Design of a FRET-based High-Throughput Screening kit for SUMO1-Ubc9 protein-protein interactions

Design of a FRET-based High-Throughput Screening kit for SUMO1-Ubc9 protein-protein interactions David Bui Richard Lauhead Randall Mello Michelle Tran Team E Advisor: Jiayu Liao June 7 th, 21 1 of 53

Table of Contents Abstract 3 Project Objectives....4 Background.. 5 Prior Art Review.. 6 Functional and Performance Specifications.... 9 Block Diagram of Problem....12 Evolution of the Final Design.......12 Detailed Description Final Solution.....33 Materials Selection.....35 Financial Considerations of Design.......37 Conclusions......38 Future Work....4 Appendix......41 2 of 53

Abstract The detection of protein-protein interactions is an important approach in many cellular processes. To meet market demands, a commercial-friendly kit must be designed to ensure costeffectiveness, easiness to use, and produce efficient and reliable results. A number of existing methods such as the SPA kit and BRET2 kit have suggested that they are relatively time consuming and have limitations and concerns in the use of their application. Therefore we designed a set of experiments in optimizing protein expression, purification, and concentrations used in running a high-throughput screening (HTS) binding assay based upon Förster Resonance Energy Transfer (FRET). We were able to demonstrate that FRET is not affected by the purity of our proteins and focused upon optimizing the yield of CYpet-SUMO1 and Ypet-Ubc9 for implementation in the binding assay. Our hypothesis of using untagged Ubc9 as a mock inhibitor failed in demonstrating a loss in FRET and further investigation is required to incorporate inhibitor studies into the binding assay. A statistically designed HTS parameter, the Z factor, was employed in the optimization of protein concentrations and FRET ratio in overall design for a HTS assay. The Z -Factor depends on dynamic range and variance parameters of the assay and can be used to evaluate the overall quality to meet a standard of greater than.5 from NIH. To obtain a Z -Factor that is approved by NIH standards, our assay is designed to measure the Ratio between the fluorescence emission of the acceptor and the fluorescent emission of the donor. The results demonstrate that the concentrations found to be at a Z factor for an excellent assay are much lower than the commercial BRET2 kit and Bioluminescence Resonance Energy Transfer (BRET) technology. Therefore at these concentrations for CYpet-SUMO1 and Ypet-Ubc9 and its statistically determined excellent assay standards in detection of hits, we expect our in-vitro FRET-based HTS assay kit will be commercially marketable and yield reproducible and high quality results. Project Objectives The objective of this project is to develop an in-vitro high-throughput FRET-based assay kit that will allow for the sensitive detection of a protein-protein interaction between SUMO-1 and UBC9, the E2 3 of 53

enzyme 1. Figure 1. The interaction scheme between CYpet-SUMO1 and Ypet-Ubc9 as applied to the assay. A negative hit occurs when there is no inhibitor present. A positive hit occurs when an inhibitor is present. No FRET Negative hit FRET Positive hit Ypet-Ubc9 CYpet-SUMO1 This kit will allow the user to screen compounds that could interfere with the conjugation of SUMO1 and Ubc9. The kit will include all necessary proteins and solutions to run the assay. In the design of this kit, it will include development of a production procedure and optimization in each step to lessen production cost and time involved to produce the screening kit, in addition a production and kit manual will be included. A Z factor is a value that is used to determine the overall quality of an assay. The binding assay kit will need to meet a Z factor standard of greater than.5 to be validated as an excellent assay as described by the National Institute of Health guidelines 2. The Z -Factor depends on dynamic range and variance parameters of the assay 3. A secondary objective is to genetically engineer CYpet- SUMO1 and Ypet-Ubc9 with an E. coli secretion sequence that will optimize our protein expression and purification protocols. Background SUMO (small ubiquitin-like modifier) belongs to a family of protein modifiers that changes the 1 Knipscheer, Puck, and William J. Van Dijk. "Noncovalent Interaction between Ubc9 and SUMO Promotes SUMO Chain Formation." The EMBO Journal 26 (27): 2797-87. 2 NIH Chemical Genomics Center. Assay Validation. http://www.ncgc.nih.gov/guidance/section2.html#reagentstability-over-time 3 Zhang, Ji-Hu, Thomas D.Y. Chung, and Kevin R. Oldenburg. "Assays A Simple Statistical Parameter for Use in Evaluation and Validation of High Throughput Screening." Journal of Biomolecular Screening 4.2 (1999): 67-73. 4 of 53

function, half-life, or location of proteins when it is covalently attached or detached from the target substrate 4. SUMO modifications play an important role in many biological pathways. Some of these pathways include regulatory gene expression and gene stability. The conjugation of SUMO to its particular substrates requires catalysis of several enzymes. These enzymes include E1 (sumo activating enzyme), E2 (SUMO- conjugating enzyme), and E3 (SUMO Ligases). The specificity between SUMO, the catalyzing enzymes, and the substrate is not fully understood. There are several ways to currently detect protein-protein interactions. A common method involves using an immunoblotting system. The process of immunoblotting requires that first the sample proteins are removed from the cells by lysing then centrifugation. From there the protein sample is treated with SDS-page after denaturing proteins down to their primary structure. Then the sample is placed into a well filled with a polyacrylamide gel with a positive electrode placed at the bottom of the well, and then the proteins will travel down the gel based on size as shown below 5. Figure 2. Schematic of a current method of screening using gel separation to determine protein interactions. The immunoblotting system takes several days, and the number of compounds that can be screened is limited to the number of wells that can fit within a gel. Screening several hundred compounds 4 Kurepa, Jasmina. "The Small Ubiquitin-like Modifier (SUMO) Protein Modification System in Arabidopsis Accumulation of SUMO1 and -2 Conjugates is Increased by Stress." Journal of Biological Chemistry. (22). 5 "Western Blot Protocol." Bioprotocols. N.p., 28 12 28. Web. 23 May 21. <http://www.ebiotek.com/protein/244-western-blot-protocol.html>. 5 of 53

with this method is tedious, slow, and costly. A FRET system requires a system of a Donor and Acceptor fluorophores. This system lends itself to studying protein-protein interactions. Attaching a donor fluorophore to one protein and an acceptor fluorophore to another, the protein interaction can be determined by FRET occurring between the donor and acceptor. FRET allows for detection of proteinprotein interactions without the use of immunoblotting. FRET relies on fluorescence, and fluorescence occurs within a frame of microseconds to seconds. So using a FRET system would allow for extremely fast screening of compounds very quickly. The number of wells within a plate only limits the number of FRET compounds. Prior Art Review Scintillation Proximity Assay (SPA) kit by Perkin Elmer The SPA kit is designed for high-throughput screening (HTS) and relies on the principle of energy conversion of radioactive decay. The main component to the kit is the specially designed beads that contain a scintillant that emits light when excited by radiolabeled molecules. The energy conversion from radioactive decay to emitted photon can only occur when the bead is in close proximity to the radioactive molecule for example in the case of binding. The emitted photons are detected in the photomultiplier tubes of Scintillation counters for analysis. Figure 3. The concept of Scintillation used in the SPA kit and the method of scintillant bead and radioligand used as a reporter in binding assays. (Adapted from http://spotlite.nih.gov/assay/index.php/section5:scintillation_proximity_assays_(spa)) For the case of HTS, the SPA kit is available with Flashplates and Scintiplates in 96 and 384 well 6 of 53

plates. The plates are designed to contain the target molecule conjugated with a scintillant bead in each well. The radiolabeled ligand designed for the sample molecule is added into the well with the sample molecule. Photons will be detected only when the combination sample-radio- ligand binds to the target conjugated scintillant bead. Figure 4. Diagram of the SPA kit using Flashplate technology that is applied to 96 and 384 well plates. (Adapted from http://las.perkinelmer.com/content/images/smallimages/flasplate_tech.jpg) The sensitivity of the SPA kit requires a minimum amount of.125 mg of SPA beads for detection of binding per well but has been shown to be at optimal levels of signal and cost to be at the minimum amount of.35 to.5 mg per well 6. The radioisotope in conjunction with the ligand can emit beta particles from a range of distances of.5um to 17um depending on which radioisotope is chosen and is a factor in the SPA kit s sensitivity. The use of a SPA kit using the FlashPlate technology enhances the sensitivity of the kit and has been demonstrated in protocols to have a concentration of radioligand around 5 pm per well with a receptor amount of.25 µg per well 7. This technology has a high level of selectivity due to the specificity of receptor-ligand binding and enzyme-substrate complexes. A disadvantage of this technology in the long incubation time required before screening can take place. It has been demonstrated that a minimum of 1 hours of incubation is required to reach data that was at a 6 http://www.ncgc.nih.gov/guidance/section5.html 7 Kahl, Steven D., and Christian C. Felder. "Basic Miniaturization of Receptor-Radioligand Interactions using 384- Well FlashPlates." Current Protocols in Neuroscience-Unit 7.15 Scintillation Proximity Assay (25): 1-55. 7 of 53

constant value and high signal to noise 2. Bioluminescence Resonance Energy Transfer (BRET2) kit by Biosignal Packard The BRET2 kit is designed for G-protein coupled receptor (GPCR) high-throughput screening (HTS). The kit uses ß-Arrestin 2 conjugated with a Green Fluorescent Protein mutant (GFP) and a V2 Vasopressin receptor conjugated with the bioluminescent enzyme Renilla Luciferase (RLuc). When a substrate, a derivative of coelenterazine DeepBlueC, is added to the sample the RLuc enzyme catalyzes a reaction with DeepBlueC giving light as a byproduct. If, during the screening process a molecule causes a conformational binding change in the GPCR, The ß- Arrstestin2-GFP fusion protein comes in close contact with the V2-Vasopressin-Rluc-DeepBlueC fusion protein enzyme complex and the fluorescence energy of the enzyme complex is transferred to the GFP fusion protein and can be measured. Figure 5. Diagram of the BRET2 kit and its adaption to the 384-well HTS assay (Adapted from "Bioluminescence Resonance Energy Transfer (BRET) for the Real-Time Detection of Protein-Protein Interactions." Nature Protocols) The kit can also be designed to measure the absence of BRET in the screening for possible inhibitors to the GPCR. The kit comes with the plasmids coding the two fusion proteins required for BRET as well as the substrate DeepBlueC. The sensitivity of the BRET2 technology lies in the crucial 8 of 53

substrate DeepBlueC where a minimum concentration of 5 µm is needed per well 8. The selectivity of the kit is similar to the SPA kit and involves the enzyme Renilla Luciferase with its substrate DeepBlueC. One disadvantage of the BRET2 kit is that the kit is specifically designed for GPCR screening only and other protein-protein interactions cannot be observed and measured. Another disadvantage is that the kit requires the consumer to transform the plasmids into cells, express the protein, and purify the protein. This is time-consuming in the overall goal to measure protein-protein interactions in GPCRs in HTS Assays. Functional and Performance Specifications In developing an in-vitro HTS FRET-based kit, standard procedure conditions were optimized, including expression and purification. To show that the FRET-based assay kit could accurately detect the protein-protein interactions between SUMO1 and Ubc9, some functional specifications are required. The sensitivity of the Flexstation II can be seen as low as 25ng; however, this value is close to the range of noise. Therefore, the amount of protein used in the assay must fluoresce at a level that will not be misinterpreted as noise. Consequently, a minimum of 5ng of each protein will be used in each well of the assay. Other tests have shown that the purity of the protein has no effect on the FRET ratio. Therefore, the purity of the protein is not critical and there will be no specification for purity. However, after several rounds of purification, it has been noted that the purity of the protein is between 45 and 55 percent. We are currently implementing the Z -factor studies into our kit design to test the assay s quality and is a value used in high-throughput screening. A Z -factor ranging from.5 to 1 indicates that the assay kit has met its standards of being deemed an excellent assay. Many requirements of the system have been determined prior to work on the assay kit. These requirements include the determination of the Kd value between CYpet-SUMO1 and Ypet-Ubc9 to be.75 µm by Surface Plasmon Resonance as well as by Fluorescence. Also, determination that FRET occurs between the two proteins CYpet-SUMO1 and Ypet-Ubc9 is another requirement that was investigated prior to kit design. The figures for these can be seen below. 8 Pfleger, Kevin DC, Ruth M. Seeber, and Karin A. Eidne. "Bioluminescence Resonance Energy Transfer (BRET) for the Real-Time Detection of Protein-Protein Interactions." Nature Protocols 1 (26): 337-45. 9 of 53

Figure 6. Diagram of CYpet-SUMO1 and Ypet-Ubc9 and their incorporation into a HTS 384- well format to investigate and analyze protein-protein interactions with the phenomena of FRET. (FRET picture adapted from www.nature.com/.../v4/n7/fig_tab/nrm1153_f2.html) Figure 7. Data obtained using FRET as a means to calculate the Kd between SUMO1 and Ubc9. (This work was completed by Yang Song from the Liao Lab) Plot of Bound Ypet-Ubc9 vs. Free Ypet Ubc9 for Kd determination 1..8 Bound YPet (um).6.4.2. 1 of 53 -.2

Figure 8. This figure illustrates that there is a FRET interaction between CYpet-SUMO1 and Ypet-Ubc9. This was determined by subtracting the signal from Ypet-Ubc9 that was directly excited by the donor excitation wavelength from the raw FRET data spectrum. Since two peaks are still clearly visible in the processed FRET spectrum, a FRET interaction is confirmed. Fluorescence(RFU) 16 14 12 1 8 6 4 2 Fluorophore Spectrums: CYpet and Ypet 45 5 55 6 Wavelength (nm) FRET Spectrum Raw Ypet Spectrum ex: 414 nm Background FRET Spectrum processed Cypet Spectrum Cypet without background 11 of 53

Block Diagram of Problem Protein Expression Cypet-SUMO1, Ypet-Ubc9, Ubc9 2 weeks Protein Purification Ni-NTA Column Chromatography/ Dialysis 2 weeks Protein Characterization Bradford assay, fluorescence characterization, SDS PAGE Protein Gel etc.. 3 weeks Protein Expression Optimization IPTG, 7-16 hour growth/expression 2 weeks Inhibitor Kd Modeling Effects of Inhibitor Kd on Bound Protein, Kdapp equations 2 weeks Ubc9 Mock Inhibitor Studies Ubc9+Ypet-Ubc9/ Ubc9+Cypet- SUMO1/Ypet- Ubc9+Cypet- SUMO1 concentration experiments 2 weeks Z Factor Studies for HTS FRET Ratio optimization Cypet-SUMO1/Ypet- Ubc9 Concentration optimization 3 weeks Kit Design/ Quality Control Lyophilization, Oxidation, and Stability studies with urea 3 weeks Evolution of the Final Design CYpet-SUMO1 and Ypet-Ubc9 Sensitivity Tests The first set of experiments were designed with a dual nature, first to analyze the sensitivity of the fluorescent proteins CYpet-SUMO1 and Ypet-Ubc9 and second to test the overall capabilities of the FlexStation II device in which our kit is designed to function with. As shown in figure 1, we were able to graph a standard set of the fluorescence vs. the amount of protein in nanograms from highly purified samples of CYpet-SUMO1 and Ypet-Ubc9 ranging from to 1ug. From the linear fit and the equation of the line thereof, we will be able to solve for the amount of fluorescent protein in the total protein concentration for future samples of CYpet-SUMO1 and Ypet-Ubc9 from the various optimization experiments. As mentioned above, this experiment also served as a sensitivity test of the FlexStation II 12 of 53

device s capabilities to find an optimum signal level of concentration that is above the noise inherent to the device and the solution that the fluorescent protein is diluted in. After the sensitivity test and standards have been analyzed, we can now proceed with optimizing the protein purification steps for each individual protein based on single fluorescent protein purity as well as FRET efficiency. Although the fit is not necessarily linear, many people do perform a linear fit, which has been done in the literature and has been done so below 9. Figure 9. Linear fit of Fluorescent intensity vs. Protein amount in nanograms for (1) CYpet-SUMO1 and (2) Ypet-Ubc9. (1) Fluorescence(RFU) 25 2 15 1 5 Cypet-SUMO1 Em475 nm y = 26.14x R² =.9963 2 4 6 8 1 Protein amount (ng) Fluorescence(RFU) Linear (Fluorescence(RFU)) 9 Wan, Jiandi, Marlon S. Thomas, Sean Guthrie, and Valentine I. Vullev. "Surface-Bound Proteins with Preserved Functionality." Annals of Biomedical Engineering 37.6 (29): 119-25. 13 of 53

(2) Fluorescence(RFU) 7 6 5 4 3 2 1 Ypet-UBC9 Em 53nm 5 1 15 Protein amount (ng) y = 632.68x R² =.984 Fluorescence(RFU) Linear (Fluorescence(RFU)) Table 1. Fluorescent intensities of CYpet-SUMO1 (1) and Ypet-Ubc9 (2) at protein amounts of to 1 µg to determine the capabilities of the FlexStation II device used in design of the SUMO kit. From these values we determined that the amount of fluorescent protein that is away from the noise level of the machine and still at low amounts for optimization is at 5 ng. (1) Cypet-Sumo1 Fluorescence Fluorescence Fluorescence Average Std dev (ng) (RFU) (RFU) (RFU) 915 2.27E+7 2.35E+7 2.35E+7 2.32E+7 4.52E+5 5 9.47E+6 1.39E+7 1.51E+7 1.45E+7 8.33E+5 25 5.35E+6 8.32E+6 9.62E+6 7.76E+6 2.19E+6 1 2.24E+6 2.56E+6 4.4E+6 3.6E+6 1.17E+6 5 1.2E+6 1.3E+6 2.7E+6 1.46E+6 5.4E+5 25 5.5E+5 6.91E+5 7.42E+5 6.46E+5 1.25E+5 1 1.48E+5 2.6E+5 2.3E+5 1.94E+5 4.22E+4 5 7.1E+4 8.17E+4 8.19E+4 7.79E+4 6.75E+3 1 1.9E+4 1.6E+4 1.94E+4 1.81E+4 1.88E+3 5 1.34E+4 1.62E+4 1.52E+4 1.49E+4 1.4E+3 25 1.72E+4 1.27E+4 1.3E+4 1.43E+4 2.5E+3 1 1.8E+4 1.12E+4 1.29E+4 1.16E+4 1.14E+3 5 1.4E+4 1.22E+4 1.24E+4 1.16E+4 1.1E+3 14 of 53

2.5 1.1E+4 1.12E+4 1.26E+4 1.16E+4 8.95E+2 1 1.15E+4 2.26E+4 1.21E+4 1.54E+4 6.19E+3 1.2E+4 9.14E+3 1.22E+4 1.11E+4 1.71E+3 (2) Ypet-Ubc9 (ng) Fluorescence Fluorescence Fluorescence Average Stdev (RFU) (RFU) (RFU) 1 5.87E+7 5.6E+7 6.2E+7 5.89E+7 3.1E+6 5 3.64E+7 3.61E+7 3.93E+7 3.73E+7 1.76E+6 25 2.2E+7 2.1E+7 2.34E+7 2.12E+7 1.84E+6 1 7.2E+6 7.62E+6 9.62E+6 8.15E+6 1.29E+6 5 4.32E+6 3.68E+6 4.15E+6 4.5E+6 3.32E+5 25 1.91E+6 1.68E+6 1.83E+6 1.81E+6 1.16E+5 1 4.41E+5 4.25E+5 4.41E+5 4.36E+5 9.4E+3 5 1.26E+5 2.9E+5 2.55E+5 1.97E+5 6.52E+4 1 1.69E+4 2.47E+4 2.9E+4 2.8E+4 3.92E+3 5 1.1E+4 2.64E+4 1.86E+4 1.87E+4 7.71E+3 25 8.62E+3 1.22E+4 1.59E+4 1.22E+4 3.65E+3 1 6.83E+3 8.77E+3 1.6E+4 8.72E+3 1.86E+3 5 6.62E+3 1.7E+4 9.1E+3 8.81E+3 2.5E+3 2.5 6.57E+3 1.14E+4 9.74E+3 9.23E+3 2.44E+3 1 6.82E+3 9.E+3 8.4E+3 8.7E+3 1.12E+3 8.25E+3 9.14E+3 7.43E+3 8.27E+3 8.55E+2 CYpet-SUMO1 and Ypet-Ubc9 Purification Optimization To optimize our protein purification steps for CYpet-SUMO1 and Ypet-Ubc9, we chose to run an experiment on the Ni-NTA bead affinity chromatography step adapted from Qiagen Ni-NTA agarose beads purification booklet. As shown in table 2, each wash protocol was designed by varying concentrations of NaCl, Tris-HCl, Imidazole, and Triton in wash buffers 1, 2, 3, and elution. For every 1- liter of expressed CYpet-SUMO1 or Ypet-Ubc9 we used 5 µl of Ni-NTA beads, 1 ml of wash buffers 1, 2, and 3, and 5 µl of elution buffer. Each sample of fluorescent protein was collected and then analyzed based upon the overall protein concentration and the fluorescent protein concentration to determine each sample s purity of Cypet-SUMO1 and Ypet-Ubc9 as discussed in the next section. 15 of 53

Table 2. Protein purification optimization protocols for Ni-NTA bead affinity chromatography adapted from Qiagen Ni-NTA agarose beads purification booklet. (2) Resuspension buffer containing NaCl, Tris- HCl, and Imidazole used to resuspend the E.coli pellets after centrifugation for lysis and subsequent Ni- NTA bead affinity chromatography. (1) Ingredients Wash1 Concentration of Solutions (M) Protocol 1 Protocol 2 Protocol 3 NaCL.3.5.4 Tris HCL.2.2.2 ph 7.4 Wash2 NaCL.3 2 1.2 Tris HCL.2.2.2 ph 7.4 Triton.5% 2.% 1.25% Wash3 NaCL.3 2 1.2 Tris HCL.2.2.2 ph 7.4 Imidiazole.2.5.35 Elution NaCL.3.3.3 Tris HCL.2.2.2 ph 7.4 Imidiazole.15.25.2 16 of 53

(2) Resuspension Buffer Concentration (M) NaCl.5 Tris-HCl ph.2 7.4 Imidiazole.5 CYpet-SUMO1 and Ypet-Ubc9 Purity Tests The three different protocols mentioned above for the protein purification optimization experiment were analyzed for the highest fluorescent protein amount over the total protein and other contaminants after Ni-NT bead affinity chromatography using the Bradford Assay and the linear fit equation of fluorescence versus protein amount determined in the CYpet-SUMO1 and Ypet-Ubc9 sensitivity tests mentioned earlier. The Bradford Assay is based upon relating the absorbance of a standard set of known concentrations of some sample protein e.g Bovine Serum Albumin to construct a graph relating the absorbance versus the amount of total protein. From the graph we can determine the linear fit and the respectable equation of the line to use for our samples of CYpet-SUMO1 and Ypet-Ubc9 based on their absorbance in the range of the standard set to determine overall protein amount in nanograms. Once the total protein amount and fluorescent protein amount are determined, we can calculate the ratio of the two to find the overall fluorescent protein purity of CYpet-SUMO1 and Ypet- Ubc9 as a percentage. Therefore determining the protein s overall purity can be used as an aid in developing the next experiments in determining if purity does effect the fluorescent proteins and their ability for FRET for optimizing our SUMO pathway-based kit. Figure 1. Standard Bradford Assay linear fit of Absorbance vs. Protein amount in nanograms. From the equation of the line we were able to determine total protein amount of CYpet-SUMO1 and 17 of 53

Ypet-Ubc9 samples from their respectable absorbance. These calculated amounts can then be used with the fluorescent protein amounts from figure 1 to determine the overall purity of our expressed CYpet- SUMO1 and Ypet-Ubc9 as shown in table 3. 1.4 1.2 Absorbance vs Protein amount y = 2E-1x 3-1E-6x 2 +.18x -.71 1.8.6 abs Poly. (abs).4.2 5 1 15 2 25 -.2 Protein amount(ng) Table 3. Purity results of (1) CYpet-SUMO1 and (2) Ypet-Ubc9 from the protein purification optimization protocols 1 through 3. (1) Cypet-SUMO1 Purification protocol Bradford concentrations (ng/µl) Fluorescent concentration (ng/µl) Purity Protocol1 4.84E+3 4.71E+3.97 Protocol2 3.19E+3 1.68E+3.53 Protocol3 4.64E+3 3.23E+3.7 (2) Ypet-UBC9 Purification protocol Bradford concentrations (ng/µl) Fluorescent concentration (ng/µl) Purity Protocol1 7.71E+3 5.9E+3.66 Protocol2 9.55E+2 6.17E+2.65 Protocol3 5.5E+3 2.78E+3.51 Purity tests on CYpet-SUMO1 and Ypet-Ubc9 based on protein s fluorescence and FRET efficiency In this experiment we wanted to determine if the purity of CYpet-SUMO1 and Ypet-Ubc9 proteins has any effect on the fluorescence of the single proteins as well as in FRET. This is important in the design of a functional kit based on FRET screening and for future optimization techniques because 18 of 53

purity can determine the range to where we need to optimize the protein expression and purification protocols for an efficient kit. The experiment was designed to incorporate BL21 E. coli cell lysate as the impure portion of our samples at varying amounts to create a range of purity percentages in which to analyze the fluorescence thereof. As shown in figure 3, CYpet-SUMO1 and Ypet-Ubc9 were kept at a constant amount of 1 µg in the single protein fluorescence studies and then 5 ng of each protein for the FRET efficiency studies in (3). It can be determined that the purity of the protein has little or no effect on the FRET ratio and thus little no effect on the assay. The FRET purity does show data that is statistically different from the 1% pure; however, the difference is small and the purities are scattered, so the difference may be due to pipette variation. Graphs of the spectrums that were used to calculate the bar graphs and p values for all the experiments can be seen in appendix VIII. Figure 11. Graph of Percent Purity vs. RFU at 475 nm for (1) CYpet-SUMO1, (2) Ypet-Ubc9, and (3) FRET of donor CYpet-SUMO1 to acceptor Ypet-Ubc9. Each colored line represents each sample at different percentages of purity to due varying amounts of BL-21 lysate impurity. Each peak represents the emission maximum for each protein e.g CYpet-SUMO1 at 475 nm and Ypet-Ubc9 at 53 nm. The figure in (4) represents the FRET efficiency of acceptor fluorescence over donor fluorescence versus the percent purity of CYpet-SUMO1 and Ypet-Ubc9. (1) Purity Effects on Cypet-SUMO1 1ug RFU at 475nm 3 25 2 15 1 5 Purity Effects on Cypet- SUMO1 1ug 1% 9% 8% 7% 6% 5% 4% 3% 2% 1% Percent purity Comparing 1% with others P-Values P<.5 indicated by * P-Values P<.1 indicated by ** 19 of 53

(2) RFU at 53 nm 7 6 5 4 3 2 1 YPET-UBC9 purity test 1 ug YPET-UBC9 purity test 1 ug 1% 9% 8% 7% 6% 5% 4% Percent purity 3% 2% 1% Comparing 1% with others P-Values P<.5 indicated by * P-Values P<.1 indicated by ** (3) 53/475 Fluorescent Ratio.9.8.7.6.5.4.3.2.1 Purity effects on FRET ratio 5ng * Purity effects on FRET ratio 5ng Percent Purity Comparing 1% with others P-Values P<.5 indicated by * P-Values P<.1 indicated by ** Lyophilization experiments of CYpet-SUMO1 and Ypet-Ubc9 for Kit Design Lyophilization is a commonly used method of increasing the concentration of the desired proteins 2 of 53

as well as enhancing the stability of the protein for long-term use in industry and therefore a possible method of optimizing our FRET based SUMO kit. The experiment was designed by lyophilizing 1 ml of CYpet-SUMO1 and Ypet-Ubc9 in 1.5 ml tubes and then broken down into two main categories: stability testing without resuspension of powder and stability testing with resuspension of powder. The following proteins were then studied in four different temperature parameters: room temperature (~27 C), 4 C, - 2 C, and -8 C. An overview of this experiment can be visualized in figure 4. To determine the overall effects of lyophilization on protein stability and FRET efficiency, we obtained the purity of CYpet- SUMO1 and Ypet-Ubc9 before lyophilization to dilute each sample at the same concentration. As shown in figure 5, after 3 days of stability testing of the lyophilized protein samples they were analyzed at their respectable emission maxima of single protein fluorescence of (1) CYpet-SUMO1, (2) Ypet-Ubc9, and (3) FRET efficiency of donor CYpet-SUMO1 and acceptor Ypet-Ubc9. Each separate line on the graphs corresponds to either resuspended or powder samples and at the four varying temperatures mentioned above. The data shows that lyophilization does have an effect on stability at room temperature and 4 degrees. However, if the protein is stored at negative 2 degrees then lyophilization has little effect on the protein stability. This shows that lyophilization is not necessary for protein storage of the assay as long as it is kept frozen at negative 2 degrees or less. After optimizing lyophilization we can then move forward to the last step in our protein optimizing pathway: Protein Expression. Figure 12. Overview of the Lyophilization Experiment for Optimization of the SUMO kit. 21 of 53

Figure 13. Protein stability after 3 days of Lyophilized samples of (1) CYpet-SUMO1, (2) Ypet- Ubc9, and (3) FRET of CYpet-SUMO1 to Ypet-Ubc9 and FRET ratio. Each line corresponds to the fluorescent proteins at different temperatures and either resuspended or in powder form. (1) RFU at 475 nm 7 6 5 4 3 2 1 Cypet-SUMO1 lyophilization effects * ** ** ** ** ** Cypet-SUMO1 lyophilization stability lyophilization test (2) 22 of 53

Ypet-UBC9 lyophilization stability 18 16 14 RFU at 53 nm 12 1 8 6 4 2 ** Ypet-UBC9 lyophilization stability Comparing non-lyophilized with others P-Values P<.5 indicated by * P-Values P<.1 indicated by ** Lyophilization test (3) FRET Ratio Lyophilization effects 1.4 Emission Ration 53nm/475 nm 1.2 1.8.6.4.2 ** FRET lyophilization Stability Lyophilization test Comparing non-lyophilized with others P-Values P<.5 indicated by * P-Values P<.1 indicated by ** Protein Expression Optimization of CYpet-SUMO1 and Ypet-Ubc9 for Kit Design 23 of 53

Expression of CYpet-SUMO1 and Ypet-Ubc9 from our genetically engineered BL21 E. coli cells is a crucial step in optimizing our kit design because it directly controls the amount of fluorescent protein expressed as well as the functionality of each respectable protein. The main parameter in this experiment is the expression-signaling molecule Isopropyl β-d-1-thiogalactopyranoside (IPTG). The experiment consists of four protocols subdivided into two different IPTG-induced expression times: same day (8 hours) and overnight (16 hours). Each protocol is also divided into three different amounts of IPTG added: 1, 5, and 1 µl. After subsequent cell lysis and purification using the optimum Ni- NTA bead affinity chromatography protocol mentioned previously, each sample of fluorescent protein CYpet-SUMO1 and Ypet-Ubc9 was then analyzed by the linear fit of the standards for each protein from the fluorescent protein fluorescence versus amount in nanograms. The resulting fluorescent protein concentrations for CYpet-SUMO1 and Ypet-Ubc9 for each protocol and IPTG parameter can be seen in table 4. It is clear that Protocol 2 is the best protocol to use for expressing both CYpet-SUMO1 and Ypet- Ubc9. It is overwhelmingly better for CYpet-SUMO1 production and slightly better for Ypet-Ubc9 production. Figure 14: results of the expression optimization studies showing that protocol 2 is the best protocol for maximizing protein expression. Fluorescent Protein concentration(ng/ul) 14 12 1 8 6 4 2 Expression optimization Cypet-SUMO1 Concentration produced Ypet-UBC9 Concentration Produced Protocol 24 of 53

Table 4. Protein expression optimization experiment overview of different protocols and varying parameters used for (1) CYpet-SUMO1 and (2) Ypet-Ubc9. The resulting fluorescent protein concentrations can be seen for each protocol to determine the optimal expression protocol. Unfortunately, due to time constraints, only one test was run for protein expression. More will be completed in the future. (1) Cypet-SUMO1 IPTG Expression Time IPTG Added (µl) Fluorescent protein concentration (ng/µl) Protocol 1 Same day 5 76 Protocol 2 Overnight 1 972 Protocol 3 Same day 1 59 Protocol 4 Overnight 5 1396 (2) Ypet-Ubc9 IPTG Expression Time IPTG added (µl) Fluorescent Protein Concentration (ng/µl) Protocol 1 Same day 5 7831 Protocol 2 Overnight 1 11579 Protocol 3 Same day 1 7827 Protocol 4 Overnight 5 9586 Stability Studies The purpose of the stability and oxidative testing was to provide evidence on how the quality of the FRET interaction varies with time under the influence of a few environmental factors such as temperature and oxygen. Our goal was to determine which storage setting would retain the protein-protein interaction at its optimal for certain time periods. The temperatures that were tested were room temperature (RT), 4 degrees Celsius, negative 2 degrees Celsius, and negative 8 degrees Celsius. From the results we obtained, we found out that the protein evaporated when there was oxygen present at room temperature and 4 C when stored for more than 1 day. This explains why there is no data for these samples. The fresh proteins that had been stored for only 6 hours had the best FRET interaction with the highest ratio of 2.28. However, we see a slight decrease in the FRET ratio as time increases. We noted that after 3 days of storage, the FRET interaction were not statistically significant with a 95% confidence among the various storage settings. Although it is not statistically significant, it is more favorable to store the proteins at -2 C since most of the data shows that it retained the highest FRET interaction. The data obtained from several of these days can be seen below in the various graphs. One error throughout the experiments that contributed to the various fluctuations among the days was pipette variations. 25 of 53

Figure 15. Stability studies showing the difference in assay conditions based on the Temperature, Oxidation (Ox) and Time. The times tested were 6 hours (1), 1 day (2), 3 days (3), 5 days (4) and 7 days (5). (1) 3 Stability and Oxidation Tests-6 Hours 2.5 53:475 Ratio 2 1.5 1 ** ** * * * ** ** 53nm/4 75nm FRET.5 RT Ox RT No Ox 4 Ox 4 No Ox -2 Ox -2 No Ox Storage Temperature, Oxidation -8 Ox -8 No Ox Comparing with -2 o C no ox P-Values P<.5 indicated by * P-Values P<.1 indicated by ** 26 of 53

(2) Stability and Oxidation Tests- 1 Day 1.4 1.2 1 * 53:475 Ratio.8.6.4 53nm/475nm FRET ratio.2 (3) RT Ox RT No Ox 4 Ox 4 No Ox -2 Ox -2 No -8 Ox -8 No Ox Ox Comparing with -2 o C no ox Storage Temp and Oxidation P-Values P<.5 indicated by * P-Values P<.1 indicated by ** Stability and Oxidation Tests-Day 3 1.4 1.2 1 * * * 53:475 Ratio.8.6.4 53nm/475nm FRET ratio.2 RT Ox RT No Ox 4 Ox 4 No Ox -2 Ox -2 No Ox Storage Temperature, Oxidation -8 Ox -8 No OxComparing with -2 o C no ox P-Values P<.5 indicated by * P-Values P<.1 indicated by ** 27 of 53

(4) 1.2 Stability and Oxidation Tests-Day 5 1 53:475 Ratio.8.6.4 53nm/475nm FRET ratio.2 RT Ox RT No Ox 4 Ox 4 No Ox -2 Ox -2 No Ox -8 Ox -8 No Ox Storage Temperature, Oxidation (5) 1.2 Stability and Oxidation Tests-Day 7 1 53:475 Ratio.8.6.4 53nm/475nm FRET ratio.2 RT Ox RT No Ox 4 Ox 4 No Ox -2 Ox -2 No -8 Ox -8 No Comparing with -2 Ox Ox o C no ox P-Values P<.5 indicated by * Storage Temperature, Oxidation P-Values P<.1 indicated by ** 28 of 53

Mock Inhibitor Studies After determining many of the parameters that would make the assay useable, the mock inhibitor should be tested for effectiveness. As there are no known inhibitors for the inhibition of SUMO1 and Ubc9, this test will require the use of an untagged acceptor protein UBC9. The untagged Ubc9 will compete with Ypet-Ubc9 for the binding site on CYpet-SUMO1. The FRET process depends on the proximity of CYpet to Ypet; therefore, if untagged Ubc9 competes for the binding site, then FRET will decrease within the system. The more Ubc9 that is added means less FRET interactions that are likely to take place. This can be measured by analyzing fluorescent emission at CYpet emission maximum and Ypet emission maximum and taking their ratios. TO determine the effectiveness of the assay, UBC9 was titrated into wells containing.21 µm CYpet-SUMO1 and.39 µm Ypet-Ubc9. The titration concentrations were determined as multiples of Ypet-Ubc9, so the titration ranges from 1 % the amount of Ypet-UBC9 to 1 times the amount of Ypet-Ubc9. The fluorescence was measured at, 15, 3 and 45 minutes. The data from 15 minutes through 45 minutes is nearly the same, so only the 3-minute data is displayed below. What can be seen is an increase in FRET ratio as there is an increase in untagged Ubc9, this is the opposite effect of what we expected. The cause of these results is still under investigation. Figure 16. This graph displays the increase in binding caused by an increase in concentration of Ubc9. 29 of 53

1.2 Average FRET RATIO time 3 min 1 FRET Ratio 53/475.8.6.4 * * Average FRET RATIO time 3 min.2.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.511.5 UBC9 concentration as a Multiple of Ypet-UBC9 concentration Comparing with the [Ubc9] well P-Values P<.5 indicated by * P-Values P<.1 indicated by ** Z Factor parameters To make a useable assay there are several parameters that must be met. These parameters are important to make a reliable High-throughput Screening assay. These parameters include the ratio of donor and acceptor that must be placed into a well and the total amount of both proteins into each well respectively. These amounts can be determined making use of Z factor calculations. Z factor values that are in the range of.5 to 1 validate the parameters for an excellent assay. 1 is the value of an ideal assay 1. Z factor is dependent on mean and standard deviation so for each of the following assays, each test was repeated 1 times. Also, the determination of Z factor depends on the use of positive and negative controls in the assay. Since the Ubc9 as a control did not work as expected, a concentration of urea was implemented to act as the substance that will provide the positive hits. Although urea is not an inhibitor, it denatures protein by hydrogen bonding. Dialysis buffer was then used for the negative hits to calculate the Z factor. To determine the correct ratio, an assay was performed that held the concentration of CYpet-SUMO1 at a constant concentration of.21 µm and varied the amount of Ypet-Ubc9. When a 1 Zhang, Ji-Hu, Thomas D.Y. Chung, and Kevin R. Oldenburg. "Assays A Simple Statistical Parameter for Use in Evaluation and Validation of High Throughput Screening." Journal of Biomolecular Screening 4.2 (1999): 67-73. 3 of 53

ratio is reached that contains a Z factor above.5, then that ratio is a valid ratio for an excellent assay. After determining the ratio of CYpet-SUMO1 to Ypet-Ubc9, the total amount of protein must be determined. For this assay, the ratio of the donor and acceptor were held constant, but the amount of each protein was increased. When the Z factor for this test reaches above.5, then the total amount of protein needed per well is determined. Calculation of Z factor in the amount assay is the same as the ratio assay. The results are shown below in figure 6. The Ratio selected for the kit is 1:2.8. Although it is not the minimum ratio that satisfies the Z-factor requirement, it is far enough above.5 to allow for protein degradation with time. The same logic is followed for selecting the minimal amount of protein needed for the well..25 µm was chosen as the Cypet-SUMO1 concentration and.7µm was chosen as the Ypet- Ubc9 concentration. Figure 17. Z factor study graphs to first in (1) to determine the optimal amount of urea needed to bring the FRET ratio down to be a model for a positive hit in the screening. (2) Graph to find the Z factor values with a range of FRET ratios to determine the optimal FRET ratio at a Z factor value that would make this assay an excellent assay. (3) Graph of calculated Z factor values with a set FRET ratio found in (2) and a range of CYpet-SUMO1 and Ypet-Ubc9 concentrations to determine the optimal amount of proteins needed for an excellent assay with a Z factor above.5. 31 of 53

(1) FRET Ratio 53/475.9.8.7.6.5.4.3 Average FRET ratios with different concentrations of Urea ** ** ** ** 1 2 3 4 Urea Concentration M Average FRET ratios with different concentrations of Urea Comparing concentrations with preceding concentrations P-Values P<.5 indicated by * P-Values P<.1 indicated by ** (2) 1.8 Z factor with each well having.21 um Cypet-Sumo1 in 6 ul DB.6 Z Factor.4.2 -.2 1:1.2 1:2 1:2.8 1:3.6 1:4.4 1:5.2 Z factor with each well having.21 um Cypet-SUmo1 in 6 ul DB -.4 [CyPet-SUMO1] : [YPet-Ubc9] ratio 32 of 53

(3) Z Factor.8.7.6.5.4.3.2.1 -.1 Z factor for determining a useable amount of protein.15.2.25.3.35.4 Amount of Cypet um Z factor for determining a useable amount of protein Detailed Description of the Final Solution Model of Solution The purpose of this assay is to show bound CYpet-SUMO1 and YPET-UBC9 based on the properties of the inhibitor. The only parameter that changes from one test compound to the next is the dissociation constant of the compound, so there should be a solution that can model Bound Protein vs. the dissociation constant of an inhibitor. To setup this problem, the formula for K d was used 1. K dapp = [CCCC1][YYYY9] [BBBB] 2. K dapp = K d (1 + [II] K i ) Where K dapp is the apparent dissociation constant, CS1 is CYpet-SUMO1, YU9 is Ypet-UBC9, I is inhibitor, K i is dissociation constant of inhibitor, K d is the dissociation constant between SUMO1 and Ubc9 and BP is the bound protein complex CYpet-SUMO1*Ypet-Ubc9. Solving the equation above results in the following: 33 of 53

[ CS1]*[ YU9] [ BP] = [ I] Kd (1 + ) Ki And with all the variables remaining constant except BP and Ki. A plot of BP vs. Ki can be constructed from the following: [ CS 1] =.25µM, [ YU 9] =.7µ M, [ I ] = 5µ M,. Figure 18: Plot of Bound protein vs. Dissociation constant of inhibitor. max Kd of Inhibitor (µm) The Concentration of Bound protein approaches the value of.233 µm as the dissociation constant approaches infinity. This is expected because.233µm is the amount of bound protein when there is no inhibitor present. The high concentration of inhibitor that is used affects the curvature of the plot, so if the assay is deemed to sensitive, the inhibitor concentration can be reduced, and the range of inhibitor dissociation constants that can be detected will also be reduced. As the assay stands now, and inhibitor with a dissociation constant in the 1 molar range could be detected. (Appendix A) 34 of 53

Materials Selection Expression In the expression process we used BL21 E.coli to express the target proteins CYpet-SUMO1 and Ypet-Ubc9 based on the fact that this bacteria as an expression agent has been well established and the bacteria does not do any post-translational modifications. In addition to this we genetically engineered a Histidine (His) tag to our proteins for the purpose of elution because it has been understood that a His tag will bind to the Nickel beads (Ni-NTA) which will be discussed in the purification process. Purification Process Wash buffers: Tris-HCl, Triton, NaCl We chose this combination for our first wash buffer based on the QIAexpressionist handbook. Tris-HCl was implemented to stabilize the ph of the overall wash buffer due to the fact that it absorbs counter ions. Triton and NaCL were used because together they act as a detergent and thereby remove weak hydrophobic interactions and help purify the proteins from any contaminants. Elution buffers: Imidazole, Ni-NTA beads Because our target proteins have a Histidine residue attached to them, as it is filtered through an affinity chromatography tube filled with Ni-NTA beads the proteins will bind strongly to the beads. To elute our proteins through this column would require a compound that s binding affinity is stronger than that of our His-tagged proteins to outcompete for binding sites on the nickel beads and therefore allow the proteins to elute through. We chose Imidiazole because within the Histidine residue it contains an imidaizole ring itself, but its binding ability to Ni-NTA beads is reduced due to interactions with the rest of the structure. Imidiazole on its own has a higher affinity to Ni-NTA beads and therefore we choose this compound for our elution buffer due to the fact that it can outcompete the Histidine residues on CYpet-SUMO1 and Ypet-Ubc9 for binding to the Ni-NTA beads. FRET Binding Assays: FlexStation II, Ypet, CYpet 35 of 53

All other testing after purification was based on fluorescence detection with the use of the FlexStation II. This detection device was chosen for its ability to quickly test for fluorescence of our proteins with tagged fluorophores. The machine also allows for the ability to excite our samples with specified wavelengths allow us to screen 384 samples at one time in a HTS format. CYpet (Cyan fluorescent Protein for energy transfer) and Ypet (Yellow fluorescent Protein for energy transfer) were chosen as the ideal fluorophores due to their high quantum yields as well as their large Stoke s shift. CYpet, when excited with a wavelength of 475 nm, emits a cyan color and Ypet emits a yellow color when excited at 53 nm and for our studies of the protein-protein interaction of SUMO1 and Ubc9 when the two proteins bind to each other at the distance of 2-1 nm, they allow for Förster resonance energy transfer (FRET) to occur 11. This interaction is why CYpet and Ypet were chosen to be the fluorophores genetically engineered onto SUMO1 and Ubc9 and the phenomena of FRET easily allow us to test whether or not the proteins have bound together. Financial Considerations of the Design As shown in table 5, each item and its respectable cost were determined based upon the brand and type used. Every item was graciously provided by our advisor Dr. Jiayu Liao and the FlexStation II device used in our fluorescence measurements and high-throughput screenings was from Dr. Jiayu Liao s laboratory. Additionally our senior design was funded by the programs 2 dollars to provide for some of the materials used e.g. NaCl, Triton, Tris-HCl, and primers used in construction of our genetically engineered proteins and for the purification optimization experiments. kit. Table 5. Overview of every item used in optimizing and designing our HTS FRET based SUMO Items Cost ($) Ni-NTA beads (25ml) 243 11 Dos Remedios, Cristobal G., and Pierre D.J. Moens. "Fluorescence Resonance Energy Transfer Spectroscopy Is a Reliable "Ruler" for Measuring Structural Changes in Proteins." Journal of Structural Biology 115 (1995): 175-85. 36 of 53

2XYT Pellets (x32) 65 NACl (1kg) 15 IPTG (25g) 2 Kanamycin (25g) 26 Tris-HCl (1L) 38.2 Triton (5 ml) 21 Imidazole (2g) 5 Primers (2) 1 384-Falcon well plate (1) 54 Falcon Tubes (25) 65 Miscellaneous 2 Total 1131.2 Conclusions By employing the phenomena of FRET in the design of a binding assay between Cypet- SUMO1 and the E2 conjugating enzyme Ypet-Ubc9, we were able to engineer a set of experiments and models for creation of a kit for High-Throughput screening. This binding assay kit will give the user the ability to screen compounds in one 384-well plate for interactions between Cypet-SUMO1 and Ypet-Ubc9 with the designated buffer and positive control of urea to be a model for a positive hit. To create such a kit that will be valid and accurate in its detection of positive and negative hits, we designed a set of experiments to optimize protein expression, purification, sensitivity of detection using the FlexStation II, and selectivity of Cypet-SUMO1 and Ypet-Ubc9. Finally, to optimize the binding assay we calculated Z factor values for the FRET ratio and then the concentrations of CYpet-SUMO1 and Ypet-Ubc9 at a determined FRET ratio. We then optimized the assay s sensitivity by choosing the FRET ratio and concentrations at the minimum values above a Z factor of.5. The values chosen from the Z factor studies as shown in table 6 below. Table 6. The determined amounts of proteins CYpet-SUMO1 and Ypet-Ubc9 in concentration per well and total volume for a 384-well plate. The optimum FRET ratio is also 37 of 53

shown. The total volume per well is 6 µl with 2 µl CYpet-SUMO1, Ypet-Ubc9, and 2 µl of negative control buffer, positive control Urea at a concentration of 3 M or compound. Optimum Values from Z-Factor CYpet-SUMO1 Ypet-Ubc9 Test (>.5) FRET Ratio (53/475) 1 2.8 Concentration per well (µm).25.7 Total Protein amount per 384 well plate (µg) 76 251 At these concentrations and at this FRET ratio, our designed binding assay functions at a Z factor value greater than.5, which validates this assay as an excellent assay for use with the FlexStation II HTS device. Every step in production of proteins CYpet-SUMO1 and Ypet-Ubc9 was optimized to yield the maximum amount of proteins per every one-liter of cells grown. It was further demonstrated that purity of protein has no effect in the FRET ratio at protein amounts at the FlexStation II device s minimum level of detection of 5 ng. Our design also meets the specifications laid out in the objectives by creating an assay that meets the standards of binding assays with the Z factor at the minimum amount of proteins needed for a stable FRET ratio to be analyzed in a HTS format. The area in our design that did not meet the criteria and specifications laid out in our objectives is the mock inhibitor studies using untagged Ubc9. Our hypothesis that Ubc9 would act as a model inhibitor for use as a positive control for inhibition binding assays was incorrect and was therefore not included in our binding assay kit design. Also our secondary objective to engineer the proteins CYpet-SUMO1 and Ypet-Ubc9 with a secretion sequence to optimize protein expression and purification was not met in the time of finishing this design and will be looked into further in our future work. 38 of 53

The concentrations and amounts of CYpet-SUMO1 and Ypet-Ubc9 that will be employed in the HTS binding assays based upon the Z factor calculations can be shown to be at the amounts similar and in some cases much lower to those in the binding assay kits commercially available and reviewed in the prior art review. One study and its conclusion shown in figure 7 in the BRET2 kit with its reporter Luciferase gives concentrations that meet the criterion of a Z factor above.5 starting at a concentration of 3 µm. Our binding assay demonstrates a Z factor of.58 at a concentration of.7 µm for reporter Ypet-Ubc9. The BRET2 kit also requires a substrate for the reporter Luciferase activation, DeepBlueC, which is at a concentration of 5 µm per well 4. In contrast, the substrate in our FRET-based HTS binding assay, CYpet-SUMO1, is at a concentration of.25 µm per well with a Z factor of.58. As mentioned in the prior art review the SPA kit with FlashPlate technology has concentrations of radioligand substrate 1 times lower than our concentrations for CYpet-SUMO1. One advantage in our design of a FRET-based technology is that where the SPA kit has a minimum reporting distance of 5 um, FRET has a minimum reporting distance of 2 nm. Another advantage of our design over the SPA kit is that our kit can give reliable values for binding assays at any time, whereas the SPA kit requires a minimum of 1 hours of incubation before the assay can be run. Therefore our design of a FRET-based HTS kit offers advantages of less reporter and substrate amounts as well as a rapid time of detection and usage when compared to conventional HTS binding assays commercially available. The experiments and optimization procedures mentioned in this report demonstrates that our design has met the objectives of our project to create a FRET-based binding assay to be used in HTS to detect positive hits in disruption of the CYpet-SUMO1 and Ypet-Ubc9 binding complex. 39 of 53

Figure 19. Table of concentration of reporter Luciferase and corresponding Z factor from 384-well plate HTS assay based upon BRET technology. Adapted from Molecular Biology in Medicinal Chemistry Volume 21, Wiley-VCH 24. Future Work To bring our project to the next level, we need to further analyze and investigate a mock inhibitor to be incorporated into our kit design. This will aid in bringing versatility in the kit where a user can not only screen for positive hits but can also compare and analyze the characteristics of the positive compound in its degree of inhibition. Also we wish to further optimize the protein expression and purification of proteins used in our kit by engineering CYpet-SUMO1 and Ypet-Ubc9 with an E. coli secretion sequence that will secrete our proteins into the extracellular environment instead of intracellularly. At the end of our design we will package the materials needed and create a user manual to bring it to the commercial level. 4 of 53

Appendix I. List of Abbreviations FRET- Forster Resonance Energy Transfer SUMO- Small Ubiquitin Like Modifier Ubc9- Ubiqutin Conjugating Enzyme 9 Ox- Oxidation HTS- High-Throughput Screening BRET- Bioluminescence Resonance Energy Transfer SPA- Scintillation Proximity Assay GPCR- G-Protein Coupled Receptor CFP- Cyan Fluorescent Protein YFP- Yellow Fluorescent Protein RFU- Relative Fluorescent Units CS1- CYpet-SUMO1 YU9- Ypet-Ubc9 BP- Binding Protein 41 of 53

Appendix II. Project Budget Items Cost ($) Ni-NTA beads (25ml) 243 2XYT Pellets (x32) 65 NACl (1kg) 15 IPTG (25g) 2 Kanamycin (25g) 26 Tris-HCl (1L) 38.2 Triton (5 ml) 21 Imidazole (2g) 5 Primers (2) 1 384-Falcon well plate (1) 54 Falcon Tubes (25) 65 Miscellaneous 2 Total 1131.2 42 of 53

Appendix III. List of Equipment and Facilities FlexStation II Benchtop Scanning Fluorometer Eppendorf Pipettes Flasks Facility: Liao Lab Autoclave Facility: Autoclave Room, Bourns College of Engineering 43 of 53

Appendix IV. Team Job Responsibilities David Bui o Protein Expression o Protein Purification o Protein Characterization Richard Lauhead o Protein expression o Protein purification o Protein characterization o Lyophilization o Ubc9 Mock inhibitor Studies o Inhibitor Kd Modeling Randall Mello o Protein expression o Protein Purification o Protein characterization o Protein expression optimization o Z Factor Studies for HTS Michelle Tran o Stability/ Oxidation Studies 44 of 53

o Team Website Design o Z Factor Studies Appendix VIII.Testing Results Spectrums for various tests: CYpet-Sumo1 Standard curve Standard curve for determining cypet sumo1 concentration RFU 3 25 2 15 1 5 y = 265.14x R² =.9936-5 5 1 Standard curve for determining cypet sumo1 concentration Linear (Standard curve for determining cypet sumo1 concentration) amount of cypet sumo1 in well in 6 ul (ng) Ypet-Ubc9 standard curve 45 of 53

Standard curve for determining Ypet UBC9 concentration 7 6 y = 632.8x R² =.984 5 RFU 4 3 2 1 Total amount of ypet ubc+'yu1 snstvty chk and prty check'!$h$184(ng)9 in 6 ul Linear (Total amount of ypet ubc+'yu1 snstvty chk and prty check'!$h$184(ng)9 in 6 ul) -2 2 6 1 Amount of Ypet Ubc9 in well (ng) Bradford assay for total protein. 46 of 53

1.4 y = 2E-1x 3-1E-6x 2 +.18x -.71 1.2 R² =.9991 Absorbance(RFU) 1.8.6.4.2 -.2 Absorbance vs Protein amount 5 1 15 2 25 Protein amount(ng) abs Poly. (abs) Purity effects on protein RFU Purity effects on Cypet-SUMO1 1ug 25 2 15 1 5 45 47 49 51 53 55 57 Wavelength(nm) 1% 9% 8% 7% 6% 5% 4% 3% 2% 1% 47 of 53

7 Ypet-UBC9 purity check 1ug RFU 6 5 4 3 2 1 49 51 53 55 57 59 61 Wavelength(nm) 1% 9% 8% 7% 6% 5% 4% 3% 2% 1% 14 Purity effects on Fret 5ng RFU 12 1 8 6 4 2 45 5 55 6 65 Wavlength(nm) 1% 9% 8% 7% 6% 5% 4% 3% 2% 1% Lyophilization effects on protein 48 of 53

Cypet-SUMO1 stability after 3 days 2 ug of protein Fluorescence 6 5 4 3 2 1 45 47 49 51 53 55 57 Wavelength(nm) non-lyophilized 4 deg resuspended RT Resuspended 4 deg Resuspended -8 deg resuspended -2 deg powder rt powder 4 deg powder -2 deg powder -8 deg blank Ypet-UBC9 stability after 3 days 2 ug of protein Fluorescence 18 16 14 12 1 8 6 4 2 49 51 53 55 57 59 61 Wavlength(nm) resuspended RT Resuspended 4 deg Resuspended -8 deg resuspended -2 deg powder rt powder 4 deg powder -2 deg powder -8 deg blank non-lyophilized 4 deg 49 of 53

FRET Lyophilization stability after 3 days 1ug per each protein Fluorescence 3 25 2 15 1 5 45 5 55 6 65 Wavelength(nm) yu csnormal 4 degree resuspended RT Resuspended 4 deg Resuspended -2 deg resuspended -8 deg powder rt powder 4 deg powder -2 deg powder -8 deg blank FRET Spectrum CS1 and YU9 16 14 12 1 8 6 4 2 Urea Effects on spectrum Effects of Urea on spectrum 45 47 49 51 53 55 57 Wavlength nm M urea.5 M urea 1 M urea 1.5 M urea 2 M urea 2.5 M urea 3 M urea Z factor determination 5 of 53

RFU 2 18 16 14 12 1 8 6 4 2 FRET Spectrum positive and negative hits at ratio of 1:2.8 [CS1]:[YU9] 45 47 49 51 53 55 57 Wavelength nanometers Negative hits Positive hits RFU 18 16 14 12 1 8 6 4 2 FRET spectrum fo determining z factor amount [CS1]=.25 [YU9]=.7 45 47 49 51 53 55 57 Wavelength nm Negative hits Positive hits 51 of 53

A) All figures were created in Mathematica Code used to make plot of Bound protein vs. inhibitor dissociation constant. xnot =.25 ynot= 2.8*xnot cnot = 5.25.7 5 xy = (xnot*ynot)/(.75(1+cnot/kc)).233333/(1+5/kc) Plot[xy,{kc,,1},PlotRange {,.25},Filling Bottom] kc=1 xy.233333 End 52 of 53