Team:IISER Berhampur/Engineering



Our proposed reporter system i.e. FRaPPe, is an engineered syn-bio device that can be used to accelerate drug development for various diseases. It is built by combining three different biological modules, a CID (Chemically Induced Dimerization) module, a FRET (Fluorescence Resonance Energy Transfer) module and a POI (Protein of Interest) module. Using the first two modules, we can deduce the efficiency of any peptide-based drug in inhibiting the interactions between the proteins present in the POI module in in-vitro conditions. Being an engineered device, it's building, testing and improving process is based on several rounds of the engineering design cycle, but due to the constraints placed by the COVID-19 pandemic, we could only proceed to a small extent in the actual implementation cycle. Most of our design cycle has been based on suggestions from experts and our own molecular modelling and simulations, which is described below. However, we will be describing all that we have achieved in these terms and hope to achieve in the future in the following sections.



To understand the predicament faced by humanity in developing effective therapeutics against deadly diseases, one need not look further than the COVID-19 pandemic waging around us right now. The existing notion is that it takes at least 10 years for a drug to be discovered and commercialised. The bitter truth : the actual journey might even take three times this time period.

We first assessed the current drug development scenario.

A new discovery takes several years to be refined and matured into a viable product. For viruses, in particular, the problems are many-fold. The arsenal is much more limited when compared to other infectious agents such as bacteria. Moreover, since viruses are technically not alive, they hijack the cells of our own body to feed their replication machinery. Any drug that targets a part of its life cycle, could end up harming the host in the process. And then comes the issue of diversity among viruses. Investing a huge deal of money and resources to study every unique virus stretches out the time needed immensely.

Figure 1: Drug Discovery and Development : A Timeline. Information sourced from U.S. Food and Drug Administration.

Specifically for Dengue, our disease of interest, the existence of different serotypes makes it very difficult for researchers to come up with a pan-serotype drug. Currently there are no clinically approved drugs against dengue. Until now most drugs that went into trial were repurposed from treatments of other diseases. For example Balapiravir which was initially developed as a drug against hepatitis C virus was entered into clinical trials but failed to show any significant effect when compared to placebos. (Shi et al., 2014)


We want to solve the Dengue scourge. The first order of business will therefore be to find what is so special about this virus that can give us an edge over conventional supportive therapies.

As it turns out, Dengue is one of the masters of an intricate phenomenon named Protein-protein interactions (PPI). In this, two proteins in close proximity come in contact with each other in a certain orientation, forming a complex (either transient or permanent) in order to perform a specific function. In a cell, interactions between the various proteins occur at many different levels and bring about numerous cellular processes and pathways. As such PPIs form a very integral process in living cells.

Especially during a viral infection, the virus can only survive and multiply in the host system by making use of numerous protein-protein interactions between its proteins and the proteins of the host cell. These host-virus PPI’s serve to perform various functions from facilitating virus entry into the cell, to disabling host immune response pathways and spreading through tissues. When talking specifically about Flaviviruses like DENV, which are RNA viruses, PPIs are employed by the virus on every step of its pathogenesis. For example, right from the point when the viruses are being secreted from the infected cells, the mature virus is composed of multiple copies of core and capsid proteins, two envelope proteins and a virus-specific set of host proteins that enables it to be properly secreted so that it can then enter and thereby infect other cells. During entry into the cells, the primary protein interactions at the cell surface either serve to concentrate virus particles or trigger the productive virus uptake program. The host factors that flaviviruses generally tend to interact with are - heparin sulfates, Dendritic Cell-Specific Intracellular Adhesion Molecule (ICAM)-3-Grabbing Non-Integrin 1 (DC-SIGN, CD209 antigen), and DC-SIGNR, which interact with N-linked glycans of the viral E glycoprotein. Also, after the cleavage of the polyproteins and formation of functional protein subunits, these proteins induce the rearrangement of the ER membrane to form membranous compartments to facilitate viral replication and assembly. Here too, host-viral PPIs are made use of to form a replication complex, stabilize it and then disable the immune response of the host. Specific examples of these interactions include - interactions between DENV NS5 and GBF1 which induces ER membrane remodelling and replication complex formation early in the DENV infection state and also those between NS4B and 12 mitochondrial proteins that induce changes of the mitochondrial architecture which is necessary to prevent viral RNA recognition by the innate immune system. Sometimes the structural proteins to help in affecting immune responses other than just facilitating entry. For example, the capsid proteins of DENV and WNV have been found to interact with host peroxisome biogenesis factor- Pex19. This leads to loss of peroxisomes in the cell which diminishes early antiviral signalling and formation of interferons. (Gerold et al., 2017)

There are many other such host-viral PPIs that are absolutely essential for entry and replication of viral particles in the cell. Thus, inhibition of host-virus PPI is a therapeutic technique that holds potential in order to develop an effective antiviral drug against Dengue and all other viral diseases where this is a common theme.


We also compared the different options available among the classes of molecules widely used as drugs before zeroing in on our optimal molecule.

Figure 2: Comparison of the existing inhibitor drug molecule classes. Bruzzoni-Giovanelli et al. 2018 ; Lau et al., 2018 ; Lee et al., 2019 ; Sorolla et al., 2019

We can see that peptides are the preferred choice for our purpose. However, they have some disadvantages too. Peptides are generally unstable and can be degraded easily, but several studies done in the last few decades paved the way to get rid of these problems through methods like termini protection and chemical modifications. Also, there are recent advances in techniques to overcome several other intrinsic disadvantages of peptide drugs such as their solubility, low membrane permeability, fast elimination and bioavailability. That is the reason why, over the past two decades, nearly 60 peptide drugs have been approved worldwide. iPEPs as biotherapeutics are therefore receiving increasing attention.

Figure 3: Cumulative number of peptides approved in major pharmaceutical markets and the number of peptides entering clinical development. Entry into clinical development is defined as the year of the first Phase 1 or pilot human study (Lau & Dunn, 2018).



One of the important mechanisms of dengue viral pathogenesis manifestation is through myriad protein-protein interactions of viral Non-structural (NS) and host proteins. Antiviral approaches explored thus far have targeted both structural and nonstructural proteins of DENV. The structural proteins are the molecules that the virus particle comes equipped with, such as the proteins comprising their capsids or coats. The non-structural proteins are encoded by the viral genome and expressed in the infected host cells. These are mostly important players in replication and translation processes. We propose to design biomolecules which targets NS proteins specifically and disrupts all their pathogenesis pathways. (Visit our modelling section to know why we chose NS proteins using mutational analysis.)

The Dengue virus boasts a repertoire of seven non-structural proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B, NS5) coded for by its 10.7 kilobase capped positive sense, single-stranded RNA genome. These are all formed as a single polyprotein, and then chopped up into functional parts by the host and viral enzymes. They are the workhorses for viral replication and can also temper host immune responses in clever ways. So which one to choose? For this, first, we did an extensive literature survey to find out various PPI associated with each of the 7 DENV NSPs and collected structural and functional data about these interactions. From these we formed representative interactomes for each DENV NSP and finally, based on several criteria such as the amount of structural data available for each PPI and importance of the function each PPI plays in DENV Pathogenesis, we finally narrowed it down to the interaction between DENV NS5 and hSTAT2. (Read more about this in our Modelling section.)


The relevance of this PPI was based on its involvement in the interferon (IFN) signalling pathway, a primary component of the innate antiviral response in mammals. To understand this, we did another round of research. The type 1 interferon (IFN) pathway acts to delay virus replication and to stimulate the activation of antiviral effector cells. The IFN receptor functions through the Janus kinase-signal transducer and activator of transcription (JAK-STAT) pathway in which JAKs phosphorylate STATs, which then translocate to the nucleus to induce the expression of IFN-stimulated genes (ISGs). Type I IFN signalling activates phosphorylated STAT1 and STAT2 to heterodimerize and associate with interferon regulatory factor-9 (IRF9) to form IFN-stimulated gene factor-3 (ISGF3), which specifically binds IFN-stimulated response elements (ISREs) within antiviral gene promoters. Over the past decade, flavivirus NS5 proteins have surfaced as potent antagonists of IFN signalling.

DENV proteins NS2A, NS4A, and NS4B are able to antagonize IFN signalling. NS4B is highly potent and prevents STAT1 phosphorylation. DENV NS5, on the other hand, binds to STAT2 and inhibits IFN dependent signalling, by targeting STAT2 for proteasome-mediated degradation only when it is expressed as a proteolytically processed precursor. More importantly, The loss of STAT2 expression is conserved across the DENV serotypes and it occurs independently of the structural proteins.

Figure 4: Diagrammatic Representation of the JAK-STAT pathway involved in the IFN response and inhibition of components by non-structural proteins of DENV. The inset shows the ubiquitin-mediated proteasomal degradation of STAT2 initiated by NS-5.

Further evidence to consolidate the role of STAT2 as an important player in IFN responses come from studies by Perry et al., 2011 which showed that STAT2 mediates STAT1-independent protection against DENV infection in mice. The combined loss of STAT1 and STAT2 results in high viral burden in tissues. It has been demonstrated by the same research group that mice lacking STAT1, a key mediator of both type I and II IFN responses, are not susceptible to DENV-mediated disease. So, what is the mechanism responsible for protection against DENV disease in the absence of STAT1? STAT2 seems to be the protein that mediates type I IFN signalling during DENV infection in the absence of STAT1. The resulting antiviral response includes amplification of type I IFN and the expression of interferon-stimulated genes (ISG) by associating with their promoter elements and leading to the antiviral response.


Hence, our aim now, is to study the Protein-protein interaction between the NS-5 viral protein and hSTAT2 host protein. How to do this?

We thought of using inhibitory peptides. And, gradually the idea formed of developing a composite reporter that can kill two birds with one stone: it can help the biologist study the fundamental molecular interaction of the proteins and can aid the chemist design an appropriate drug and screen its efficiency as well.


1. Chemically Induced Dimerization

In order to study our target interaction, we would need a viral infected cell. This is where Synthetic Biology came to our rescue. Instead of actual viral infection (which would require specialized laboratory set up and additional safety protocols), we decided to assemble the viral protein and the host protein in plasmid constructs. FRaPPe this helps us achieve the replication state in mammalian cells without needing patient samples.

Module 1 : CID

Purpose : Bring the two proteins in proximity and keep control on their interaction.

Challenge : The proximity and the strength of the interaction between the proteins may be difficult to control or modulate.

Solution : We decided to use our first weapon, which is a chemical biology technique called, Chemically Induced Dimerization, wherein two particular proteins which can be induced to dimerize with one another, are fused with our selected proteins and expressed. In this way, our chosen proteins can be made to interact by bringing them in close proximity. Thus, we can replicate the viral infection with respect to host-viral protein interaction in a relatively safe setting.

In our system we aim to use FKBP12 and FRB protein domains and Rapamycin, that will induce their dimerization (Chemical Inducer of Dimerization). With the CID Reporter, we can ensure that our chosen PPI takes place in the cell and we can localize and observe it more closely.

Figure 5: The binding affinities of the different components in the FKBP-Rapamycin-FRB system. Binding affinities are greatly increased in the presence of Rapamycin and a ternary complex is formed via two possible routes. Kd values taken from Banaszynski et al., 2005

2. Fluorescence or Förster Resonance Energy Transfer

Module 2: FRET

Purpose : Detect PPI and offer a quantifiable measure of strength of interaction in presence or absence of Peptide Inhibitor.

Challenge : To quantify the extent of the interactions, we need a second missile, a readout of the interactions.

Solution : We plan on using a FRET Reporter as an assay system to understand the interaction. From the literature survey, it was clear that PPI followed by Immuno pull down and western blotting will be time consuming. And if we wanted an immediate readout, fluorescent tagged proteins could do the trick. FRET (Forster’s or Fluorescence Resonance Energy Transfer) is a phenomenon that essentially functions as a spectroscopic ruler, by employing two fluorophores (donor and acceptor) wherein the excitation energy of the donor is transmitted to the acceptor in its vicinity, resulting in the distinctive fluorescence emission spectra of the acceptor.

Such biosensor systems are highly sensitive to the distance of separation between the interacting molecules (within 1 to 10nm range). Due to this property, FRET is used to quantify many biological processes that are dependent on molecular proximity such as Protein-Protein Interactions. FRET is more advantageous than other Reporter systems because it allows observation of live cells in a non-destructive and minimally invasive way. Efficiency can be used as a measure of quantification.

The FRET Efficiency is basically the percentage of energy transfer from the donor to acceptor fluorophores at a given state and depends on the distance between donor and acceptor dipoles, while the FRET Dynamic Range refers to the range at which the reporter operates and is essential for detection of highly sensitive cellular events.


In our project, we are going to use a FRET-CID coupled system where, after the CID reporter ensures the PPI, the FRET reporter will measure the extent of interaction and also act as a measure for the peptide inhibitor efficiency by showing a change in the acceptor fluorophore’s emission spectra. The CFP will be our donor fluorophore and YFP, our acceptor fluorophore. The measurable parameters generated from the FRET system, namely the efficiency and dynamic range will be our tools in gauging the interaction and its subsequent inhibition.

Figure 6: Schematic diagram of the building blocks of FRaPPe. Modules used include CID, FRET and POI.



To assemble the modules into our plasmid constructs we designed several possible combinations. The molecular docking experiments were the starting point. Using the visualization of the structure based on the experimental data, we chalked out a rough design for our plasmids. Since the experimental data was limited, and did not resemble the entire length of the proteins, we used the reconstructed structure using homology modelling to obtain a clearer picture. This gave us insights about selecting appropriate N or C terminals to position our FRET fluorophores. (To read more about this, visit our Modelling section).


Using these inputs we narrowed down our constructs from an initial 36 to 10. These are shown in the constructs table below along with their respective codes.



  • For our first set of constructs, we will insert our genes of interest in a suitable CID plasmid backbone, so as to create fusion constructs, FRaPPE C5 and FRaPPe C6 (See construct table).

  • The interactions between any DENV NS and host protein or between two DENV NS proteins can be studied by ligating their coding sequences with the CID Reporter.

  • The plasmid will be cloned in E. coli and the resulting reporter system will be expressed in HEK-293 cell lines. We have chosen HEK293 cells due to their ease of transfection and maintenance in culture. They have a rapid doubling time (36 hours) and can be reliably used for both stable and transient transfection (we will be doing the latter). Thus, compared to more cancerous cell lines like HeLa, HEK cells are the more feasible choice for studying our reporter in a mammalian system.

  • In addition to the previous CID reporter, the DENV NS protein and its host target will be ligated with CFP and YFP respectively. Test construct series FRaPPE T1-T4 will be generated to account for different orientations. We also propose the control constructs FRaPPE C1-C4 to take into account effects of the CID alone and the interaction between our proteins of interest in the absence of CID. (See Construct Table)

  • These plasmids will also be cloned in E. coli and the whole system, in turn, will be transfected into HEK 293 cell line, where it can be used as the control for the assay platform for screening of small molecules/activators/peptide inhibitors.



Since we did not have access to lab facilities during the lockdown, we were not able to carry out our wet lab completely. We have started our preliminary experiments recently. (Refer to Experiments Section). Nevertheless, we did plan how we would be conducting the experiments if the opportunity came up.

  • We ordered CID-FRET backbone plasmids from Addgene [Plasmid #20160, CFP-FKBP (CF) and Plasmid #20148, YFP-FRB (YR)], which contain the coding sequences of CFP and YFP fluorophores that we have decided to use as our FRET pair, in addition to coding sequences for FKBP and FRB (FKBP-Rapamycin Binding Domain) under the control of a CMV promoter and enhancer. It used Neomycin/Kanamycin resistance as a marker. Coding sequences of POIs were ordered from Twist Bioscience.

  • Based on the Restriction Enzyme sites available in the plasmids at the appropriate locations for inserting our POI modules, we chose those which were not present within the coding sequences of NS-5 and hSTAT2 to avoid fragmentation during cloning. We modified the gene sequences for our Proteins of Interest as follows:

  • For NS-5 (GenBank : KR919821.1, DENV TSV08 Serotype 1) . To be cloned into Addgene plasmid # 20148 using HindIII at 5’ and SalI at 3’

  • For hSTAT2 (GenBank : U18671.1) . To be cloned into Addgene plasmid # 20160 using XhoI at 5’ and EcoRI at 3’

  • Plasmid maps were prepared accordingly. (See Parts Section)

  • Since the coding sequences of the proteins are quite large, we were advised by Twist Bioscience to use Splicing by Overlap Extension to get them synthesized, hence we modified the strategy accordingly.


  • Once all our reagents and components arrive, we will first be performing the above mentioned restriction digests for the corresponding plasmids as per our constructs table. This will be followed by ligation to create our test constructs. For our control constructs, similar procedures will be used. In C3 and C4, the modules for CID will be excised out by appropriate restriction digests and the gene blocks inserted instead. For C5 and C6, the fluorophore sequences will be excised and replaced by gene blocks.

  • Given that the YR plasmid has an MCS between the FRB and YFP sequences, construct T2 will be made, and for the CF plasmid, insertion between FKBP and CFP may not be reliable due to absence of an MCS, hence construct T3 will be made.


  • Transformation of the plasmid into competent DH5α E. coli cells will follow and positive recombinants for FRaPPe constructs will be confirmed by colony PCR or restriction digestion and sequencing.

  • Confirmed plasmids will be amplified by Maxiprep to prepare DNA for transfection into HEK293 cells. For test cell lines, constructs T2 and T3 will be co-transfected, and for control cell lines, constructs C1, C2 ; C3, C4 ; C5, C6 will be pairwise co-transfected in different cultures. Presence of reporter will be screened by qPCR.

  • Cells will then be used for a dose dependent Rapamycin assay and corresponding FRET measurement to choose an optimum Rapa concentration.

  • After procuring the chemically synthesized peptide inhibitor, the FRaPPe T2 and T3 co-transfected cells will be treated with the inhibitor and the efficiency of each peptide inhibitor in inhibiting the PPI will be quantified. The various approaches for this quantification are discussed in the section below.

Figure 7: Overview of the Reporter System showing the overall test procedure for experimentation and characterization of its working.


For the quantification of peptide inhibitor efficiency, the following 2 approaches have been suggested-

  • Approach 1 - This approach focuses on the quantification of the inhibition based on the FRET readout with respect to different concentration of peptide inhibitors

  • Approach 2 - This methodology focuses on controlling the level of Rapamycin for ranking of the peptide inhibitors.

Approach 1

  • Step 1 : Initially FRaPPe is kept at a fixed concentration of Rapamycin ( Assumed to be x) .
  • Step 2 : Now we titrate the cell culture containing our reporter system with the peptide inhibitor and note the FRET emission after each addition. From here we obtain a plot of FRET emission versus the log(concentration) of the peptide inhibitor.
  • We can then obtain the concentration of the peptide inhibitor for which the FRET emission gets reduced by 50% of the initial value (without iPEP) which is the IC50 value. IC50 values for each inhibitor can then be compared and ranked.

Figure 8: Representative diagram for obtaining IC50 values

Approach 2

Since the reporter system consists of a CID system in which the interactions between FKBP-FRB are mostly regulated by the addition of Rapamycin, it can act as a good indicator to find out the efficiency of the peptide inhibitors and rank them accordingly.

  • Step 1 : The basic principle behind this method is that with increasing concentration of Rapamycin, we can increase the amount of FKBP-FRB dimerization that takes place and hence increase the FRET efficiency values readout. Initially, with some amount of rapamycin (assumed to be x) we will keep the FRET efficiency at a certain fixed value.
  • Step 2 : A certain concentration ( assume it to be z) of peptide inhibitor is added to the system. Now the concentration of the Ramamycin is increased until the FRET efficiency value reaches back to our fixed value .

Since with the addition of peptide inhibitors the interactions between the proteins will be inhibited, to reach the initial FRET efficiency value, more amount of Rapamycin will be required. Let it be x+y. Then the value of y gives us a proxy of the efficiency of the peptide inhibitors. A higher value of y will indicate that the peptide is a better inhibitor of the PPI.


We have tried to analyse the points where we can encounter possible challenges during our reporter creation.


Potential challenge: Since there is a co-transfection, the expression level of the two proteins may show differences.

Solution: The expression and localisation pattern of the proteins in the engineered cells will be first understood using the control constructs.


Potential challenge: Transfection, our method to introduce the plasmids into HEK cells, may generate heterogeneity as not all cells will take up the plasmids uniformly.

Solution: We plan to use electroporation after preliminary optimization with the transfection.


Based on the anticipated challenges that we may face, some possible improvements that we can introduce have also been brainstormed. Many of these inputs were provided to us in different interviews with researchers working in CID, FRET and related fields


Potential challenge: Transient transfection may not provide reliable results while working with a CID system. Control over expression.

Solution: Lentiviral based transduction approach for generating stable cell lines and also for introduction of a single construct with both the POIs separated by linkers.

Strategy: See construct info of FRaPPe T1 and T2, the gene hSTAT2 and NS5 have been synthesized with additional REs AgeI and MluI to facilitate easy excision and introduction to lentiviral vector (addgene 17446, pLenti CMV GFP Hygro).


Potential challenge: Use of CID for the inhibitor screening system.

The FKBP-FRB based CID system is considered to be a very strong interaction system. (See Figure 4 under design tab). The Kd values reflect an irreversible interaction, and hence it is possible that they can affect the interaction between the POIs (in terms of peptide efficiency determination).

To elaborate, based on FRaPPe constructs T2 and T3, the FRET efficiency readout will actually be a function of two interactions:

(i) FKBP-FRB interaction

(ii) NS5-STAT2 interaction

But suppose, the interaction between FKBP-FRB dominates strongly over the NS5-STAT2, then the FRET readouts may not decrease even if the peptide inhibitors are working very efficiently in inhibiting interaction between NS5 and STAT2.

Solution: Use of control constructs to identify the relevant interaction and corresponding inhibition.


  • The control constructs FRaPPe C1, C2, C3, C4 will help in solving the above problem.
  • Additionally, rapamycin concentration can be titrated to modulate the FKBP-FRB interaction.
  • In FRaPPe T1 and T2 CFP and YFP are attached directly to NS5 and hSTAT2 and not FKBP and FRB, thus may not be influenced by the strength of FKBP-FRB interaction alone.
  • An alternative reversible CID system which can bring the target proteins into proximity but will not affect their interactions significantly. While such systems are not very widely available, research has revealed some systems that can suit this purpose, such as that using a synthetic ligand of FKBP (SLF’). SLF’ binds to FKBP and is also capable of binding E.coli dihydrofolate reductase (eDHFR) via its TMP (Trimethoprim) moiety. The rapid SLF’-TMP induced heterodimerization of FKBP and eDHFR can be easily reversed by using TMP itself as a competitive inhibitor. (Liu et al., 2014)

However, our Molecular Dynamics Simulation (visit our Modelling section) shows that the interaction of NS5 and hSTAT2 is relatively weak, thus strengthening the need for a regulated mechanism of PPI. The use of CID should circumvent this problem and hence we are making an informed choice of using the CID system for creating FRaPPe.


1. Banaszynski et al. (2005) Journal of the American Chemical Society, 127(13), 4715-4721.

2. Bruzzoni-Giovanelli et al. Drug Discovery Today 23.2 (2018): 272-285.

3. Gerold, G. et al. (2017) Molecular & cellular proteomics : MCP, 16(4 suppl 1), S75–S91.

4. Gu F, Shi Y (2014). Clin. Invest. (Lond.) (2014) 4(8), 683–685.

5.Lau et al. Bioorganic & medicinal chemistry 26.10 (2018): 2700-2707.

6. Lee et al. International journal of molecular sciences 20.10 (2019): 2383.

7. Liu et al. (2014) Angewandte Chemie International Edition, 53(38), 10049-10055.

8. Sorolla et al. Oncogene (2019): 1-18.

©iGEM IISER Berhampur