmid insert via sequencing before transformation.
Double NanoLuciferase
Cycle 1
Test:
For NB-NL-NL, transformed BL21 (DE3) colonies had successfully grow on plate (Figire 4.1). In SDS-PAGE, a dark band was observed at about 60 kDa similar to the length of NB-NL-NL (54.7 kDa) (Figure 4.2), suggesting successful protein expressions. From Bradford assay, the concentration of NB-NL-NL elution samples were determined to be ranging from 182.58 μg/mL to 944.25 μg/mL depending on the order of eluted samples.
Figure 4.1, Transformed BL21 (DE3) plate of NB-NL-NL.
Figure 4.2, SDS-PAGE result of NB-NL-NL.
For NL-NL, a gel electrophoresis was conducted to check if the excess His-tag, terminator, and BsaI reversed site had been trimmed off to give NL first position amplicon. Consistent with our expectation, the NL first position amplicon was smaller than the original NL gBlock (Figure 4.3). This suggested a successful PCR. We sent the results for sequencing while proceeding to transformation. Two plates of the NL-NL constructs were plated and one plate of NB-NL-NL was made (Figure 4.4) .
Figure 4.3, Gel electrophoresis of NL PCR amplicon.
Figure 4.4, Transformed DH5α Plates of NL-NL with few blue colonies.
Learn:
For NL-NL, although its plates had white colonies with a few blue colonies, the sequencing results suggest that linker-NL has not been inserted. We expected a length of 3851 bp, but the sequenced plasmid had a length of 2867 bp matching the length of a plasmid without the linker-NL insert. Reviewing back to the gel analysis of the first position amplicon of NL, there were some un-trimmed NL impurities present in the sample (Figure 4.5). From this we generated two possible explanations. Firstly, this outcome might be due to chance by picking up a colony that does not have the correct insert. The NL amplicon had a much higher concentration than the impurities from the gel analysis. Therefore, some of the colonies in the plates might contain the correct insert. Secondly, the NL impurities outcompeted the NL amplicon during Golden Gate cloning. Furthermore, as NL is smaller than linker-NL, it would be inserted more readily into the plasmid backbone, outcompeting double NL.
Figure 4.5: Sequencing result of NL-NL plasmid.
These two reasonings were conflicting in essence because white colonies were observed on plates suggesting plasmids with inserts had been cloned into the cell. If PCR was successful, the overhang of the non-PCR NL would not be complimentary to the overhang of the plasmid backbone resulting in many blue colonies. This suggested that it was more likely that the issue occurred in PCR amplification.
Cycle 2
Build:
PCR amplification of NL was repeated with adjusted elongation time to fit the size of gBlock precisely. Then, we used gel extraction instead of PCR clean up to extract the desirable band. An amplicon sample was taken for sequencing while the rest of the PCR product was used for golden gate cloning and transformation.
Test:
Two plates of NL-NL both had white colonies with a few blue colonies (Figure 4.6). To reduce the possibility that a colony with faulty plasmid was picked by chance, we picked a total of 16 colonies and sent for sequencing. By analysing the plasmid length (Figure 4.7), we confirmed that except one sample did not have linker-NL inserted, the length of the rest samples, 3425 bp, were similar to that of the correct double NL plasmid (3429 bp). And the sequenced plasmid matched mostly with our design (Figure 4.8). This suggests that a successful Golden Gate cloning.
Figure 4.6: DH5α transformed plates of NL-NL with a few blue colonies.
Figure 4.7: A summary of the sequence length of the 16 NL-NL miniprep samples.
Figure 4.8: Alignment of the plasmid NL-NL insert with its design on Benchling.
Learn:
Compared to our two previous attempts with PCR and Golden Gate cloning, the success of the above process demonstrated that gel extraction can more effectively eliminate incorrect amplicons than PCR cleanup, thus increasing the likelihood of successful Golden Gate cloning. Additionally, we proposed that Golden Gate cloning efficiency could be further enhanced by conducting it in two steps: first, combining the two fragments without the plasmid backbone, followed by a second Golden Gate cloning step with the plasmid backbone to insert the fragments. Due to time constraints and the fact that double NL serves only as a control in the enzymatic assay rather than a fusion protein for biomarker detection, we did not proceed with its expression and purification.References
- Mi L, Wang P, Yan J, Qian J, Lu J, Yu J, et al. A novel photoelectrochemical immunosensor by integration of nanobody and TiO 2 nanotubes for sensitive detection of serum cystatin C. Analytica chimica acta. 2016 Jan 1;902:107–14.
- Alieva NO, Konzen KA, Field SF, Meleshkevitch EA, Hunt ME, Beltran-Ramirez V, et al. Diversity and Evolution of Coral Fluorescent Proteins. El-Shemy HA, editor. PLoS ONE. 2008 Jul 16;3(7):e2680.
- Liljeruhm J, Funk SK, Tietscher S, Edlund AD, Jamal S, Wistrand-Yuen P, et al. Engineering a palette of eukaryotic chromoproteins for bacterial synthetic biology. Journal of Biological Engineering [Internet]. 2018 May 10 [cited 2019 Oct 20];12(1). Available from: https://jbioleng.biomedcentral.com/articles/10.1186/s13036-018-0100-0
- NanoLuc Sequence [Internet]. Nanolight.com. 2019 [cited 2024 Sep 23]. Available from: https://nanolight.com/content/nanoluc-sequence/
- 5.Sato W, Rasmussen M, Deich C, Engelhart AE, Adamala KP. Expanding luciferase reporter systems for cell-free protein expression. Scientific Reports [Internet]. 2022 Jul 7 [cited 2022 Nov 18];12(1):11489. Available from: https://www.nature.com/articles/s41598-022-15624-6