Team:IISc-Bangalore/Results
Results
Here, we have shown the results of the different parts - protein production and Bacterial Cellulose or Microcrystalline Cellulose Sheet Production - of our project. Go through our journey of experiments and their results as we went through multiple engineering design cycles. We present several documented results we obtained from our experiments that show the readers how far we had been able to progress.
Cloning and protein production
For the functionalization aspect of our project, we have two constructs - C1 (SpyCatcher002-dCBD) and C2 (OpdA-SpyTag002). This section mentions our efforts in order to produce, purify and characterize these fusion proteins. We first cloned these modules in a suitable vector, carried out some expression tests where we standardized the induction conditions, and then performed some solubility tests as well, in order to check if these novel fusion proteins could be purified. We were successful in figuring out suitable conditions for expression and purification of one of our constructs, while it currently is a work-in-progress for the other construct.
Cloning of C1 and C2
Constructs C1 (SpyCatcher002-dCBD) and C2 (OpdA-SpyTag002) were synthesized by GenScript and shipped to us after being cloned into pUC57-Kan vectors. We amplified the constructs using suitable primers in order to carry out Gibson assembly, by performing PCR with KOD polymerase. Gel electrophoresis results are given below:
In order to express our constructs in E. coli, we chose a pETMCN-T7 (KanR) vector, because it has a high copy number and transcription using the T7 RNA polymerase can be strongly induced upon addition of IPTG.
pETMCN-T7 was linearized by digestion with NdeI. PCR primers for amplification of C1 and C2 were designed to have overlaps with pETMCN-T7 on either side of the NdeI site. Both the PCR amplified constructs and the linearized vector were gel extracted. Gibson Assembly was performed on these samples to yield C1_pETMCN-T7 and C2_pETMCN-T7.
The above samples were transformed into E. coli DH5 \(\alpha\), inoculated into liquid culture, and miniprepped. Double digestion was performed using NheI and BglII to verify positive clones by insert release. Gel electrophoresis results are given below:
All C1 samples contained an insert of appropriate size (~1 kb) and were thus positive clones. Samples 7 and 8 were chosen for further experiments (based on concentrations).
C2 samples 3 and 5 showed anomalous bands, while the rest were positive clones (insert band around ~2kb). Samples 6 and 7 were chosen for further experiments.
Expression Tests
The chosen positive clones were transformed into E. coli BL21(DE3) for protein production. Secondary liquid cultures were induced with 0.25 mM or 0.5 mM IPTG and incubated at 23°C for 6 hrs or 30°C for 4 hrs. Cell pellets were lysed by boiling in SDS Laemmli buffer. SDS-PAGE was performed on all the samples.
No prominent band at ~25kDa (C1 molecular weight) could be observed for induced samples as compared to uninduced samples for any of the tested induction conditions.
Brighter bands at ~75kDa (C2 molecular weight) could be observed for induced samples compared to uninduced for 0.25 mM IPTG at both temperatures.
A Western blot was performed using anti-His primary antibody to detect C1, and C2 samples were taken as well.
For C1, prominent bands were observed at ~25kDa, with induction being strongest for 0.5 mM IPTG incubated at 30°C. This was taken to be the induction condition for a large-scale culture. The same conditions were finalized for C2 also since there was a prominent band at ~75kDa.
Solubility Tests
Large scale E. coli BL21(DE3) cultures (50 mL each) were set up for both C1 and C2, with induction by 0.5 mM IPTG at 30°C. After cells were harvested, they were lysed and subjected to Ni-NTA pulldown to check solubility. SDS-PAGE was performed on the boiled samples taken from different fractions (lysate, pellet, supernatant, washed and bead-bound).
Amounts of protein loaded were not normalized, and hence large blobs were observed in the pellet fractions. The bead-bound fraction for C1 showed a large band at ~25kDa, which implied that C1 was appreciably soluble. No such observation could be made for C2.
By this point, we had standardized the expression conditions for C1 and C2, and also verified that C1 was indeed soluble. To verify the same for C2, two Western blots were performed on the C2 pulldown samples, using anti-His and anti-GFP primary antibody respectively.
In the anti-GFP blot, a prominent band at ~75kDa could be observed, but there were a lot of other bands too, and similar lower molecular weight bands were observed for the anti-His blot too. This suggested that there was a lot of non-specific antibody binding, contamination, or protein degradation in case of C2. This was repeated for more dilute antibody treatments, but results looked similar, which led us to believe that C2 was indeed getting degraded.
The pulldown of C2 was performed again with an anti-GFP nanobody coupled to GST tagged glutathione-sepharose beads. SDS-PAGE was performed on the samples.
Prominent bands at ~37kDa were observed corresponding to the nanobody, but none at ~75kDa, which implied that pulldown was unsuccessful. We suspected that C2 might not be properly soluble and hence might not have been interacting with the anti-GFP nanobody tagged beads. We thus decided to modify the C2 construct.
C2(2.0)
In order to make C2 less bulky, we decided to take out the sfGFP module. This was taken into account in the design phase of the project, and KpnI restriction sites had been included on either end of the sfGFP CDS. C2_pETMCN-T7 was digested with KpnI followed by ligation to yield C2(2.0)_pETMCN-T7.
Positive clones were identified using insert release by digestion with BamHI and BglII. Transformation was done into E. coli BL21(DE3) followed by Ni-NTA pulldown as before.
No prominent band at ~50kDa (C2(2.0) molecular weight) could be observed.
Pulldown was performed again using SpyDock linked Ni-NTA beads.
Although a prominent band could be observed above ~15kDa (SpyDock molecular weight), none could be observed at ~50kDa. Till this point, we had not been able to verify solubility of either C2 or C2(2.0). We suspected that the hydrophobic signal peptide attached to OpdA might be leading to aggregation, so we decided to modify C2 yet again.
In the meanwhile, we also decided to conduct some assays on the crude extract of the large-scale secondary cultures. The assays we did were the Bradford Assay and the assay of the OpdA domain using Coumaphos. We also decided to grow the bacteria in both LB and Terrific Broth (TB), because a short literature review showed that some people have used TB as the growth medium for the secondary culture for the OpdA production
We used both the C2 and the C2 (2.0) constructs for this purpose.
For this, we harvested the protein from the cells. On the crude cell extract, we conducted the Bradford assay, using BSA as the reference for the standard curve. The standard curve is shown below:
| Sample | Adjusted Absorbance | Estimated Concentration (mg/mL) |
|---|---|---|
| LB EV | 0.9034 | 0.5937 |
| LB C2 | 0.8659 | 0.5445 |
| LB C2 (2.0) | 0.8607 | 0.5377 |
| LB NT | 0.9296 | 0.6281 |
| TB EV | 0.9449 | 0.6482 |
| TB C2 | 1.0485 | 0.7841 |
| TB C2 (2.0) | 0.729 | 0.3648 |
| TB NT | 1.109 | 0.8636 |
In the above table, EV refers to Empty Vector samples. NT refers to non-transformed samples.
Surprisingly, the non-transformed TB culture had the highest protein concentration, while the C2 (2.0) cultures had the lowest in their corresponding set, be it LB or TB.
For Enzyme assay, we dissolved Coumaphos in Ches Buffer and used it as a substrate. The product, chlorferon has an absorption peak at 348 nm. The absorbance values of the different samples are as follows:
| Culture Medium | EV | C2 | C2 (2.0) | NT |
|---|---|---|---|---|
| LB 1 | -0.189 | 0.111 | 0.188 | Used as Blank |
| LB 2 | 0.824 | 0.314 | 0.09 | Used as Blank |
| TB 1 | 0.241 | 0.647 | 1.599 | Used as Blank |
| TB 2 | 1.333 | 1.240 | 0.683 | Used as Blank |
LB 1 and LB2 represent replicates, so do TB1 and TB2 for the corresponding cultures.
A very broad trend that was easily seen was that enzyme production seemed to be higher in cultures grown in TB than those grown in LB. So, we can continue using TB as the growth medium for our protein's expression.
We normalized the absorbance of the test samples at 348 nm (corresponding to lambda max of chlorferon, the degradation product of assay substrate coumaphos) with their respective protein concentrations as deduced from the Bradford assay. In OpdA assay, NT was used as the blank (thus, futile to deal with its absorbance, be it normalized or not). The normalized absorbance (in arbitrary units) values are tabulated below:
| Culture Medium | EV | C2 | C2(2.0) | NT |
|---|---|---|---|---|
| LB 1 | -0.31834 | 0.20386 | 0.34964 | Used as Blank |
| LB 2 | 1.38791 | 0.57668 | 0.16738 | Used as Blank |
| Average of the values | 0.53479 | 0.39027 | 0.25851 | - |
| TB 1 | 0.37180 | 0.82515 | 4.38322 | Used as Blank |
| TB 2 | 2.05646 | 1.58143 | 1.87226 | Used as Blank |
| Average of the values | 1.21413 | 1.203 | 3.1277 | - |
Here '1' and '2' simply denote two replicates of the OpdA assay with the samples mentioned, in chronological order.
To understand the trend exhibited by the average values, a bar chart was plotted:
The normalized absorbances of the degradation product can be considered as a measure for enzyme activity in the crude cell extract. As can be observed from the graph, when grown in LB, the enzyme activity seemed to decrease from the Empty Vector control to C2 to C2 (2.0). However, when grown in Terrific Broth (TB), while there is almost no change in enzyme activity between C2 and Empty Vector control, there is a huge increase in enzymatic activity in case of C2 (2.0).
For our interpretation of these results, please visit the Proof of Concept page.
C2(3.0)
In order to take out the OpdA signal peptide, a PCR was performed on C2(2.0)_pETMCN-T7 to amplify C2(2.0) minus the N-terminal signal. This PCR fragment was cloned into an empty pETMCN-T7 vector to yield C2(3.0)_pETMCN-T7.
We are in the "Build" phase of this engineering cycle. We are yet to clone the bacteria into BL21(DE3) cells. So, the progress on this is not significant enough to give us any conclusive findings
Making microcrystalline cellulose sheet from MCC powder
We also tried to make a sheet of microcrystalline cellulose (MCC) from the commercially available powdered form. Powder of MCC was purchased from Sigma Aldrich & Co. Initially, we started out on finding a suitable solvent for dissolving MCC powder. Although data is available in literature in this regard, they are not exhaustive and are often ambiguous. Based on literature review, we decided to focus on DMSO (dimethyl sulfoxide), NMP (N-methyl pyrrolidinone and a solution of 7% LiCl in DMAc (dimethyl acetamide) as probable solvents for further investigation.
MCC was mixed with these solvents and stirred on a magnetic stirrer overnight. The results were as follows:
| Solution | Solute | Solvent | Stirring Speed | Results | Inference |
|---|---|---|---|---|---|
| Cellulose + DMSO | 1 g MCC powder | 10 mL DMSO | 450 rpm | Maximum cloudiness | Maximum solubility of MCC |
| Cellulose + 7% LiCl in DMAc | 1 g MCC powder | 7% LiCl in DMAc | 450 rpm | Intermediate cloudiness | Intermediate solubility of MCC |
| Cellulose + NMP | 1 g MCC powder | 10 mL NMP | 460 rpm | Minimum cloudiness | Minimum solubility of MCC |
To test whether it was possible to form a mechanically strong film of bacterial cellulose from the solutions in these solvents, we decided to perform a standard test used by materials engineers for understanding this - drop casting. This small scale test saves wastage of resources by doing a random hit-and-trial with huge volumes of samples. A drop of the sample is spread out over a glass slide and heated at a temperature about 20 oC below its boiling point, and then subjected to a variety of treatment conditions. The film formed on the glass slide is observed.
Drop-castings were performed with different sample volumes. Since MCC was most soluble in DMSO among the three solvents, we decided to investigate the influence of different heating durations on the integrity of the film formed from this solution. The results of our drop-castings using the above mentioned solvents are enlisted below:
| Label | Solvent | Temp. used | Volume used | Time of heating | Observations |
| DMSO 1 | DMSO | 170°C | 1 mL | 30 s | Cracked when quenched |
| DMSO 2 | DMSO | 170°C | 0.5 mL | 20 s | Large cracks; Highly non-uniform thickness |
| DMSO 3 | DMSO | 170°C | 1 mL | 35 s | Large cracks |
| NMP 1 | NMP | 175°C | 0.6 mL | 20 s | Microcracks on visual examination after annealing. Excessively flaky. |
| LiCl/DMAc 1 | 7% LiCl in DMAc | 140°C | 1 mL | 45 s | Negligible presence of cracks. A few small bubbles; absorbed moisture after annealing; after heating for more than 850 s, turned yellow |
| LiCl/DMAc 2 | 7% LiCl in DMAc | 140°C | 0.5 mL | 420 s | Almost no cracks; bubbles; absorbed moisture |
Next, we sought to see if the addition of a binder like PVA (polyvinyl alcohol) could influence the mechanical integrity of the film and strengthen it. We tried to carry out drop-castings using the same solutions, but with 10 % w/v PVA added to it. Since quenching gave better results previously, all the samples were quenched on a pool of acetone after heating to evaporate the solvent. Temperatures for heating were same as before. However, we found that all the drop-casted films were patchy and flaky and had turned brownish, except for the solution containing DMSO as solvent, which formed a homogenous sheet, albeit with a brownish hue.
Having been slightly surprised at these results, we were intrigued to find out the possible cause(s) behind this. PVA is known to be strongly preferring the formation of mechanically strong films. The results therefore seem counter-intuitive at the first glance. However, a survey of existing literature and the materials data sheets provided by companies led us to conclude that the heating regime used for most of the samples had exceeded the degradation temperature of PVA, while the heating temperature for DMSO was slightly lower than the average degradation temperature. The brownish colour was most likely imparted due to the influence of the degradation products of PVA.
We therefore decided to redo our experiment for PVA by using a lower temperature. Although this would prolong the heating duration, we expected to see a stable film being formed, devoid of the brownish hue of the previous case. Our hypothesis was indeed validated in the next experiment
| Solution | Heating | Observations |
| DMSO + 10% PVA + MCC | Heated on Hot Plate gradually till 90°C for 8.5 min and then rapidly quenched | Homogenous sheet |
We have thus tried to optimise the conditions for making MCC sheet from commercially available powder. Further studies need to be done by drop-casting on a block of PDMS or inside a PDMS tube instead of a glass slide. Detailed studies need to also be performed on the interesting solvent system of 7% LiCl in DMAc, because the film formed in this case absorbs moisture due to the presence of hygroscopic LiCl and spontaneously gets healed of any cracks present.
Overproduction of bacterial cellulose
Knocking out celG from A. tumefaciens for BC overproduction
We decided to knock out the celG gene from the chvB- mutant A. tumefaciens C58 via FLP-FRT recombination.
For this purpose, the lambda red recombination cassette of pKD46(AmpR) was amplified using gradient PCR and put under tetR promoter in pST-KT. We also planned to amplify the complement1-FRT-CamR-FRT-complement2 fragment which would replace the part of the celG sequence we wished to remove. The complement1 and complement2 are DNA sequences that are complementary to the regions upstream and downstream of the portion we want to knock out.
We were able to successfully amplify the lambda red recombination cassette. Based on the gel images, we can see that the PCR product amount was highest at 50°C. No product was obtained at all at 52°C and 58°C.
We couldn't amplify the FRT-complements fragment. Based on the gel images, we believe this is probably due to the formation of primer-dimers.
The bacterial cultures were received from Prof. Ann G. Matthysse. However, we could not revive the bacteria on the LB Agar plates despite following the instructions sent by Prof. Matthysse. She said that this might be because the cultures sent were too old.
We are still figuring out how to proceed with this part of the plan. In the meanwhile, we decided to go for an alternate strategy for producing bacterial cellulose, the results of which have been provided at the end.
Using K. xylinus for BC production
To produce Bacterial Cellulose sheet, we decided to grow a culture of Komagataeibacter xylinus (also known as A. xylinum) as it is known to overproduce bacterial cellulose when grown in the Hestrin-Schramm (HS) culture medium at 28°C, as per the paper Keshk, S. M. A. S. (2006)
Thus, we decided to try growing the bacteria in the HS medium.
We first ordered an Agar slant of the ATCC 11142 strain of the bacteria from the National Chemical Laboratories in Pune.
As per the HS media preparation protocol of the 2020 UCSC iGEM Team, we prepared the HS medium and inoculated our bacteria into that. However, even after 3-4 days, no growth was observed, which was concerning given that the bacterium has a generation time of about 3 days.
We then decided to try the medium, which was recommended by NCL, Pune to be used to revive or store the culture. This medium, unlike the HS medium, has no buffer components and has sorbitol instead of glucose as the carbon source. This medium, we hereafter refer to as the "Sorbitol medium".
We tried to grow our bacteria in the Sorbitol medium, and the bacteria successfully grew in the medium.
However, there were a host of other complications:
- Being an extremely general medium containing nothing but yeast extract, sorbitol medium is highly prone to contamination. Many of our samples and even some of the medium prepared by us got contaminated (Figure 2 and Figure 3).
- Bacterial Cellulose production was not observed in Sorbitol medium as well despite the bacteria growing. This is probably because Sorbitol is not such a good carbon source for cellulose production.
After several unsuccessful attempts at growing the bacteria in HS media or getting to produce bacterial cellulose in Sorbitol medium, we talked to Dr. Mudrika Khandelwal and Dr. Shivakalyani Adepu from The Cellulose Group at the Department of Materials Science and Metallurgical Engineering of IIT Hyderabad.
Based on their suggestion and using the protocol provided in their paper (Adepu, S. 2020), we changed the composition of the HS medium to have 3.4 g of Na2HPO4 instead of 2.7 g and pH 4.5 instead of 6. We then tried to grow our bacteria in this medium and we were successful!
In fact, we inoculated two cultures, one in the HS medium and one in the Sorbitol medium. While the one in the Sorbitol medium got contaminated as shown by the rapid increase in OD600, the one in HS medium grew at a slow and steady state, showing that Sorbitol medium is much more prone to contamination.
However, we now tried to follow their protocol to setup secondary cultures for bacterial cellulose production. But, even 2 weeks later, there were no changes observed in the secondary culture (Figure 4).
References:
- Keshk, S. M. A. S. "Physical properties of bacterial cellulose sheets produced in presence of lignosulfonate." Enzyme and Microbial Technology 40.1 (2006): 9-12. doi.org/10.1016/j.enzmictec.2006.07.038
- Adepu, S., & Khandelwal, M. (2020). Ex-situ modification of bacterial cellulose for immediate and sustained drug release with insights into release mechanism. Carbohydrate Polymers, 249, 116816. doi.org/10.1016/j.carbpol.2020.116816