Team:Edinburgh/Results
Results
What did we aim to do?
We aimed to:
1. Characterise the different silica-binding tags (L2, L2NC, Car9, Sb7)
2. Clone and express enzymes fused with silica-binding tags
3. Successfully immobilise our enzymes to silica surfaces
4. Characterise the effect of tag addition and immobilisation on activity and stability (to conditions/over time)
5. Apply directed evolution methods to increase the activity of our tagged enzymes for application to the SuperGrinder
What did we find?
We will summarise our findings relating to the immobilisation of the cellulase Cex:
Cex-Car9 showed low solubility at the SDS page assay and reduced activity against the substrate. Indeed, the negative-charges of Car9 seem to hinder the catalytic site (rich in positive charges). On the other hand, L2NC tag seems to increase the solubility of Cex and consequently the activity, even rescuing the non-functional Cex phenotype we came across. L2NC tag very similar results were also observed when fused to LCC and PETase enzymes. This leads to the conclusion that L2NC is the best candidate for the increase, or in any case lack of detrimental effect, of activity for these enzymes.
Although data were insufficient to quantify binding affinities and thus state which tag would bind to the silica beads or glass with the strongest affinity, preliminary qualitative data seem to suggest the L2 tag and the L2NC tag to be the most efficient at binding the acid-scarified glass surface. More studies with a higher number of replicates and normalisation adjustment will be needed in order to define the best candidate. The “proof of concept” page of our wiki shows more detailed data about the immobilisation of the enzymes on glass and silica beads, showcasing promising results.
Select the section of results you would like to read more about...
To first assess the silica-binding tags, we assembled constructs to express GFP under the T7 promoter, with each tag at the C-terminal (except for L2 which was received from Marcos Valenzuela-Ortega as an N-terminal fusion already). We transformed assemblies into E. coli DH5α. As the intended final assembly contained sfGFP, we used a plasmid backbone with a lacZ’ reporter gene (pJUMP29-1A (lacZ’)) to enable blue-white screening. After transformation into E. coli DH5α, colonies with an insert replacing the lacZ’ reporter gene were unable to produce functional β-galactosidase to metabolise X-gal to an insoluble blue pigment, thus appeared white (Figure 1). Screening plates had an average ratio of 1:12 blue to white colonies, showing that assembly was successful. Colony PCR performed on selected white colonies and analysed by gel electrophoresis showed a single band of product for each colony matching the expected insert sizes (Figure 2A). Nucleotide sequence was confirmed by Sanger sequencing.
Figure 1 Example of blue-white screening. 10% cells plated on selective LB agar (kanamycin 50 μg/ml).Left: Uncut backbone transformation (no enzyme control). All colonies blue. Right: selection plate for assembly of sfGFP-linker-L2NC with 19 blue colonies and 239 white.
We used E. coli BL21 (DE3) for expression of constructs, induced with IPTG (Figure 2B). Cells expressing L2-sfGFP (blue line) showed a notably slower growth rate than those expressing the smaller tags, hypothesised to be due to the large size, or some toxic effects of overexpression (Figure 2C). L2-sfGFP cultures also demonstrated lower fluorescence activity. This was supported by observations by Taniguchi et al, 2007 [1] who found that L2 reduced fluorescence activity of GFP by 30% when used as a fusion tag. Insoluble fractions recovered after lysis were also larger (data not shown), suggesting that the L2 tag decreased solubility of the sfGFP, reducing fluorescence activity and perhaps putting higher metabolic burden on cells. This would not be surprising as the L2 tag is larger than sfGFP (273 vs 237 amino acids), so is likely to impact folding. Fluorescence activity and growth seemed to be less impacted by the smaller tags L2NC and Car9.
Figure 2 A Colony PCR using PS1/PS2 primers showed expected band sizes for all inserts (3 colonies tested per strain). B. Images post-induction by 1 mM IPTG (top= cell pellet, bottom=cell culture). Cultures with different tags show varying GFP fluorescence levels. C. Growth curves of strains during induction (time of induction marked with blue arrow, OD600=0.5-0.6). Induced cultures are marked by dots/lines, uninduced cultures marked by crosses/dashed lines D. SDS-PAGE analysis of solubility shows tagged GFP present in both soluble and insoluble fractions Expected sizes sfGFP (27.1kDa), L2-sfGFP (57.7kDa), sfGFP-Car9 (28.7kDa), sfGFP-L2NC (41.3kDa). Compared to NEB protein ladder.
Cells were lysed by BugBuster (as a quick-and-easy method), or sonication (allowing us to control the “binding buffer” composition), and lysates incubated with different types of silica beads for different incubation times. Beads were washed using “wash buffer” and then visualised to observe residual GFP fluorescence. We also compared the difference in fluorescence between lysate samples incubated in the presence and absence of silica beads to quantify the amount of GFP removed from solution when beads were present (Table 1).
| Removal of GFP fluorescence from solution (%) | ||
| Bead type | L2-sfGFP | Untagged sfGFP |
| 500 μm | 27.0 | 12.0 |
| Celite 545 | 23.0 | 5.5 |
Proteins successfully immobilised to beads were detected by SDS-PAGE. Beads were boiled in reducing sample buffer to release proteins adsorbed to the surface. Protein bands of the expected sizes were released from bead samples, especially Celite beads, after boiling in reducing sample buffer (Figure 3), although other contaminating bands with lower intensities were also observed in some cases (data not shown). No sfGFP was detectable on beads incubated with untagged samples. This suggests that the reduction in fluorescence observed for untagged sfGFP (5.5-12%) in Table 1 could result from natural reduction in fluorescence activity over time, rather than from immobilisation.
Figure 3 SDS-PAGE of protein samples released from Celite beads, enabling comparison of silica binding tags. Boxes surround bands of expected sizes.
We also tested removal of fluorescence from solution using different quantities of silica beads, which showed that the fluorescence removal scaled with bead quantity (Figure 4).
Figure 4 A. GFP removal from solution after incubation of 600 μL clarified cell lysate with three different quantities of beads, overnight at 4°C (low 10 mg, medium 50 mg, high 100 mg). Fluorescence values are normalised to an arbitrary value of 1 in the absence of silica.
Overall our results indicated that the L2NC tag with the addition of the short amino-acid linker was the silica-binding tag with the highest affinity, and the immobilisation worked best with the Celite 545 beads. Further results and comparisons of the tags are demonstrated on the “Proof of concept” page.
References:
[1] Taniguchi K, Nomura K, Hata Y, Nishimura T, Asami Y, Kuroda A. The Si-tag for immobilizing proteins on a silica surface. Biotechnol Bioeng [Internet]. 2007 Apr 15 [cited 2021 Apr 23];96(6):1023–9. Available from: http://doi.wiley.com/10.1002/bit.21208
To successfully heterologously express proteases such as KerA, it was necessary to screen different secretion signal peptides. Secretion is important to reduce host toxicity of over-expressed proteases and avoid expression in intracellular inclusion bodies.
Protein secretion screen
We screened for KerA-secreting clones by comparing zones of clearance on milk agar plates. Whilst hydrolysis of milk does not test specificity toward keratinous substrates, the screen is a quick and easy measure of secreted proteolytic activity. A control plate comparing wild-type strains of Bacillus subtilis which had known protease secretion abilities confirmed that the screen worked: zones of clearance were observed around protease secreting strains, but not around those that were deficient in secreted proteases (B. subtilis SCK6 and IIG-Bs27) (Figure 1)
Figure 1 Protease secretion from three different strains of Bacillus subtilis, observed by measuring zones of clearance on a milk agar plate (1% skimmed milk powder). Protease secretion was only observed from B. subtilis 168 as expected.
We thus tested the addition of eight different signal peptides to KerA, expressed in E. coli. Signal peptides were already assembled in JUMP Level 0 plasmids by Marcos Valenzuela-Ortega (Chris French lab, University of Edinburgh), summarised in Table 1. We also tested secretion of KerA using the native KerA signal prepeptide from B. licheniformis. Colony PCR after JUMP assembly and transformation verified that insert sizes of positive transformants were as expected (Figure 2). Nucleotide sequences were confirmed by Sanger sequencing.
Figure 2 Figure 2: Colony PCR from positive transformants using PS1/PS2 primer pair. Gel showed expected insert sizes for KerA assembled with different signal peptides. Signal peptides used are as follows, with expected size in parentheses: K1.PelB (1581), K2.PelB_D5 (1599), K3.CelCD_G20 (1587), K4.CelCD_Gdom (2649), K5.CelCD_N29 (1599), K6.OsmY (2130), K7.YebF (1872), K8.Spy (2001), KK. KerA native signal peptide (1602), G. Empty backbone (pJUMP29-1B) (1161) - negative control with no template DNA added (just H2O).
Wild-type (untransformed) E. coli BL21 (DE3), and transformants with the empty pJUMP29-1B backbone did not produce a halo, however some assemblies of KerA with signal peptides produced significant zones of clearance on milk plates after incubation (30°C, 48-72 hrs), therefore the screening methodology was considered an appropriate initial selection. Secretion-positive cultures were selected for further assays on proteolytic activity.
The results shown in Figure 3 show the halo screen of both the individually assembled SP-KerA constructs, and a second assembly method in which all 8 signal peptides were added to the same reaction (SP-mix). The assembly product was then directly transformed into E.coli BL21 (DE3) to enable expression and screening on the primary transformation plate. The two methods showed slightly different results for unknown reasons. Figure 3E summarises the differences. Overall the secretion signal with most evidence for secretion was Cel-CD_G20.
Figure 3 A Colony PCR using PS1/PS2 primer pair on assemblies of KerA with different signal peptides in E. coli DH5a. Expected insert sizes are shown above in bp. Three colonies were selected for each assembly. B. Individually assembled KerA with different signal peptides, including green colonies (1G, bottom right) as negative control. Incubated at 30°C, 60 h. C. Primary screening plate from mixed signal peptide assembly transformed directly into E. coli BL21 (DE3) (1% milk agar plate with kanamycin selection, incubated 28°C, 72 h). D. Re-patched colonies selected from primary screening plate, numbered matches labelling on primary screening plate. Three “halo negative” colonies from the primary plate were also selected (1N, 2N, 3N) and a green colony with no insert (1G) as negative controls. Imaged after incubation at 30°C, 72 h. E. Table of clones positive for protease secretion on either the individually assembled plate or the mix plate. X signifies that a reproducible halo was observed. (x) signifies that a halo was observed but that it was not reproducible.
Method discussion:
To screen for secretion from B. subtilis we would use B. subtilis IIG-Bs27-31 as a host because untransformed cells showed no clearance zone in initial tests, therefore had no background activity. Interestingly, halos were observed after the same period of time regardless of whether IPTG was applied for induction (40 μl to top, 100 mM). This could suggest that the milk in the plate could substitute for IPTG (an analogue of lactose). This finding is potentially useful for lowering production costs of inducible enzymes in industry.
Table 1
| Signal Peptide | Sequence | Ref. | |
|---|---|---|---|
| K1 | PelB_N (Erwinia carotovora) | AATGAAATACCTGCTGCCGACCGCTGCTGCTGGTCTGCTGCTCCTCGCTGCCCAGCCGGCGATGGCAGCC | (Wang et al, 2016)
22 residues as in pET27b. Added 5’ A |
| K2 | PelB-D5 (Erwinia carotovora, modified) | AATGAAATACCTGCTGCCGACCGCTGCTGCTGGTCTGCTGCTCCTCGCTGCCCAGCCGGCGATGGCAGATGATGACGATGATGGAGCC | (Kim et al, 2015)
22 residues as in pet27b. Added 5’ DDDDDGA |
| K3 | Cel-CD-Gdom (Bacillus sp.) | AATGGAAGGAAACACTCGTGAAGATAATTTTAAACATTTATTAGGTAATGACAATGTTAAACGCCCTTCTGAGGCTGGCGCATTACAATTACAAGAAGTCGATGGACAAATGACATTAGTAGATCAACATGGAGAAAAAATTCAATTACGTGGAATGAGTACACACGGATTACAATGGTTTCCTGAGATCTTGAATGATAACGCATACAAAGCTCTTGCTAACGATTGGGAATCAAATATGATTCGTCTAGCTATGTATGTCGGTGAAAATGGCTATGCTTCAAATCCAGAGCTGATTAAAAGCAGAGTCATTAAAGGAATAGATCTTGCTATTGAAAATGACATGTATGTCATCGTTGATTGGCATGTACATGCACCTGGTGATCCTAGAGATCCCGTTTACGCTGGAGCAGAAGATTTCTTTAGAGATATTGCAGCATTATATCCTAACAATCCACACATTATTTATGAGTTAGCGAATGAGCCAAGTAGTAACAATAATGGTGGAGCTGGGATTCCAAATAATGAAGAAGGTTGGAATGCGGTAAAAGAATACGCTGATCCAATTGTAGAAATGTTACGTGATAGCGGGAACGCAGATGACAATATTATCATTGTGGGTAGTCCAAACTGGAGTCAGCGTCCTGACTTAGCAGCTGATAATCCAATTGATGATCACCATACAATGTATACTGTTCACTTCTACACTGGTTCACATGCTGCTTCAACTGAAAGCTATCCGCCTGAAACTCCTAACTCTGAAAGAGGAAACGTAATGAGTAACACTCGTTATGCGTTAGAAAACGGAGTAGCAGTATTTGCAACAGAGTGGGGAACTAGCCAAGCAAATGGAGATGGTGGTCCTTACTTTGATGAAGCAGATGTATGGATTGAGTTTTTAAATGAAAACAACATTAGCTGGGCTAACTGGTCTTTAACGAATAAAAATGAAGTATCTGGTGCATTTACACCATTCGAGTTAGGTAAGTCTAACGCAACAAGTCTTGACCCAGGGCCAGACCAAGTATGGGTACCAGAAGAGTTAAGTCTTTCTGGAGAATATGTACGTGCTCGTATTAAAGGTGTGAACTATGAGCCAATCGACCGTACAAAATACACGAAAGTACTTGGCGGAGCC | (Gao et al, 2016)
Domain as used by authors (M + 375 from catalytic domain), +GGA M15743.1
|
| K4 | Cel-CD-G20 (Bacillus sp.) | AATGGAAGGAAACACTCGTGAAGATAATTTTAAACATTTATTAGGTAATGACAATGTTAAACGCAAACGCGGAGCC | (Gao et al, 2016)
Peptide used by authors: M+20 residues from catalytic domain + KRGA GeneBank ID M15743.1 |
| K5 | CelCD-N29 (Bacillus sp.) | AATGCTCAGAAAGAAAACAAAGCAGTTGATTTCTTCCATTCTTATTTTAGTTTTACTTCTATCTTTATTTCCGACAGCTCTTGCAGCC | (Gao et al, 2016)
29 residue signal peptide of original gene GeneBank ID M15743.1 |
| K6 | OsmY (E. coli MG1655) | AATGACTATGACAAGACTGAAGATTTCGAAAACTCTGCTGGCTGTAATGTTGACCTCTGCCGTCGCGACCGGCTCTGCCTACGCGGAAAACAACGCGCAGACTACCAATGAAAGCGCAGGGCAAAAAGTCGATAGCTCTATGAATAAAGTCGGTAATTTCATGGATGACAGCGCCATCACCGCGAAAGTGAAGGCGGCCCTGGTGGATCATGACAACATCAAGAGCACCGATATCTCTGTAAAAACCGATCAAAAAGTCGTGACCCTGAGCGGTTTCGTTGAAAGCCAGGCCCAGGCCGAAGAGGCAGTGAAAGTGGCGAAAGGCGTTGAAGGGGTGACCTCTGTCAGCGACAAACTGCACGTTCGCGACGCTAAAGAAGGCTCGGTGAAGGGCTACGCGGGTGACACCGCCACCACCAGTGAAATCAAAGCCAAACTGCTGGCGGACGATATCGTCCCTTCCCGTCATGTGAAAGTTGAAACCACCGACGGCGTGGTTCAGCTCTCCGGTACCGTCGATTCTCAGGCACAAAGTGACCGTGCTGAAAGTATCGCCAAAGCGGTAGATGGTGTGAAAAGCGTTAAAAATGATCTGAAAACTAAGGGCAGCGGATCAGCC | (Bokinsky et al, 2011)
Complete gene (201 residues) + GSGSA GeneID: 948895 |
| K7 | YebF (E. coli MG1655) | AATGAAAAAAAGAGGGGCGTTTTTAGGGCTGTTGTTGGTTTCTGCCTGCGCATCAGTTTTCGCTGCCAATAATGAAACCAGCAAGTCGGTCACTTTCCCAAAGTGTGAAGATCTGGATGCTGCCGGAATTGCCGCGAGCGTAAAACGTGATTATCAACAAAATCGCGTGGCGCGTTGGGCAGATGATCAAAAAATTGTCGGTCAGGCCGATCCCGTGGCTTGGGTCAGTTTGCAGGACATTCAGGGTAAAGATGATAAATGGTCAGTACCGCTAACCGTGCGTGGTAAAAGTGCCGATATTCATTACCAGGTCAGCGTGGACTGCAAAGCGGGAATGGCGGAATATCAGCGGCGTTCAGCC | (Natarajan et al, 2017)
Complete gene (118 residues) + SA GeneID: 946363
|
| K8 | Spy (E. coli MG1655) | AATGCGTAAATTAACTGCACTGTTTGTTGCCTCTACCCTGGCTCTTGGCGCGGCTAACCTGGCCCATGCCGCAGACACCACTACCGCAGCACCGGCTGACGCGAAGCCGATGATGCACCACAAAGGCAAGTTCGGTCCGCATCAGGACATGATGTTCAAAGACCTGAACCTGACCGACGCGCAGAAACAGCAGATCCGCGAAATCATGAAAGGCCAGCGTGACCAGATGAAACGTCCGCCGCTGGAAGAACGCCGCGCAATGCATGACATCATTGCCAGCGATACCTTCGATAAAGTAAAAGCTGAAGCGCAGATCGCAAAAATGGAAGAACAGCGCAAAGCTAACATGCTGGCGCACATGGAAACCCAGAACAAAATTTACAACATCCTGACGCCGGAACAGAAAAAGCAATTTAATGCTAATTTTGAGAAGCGTCTGACAGAACGTCCAGCGGCAAAAGGTAAAATGCCTGCAACTGCTGAATCAGCC | (Tsirigotaki et al, 2017)
Complete gene (161 residues) + SA GeneID: 946253
|
| KK | KerA_SP (Bacillus licheniformis) | ATGATGAGGAAAAAGAGTTTTTGGCTTGGGATGCTGACGGCCTTCATGCTCGTGTTCACGATGGCATTCAGCGATTCCGCTTCAGCAGCC | (Lin et al, 1995)
Annotated prepeptide from native KerA gene |
Immobilisation of tagged enzymes to the beads was mainly detected using activity assays after immobilisation, compared to an untagged version (see next section).
We also began to develop a protocol for high-throughput screening of immobilised enzymes by immobilising tagged enzymes recovered from a directed evolution library to glass slides. This can be done by immersion in lysis solution. The immobilised enzymes on the glass slide could then be used for enzymatic assays in situ, and matched back to the parent colonies from position on the plate. This would be a highly valuable application of these silica-binding tags., enabling screening for hundreds of variants per plate. Moreover, when selecting for enzymes with higher activity, it is beneficial to be able to screen in their immobilised state as this is the condition for their final application.
Immobilisation to a Glass Surface results
In order to assess whether Si-tagged proteins were immobilising to the glass surface, various Si-tagged sfGFP constructs were tested due to the fluorescent protein’s easy readout in a UV transilluminator. Through iterations, various factors such as pH of the Tris-HCl binding buffer, salt concentration upon binding and the acid-treatment of the glass surface, proved to be important to the effective binding of Si-tagged sfGFP.
Figure 1 Immobilisation of Si-tagged sfGFP to a Treated Glass Surface. sfGFP tagged with four different Si-tags (Car9, L2NC-linked, L2NC and L2) as well as a negative control of untagged sfGFP were aliquoted on to a glass slide, washed and a visualised under UV light. After wash acquisition is longer than the others to effectively visualise the fluorescence on the slide.
Figure 1 suggests that there is evidence of the Si-tagged protein binding to the treated glass surface. Different methods were tried, although the final method closely followed Taniguchi et al. with some modifications [1]. Infact, increasing the pH of the Tris buffer was believed to be appropriate considering the high prevalence of positively charged amino acid residues in the Si-tags and is more common in the supporting literature [2,3,4]. Additionally, the acid-treatment of the microscope glass slides improved the attachment on the glass slide. Indeed, the exposure to acid removes alkaline salts from the crystalline structure of the glass, being replaced by H+ ions, and increasing silanol groups exposition [5]. A similar experiment was run with the cells lysed by sonication in the 25mM Tris-HCl (pH 8.0) binding buffer without the added salts. None of the samples were able to adhere to the glass. Additionally, an alternative lysing solution, BugBuster, was used instead of B-PER as suggested by Taniguchi and colleagues for cost and availability reasons. No other lysing solutions were tested, and considering BugBuster’s proprietary contents, it may have affected the appropriate binding of the Si-tag to the surface. However, both the lysis method and solutions vary in this experiment and therefore cannot be directly compared, we hypothesized that the added salt increases the specificity of the Si-tag to the silanol group of the glass surface, but more experiments must follow to confirm or deny it.
Eventually, the L2-tagged sfGFP showed the highest fluorescence in comparison to L2NC and Car9 after the wash step, pointing to the conclusion that it most strongly binds to the glass surface. Nonetheless, these results must be considered qualitative as this reflects a degree of fluorescence after a wash step, only. In addition, protein concentrations were not standardised, which may have a large impact on the amount of fluorescence visualised. Instead, the resuspension of the wet cell weight was taken to standardise the samples. Moreover, the samples were only washed in the wash buffer for 15 minutes, whereas Taniguchi and colleagues washed their slide for 24 hours and still visualised their Si-tagged sfGFP constructs [1]. This could be explained by the authors having more sophisticated equipment for their experiment. In addition, longer washings drastically decrease the chance of non-specific binding, allowing for more appropriate conclusions to be drawn.
Future Perspectives for the High-Throughput Screening of Immobilised Enzymes
The evidence of immobilisation of Si-tagged proteins on a glass surface is a promising first step towards the high-throughput screening of immobilised enzymes. Future experiments would include binding buffer optimization by varying the concentrations of Tris-HCl (pH 8.0), NaCl and other lysis buffers including the B-PER lysing solution. Indeed, as the lysing solution comprises nearly half of the original mixture, its alteration could have a large impact on the binding capability of the tag.
Additionally, as the treatment of the glass surfaces was suggested to have found a positive impact on protein immobilisation, further treatment techniques are worth pursuing. Infact, increasing the concentration of HCl and exposing the glass slide to the acid for longer amounts of time could further activate the glass slide, thereby increasing the yield of protein immobilisation. Future experiments with this consideration may prove pivotal
1. Taniguchi K, Nomura K, Hata Y, Nishimura T, Asami Y, Kuroda A. The Si-tag for immobilizing proteins on a silica surface. Biotechnol Bioeng [Internet]. 2007 Apr 15 [cited 2021 Apr 23];96(6):1023–9. Available from: http://doi.wiley.com/10.1002/bit.21208
2. Kim S, Joo K Il, Jo BH, Cha HJ. Stability-Controllable Self Immobilization of Carbonic Anhydrase Fused with a Silica-Binding Tag onto Diatom Biosilica for Enzymatic CO2Capture and Utilization. ACS Appl Mater Interfaces [Internet]. 2020 Jun 17 [cited 2021 Apr 24];12(24):27055–63. Available from: https://dx.doi.org/10.1021/acsami.0c03804
3. Coyle BL, Baneyx F. A cleavable silica-binding affinity tag for rapid and inexpensive protein purification. Biotechnol Bioeng [Internet]. 2014 Oct 1 [cited 2021 Apr 23];111(10):2019–26.
4. Rauwolf S, Steegmüller T, Schwaminger SP, Berensmeier S. Purification of a peptide tagged protein via an affinity chromatographic process with underivatized silica. Eng Life Sci [Internet]. 2021 [cited 2021 Aug 13]; Available from: https://onlinelibrary.wiley.com/doi/full/10.1002/elsc.202100019 32.
5. Szalóki M, Hegedűs V, Fodor T, Martos R, Radics T, Hegedűs C, et al. Evaluation of the Effect of the Microscopic Glass Surface Protonation on the Hard Tissue Thin Section Preparation. Appl Sci 2020, Vol 10, Page 7742 [Internet]. 2020 Nov 1 [cited 2021 Aug 18];10(21):7742. Available from: https://www.mdpi.com/2076-3417/10/21/7742/htm
Adding fusion tags to an enzyme, and immobilising it to a surface can impact activity. Taniguchi et al, 2007 [1] suggested that the largest tag (L2) decreased luciferase activity by 60-70%, suggesting that the large tag may have a greater impact on enzymes which require conformational changes for activity than it had on sfGFP fluorescence activity (demonstrating just 30% reduction). It is important to test the impact on our enzymes to aid our assessment of how feasible the overall SuperGrinder will be based on observed performance. Our preliminary results from activity assays are below, with final results on the Proof of Concept page.
Cex (from Cellulomonas fimi) Immobilisation to Silica Beads
The immobilisation protocol was followed with untagged Cex, Cex-Car9 and Cex-L2NC and the enzymatic activity of the supernatant was tested before and after incubation with the beads, the subsequent wash buffers, and the beads themselves. Two assays were performed on the samples to test both the cellulolytic and xylanolytic activity of Cex. In order to assess the activity of the enzyme on the beads, the reaction should ideally be constantly kept in suspension, so as to avoid the sedimentation of the beads to the bottom of the reaction tube. After an incubation period, the reaction should be spun down, and the supernatant taken so as to avoid possible interference of the beads in the spectrophotometer.
Recombinant Cex Methylumbelliferyl-cellobioside assay
Methylumbelliferyl-cellobioside (MUC) is a common assay for the activity of exoglucanases. Once digested by the exoglucanase, MUC is converted into cellobiose and 4-methylumbelliferone (4-MU), a fluorogenic compound which has excitation and emission wavelengths at 360/460nm, however the absorbance of which can also be read at 348nm for the monitoring of MU release related to enzyme [2]. The results are displayed in Figure 1. These are interesting preliminary results, but more repeats would allow for more decisive conclusions to be made. These can be viewed on the Proof of concept page.
Figure 1 Enzyme Activity Before and After Immobilisation to MUC. Enzyme activity was determined by calculating ε of 4-MU = 9954M-1cm-1. 1 U is equal to 1μmol of substrate released per minute. sfGFP-L2NC is a control which can bind Silica beads, but does not impact the screening of the substrate. Cex-L2NC showed increased activity after immobilisation, whereas untagged Cex and Cex-Car9 showed no change. GFP-L2NC results from Wash 3 step cannot be explained.
Nonetheless, while Cex-Car9 appears to be non-functional, as its absorbance is lower than the controls (untagged Cex), the data appear to suggest that the Cex-L2NC can interact with the MUC substrate to produce fluorogenic 4-MU. Curiously, untagged Cex does not appear to have any activity on the substrate. This could be due to protein aggregation after induction with IPTG, thereby nullifying the activity of the enzyme. SDS-PAGE was performed on all samples after 18oC overnight induction, however the results are unclear and inconclusive.We investigated the reasons behind this further using structural modelling.
PETase and LCC pNPB assays
Para-Nitrophenol butyrate (pNPB) is a model substrate which is hydrolysed by PETase and LCC enzymes into para-Nitrophenol (having a characteristic yellow colour) (Figure 2) with maximum absorbance at 415 nm. This was considered as a preliminary assay to check the presence of active enzymes expressed (3). The calibration curve was then constructed, and the concentration of unknown samples was calculated using the trendline formula.
Figure 2 Results of para-Nitrophenol Butyrate assay. (A) 1-Without level 1 plasmid as a control, 2- PETase enzyme, 3-PETase tagged with Car9, 4- PETase with L2NC tag, and 5-PETase with SB7 tag. (B) 1-Tube without any level 1 plasmid as a control, 2- LCC enzyme, 3-LCC tagged with Car9, 4- LCC with L2NC tag, and 5-LCC with SB7 tag.
Figure 3 para-Nitrophenol (mM) produced by different proteins for the enzymatic activity analysis.
This experiment demonstrated all the protein constructs to be active. PETase tagged with L2NC was shown to be the most active enzyme, with around four-fold higher enzymatic activity than untagged PETase. Additionally, this test presented a significant increased activity of the PETase enzyme when tags are used. Si-tags might be increasing the enzyme’s affinity for the substrate, resulting in increased activity [4]. This was also true for the LCC enzyme, where all of the tags increased its activity. This can be argued that the tags are contributing to the LCC enzyme’s increased thermal stability (at 72°C), as suggested by [5]. Interestingly, the tube containing cellular proteins without a level 1 plasmid and therefore without cutinases produced a positive pNPB result at 30°C, suggesting the presence of proteins in E. coli BL21 (DE3) cells with cutinase or lipase activity (both of which produce a positive pNPB result).This was significantly enhanced in the presence of a higher temperature (72°C), suggesting that some enzymes are temperature-dependent [6].
The major drawback of this assay is the low structural similarity between the pNPB and PET polymers [3], making it impossible to identify the actual activity on PET, which remains a comparative speculation.
Assay with PET film
An absorbance value for all the major degradation products of PET (MHET and TPA), was selected measured with a spectrophotometer at 260 nm absorbance [7]. Unfortunately, since MHET was not commercially available, BHET was used to calculate the calibration curve for its detection,. It is important to note that this method harbours a heavy drawback. Infact, the activity detected might be due to the hydrolysis of molecules on the PET surface rather than degradation of the PET core [3]. For this reason, electron microscopy scanning should be employed in the future for observing the mass loss of the PET film after hydrolysis.
Based on the absorbance profiles of MHET and TPA, they have distinct extinction coefficients at 260 nm of 5500 M−1 cm−1 and 4200 M−1 cm−1, respectively. Concentrations were calculated using the Beer-Lambert Law (Figure 4) [7].
Figure 4 MHET and TPA (mM) produced by proteins after the degradation of PET film.
According to Figure 4, PETase tagged with L2NC, produced about three times more monomers (MHET and TPA) than PETase alone, indicating that L2NC increased the activity of the PETase enzyme. For the LCC enzyme, all the tags enhanced its activity, although the increase is not statistically significant. These results are debatable since the literature indicates that increasing hydrophobicity of proteins accounts for an increased affinity of enzymes for their substrate [8]. And modelling predicted that, PETase without a tag was the most hydrophobic, while PETase with L2NC was the least one. However, the findings of this assay were the opposite. On the other side, the activity of the LCC enzyme was enhanced in the presence of an SB7 tag that was more hydrophobic than the enzyme alone. However, the findings are not valid for Car9 and L2NC tags. Additionally, as shown by the control reaction (untransformed), there was also an indication of the presence of MHET and TPA produced. Thus, this assay may not be sensitive for detecting these two monomers, as many proteins also absorb at 260 nm and may provide false-positive results after the degradation of PET [7].
1. Taniguchi K, Nomura K, Hata Y, Nishimura T, Asami Y & Kuroda A (2007) The Si-tag for immobilizing proteins on a silica surface. Biotechnol. Bioeng. 96: 1023–1029
2. Duedu KO, French CE. Characterization of a Cellulomonas fimi exoglucanase/xylanase-endoglucanase gene fusion which improves microbial degradation of cellulosic biomass. 2016 [cited 2021 Aug 13]; Available from: https://doi.org/10.1016/j.enzmictec.2016.08.005
3. Pirillo, V., Pollegioni, L. and Molla, G. (2021) “Analytical methods for the investigation of enzyme-catalyzed degradation of polyethylene terephthalate,” The FEBS Journal. doi: 10.1111/FEBS.15850
4. Costa, S. et al. (2014) “Fusion tags for protein solubility, purification and immunogenicity in Escherichia coli: the novel Fh8 system,” Frontiers in Microbiology, 0(FEB), p. 63. doi: 10.3389/FMICB.2014.00063.
5. A, Su et al. (2018) “Immobilized cutinases: Preparation, solvent tolerance and thermal stability,” Enzyme and microbial technology, 116, pp. 33–40. Doi: 10.1016/J.ENZMICTEC.2018.05.010.
6. Peterson, M. E. et al. (2007) “The dependence of enzyme activity on temperature: determination and validation of parameters,” Biochemical Journal, 402(Pt 2), p. 331. doi: 10.1042/BJ20061143
7. Zhong-Johnson, E. Z. L., Voigt, C. A. and Sinskey, A. J. (2021a) “An absorbance method for analysis of enzymatic degradation kinetics of poly(ethylene terephthalate) films,” Scientific Reports 2021 11:1, 11(1), pp. 1–9. doi: 10.1038/s41598-020-79031-5.
8. Sarkar, A. and Kellogg, G. E. (2010) “Hydrophobicity – Shake Flasks, Protein Folding and Drug Discovery,” Current topics in medicinal chemistry, 10(1), p. 67. Available at: /pmc/articles/PMC2851852/
AlphaFold2 prediction of recombinant Cex constructs.
Possible reasons for Cex-Car9’s lack of activity were assessed by using the recent AlphaFold2 package. Images of the models can be viewed in Figure 1.
Figure 1 3D Structures of Cex-Car9 and Cex-L2NC Structures generated with the AlphaFold2 package and visualised on UCSF ChimeraX. A: active site of Cex. B: Car9 C terminal tag. C: L2NC C-terminal tag.
It is important to note that both tags have a low local distance difference test (lDDT), indicating a low average accuracy of the residue positions. However, if the models in Figure 2 are to be considered, the positively charged Car9 tags are extremely close to the entrance of the negatively charged active site. On the other hand, the much larger L2NC tag coils around the enzyme and away from the active site, possibly explaining its increased activity. The Car9 tag could be impeding the binding of the substrate to the active site, thereby impacting the ability of Cex to catalyse the reaction.
Further discussion
Another possible explanation for the increased activity of Cex-L2NC is the improved solubility of Cex-L2NC in comparison to Cex-Car9 and untagged Cex. An initial solubility experiment with induction at 37oC was run to assess the solubility of the recombinant protein in E. coli BL21 (DE3). The results of the resulting SDS-PAGE are shown in Figure 2.
Figure 2 Solubility test of Cex-Car and Cex-L2NC after incubation overnight at 37oC Expected weight of Cex-Car9: 49.1 kDa. Expected weight of Cex-L2NC: 61.75 kDa. W: whole cell, S: soluble fraction, I: insoluble fraction.
The acrylamide gel shows that the majority of the recombinant protein is insoluble when induced at 37oC and therefore most likely aggregated and unfunctional. Several repeats of the solubility experiment were attempted with the 18oC however the results were consistently inconclusive. The results of the SDS-PAGE appear to show a greater proportion of the recombinant protein in the soluble fraction for Cex-L2NC than Cex-Car9. It is possible that the highly positively charged residues that comprise the L2NC tag help to act as a solubilising agent, thereby increasing the recombinant enzyme’s activity. Future experiments should continue with L2NC due to its increased solubility, activity and lower dissociation constant [1]
Hydrophobicity of recombinant PETases and LCCs
For LCC and PETase enzymes, the hydrophobicity of the proteins was calculated with the Kyte and Doolittle scale model which determines it according to the hydropathy index of its amino acids, representing the hydrophobicity of their side chain. According to the graph (Figure 3), L2NC notably increased the solubility of both LCC and PETase, similarly to what previously observed when fused to Cex enzyme.
Figure 3 Hydrophobicity of the proteins as measured by the Kyte and Doolittle scale. More negative scores indicate higher hydrophobicity.
References:
[1] Taniguchi K, Nomura K, Hata Y, Nishimura T, Asami Y, Kuroda A. The Si-tag for immobilizing proteins on a silica surface. Biotechnol Bioeng [Internet]. 2007 Apr 15 [cited 2021 Apr 23];96(6):1023–9. Available from: http://doi.wiley.com/10.1002/bit.21208