6.2 B IACORE - MEASUREMENTS
6.2.1 Improving the regeneration
To ensure that no Juno was left in between analyte injections and to avoid problems with baseline decay, a multi-cycle approach was chosen. The plan was to regenerate the anti-FLAG surface after each cycle, i.e. remove both Juno and Izumo1, hence an effective regeneration procedure had to be found. Several regeneration cycles without injecting Juno was performed. Firstly, 10 mM HCl was used, but the baseline at the end and start of each cycle was not the same, suggesting that the regeneration method was not optimal, see figure 18.
34
Figure 18: Regeneration test with HCl. The RU-increase at 250 s is injection of Izumo1, the dip at 1550 s is injection of HCl.
A new regeneration test was made using 20 mM NaOH followed by 20 mM HCl. After several multi-cycle procedures it was observed that less Izumo1 could be captured for each multi-cycle, indicating that the anti-FLAG antibodies had been damaged from the regeneration. Suspecting that the NaOH had caused the damage, a new regeneration method was tested using 0.1% SDS and 20 mM HCl. However, the RU increase of Izumo1 decreased drastically, and after 4 cycles Izumo1 could not be captured (figure 19).
It seems that the SDS damaged the antibodies, since HCl had been used earlier with less drastic results.
Figure 19: Sensorgram of regeneration test with SDS and HCl. The decrease for each cycle suggest damage to the anti-FLAGs.
This sensorgram is zoomed in for better visualisation (the regeneration pulses are not visible here, see appendix).
35 New anti-FLAG antibodies were applied to flow cell 3 and 4 of the chip (with 3 as reference), since flow cell 2 now was destroyed. Glycine 10 mM pH 2.5 was tried as the next regeneration media, but still less Izumo1 could be added for each cycle. Next approach was to use only alkaline regeneration, hence 20 mM NaOH was tried. Finally, it seemed that a good regeneration method was found; the baseline was similar at start and end and the capture of Izumo1 was similar for each cycle, see figure 20.
Figure 20: Sensorgram of regeneration test with NaOH.
From the optimisation of regeneration methods, it could be concluded that the anti-FLAG antibodies are rather sensitive, but that the regeneration with 20 mM NaOH worked well.
6.2.2 Measurements with recombinant human Juno produced in E. coli
As can be seen in figure 21, three cycles were performed with 0, 0.8 and 1.6 µM Juno respectively. Each cycle started with capture of Izumo1 followed by baseline stabilisation, injection of Juno and regeneration. As this was before the optimal regeneration had been found, the regeneration was made with NaOH and HCl and as the sensogram shows; less Izumo1 can be captured for each cycle.
Injection of Juno generated signals that indicate binding and are proportional to the Juno concentrations (after subtraction of the reference cell). Further runs could not be made with this Juno due to limited amount of protein, and therefore it was not possible to calculate a value of the affinity. However, the measurements clearly show that the proteins do bind which is a nice result since there are no previously published studies using E. coli-expressed Juno in combination with Izumo1 produced in mammalian
36 cells. The squared shape of the curve indicates that the proteins dissociate fast; the relative response stop increasing shortly after the start of injection and stabilises since association and dissociation occur at the same time, and at the end of the injection the steep decrease is a result of an instant dissociation. Based on the shape of the sensorgram, a hypothesis was made that Kd probably lies within the micromolar-range. This is widely different from a study in 2016 where both recombinant human proteins were expressed in insect cells and a Kd of 48 nM was found7.
Figure 21: Sensorgram of Juno-Izumo1 binding.
6.2.3 Measurements with recombinant human Juno produced in mammalian cells
6.2.3.1 Single-cycle approachEven though optimisation of the regeneration had been investigated the regeneration was not perfect, and it was therefore decided to try a single-cycle approach. Since the Juno that was used in the previous runs indicated a weak binding in the micro-molar range, the first single-cycle run was made with rather high Juno-concentrations of 0.05 – 30 µM. Interactions between the two proteins was measured, but surprisingly the curves had a widely different shape, indicating a slower dissociation than with the previously used Juno. Furthermore, the relative response of the peaks only differed for the lower concentrations and remained constant for the higher concentrations, see figure 22. Since these curves indicated a stronger binding, it was suspected that this run was using concentrations way above the dissociation constant Kd, which would explain why the relative response was unchanged.
37
Figure 22: Sensorgram of single-cycle run with Juno at 0.05, 0.1, 0.2, 0.4, 0.8, 1.7, 3.3, 7.5 and 15 µM. The relative response of the peaks only changes slightly, and there is a constant drift of the baseline.
A new single-cycle run was performed using concentrations of 2 – 1000 nM, see figure 23. Here there was a clear increase in relative response as the concentration of Juno increased, indicating that we were now covering the right concentration span to find Kd. Since it is easier to do kinetics calculations from multi-cycle data in the software, the same protein concentrations were used in two multi-cycle runs.
Figure 23: Sensorgram of a single-cycle run with Juno-concentrations of 2, 4, 8, 16, 32, 63, 125, 250, 500 and 1000 nM.
1840 1890 1940 1990 2040
0 1000 2000 3000 4000 5000 6000 7000 8000
Tim e s
0 2000 4000 6000 8000 10000
Tim e s
Resp. Diff.
RU
38 6.2.3.2 Multi-cycle 1
Twelve cycles were performed (fig. 24 and 25) as previously described in the method. During the first run, it was clear that the regeneration of the sensor surface was a problem since the amount of Izumo1 that could be captured at the start of each cycle was not the same, see figure 24. The first 9 cycles captured Izumo1 with varying relative responses between 250-300 RU and for the last three cycles the capture was 200, 140 and 100 RU respectively. It seems that the regeneration of 20 mM NaOH was less effective towards the higher concentrations of Juno, maybe because these cycles would have needed more time for dissociation. This was an issue since the calculations used for calculating Kd are based on that the concentration of one of the proteins is kept constant while the other one is varied. Moreover, it is difficult to compensate for the varying Izumo1 capture in the calculations since it is unknown how this affects the binding assay. The ineffective regeneration could lead to that active Izumo1 is left on the surface so that more protein will be able to bind to Juno during the next cycle, or it could mean that inactive or still bound Izumo1 is left on the surface which would lead to that less active Izumo1 can be captured in the next cycle. A new multi-cycle run was made where a harsher regeneration method was applied in order to solve this issue.
Figure 24: Sensorgram showing multi-cycle run 1; the arrows indicate the amount of Izumo1 captured for the different cycles.
6.2.3.3 Multi-cycle 2
During the second multi-cycle run the regeneration was better and the capture of Izumo1 more even, but there is still a gap of ~50 RU in between cycles. The curves for the higher Juno pulses are closer together (see figure 26) which indicates that the maximum response where all binding sites are occupied is near, something that was not seen in the first multi-cycle run.
39
Figure 25: Sensorgram of multi-cycle 1 showing the association and dissociation of Juno and Izumo1.
Figure 26: Sensorgram of multi-cycle 2 showing the association and dissociation of Juno and Izumo1. Cycle 3 containing Juno 4 nM is omitted since there were problems with air bubbles during the injection of Izumo1.
-40 10 60 110 160
1000 1100 1200 1300 1400 1500 1600 1700
Tim e s
Resp. Diff.
RU
Juno 0 nM Juno 2 nM Juno 4 nM Juno 8 nM Juno 16 nM Juno 32 nM Juno 63 nM Juno 125 nM Juno 250 nM Juno 500 nM Juno 1000 nM Juno 2000 nM
-40 10 60 110 160
1000 1100 1200 1300 1400 1500 1600 1700 1800
Tim e s
Resp. Diff.
RU
Juno 0 nM Juno 2 nM Juno 8 nM Juno 16 nM Juno 32 nM Juno 63 nM Juno 125 nM Juno 250 nM Juno 500 nM Juno 1000 nM Juno 2000 nM
40 6.2.3.4 Calculations of the equilibrium dissociation constant
The sensorgrams where analysed using two approaches and the results are presented in table 10.
Table 10: Summary of calculated values of the dissociation constant.
Run Single-cycle Multi-cycle 1 Multi-cycle 2
Kd [nM] (saturation model) 33 17 12*
33 33*
Kd [nM] (Langmuir model) - 7.1 7.2
*the relative response at equilibrium (RUeq) was multiplied with the relative response of captured Izumo1 divided by the relative response of captured Izumo1 of the first cycle with 0 nM Juno; RUeq x RUIzumo1/RUIzumo1(cycle 1).
6.2.3.4.1 Equilibrium analysis
The data points from the single-cycle (figure 27) and the second multi-cycle (figure 28 and 29) run could be fitted well to a saturation binding model and both led to a dissociation constant of 33 nM. To calculate the same Kd from two separate measurements is very promising and the good fit of the data points and the fitting curve indicates a small error. Furthermore, the value is not that far from the previously published Kd of 48 nM that was measured of recombinant human Juno and Izumo1 expressed in insect cells7.
Figure 27: Equilibrium analysis of the single-cycle run.
0
0 200 400 600 800 1000 1200
RU
41
Figure 28: Equilibrium analysis of the second multi-cycle run.
Figure 29: Equilibrium analysis of the second multi-cycle run. *the relative response at equilibrium (RUeq) was multiplied with the relative response of captured Izumo1 divided by the relative response of captured Izumo1 of the first cycle with 0 nM Juno; RUeq x RUIzumo1/RUIzumo1(cycle 1).
The first multi-cycle run generated data points that did not fit as well to the binding model, since the relative responses for the higher Juno-concentrations do not converge (figure 30). This was noted already in the sensorgrams in figure 25 and 26 where the higher concentrations indicate saturation in multi-cycle 2 but not in multi-cycle 1. The deviating results in multi-cycle 2 is most likely due to the uneven regeneration. Since the sensor surface was not entirely cleaned, it is possible that bound Izumo1-Juno complex remained on the surface which could explain why the responses for the last cycles are higher than expected. An attempt was made to compensate for the uneven Izumo1-capture during the multi-cycle procedures in figure 29 and 31.
0
0 500 1000 1500 2000 2500
RU
0 500 1000 1500 2000 2500
RU
42
Figure 30: Equilibrium analysis of the first multi-cycle run.
Figure 31: Equilibrium analysis of the first multi-cycle run.*the relative response at equilibrium (RUeq) was multiplied with the relative response of captured Izumo1 divided by the relative response of captured Izumo1 of the first cycle with 0 nM Juno; RUeq x RUIzumo1/RUIzumo1(cycle 1).
0 500 1000 1500 2000 2500
RU
0 500 1000 1500 2000 2500
RU
43 6.2.3.4.2 Kinetic analysis
The dissociation constant calculated using the Langmuir 1:1 binding model was lower and both multi-cycles gave a similar value (fig. 32 & 33). The calculations led to a Kd of 7.1 and 7.2 nM respectively which indicates an even stronger binding between Juno and Izumo1.
Figure 32: Langmuir binding analysis of multi-cycle 1 where the black curves represent the fitted curves while the colorfull ones are primary data.
Figure 33: Langmuir binding analysis of multi-cycle 2 where the black curves represent the fitted curves while the colorfull ones are primary data.
800 1000 1200 1400 1600 1800 2000
Tim e s
800 1000 1200 1400 1600 1800 2000
Tim e s
44
6.2.4 Control protein
To ensure that the binding of Juno to immobilised Izumo1 was specific, BSA was used as a control.
Injection of two high concentrations of BSA did not show any indication of binding, see figure 34.
The decrease upon injection of BSA is probably due to a bulk effect that is present in both flow cells to different extents.
Figure 34: Sensorgram of control with BSA.
45
7 To Conclude
Looking back at the objectives at the start of the project, one can say that they were partly fulfilled. An immobilisation procedure with gold nanoparticles was systematically developed and improvements were made along the way. Lessons learned here were that it is vital to find working pH, ionic strength and concentrations to prevent agglomeration of the nanoparticles in all three steps. Parameters such as incubation time and centrifugation speed affected the product yield. It was also discovered that cobalt forms insoluble salts with PBS and that too much shaking or rotation of the incubated samples led to precipitation. In the future, it would be interesting to try the immobilisation with other buffers, perhaps one with low ionic strength. It would also be better to work with bigger volumes, if the resources allow, since the pellet was often so small that it was hard to retrieve.
The interaction between human recombinant Juno and Izumo1 was not measured using the nanoparticle approach, but it was successfully measured with Biacore. The interaction of Izumo1 produced in mammalian cells with Juno produced in bacteria (Junobac) and mammalian cells (Junomam) respectively could be detected. Based on the affinity profiles in the sensorgrams, it seems that Izumo1 binds weaker to Junobac compared to Junomam, which is reasonable since the glycosylations in Junomam should have a closer resemblance to the human wildtype protein.
The dissociation constants calculated using the two binding models differ and it is difficult to say which is the most probable value without doing further measurements. In the future it would be good to do more binding assays with an optimised regeneration method. However, the values are not too far apart, and it can be concluded that the Kd lies in the lower nano-molar range, probably around 7-33 nM. This result indicates a stronger binding than what was found by Aydin et al. in 2016 where a dissociation constant of 48 nM was calculated through SPR-measurements with proteins expressed in insect cells7. Comparing the binding assays made in this project and by Aydin et al. it can be concluded that the affinity of Juno and Izumo1 is dependent on the expression host which indicates that the glycosylation is important for the binding of the two proteins. Since little is known about the mechanism between Juno and Izumo1 today, this is a very interesting observation that takes us one step closer in the course of establishing the binding mechanism between these essential fertilisation proteins.
46
8 Appendix
8.1 Protein sequence Juno
Full length:
GDELLNICMNAKHHKRVPSPEDKLYEECIPWKDNACCTLTTSWEAHLDVSPLYNFSLFHCGL LMPGCRKHFIQAICFYECSPNLGPWIQPVGSLGWEVAPSGQGERVVNVPLCQEDCEEWWEDC RMSYTCKSNWRGGWDWSQGKNRCPKGAQCLPFSHYFPTPADLCEKTWSNSFKASPERRNSG RCLQKWFEPAQGNPNVAVARLFASSAPSWELSYTIMVCSLFLPFLS
8.2 Gels from SDS-PAGE
Gel of Juno (produced in E. coli) before and after buffer exchange to HEPES-buffer as described in the method. The concentrations seem to be similar before and after.
Gel of Juno (produced in mammalian cells) before and after buffer exchange to HEPES-buffer as described in the method. The concentrations seem to be similar before and after but it is a bit difficult to tell due to the broad bands.
Before After kDa
36 28
47 Gel of Izumo1. The concentration was too low to detect strong bands, but we see that there are several bands, which is expected since it is a lysate. It was filtrated prior to SPR-measurements.
kDa
36 28
Before After
48
8.3 More sensorgrams from Biacore
Regeneration 10 mM HCl, Flow cell (Fc) 2-1
Regeneration 0.01% SDS + 20 mM HCl, Fc2-Fc1
-1000
Regenereringstest_0n Fc=2-1 - 1 Regenereringstest_0n Fc=2-1 - 2
-500
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Tim e s
Resp. Diff.
RU
Testkörning_BSA_SA_H Fc=2-1 - 1 Testkörning_BSA_SA_H Fc=2-1 - 2 Testkörning_BSA_SA_H Fc=2-1 - 3 Testkörning_BSA_SA_H Fc=2-1 - 4 Testkörning_BSA_SA_H Fc=2-1 - 5 Testkörning_BSA_SA_H Fc=2-1 - 6
49
-200 -100 0 100 200 300 400 500
0 200 400 600 800 1000 1200 1400 1600
Tim e s
Resp. Diff.
RU
Testkörning_BSA_SA_H Fc=2-1 - 1 Testkörning_BSA_SA_H Fc=2-1 - 2 Testkörning_BSA_SA_H Fc=2-1 - 3 Testkörning_BSA_SA_H Fc=2-1 - 4 Testkörning_BSA_SA_H Fc=2-1 - 5 Testkörning_BSA_SA_H Fc=2-1 - 6
50 Regeneration NaOH 20 mM, Fc2-Fc1
Binding assay Juno-Izumo1 (Juno produced in E. coli) Flow cell 2-1
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Tim e s
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Tim e s
Resp. Diff.
RU
51 Flow cell 1
Flow cell 2
-2500 -2000 -1500 -1000 -500 0 500
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Tim e s
Response
RU
-3500 -3000 -2500 -2000 -1500 -1000 -500 0 500 1000
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Tim e s
Response
RU
52 Control protein BSA, Fc2-Fc1
Immobilisation anti-FLAG Fc2
-50 50 150 250 350 450
0 200 400 600 800 1000 1200 1400 1600
Tim e s
Resp. Diff.
RU
Testkörning_BSA_SA_H Fc=2-1 - 1 Testkörning_BSA_SA_H Fc=2-1 - 2
15000 20000 25000 30000 35000 40000 45000
0 1000 2000 3000 4000 5000 6000
Tim e s
Response
RU
Immobilization Fc2 ant Fc=2 - 1
53
9 References
1. WHO | Infertility is a global public health issue.
WHO (2019).
2. Epstein, R. H. Emotions , Fertility, and the 1940s Woman. J. Public Health Policy 24, 195–211 (2003).
3. Spermosens AB. Male Infertility - Spermosens.
Available at: https://www.spermosens.com/.
(Accessed: 12th August 2019)
4. Spermosens AB. Male Infertility - Spermosens.
5. Patrizio, P., Esposito, M., Kulshrestha, S. &
Khorram, O. Male infertility. Handb. Assist.
Reprod. Lab. 28, 97–138 (2000).
6. Bianchi, E., Doe, B., Goulding, D. & Wright, G. J.
Juno is the egg Izumo receptor and is essential for mammalian fertilization. Nature 508, 483–
487 (2014).
7. Aydin, H., Sultana, A., Li, S., Thavalingam, A. &
Lee, J. E. Molecular architecture of the human sperm IZUMO1 and egg JUNO fertilization complex. Nature 534, 562–565 (2016).
8. Inoue, N., Ikawa, M., Isotani, A. & Okabe, M.
The immunoglobulin superfamily protein Izumo is required for sperm to fuse with eggs. Nature 434, 234–238 (2005).
9. Bianchi, E., Doe, B., Goulding, D. & Wright, G. J.
Juno is the egg Izumo receptor and is essential for mammalian fertilization. Nature 508, 483–
487 (2014).
10. Papale, L., Fiorentino, A., Montag, M. & Tomasi, G. The zygote. Hum. Reprod. 27, 22–49 (2012).
11. Bayer, S., Alper, M. & Penzias, A. The Boston IVF Handbook of Infertility: A Practical Guide for Practitioners Who Care for Infertile Couples (Reproductive Medicine and Asst.
Reproduction). (2011).
12. Bayer, S., Alper, M. & Penzias, A. The Boston IVF Handbook of Infertility: A Practical Guide for Practitioners Who Care for Infertile Couples (Reproductive Medicine and Asst.
Reproduction). (2011).
13. Q-IVF (Nationellt kvalitetsregister för assisterad befruktning). Fertilitetsbehandlingar i Sverige Årsrapport 2019.
14. 1177 Vårdguiden. IVF, provrörsbefruktning - 1177 Vårdguiden. Available at:
https://www.1177.se/barn--gravid/barnloshet/ivf-provrorsbefruktning/.
(Accessed: 23rd October 2019)
15. Sophiahemmet sjukhus. Priser | IVF- och Fertilitetsmottagningen. Available at:
https://ivf-gruppen.se/priser/. (Accessed: 23rd October 2019)
16. Healthcare, G. E. & Sciences, L. Biacore TM Assay Handbook.
17. Cytodiagnistics Inc. Gold Nanoparticle Properties - Cytodiagnostics. Available at:
http://www.cytodiagnostics.com/store/pc/view content.asp?idpage=2. (Accessed: 7th August 2019)
18. Wu, X., Hou, L., Lin, X. & Xie, Z. Application of Novel Nanomaterials for Chemo- and Biosensing of Algal Toxins in Shellfish and Water. Novel Nanomaterials for Biomedical, Environmental and Energy Applications (Elsevier Inc., 2019).
doi:10.1016/b978-0-12-814497-8.00012-6 19. Rycenga, M. et al. Controlling the Synthesis and
Assembly of Silver Nanostructures for Plasmonic Applications. Chem. Rev. 111, 3669–3712 (2011).
20. Cytodiagnostics Inc. Introduction to Gold Nanoparticle Characterization - Cytodiagnostics. surface plasmon resonance biosensing: Current challenges and approaches. Sensors
(Switzerland) 15, 15684–15716 (2015).
22. Unser, S., Holcomb, S., Cary, R. & Sagle, L.
Collagen-gold nanoparticle conjugates for versatile biosensing. Sensors (Switzerland) 17, (2017).
23. Kitai, T. et al. Simple method of synthesizing nickel-nitrilotriacetic acid gold nanoparticles with a narrow size distribution for protein labeling. Jpn. J. Appl. Phys. 50, (2011).