Team:St Andrews/Results



Antibody Expression Results

Table 1. Abbreviations and classifications for each antibody fragment studied during our project

Antibody 1 (AB1) = 4B53, Human IgG CH3 domain edit 4

Antibody 2 (AB2) = 4B53, Human IgG CH3 domain edit 5

Antibody 3 (AB3) = 1CQK, Mouse IgG CH3 domain edit 4

Antibody 4 (AB4) = 1CQK, Mouse IgG CH3 domain edit 5

Antibody 5 (AB5) = Human IgE CH3 domain edit 4

Antibody 6 (AB6) = Human IgE CH3 domain edit 5

Antibody 7 (AB7) = 4B53, Human IgG CH3 domain wild type

Antibody 8 (AB8) = 4B53, Human IgG CH3 domain edit 1

Antibody 9 (AB9) = 4B53, Human IgG CH3 domain edit 2

Antibody 10 (AB10) = 1CQK, Mouse IgG CH3 domain edit wild type

Antibody 11 (AB11) = 1CQK, Mouse IgG CH3 domain edit 1

Antibody 12 (AB12) = 1CQK, Mouse IgG CH3 domain edit 2

Antibody 13 (AB13) = Human IgE CH3 domain wild type

Antibody 14 (AB14) = Human IgE CH3 domain edit 1

Antibody 15 (AB15) = Human IgE CH3 domain edit 2


Cloning Antibody Fragments into pEHISTEV

The antibody wild-type CH3 domains and the four mutant edits for each antibody-fragment were successfully cloned into the pEHISTEV plasmid. The pEHISTEV plasmid is designed to express proteins with a His-tag that is removable by the TEV protease. The expression is under a T7 promoter and control by the lac repressor. The plasmids were first transformed into SoluBL21 E. coli cells and the success of the cloning was determined by digestion with restriction enzymes and colony-PCR.

The re-digestion of AB3, AB4, AB5 and AB6 by BamHI and NcoI restriction enzymes that were used for the initial double-digest showed the correctly sized fragments at ~350 base pairs and 6.5 kilobases.

Figure 1. Digestion of the antibody fragments cloned into pEHISTEV using the restriction enzymes that were used of the initial double digest. All four cloned fragments showed correctly sized inserts, so the cloning was successful.

The cloning was confirmed by colony-PCR for AB3, AB4, AB5 and AB6 as well, where primers on the two ends of our inserts were used to PCR-up ~350 base-pair long fragments.

Figure 2. Colony PCR for the four cloned antibody fragments, showing correct colonies for AB3 in lane 2, for AB5 in lane 6 and for AB6 in lane 8. AB4 has no correct colonies based on this colony PCR.

The colonies used for inoculation for 3, 5, and 6 were correct for these fragments based on colony-PCR, but AB4 showed no amplified product.

Figure 3. AB4 colony PCR showing multiple amplified products, including ones around 1500 base pairs, while the expected product is around 350 base pairs. It is possible that none of the screened colonies contain correct ligations but nonspecific primer binding can also be an issue.

More colonies were screened, buts Since no colonies showed only one amplified product of the correct size, the extracted plasmids were digested by an restriction enzyme (AflII) which should produce three fragments (4 kilobases, 1.5 kilobases and 470 bases) for our plasmids to see if the extra amplified fragment of 1.5 kilobases is made due to improper ligation or non-specific primer-binding.

Figure 4. AFlII digestion for AB4 cloned into pEHISTEV. The digestion produced three fragments of the correct sizes (4000, 1500 and 470 base pairs, so the ligation was correct.

The digestion showed correctly sized inserts, so these plasmids were used for future transformations.

The cloning success was determined by colony PCR for all the other antibody fragments, where similar primers annealing to the end of our inserts were used to asses the presence of a correct insert.


Figure 5. Colony PCRs for AB7, AB10, AB1, AB2, AB8, AB9, AB11, AB12, AB13, AB14 and AB15. All antibody fragments have at least one colony with the correct cloning, based on the correctly amplified ~350 bas pairs long fragments, these will be used for future inoculation steps.


All the antibody fragments were cloned correctly in pEHISTEV so expression was carried out using SoluBL21 cells that express sequences under the T7 promoter.

The cloning success was determined by sequencing for AB7 as well, where the T7 promoter was used as a sequencing primer to examine the sequences outside the coding sequence as well. The sequencing results showed no point mutations and the AB7 plasmid showed a 100% correct sequence.


Expressing Human IgG Wild Type (AB7)

After initial expression tests were unsuccessful due to probable problems with cell lines or IPTG, we focused our attention on expressing the Human IgG wild-type (AB7) and its mutants.

The ideal expression conditions were determined for the wild-type, using induction with IPTG at different time-intervals, temperatures and IPTG concentrations.

Figure 6. Expression tests showing for the wild type (iGEM7) under different conditions. The protein produced the strongest band, hence expressed the best at 25°C overnight expression with 0.2 mM IPTG used for induction.

The expression test showed the highest amount of protein for AB7 at 25°C with 0.2 mM IPTG concentration at overnight so the one-litre protein-expressions were carried out using these conditions.

Figure 7. Purification of the human IgG wild type using nickel affinity chromatography and TEV cleavage. The expressed protein is highly soluble and was purified well via this method. The reverse nickel purified fraction only contains one band at the expected 12 kDa mass, so pure recombinant protein was obtained.

The wild type (AB7) was expressed in a high amount in the one litre culture and most of the protein is in the soluble fraction, with the protein being only partially insoluble. The elution fraction contained a high amount of recombinant protein at 15 kDa, and the cleavage was successful, although partial, as the 3 kDa size decrease is visible in the post-cleavage fraction. The reverse nickel purified fraction contains pure, recombinant protein at 12 kDa. The identity was confirmed by electrospray mass-spectrometry with an expected mass of 12373.8 Da.


Figure 8. Mass-spectrum for the expressed and purified wild type human IgG fragment. The spectrum shows a very strong and clear signal at the expected molecular mass (12373 Da) which correspond to the recombinant protein.


The correct recombinant protein was expressed based on the SDS-Page gel and the mass-spectrometry results.


Expressing Mutant Antibody Fragments

Figure 9. Whole cell lysates showing clear bands at around 15 kDa for all mutant antibody fragment expressions. All four mutants expressed, but the solubility cannot be assessed based on this gel alone.

four IgG mutants (AB1, AB2, AB8 and AB9) expressed well and produced a double band at around 15 kDa. Edit 2 shows the strongest band, and hence probably the most protein. This culture was induced at a lower optical density (0.22 instead of >0.6) since the cells were not growing over a long period of time.


Expressing Human IgG Edit 1 (AB8)

The mutant protein was induced under the same condition and the purified protein sample was analysed via mass-spectroscopy with an expected mass of 12303.64 Da.

Figure 10. Mass-spectrometry analysis of purified edit 1 sample. The spectrum showed no signal around the expected 12 kDa molar mass, only a contamination at 43 kDa that cannot be related to a dimeric state of the recombinant fragment based on molar mass.

Mass-spectrometry analysis only showed a contamination and no mutant protein present, even though the gel showed a correctly sized band.


Expressing Human IgG Edit 2 (AB9)

AB9 was induced under the same condition and purified using the same procedures.

Figure 11. Reverse nickel affinity purification for edit 2 recombinant protein expression. The elution fraction shows a band at around the expected 15 kDa molar mass, while the concentrated reverse nickel purified fraction shows an expected band at 12 kDa. Edit 2 expressed and was successfully purified by this method, although some contaminations are still present in the final fraction, around 40 kDa and 70 kDa.

Edit 2 was mostly insoluble and only a fraction of the expressed protein was in the soluble fraction, while the elution contained less protein than the wild type as well. The TEV cleavage is still visible, with a 3 kDa decrease in size from 15 kDa to 12 kDa between the elution fraction and the concentrated reverse Ni-purified sample. An impurity around 60 kDa is also present in the purified sample.

The mutant expressed in a smaller quantity but it was still analysed via mass-spectrometry, with an expected mass of 12261.56 Da.

Figure 12. Mass-spectra for edit 2 showing a clear signal at the expected molecular mass at 12258.8. corresponding to a fragment with a disulphide bond. No secondary signal is present that would indicate isopeptide bond formation.

Mass-spectrometry confirmed the expression but the contamination is not visible so the sample was submitted to mass-spectrometry again the next day. The mass-spectrometry analysis required more sample than expected based on the concentration calculated from the absorption.

Figure 13. Second submission of the same sample for edit 2, on this spectrum the original signal corresponding to the recombinant protein is not visible. New signal for contamination is picked up around 68 kDa, but this cannot be corresponded to any multimeric state of the recombinant protein.

The contamination is visible on the spectrum but the mutant protein is no longer present in the same sample.


Expressing Human IgG Edit 4 (AB1)

AB1 was induced under the same condition and purified using the same procedures.

Figure 14. Reverse nickel purification of edit 4, including the nickel affinity purified elution fraction and the concentrated solutions after TEV protease cleavage. The gel shows no band in the 12-15 kDa range, so edit 4 might not have expressed at all.

The gel showed no clear band at the right size in any of the purified fraction, but the insoluble fraction contains a strong band of the right size. Edit 4 might be entirely insoluble so no mass-spectrometry sample was run.


Expressing Human IgG Edit 5 (AB2) – Cysteine-free mutant

Figure 15.Nickel affinity chromatography purification fractions for edit 5 expression. A clear and strong band is present in the soluble fraction at around 15 kDa, which is also present in a smaller quantity in the elution fraction. Edit 5 expressed well and Is soluble, although less so than the wild type.

The cysteine-free mutant was expressed under the same conditions, and the band corresponding to the mutant protein at 15 kDa is strong in the soluble fraction and present in the elution too. AB2 is soluble but it is expressed in a smaller quantity than the wild type.

Figure 16. Reverse nickel purification for the expressed edit 5 solutions. No strong band associated with recombinant protein is visible in either of the purified or elution fractions, even though the elution showed clear bands previously. The 10x concentrated final sample shows a band at around 12 kDa where the recombinant protein is expected.

Edit 5 was only present in a very small quantity after TEV cleavage and reverse nickel chromatography, but the concentrated elution fraction still showed a correctly sized band at 12 kDa so a mass-spectrometry sample was run with an expected mass of 12229.52 Da.

Figure 17. Mass-spectrometry result for edit 5. No signal is present around 12 kDa where our protein is expected, only one clear signal at 43320 kDa is present, which is most likely an E. coli protein.

The spectrum shows no recombinant protein present only a contamination at 43 kDa. Even though the gel clearly showed a band, no protein is picked-up by mass-spectrometry.


Expressing the Human IgG Wild-Type and its Mutants

Figure 18. Expressed and purified wild type and mutant human IgG CH3 domains. The wild type is present in a high quantity, containing a small amount of un-cleaved protein, while only edit 1 (AB8) and edit 2 (AB9) show very weak bands at the correct size. Edit 4 (AB1) and edit 5 (AB2) show no band corresponding to recombinant protein. The gel also shows no difference for AB7 between a sample buffer with or without reducing agent.

A gel was run for all the expressed protein solutions. The wild type (AB7) produced a strong band corresponding the recombinant protein before TEV cleavage (~15 kDa) and after cleavage (~12 kDa). The mutant proteins were expressed in a significantly smaller quantity and this final gel only shows very weak bands for AB8 and AB9 while the AB1 and AB2 fractions contain no band at the correct size.

The mutant proteins expressed less than the wild type under the same conditions and were also less soluble.

Table 2. Concentrations for each expressed antibody fragment for the initial elution step after nickel affinity chromatography and for each concentrated reverse-nickel purified protein solutions. The wild type expressed in a much greater quantity while edit 2, edit 4 and edit 5 produced a similar amount of protein after nickel purification. Protein concentration is slightly higher for edit 1, but still lower than the wild type.


Wild type (AB7)

Edit 1 (AB8)

Edit 2 (AB9)

Edit 4 (AB1)

Edit 5 (AB2)

Elution (20-25 mL)

5.3 mg/mL

0.70 mg/mL

0.27 mg/mL

0.21 mg/mL

0.23 mg/mL

2.5 mL concentrated


0.21 mg/mL

0.30 mg/mL

0.23 mg/mL

0.15 mg/mL

300 µL concentrated


0.17 mg/mL

0.37 mg/mL


0.21 mg/mL

150 µL concentrated





0.32 mg/mL


The mass-spectrometry analysis proved to be inconclusive for the mutants as only one mutant protein (AB9) was visible on the mass-spectra and only for during the first analysis. The AB8 and AB2 sample, as well as the second AB9 sample, contained correct bands on the SDS-Page gels that were not visible on the mass-spectra.

The cysteine-free edit (AB2) seemed highly soluble based on the first gel run after nickel chromatography but subsequent gels showed less and less protein after each purification or concentration steps as this protein precipitated from the solution.

No data supports the presence of an isopeptide bond in any of the human IgG CH3 domain edits based on SDS-Page gels and mass-spectrometry analysis.


Expressing Mouse IgG CH3 Domain Wild-Types and Mutants in 10 mL Cultures

The mouse IgG CH3 domain wild-type and edit 2 showed no soluble protein expression based on small scale (10 mL) E.coli cultures with 0.2 mM IPTG overnight expression at 25°C.

The wild type expressed well, but the protein is insoluble based on small scale SDS-page gels (Figure 19.)

Figure 19. Mouse IgG wild-type expression duplicates showing insoluble recombinant protein at 15 kDa and no soluble expression based on the soluble fraction.

Edit 2 is not visible on the gel even in the insoluble fraction, so if this protein expresses, it only does so in a very low quantity.

Figure 20. Small-scale mouse IgG CH3 domain edit 2 expression, showing no clear band corresponding to a recombinant protein.

Expressing Human IgE CH3 Domain Wild-Types and Mutants in 10 mL Cultures

Human IgE CH3 domain wild-type expression showed strong bands corresponding to insoluble protein expression, while no soluble recombinant protein is observable.

Figure 21. Human IgE wild-type expression, showing clear bands corresponding to insoluble recombinant protein for the induced fraction. Soluble recombinant protein expression is not visible on this gel.

None of the IgE CH3 domain mutants showed clear soluble protein expression based on the small-scale expression. All of the mutants expressed well insolubly, as strong bands are visible in the insoluble factions at 15 kDa corresponding to recombinant protein. As these proteins are insoluble no isopeptide bond formation is predicted.

Figure 22. Human IgE CH3 domain mutant expressions. All mutant expressed well, but no soluble protein is visible on these gels. The strong band at 15 kDa corresponds to the recombinantly expressed protein in the insoluble fraction.

Thermatoga Results



To test the idea that substitution of an inert amino acid into a double salt bridge could induce isopeptide bond formation, we identified a protein with a double salt bridge motif, the beta-glucosidase enzyme from the thermophile Thermatoga maritima (figure 1), which we referred to as TMBG.

Figure 1:Triple salt bridge motif in the beta-glucosidase from Thermatoga maritima (PDB ID: 2J77). Hydrogen bonds, forming salt bridges between residues, are shown as yellow dashed lines. Figure made using PyMol.

To induce and isopeptide bond, we created the substitutions R371M and R413M, with the intention that Methionine would occupy a similar space in the protein but not form Hydrogen bonds with any residues, preventing isopeptide bond formation. We also introduced the mutant R371M, R413Y to shield the bond from solvent with a Tyrosine.


As we supposed the correctly folded protein was soluble, we surmised that checking the mutant’s solubility would give a quick indicator of their ability to form isopeptide bonds. As a positive control, we expressed the wild type protein to check its solubility in our system.

Samples were prepared as outline in Methods. The wild-type protein was given to us in a pET28a vector, mutant sequences were ordered in from IDT then cloned into pEHisTEV vectors (see Methods). Expression was induced using 1mM iPTG.

Figure 2. shows soluble expression for the wild-type protein, the pET28a vector was known to have a leaky promoter so the bands at ~55kDa (expected mass 52kDa) indicate expression, as they are absent in lanes used for mOrange and Antibody expression (lanes 2-10).

Figures 3 shows expression tests for the TMBG mutants. There is no visible expression for TMBG R371M R413M (Edit 1) with no bands around the 55kDa mark. In both, expression for TMBG R371M R413Y (Edit 2) appears to be dominantly insoluble, with no discernible difference at 55kDa between the soluble bands but a large 55kDa band in the insoluble fraction.

Figure 3: Expression tests for TMBG Edit 1 (Lanes 2-5) and TMBG Edit 2 (Lanes 6-9). Lane 1 = Ladder, Lane 2 = Edit 1 uninduced soluble fraction, Lane 3 = Edit 1 uninduced insoluble fraction, Lane 4 = Edit 1 induced soluble fraction, Lane 5 = Edit 1 induced insoluble fraction, Lane 6 = Edit 2 uninduced soluble fraction, Lane 7 = Edit 2 uninduced insoluble fraction, Lane 8 = Edit 2 induced soluble fraction, Lane 9 = Edit 2 induced insoluble fraction. Expression of TMBG Edit 2 is highlighted in the red circle.


Thus, as the wild-type TMBG is soluble but TMBG R371M R413Y is dominantly insoluble, we predict the mutant to not be capable of forming isopeptide bonds. Absence of expression for TMBG R371M R413Y means we can draw no conclusions as to its solubility.

School of Biology

School of Chemistry

School of Mathematics

School of CS

School of Physics

School of Philosophy

Sir Kenneth Murray Endowment Fund

iGEM St Andrews 2019