Motivation: Antimicrobial peptides (AMPs) are promising alternatives to conventional antibiotics; however, bacterial resistance mechanisms can limit their effectiveness. In Bacillus subtilis, membrane modification pathways such as dlt and mprF alter the cell surface charge, reducing AMP binding and promoting resistance. Conventional bulk assays often mask cell-to-cell heterogeneity, making it difficult to identify transient or rare resistance phenotypes. This project aims to establish a single-cell experimental platform to investigate the molecular mechanisms underlying AMP resistance and quantify phenotypic heterogeneity at single-cell resolution.

Work: This interdisciplinary project integrates genetic engineering, microfluidics, fluorescence microscopy, and image analysis to investigate bacterial resistance at the single-cell level. Gene knockout and complementation strains were generated to study membrane modification pathways involved in AMP resistance. A mother-machine microfluidic platform was developed through optimization of PDMS fabrication, plasma bonding, valve-recirculator-controlled media switching, and bubble-free fluid handling to enable long-term live-cell imaging under dynamic antimicrobial exposure. Automated fluorescence microscopy is used to quantify growth dynamics, membrane surface charge, and antimicrobial susceptibility, while a Python-based image analysis pipeline using DeLTA was developed to perform single-cell segmentation, tracking, growth curve extraction, and fluorescence quantification. The resulting genotype–phenotype relationships will provide mechanistic insights into bacterial resistance and guide the development of future antimicrobial strategies.

Current Status: Ongoing DPhil research, University of Oxford.

Motivation: Membrane proteins mediate essential cellular processes and constitute the targets of more than half of all therapeutic drugs. However, their structural characterization remains challenging because they require membrane-mimetic environments for stability, while conventional detergents used during purification often disrupt native protein interactions. Peptidiscs are detergent-free membrane mimetics that stabilize membrane proteins in solution and have emerged as a promising alternative for structural studies. This project aimed to establish and validate peptidisc-based reconstitution workflows compatible with native mass spectrometry (native MS) for analysing membrane protein complexes in their native-like state.

Work: The project focused on optimizing membrane protein purification, detergent removal, and peptidisc reconstitution using both on-column and on-bead assembly strategies. Membrane proteins, including the multidrug efflux transporter AceI and the β-barrel assembly machinery (BAM) complex, were reconstituted into peptidiscs and analysed by native mass spectrometry to evaluate complex integrity, oligomeric state, and stability. Different reconstitution workflows were systematically compared to identify conditions that preserved native protein assemblies while maintaining compatibility with native MS, establishing a robust platform for structural characterization of membrane protein complexes in a native-like lipid environment.

Publication: Deedwania A, Wang Y, Robinson CV, Bolla JR. Native mass spectrometry of membrane proteins reconstituted in peptidiscs. RSC Chemical Biology. 2025;7(1):120–128. doi:10.1039/d5cb00236b. Link to the paper

Motivation: Directed evolution relies on the generation of diverse mutant protein libraries to identify variants with improved or novel functions. Conventional random mutagenesis methods often produce a large proportion of non-functional variants, while existing targeted mutagenesis approaches can be laborious and difficult to scale for multiple mutation sites. This project aimed to establish an efficient and robust methodology for constructing large, high-quality protein libraries using SLUPT (Synthesis of Libraries via deoxyuridine-containing PCR Templates), enabling systematic exploration of protein sequence-function relationships.

Work: The project focused on optimizing the complete SLUPT workflow, including single-stranded DNA template preparation, targeted multi-site mutagenesis, cloning, and library amplification. Protocol modifications were introduced to improve ssDNA recovery, PCR amplification efficiency, and cloning fidelity, resulting in a robust methodology for generating large mutant libraries. Green Fluorescent Protein (GFP) was used as a model system to validate the approach by introducing targeted mutations within key fluorescence-determining residues. The resulting mutant library was screened using high-throughput fluorescence spectroscopy to identify variants with altered excitation and emission spectra, demonstrating the effectiveness of the workflow for protein engineering and future directed evolution studies.

Report: Creating Large Protein Libraries: An Efficient Engineering Approach, Second Rotation Report, University of Oxford (2024).

Motivation: Temperature-sensitive (TS) plasmids are those plasmids that can replicate till a permissible temperature (say 30°C), but they lose their ability to replicate at higher temperature (say 35°C), and thus, the plasmid is lost by the cell in successive replication cycles. TS plasmids are important in synthetic biology as they give the ability to control the burden on any cell by altering the plasmid replication as a function of temperature. They can be used in genetic engineering and manipulation as delivery vectors, for inducible replication by supplying protein for transient expression and various other features. Once the plasmid has done the work and has served the purpose, it becomes a burden to the cell, and thus, to reduce the burden, it is important to make the cell plasmid free. The TS plasmid can be cured by raising the temperature.

Work: The goal of this project was to design and construct a temperature-sensitive plasmid that would later contain genes for recombineering (genome engineering). The design of this TS plasmid involved protein engineering steps: predicting the TS mutation, predicting the 3D structure of the replication protein, performing docking with the origin of replication and then proving these predictions by experiments: cloning and screening for the TS phenotype.

Publication: Deedwania A, Karmakar S, Kumar V, Shefrin S, Sundar D, Srivastava P. Construction and characterization of a temperature-sensitive pRC4 replicon for Rhodococcus and Gordonia. Gene. 2024 Feb 20;896:147990. doi: 10.1016/j.gene.2023.147990. Epub 2023 Nov 15. PMID: 37977321. Link to the paper

Motivation: Cell fusion is an important process for various cellular processes including cell-cell communication, drug delivery, cell transfection and other therapeutic processes. For synthetic cells, designing and controlling spatial and temporal biological reactions have been a challenging process. Targeted vesicle fusion has been shown as a promising approach to control interactions between compartments selectively. Membrane fusion is also important to initiate protein synthesis in synthetic cells. Thus, modeling and designing a biomolecular circuit is an important step.

Work: I designed, modeled and simulated a biomolecular subsystem that would export single-stranded DNA (ssDNA) encapsulated inside a synthetic cell (vesicle) and aid the linker-induced fusion of two DNA-tethered synthetic cells that are externally embedded with the complementary strand of the exported ssDNA. Upon the formation of double-stranded DNA, the two cells would fuse. The CRN models were made using BioCRNpyler, compartmentalization was introduced through subSBML and simulations were done using Bioscrape.
An example of the designed subsystem is- The vesicle is encapsulated with cell extract, linker ssDNA and plasmid that can express aHL protein. The input signal, aTc in the external environment will induce the expression of the gene, producing aHL. The protein binds to the membrane and forms a channel through which ssDNA can be exported. This exported ssDNA acts as linker which helps in the controlled fusion of vesicles. This controlled fusion can help us initiate protein expression in other vesicles. For a simple model to initiate protein expression, the liposome population 1 is encapsulated with gene expression of exsA, induced in presence of aTc (present in the external environment). Liposome population 2 is encapsulated with a gene which expresses GFP, which is activated only in presence of exsA, so there is no protein in this liposome as exsA is being expressed in liposome population 1. These are DNA-tethered liposomes. When the exported linker DNA reacts with its complementary strand, it initiates fusion of two populations of liposome, thus exsA can induce the expression of GFP in the fused cell.

The code and models are available at my GitHub repository.

Motivation: RNA Thermometers (RNATs) are temperature-sensitive RNA molecules that can control protein expression. RNATs alter their secondary structure in response to temperature fluctuations. They are present in the 5’ UTR region of mRNA sequence. As temperature increases, the hydrogen bonds break/ melt and thus, expose the regions like ribosome binding sites that affect translation, and permit translation. This structural transition can depend on either the heat shock or cold shock response resulting in either exposing or hiding of RBS. Thus, it was important to understand the molecular mechanism or the structural transition that takes place by temperature change in some naturally occurring RNATs.

Work: I studied the temperature dependence of protein expression of naturally occurring RNATs like rpoH, cspA and agsA. The RNATs sequence was also used to analyze the expression and melt profile via the software tool NUPACK. I also constructed a library of synthetic RNA thermometers expected to give a better melt and expression profile. A single base change was made in the natural RNATs, and the sequence was again analyzed; the ones giving better melt profile were synthesized and tested. .

Project poster is available at this link.

Motivation: Lead poisoning is a catastrophe affecting millions of people around the world. Having a biosensor that could sense lead and also recover it from water could solve this major problem. Previously developed whole-cell lead biosensors cater to a limited sample type i.e. soil samples. Only a few studies have expanded the sample type to wastewater samples, which is one of the main focuses of our biosensor. Whole-cell lead biosensors are intensity-based. Also, industries use various physicochemical methods like chemical precipitation etc for lead removal from industrial wastewater, which brings down the lead concentration from 200-500 mg/L to 0.4 mg/L, but is still ten-fold more than the permissible limit and resulting in poor recovery efficiency beyond this concentration.

Work: We made a system that would assess and monitor lead contamination, and also adsorb lead from water samples and recover it for viable commercial use. We intend to develop a frequency-based biosensor that provides many advantages over intensity-based biosensors like robustness from environmental factors, and also developed a new recovery system for the same using cell surface adsorbtion. A lead binding protein would be expressed on the cell surface that would capture lead from the solution, thus recovering lead even at lower concentrations.

Role: Advisor and mentor to the iGEM IIT Delhi team. Guided in ideation of the project and mentored, taught undergraduate students in the experimental lab and modeling techniques. .

Won the gold medal at the iGEM 2022 competition. The work is available at the iGEM wiki.