Logo: to the web site of Uppsala University

Drug-resistant tuberculosis (DR-TB) kills similar to 200,000 people every year. A contributing factor is the slow turnaround time (TAT) associated with drug susceptibility diagnostics. The prevailing gold standard for phenotypic drug susceptibility testing (pDST) takes at least two weeks. Here we show that growth-based pDST for slow-growing mycobacteria can be conducted in 12 h. We use Mycobacterium tuberculosis variant bovis Bacillus Calmette-Guerin (BCG) and Mycobacterium smegmatis as the mycobacterial pathogen models and expose them to antibiotics used in (multidrug-resistant) tuberculosis (TB) treatment regimens - i.e., rifampicin (RIF), isoniazid (INH), ethambutol (EMB), linezolid (LZD), streptomycin (STR), bedaquiline (BDQ), and levofloxacin (LFX). The bacterial growth in a microfluidic chip is tracked by time-lapse phase-contrast microscopy. A deep neural network-based segmentation algorithm is used to quantify the growth rate and to determine how the strains responded to drug treatments. Most importantly, a panel of susceptible and resistant M. bovis BCG are tested at critical concentrations for INH, RIF, STR, and LFX. The susceptible strains could be identified in less than 12 h. These findings are comparable to what we expect for pathogenic M. tuberculosis as they share 99.96% genetic identity.

The intracellular position of genes may impact their expression, but it has not been possible to accurately measure the 3D position of chromosomal loci. In 2D, loci can be tracked using arrays of DNA-binding sites for transcription factors (TFs) fused with fluorescent proteins. However, the same 2D data can result from different 3D trajectories. Here, we have developed a deep learning method for super-resolved astigmatism-based 3D localization of chromosomal loci in live E. coli cells which enables a precision better than 61 nm at a signal-to-background ratio of ~4 on a heterogeneous cell background. Determining the spatial localization of chromosomal loci, we find that some loci are at the periphery of the nucleoid for large parts of the cell cycle. Analyses of individual trajectories reveal that these loci are subdiffusive both longitudinally (x) and radially (r), but that individual loci explore the full radial width on a minute time scale.

Most of the biological sciences deal with understanding the relationships between phenotypes and the underlying molecular mechanisms of organisms. This thesis is an engineering, computational, and experimental exercise in expanding the scope and scale of phenotype-genotype mapping techniques in single-cell microbiology using microscopy, microfluidics, and image processing. To this end, we use mother-machine-based microfluidic devices together with recently developed techniques in deep learning and optics. We use optical microscopes to observe cells of different genotypes, physically move cells, and image molecules inside them.

We have designed a novel microfluidic device to expand the throughput of single-cell lineage tracing an order of magnitude compared to existing methods. We demonstrate the ability to isolate single cells from such a device using optical tweezers after phenotypic characterization in real time. We have developed analysis algorithms of various kinds with the prime intention of performing high-throughput real-time image processing in conjunction with experimental runs to identify interesting cells for further investigation.

We have also developed an experimental protocol for bacterial species identification using fluorescence-in-situ hybridization (FISH) in microfluidic chips to complement an existing phenotype-based antibiotic-susceptibility test (AST). We apply this method together with deep-learning-based cell segmentation and tracking algorithms, and image classification methods to perform species-ID of up to 10 species in 2-3 hrs.

Lastly, we have developed a 3D dot localization method to investigate how the chromosome structure changes during the E. coli cell cycle. Different loci on the E. coli chromosome were labeled using DNA-binding fluorescent proteins and imaged using an optical setup with an astigmatic point-spread-function. Mother-machine devices were used to constrain the movement of cells to the lateral plane during growth. A deep-learning-based single-molecule localization method was adapted for this application and used to map the chromosomal loci’s physical position in 3D as a function of cell size during the E. coli cell cycle.

Antimicrobial resistance is an increasing problem on a global scale. Rapid antibiotic susceptibility testing (AST) is urgently needed in the clinic to enable personalized prescriptions in high-resistance environments and to limit the use of broad-spectrum drugs. Current rapid phenotypic AST methods do not include species identification (ID), leaving time-consuming plating or culturing as the only available option when ID is needed to make the sensitivity call. Here we describe a method to perform phenotypic AST at the single-cell level in a microfluidic chip that allows subsequent genotyping by in situ FISH. By stratifying the phenotypic AST response on the species of individual cells, it is possible to determine the susceptibility profile for each species in a mixed sample in 2 h. In this proof-of-principle study, we demonstrate the operation with four antibiotics and mixed samples with combinations of seven species.

Transposon mutagenesis is a powerful method to create deep libraries of genetically diverse cells. It has previously not been possible to analyze transposon libraries with respect to complex phenotypes. Here, we use optical pooled screening to characterize a transposon library using high-resolution time-lapse imaging, which is analyzed in real time such that we can use an optical tweezer to isolate cells with interesting phenotypes. We used the method to identify mutants with perturbations in replication initiation control in Escherichia coli, but it can be used to identify genetic elements connected to any type of complex or dynamic single-cell phenotype.