Postdoctoral research

Developing Methods to Analyze and Improve Control

Biological control systems enables cells to maintain homeostasis in the face of external fluctuations of conditions and internal fluctuations arising from reactions involving small copy-number players. We have developed a novel microfluidic platform to count proteins with extremely low abundances in live cells. These low-copy number proteins are the key players for fluctuations in gene-expression in bacteria. Cells have elaborate control circuits in order to suppress or to exploit the internal fluctuations arising from noisy gene-expression. We have developed a high-throughput time-lapse imaging technology that enables us to perform a systems-level study across the entire genome to analyze such control circuits. The same platform and a library of reporters allows us to characterize the performances of a library of synthetic control circuits and improve the designs.

Understanding Microbial Stress Response

Microbial stress response systems enable bacteria to exploit nutrient-booms and survive stresses during nutrient-busts. We are developing high-throughput time-lapse imaging technologies to investigate gene expression dynamics during stress responses in single bacterial cells. We are also developing cloning-free barcoding systems in order to track such dynamics across all the genes. The barcodes help us to study individual strains from a genomic library of transcription or translation reporters.

Counting Single-molecules in Bacteria

Gene-expression is noisy as the key control molecules are often present in extremely low abundances. More than 10% of proteins in bacteria are present in less than 10 copies per cell. In collaboration with Dr. Burak Okumus, I was involved in developing a microfluidic platform that enables us to count extremely low-copy proteins in live bacterial cells. Specifically, we have developed a PDMS based multilayer device for continually flowing cell culture into observation chambers. By actuating on-chip valves, the chamber ceiling can be lowered to immobilize cells on the coverslip, allowing us to image and replace cells every few seconds. We found that slightly compressing cells can slow down diffusion of proteins almost 100-fold, allowing us to detect individual fluorescent protein (FP) molecules with minimal perturbations until the moment of imaging. The slight flattening of cells ensures that all molecules are in focusand reduces the risk of spot overlap. Extensive validations show that this method accurately and reliably counts proteins present in low numbers per cell and has enabled us to study proteins that previously were invisible.

1. Okumus B, Landgraf D, Lai GC, Bakshi S, Arias-Castro JC, Yildiz S, Huh D, Fernandez-Lopez R, Peterson CN, Toprak E, Karoui ME, Paulsson J. Mechanical slowing-down of cytoplasmic diffusion allows in vivo counting of proteins in individual cells. Nature communications; 7:12130. Link
2. Okumus B, Baker CJ, Arias-Castro JC, Lai GC, Leoncini E, Bakshi S, Luro S, Landgraf D, Paulsson J. Single-cell microscopy of suspension cultures using a microfluidics-assisted cell screening platform. Nature Protocols; 13:170. Link

Spatio-temporal Organization of Bacteria Nucleoid (2014)

The structure and dynamics of bacterial chromosome are not well understood. Bacterial chromosomes do not fill the entire cytoplasm, instead occupy an irregular sub-region of the cytoaplasm called nucleoids. The overall spatial extent of bacterial nucleoids has been suggested to arise from a balance of nucleoid-expanding forces (e.g. transertion) and nucleoid-compacting forces (e.g. configurational entropy of the confined DNA polymer, macromolecular crowing, supercoiling etc.). By using time-lapse imaging of nucleoid dynamics in live cell, and correlating it with dynamic changes of spatial distribution of ribosomes, we have discovered a novel force that governs the spatial extent of nucleoids. This force relies on relative tendencies of ribosome and nucleoid to mix together in bacterial cytoplasm. In normal conditions, translating ribosomes (70S-polysomes) and the chromosomal DNA segregate, while 30S and 50S ribosomal subunits are able to penetrate the nucleoids. Growth conditions and drug treatments determine the partitioning of ribosomes into 70S-polysomes versus free 30S and 50S subunits. Entropic and excluded volume effects then dictate the resulting chromosome and ribosome spatial distributions. For example, transcription inhibition by rifampicin results in conversion of 70S-polysomes into free 30S and 50S subunits, due to loss of transcripts. As 30S and 50S can mix well with the nuceloid, transcription halt causes nucleoid expansion..

Bakshi S, Choi H, Mondal J, Weisshaar JC. Time-dependent effects of transcription-and translation-halting drugs on the spatial distributions of the Escherichia coli chromosome and ribosomes.

Molecular Microbiology; 94 (4): 871-887 (2014). Link

Time-lapse imaging of nucleoid (chromosomal DNA) dynamics in live cell requires a permeable, non-perturbative, long-lived fluorescent strain specific to the nucoeoids. The ideal stain would not affect the cell growth behavior, nucleoid morphology and dynamics, even under the imaging conditions. By using microfluidic devices and time-lapse imaging we have critically tested available "live-cell stains" and "dead-cell stains" for their utility. Ironically, we found that the "dead-cell stain" SYTOX Orange is nearly ideal for imaging nucleoid morphology in live E. coli cells, while the "live-cell stains" affect nucleoid morphology and cell growth. SYTOX Orange performed really well in both Gram negative E. coli and Gram positive B. subtilis. SYTOX Orange was used to investigate nucleoid dynamics in E. coli cells in the study mentioned above.

Bakshi S, Choi H, Rangarajan N, Barns KJ, Weisshaar JC. Non-perturbative Imaging of Nuceloid Morphology in Live Bacterial Cells during Antimicrobial Peptide Attack.

Applied and Environmental Microbiology; 00989-14 (2014). Link

Graduate research

Partitioning of RNA Polymerase Activity in Live E. coli (2013)

RNA Polymerase (RNAP) can be freely diffusing in the cytoplasm, be non-specifically bound to DNA, or transcribing a gene. Partitioning of the RNAPs among these states plays a key role in regulating the gene expression in bacteria. Using fluorescent protein labeled RNAP molecules, we have probed the diffusion of single RNAP molecules in live E. coli cells, to understand RNAP partitioning. A quantitative description of the partitioning of RNAP among its various roles in live E. coli provides an important test of global models of cellular processes. In particular, the concentration of free RNA polymerase, which to our knowledge has never before been measured directly, directly affects the expression level of most housekeeping genes and of rrn operons. The new data enable us to partition RNAP activity into the fraction transcribing (f_trxn, which includes initiation, elongation, pausing, and termination), the fraction nonspecifically bound to DNA (f_ns), the fraction able to bind DNA but freely diffusing in three dimensions (f_free), and a small fraction that evidently cannot bind to DNA on a 1-s timescale (f_nb). Our best estimates of these fractions disagree with earlier model results that relied on estimates of free RNAP from experiments on mini-cell preparations Studying this partitioning behavior of RNAP in different growth conditions can help us understand transcriptional control of gene expression in adaptive bacteria like E. coli.

Bakshi S, Dalrymple R, Li W, Choi H, Weisshaar JC. Partitioning of RNA Polymerase Activity in Live E. coli from Analysis of Single-molecule Di.usive Trajectories.

Biophysical Journal; 105 (12): 2676-2686 (2013). Link

Superresolution Imaging of Ribosome and RNA Polymerase in Live Bacteria (2012)

Spatial organization of the transcription and translation machinery (RNA Polymerase and ribosomes) dictates how genetic information is transferred from chromosomal DNA to the proteins. We have measured the quantitative spatial distributions of ribosomes and RNA polymerase (RNAP) in live E. coli by superresolution fluorescence microscopy. We found that ribosomes are strongly segregated from nucleoid, but RNAP localizes to the nucleoid lobes. Single molecule tracking of ribosome suggests that ~80% of the ribosomes are diffusing slowly. We have assigned this fraction of ribosome to the translating polysomes (70S). Most of these slowly diffusing ribosomes are distributed away from the nucleoid, in the ribosome-rich regions near end-cap and between the nucleoid lobes. This suggests that most of the translation activities are taking place in the ribosome-rich regions away from nucleoid. On the other hand, transcription activities are concentrated within the bulk nucleoid as RNAP distribution closely mimics the nucleoid distribution. The degree of DNA-ribosome segregation argues persuasively against co-transcriptional translation as the primary means of protein expression. It seems that most of the newly synthesized mRNA copies, decorated with ribosomes and other protective species, freely diffuse to the ribosome-rich regions where the bulk of translation occurs.

Bakshi S, Siryaporn A, Goulian M, Weisshaar JC. Superresolution Imaging of Ribosomes and RNA Polymerase in Live Escherichia coli Cells

Molecular Microbiology; 85(1): 21-38 (2012). Link

Tracking A Single Fluorescent Protein in Live Bacteria (2011)

It is important to segregate underlying biological phenomena of the protein of interest from the perturbations due to the fluorescent labeling strategies used to monitor the protein. To investigate the spatial biology of the fluorescent tag itself, we needed to examine the spatial distribution and dynamics of a fluorescent protein tag. We have used two-dimensional photoactivation localization microscopy (PALM) to study the spatial distribution and diffusion of the protein Kaede in the cytoplasm of live E. coli. We found that Kaede is distributed homogeneously inside the E. coli cytoplasm. Sub-diffraction limit tracking of single copies of diffusing Kaede molecules was used to understand its dynamics inside cytoplasm. Comparison of single molecule tracking studies and Monte Carlo simulations of particle diffusion in a confined space, shaped as E. coli cytoplasm, suggests that the dynamics of Kaede molecules is consistent with homogeneous diffusion. Homogeneous Brownian diffusion, and uniform spatial distribution of a fluorescent tag makes it a good candidate as a label, as the deviations from Brownian motion and/or uniform distribution of the fusion proteins can be associated with important biological properties of the protein of interest.

Bakshi S, Bratton BP., Weisshaar JC. Subdi.raction-limit Study of Kaede Diffusion and Spatial distribution in Live Escherichia coli

Biophysical Journal; 101(10):2535-2544 (2011). Link