Professor Achillefs Kapanidis

Abstract:

Abstract The understanding of cellular mechanisms benefits substantially from accurate determination of protein diffusive properties. Prior work in this field primarily focuses on traditional methods, such as mean square displacements, for calculation of protein diffusion coefficients and biological states. This proves difficult and error-prone for proteins undergoing heterogenous behaviour, particularly in complex environments, limiting the exploration of new biological behaviours. The importance of determining protein diffusion coefficients, anomalous exponents, and biological behaviours led to the Anomalous Diffusion Challenge 2024, exploring machine learning methods to infer these variables in heterogenous trajectories with time-dependent changepoints. In response to the challenge, we present M3, a machine learning method for pointwise inference of diffusive coefficients, anomalous exponents, and states along noisy heterogenous protein trajectories. M3 makes use of long short-term memory cells to achieve small mean absolute errors for the diffusion coefficient and anomalous exponent alongside high state accuracies (>90%). Subsequently, we implement changepoint detection to determine timepoints at which protein behaviour changes. M3 removes the need for expert fine-tuning required in most conventional statistical methods while being computationally inexpensive to train. The model finished in the Top 5 of the Anomalous Diffusive Challenge 2024, with small improvements made since challenge closure.

Abstract:

A main limitation of single-molecule fluorescence (SMF) measurements is the 'high concentration barrier', describing the maximum concentration of fluorescent species tolerable for sufficient signal-to-noise ratio. To address this barrier in several SMF applications, we design fluorogenic probes based on short single-stranded DNAs, fluorescing only upon hybridizing to their complementary target sequence. We engineer the quenching efficiency and fluorescence enhancement upon duplex formation through screening several fluorophore-quencher combinations, label lengths, and sequence motifs, which we utilize as tuning screws to adapt our labels to different experimental designs. Using these fluorogenic probes, we can perform SMF experiments at concentrations of 10 μM fluorescent labels; this concentration is 100-fold higher than the operational limit for standard TIRF experiments. We demonstrate the ease of implementing these probes into existing protocols by performing super-resolution imaging with DNA-PAINT, employing a fluorogenic 6-nt-long imager; through the faster acquisition of binding events, the imaging of viral genome segments could be sped up significantly to achieve extraction of 20-nm structural features with only ∼150 s of imaging. The exceptional tunability of our probe design will overcome concentration barriers in SMF experiments and unlock new possibilities in super-resolution imaging, molecular tracking, and single-molecule fluorescence energy transfer (smFRET).

Abstract:

Efficient DNA repair is essential for maintaining genome integrity and ensuring cell survival. In Escherichia coli, RecBCD plays a crucial role in processing DNA ends, following a DNA double-strand break (DSB), to initiate repair. While RecBCD has been extensively studied in vitro, less is known about how it contributes to rapid and efficient repair in living bacteria. Here, we use single-molecule microscopy to investigate DNA repair in real time in E. coli. We quantify RecB single-molecule mobility and monitor the induction of the DNA damage response (SOS response) in individual cells. We show that RecB binding to DNA ends caused by endogenous processes leads to efficient repair without SOS induction. In contrast, repair is less efficient in the presence of exogenous damage or in a mutant strain with modified RecB activities, leading to high SOS induction. Our findings reveal how subtle alterations in RecB activity profoundly impact the efficiency of DNA repair in E. coli.

Abstract:

Rapid antibiotic susceptibility tests (ASTs) are an increasingly important part of clinical care as antimicrobial resistance (AMR) becomes more common in bacterial infections. Here, we use the spatial distribution of fluorescently labelled ribosomes to detect intracellular changes associated with antibiotic susceptibility in E. coli cells using a convolutional neural network (CNN). By using ribosome-targeting probes, one fluorescence image provides data for cell segmentation and susceptibility phenotyping. Using 60,382 cells from an antibiotic-susceptible laboratory strain of E. coli, we showed that antibiotics with different mechanisms of action result in distinct ribosome phenotypes, which can be identified by a CNN with high accuracy (99%, 98%, 95%, and 99% for ciprofloxacin, gentamicin, chloramphenicol, and carbenicillin). With 6 E. coli strains isolated from bloodstream infections, we used 34,205 images of ribosome phenotypes to train a CNN that could classify susceptible cells with 91% accuracy and resistant cells with 99% accuracy. Such accuracies correspond to the ability to differentiate susceptible and resistant samples with 99% confidence with just 2 cells, meaning that this method could eliminate lengthy culturing steps and could determine susceptibility with 30 min of antibiotic treatment. The ribosome phenotype method should also be able to identify phenotypes in other strains and species.